Compare commits

...

66 commits

Author SHA1 Message Date
Jérémy Lecour 145b279a12
Merge branch 'unstable' into stable
All checks were successful
gitea/ansible-roles/pipeline/head This commit looks good
Ansible Lint |Total|New|Outstanding|Fixed|Trend |:-:|:-:|:-:|:-:|:-: |2710|0|2710|0|:zzz:
gitea/ansible-roles/pipeline/tag This commit looks good
2024-04-30 17:42:11 +02:00
Jérémy Lecour 6ee1e609ac
Release 24.04 2024-04-30 17:41:15 +02:00
Jérémy Lecour 2ad2ae8521
amend CHANGELOG 2024-04-30 17:38:26 +02:00
Jérémy Lecour e3746d18fb
proftpd: remove whitelist block if feature is disabled 2024-04-30 17:38:14 +02:00
Jérémy Lecour 1d5415237c
sort CHANGELOG 2024-04-30 17:16:22 +02:00
Jérémy Lecour e5bbb601d3
evobackup-client: replace non-functional role with install tasks 2024-04-30 17:08:44 +02:00
Mathieu Trossevin 4c2548b21d
fix(certbot): Fix HAPEE renewal hook 2024-04-26 09:29:15 +02:00
Ludovic Poujol 5d11468327
docker-host: lint 2024-04-18 16:10:26 +02:00
Ludovic Poujol a41e78b556
docker-host: Removed setting docker_conf_use_iptables (iptable usage forced to true 2024-04-18 15:38:11 +02:00
Jérémy Lecour 8cd887ee21
minifirewall: fix task name 2024-04-18 15:18:58 +02:00
Jérémy Lecour 42ad242aaf
vrrpd: configure minifirewall with blocks instead of lines 2024-04-18 15:18:42 +02:00
Jérémy Lecour f8e92d2eeb
haproxy: support bookworm for backport packages 2024-04-18 15:11:34 +02:00
Jérémy Lecour 9a65312190
evolinux-base: disable logcheck monitoring of journald only if journald.logfiles exists 2024-04-18 15:10:01 +02:00
Jérémy Dubois 16394060c9 openvpn: let tasks using openbsd_pkg but commented 2024-04-18 11:11:03 +02:00
Jérémy Dubois f4e6aabe8a openvpn: install packages manually
Because openbsd_pkg module is broken since OpenBSD 7.4 with the version of Ansible we currently use
2024-04-18 11:04:08 +02:00
William Hirigoyen 0ec343766d postfix/amavis: max servers is now 3 (previously 2) 2024-04-17 17:09:12 +02:00
William Hirigoyen d1410e38a1 evolinux-base/logcheck: fix conf patch, journal check was not disabled when asked 2024-04-16 10:25:13 +02:00
mgauthier 7272106bce Merge pull request 'Add option to return the amp values' (#177) from munin_plugin_ipmi_with_amp into unstable
Reviewed-on: #177
2024-04-12 17:04:50 +02:00
Mathieu Gauthier-Pilote d11cf4987b Add option to return the amp values 2024-04-12 17:03:36 +02:00
Jérémy Lecour 2a264dd2bc
redis: replace inline argument with environment variable for the password 2024-04-12 15:54:20 +02:00
Brice Waegeneire 5708e7205d nrpe: !disk1 exclude filesystem type overlay 2024-04-11 15:48:37 +02:00
Eric Morino 4bbe2f4f72 Delete 'state' option on template for PGDG repo 2024-04-09 09:12:01 +02:00
Jérémy Lecour 7a9be8d6fa
vrrpd : configure and restart minifirewall before starting VRRP 2024-04-04 18:40:39 +02:00
William Hirigoyen 5acb1956f5 packweb: fix old bug (2017!) .orig file created by module patch and taken in account by ProFTPd 2024-03-27 15:17:48 +01:00
Alexis Ben Miloud--Josselin 96504b1deb evolinux-users: Add sudo mvcli for nagios user 2024-03-27 12:13:49 +01:00
William Hirigoyen 56eef89084 nagios-nrpe: create /etc/bash_completion.d if missing 2024-03-22 11:09:06 +01:00
William Hirigoyen b30b7c884a fail2ban: SQLite purge script didn't vacuum as expected + error when vacuum cannot be done 2024-03-19 18:09:56 +01:00
David Prevot fe66ad9c4f Changelog entry for roundcube log change 2024-03-19 17:44:12 +01:00
David Prevot 96c1017b5d roundcube: Use /var/log/roundcube directly
/home/roundcube/log does not even exists…
2024-03-19 17:07:43 +01:00
David Prevot 23f4f9690f roundcube: Use /var/log/roundcube directly
/var/lib/roundcube/logs is already a symlink to /var/log/roundcube.
2024-03-19 16:54:14 +01:00
Jérémy Lecour d758afdd4b
autosysadmin-agent: upstream release 24.03.2 2024-03-19 08:19:31 +01:00
Ludovic Poujol ae2e447bc4
evolinux-base: Add new variable to disable global customisation of bash config 2024-03-18 15:30:23 +01:00
Eric Morino 8e3724d5e7 Add documentation for add account proftpd with ips whitelist 2024-03-18 09:50:43 +01:00
Eric Morino ff19df2444 Add proftpd_sftp_enable_user_whitelist : False by default 2024-03-18 09:24:13 +01:00
Eric Morino eda69725d5 proftpd: add whitelist ip in virtualhost sftp 2024-03-15 09:20:32 +01:00
Alexis Ben Miloud--Josselin bc9609ce48 nextcloud: use latest version by default 2024-03-14 10:16:40 +01:00
Ludovic Poujol aca146adbc
memcached: Fix conditions not properly writen (installation was always in multi-instance mode) 2024-03-11 10:54:36 +01:00
Gregory Colpart ccff3b2105 Suppression de fichiers désormais inutilisés (ils sont vides) 2024-03-10 12:38:13 +01:00
Tom David--Broglio 2fe0d25277 correction tasks keepalived 2024-03-07 15:47:39 +01:00
Jérémy Lecour 501f5e7577
autosysadmin-agent: upstream release 24.03.1 2024-03-05 16:54:10 +01:00
Alexis Ben Miloud--Josselin 8b68f03910 certbot: Fix HAProxy renewal hook 2024-03-04 15:00:01 +01:00
Jérémy Lecour 78326e43e8
wording in the CHANGELOG 2024-03-01 09:09:52 +01:00
Jérémy Lecour 2a856d579e
Merge branch 'unstable' into stable
All checks were successful
gitea/ansible-roles/pipeline/head This commit looks good
Ansible Lint |Total|New|Outstanding|Fixed|Trend |:-:|:-:|:-:|:-:|:-: |2712|0|2712|0|:zzz:
gitea/ansible-roles/pipeline/tag This commit looks good
2024-03-01 09:06:08 +01:00
Jérémy Lecour beea53aa64
Merge branch 'stable' into unstable 2024-03-01 09:04:25 +01:00
Jérémy Lecour 342380876a
Release 24.03 2024-03-01 09:00:49 +01:00
Jérémy Lecour 24cbbf2f54
fix CHANGELOG 2024-03-01 08:55:07 +01:00
Jérémy Lecour 56237bb3c6
evolinux-base: execute autosysadmin-agent and autosysadmin-restart_nrpe roles
All checks were successful
Ansible Lint |Total|New|Outstanding|Fixed|Trend |:-:|:-:|:-:|:-:|:-: |2711|9|2702|6|:-1: Reference build: <a href="https://jenkins.evolix.org/job/gitea/job/ansible-roles/job/unstable/13//ansiblelint">Evolix » ansible-roles » unstable #13</a>
gitea/ansible-roles/pipeline/head This commit looks good
2024-03-01 08:35:16 +01:00
Jérémy Lecour abd329b9c1
autosysadmin-restart_nrpe: add role 2024-03-01 08:32:47 +01:00
Jérémy Lecour 037ec9d376
autosysadmin-agent: upstream release 24.03 2024-03-01 08:26:43 +01:00
Jérémy Lecour c333970606
autosysadmin-agent: rename /usr/share/scripts/autosysadmin/{auto,restart} 2024-02-29 19:16:18 +01:00
Jérémy Lecour 10b507adc4
autosysadmin-agent: logs clearing is done weekly
All checks were successful
Ansible Lint |Total|New|Outstanding|Fixed|Trend |:-:|:-:|:-:|:-:|:-: |2708|6|2702|4|:-1: Reference build: <a href="https://jenkins.evolix.org/job/gitea/job/ansible-roles/job/unstable/12//ansiblelint">Evolix » ansible-roles » unstable #12</a>
gitea/ansible-roles/pipeline/head This commit looks good
2024-02-29 18:50:14 +01:00
Jérémy Lecour b2e22413bc
autosysadmin-agent: upstream release 24.02.3
All checks were successful
Ansible Lint |Total|New|Outstanding|Fixed|Trend |:-:|:-:|:-:|:-:|:-: |2706|22|2684|5|:-1: Reference build: <a href="https://jenkins.evolix.org/job/gitea/job/ansible-roles/job/unstable/11//ansiblelint">Evolix » ansible-roles » unstable #11</a>
gitea/ansible-roles/pipeline/head This commit looks good
2024-02-28 15:40:39 +01:00
William Hirigoyen bec868009c nagios: add option --full to check pressure IO and mem to avoid flaps
All checks were successful
Ansible Lint |Total|New|Outstanding|Fixed|Trend |:-:|:-:|:-:|:-:|:-: |2689|6|2683|6|:-1: Reference build: <a href="https://jenkins.evolix.org/job/gitea/job/ansible-roles/job/unstable/10//ansiblelint">Evolix » ansible-roles » unstable #10</a>
gitea/ansible-roles/pipeline/head This commit looks good
2024-02-27 10:33:49 +01:00
David Prevot aea710cb25 redis: Update munin plugin
All checks were successful
Ansible Lint |Total|New|Outstanding|Fixed|Trend |:-:|:-:|:-:|:-:|:-: |2689|4|2685|3|:-1: Reference build: <a href="https://jenkins.evolix.org/job/gitea/job/ansible-roles/job/unstable/9//ansiblelint">Evolix » ansible-roles » unstable #9</a>
gitea/ansible-roles/pipeline/head This commit looks good
2024-02-22 09:44:50 +01:00
Ludovic Poujol b0ba70f06c
certbot: Renewal hook for NRPE
All checks were successful
Ansible Lint |Total|New|Outstanding|Fixed|Trend |:-:|:-:|:-:|:-:|:-: |2688|4|2684|6|:+1: Reference build: <a href="https://jenkins.evolix.org/job/gitea/job/ansible-roles/job/unstable/8//ansiblelint">Evolix » ansible-roles » unstable #8</a>
gitea/ansible-roles/pipeline/head This commit looks good
2024-02-21 12:27:18 +01:00
Jérémy Dubois 0a4a220bdf openvpn: earlier alert for CA expiration
All checks were successful
Ansible Lint |Total|New|Outstanding|Fixed|Trend |:-:|:-:|:-:|:-:|:-: |2690|4|2686|3|:-1: Reference build: <a href="https://jenkins.evolix.org/job/gitea/job/ansible-roles/job/unstable/7//ansiblelint">Evolix » ansible-roles » unstable #7</a>
gitea/ansible-roles/pipeline/head This commit looks good
2024-02-21 10:51:08 +01:00
Jérémy Lecour 282dcb28f4
apt: add comments to deb822 migration scripts
All checks were successful
Ansible Lint |Total|New|Outstanding|Fixed|Trend |:-:|:-:|:-:|:-:|:-: |2689|4|2685|3|:-1: Reference build: <a href="https://jenkins.evolix.org/job/gitea/job/ansible-roles/job/unstable/6//ansiblelint">Evolix » ansible-roles » unstable #6</a>
gitea/ansible-roles/pipeline/head This commit looks good
2024-02-20 18:50:39 +01:00
Alexis Ben Miloud--Josselin a0fc763a0c certbot: Utiliser pkey pour tester clé
All checks were successful
Ansible Lint |Total|New|Outstanding|Fixed|Trend |:-:|:-:|:-:|:-:|:-: |2688|3|2685|3|:-1: Reference build: <a href="https://jenkins.evolix.org/job/gitea/job/ansible-roles/job/unstable/5//ansiblelint">Evolix » ansible-roles » unstable #5</a>
gitea/ansible-roles/pipeline/head This commit looks good
2024-02-20 16:12:26 +01:00
David Prevot a56e8c27ee lxc-php, php: Update sury PGP key
All checks were successful
Ansible Lint |Total|New|Outstanding|Fixed|Trend |:-:|:-:|:-:|:-:|:-: |2688|4|2684|5|:+1: Reference build: <a href="https://jenkins.evolix.org/job/gitea/job/ansible-roles/job/unstable/4//ansiblelint">Evolix » ansible-roles » unstable #4</a>
gitea/ansible-roles/pipeline/head This commit looks good
2024-02-20 13:57:17 +01:00
Jérémy Lecour 56db6e1fbc
apt: add ftp.evolix.org as recognized system source
All checks were successful
Ansible Lint |Total|New|Outstanding|Fixed|Trend |:-:|:-:|:-:|:-:|:-: |2689|0|2689|0|:zzz:
gitea/ansible-roles/pipeline/head This commit looks good
2024-02-20 09:49:41 +01:00
Jérémy Lecour 015cac688e
redis: create sysfs config file if missing 2024-02-20 09:48:58 +01:00
Jérémy Lecour c12c581f63
update CHANGELOG
Some checks failed
gitea/ansible-roles/pipeline/head There was a failure building this commit
2024-02-12 19:07:20 +01:00
Alexis Ben Miloud--Josselin 7c2fd5e394
kvm-host: Add firewall rule for DRBD
Some checks failed
gitea/ansible-roles/pipeline/head There was a failure building this commit
2024-02-12 18:52:29 +01:00
Jérémy Lecour 9402458304
Merge branch 'unstable' into stable
Some checks failed
gitea/ansible-roles/pipeline/head There was a failure building this commit
gitea/ansible-roles/pipeline/tag There was a failure building this commit
2024-02-08 11:08:47 +01:00
Jérémy Lecour cf0fab1e22
Release 24.02.1 2024-02-08 11:08:28 +01:00
Jérémy Lecour 13284645de
fail2ban: fix Ansible syntax
Some checks failed
gitea/ansible-roles/pipeline/head There was a failure building this commit
2024-02-08 11:03:14 +01:00
131 changed files with 7319 additions and 2565 deletions

View file

@ -6,8 +6,8 @@ The format is based on [Keep a Changelog](http://keepachangelog.com/en/1.0.0/).
This project does not follow semantic versioning.
The **major** part of the version is the year
The **minor** part changes is the month
The **patch** part changes is incremented if multiple releases happen the same month
The **minor** part is the month
The **patch** part is incremented if multiple releases happen the same month
## [Unreleased]
@ -21,6 +21,76 @@ The **patch** part changes is incremented if multiple releases happen the same m
### Security
## [24.04] 2024-04-30
### Added
proftpd: optional configuration of IP whitelists per groups of users
### Changed
* autosysadmin-agent: upstream release 24.03.2
* evobackup-client: replace non-functional role with install tasks
* evobackup-client: upstream release 24.04.1
* evolinux-base: Add new variable to disable global customisation of bash config
* evolinux-base: Disable logcheck monitoring of journald only if journald.logfiles exists
* evolinux-users: Add sudo mvcli for nagios user
* haproxy: support bookworm for backport packages
* nrpe: !disk1 exclude filesystem type overlay
* postfix/amavis: max servers is now 3 (previously 2)
* roundcube: Use /var/log/roundcube directly
* vrrpd: configure and restart minifirewall before starting VRRP
* vrrpd: configure minifirewall with blocks instead of lines
### Fixed
* certbot: Fix HAPEE renewal hook
* certbot: Fix HAProxy renewal hook
* evolinux-base/logcheck: fix conf patch, journal check was not disabled when asked
* fail2ban: SQLite purge script didn't vacuum as expected + error when vacuum cannot be done
* keepalived: Fix tasks that use file instead of copy
* memcached: Fix conditions not properly writen (installation was always in multi-instance mode)
* nagios-nrpe: create /etc/bash_completion.d if missing
* openvpn: install packages manually, because openbsd_pkg module is broken since OpenBSD 7.4 with the version of Ansible we currently use
* packweb: fix old bug (2017!) .orig file created by module patch and taken in account by ProFTPd
* redis: replace inline argument with environment variable for the password
### Removed
* docker-host: Removed `docker_conf_use_iptables` variable (iptable usage forced to true)
## [24.03] 2024-03-01
### Added
* autosysadmin-agent: upstream release 24.03
* autosysadmin-restart_nrpe: add role
* certbot: Renewal hook for NRPE
* kvm-host: add minifirewall rules if DRBD interface is configured
* proftpd: add whitelist ip
### Changed
* apt: add ftp.evolix.org as recognized system source
* autosysadmin-agent: logs clearing is done weekly
* autosysadmin-agent: rename /usr/share/scripts/autosysadmin/{auto,restart}
* certbot: use pkey to test the key
* evolinux-base: execute autosysadmin-agent and autosysadmin-restart_nrpe roles
* lxc-php, php: Update sury PGP key
* openvpn: earlier alert for CA expiration
* redis: create sysfs config file if missing
* nextcloud: use latest version by default
### Removed
* autosysadmin: replaced by autosysadmin-agent
## [24.02.1] 2024-02-08
### Fixed
* fail2ban: fix Ansible syntax
## [24.02] 2024-02-08
### Added
@ -71,6 +141,7 @@ The **patch** part changes is incremented if multiple releases happen the same m
* nagios: add dockerd check in nrpe check template
* nagios: cleaning nrpe check template
* nagios: rename var `nagios_nrpe_process_processes` into `nagios_nrpe_processes` and check systemd-timesyncd instead of ntpd in Debian 12
* nagios: add option --full to check pressure IO and mem to avoid flaps
* proftpd: in SFTP vhost, enable SSH keys login, enable ed25549 host key for Debian >= 11
* redis: manage config template inside a block, to allow custom modifications outside
* spamassassin: Use spamd starting with Bookworm

View file

@ -39,7 +39,7 @@ $sa_spam_subject_tag = '[SPAM]';
$log_level = 2;
# En fonction besoin/ressources, on a juste le nbre de process
$max_servers = 2;
$max_servers = 3;
$enable_ldap = 1;
$default_ldap = {

View file

@ -1,5 +1,11 @@
#!/usr/bin/env python3
##########
# This script takes a multi-lines input of "oneliner-style" APT sources definitions.
# It converts them into "deb822-style" sources.
# Each generated file will have only one stanza, possibly with multiple Types/Suites/Components
##########
import re
import sys
import os
@ -10,11 +16,16 @@ import apt_pkg
# Order matters !
destinations = {
"debian-security": "security.sources",
".*-backports": "backports.sources",
".debian.org": "system.sources",
"mirror.evolix.org": "system.sources",
"ftp.evolix.org": "system.sources",
"pub.evolix.net": "evolix_public_old.sources.bak",
"pub.evolix.org": "evolix_public.sources",
"artifacts.elastic.co": "elastic.sources",
"download.docker.com": "docker.sources",
"downloads.linux.hpe.com": "hp.sources",
@ -76,6 +87,11 @@ def prepare_sources(lines):
key, value = option.split("=")
options[key] = value
### WARNING ###
# if there are multiple lines with different URIS for a given destination (eg. "system")
# each one will overwrite the previous one
# and the last evaluated will be what remains.
if dest in sources:
sources[dest]["Types"].add(matches["type"])
sources[dest]["URIs"] = matches["uri"]

View file

@ -1,5 +1,11 @@
#!/bin/sh
##########
# This script changes all "one-line" APT sources into "deb822" sources.
# It is responsible for searching and processing the files.
# The actual format migration is done by a python script.
##########
deb822_migrate_script=$(command -v deb822-migration.py)
if [ -z "${deb822_migrate_script}" ]; then
@ -46,4 +52,4 @@ for file in $(find /etc/apt/sources.list.d -mindepth 1 -maxdepth 1 -type f -name
done
echo "${count} file(s) migrated"
exit ${rc}
exit ${rc}

View file

@ -0,0 +1,17 @@
---
general_scripts_dir: "/usr/share/scripts"
autosysadmin_agent_bin_dir: "/usr/local/bin/autosysadmin"
autosysadmin_agent_lib_dir: "/usr/local/lib/autosysadmin"
autosysadmin_agent_auto_dir: "{{ general_scripts_dir }}/autosysadmin/restart"
autosysadmin_agent_crontab_enabled: true
autosysadmin_agent_log_retention_days: 365
autosysadmin_config: []
### All repair are disabled if set to 'off'
### even if a specific repair value is 'on'
# repair_all: 'on'
### Default values for checks
# repair_foo: 'off'

View file

@ -0,0 +1,25 @@
#!/bin/bash
days=${1:-365}
log_dir="/var/log/autosysadmin/"
if [ -d "${log_dir}" ]; then
find_run_dirs() {
find "${log_dir}" \
-mindepth 1 \
-maxdepth 1 \
-type d \
-ctime "+${days}" \
-print0
}
log() {
/usr/bin/logger -p local0.notice -t autosysadmin "${1}"
}
while IFS= read -r -d '' run_dir; do
rm --recursive --force "${run_dir}"
log "Delete ${run_dir} (older than ${days} days)"
done < <(find_run_dirs)
fi
exit 0

View file

@ -0,0 +1,907 @@
#!/bin/bash
VERSION="24.03.2"
# Common functions for "repair" and "restart" scripts
set -u
# Initializes the program, context, configuration…
initialize() {
PATH="${PATH}":/usr/sbin:/sbin
# Used in many places to refer to the program name.
# Examples: repair_mysql, restart_nrpe…
PROGNAME=$(basename "${0}")
# find out if running in interactive mode, or not
if [ -t 0 ]; then
INTERACTIVE=1
else
INTERACTIVE=0
fi
readonly INTERACTIVE
# Default empty value for Debug mode
DEBUG="${DEBUG:-""}"
# Repair scripts obey to the value of a variable named after the script
# You can set the value ("on" or "off") in /etc/evolinux/autosysadmin
# Here we set the default value to "on".
declare -g "${PROGNAME}"=on # dynamic variable assignment ($PROGNAME == repair_*)
PID=$$
readonly PID
# Each execution (run) gets a unique ID
RUN_ID="$(date +"%Y-%m-%d_%H-%M")_${PROGNAME}_${PID}"
readonly RUN_ID
# Main log directory
MAIN_LOG_DIR="/var/log/autosysadmin"
readonly MAIN_LOG_DIR
# shellcheck disable=SC2174
mkdir --mode=750 --parents "${MAIN_LOG_DIR}"
chgrp adm "${MAIN_LOG_DIR}"
# Each execution store some information
# in a unique directory based on the RUN_ID
RUN_LOG_DIR="${MAIN_LOG_DIR}/${RUN_ID}"
readonly RUN_LOG_DIR
# shellcheck disable=SC2174
mkdir --mode=750 --parents "${RUN_LOG_DIR}"
chgrp adm "${RUN_LOG_DIR}"
# This log file contains all events
RUN_LOG_FILE="${RUN_LOG_DIR}/autosysadmin.log"
readonly RUN_LOG_FILE
# This log file contains notable actions
ACTIONS_FILE="${RUN_LOG_DIR}/actions.log"
readonly ACTIONS_FILE
touch "${ACTIONS_FILE}"
# This log file contains abort reasons (if any)
ABORT_FILE="${RUN_LOG_DIR}/abort.log"
readonly ABORT_FILE
# touch "${ABORT_FILE}"
# Date format for log messages
DATE_FORMAT="%Y-%m-%d %H:%M:%S"
# This will contain lock, last-run markers…
# It's ok to lose the content after a reboot
RUN_DIR="/run/autosysadmin"
readonly RUN_DIR
mkdir -p "${RUN_DIR}"
# Only a singe instace of each script can run simultaneously
# We use a customizable lock name for this.
# By default it's the script's name
LOCK_NAME=${LOCK_NAME:-${PROGNAME}}
# If a lock is found, we can wait for it to disappear.
# The value must be understood by sleep(1)
LOCK_WAIT="0"
# Default values for email headers
EMAIL_FROM="equipe+autosysadmin@evolix.net"
EMAIL_INTERNAL="autosysadmin@evolix.fr"
LOCK_FILE="${RUN_DIR}/${LOCK_NAME}.lock"
readonly LOCK_FILE
# Remove lock file at exit
cleanup() {
# shellcheck disable=SC2317
rm -f "${LOCK_FILE}"
}
trap 'cleanup' 0
# Load configuration
# shellcheck disable=SC1091
test -f /etc/evolinux/autosysadmin && source /etc/evolinux/autosysadmin
log_all "Begin ${PROGNAME} RUN_ID: ${RUN_ID}"
log_all "Log directory is ${RUN_LOG_DIR}"
}
# Executes a list of tasks before exiting:
# * prepare a summary of actions and possible abort reasons
# * send emails
# * do some cleanup
quit() {
log_all "End ${PROGNAME} RUN_ID: ${RUN_ID}"
summary="RUN_ID: ${RUN_ID}"
if [ -s "${ABORT_FILE}" ]; then
# Add abort reasons to summary
summary="${summary}\n$(print_abort_reasons)"
hook_mail "abort"
return_code=1
else
if [ -s "${ACTIONS_FILE}" ]; then
# Add notable actions to summary
summary="${summary}\n$(print_actions "Aucune action")"
hook_mail "success"
fi
return_code=0
fi
hook_mail "internal"
if is_interactive; then
# shellcheck disable=SC2001
echo "${summary}" | sed -e 's/\\n/\n/g'
else
/usr/share/scripts/evomaintenance.sh --auto --user autosysadmin --message "${summary}" --no-commit --no-mail
fi
teardown
# shellcheck disable=SC2086
exit ${return_code}
}
teardown() {
:
}
# Return true/false
is_interactive() {
test "${INTERACTIVE}" -eq "1"
}
save_server_state() {
DUMP_SERVER_STATE_BIN="$(command -v dump-server-state || command -v backup-server-state)"
if [ -z "${DUMP_SERVER_STATE_BIN}" ]; then
log_all "Warning: dump-server-state is not present. No server state recorded."
fi
if [ -x "${DUMP_SERVER_STATE_BIN}" ]; then
DUMP_DIR=$(file_path_in_log_dir "server-state")
# We don't want the logging to take too much time,
# so we kill it if it takes more than 20 seconds.
timeout --signal 9 20 \
"${DUMP_SERVER_STATE_BIN}" \
--dump-dir="${DUMP_DIR}" \
--df \
--dmesg \
--iptables \
--lxc \
--netcfg \
--netstat \
--uname \
--processes \
--systemctl \
--uptime \
--virsh \
--disks \
--mysql-processes \
--no-apt-states \
--no-apt-config \
--no-dpkg-full \
--no-dpkg-status \
--no-mount \
--no-packages \
--no-sysctl \
--no-etc
log_run "Server state saved in \`server-state' directory."
fi
}
is_debug() {
# first time: do the check…
# other times: pass
if [ -z "${DEBUG:-""}" ]; then
debug_file="/etc/evolinux/autosysadmin.debug"
if [ -e "${debug_file}" ]; then
last_change=$(stat -c %Z "${debug_file}")
limit_date=$(date --date "14400 seconds ago" +"%s")
if [ $(( last_change - limit_date )) -le "0" ]; then
log_run "Debug mode disabled; file is too old (%{last_change} seconds)."
rm "${debug_file}"
# Debug mode disabled
DEBUG="0"
else
log_run "Debug mode enabled."
# Debug mode enabled
DEBUG="1"
fi
else
# log_run "Debug mode disabled; file is absent."
# Debug mode disabled
DEBUG="0"
fi
fi
# return the value
test "${DEBUG}" -eq "1"
}
# Uses the who(1) definition of "active"
currently_active_users() {
LC_ALL=C who --users | grep --extended-regexp "\s+\.\s+" | awk '{print $1}' | sort --human-numeric-sort | uniq
}
# Users active in the last 29 minutes
recently_active_users() {
LC_ALL=C who --users | grep --extended-regexp "\s+00:(0|1|2)[0-9]\s+" | awk --field-separator ' ' '{print $1,$6}'
}
# Save the list of users to a file in the log directory
save_active_users() {
LC_ALL=C who --users | save_in_log_dir "who-users"
}
# An autosysadmin must not perform actions if a user is active or was active recently.
#
# This can by bypassed in interactive mode.
# It's OK to lose this data after a reboot.
ensure_no_active_users_or_exit() {
# Save all active users
save_active_users
if is_debug; then
log_run "Debug mode enabled: continue without checking active users."
return 0;
fi
# Is there any currently active user?
currently_active_users=$(currently_active_users)
if [ -n "${currently_active_users}" ]; then
# shellcheck disable=SC2001
users_oneliner=$(echo "${currently_active_users}" | sed -e 's/\n/ /')
log_run "Currently active users: ${users_oneliner}"
if is_interactive; then
echo "Some users are currently active:"
# shellcheck disable=SC2001
echo "${currently_active_users}" | sed -e 's/\(.\+\)/* \1/'
answer=""
while :; do
printf "> Continue? [Y,n,?] "
read -r answer
case ${answer} in
[Yy]|"" )
log_run "Active users check bypassed manually in interactive mode."
return
;;
[Nn] )
log_run "Active users check confirmed manually in interactive mode."
log_abort_and_quit "Active users detected: ${users_oneliner}"
;;
* )
printf "y - yes, continue\n"
printf "n - no, exit\n"
printf "? - print this help\n"
;;
esac
done
else
log_abort_and_quit "Currently active users detected: ${users_oneliner}."
fi
else
# or recently (the last 30 minutes) active user?
recently_active_users=$(recently_active_users)
if [ -n "${recently_active_users}" ]; then
# shellcheck disable=SC2001
users_oneliner=$(echo "${recently_active_users}" | sed -e 's/\n/ /')
log_run "Recently active users: ${users_oneliner}"
if is_interactive; then
echo "Some users were recently active:"
# shellcheck disable=SC2001
echo "${recently_active_users}" | sed -e 's/\(.\+\)/* \1/'
answer=""
while :; do
printf "> Continue? [Y,n,?] "
read -r answer
case ${answer} in
[Yy]|"" )
log_run "Active users check bypassed manually in interactive mode."
return
;;
[Nn] )
log_run "Active users check confirmed manually in interactive mode."
log_abort_and_quit "Recently active users detected: ${users_oneliner}."
;;
* )
printf "y - yes, continue\n"
printf "n - no, exit\n"
printf "? - print this help\n"
;;
esac
done
else
log_abort_and_quit "Recently active users detected: ${users_oneliner}."
fi
fi
fi
}
# Takes an NRPE command name as 1st parameter,
# and executes the full command if found in the configuration.
# Return the result and the return code of the command.
check_nrpe() {
check="$1"
nrpe_files=""
# Check if NRPE config is found
if [ -f "/etc/nagios/nrpe.cfg" ]; then
nrpe_files="${nrpe_files} /etc/nagios/nrpe.cfg"
else
msg="NRPE configuration not found: /etc/nagios/nrpe.cfg"
log_run "${msg}"
echo "${msg}"
return 3
fi
# Search for included files
# shellcheck disable=SC2086
while IFS= read -r include_file; do
nrpe_files="${nrpe_files} ${include_file}"
done < <(grep --extended-regexp '^\s*include=.+' ${nrpe_files} | cut -d = -f 2)
# Search for files in included directories
# shellcheck disable=SC2086
while IFS= read -r include_dir; do
nrpe_files="${nrpe_files} ${include_dir}/*.cfg"
done < <(grep --extended-regexp '^\s*include_dir=.+' ${nrpe_files} | cut -d = -f 2)
# Fetch uncommented commands in (sorted) config files
# shellcheck disable=SC2086
nrpe_commands=$(grep --no-filename --exclude=*~ --fixed-strings "[${check}]" ${nrpe_files} | grep --invert-match --extended-regexp '^\s*#\s*command' | cut -d = -f 2)
nrpe_commands_count=$(echo "${nrpe_commands}" | wc -l)
if is_debian_version "9" "<=" && [ "${nrpe_commands_count}" -gt "1" ]; then
# On Debian <= 9, NRPE loading was not sorted
# we need to raise an error if we have multiple defined commands
msg="Unable to determine which NRPE command to run"
log_run "${msg}"
echo "${msg}"
return 3
else
# On Debian > 9, use the last command
nrpe_command=$(echo "${nrpe_commands}" | tail -n 1)
nrpe_result=$(${nrpe_command})
nrpe_rc=$?
log_run "NRPE command (exited with ${nrpe_rc}): ${nrpe_command}"
log_run "${nrpe_result}"
echo "${nrpe_result}"
return "${nrpe_rc}"
fi
}
# An autosysadmin script must not run twice (or more) simultaneously.
# We use a customizable (with LOCK_NAME) lock file to keep track of this.
# A wait time can be configured.
#
# This can by bypassed in interactive mode.
# It's OK to lose this data after a reboot.
acquire_lock_or_exit() {
lock_file="${1:-${LOCK_FILE}}"
lock_wait="${2:-${LOCK_WAIT}}"
# lock_wait must be compatible with sleep(1), otherwise fallback to 0
if ! echo "${lock_wait}" | grep -Eq '^[0-9]+[smhd]?$'; then
log_run "Lock wait: incorrect value '${lock_wait}', fallback to 0."
lock_wait=0
fi
if [ "${lock_wait}" != "0" ] && [ -f "${lock_file}" ]; then
log_run "Lock file present. Let's wait ${lock_wait} and check again."
sleep "${lock_wait}"
fi
if [ -f "${lock_file}" ]; then
log_abort_and_quit "Lock file still present."
else
log_run "Lock file absent. Let's put one."
touch "${lock_file}"
fi
}
# If a script has been run in the ast 30 minutes, running it again won't fix the issue.
# We use a /run/ausosysadmin/${PROGNAME}_lastrun file to keep track of this.
#
# This can by bypassed in interactive mode.
# This is bypassed in debug mode.
# It's OK to lose this data after a reboot.
ensure_not_too_soon_or_exit() {
if is_debug; then
log_run "Debug mode enabled: continue without checking when was the last run."
return 0;
fi
lastrun_file="${RUN_DIR}/${PROGNAME}_lastrun"
if [ -f "${lastrun_file}" ]; then
lastrun_age="$(($(date +%s)-$(stat -c "%Y" "${lastrun_file}")))"
log_run "Last run was ${lastrun_age} seconds ago."
if [ "${lastrun_age}" -lt 1800 ]; then
if is_interactive; then
echo "${PROGNAME} was run ${lastrun_age} seconds ago."
answer=""
while :; do
printf "> Continue? [Y,n,?] "
read -r answer
case ${answer} in
[Yy]|"" )
log_run "Last run check bypassed manually in interactive mode."
break
;;
[Nn] )
log_run "Last run check confirmed manually in interactive mode."
log_abort_and_quit 'Last run too recent.'
;;
* )
printf "y - yes, continue\n"
printf "n - no, exit\n"
printf "? - print this help\n"
;;
esac
done
else
log_abort_and_quit "Last run too recent."
fi
fi
fi
touch "${lastrun_file}"
}
# Populate DEBIAN_VERSION and DEBIAN_RELEASE variables
# based on gathered information about the operating system
detect_os() {
DEBIAN_RELEASE="unknown"
DEBIAN_VERSION="unknown"
LSB_RELEASE_BIN="$(command -v lsb_release)"
if [ -e /etc/debian_version ]; then
DEBIAN_VERSION="$(cut -d "." -f 1 < /etc/debian_version)"
if [ -x "${LSB_RELEASE_BIN}" ]; then
DEBIAN_RELEASE="$("${LSB_RELEASE_BIN}" --codename --short)"
else
case "${DEBIAN_VERSION}" in
7) DEBIAN_RELEASE="wheezy" ;;
8) DEBIAN_RELEASE="jessie" ;;
9) DEBIAN_RELEASE="stretch" ;;
10) DEBIAN_RELEASE="buster" ;;
11) DEBIAN_RELEASE="bullseye" ;;
12) DEBIAN_RELEASE="bookworm" ;;
13) DEBIAN_RELEASE="trixie" ;;
esac
fi
# log_run "Detected OS: Debian version=${DEBIAN_VERSION} release=${DEBIAN_RELEASE}"
# else
# log_run "Detected OS: unknown (missing /etc/debian_version)"
fi
}
is_debian_wheezy() {
test "${DEBIAN_RELEASE}" = "wheezy"
}
is_debian_jessie() {
test "${DEBIAN_RELEASE}" = "jessie"
}
is_debian_stretch() {
test "${DEBIAN_RELEASE}" = "stretch"
}
is_debian_buster() {
test "${DEBIAN_RELEASE}" = "buster"
}
is_debian_bullseye() {
test "${DEBIAN_RELEASE}" = "bullseye"
}
is_debian_bookworm() {
test "${DEBIAN_RELEASE}" = "bookworm"
}
is_debian_trixie() {
test "${DEBIAN_RELEASE}" = "trixie"
}
is_debian_version() {
local version=$1
local relation=${2:-"eq"}
if [ -z "${DEBIAN_VERSION:-""}" ]; then
detect_os
fi
dpkg --compare-versions "${DEBIAN_VERSION}" "${relation}" "${version}"
}
# List systemd services (only names), even if stopped
systemd_list_services() {
pattern=$1
systemctl list-units --all --no-legend --type=service "${pattern}" | grep --only-matching --extended-regexp '\S+\.service'
}
is_systemd_enabled() {
systemctl --quiet is-enabled "$1" 2> /dev/null
}
is_systemd_active() {
systemctl --quiet is-active "$1" 2> /dev/null
}
is_sysvinit_enabled() {
find /etc/rc2.d/ -name "$1" > /dev/null
}
get_fqdn() {
# shellcheck disable=SC2155
local system=$(uname -s)
if [ "${system}" = "Linux" ]; then
hostname --fqdn
elif [ "${system}" = "OpenBSD" ]; then
hostname
else
log_abort_and_quit "System '${system}' not recognized."
fi
}
get_complete_hostname() {
REAL_HOSTNAME="$(get_fqdn)"
if [ "${HOSTNAME}" = "${REAL_HOSTNAME}" ]; then
echo "${HOSTNAME}"
else
echo "${HOSTNAME} (${REAL_HOSTNAME})"
fi
}
# Fetch values from evomaintenance configuration
get_evomaintenance_mail() {
grep "EVOMAINTMAIL=" /etc/evomaintenance.cf | cut -d '=' -f2
}
get_evomaintenance_emergency_mail() {
grep "URGENCYFROM=" /etc/evomaintenance.cf | cut -d '=' -f2
}
get_evomaintenance_emergency_tel() {
grep "URGENCYTEL=" /etc/evomaintenance.cf | cut -d '=' -f2
}
# Log a message to the log file in the log directory
log_run() {
local msg="${1:-$(cat /dev/stdin)}"
# shellcheck disable=SC2155
local date=$(/bin/date +"${DATE_FORMAT}")
printf "[%s] %s[%s]: %s\\n" \
"${date}" "${PROGNAME}" "${PID}" "${msg}" \
>> "${RUN_LOG_FILE}"
}
# Log a message in the system log file (syslog or journald)
log_global() {
local msg="${1:-$(cat /dev/stdin)}"
echo "${msg}" \
| /usr/bin/logger -p local0.notice -t autosysadmin
}
# Log a message in both places
log_all() {
local msg="${1:-$(cat /dev/stdin)}"
log_global "${msg}"
log_run "${msg}"
}
# Log a notable action in regular places
# and append it to the dedicated list
log_action() {
log_all "$*"
append_action "$*"
}
# Append a line in the actions.log file in the log directory
append_action() {
echo "$*" >> "${ACTIONS_FILE}"
}
# Print the content of the actions.log file
# or a fallback content (1st parameter) if empty
# shellcheck disable=SC2120
print_actions() {
local fallback=${1:-""}
if [ -s "${ACTIONS_FILE}" ]; then
cat "${ACTIONS_FILE}"
elif [ -n "${fallback}" ]; then
echo "${fallback}"
fi
}
# Log a an abort reason in regular places
# and append it to the dedicated list
log_abort() {
log_all "$*"
append_abort_reason "$*"
}
# Append a line in the abort.log file in the log directory
append_abort_reason() {
echo "$*" >> "${ABORT_FILE}"
}
# Print the content of the abort.log file
# or a fallback content (1st parameter) if empty
# shellcheck disable=SC2120
print_abort_reasons() {
local fallback=${1:-""}
if [ -s "${ABORT_FILE}" ]; then
cat "${ABORT_FILE}"
elif [ -n "${fallback}" ]; then
echo "${fallback}"
fi
}
# Print the content of the main log from the log directory
print_main_log() {
cat "${RUN_LOG_FILE}"
}
# Log an abort reason and quit the script
log_abort_and_quit() {
log_abort "$*"
quit
}
# Store the content from standard inpu
# into a file in the log directory named after the 1st parameter
save_in_log_dir() {
local file_name=$1
local file_path="${RUN_LOG_DIR}/${file_name}"
cat /dev/stdin > "${file_path}"
log_run "Saved \`${file_name}' file."
}
# Return the full path of the file in log directory
# based on the name in the 1st parameter
file_path_in_log_dir() {
echo "${RUN_LOG_DIR}/${1}"
}
format_mail_success() {
cat <<EOTEMPLATE
From: AutoSysadmin Evolix <${EMAIL_FROM}>
Content-Type: text/plain; charset=UTF-8
MIME-Version: 1.0
Content-Transfer-Encoding: 8bit
X-Script: ${PROGNAME}
X-RunId: ${RUN_ID}
To: ${EMAIL_CLIENT:-alert5@evolix.fr}
Cc: ${EMAIL_INTERNAL}
Subject: [autosysadmin] Intervention automatisée sur ${HOSTNAME_TEXT}
Bonjour,
Une intervention automatisée vient de se terminer.
Nom du serveur : ${HOSTNAME_TEXT}
Heure d'intervention : $(LC_ALL=fr_FR.utf8 date)
Script déclenché : ${PROGNAME}
### Actions réalisées
$(print_actions "Aucune")
### Réagir à cette intervention
Vous pouvez répondre à ce message (${EMAIL_FROM}).
En cas d'urgence, utilisez l'adresse ${EMERGENCY_MAIL}
ou notre ligne d'astreinte (${EMERGENCY_TEL})
--
Votre AutoSysadmin
EOTEMPLATE
}
format_mail_abort() {
cat <<EOTEMPLATE
From: AutoSysadmin Evolix <${EMAIL_FROM}>
Content-Type: text/plain; charset=UTF-8
MIME-Version: 1.0
Content-Transfer-Encoding: 8bit
X-Script: ${PROGNAME}
X-RunId: ${RUN_ID}
To: ${EMAIL_CLIENT:-alert5@evolix.fr}
Cc: ${EMAIL_INTERNAL}
Subject: [autosysadmin] Intervention automatisée interrompue sur ${HOSTNAME_TEXT}
Bonjour,
Une intervention automatisée a été déclenchée mais s'est interrompue.
Nom du serveur : ${HOSTNAME_TEXT}
Heure d'intervention : $(LC_ALL=fr_FR.utf8 date)
Script déclenché : ${PROGNAME}
### Actions réalisées
$(print_actions "Aucune")
### Raison(s) de l'interruption
$(print_abort_reasons "Inconnue")
### Réagir à cette intervention
Vous pouvez répondre à ce message (${EMAIL_FROM}).
En cas d'urgence, utilisez l'adresse ${EMERGENCY_MAIL}
ou notre ligne d'astreinte (${EMERGENCY_TEL})
--
Votre AutoSysadmin
EOTEMPLATE
}
# shellcheck disable=SC2028
print_report_information() {
echo "**Uptime**"
echo ""
uptime
echo ""
echo "**Utilisateurs récents**"
echo ""
who_file=$(file_path_in_log_dir "who-users")
if [ -s "${who_file}" ]; then
cat "${who_file}"
else
who --users
fi
echo ""
echo "**Espace disque**"
echo ""
df_file=$(file_path_in_log_dir "server-state/df.txt")
if [ -s "${df_file}" ]; then
cat "${df_file}"
else
df -h
fi
echo ""
echo "**Dmesg**"
echo ""
dmesg_file=$(file_path_in_log_dir "server-state/dmesg.txt")
if [ -s "${dmesg_file}" ]; then
tail -n 5 "${dmesg_file}"
else
dmesg | tail -n 5
fi
echo ""
echo "**systemd failed services**"
echo ""
failed_services_file=$(file_path_in_log_dir "server-state/systemctl-failed-services.txt")
if [ -s "${failed_services_file}" ]; then
cat "${failed_services_file}"
else
systemctl --no-legend --state=failed --type=service
fi
if command -v lxc-ls > /dev/null 2>&1; then
echo ""
echo "**LXC containers**"
echo ""
lxc_ls_file=$(file_path_in_log_dir "server-state/lxc-list.txt")
if [ -s "${lxc_ls_file}" ]; then
cat "${lxc_ls_file}"
else
lxc-ls --fancy
fi
fi
apache_errors_file=$(file_path_in_log_dir "apache-errors.log")
if [ -f "${apache_errors_file}" ]; then
echo ""
echo "**Apache errors**"
echo ""
cat "${apache_errors_file}"
fi
nginx_errors_file=$(file_path_in_log_dir "nginx-errors.log")
if [ -f "${nginx_errors_file}" ]; then
echo ""
echo "**Nginx errors**"
echo ""
cat "${nginx_errors_file}"
fi
}
format_mail_internal() {
cat <<EOTEMPLATE
From: AutoSysadmin Evolix <${EMAIL_FROM}>
Content-Type: text/plain; charset=UTF-8
MIME-Version: 1.0
Content-Transfer-Encoding: 8bit
X-Script: ${PROGNAME}
X-RunId: ${RUN_ID}
To: ${EMAIL_INTERNAL}
Subject: [autosysadmin] Rapport interne d'intervention sur ${HOSTNAME_TEXT}
Bonjour,
Une intervention automatique vient de se terminer.
Nom du serveur : ${HOSTNAME_TEXT}
Heure d'intervention : $(LC_ALL=fr_FR.utf8 date)
Script déclenché : ${PROGNAME}
### Actions réalisées
$(print_actions "Aucune")
### Raison(s) de l'interruption
$(print_abort_reasons "Aucune")
### Log autosysadmin
$(print_main_log)
### Informations additionnelles
$(print_report_information)
--
Votre AutoSysadmin
EOTEMPLATE
}
# Generic function to send emails at the end of the script.
# Takes a template as 1st parameter
hook_mail() {
if is_debug; then
log_run "Debug mode enabled: continue without sending mail."
return 0;
fi
HOSTNAME="${HOSTNAME:-"$(get_fqdn)"}"
HOSTNAME_TEXT="$(get_complete_hostname)"
EMAIL_CLIENT="$(get_evomaintenance_mail)"
EMERGENCY_MAIL="$(get_evomaintenance_emergency_mail)"
EMERGENCY_TEL="$(get_evomaintenance_emergency_tel)"
MAIL_CONTENT="$(format_mail_"$1")"
SENDMAIL_BIN="$(command -v sendmail)"
if [ -z "${SENDMAIL_BIN}" ]; then
log_global "ERROR: No \`sendmail' command has been found, can't send mail."
fi
if [ -x "${SENDMAIL_BIN}" ]; then
echo "${MAIL_CONTENT}" | "${SENDMAIL_BIN}" -oi -t -f "equipe@evolix.fr"
log_global "Sent '$1' mail for RUN_ID: ${RUN_ID}"
fi
}
is_holiday() {
# gcal mark today as a holiday by surrounding with < and > the day
# of the month of that holiday line. For example if today is 2022-05-01 we'll
# get among other lines:
# Fête du Travail (FR) + Di, < 1>Mai 2022
# Jour de la Victoire (FR) + Di, : 8:Mai 2022 = +7 jours
LANGUAGE=fr_FR.UTF-8 TZ=Europe/Paris gcal --cc-holidays=fr --holiday-list=short | grep -E '<[0-9 ]{2}>' --quiet
}
is_weekend() {
day_of_week=$(date +%u)
if [ "${day_of_week}" != 6 ] && [ "${day_of_week}" != 7 ]; then
return 1
fi
}
is_workday() {
if is_holiday || is_weekend; then
return 1
fi
}
is_worktime() {
if ! is_workday; then
return 1
fi
hour=$(date +%H)
if [ "${hour}" -lt 9 ] || { [ "${hour}" -ge 12 ] && [ "${hour}" -lt 14 ] ; } || [ "${hour}" -ge 18 ]; then
return 1
fi
}

View file

@ -0,0 +1,112 @@
#!/bin/bash
# Specific functions for "repair" scripts
is_all_repair_disabled() {
# Fetch values from the config
# and if it is not defined or has no value, then assign "on"
local status=${repair_all:=on}
test "${status}" = "off" || test "${status}" = "0"
}
is_current_repair_disabled() {
# Fetch values from the config
# and if it is not defined or has no value, then assign "on"
local status=${!PROGNAME:=on}
test "${status}" = "off" || test "${status}" = "0"
}
ensure_not_disabled_or_exit() {
if is_all_repair_disabled; then
log_global 'All repair scripts are disabled.'
exit 0
fi
if is_current_repair_disabled; then
log_global "Current repair script (${PROGNAME}) is disabled."
exit 0
fi
}
# Set of actions to do at the begining of a "repair" script
pre_repair() {
initialize
# Are we supposed to run?
ensure_not_disabled_or_exit
# Has it recently been run?
ensure_not_too_soon_or_exit
# Can we acquire a lock?
acquire_lock_or_exit
# Is there any active user?
ensure_no_active_users_or_exit
# Save important information
save_server_state
}
# Set of actions to do at the end of a "repair" script
post_repair() {
quit
}
repair_lxc_php() {
container_name=$1
if is_systemd_enabled 'lxc.service'; then
lxc_path=$(lxc-config lxc.lxcpath)
if lxc-info --name "${container_name}" > /dev/null; then
rootfs="${lxc_path}/${container_name}/rootfs"
case "${container_name}" in
php56) fpm_log_file="${rootfs}/var/log/php5-fpm.log" ;;
php70) fpm_log_file="${rootfs}/var/log/php7.0-fpm.log" ;;
php73) fpm_log_file="${rootfs}/var/log/php7.3-fpm.log" ;;
php74) fpm_log_file="${rootfs}/var/log/php7.4-fpm.log" ;;
php80) fpm_log_file="${rootfs}/var/log/php8.0-fpm.log" ;;
php81) fpm_log_file="${rootfs}/var/log/php8.1-fpm.log" ;;
php82) fpm_log_file="${rootfs}/var/log/php8.2-fpm.log" ;;
php83) fpm_log_file="${rootfs}/var/log/php8.3-fpm.log" ;;
*)
log_abort_and_quit "Unknown container '${container_name}'"
;;
esac
# Determine FPM Pool path
php_path_pool=$(find "${lxc_path}/${container_name}/" -type d -name "pool.d")
# Save LXC info (before restart)
lxc-info --name "${container_name}" | save_in_log_dir "lxc-${container_name}.before.status"
# Save last lines of FPM log (before restart)
tail "${fpm_log_file}" | save_in_log_dir "$(basename "${fpm_log_file}" | sed -e 's/.log/.before.log/')"
# Save NRPE check (before restart)
/usr/local/lib/nagios/plugins/check_phpfpm_multi "${php_path_pool}" | save_in_log_dir "check_fpm_${container_name}.before.out"
lxc-stop --timeout 20 --name "${container_name}"
lxc-start --daemon --name "${container_name}"
rc=$?
if [ "${rc}" -eq "0" ]; then
log_action "Restart LXC container '${container_name}: OK"
else
log_action "Restart LXC container '${container_name}: failed"
fi
# Save LXC info (after restart)
lxc-info --name "${container_name}" | save_in_log_dir "lxc-${container_name}.after.status"
# Save last lines of FPM log (after restart)
tail "${fpm_log_file}" | save_in_log_dir "$(basename "${fpm_log_file}" | sed -e 's/.log/.after.log/')"
# Save NRPE check (after restart)
/usr/local/lib/nagios/plugins/check_phpfpm_multi "${php_path_pool}" | save_in_log_dir "check_fpm_${container_name}.after.out"
else
log_abort_and_quit "LXC container '${container_name}' doesn't exist."
fi
else
log_abort_and_quit 'LXC not found.'
fi
}

View file

@ -0,0 +1,76 @@
#!/bin/bash
# Specific functions for "restart" scripts
running_custom() {
# Placeholder that returns 1, to prevent running if not redefined
log_global "running_custom() function has not been redefined! Let's quit."
return 1
}
# Examine RUNNING variable and decide if the script should run or not
is_supposed_to_run() {
if is_debug; then return 0; fi
case ${RUNNING} in
never)
# log_global "is_supposed_to_run: no (never)"
return 1
;;
always)
# log_global "is_supposed_to_run: yes (always)"
return 0
;;
nwh-fr)
! is_worktime
rc=$?
# if [ ${rc} -eq 0 ]; then
# log_global "is_supposed_to_run: yes (nwh-fr returned ${rc})"
# else
# log_global "is_supposed_to_run: no (nwh-fr returned ${rc})"
# fi
return ${rc}
;;
nwh-ca)
# Not implemented yet
return 0
;;
custom)
running_custom
rc=$?
# if [ ${rc} -eq 0 ]; then
# log_global "is_supposed_to_run: yes (custom returned ${rc})"
# else
# log_global "is_supposed_to_run: no (custom returned ${rc})"
# fi
return ${rc}
;;
esac
}
ensure_supposed_to_run_or_exit() {
if ! is_supposed_to_run; then
# simply quit (no logging, no notifications…)
# log_global "${PROGNAME} is not supposed to run (RUNNING=${RUNNING})."
exit 0
fi
}
# Set of actions to do at the begining of a "restart" script
pre_restart() {
initialize
# Has it recently been run?
ensure_not_too_soon_or_exit
# Can we acquire a lock?
acquire_lock_or_exit
# Save important information
save_server_state
}
# Set of actions to do at the end of a "restart" script
post_restart() {
quit
}

View file

@ -0,0 +1,157 @@
#!/bin/bash
: "${AUTOSYSADMIN_LIB:=/usr/local/lib/autosysadmin}"
source "${AUTOSYSADMIN_LIB}/common.sh" || exit 1
source "${AUTOSYSADMIN_LIB}/repair.sh" || exit 1
pre_repair
# We always keep some reserved blocks to avoid missing some logs
# https://gitea.evolix.org/evolix/autosysadmin/issues/22
RESERVED_BLOCKS_MIN=1
get_mountpoints() {
# the $(...) get the check_disk1 command
# the cut command selects the critical part of the check_disk1 output
# the grep command extracts the mountpoints and available disk space
# the last cut command selects the mountpoints
check_disk1_command=$(grep check_disk1 /etc/nagios/nrpe.d/evolix.cfg | cut -d'=' -f2-)
${check_disk1_command} -e | cut -d'|' -f1 | grep --extended-regexp --only-matching '/[[:graph:]]* [0-9]+ [A-Z][A-Z]' | cut -d' ' -f1
}
is_reserved_blocks_nominal() {
partition=${1}
fs_type="$(findmnt -n --output=fstype "${partition}")"
if [ "${fs_type}" = "ext4" ]; then
device="$(findmnt -n --output=source "${partition}")"
reserved_block_count="$(tune2fs -l "${device}" | grep 'Reserved block count' | awk -F':' '{ gsub (" ", "", $0); print $2}')"
block_count="$(tune2fs -l "${device}" | grep 'Block count' | awk -F':' '{ gsub (" ", "", $0); print $2}')"
percentage=$(awk "BEGIN { pc=100*${reserved_block_count}/${block_count}; i=int(pc); print (pc-i<0.5)?i:i+1 }")
log_run "Reserved blocks for ${partition} is currently at ${percentage}%"
if [ "${percentage}" -gt "${RESERVED_BLOCKS_MIN}" ]; then
log_run "Allowing tune2fs action to reduce the number of reserved blocks"
return 0
else
log_run "Reserved blocks already at or bellow ${RESERVED_BLOCKS_MIN}%, no automatic action possible"
return 1
fi
else
log_run "Filesystem for ${partition} (${fs_type}) is incompatible with reserved block reduction."
return 1
fi
}
reduce_reserved_blocks() {
partition=${1}
device=$(findmnt -n --output=source "${partition}")
tune2fs -m "${RESERVED_BLOCKS_MIN}" "${device}"
log_action "Reserved blocks for ${partition} changed to ${RESERVED_BLOCKS_MIN} percent"
}
is_tmp_to_delete() {
size="$(find /var/log/ -type f -ctime +1 -exec du {} \+ | awk '{s+=$1}END{print s / 1024}')"
if [ -n "${size}" ]; then
return 0
else
return 1
fi
}
is_log_to_delete() {
size="$(find /var/log/ -type f -mtime +365 -exec du {} \+ | awk '{s+=$1}END{print s / 1024}')"
if [ -n "${size}" ]; then
return 0
else
return 1
fi
}
clean_apt_cache() {
for container in $(lxc-ls -1); do
if [ -e "$(lxc-config lxc.lxcpath)/${container}/rootfs/var/cache" ]; then
lxc-attach --name "${container}" -- apt-get clean
log_action "Clean apt cache in LXC container ${container}";
fi
done
# NOTE: "head -n 1" might be superfluous, but let's be sure to have only the first returned value
biggest_subdir=$(du --summarize --one-file-system "/var/*" | sort --numeric-sort --reverse | sed 's/^[0-9]\+[[:space:]]\+//;q' | head -n 1)
case "${biggest_subdir}" in
'/var/cache')
apt-get clean
log_action 'Clean apt cache'
;;
esac
}
clean_amavis_virusmails() {
if du --inodes /var/lib/* | sort --numeric-sort | tail -n 3 | grep --quiet 'virusmails$'; then
find /var/lib/amavis/virusmails/ -type f -atime +30 -delete
log_action 'Clean amavis infected mails'
fi
}
critical_mountpoints=$(get_mountpoints)
if [ -z "${critical_mountpoints}" ]; then
log_abort_and_quit "No partition is in critical state, nothing left to do."
else
for mountpoint in ${critical_mountpoints}; do
case "${mountpoint}" in
/var)
#if is_log_to_delete
#then
# find /var/log/ -type f -mtime +365 -delete
# log_action "$size Mo of disk space freed in /var"
#fi
if is_reserved_blocks_nominal /var; then
reduce_reserved_blocks /var
clean_apt_cache
clean_amavis_virusmails
fi
;;
/tmp)
#if is_tmp_to_delete
#then
# find /tmp/ -type f -ctime +1 -delete
# log_action "$size Mo of disk space freed in /tmp"
#fi
if is_reserved_blocks_nominal /tmp; then
reduce_reserved_blocks /tmp
fi
;;
/home)
if is_reserved_blocks_nominal /home; then
reduce_reserved_blocks /home
fi
;;
/srv)
if is_reserved_blocks_nominal /srv; then
reduce_reserved_blocks /srv
fi
;;
/filer)
if is_reserved_blocks_nominal /filer; then
reduce_reserved_blocks /filer
fi
;;
/)
if is_reserved_blocks_nominal /; then
reduce_reserved_blocks /
# Suggest remove old kernel ?
fi
;;
*)
# unknown
log_run 'Unknown partition (or weird case) or nothing to do'
;;
esac
done
fi
post_repair

View file

@ -0,0 +1,35 @@
#!/bin/bash
: "${AUTOSYSADMIN_LIB:=/usr/local/lib/autosysadmin}"
source "${AUTOSYSADMIN_LIB}/common.sh" || exit 1
source "${AUTOSYSADMIN_LIB}/repair.sh" || exit 1
pre_repair
service="elasticsearch.service"
service_name="elasticsearch"
if is_systemd_enabled "${service}"; then
if is_systemd_active "${service}"; then
log_abort_and_quit "${service} is active, nothing left to do."
else
# Save service status before restart
systemctl status "${service}" | save_in_log_dir "${service_name}.before.status"
# Try to restart
timeout 20 systemctl restart "${service}" > /dev/null
rc=$?
if [ "${rc}" -eq "0" ]; then
log_action "Restart ${service_name}: OK"
else
log_action "Restart ${service_name}: failed"
fi
# Save service status after restart
systemctl status "${service}" | save_in_log_dir "${service_name}.after.status"
fi
else
log_abort_and_quit "${service} is disabled (or missing), nothing left to do."
fi
post_repair

View file

@ -0,0 +1,131 @@
#!/bin/bash
: "${AUTOSYSADMIN_LIB:=/usr/local/lib/autosysadmin}"
source "${AUTOSYSADMIN_LIB}/common.sh" || exit 1
source "${AUTOSYSADMIN_LIB}/repair.sh" || exit 1
pre_repair
## Apache
service="apache2.service"
service_name="apache2"
if is_systemd_enabled "${service}"; then
if is_systemd_active "${service}"; then
log_all "${service} is active. Skip."
else
# Save service status before restart
systemctl status "${service}" | save_in_log_dir "${service_name}.before.status"
# check syntax
if apache2ctl -t > /dev/null 2>&1; then
# Try to restart
timeout 20 systemctl restart "${service}" > /dev/null
rc=$?
if [ "${rc}" -eq "0" ]; then
log_action "Restart ${service_name}: OK"
else
log_action "Restart ${service_name}: failed"
fi
# Save service status after restart
systemctl status "${service}" | save_in_log_dir "${service_name}.after.status"
# Save error logs
date=$(LANG=en_US.UTF-8 date '+%b %d')
grep "${date}" /home/*/log/error.log /var/log/apache2/*error.log \
| grep -v \
-e "Got error 'PHP message:" \
-e "No matching DirectoryIndex" \
-e "client denied by server configuration" \
-e "server certificate does NOT include an ID which matches the server name" \
| save_in_log_dir "apache-errors.log"
else
log_action "Restart ${service_name}: skip (invalid configuration)"
fi
fi
else
log_all "${service} is disabled (or missing). Skip."
fi
## Nginx
service="nginx.service"
service_name="nginx"
if is_systemd_enabled "${service}"; then
if is_systemd_active "${service}"; then
log_all "${service} is active. Skip."
else
# Save service status before restart
systemctl status "${service}" | save_in_log_dir "${service_name}.before.status"
# check syntax
if nginx -t > /dev/null 2>&1; then
# Try to restart
timeout 20 systemctl restart "${service}" > /dev/null
rc=$?
if [ "${rc}" -eq "0" ]; then
log_action "Restart ${service_name}: OK"
else
log_action "Restart ${service_name}: failed"
fi
# Save service status after restart
systemctl status "${service}" | save_in_log_dir "${service_name}.after.status"
# Save error logs
### Consider doing for Nginx the same as Apache
else
log_action "Restart ${service_name}: skip (invalid configuration)"
fi
fi
else
log_all "${service} is disabled (or missing). Skip."
fi
## LXC
if is_systemd_enabled 'lxc.service'; then
for container in $(lxc-ls -1 | grep --fixed-strings 'php' | grep --extended-regexp --invert-match --regexp '\bold\b' --regexp '\bdisabled\b'); do
repair_lxc_php "${container}"
done
else
log_all "LXC is disabled (or missing). Skip."
fi
## FPM
fpm_services=$(systemd_list_services 'php*-fpm*')
if [ -n "${fpm_services}" ]; then
for service in ${fpm_services}; do
service_name="${service//.service/}"
if is_systemd_enabled "${service}"; then
if is_systemd_active "${service}"; then
log_all "${service} is active. Skip."
else
# Save service status before restart
systemctl status "${service}" | save_in_log_dir "${service_name}.before.status"
# Try to restart
timeout 20 systemctl restart "${service}" > /dev/null
rc=$?
if [ "${rc}" -eq "0" ]; then
log_action "Restart ${service_name}: OK"
else
log_action "Restart ${service_name}: failed"
fi
# Save service status after restart
systemctl status "${service}" | save_in_log_dir "${service_name}.after.status"
fi
else
log_all "${service} is disabled (or missing). Skip."
fi
done
else
log_all "PHP FPM not found. Skip."
fi
post_repair

View file

@ -0,0 +1,69 @@
#!/bin/bash
: "${AUTOSYSADMIN_LIB:=/usr/local/lib/autosysadmin}"
source "${AUTOSYSADMIN_LIB}/common.sh" || exit 1
source "${AUTOSYSADMIN_LIB}/repair.sh" || exit 1
pre_repair
if is_debian_version "8" "<="; then
if is_sysvinit_enabled '*mysql*'; then
if ! pgrep -u mysql mysqld > /dev/null; then
# Save service status before restart
timeout 2 mysqladmin status 2>&1 | save_in_log_dir "mysql.before.status"
timeout 20 /etc/init.d/mysql restart > /dev/null
rc=$?
if [ "${rc}" -eq "0" ]; then
log_action "Restart mysql: OK"
else
log_action "Restart mysql: failed"
fi
# Save service status after restart
timeout 2 mysqladmin status 2>&1 | save_in_log_dir "mysql.after.status"
else
log_abort_and_quit "mysqld process alive. Aborting"
fi
else
log_abort_and_quit "MySQL not enabled. Aborting"
fi
else
if is_debian_version "12" ">="; then
service="mariadb.service"
service_name="mariadb"
else
service="mysql.service"
service_name="mysql"
fi
if is_systemd_enabled "${service}"; then
if is_systemd_active "${service}"; then
log_abort_and_quit "${service} is active, nothing left to do."
else
# Save service status before restart
systemctl status "${service}" | save_in_log_dir "${service_name}.before.status"
# Try to restart
timeout 20 systemctl restart "${service}" > /dev/null
rc=$?
if [ "${rc}" -eq "0" ]; then
log_action "Restart ${service_name}: OK"
else
log_action "Restart ${service_name}: failed"
fi
# Save service status after restart
systemctl status "${service}" | save_in_log_dir "${service_name}.after.status"
fi
else
log_abort_and_quit "${service} is disabled (or missing), nothing left to do."
fi
fi
post_repair

View file

@ -0,0 +1,35 @@
#!/bin/bash
: "${AUTOSYSADMIN_LIB:=/usr/local/lib/autosysadmin}"
source "${AUTOSYSADMIN_LIB}/common.sh" || exit 1
source "${AUTOSYSADMIN_LIB}/repair.sh" || exit 1
pre_repair
service="opendkim.service"
service_name="opendkim"
if is_systemd_enabled "${service}"; then
if is_systemd_active "${service}"; then
log_abort_and_quit "${service} is active, nothing left to do."
else
# Save service status before restart
systemctl status "${service}" | save_in_log_dir "${service_name}.before.status"
# Try to restart
timeout 20 systemctl restart "${service}" > /dev/null
rc=$?
if [ "${rc}" -eq "0" ]; then
log_action "Restart ${service_name}: OK"
else
log_action "Restart ${service_name}: failed"
fi
# Save service status after restart
systemctl status "${service}" | save_in_log_dir "${service_name}.after.status"
fi
else
log_abort_and_quit "${service} is disabled (or missing). Abort."
fi
post_repair

View file

@ -0,0 +1,14 @@
#!/bin/bash
: "${AUTOSYSADMIN_LIB:=/usr/local/lib/autosysadmin}"
source "${AUTOSYSADMIN_LIB}/common.sh" || exit 1
source "${AUTOSYSADMIN_LIB}/repair.sh" || exit 1
LOCK_WAIT="15s"
LOCK_NAME="repair_http"
pre_repair
repair_lxc_php php56
post_repair

View file

@ -0,0 +1,14 @@
#!/bin/bash
: "${AUTOSYSADMIN_LIB:=/usr/local/lib/autosysadmin}"
source "${AUTOSYSADMIN_LIB}/common.sh" || exit 1
source "${AUTOSYSADMIN_LIB}/repair.sh" || exit 1
LOCK_WAIT="15s"
LOCK_NAME="repair_http"
pre_repair
repair_lxc_php php70
post_repair

View file

@ -0,0 +1,14 @@
#!/bin/bash
: "${AUTOSYSADMIN_LIB:=/usr/local/lib/autosysadmin}"
source "${AUTOSYSADMIN_LIB}/common.sh" || exit 1
source "${AUTOSYSADMIN_LIB}/repair.sh" || exit 1
LOCK_WAIT="15s"
LOCK_NAME="repair_http"
pre_repair
repair_lxc_php php73
post_repair

View file

@ -0,0 +1,14 @@
#!/bin/bash
: "${AUTOSYSADMIN_LIB:=/usr/local/lib/autosysadmin}"
source "${AUTOSYSADMIN_LIB}/common.sh" || exit 1
source "${AUTOSYSADMIN_LIB}/repair.sh" || exit 1
LOCK_WAIT="15s"
LOCK_NAME="repair_http"
pre_repair
repair_lxc_php php74
post_repair

View file

@ -0,0 +1,14 @@
#!/bin/bash
: "${AUTOSYSADMIN_LIB:=/usr/local/lib/autosysadmin}"
source "${AUTOSYSADMIN_LIB}/common.sh" || exit 1
source "${AUTOSYSADMIN_LIB}/repair.sh" || exit 1
LOCK_WAIT="15s"
LOCK_NAME="repair_http"
pre_repair
repair_lxc_php php80
post_repair

View file

@ -0,0 +1,14 @@
#!/bin/bash
: "${AUTOSYSADMIN_LIB:=/usr/local/lib/autosysadmin}"
source "${AUTOSYSADMIN_LIB}/common.sh" || exit 1
source "${AUTOSYSADMIN_LIB}/repair.sh" || exit 1
LOCK_WAIT="15s"
LOCK_NAME="repair_http"
pre_repair
repair_lxc_php php81
post_repair

View file

@ -0,0 +1,14 @@
#!/bin/bash
: "${AUTOSYSADMIN_LIB:=/usr/local/lib/autosysadmin}"
source "${AUTOSYSADMIN_LIB}/common.sh" || exit 1
source "${AUTOSYSADMIN_LIB}/repair.sh" || exit 1
LOCK_WAIT="15s"
LOCK_NAME="repair_http"
pre_repair
repair_lxc_php php82
post_repair

View file

@ -0,0 +1,14 @@
#!/bin/bash
: "${AUTOSYSADMIN_LIB:=/usr/local/lib/autosysadmin}"
source "${AUTOSYSADMIN_LIB}/common.sh" || exit 1
source "${AUTOSYSADMIN_LIB}/repair.sh" || exit 1
LOCK_WAIT="15s"
LOCK_NAME="repair_http"
pre_repair
repair_lxc_php php83
post_repair

View file

@ -0,0 +1,32 @@
#!/bin/bash
: "${AUTOSYSADMIN_LIB:=/usr/local/lib/autosysadmin}"
source "${AUTOSYSADMIN_LIB}/common.sh" || exit 1
source "${AUTOSYSADMIN_LIB}/repair.sh" || exit 1
pre_repair
for service in $(systemd_list_services 'redis-server*'); do
service_name="${service//.service/}"
if is_systemd_active "${service}"; then
log_all "${service} is active. Skip."
else
# Save service status before restart
systemctl status "${service}" | save_in_log_dir "${service_name}.before.status"
# Try to restart
timeout 20 systemctl restart "${service}" > /dev/null
rc=$?
if [ "${rc}" -eq "0" ]; then
log_action "Restart ${service_name}: OK."
else
log_action "Restart ${service_name}: failed."
fi
# Save service status after restart
systemctl status "${service}" | save_in_log_dir "${service_name}.after.status"
fi
done
post_repair

View file

@ -1,43 +1,24 @@
#!/bin/bash
# Source functions file
# shellcheck source=./functions.sh
source /usr/share/scripts/autosysadmin/functions.sh
: "${AUTOSYSADMIN_LIB:=/usr/local/lib/autosysadmin}"
source "${AUTOSYSADMIN_LIB}/common.sh" || exit 1
source "${AUTOSYSADMIN_LIB}/repair.sh" || exit 1
init_autosysadmin
load_conf
test "${repair_tomcat_instance:=off}" = off && log_error_exit 'Script disabled, nothing to do here!'
# Has it recently been run?
is_too_soon
lockfile="/run/lock/repair_tomcat_instance"
cleanup() {
rm -f "${lockfile}"
}
trap 'cleanup' 0
acquire_lock_or_exit "${lockfile}"
ensure_no_active_users_or_exit
# The actual work starts below !
log_system_status
pre_repair
repair_tomcat_instance_handle_tomcat() {
if /bin/su - "${1}" -c "/bin/systemctl --quiet --user is-active tomcat.service" ; then
if ! /bin/su - "${1}" -c "/usr/bin/timeout 20 /bin/systemctl --quiet --user restart tomcat.service"
then
log_error_exit "Echec de redémarrage instance tomcat utilisateur ${1}"
log_abort_and_quit "Echec de redémarrage instance tomcat utilisateur ${1}"
else
log_action "Redémarrage instance tomcat utilisateur ${1}"
fi
elif /bin/systemctl --quiet is-active "${1}".service ; then
if ! /usr/bin/timeout 20 systemctl --quiet restart "${1}".service
then
log_error_exit "Echec de redémarrage instance tomcat ${1}"
log_abort_and_quit "Echec de redémarrage instance tomcat ${1}"
else
log_action "Redémarrage instance tomcat ${1}"
fi
@ -50,4 +31,4 @@ do
repair_tomcat_instance_handle_tomcat "${instance}"
done
AUTOSYSADMIN=1 /usr/share/scripts/evomaintenance.sh -m "$0: done" --no-commit --no-mail
post_repair

View file

@ -0,0 +1,41 @@
#!/bin/bash
: "${AUTOSYSADMIN_LIB:=/usr/local/lib/autosysadmin}"
source "${AUTOSYSADMIN_LIB}/common.sh" || exit 1
source "${AUTOSYSADMIN_LIB}/repair.sh" || exit 1
## Custom lock wait and/or lock name
# LOCK_WAIT="15s"
# LOCK_NAME="repair_http"
pre_repair
## The name of the service, mainly for logging
service_name="example"
## The systemd service name
systemd_service="${service_name}.service"
if is_systemd_enabled "${systemd_service}"; then
if is_systemd_active "${systemd_service}"; then
log_abort_and_quit "${systemd_service} is active, nothing left to do."
else
# Save service status before restart
systemctl status "${systemd_service}" | save_in_log_dir "${service_name}.before.status"
# Try to restart
timeout 20 systemctl restart "${systemd_service}" > /dev/null
rc=$?
if [ "${rc}" -eq "0" ]; then
log_action "Restart ${service_name}: OK"
else
log_action "Restart ${service_name}: failed"
fi
# Save service status after restart
systemctl status "${systemd_service}" | save_in_log_dir "${service_name}.after.status"
fi
else
log_abort_and_quit "${service_name} is disabled (or missing), nothing left to do."
fi
post_repair

View file

@ -0,0 +1,19 @@
Autosysadmin "restart auto" scripts
===================================
In this directory you can place scripts that will be executed automatically by a cron job (stored in `/etc/cron.d/autosysadmin`).
They must satisfy the default `run-parts(8)` constraints :
* be "executable"
* belong to the Debian cron script namespace (`^[a-zA-Z0-9_-]+$`), example: `restart_amavis`
Warning: scripts that do not satisfy those criteria will NOT be run (silently)!
You can print the names of the scripts which would be run, without actually running them, with this command :
```
$ run-parts --test /usr/share/scripts/autosysadmin/restart
```
You can use `zzz-restart_example.template` as boilerplate code to make your own "restart" script.

View file

@ -0,0 +1,120 @@
#!/bin/bash
: "${AUTOSYSADMIN_LIB:=/usr/local/lib/autosysadmin}"
source "${AUTOSYSADMIN_LIB}/common.sh" || exit 1
source "${AUTOSYSADMIN_LIB}/restart.sh" || exit 1
# shellcheck disable=SC2034
RUNNING="nwh-fr"
## Possible values for RUNNING :
## never => disabled
## always => enabled
## nwh-fr => enabled during non-working-hours in France
## nwh-ca => enabled during non-working-hours in Canada (not supported yet)
## custom => enabled if `running_custom()` function returns 0, otherwise disabled.
## Uncomment and customize this method if you want to have a special logic :
##
## return 1 if we should not run
## return 0 if we should run
##
## Some available functions :
## is_weekend() : Saturday or Sunday
## is_holiday() : holiday in France (based on `gcal(1)`)
## is_workday() : not weekend and not holiday
## is_worktime() : work day between 9-12h and 14-18h
#
# running_custom() {
# # implement your own custom method to decide if we should run or not
# }
## The name of the service, mainly for logging
service_name="example"
## The SysVinit script name
sysvinit_script="${service_name}"
## The systemd service name
systemd_service="${service_name}.service"
is_service_alive() {
## this must return 0 if the service is alive, otherwise return 1
## Example:
pgrep -u USER PROCESS_NAME > /dev/null
}
## Action for SysVinit system
sysvinit_action() {
# Save service status before restart
timeout 2 "/etc/init.d/${sysvinit_script}" status | save_in_log_dir "${service_name}.before.status"
# Try to restart
timeout 20 "/etc/init.d/${sysvinit_script}" restart > /dev/null
rc=$?
if [ "${rc}" -eq "0" ]; then
log_action "Restart ${service_name}: OK"
else
log_action "Restart ${service_name}: failed"
fi
# Save service status after restart
timeout 2 "/etc/init.d/${sysvinit_script}" status | save_in_log_dir "${service_name}.after.status"
}
## Action for systemd system
systemd_action() {
# Save service status before restart
systemctl status "${systemd_service}" | save_in_log_dir "${service_name}.before.status"
# Try to restart
# systemctl (only for NRPE ?) sometimes returns 0 even if the service has failed to start
# so we check the status explicitly
timeout 20 systemctl restart "${systemd_service}" > /dev/null \
&& sleep 1 \
&& systemctl status "${systemd_service}" > /dev/null
rc=$?
if [ "${rc}" -eq "0" ]; then
log_action "Restart ${service_name}: OK"
else
log_action "Restart ${service_name}: failed"
fi
# Save service status after restart
systemctl status "${systemd_service}" | save_in_log_dir "${service_name}.after.status"
}
# Should we run?
if ! is_supposed_to_run; then
# log_global "${PROGNAME} is not supposed to run (RUNNING=${RUNNING})."
exit 0
fi
if is_service_alive; then
# log_global "${service_name} process alive. Aborting"
exit 0
fi
# Yes we do, so check for sysvinit or systemd
if is_debian_version "8" "<="; then
if ! is_sysvinit_enabled "*${sysvinit_script}*"; then
# log_global "${service_name} not enabled. Aborting"
exit 0
fi
# Let's finally do the action
pre_restart
sysvinit_action
post_restart
else
if ! is_systemd_enabled "${systemd_service}"; then
# log_global "${service_name} is disabled (or missing), nothing left to do."
exit 0
fi
if is_systemd_active "${systemd_service}"; then
# log_global "${service_name} is active, nothing left to do."
exit 0
fi
# Let's finally do the action
pre_restart
systemd_action
post_restart
fi

View file

@ -0,0 +1,16 @@
---
- name: restart nagios-nrpe-server
service:
name: nagios-nrpe-server
state: restarted
- name: restart nrpe
service:
name: nrpe
state: restarted
- name: restart rsyslog
service:
name: rsyslog
state: restarted

View file

@ -0,0 +1,25 @@
---
- name: "Add begin marker if missing"
ansible.builtin.lineinfile:
path: "/etc/cron.d/autosysadmin"
line: "# BEGIN ANSIBLE MANAGED SECTION FOR AUTOSYSADMIN"
insertbefore: BOF
create: yes
- name: "Add end marker if missing"
ansible.builtin.lineinfile:
path: "/etc/cron.d/autosysadmin"
line: "# END ANSIBLE MANAGED SECTION FOR AUTOSYSADMIN"
insertbefore: "EOF"
create: yes
- name: "Create config if missing"
ansible.builtin.blockinfile:
path: "/etc/cron.d/autosysadmin"
marker: "# {mark} ANSIBLE MANAGED SECTION FOR AUTOSYSADMIN"
block: "{{ lookup('ansible.builtin.template', '../templates/autosysadmin.cron.j2') }}"
owner: root
group: root
mode: "0750"
create: yes

View file

@ -0,0 +1,114 @@
---
- name: "Remount /usr if needed"
ansible.builtin.include_role:
name: remount-usr
- name: Previous autosysadmin restart directory is renamed
command:
cmd: mv "/usr/share/scripts/autosysadmin/auto" "{{ autosysadmin_agent_auto_dir }}"
removes: "/usr/share/scripts/autosysadmin/auto"
creates: "{{ autosysadmin_agent_auto_dir }}"
- name: Create autosysadmin directories
ansible.builtin.file:
path: "{{ item }}"
state: directory
owner: "root"
group: "root"
mode: "0750"
loop:
- "{{ autosysadmin_agent_bin_dir }}"
- "{{ autosysadmin_agent_lib_dir }}"
- "{{ autosysadmin_agent_auto_dir }}"
- name: Copy libraries
ansible.builtin.copy:
src: "upstream/lib/"
dest: "{{ autosysadmin_agent_lib_dir }}/"
owner: root
group: root
mode: "0750"
- name: Copy repair scripts
ansible.builtin.copy:
src: "upstream/repair/"
dest: "{{ autosysadmin_agent_bin_dir }}/"
owner: root
group: root
mode: "0750"
- name: Copy other utilities
ansible.builtin.copy:
src: "upstream/bin/"
dest: "{{ autosysadmin_agent_bin_dir }}/"
owner: root
group: root
mode: "0750"
### WARNING: thos files are explicitly marked as non-executable
### to prevent them from being run automatically by run-parts
- name: Copy restart scripts
ansible.builtin.copy:
src: "upstream/restart/"
dest: "{{ autosysadmin_agent_auto_dir }}/"
owner: root
group: root
mode: "0640"
- name: Ensure /etc/evolinux folder exists
ansible.builtin.file:
path: "/etc/evolinux"
state: directory
owner: "root"
group: "root"
mode: "0700"
- name: Copy the configuration file if missing
ansible.builtin.template:
src: "autosysadmin.cf.j2"
dest: "/etc/evolinux/autosysadmin"
owner: root
group: root
mode: "0640"
force: no
# Repair scripts are supposed to be 'on' by default
# A line "repair_XXX=off" is added to the file only if the script is to be disabled.
# That's why all the ternary logic for the state is reversed.
- name: Update value per variable
ansible.builtin.lineinfile:
dest: "/etc/evolinux/autosysadmin"
line: "{{ item }}={{ autosysadmin_config[item] | default(true) | bool | ternary('on', 'off') }}"
regexp: '^(#\s*)?{{ item }}=.*'
state: "{{ autosysadmin_config[item] | default(true) | bool | ternary('absent', 'present') }}"
register: _line
loop: "{{ autosysadmin_repair_scripts | union(['repair_all']) }}"
- name: Ensure restart folder exists
ansible.builtin.file:
path: "auto"
state: directory
owner: "root"
group: "root"
mode: "0700"
- name: Legacy scripts are removed
ansible.builtin.file:
path: "{{ general_scripts_dir }}/autosysadmin/{{ item }}"
state: absent
loop:
- repair_amavis.sh
- repair_disk.sh
- repair_elasticsearch.sh
- repair_http.sh
- repair_mysql.sh
- repair_opendkim.sh
- repair_php_fpm56.sh
- repair_php_fpm70.sh
- repair_php_fpm73.sh
- repair_php_fpm74.sh
- repair_php_fpm80.sh
- repair_php_fpm81.sh
- repair_redis.sh
- repair_tomcat_instance.sh

View file

@ -1,10 +1,8 @@
---
- name: Copy logrotate configuration for autosysadmin
ansible.builtin.copy:
src: "files/logrotate_autosysadmin.conf"
src: "files/autosysadmin.logrotate.conf"
dest: "/etc/logrotate.d/autosysadmin"
owner: root
group: root
mode: "0644"
tags:
- autosysadmin

View file

@ -0,0 +1,31 @@
---
- name: The list of all repair scripts is composed.
set_fact:
autosysadmin_repair_scripts: "{{ lookup('ansible.builtin.fileglob', '../../../autosysadmin/agent/repair/repair_*', wantlist=True) | map('basename') | sort }}"
- name: Install dependencies
ansible.builtin.include_tasks: dependencies.yml
- name: Install autosysadmin
ansible.builtin.include_tasks: install.yml
- name: Crontab configuration
ansible.builtin.include_tasks: crontab.yml
- name: NRPE configuration
ansible.builtin.include_tasks: nrpe.yml
- name: sudo configuration
ansible.builtin.include_tasks: sudo.yml
- name: rsyslog configuration
ansible.builtin.include_tasks: rsyslog.yml
- name: logrotate configuration
ansible.builtin.include_tasks: logrotate.yml
- name: Install latest version of dump-server-state
ansible.builtin.include_role:
name: evolinux-base
tasks_from: dump-server-state.yml

View file

@ -0,0 +1,9 @@
---
- name: custom configuration is present
ansible.builtin.template:
src: autosysadmin.nrpe.cfg.j2
dest: /etc/nagios/nrpe.d/autosysadmin.cfg
group: nagios
mode: "0640"
force: yes
notify: restart nagios-nrpe-server

View file

@ -1,11 +1,9 @@
---
- name: Copy rsyslog configuration for autosysadmin
ansible.builtin.copy:
src: "files/rsyslog_autosysadmin.conf"
src: "files/autosysadmin.rsyslog.conf"
dest: "/etc/rsyslog.d/autosysadmin.conf"
owner: root
group: root
mode: "0644"
notify: Restart rsyslog
tags:
- autosysadmin
notify: restart rsyslog

View file

@ -1,9 +1,7 @@
---
- name: Add autosysadmin sudoers file
ansible.builtin.template:
src: sudoers.j2
src: autosysadmin.sudoers.j2
dest: /etc/sudoers.d/autosysadmin
mode: "0600"
validate: "visudo -cf %s"
tags:
- autosysadmin

View file

@ -0,0 +1,12 @@
# This configuration is partially managed by Ansible
# You can change specific values manually, but they may be overridden by Ansible
#
# To be safe, update the hosts_vars/group_vars in the autosysadmin project
# https://gitea.evolix.org/evolix/autosysadmin/src/branch/master
# then use the "agent" playbook to deploy.
#
# Configuration for autosysadmin
# Use this file to change configuration values defined in repair scripts
# To disable all repair scripts : repair_all=off
# To disable "repair_http" : repair_http=off
#

View file

@ -0,0 +1,7 @@
PATH=/usr/local/sbin:/usr/local/bin:/usr/sbin:/usr/bin:/sbin:/bin
# Run each enabled script
*/5 * * * * root run-parts /usr/share/scripts/autosysadmin/restart
# Clean run log files
@weekly root {{ autosysadmin_agent_bin_dir | mandatory }}/delete_old_logs.sh {{ autosysadmin_agent_log_retention_days | default('365') }}

View file

@ -0,0 +1,8 @@
#
# Ansible managed - DO NOT MODIFY, your changes will be overwritten !
#
# Autosysadmin repair commands
{% for script in lookup('ansible.builtin.fileglob', '../../../autosysadmin/agent/repair/repair_*', wantlist=True) | map("basename") | sort %}
command[{{ script }}]=sudo {{ autosysadmin_agent_bin_dir }}/{{ script }}
{% endfor %}

View file

@ -0,0 +1,7 @@
#
# Ansible managed - DO NOT MODIFY, your changes will be overwritten !
#
{% for script in lookup('ansible.builtin.fileglob', '../../../autosysadmin/agent/repair/repair_*', wantlist=True) | map("basename") | sort %}
nagios ALL = NOPASSWD: {{ autosysadmin_agent_bin_dir }}/{{ script }}
{% endfor %}

View file

@ -0,0 +1,8 @@
---
general_scripts_dir: "/usr/share/scripts"
restart_nrpe_path: "{{ general_scripts_dir }}/autosysadmin/restart/restart_nrpe"
# Change this to customize the RUNNING value in the script
restart_nrpe_running: Null

View file

@ -0,0 +1,105 @@
#!/bin/bash
: "${AUTOSYSADMIN_LIB:=/usr/local/lib/autosysadmin}"
source "${AUTOSYSADMIN_LIB}/common.sh" || exit 1
source "${AUTOSYSADMIN_LIB}/restart.sh" || exit 1
## Possible values for RUNNING :
## never => disabled
## always => enabled
## nwh-fr => enabled during non-working-hours in France
## nwh-ca => enabled during non-working-hours in Canada (not supported yet)
## custom => enabled if `running_custom()` function return 0, otherwise disabled.
# shellcheck disable=SC2034
RUNNING="nwh-fr"
## The name of the service, mainly for logging
service_name="nagios-nrpe-server"
## The SysVinit script name
sysvinit_script="${service_name}"
## The systemd service name
systemd_service="${service_name}.service"
is_service_alive() {
## this must return 0 if the service is alive, otherwise return 1
## Example:
pgrep -u nagios nrpe > /dev/null
}
## Action for SysVinit system
sysvinit_action() {
# Save service status before restart
timeout 2 "/etc/init.d/${sysvinit_script}" status | save_in_log_dir "${service_name}.before.status"
# Try to restart
timeout 20 "/etc/init.d/${sysvinit_script}" restart > /dev/null
rc=$?
if [ "${rc}" -eq "0" ]; then
log_action "Restart ${service_name}: OK"
else
log_action "Restart ${service_name}: failed"
fi
# Save service status after restart
timeout 2 "/etc/init.d/${sysvinit_script}" status | save_in_log_dir "${service_name}.after.status"
}
## Action for systemd system
systemd_action() {
# Save service status before restart
systemctl status "${systemd_service}" | save_in_log_dir "${service_name}.before.status"
# Try to restart
# systemctl (only for NRPE ?) sometimes returns 0 even if the service has failed to start
# so we check the status explicitly
timeout 20 systemctl restart "${systemd_service}" > /dev/null \
&& sleep 1 \
&& systemctl status "${systemd_service}" > /dev/null
rc=$?
if [ "${rc}" -eq "0" ]; then
log_action "Restart ${service_name}: OK"
else
log_action "Restart ${service_name}: failed"
fi
# Save service status after restart
systemctl status "${systemd_service}" | save_in_log_dir "${service_name}.after.status"
}
# Should we run?
if ! is_supposed_to_run; then
# log_global "${PROGNAME} is not supposed to run (RUNNING=${RUNNING})."
exit 0
fi
if is_service_alive; then
# log_global "${service_name} process alive. Aborting"
exit 0
fi
# Yes we do, so check for sysvinit or systemd
if is_debian_version "8" "<="; then
if ! is_sysvinit_enabled "*${sysvinit_script}*"; then
# log_global "${service_name} not enabled. Aborting"
exit 0
fi
# Let's finally do the action
pre_restart
sysvinit_action
post_restart
else
if ! is_systemd_enabled "${systemd_service}"; then
# log_global "${service_name} is disabled (or missing), nothing left to do."
exit 0
fi
if is_systemd_active "${systemd_service}"; then
# log_global "${service_name} is active, nothing left to do."
exit 0
fi
# Let's finally do the action
pre_restart
systemd_action
post_restart
fi

View file

@ -0,0 +1,24 @@
---
- name: "Remount /usr if needed"
ansible.builtin.include_role:
name: remount-usr
- name: "Copy restart_nrpe"
ansible.builtin.copy:
src: upstream/restart_nrpe
dest: "{{ restart_nrpe_path }}"
owner: "root"
group: "root"
mode: "0750"
- name: "Customize RUNNING value"
ansible.builtin.lineinfile:
path: "{{ restart_nrpe_path }}"
line: "RUNNING=\"{{ restart_nrpe_running }}\""
regexp: "^ *RUNNING="
create: False
when:
- restart_nrpe_running is defined
- restart_nrpe_running != None
- restart_nrpe_running | length > 0

View file

@ -1,22 +0,0 @@
---
general_scripts_dir: "/usr/share/scripts"
autosysadmin_dir: "{{ general_scripts_dir }}/autosysadmin"
# Default values for enabled checks
repair_amavis: 'on'
repair_disk: 'on'
repair_elasticsearch: 'on'
repair_http: 'on'
repair_mysql: 'on'
repair_opendkim: 'off'
repair_php_fpm56: 'off'
repair_php_fpm70: 'off'
repair_php_fpm73: 'off'
repair_php_fpm74: 'off'
repair_php_fpm80: 'off'
repair_php_fpm81: 'off'
repair_php_fpm82: 'off'
repair_php_fpm83: 'off'
repair_redis: 'off'
repair_tomcat_instance: 'off'

View file

@ -1,478 +0,0 @@
#!/bin/bash
get_system() {
uname -s
}
get_fqdn() {
if [ "$(get_system)" = "Linux" ]; then
hostname --fqdn
elif [ "$(get_system)" = "OpenBSD" ]; then
hostname
else
log_error_exit "OS not detected!"
fi
}
get_complete_hostname() {
REAL_HOSTNAME="$(get_fqdn)"
if [ "${HOSTNAME}" = "${REAL_HOSTNAME}" ]; then
echo "${HOSTNAME}"
else
echo "${HOSTNAME} (${REAL_HOSTNAME})"
fi
}
get_evomaintenance_mail() {
email="$(grep "EVOMAINTMAIL=" /etc/evomaintenance.cf | cut -d '=' -f2)"
if [[ -z "$email" ]]; then
email='alert5@evolix.fr'
fi
echo "${email}"
}
arguments="${*}"
get_argument() {
no_found=1
for argument in ${arguments} ; do
if [ "${argument}" = "${1}" ] ;
then
no_found=0
fi
done
return ${no_found}
}
internal_info() {
INTERNAL_INFO="$(printf '%b\n%s' "${INTERNAL_INFO}" "$*")"
}
log_action() {
log "Action : $*"
ACTIONS="$(printf '%s\n%s' "${ACTIONS}" "$*")"
}
log() {
INTERNAL_LOG="$(printf '%s\n%s %s %s %s' "${INTERNAL_LOG}" "$(date -Isec)" "$(hostname)" "$(basename "$0")" "$*")"
printf '%s %s %s %s\n' "$(date -Isec)" "$(hostname)" "$(basename "$0")" "$*" | tee -a "${LOG_DIR}/autosysadmin.log"
echo "$*" | /usr/bin/logger -p local0.notice -t autosysadmin."$0"
}
log_error_exit() {
log "ERROR : $*"
AUTOSYSADMIN=1 /usr/share/scripts/evomaintenance.sh -m "$0: $*" --no-commit --no-mail
exit 1
}
log_check_php_fpm() {
# Extraire seulement les chiffres du nom du script exécuté
# ./repair_php_fpm81.sh ==> 81
PHP_VERSION="${0//[^0-9]/}"
PHP_PATH_POOL=$(find /var/lib/lxc/php"${PHP_VERSION}"/ -type d -name "pool.d")
/usr/local/lib/nagios/plugins/check_phpfpm_multi "${PHP_PATH_POOL}" > "${LOG_DIR}/nrpe.txt"
}
log_system_status() {
DUMP_SERVER_STATE_BIN="$(command -v dump-server-state || command -v backup-server-state)"
if [ -z "${DUMP_SERVER_STATE_BIN}" ]; then
log "Warning: dump-server-state is not present. No server state recorded...."
fi
if [ -x "${DUMP_SERVER_STATE_BIN}" ]; then
# NOTE We don't want the logging to take too much time, so we kill it
# if it take more than 20 seconds.
timeout --signal 9 20 \
"${DUMP_SERVER_STATE_BIN}" \
--dump-dir="$LOG_DIR" \
--df \
--dmesg \
--iptables \
--lxc \
--netcfg \
--netstat \
--uname \
--processes \
--systemctl \
--uptime \
--virsh \
--disks \
--mysql-processes \
--no-apt-states \
--no-apt-config \
--no-dpkg-full \
--no-dpkg-status \
--no-mount \
--no-packages \
--no-sysctl \
--no-etc
log "System status logged in ${LOG_DIR}"
fi
}
read_log_system_status(){
files="df.txt dmesg.txt lxc-list.txt netstat-legacy.txt netstat-ss.txt pstree.txt ps.txt systemctl-failed-services.txt"
echo -e "\n\n#### Détails de dump-server-state"
for file in ${files} ; do
echo -e "\n### cat ${LOG_DIR}/${file} :"
tail -n 1000 "${LOG_DIR}"/"${file}"
done
}
ensure_no_active_users_or_exit() {
if is_debug; then return; fi
# Is there any active user ?
for user in $(LC_ALL=C who --users|awk '{print $1}'); do
idle_time="$(LC_ALL=C who --users | grep "${user}" | awk '{ print $6}')"
for sameusertime in $(LC_ALL=C who --users | grep "${user}" | awk '{ print $6}'); do
if is_active_user "$sameusertime"; then
hook_mail abort_active_users
log_error_exit 'At least one user was recently active. That requires human intervention. Nothing to do here!'
fi
done
done
}
is_active_user() {
# Check if a user was active in the last 30 minutes
idle_time="$1"
if [ "${idle_time}" = "old" ];
then
return 1
elif [ "${idle_time}" = "." ];
then
return 0
else
hh="$(echo "${idle_time}" | awk -F':' '{print $1}')"
mm="$(echo "${idle_time}" | awk -F':' '{print $2}')"
idle_minutes="$(( 60 * "${hh}" + "${mm}" ))"
if [ "${idle_minutes}" -ge 30 ];
then
return 1
else
return 0
fi
fi
}
is_debug() {
debug_file="/etc/evolinux/autosysadmin.debug"
if [ -e "${debug_file}" ]; then
last_change=$(stat -c %Z "${debug_file}")
limit_date=$(date --date "14400 seconds ago" +"%s")
if [ $(( last_change - limit_date )) -le "0" ]; then
rm "${debug_file}"
else
return 0
fi
fi
return 1
}
check_nrpe() {
check="$1"
list_command_nrpe=$( grep --exclude=*~ -E "\[${check}\]" -r /etc/nagios/ | grep -v '#command' )
command_nrpe_primary=$( echo "${list_command_nrpe}" | grep "/etc/nagios/nrpe.d/evolix.cfg" | cut -d'=' -f2- )
command_nrpe_secondary=$( echo "${list_command_nrpe}" | head -n1 | cut -d'=' -f2- )
if [ -z "${command_nrpe_primary}" ] && [ -z "${command_nrpe_secondary}" ]
then
return 1
else
if [ -n "${command_nrpe_primary}" ]
then
${command_nrpe_primary}
else
${command_nrpe_secondary}
fi
fi
}
acquire_lock_or_exit() {
lockfile="$1"
waittime="$2"
# si le temps dattente nest pas compréhensible par sleep(1), il vaut 0
if ! echo "${waittime}" | grep -Eq '^[0-9]+[smhd]?$'
then
waittime=0
fi
# si le temps dattente est supérieur à 0 et si le lock existe, on attend
if test "${waittime}" -gt 0 && test -f "${lockfile}"
then
sleep "${waittime}"
fi
# si le lock existe, on sarrête
if test -f "${lockfile}"
then
log_error_exit "lock file ${lockfile} exists"
fi
touch "${lockfile}"
}
is_too_soon() {
if is_debug; then return; fi
witness="/tmp/autosysadmin_witness_$(basename "$0")"
if test -f "${witness}"
then
compare="$(($(date +%s)-$(stat -c "%Y" "${witness}")))"
if [ "${compare}" -lt 1800 ];
then
log_error_exit 'already executed less than 30 minutes ago'
fi
rm "${witness}"
fi
touch "${witness}"
}
init_autosysadmin() {
PATH="${PATH}":/usr/sbin:/sbin↩
unset ACTIONS
SCRIPTNAME=$(basename "$0")
PROGNAME=${SCRIPTNAME%.sh}
RUN_ID="$(date +"%Y-%m-%d_%H-%M")_${SCRIPTNAME}_$(openssl rand -hex 6)"
LOG_DIR="/var/log/autosysadmin/${RUN_ID}"
mkdir -p "${LOG_DIR}"
log "Autosysadmin : Script ${SCRIPTNAME} triggered"
# Detect operating system name, version and release↩
detect_os
}
load_conf() {
# Load conf and enable script by default.
# To disable script locally, set "$PROGNAME"=off in /etc/evolinux/autosysadmin.
# To disable script globally, set "$PROGNAME"=off in the script, after load_conf() call.
declare -g "$PROGNAME"=on # dynamic variable assignment ($PROGNAME == repair_*)
# Source configuration file
# shellcheck source=../roles/deploy_autosysadmin/templates/autosysadmin.cfg.j2
test -f /etc/evolinux/autosysadmin && source /etc/evolinux/autosysadmin
}
detect_os() {
# OS detection
DEBIAN_RELEASE=""
LSB_RELEASE_BIN="$(command -v lsb_release)"
if [ -e /etc/debian_version ]; then
DEBIAN_VERSION="$(cut -d "." -f 1 < /etc/debian_version)"
if [ -x "${LSB_RELEASE_BIN}" ]; then
DEBIAN_RELEASE="$("${LSB_RELEASE_BIN}" --codename --short)"
else
case "${DEBIAN_VERSION}" in
8) DEBIAN_RELEASE="jessie";;
9) DEBIAN_RELEASE="stretch";;
10) DEBIAN_RELEASE="buster";;
11) DEBIAN_RELEASE="bullseye";;
esac
fi
fi
}
is_debian_jessie() {
test "${DEBIAN_RELEASE}" = "jessie"
}
is_debian_stretch() {
test "${DEBIAN_RELEASE}" = "stretch"
}
is_debian_buster() {
test "${DEBIAN_RELEASE}" = "buster"
}
is_debian_bullseye() {
test "${DEBIAN_RELEASE}" = "bullseye"
}
systemd_list_service_failed() {
systemctl list-units --failed --no-legend --full --type=service "$1" |
awk '{print $1}'
}
systemd_list_units_enabled() {
list_units_enabled=$(systemctl list-unit-files --state=enabled --no-legend | awk "/$1/{print \$1}")
if [ -z "${list_units_enabled}" ]
then
return 1
else
echo "${list_units_enabled}"
fi
}
format_mail_success() {
cat <<EOTEMPLATE
From: AutoSysadmin Evolix <equipe+autosysadmin@evolix.net>
Content-Type: text/plain; charset=UTF-8
MIME-Version: 1.0
Content-Transfer-Encoding: 8bit
X-Script: $(basename "$0")
X-RunId: ${RUN_ID}
To: ${EMAIL_CLIENT:-alert5@evolix.fr}
Cc: autosysadmin@evolix.fr
Subject: [autosysadmin] Intervention sur ${HOSTNAME_TEXT}
Bonjour,
Une intervention automatique vient de se terminer.
Nom du serveur : ${HOSTNAME_TEXT}
Heure d'intervention : $(LC_ALL=fr_FR.utf8 date)
### Renseignements sur l'intervention
${ACTIONS}
### Réagir à cette intervention
Vous pouvez répondre à ce message (sur l'adresse mail equipe@evolix.net).
En cas d'urgence, utilisez l'adresse maintenance@evolix.fr ou
notre téléphone portable d'astreinte (04.26.99.99.26)
--
Votre AutoSysadmin
EOTEMPLATE
}
format_mail_abort_active_users() {
cat <<EOTEMPLATE
From: AutoSysadmin Evolix <equipe+autosysadmin@evolix.net>
Content-Type: text/plain; charset=UTF-8
MIME-Version: 1.0
Content-Transfer-Encoding: 8bit
X-Script: $(basename "$0")
X-RunId: ${RUN_ID}
To: ${EMAIL_CLIENT:-alert5@evolix.fr}
Cc: autosysadmin@evolix.fr
Subject: [autosysadmin] Intervention interrompue sur ${HOSTNAME_TEXT}
Bonjour,
Une intervention automatique a été interrompue en raison
d'un utilisateur actuellement actif sur le serveur.
Nom du serveur : ${HOSTNAME_TEXT}
Heure d'intervention : $(LC_ALL=fr_FR.utf8 date)
### Utilisateur(s) connecté(s)
$(w)
--
Votre AutoSysadmin
EOTEMPLATE
}
format_mail_internal_info() {
cat <<EOTEMPLATE
From: AutoSysadmin Evolix <equipe+autosysadmin@evolix.net>
Content-Type: text/plain; charset=UTF-8
MIME-Version: 1.0
Content-Transfer-Encoding: 8bit
X-Script: $(basename "$0")
X-RunId: ${RUN_ID}
To: autosysadmin@evolix.fr
Subject: [autosysadmin] Complements (interne) - Intervention sur ${HOSTNAME_TEXT}
Bonjour,
Une intervention automatique vient de se terminer.
Nom du serveur : ${HOSTNAME_TEXT}
Heure d'intervention : $(LC_ALL=fr_FR.utf8 date)
Script déclenché : $(basename "$0")
### Actions effectuées
${ACTIONS}
### Logs autosysadmin
${INTERNAL_LOG}
### Utilisateur(s) connecté(s)
$(w)
### Informations additionnelles données par le script $(basename "$0")
${INTERNAL_INFO}
--
Votre AutoSysadmin
EOTEMPLATE
}
hook_mail() {
if is_debug; then return; fi
HOSTNAME="${HOSTNAME:-"$(get_fqdn)"}"
HOSTNAME_TEXT="$(get_complete_hostname)"
EMAIL_CLIENT="$(get_evomaintenance_mail)"
MAIL_CONTENT="$(format_mail_"$1")"
SENDMAIL_BIN="$(command -v sendmail)"
if [ -z "${SENDMAIL_BIN}" ]; then
log "No \`sendmail' command has been found, can't send mail."
fi
if [ -x "${SENDMAIL_BIN}" ]; then
echo "${MAIL_CONTENT}" | "${SENDMAIL_BIN}" -oi -t -f "equipe@evolix.net"
fi
}
# We need stable output for gcal, so we force some language environment variables
export TZ=Europe/Paris
export LANGUAGE=fr_FR.UTF-8
is_holiday() {
# gcal mark today as a holiday by surrounding with < and > the day
# of the month of that holiday line. For exemple if today is 2022-05-01 we'll
# get among other lines:
# Fête du Travail (FR) + Di, < 1>Mai 2022
# Jour de la Victoire (FR) + Di, : 8:Mai 2022 = +7 jours
gcal --cc-holidays=fr --holiday-list=short | grep -E '<[0-9 ]{2}>' --quiet
}
is_weekend() {
day_of_week=$(date +%u)
if [ "$day_of_week" != 6 ] && [ "$day_of_week" != 7 ]; then
return 1
fi
}
is_workday() {
if is_holiday || is_weekend; then
return 1
fi
}
is_worktime() {
if ! is_workday; then
return 1
fi
hour=$(date +%H)
if [ "${hour}" -lt 9 ] || { [ "${hour}" -ge 12 ] && [ "${hour}" -lt 14 ] ; } || [ "${hour}" -ge 18 ]; then
return 1
fi
}

View file

@ -1,33 +0,0 @@
#!/bin/bash
# Source functions file
# shellcheck source=./functions.sh
source /usr/share/scripts/autosysadmin/functions.sh
# shellcheck source=./restart_amavis.sh
source /usr/share/scripts/autosysadmin/restart_amavis.sh
init_autosysadmin
load_conf
test "${repair_amavis:=off}" = off && log_error_exit 'Script disabled, nothing to do here!'
# Verify if check_nrpe are not OK
check_nrpe "check_amavis" && log_error_exit 'check_amavis is OK, nothing to do here!'
# Has it recently been run?
get_argument "--no-delay" || is_too_soon
lockfile="/run/lock/repair_amavis"
cleanup() {
rm -f "${lockfile}"
}
trap 'cleanup' 0
acquire_lock_or_exit "${lockfile}"
ensure_no_active_users_or_exit
# The actual work starts below !
restart_amavis
hook_mail success
AUTOSYSADMIN=1 /usr/share/scripts/evomaintenance.sh -m "$0: done" --no-commit --no-mail

View file

@ -1,173 +0,0 @@
#!/bin/bash
# Source functions file
# shellcheck source=./functions.sh
source /usr/share/scripts/autosysadmin/functions.sh
init_autosysadmin
load_conf
test "${repair_disk:=off}" = off && log_error_exit 'Script disabled, nothing to do here!'
# Has it recently been run?
is_too_soon
lockfile="/run/lock/repair_disk"
cleanup() {
rm -f "${lockfile}"
}
trap 'cleanup' 0
acquire_lock_or_exit "${lockfile}"
ensure_no_active_users_or_exit
# The actual work starts below !
get_mountpoints() {
# the $(...) get the check_disk1 command
# the cut command selects the critical part of the check_disk1 output
# the grep command extracts the mountpoints and available disk space
# the last cut command selects the mountpoints
$(grep check_disk1 /etc/nagios/nrpe.d/evolix.cfg | cut -d'=' -f2-) -e | cut -d'|' -f1 | grep -Eo '/[[:graph:]]* [0-9]+ [A-Z][A-Z]' | cut -f1 -d' '
}
is_reserved-blocks() {
fs_type="$(findmnt -n --output=fstype "$1")"
if [ "${fs_type}" = "ext4" ];
then
device="$(findmnt -n --output=source "$1")"
reserved_block_count="$(tune2fs -l "${device}" | grep 'Reserved block count' | awk -F':' '{ gsub (" ", "", $0); print $2}')"
block_count="$(tune2fs -l "${device}" | grep 'Block count' | awk -F':' '{ gsub (" ", "", $0); print $2}')"
percentage=$(awk "BEGIN { pc=100*${reserved_block_count}/${block_count}; i=int(pc); print (pc-i<0.5)?i:i+1 }")
log "Reserved blocks for $1 is curently at $percentage%"
if [ "${percentage}" -gt "1" ]
then
log "Allowing tune2fs action to reduce the number of reserved blocks"
return 0
else
log "Reserved blocks already at or bellow 1%, no automatic action possible"
return 1
fi
else
log "Filesystem for $1 partition is not ext4"
return 1
fi
}
change_reserved-blocks() {
# We alwasy keep some reserved blocks to avoid missing some logs
# https://gitea.evolix.org/evolix/autosysadmin/issues/22
tune2fs -m 1 "$(findmnt -n --output=source "$1")"
log_action "Reserved blocks for $1 changed to 1 percent"
}
is_tmp_to_delete() {
size="$(find /var/log/ -type f -ctime +1 -exec du {} \+ | awk '{s+=$1}END{print s / 1024}')"
if [ -n "${size}" ]
then
return 0
else
return 1
fi
}
is_log_to_delete() {
size="$(find /var/log/ -type f -mtime +365 -exec du {} \+ | awk '{s+=$1}END{print s / 1024}')"
if [ -n "${size}" ]
then
return 0
else
return 1
fi
}
clean_apt_cache() {
for lxc in $(du -ax /var | sort -nr | head -n10 | grep -E '/var/lib/lxc/php[0-9]+/rootfs/var/cache$' | grep -Eo 'php[0-9]+')
do
lxc-attach --name "${lxc}" -- apt-get clean
log_action '[lxc/'"${lxc}"'] Clean apt cache'
done
case "$(du -sx /var/* | sort -rn | sed 's/^[0-9]\+[[:space:]]\+//;q')" in
'/var/cache')
apt-get clean
log_action 'Clean apt cache'
;;
esac
}
clean_amavis_virusmails() {
if du --inodes /var/lib/* | sort -n | tail -n3 | grep -q 'virusmails$'
then
find /var/lib/amavis/virusmails/ -type f -atime +30 -delete
log_action 'Clean /var/lib/amavis/virusmails'
fi
}
for mountpoint in $(get_mountpoints)
do
case "${mountpoint}" in
/var)
#if is_log_to_delete
#then
# find /var/log/ -type f -mtime +365 -delete
# log_action "$size Mo of disk space freed in /var"
#fi
if is_reserved-blocks /var
then
change_reserved-blocks /var
clean_apt_cache
clean_amavis_virusmails
hook_mail success
fi
;;
/tmp)
#if is_tmp_to_delete
#then
# find /tmp/ -type f -ctime +1 -delete
# log_action "$size Mo of disk space freed in /tmp"
#fi
if is_reserved-blocks /tmp
then
change_reserved-blocks /tmp
hook_mail success
fi
;;
/home)
if is_reserved-blocks /home
then
change_reserved-blocks /home
hook_mail success
fi
;;
/srv)
if is_reserved-blocks /srv
then
change_reserved-blocks /srv
hook_mail success
fi
;;
/filer)
if is_reserved-blocks /filer
then
change_reserved-blocks /filer
hook_mail success
fi
;;
/)
if is_reserved-blocks /
then
change_reserved-blocks /
hook_mail success
# Suggest remove old kernel ?
fi
;;
*)
# unknown
log 'Unknown partition (or weird case) or nothing to do'
;;
esac
done
AUTOSYSADMIN=1 /usr/share/scripts/evomaintenance.sh -m "$0: done" --no-commit --no-mail

View file

@ -1,57 +0,0 @@
#!/bin/bash
# Source functions file
# shellcheck source=./functions.sh
source /usr/share/scripts/autosysadmin/functions.sh
init_autosysadmin
load_conf
test "${repair_elasticsearch:=off}" = off && log_error_exit 'Script disabled, nothing to do here!'
# Has it recently been run?
is_too_soon
lockfile="/run/lock/repair_elasticsearch"
cleanup() {
rm -f "${lockfile}"
}
trap 'cleanup' 0
acquire_lock_or_exit "${lockfile}"
ensure_no_active_users_or_exit
# The actual work starts below !
elasticsearch_is_enabled() {
systemd_list_units_enabled "elasticsearch.service"
}
elasticsearch_restart() {
if ! timeout 60 systemctl restart elasticsearch.service > /dev/null
then
log_error_exit 'failed to restart elasticsearch'
fi
}
# Test functions
test_elasticsearch_process_present() {
pgrep -u elasticsearch > /dev/null
}
if elasticsearch_is_enabled
then
if ! test_elasticsearch_process_present
then
log_action "Redémarrage de elasticsearch"
elasticsearch_restart
hook_mail success
else
log_error_exit "Elasticsearch process alive. Aborting"
fi
else
log_error_exit "Elasticsearch is not enabled. Aborting"
fi
AUTOSYSADMIN=1 /usr/share/scripts/evomaintenance.sh -m "$0: done" --no-commit --no-mail

View file

@ -1,141 +0,0 @@
#!/bin/bash
# Source functions file
# shellcheck source=./functions.sh
source /usr/share/scripts/autosysadmin/functions.sh
init_autosysadmin
load_conf
test "${repair_http:=off}" = off && log_error_exit 'Script disabled, nothing to do here!'
# Has it recently been run?
is_too_soon
lockfile="/run/lock/repair_http"
cleanup() {
rm -f "${lockfile}"
}
trap 'cleanup' 0
acquire_lock_or_exit "${lockfile}"
ensure_no_active_users_or_exit
# The actual work starts below !
log_system_status
http_detect_service() {
# check whether nginx, apache or both are supposed to be running
if is_debian_jessie; then
find /etc/rc2.d/
else
systemctl list-unit-files --state=enabled
fi | awk '/nginx/ { nginx = 1 } /apache2/ { apache2 = 1 } END { if (nginx && apache2) { print "both" } else if (nginx) { print "nginx" } else if (apache2) { print "apache2" } }'
# The previous awk command looks for two patterns: "nginx"
# and "apache2". If a line matches the patterns, a variable
# "nginx" or "apache2" is set to 1 (true). The "END" checks
# if one or both patterns has been found.
}
http_handle_apache() {
# check syntax
if ! apache2ctl -t > /dev/null 2> /dev/null
then
log_error_exit 'apache2 configuration syntax is not valid'
fi
# try restart
if ! timeout 20 systemctl restart apache2.service > /dev/null 2> /dev/null
then
log_error_exit 'failed to restart apache2'
fi
log_action "Redémarrage de Apache"
internal_info "#### grep $(LANG=en_US.UTF-8 date '+%b %d') /home/*/log/error.log /var/log/apache2/*error.log (avec filtrage)"
ERROR_LOG=$(grep "$(LANG=en_US.UTF-8 date '+%b %d')" /home/*/log/error.log /var/log/apache2/*error.log | grep -v -e "Got error 'PHP message:" -e "No matching DirectoryIndex" -e "client denied by server configuration" -e "server certificate does NOT include an ID which matches the server name" )
internal_info "$ERROR_LOG"
}
http_handle_nginx() {
# check syntax
if ! nginx -t > /dev/null 2> /dev/null
then
log_error_exit 'nginx configuration syntax is not valid'
fi
# try restart
if ! timeout 20 systemctl restart nginx.service > /dev/null 2> /dev/null
then
log_error_exit 'failed to restart nginx'
fi
log_action "Redémarrage de Nginx"
}
http_handle_lxc_php() {
# check whether containers are used for PHP and reboot them if so
if systemd_list_units_enabled 'lxc'
then
for php in $(lxc-ls | grep 'php'); do
lxc-stop -n "$php"
lxc-start --daemon -n "$php"
log_action "lxc-fpm - Redémarrage container ${php}"
done
fi
}
http_handle_fpm_php() {
# check whether php-fpm is installed and restart it if so
if enabled_units="$(systemd_list_units_enabled "php.*-fpm")"
then
systemctl restart "${enabled_units}"
log_action 'php-fpm - Redémarrage de php-fpm'
fi
}
case "$(http_detect_service)" in
nginx)
http_handle_nginx
http_handle_lxc_php
http_handle_fpm_php
hook_mail success
hook_mail internal_info
;;
apache2)
http_handle_apache
http_handle_lxc_php
http_handle_fpm_php
hook_mail success
hook_mail internal_info
;;
both)
http_handle_nginx
http_handle_apache
http_handle_lxc_php
http_handle_fpm_php
hook_mail success
hook_mail internal_info
;;
*)
# unknown
log 'nothing to do'
;;
esac
AUTOSYSADMIN=1 /usr/share/scripts/evomaintenance.sh -m "$0: done" --no-commit --no-mail

View file

@ -1,71 +0,0 @@
#!/bin/bash
# Source functions file
# shellcheck source=./functions.sh
source /usr/share/scripts/autosysadmin/functions.sh
init_autosysadmin
load_conf
test "${repair_mysql:=off}" = off && log_error_exit 'Script disabled, nothing to do here!'
# Has it recently been run?
is_too_soon
lockfile="/run/lock/repair_mysql"
cleanup() {
rm -f "${lockfile}"
}
trap 'cleanup' 0
acquire_lock_or_exit "${lockfile}"
ensure_no_active_users_or_exit
# The actual work starts below !
log_system_status
mysql_is_enabled() {
if is_debian_jessie
then
find /etc/rc2.d/ -name '*mysql*' > /dev/null
else
systemd_list_units_enabled "mysql.service"
fi
}
mysql_restart() {
if is_debian_jessie
then
if ! timeout 60 /etc/init.d/mysql restart > /dev/null
then
log_error_exit 'failed to restart mysql'
fi
else
if ! timeout 60 systemctl restart mysql.service > /dev/null
then
log_error_exit 'failed to restart mysql'
fi
fi
}
# Test functions
test_mysql_process_present() {
pgrep -u mysql mysqld > /dev/null
}
if mysql_is_enabled
then
if ! test_mysql_process_present
then
log_action "Redémarrage de MySQL"
mysql_restart
hook_mail success
else
log_error_exit "mysqld process alive. Aborting"
fi
else
log_error_exit "MySQL/MariaDB not enabled. Aborting"
fi
AUTOSYSADMIN=1 /usr/share/scripts/evomaintenance.sh -m "$0: done" --no-commit --no-mail

View file

@ -1,61 +0,0 @@
#!/bin/bash
# Source functions file
# shellcheck source=./functions.sh
source /usr/share/scripts/autosysadmin/functions.sh
init_autosysadmin
load_conf
test "${repair_opendkim:=off}" = off && log_error_exit 'Script disabled, nothing to do here!'
# Has it recently been run?
is_too_soon
lockfile="/run/lock/repair_opendkim"
cleanup() {
rm -f "${lockfile}"
}
trap 'cleanup' 0
acquire_lock_or_exit "${lockfile}"
ensure_no_active_users_or_exit
log_system_status
# Functions dedicated to this repair script
opendkim_is_enabled() {
systemd_list_units_enabled "opendkim.service"
}
opendkim_restart() {
if ! timeout 60 systemctl restart opendkim.service > /dev/null
then
log_error_exit 'failed to restart opendkim'
fi
}
opendkim_test_process_present() {
pgrep -u opendkim > /dev/null
}
# Main logic
if opendkim_is_enabled
then
if ! opendkim_test_process_present
then
log_action "Redémarrage de opendkim"
opendkim_restart
hook_mail success
else
log_error_exit "opendkim process alive. Aborting"
fi
else
log_error_exit "opendkim is not enabled. Aborting"
fi
AUTOSYSADMIN=1 /usr/share/scripts/evomaintenance.sh -m "$0: done" --no-commit --no-mail

View file

@ -1,53 +0,0 @@
#!/bin/bash
# Source functions file
# shellcheck source=./functions.sh
source /usr/share/scripts/autosysadmin/functions.sh
init_autosysadmin
load_conf
test "${repair_php_fpm56:=off}" = off && log_error_exit 'Script disabled, nothing to do here!'
# Has it recently been run?
is_too_soon
lockfile="/run/lock/repair_http"
cleanup() {
rm -f "${lockfile}"
}
trap 'cleanup' 0
acquire_lock_or_exit "${lockfile}" 15s
ensure_no_active_users_or_exit
# The actual work starts below !
log_system_status
log_check_php_fpm
if systemd_list_units_enabled 'lxc'
then
if lxc-ls | grep -q php56
then
lxc-stop -n php56
lxc-start --daemon -n php56
log_action "lxc-fpm - Redémarrage container php56"
internal_info "#### tail /var/lib/lxc/php56/rootfs/var/log/php5-fpm.log"
FPM_LOG=$(tail /var/lib/lxc/php56/rootfs/var/log/php5-fpm.log)
internal_info "$FPM_LOG" "$(read_log_system_status)"
hook_mail success
hook_mail internal_info
else
log 'Not possible :v'
fi
else
log 'Error, not a multi-php install'
fi
AUTOSYSADMIN=1 /usr/share/scripts/evomaintenance.sh -m "$0: done" --no-commit --no-mail

View file

@ -1,53 +0,0 @@
#!/bin/bash
# Source functions file
# shellcheck source=./functions.sh
source /usr/share/scripts/autosysadmin/functions.sh
init_autosysadmin
load_conf
test "${repair_php_fpm70:=off}" = off && log_error_exit 'Script disabled, nothing to do here!'
# Has it recently been run?
is_too_soon
lockfile="/run/lock/repair_http"
cleanup() {
rm -f "${lockfile}"
}
trap 'cleanup' 0
acquire_lock_or_exit "${lockfile}" 15s
ensure_no_active_users_or_exit
# The actual work starts below !
log_system_status
log_check_php_fpm
if systemd_list_units_enabled 'lxc'
then
if lxc-ls | grep -q php70
then
lxc-stop -n php70
lxc-start --daemon -n php70
log_action "lxc-fpm - Redémarrage container php70"
internal_info "#### tail /var/lib/lxc/php70/rootfs/var/log/php7.0-fpm.log"
FPM_LOG=$(tail /var/lib/lxc/php70/rootfs/var/log/php7.0-fpm.log)
internal_info "$FPM_LOG" "$(read_log_system_status)"
hook_mail success
hook_mail internal_info
else
log 'Not possible :v'
fi
else
log 'Error, not a multi-php install'
fi
AUTOSYSADMIN=1 /usr/share/scripts/evomaintenance.sh -m "$0: done" --no-commit --no-mail

View file

@ -1,53 +0,0 @@
#!/bin/bash
# Source functions file
# shellcheck source=./functions.sh
source /usr/share/scripts/autosysadmin/functions.sh
init_autosysadmin
load_conf
test "${repair_php_fpm73:=off}" = off && log_error_exit 'Script disabled, nothing to do here!'
# Has it recently been run?
is_too_soon
lockfile="/run/lock/repair_http"
cleanup() {
rm -f "${lockfile}"
}
trap 'cleanup' 0
acquire_lock_or_exit "${lockfile}" 15s
ensure_no_active_users_or_exit
# The actual work starts below !
log_system_status
log_check_php_fpm
if systemd_list_units_enabled 'lxc'
then
if lxc-ls | grep -q php73
then
lxc-stop -n php73
lxc-start --daemon -n php73
log_action "lxc-fpm - Redémarrage container php73"
internal_info "#### tail /var/lib/lxc/php73/rootfs/var/log/php7.3-fpm.log"
FPM_LOG=$(tail /var/lib/lxc/php73/rootfs/var/log/php7.3-fpm.log)
internal_info "$FPM_LOG" "$(read_log_system_status)"
hook_mail success
hook_mail internal_info
else
log 'Not possible :v'
fi
else
log 'Error, not a multi-php install'
fi
AUTOSYSADMIN=1 /usr/share/scripts/evomaintenance.sh -m "$0: done" --no-commit --no-mail

View file

@ -1,53 +0,0 @@
#!/bin/bash
# Source functions file
# shellcheck source=./functions.sh
source /usr/share/scripts/autosysadmin/functions.sh
init_autosysadmin
load_conf
test "${repair_php_fpm74:=off}" = off && log_error_exit 'Script disabled, nothing to do here!'
# Has it recently been run?
is_too_soon
lockfile="/run/lock/repair_http"
cleanup() {
rm -f "${lockfile}"
}
trap 'cleanup' 0
acquire_lock_or_exit "${lockfile}" 15s
ensure_no_active_users_or_exit
# The actual work starts below !
log_system_status
log_check_php_fpm
if systemd_list_units_enabled 'lxc'
then
if lxc-ls | grep -q php74
then
lxc-stop -n php74
lxc-start --daemon -n php74
log_action "lxc-fpm - Redémarrage container php74"
internal_info "#### tail /var/lib/lxc/php74/rootfs/var/log/php7.4-fpm.log"
FPM_LOG=$(tail /var/lib/lxc/php74/rootfs/var/log/php7.4-fpm.log)
internal_info "$FPM_LOG" "$(read_log_system_status)"
hook_mail success
hook_mail internal_info
else
log 'Not possible :v'
fi
else
log 'Error, not a multi-php install'
fi
AUTOSYSADMIN=1 /usr/share/scripts/evomaintenance.sh -m "$0: done" --no-commit --no-mail

View file

@ -1,53 +0,0 @@
#!/bin/bash
# Source functions file
# shellcheck source=./functions.sh
source /usr/share/scripts/autosysadmin/functions.sh
init_autosysadmin
load_conf
test "${repair_php_fpm80:=off}" = off && log_error_exit 'Script disabled, nothing to do here!'
# Has it recently been run?
is_too_soon
lockfile="/run/lock/repair_http"
cleanup() {
rm -f "${lockfile}"
}
trap 'cleanup' 0
acquire_lock_or_exit "${lockfile}" 15s
ensure_no_active_users_or_exit
# The actual work starts below !
log_system_status
log_check_php_fpm
if systemd_list_units_enabled 'lxc'
then
if lxc-ls | grep -q php80
then
lxc-stop -n php80
lxc-start --daemon -n php80
log_action "lxc-fpm - Redémarrage container php80"
internal_info "#### tail /var/lib/lxc/php80/rootfs/var/log/php8.0-fpm.log"
FPM_LOG=$(tail /var/lib/lxc/php80/rootfs/var/log/php8.0-fpm.log)
internal_info "$FPM_LOG" "$(read_log_system_status)"
hook_mail success
hook_mail internal_info
else
log 'Not possible :v'
fi
else
log 'Error, not a multi-php install'
fi
AUTOSYSADMIN=1 /usr/share/scripts/evomaintenance.sh -m "$0: done" --no-commit --no-mail

View file

@ -1,53 +0,0 @@
#!/bin/bash
# Source functions file
# shellcheck source=./functions.sh
source /usr/share/scripts/autosysadmin/functions.sh
init_autosysadmin
load_conf
test "${repair_php_fpm81:=off}" = off && log_error_exit 'Script disabled, nothing to do here!'
# Has it recently been run?
is_too_soon
lockfile="/run/lock/repair_http"
cleanup() {
rm -f "${lockfile}"
}
trap 'cleanup' 0
acquire_lock_or_exit "${lockfile}" 15s
ensure_no_active_users_or_exit
# The actual work starts below !
log_system_status
log_check_php_fpm
if systemd_list_units_enabled 'lxc'
then
if lxc-ls | grep -q php81
then
lxc-stop -n php81
lxc-start --daemon -n php81
log_action "lxc-fpm - Redémarrage container php81"
internal_info "#### tail /var/lib/lxc/php81/rootfs/var/log/php8.1-fpm.log"
FPM_LOG=$(tail /var/lib/lxc/php81/rootfs/var/log/php8.1-fpm.log)
internal_info "$FPM_LOG" "$(read_log_system_status)"
hook_mail success
hook_mail internal_info
else
log 'Not possible :v'
fi
else
log 'Error, not a multi-php install'
fi
AUTOSYSADMIN=1 /usr/share/scripts/evomaintenance.sh -m "$0: done" --no-commit --no-mail

View file

@ -1,53 +0,0 @@
#!/bin/bash
# Source functions file
# shellcheck source=./functions.sh
source /usr/share/scripts/autosysadmin/functions.sh
init_autosysadmin
load_conf
test "${repair_php_fpm82:=off}" = off && log_error_exit 'Script disabled, nothing to do here!'
# Has it recently been run?
is_too_soon
lockfile="/run/lock/repair_http"
cleanup() {
rm -f "${lockfile}"
}
trap 'cleanup' 0
acquire_lock_or_exit "${lockfile}" 15s
ensure_no_active_users_or_exit
# The actual work starts below !
log_system_status
log_check_php_fpm
if systemd_list_units_enabled 'lxc'
then
if lxc-ls | grep -q php82
then
lxc-stop -n php82
lxc-start --daemon -n php82
log_action "lxc-fpm - Redémarrage container php82"
internal_info "#### tail /var/lib/lxc/php82/rootfs/var/log/php8.2-fpm.log"
FPM_LOG=$(tail /var/lib/lxc/php82/rootfs/var/log/php8.2-fpm.log)
internal_info "$FPM_LOG" "$(read_log_system_status)"
hook_mail success
hook_mail internal_info
else
log 'Not possible :v'
fi
else
log 'Error, not a multi-php install'
fi
AUTOSYSADMIN=1 /usr/share/scripts/evomaintenance.sh -m "$0: done" --no-commit --no-mail

View file

@ -1,53 +0,0 @@
#!/bin/bash
# Source functions file
# shellcheck source=./functions.sh
source /usr/share/scripts/autosysadmin/functions.sh
init_autosysadmin
load_conf
test "${repair_php_fpm83:=off}" = off && log_error_exit 'Script disabled, nothing to do here!'
# Has it recently been run?
is_too_soon
lockfile="/run/lock/repair_http"
cleanup() {
rm -f "${lockfile}"
}
trap 'cleanup' 0
acquire_lock_or_exit "${lockfile}" 15s
ensure_no_active_users_or_exit
# The actual work starts below !
log_system_status
log_check_php_fpm
if systemd_list_units_enabled 'lxc'
then
if lxc-ls | grep -q php83
then
lxc-stop -n php83
lxc-start --daemon -n php83
log_action "lxc-fpm - Redémarrage container php83"
internal_info "#### tail /var/lib/lxc/php83/rootfs/var/log/php8.3-fpm.log"
FPM_LOG=$(tail /var/lib/lxc/php83/rootfs/var/log/php8.3-fpm.log)
internal_info "$FPM_LOG" "$(read_log_system_status)"
hook_mail success
hook_mail internal_info
else
log 'Not possible :v'
fi
else
log 'Error, not a multi-php install'
fi
AUTOSYSADMIN=1 /usr/share/scripts/evomaintenance.sh -m "$0: done" --no-commit --no-mail

View file

@ -1,58 +0,0 @@
#!/bin/bash
# Source functions file
# shellcheck source=./functions.sh
source /usr/share/scripts/autosysadmin/functions.sh
init_autosysadmin
load_conf
test "${repair_redis:=off}" = off && log_error_exit 'Script disabled, nothing to do here!'
# Has it recently been run?
is_too_soon
lockfile="/run/lock/repair_redis"
cleanup() {
rm -f "${lockfile}"
}
trap 'cleanup' 0
acquire_lock_or_exit "${lockfile}"
ensure_no_active_users_or_exit
# The actual work starts below !
handle_redis() {
for service in $(systemd_list_service_failed redis*)
do
# ne rien faire si le service est désactivé
if ! systemctl is-enabled --quiet "${service}"
then
continue
fi
# ne rien faire si le service est actif
if systemctl is-active --quiet "${service}"
then
continue
fi
if ! timeout 20 systemctl restart redis.service > /dev/null 2> /dev/null
then
log_error_exit "failed to restart redis ${service}"
fi
log_action "Redémarrer service ${service}"
done
}
if ( systemd_list_units_enabled 'redis.*\.service$' ) > /dev/null
then
handle_redis
hook_mail success
else
log 'Error: redis service is not enabled'
fi
AUTOSYSADMIN=1 /usr/share/scripts/evomaintenance.sh -m "$0: done" --no-commit --no-mail

View file

@ -1,63 +0,0 @@
#!/bin/bash
# Source functions file
# shellcheck source=./functions.sh
source /usr/share/scripts/autosysadmin/functions.sh
init_autosysadmin
load_conf
# Comment this line to enable
repair_template=off
test "${repair_template:=off}" = off && log_error_exit 'Script disabled, nothing to do here!'
# Has it recently been run?
is_too_soon
lockfile="/run/lock/repair_template"
cleanup() {
rm -f "${lockfile}"
}
trap 'cleanup' 0
acquire_lock_or_exit "${lockfile}"
ensure_no_active_users_or_exit
log_system_status
# Functions dedicated to this repair script
template_is_enabled() {
systemd_list_units_enabled "template.service"
}
template_restart() {
if ! timeout 60 systemctl restart template.service > /dev/null
then
log_error_exit 'failed to restart template'
fi
}
template_test_process_present() {
pgrep -u template > /dev/null
}
# Main logic
if template_is_enabled
then
if ! template_test_process_present
then
log_action "Redémarrage de template"
template_restart
hook_mail success
else
log_error_exit "template process alive. Aborting"
fi
else
log_error_exit "template is not enabled. Aborting"
fi
AUTOSYSADMIN=1 /usr/share/scripts/evomaintenance.sh -m "$0: done" --no-commit --no-mail

View file

@ -1,35 +0,0 @@
#!/bin/bash
restart_amavis() {
/etc/init.d/amavis stop 2>/dev/null
/etc/init.d/clamav-freshclam stop 2>/dev/null
/etc/init.d/clamav-daemon stop 2>/dev/null
if systemctl is-enabled --quiet 'clamav-freshclam.service'
then
freshclam
log_action "Mise à jour des définitions antivirus"
fi
if systemctl is-enabled --quiet 'clamav-daemon.service'
then
/etc/init.d/clamav-daemon start
log_action "Redémarrage de clamav-daemon"
else
log 'Error, clamav not installed'
fi
if systemctl is-enabled --quiet 'clamav-freshclam.service'
then
/etc/init.d/clamav-freshclam start
log_action "Redémarrage de clamav-freshclam"
fi
if systemctl is-enabled --quiet 'amavis.service'
then
/etc/init.d/amavis start
log_action "Redémarrage de amavis"
else
log 'Error, amavis not installed'
fi
}

View file

@ -1,16 +0,0 @@
---
- name: Restart nagios-nrpe-server
ansible.builtin.service:
name: nagios-nrpe-server
state: restarted
- name: Restart nrpe
ansible.builtin.service:
name: nrpe
state: restarted
- name: Restart rsyslog
ansible.builtin.service:
name: rsyslog
state: restarted

View file

@ -1,61 +0,0 @@
---
- name: "Remount /usr if needed"
ansible.builtin.import_role:
name: remount-usr
- name: Create autosysadmin directory
ansible.builtin.file:
path: "{{ autosysadmin_dir }}"
state: directory
owner: "root"
group: "root"
mode: "0750"
tags:
- autosysadmin
- name: Copy scripts
ansible.builtin.copy:
src: "files/scripts/{{ item }}"
dest: "{{ autosysadmin_dir }}/{{ item }}"
owner: root
group: root
mode: "0750"
loop:
- "functions.sh"
- "restart_amavis.sh"
- "repair_amavis.sh"
- "repair_disk.sh"
- "repair_elasticsearch.sh"
- "repair_http.sh"
- "repair_mysql.sh"
- "repair_php_fpm56.sh"
- "repair_php_fpm70.sh"
- "repair_php_fpm73.sh"
- "repair_php_fpm74.sh"
- "repair_php_fpm80.sh"
- "repair_php_fpm81.sh"
- "repair_php_fpm82.sh"
- "repair_php_fpm83.sh"
- "repair_tomcat_instance.sh"
tags:
- autosysadmin
- name: Ensure /etc/evolinux folder exists
ansible.builtin.file:
path: "/etc/evolinux"
state: directory
owner: "root"
group: "root"
mode: "0700"
tags:
- autosysadmin
- name: Copy the configuration file
ansible.builtin.template:
src: "autosysadmin.cf.j2"
dest: "/etc/evolinux/autosysadmin"
owner: root
group: root
mode: "0640"
tags:
- autosysadmin

View file

@ -1,37 +0,0 @@
---
- name: Install dependencies
ansible.builtin.import_tasks: dependencies.yml
tags:
- autosysadmin
- name: Install autosysadmin scripts
ansible.builtin.import_tasks: autosysadmin_scripts.yml
tags:
- autosysadmin
- name: Amend NRPE configuration
ansible.builtin.import_tasks: nrpe.yml
tags:
- autosysadmin
- name: Amend sudo configuration
ansible.builtin.import_tasks: sudo.yml
tags:
- autosysadmin
- name: Amend rsyslog configuration
ansible.builtin.import_tasks: rsyslog.yml
tags:
- autosysadmin
- name: Amend logrotate configuration
ansible.builtin.import_tasks: logrotate.yml
tags:
- autosysadmin
- name: Install last version of dump-server-state
ansible.builtin.import_role:
name: evolinux-base
tasks_from: dump-server-state.yml
tags:
- autosysadmin

View file

@ -1,11 +0,0 @@
---
- name: Custom configuration is present
ansible.builtin.template:
src: autosysadmin.cfg.j2
dest: /etc/nagios/nrpe.d/autosysadmin.cfg
group: nagios
mode: "0640"
force: true
notify: Restart nagios-nrpe-server
tags:
- autosysadmin

View file

@ -1,74 +0,0 @@
#
# Ansible managed - DO NOT MODIFY, your changes will be **overwritten** !
#
# Update the hosts_vars/group_vars on the autosysadmin project
# https://gitea.evolix.org/evolix/autosysadmin/src/branch/master
#
# Configuration for autosysadmin
# Use this file to change configuration values defined in repair scripts
# Ex : repair_http=off
{% if repair_amavis == "off" %}
repair_amavis=off
{% endif %}
{% if repair_disk == "off" %}
repair_disk=off
{% endif %}
{% if repair_elasticsearch == "off" %}
repair_elasticsearch=off
{% endif %}
{% if repair_http == "off" %}
repair_http=off
{% endif %}
{% if repair_mysql == "off" %}
repair_mysql=off
{% endif %}
{% if repair_opendkim == "off" %}
repair_opendkim=off
{% endif %}
{% if repair_php_fpm56 == "off" %}
repair_php_fpm56=off
{% endif %}
{% if repair_php_fpm70 == "off" %}
repair_php_fpm70=off
{% endif %}
{% if repair_php_fpm73 == "off" %}
repair_php_fpm73=off
{% endif %}
{% if repair_php_fpm74 == "off" %}
repair_php_fpm74=off
{% endif %}
{% if repair_php_fpm80 == "off" %}
repair_php_fpm80=off
{% endif %}
{% if repair_php_fpm81 == "off" %}
repair_php_fpm81=off
{% endif %}
{% if repair_php_fpm82 == "off" %}
repair_php_fpm82=off
{% endif %}
{% if repair_php_fpm83 == "off" %}
repair_php_fpm83=off
{% endif %}
{% if repair_redis == "off" %}
repair_redis=off
{% endif %}
{% if repair_tomcat_instance == "off" %}
repair_tomcat_instance=off
{% endif %}

View file

@ -1,22 +0,0 @@
#
# Ansible managed - DO NOT MODIFY, your changes will be overwritten !
#
# Autosysadmin repair commands
command[repair_amavis]=sudo {{ autosysadmin_dir }}/repair_amavis.sh
command[repair_disk]=sudo {{ autosysadmin_dir }}/repair_disk.sh
command[repair_elasticsearch]=sudo {{ autosysadmin_dir }}/repair_elasticsearch.sh
command[repair_http]=sudo {{ autosysadmin_dir }}/repair_http.sh
command[repair_mysql]=sudo {{ autosysadmin_dir }}/repair_mysql.sh
command[repair_opendkim]=sudo {{ autosysadmin_dir }}/repair_opendkim.sh
command[repair_php_fpm56]=sudo {{ autosysadmin_dir }}/repair_php_fpm56.sh
command[repair_php_fpm70]=sudo {{ autosysadmin_dir }}/repair_php_fpm70.sh
command[repair_php_fpm73]=sudo {{ autosysadmin_dir }}/repair_php_fpm73.sh
command[repair_php_fpm74]=sudo {{ autosysadmin_dir }}/repair_php_fpm74.sh
command[repair_php_fpm80]=sudo {{ autosysadmin_dir }}/repair_php_fpm80.sh
command[repair_php_fpm81]=sudo {{ autosysadmin_dir }}/repair_php_fpm81.sh
command[repair_php_fpm82]=sudo {{ autosysadmin_dir }}/repair_php_fpm82.sh
command[repair_php_fpm83]=sudo {{ autosysadmin_dir }}/repair_php_fpm83.sh
command[repair_redis]=sudo {{ autosysadmin_dir }}/repair_redis.sh
command[repair_tomcat_instance]=sudo {{ autosysadmin_dir }}/repair_tomcat_instance.sh

View file

@ -1,21 +0,0 @@
#
# Ansible managed - DO NOT MODIFY, your changes will be overwritten !
#
nagios ALL = NOPASSWD: {{ autosysadmin_dir }}/repair_amavis.sh
nagios ALL = NOPASSWD: {{ autosysadmin_dir }}/repair_disk.sh
nagios ALL = NOPASSWD: {{ autosysadmin_dir }}/repair_elasticsearch.sh
nagios ALL = NOPASSWD: {{ autosysadmin_dir }}/repair_http.sh
nagios ALL = NOPASSWD: {{ autosysadmin_dir }}/repair_mysql.sh
nagios ALL = NOPASSWD: {{ autosysadmin_dir }}/repair_opendkim.sh
nagios ALL = NOPASSWD: {{ autosysadmin_dir }}/repair_php_fpm56.sh
nagios ALL = NOPASSWD: {{ autosysadmin_dir }}/repair_php_fpm70.sh
nagios ALL = NOPASSWD: {{ autosysadmin_dir }}/repair_php_fpm73.sh
nagios ALL = NOPASSWD: {{ autosysadmin_dir }}/repair_php_fpm74.sh
nagios ALL = NOPASSWD: {{ autosysadmin_dir }}/repair_php_fpm80.sh
nagios ALL = NOPASSWD: {{ autosysadmin_dir }}/repair_php_fpm81.sh
nagios ALL = NOPASSWD: {{ autosysadmin_dir }}/repair_php_fpm82.sh
nagios ALL = NOPASSWD: {{ autosysadmin_dir }}/repair_php_fpm83.sh
nagios ALL = NOPASSWD: {{ autosysadmin_dir }}/repair_redis.sh
nagios ALL = NOPASSWD: {{ autosysadmin_dir }}/repair_tomcat_instance.sh

View file

@ -39,8 +39,8 @@ concat_files() {
chown root: "${hapee_cert_file}"
}
cert_and_key_mismatch() {
hapee_cert_md5=$(openssl x509 -noout -modulus -in "${hapee_cert_file}" | openssl md5)
hapee_key_md5=$(openssl rsa -noout -modulus -in "${hapee_cert_file}" | openssl md5)
hapee_cert_md5=$(openssl x509 -noout -pubkey -in "${hapee_cert_file}" | openssl md5)
hapee_key_md5=$(openssl pkey -pubout -in "${hapee_cert_file}" | openssl md5)
test "${hapee_cert_md5}" != "${hapee_key_md5}"
}

View file

@ -29,8 +29,8 @@ concat_files() {
chown root: "${haproxy_cert_file}"
}
cert_and_key_mismatch() {
haproxy_cert_md5=$(openssl x509 -noout -modulus -in "${haproxy_cert_file}" | openssl md5)
haproxy_key_md5=$(openssl rsa -noout -modulus -in "${haproxy_cert_file}" | openssl md5)
haproxy_cert_md5=$(openssl x509 -noout -pubkey -in "${haproxy_cert_file}" | openssl md5)
haproxy_key_md5=$(openssl pkey -pubout -in "${haproxy_cert_file}" | openssl md5)
test "${haproxy_cert_md5}" != "${haproxy_key_md5}"
}

View file

@ -0,0 +1,44 @@
#!/bin/sh
error() {
>&2 echo "${PROGNAME}: $1"
exit 1
}
debug() {
if [ "${VERBOSE}" = "1" ] && [ "${QUIET}" != "1" ]; then
>&2 echo "${PROGNAME}: $1"
fi
}
daemon_found_and_running() {
test -n "$(pidof nrpe)"
}
letsencrypt_lineaged_used() {
grep -r "^ssl_cert_file" /etc/nagios/ | grep "letsencrypt" | grep -q "$(basename "${RENEWED_LINEAGE}")"
}
copy_letsencrypt_cert() {
DEST_CERTIFICATE=$(grep -r "^ssl_cert_file" /etc/nagios/ | awk -F'=' '{print $2}')
DEST_PRIVATE_KEY=$(grep -r "^ssl_privatekey_file" /etc/nagios/ | awk -F'=' '{print $2}')
install --mode 440 --group nagios ${RENEWED_LINEAGE}/fullchain.pem ${DEST_CERTIFICATE}
install --mode 440 --group nagios ${RENEWED_LINEAGE}/privkey.pem ${DEST_PRIVATE_KEY}
}
main() {
if daemon_found_and_running; then
if letsencrypt_lineaged_used; then
debug "NRPE detected... Copying certificates to the right place & permissions"
copy_letsencrypt_cert
debug "Restarting NRPE"
systemctl restart nagios-nrpe-server
else
debug "NRPE doesn't use the given Let's Encrypt certificate. Skip."
fi
else
debug "NRPE is not running or missing. Skip."
fi
}
readonly PROGNAME=$(basename "$0")
readonly VERBOSE=${VERBOSE:-"0"}
readonly QUIET=${QUIET:-"0"}
main

View file

@ -3,28 +3,25 @@
docker_home: /var/lib/docker
docker_tmpdir: "{{ docker_home }}/tmp"
# Chose to use iptables instead of docker-proxy userland process
docker_conf_use_iptables: False
# Disable the possibility for containers processes to gain new privileges
docker_conf_no_newprivileges: False
docker_conf_no_newprivileges: false
# Toggle live restore (need to be disabled in swarm mode)
docker_conf_live_restore: True
docker_conf_live_restore: true
# Toggle user namespace
docker_conf_user_namespace: True
docker_conf_user_namespace: true
# Disable all default network connectivity
docker_conf_disable_default_networking: False
docker_conf_disable_default_networking: false
# Remote access
docker_remote_access_enabled: False
docker_remote_access_enabled: false
docker_daemon_port: 2376
docker_daemon_listening_ip: 0.0.0.0
# TLS
docker_tls_enabled: False
docker_tls_enabled: false
docker_tls_path: "{{ docker_home }}/tls"
docker_tls_ca: ca/ca.pem
docker_tls_ca_key: ca/ca-key.pem
@ -32,4 +29,4 @@ docker_tls_cert: server/cert.pem
docker_tls_key: server/key.pem
docker_tls_csr: server/server.csr
apt_keyring_dir: "{{ ansible_distribution_major_version is version('12', '<') | ternary('/etc/apt/trusted.gpg.d', '/etc/apt/keyrings') }}"
apt_keyring_dir: "{{ ansible_distribution_major_version is version('12', '<') | ternary('/etc/apt/trusted.gpg.d', '/etc/apt/keyrings') }}"

View file

@ -1,5 +1,14 @@
# This role installs the docker daemon
---
- name: Fail if docker_conf_use_iptables is defined
ansible.builtin.fail:
msg: "Variable docker_conf_use_iptables is deprecated and not configurable anymore. Please remove it from your variables. Also double-check the daemon.json config for docker"
when:
- docker_conf_use_iptables is defined
tags:
- always
- name: Remove older docker packages
ansible.builtin.apt:
name:
@ -23,7 +32,7 @@
when: ansible_distribution_major_version is version('10', '<')
- name: "Ensure {{ apt_keyring_dir }} directory exists"
file:
ansible.builtin.file:
path: "{{ apt_keyring_dir }}"
state: directory
mode: "755"
@ -44,35 +53,34 @@
repo: 'deb [signed-by={{ apt_keyring_dir }}/docker-debian.asc] https://download.docker.com/linux/debian {{ ansible_distribution_release }} stable'
filename: docker
state: present
update_cache: yes
update_cache: true
when: ansible_distribution_major_version is version('12', '<')
- name: Add Docker repository (Debian >=12)
ansible.builtin.template:
src: docker.sources.j2
dest: /etc/apt/sources.list.d/docker.sources
register: docker_sources
owner: root
group: root
mode: "0644"
when: ansible_distribution_major_version is version('12', '>=')
- name: Update APT cache
ansible.builtin.apt:
update_cache: yes
when: docker_sources is changed
- name: Install Docker
ansible.builtin.apt:
name:
- docker-ce
- docker-ce-cli
- containerd.io
update_cache: true
cache_valid_time: 3600
- name: python-docker is installed
- name: Package python-docker is installed
ansible.builtin.apt:
name: python-docker
state: present
when: ansible_python_version is version('3', '<')
- name: python3-docker is installed
- name: Package python3-docker is installed
ansible.builtin.apt:
name: python3-docker
state: present
@ -82,6 +90,9 @@
ansible.builtin.template:
src: daemon.json.j2
dest: /etc/docker/daemon.json
owner: root
group: root
mode: "0644"
notify: restart docker
- name: Creating Docker tmp directory

View file

@ -1,5 +1,6 @@
{
"debug": false
,"iptables": true
{# Docker data-dir (default to /var/lib/docker) #}
,"data-root": "{{ docker_home }}"
{# Keep containers running while docker daemon downtime #}
@ -7,11 +8,6 @@
{% if docker_conf_user_namespace %}
{# Turn on user namespace remaping #}
,"userns-remap": "default"
{% endif %}
{% if docker_conf_use_iptables %}
{# Use iptables instead of docker-proxy #}
,"userland-proxy": false
,"iptables": true
{% endif %}
{# Disable the possibility for containers processes to gain new privileges #}
,"no-new-privileges": {{ docker_conf_no_newprivileges | to_json }}

View file

@ -1,23 +1,16 @@
# evobackup-client
Allows the configuration of backups to one or more remote filesystems.
Install the necessary libraries and script to configure backup scripts.
The backup hosts and the ports in use need to be defined in
evobackup-client__hosts before running it.
Additional information:
The default zzz_evobackup.sh configures a system only backup, but the
template can be overriden to configure a full backup instead. If
you change the variables in defaults/main.yml you can easily run
this again and configure backups to a second set of hosts.
* [evobackup-client documentation](https://gitea.evolix.org/evolix/evobackup/src/branch/master/client/README.md)
* canary
Do not forget to set the evobackup-client__mail variable to an
email adress you control.
## Available variables
You can add this example to an installation playbook to create the
ssh key without running the rest of the role.
~~~
post_tasks:
- include_role:
name: evobackup-client tasks_from: ssh_key.yml
~~~
* `evobackup_client__lib_dir` : directory for libraries (default: `/usr/local/lib/evobackup`)
* `evobackup_client__bin_dir` : directory for scripts/binaries (default: `/usr/local/bin`)
* `evobackup_client__update_canary_enable` : should the canary be updated (default: `True`)
* `evobackup_client__update_canary_path` : path for the canary update script (default: `/etc/cron.daily/000-update-evobackup-canary`)
* `evobackup_client__update_canary_who` : who the canary update must be attributed to (default: `@daily`)

View file

@ -1,15 +1,22 @@
---
evobackup_client__root_key_path: "/root/.ssh/id_ed25519"
evobackup_client__root_key_type: "ed25519"
evobackup_client__cron_path: "/etc/cron.daily/zzz_evobackup"
evobackup_client__cron_template_name: "zzz_evobackup"
evobackup_client__mail: null
evobackup_client__servers_fallback: -1
evobackup_client__pid_path: "/var/run/evobackup.pid"
evobackup_client__log_path: "/var/log/evobackup.log"
evobackup_client__backup_path: "/home/backup"
evobackup_client__hosts: null
# evobackup_client__root_key_path: "/root/.ssh/id_ed25519"
# evobackup_client__root_key_type: "ed25519"
# evobackup_client__cron_path: "/etc/cron.daily/zzz_evobackup"
# evobackup_client__cron_template_name: "zzz_evobackup"
# evobackup_client__mail: null
# evobackup_client__servers_fallback: -1
# evobackup_client__pid_path: "/var/run/evobackup.pid"
# evobackup_client__log_path: "/var/log/evobackup.log"
# evobackup_client__backup_path: "/home/backup"
# evobackup_client__hosts: null
# - name: "backups.example.org"
# ip: "xxx.xxx.xxx.xxx"
# fingerprint: "ecdsa-sha2-nistp256 ..."
# port: xxxx
evobackup_client__lib_dir: "/usr/local/lib/evobackup"
evobackup_client__bin_dir: "/usr/local/bin"
evobackup_client__update_canary_enable: True
evobackup_client__update_canary_path: /etc/cron.daily/000-update-evobackup-canary
evobackup_client__update_canary_who: "@daily"

View file

@ -0,0 +1,82 @@
# Changelog
All notable changes to this project will be documented in this file.
The format is based on [Keep a Changelog](http://keepachangelog.com/en/1.0.0/).
This project does not follow semantic versioning.
The **major** part of the version is the year
The **minor** part changes is the month
The **patch** part changes is incremented if multiple releases happen the same month
## [Unreleased]
### Added
### Changed
### Deprecated
### Removed
### Fixed
### Security
## [24.04.1]
### Fixed
* evobackupctl: quote ARGS variable for options parsing.
## [24.04]
### Added
* Vagrant definition for manual tests
### Changed
* split functions into libraries
* add evobackupctl script
* change the "zzz_evobackup" script to a template, easy to copy with evobackupctl
* use env-based shebang for shell scripts
* use $TMPDIR if available
### Removed
* update-evobackup-canary is managed by ansible-roles.git
* deployment by Ansible is managed elsewhere (now in evolix-private.git, later in ansible-roles.git)
### Fixed
* don't exit the whole program if a sync task can't be done
## [22.12]
### Changed
* Use --dump-dir instead of --backup-dir to suppress dump-server-state warning
* Do not use rsync compression
* Replace rsync option --verbose by --itemize-changes
* Add canary to zzz_evobackup
* update-evobackup-canary: do not use GNU date, for it to be compatible with OpenBSD
* Add AGPL License and README
* Script now depends on Bash
* tolerate absence of mtr or traceroute
* Only one loop for all Redis instances
* remodel how we build the rsync command
* use sub shells instead of moving around
* Separate Rsync for the canary file if the main Rsync has finished without errors
### Removed
* No more fallback if dump-server-state is missing
### Fixed
* Make start_time and stop_time compatible with OpenBSD
## [22.03]
Split client and server parts of the project

View file

@ -0,0 +1,153 @@
#!/usr/bin/env bash
# shellcheck disable=SC2155
readonly PROGNAME=$(basename "${0}")
# shellcheck disable=SC2155
readonly PROGDIR=$(readlink -m "$(dirname "${0}")")
# shellcheck disable=SC2124
readonly ARGS=$@
# Change this to wherever you install the libraries
readonly LIBDIR="/usr/local/lib/evobackup"
source "${LIBDIR}/main.sh"
show_version() {
cat <<END
${PROGNAME} version ${VERSION}
Copyright 2024 Evolix <info@evolix.fr>,
Jérémy Lecour <jlecour@evolix.fr>.
${PROGNAME} comes with ABSOLUTELY NO WARRANTY. This is free software,
and you are welcome to redistribute it under certain conditions.
See the GNU General Public License v3.0 for details.
END
}
show_help() {
cat <<END
${PROGNAME} helps managing evobackup scripts
Options
-h, --help print this message and exit
-V, --version print version and exit
--jail-init-commands print jail init commands
--copy-template=PATH copy the backup template to PATH
END
}
jail_init_commands() {
if [ ! -f /root/.ssh/id_ed25519.pub ]; then
ssh-keygen -t ed25519 -f /root/.ssh/id_ed25519 -N ''
echo ""
fi
SSH_KEY=$(cat /root/.ssh/id_ed25519.pub)
SERVER_NAME=$(hostname -s)
if [ "$(uname -s)" = "OpenBSD" ]; then
SERVER_IP=$(ifconfig egress | grep "inet " | head -1 | awk '{ print $2}')
else
SERVER_IP=$(curl -4 https://ifconfig.me 2> /dev/null || hostname -I | awk '{ print $1}')
fi
echo "Copy-paste those lines on backup server(s) :"
echo "----------"
echo "SERVER_NAME=${SERVER_NAME}"
echo "SERVER_IP=${SERVER_IP}"
echo "echo '${SSH_KEY}' > /root/\${SERVER_NAME}.pub"
echo "bkctld init \${SERVER_NAME}"
echo "bkctld key \${SERVER_NAME} /root/\${SERVER_NAME}.pub"
echo "bkctld ip \${SERVER_NAME} \${SERVER_IP}"
echo "bkctld start \${SERVER_NAME}"
echo "bkctld status \${SERVER_NAME}"
echo "grep --quiet --extended-regexp \"^\\s?NODE=\" /etc/default/bkctld && bkctld sync \${SERVER_NAME}"
echo "----------"
}
copy_template() {
dest_path=${1}
dest_dir="$(dirname "${dest_path}")"
if [ -e "${dest_path}" ]; then
printf "Path for new evobackup script '%s' already exists.\n" "${dest_path}" >&2
exit 1
elif [ ! -e "${dest_dir}" ]; then
printf "Parent directory '%s' doesn't exist. Create it first.\n" "${dest_dir}" >&2
exit 1
else
if cp "${LIBDIR}/zzz_evobackup.sh" "${dest_path}"; then
chmod 750 "${dest_path}"
sed -i "s|@COMMAND@|${PROGDIR}/${PROGNAME} ${ARGS}|" "${dest_path}"
sed -i "s|@DATE@|$(date --iso-8601=seconds)|" "${dest_path}"
sed -i "s|@VERSION@|${VERSION}|" "${dest_path}"
printf "New evobackup script has been saved to '%s'.\n" "${dest_path}"
printf "Remember to customize it (mail notifications, backup servers…).\n"
exit 0
fi
fi
}
main() {
# If no argument is provided, print help and exit
# shellcheck disable=SC2086
if [ -z "${ARGS}" ]; then
show_help
exit 0
fi
# Parse options, based on https://gist.github.com/deshion/10d3cb5f88a21671e17a
while :; do
case ${1:-''} in
-V|--version)
show_version
exit 0
;;
-h|--help)
show_help
exit 0
;;
--jail-init-commands)
jail_init_commands
exit 0
;;
--copy-template)
# copy-template option, with value separated by space
if [ -n "$2" ]; then
copy_template "${2}"
shift
else
printf "'%s' requires a non-empty option argument.\n" "--copy-template" >&2
exit 1
fi
;;
--copy-template=?*)
# copy-template option, with value separated by =
copy_template "${1#*=}"
;;
--copy-template=)
# copy-template option, without value
printf "'%s' requires a non-empty option argument.\n" "--copy-template" >&2
exit 1
;;
--)
# End of all options.
shift
break
;;
-?*|[[:alnum:]]*)
# ignore unknown options
printf "unknown option '%s'.\n" "${1}" >&2
exit 1
;;
*)
# Default case: If no more options then break out of the loop.
break
;;
esac
shift
done
}
main ${ARGS}

View file

@ -0,0 +1,301 @@
#!/usr/bin/env bash
# shellcheck disable=SC2034,SC2317,SC2155
#######################################################################
# Snapshot Elasticsearch data
#
# Arguments:
# --protocol=<http|https> (default: http)
# --cacert=[String] (default: <none>)
# path to the CA certificate to use when using https
# --host=[String] (default: localhost)
# --port=[Integer] (default: 9200)
# --user=[String] (default: <none>)
# --password=[String] (default: <none>)
# --repository=[String] (default: snaprepo)
# --snapshot=[String] (default: snapshot.daily)
#######################################################################
dump_elasticsearch() {
local option_protocol="http"
local option_cacert=""
local option_host="localhost"
local option_port="9200"
local option_user=""
local option_password=""
local option_repository="snaprepo"
local option_snapshot="snapshot.daily"
local option_others=""
# Parse options, based on https://gist.github.com/deshion/10d3cb5f88a21671e17a
while :; do
case ${1:-''} in
--protocol)
# protocol options, with value separated by space
if [ -n "$2" ]; then
option_protocol="${2}"
shift
else
log_error "LOCAL_TASKS - ${FUNCNAME[0]}: '--protocol' requires a non-empty option argument."
exit 1
fi
;;
--protocol=?*)
# protocol options, with value separated by =
option_protocol="${1#*=}"
;;
--protocol=)
# protocol options, without value
log_error "LOCAL_TASKS - ${FUNCNAME[0]}: '--protocol' requires a non-empty option argument."
exit 1
;;
--cacert)
# cacert options, with value separated by space
if [ -n "$2" ]; then
option_cacert="${2}"
shift
else
log_error "LOCAL_TASKS - ${FUNCNAME[0]}: '--cacert' requires a non-empty option argument."
exit 1
fi
;;
--cacert=?*)
# cacert options, with value separated by =
option_cacert="${1#*=}"
;;
--cacert=)
# cacert options, without value
log_error "LOCAL_TASKS - ${FUNCNAME[0]}: '--cacert' requires a non-empty option argument."
exit 1
;;
--host)
# host options, with value separated by space
if [ -n "$2" ]; then
option_host="${2}"
shift
else
log_error "LOCAL_TASKS - ${FUNCNAME[0]}: '--host' requires a non-empty option argument."
exit 1
fi
;;
--host=?*)
# host options, with value separated by =
option_host="${1#*=}"
;;
--host=)
# host options, without value
log_error "LOCAL_TASKS - ${FUNCNAME[0]}: '--host' requires a non-empty option argument."
exit 1
;;
--port)
# port options, with value separated by space
if [ -n "$2" ]; then
option_port="${2}"
shift
else
log_error "LOCAL_TASKS - ${FUNCNAME[0]}: '--port' requires a non-empty option argument."
exit 1
fi
;;
--port=?*)
# port options, with value separated by =
option_port="${1#*=}"
;;
--port=)
# port options, without value
log_error "LOCAL_TASKS - ${FUNCNAME[0]}: '--port' requires a non-empty option argument."
exit 1
;;
--user)
# user options, with value separated by space
if [ -n "$2" ]; then
option_user="${2}"
shift
else
log_error "LOCAL_TASKS - ${FUNCNAME[0]}: '--user' requires a non-empty option argument."
exit 1
fi
;;
--user=?*)
# user options, with value separated by =
option_user="${1#*=}"
;;
--user=)
# user options, without value
log_error "LOCAL_TASKS - ${FUNCNAME[0]}: '--user' requires a non-empty option argument."
exit 1
;;
--password)
# password options, with value separated by space
if [ -n "$2" ]; then
option_password="${2}"
shift
else
log_error "LOCAL_TASKS - ${FUNCNAME[0]}: '--password' requires a non-empty option argument."
exit 1
fi
;;
--password=?*)
# password options, with value separated by =
option_password="${1#*=}"
;;
--password=)
# password options, without value
log_error "LOCAL_TASKS - ${FUNCNAME[0]}: '--password' requires a non-empty option argument."
exit 1
;;
--repository)
# repository options, with value separated by space
if [ -n "$2" ]; then
option_repository="${2}"
shift
else
log_error "LOCAL_TASKS - ${FUNCNAME[0]}: '--repository' requires a non-empty option argument."
exit 1
fi
;;
--repository=?*)
# repository options, with value separated by =
option_repository="${1#*=}"
;;
--repository=)
# repository options, without value
log_error "LOCAL_TASKS - ${FUNCNAME[0]}: '--repository' requires a non-empty option argument."
exit 1
;;
--snapshot)
# snapshot options, with value separated by space
if [ -n "$2" ]; then
option_snapshot="${2}"
shift
else
log_error "LOCAL_TASKS - ${FUNCNAME[0]}: '--snapshot' requires a non-empty option argument."
exit 1
fi
;;
--snapshot=?*)
# snapshot options, with value separated by =
option_snapshot="${1#*=}"
;;
--snapshot=)
# snapshot options, without value
log_error "LOCAL_TASKS - ${FUNCNAME[0]}: '--snapshot' requires a non-empty option argument."
exit 1
;;
--)
# End of all options.
shift
option_others=${*}
break
;;
-?*|[[:alnum:]]*)
# ignore unknown options
log_error "LOCAL_TASKS - ${FUNCNAME[0]}: unknown option '${1}' (ignored)"
;;
*)
# Default case: If no more options then break out of the loop.
break
;;
esac
shift
done
# Use the default Elasticsearch CA certificate when using HTTPS, if not specified directly
local default_cacert="/etc/elasticsearch/certs/http_ca.crt"
if [ "${option_protocol}" = "https" ] && [ -z "${option_cacert}" ] && [ -f "${default_cacert}" ]; then
option_cacert="${default_cacert}"
fi
local errors_dir="${ERRORS_DIR}/elasticsearch-${option_repository}-${option_snapshot}"
rm -rf "${errors_dir}"
mkdir -p "${errors_dir}"
# No need to change recursively, the top directory is enough
chmod 700 "${errors_dir}"
log "LOCAL_TASKS - ${FUNCNAME[0]}: start ${option_snapshot}"
## Take a snapshot as a backup.
## Warning: You need to have a path.repo configured.
## See: https://wiki.evolix.org/HowtoElasticsearch#snapshots-et-sauvegardes
local base_url="${option_protocol}://${option_host}:${option_port}"
local repository_url="${base_url}/_snapshot/${option_repository}"
local snapshot_url="${repository_url}/${option_snapshot}"
# Verify snapshot repository
local error_file="${errors_dir}/verify.err"
declare -a connect_options
connect_options=()
if [ -n "${option_cacert}" ]; then
connect_options+=(--cacert "${option_cacert}")
fi
if [ -n "${option_user}" ] || [ -n "${option_password}" ]; then
local connect_options+=("--user ${option_user}:${option_password}")
fi
if [ -n "${option_others}" ]; then
# word splitting is deliberate here
# shellcheck disable=SC2206
connect_options+=(${option_others})
fi
# Add the http return code at the end of the output
connect_options+=(--write-out '%{http_code}\n')
connect_options+=(--silent)
declare -a dump_options
dump_options=()
dump_options+=(--request POST)
dump_cmd="curl ${connect_options[*]} ${dump_options[*]} ${repository_url}/_verify?pretty"
log "LOCAL_TASKS - ${FUNCNAME[0]}: ${dump_cmd}"
${dump_cmd} > "${error_file}"
# test if the last line of the log file is "200"
tail -n 1 "${error_file}" | grep --quiet "^200$" "${error_file}"
local last_rc=$?
# shellcheck disable=SC2086
if [ ${last_rc} -ne 0 ]; then
log_error "LOCAL_TASKS - ${FUNCNAME[0]}: repository verification returned an error ${last_rc}" "${error_file}"
GLOBAL_RC=${E_DUMPFAILED}
else
rm -f "${error_file}"
# Delete snapshot
declare -a dump_options
dump_options=()
dump_options+=(--request DELETE)
dump_cmd="curl ${connect_options[*]} ${dump_options[*]} ${snapshot_url}"
log "LOCAL_TASKS - ${FUNCNAME[0]}: ${dump_cmd}"
${dump_cmd} > /dev/null
# Create snapshot
local error_file="${errors_dir}/create.err"
declare -a dump_options
dump_options=()
dump_options+=(--request PUT)
dump_cmd="curl ${connect_options[*]} ${dump_options[*]} ${snapshot_url}?wait_for_completion=true"
log "LOCAL_TASKS - ${FUNCNAME[0]}: ${dump_cmd}"
${dump_cmd} > "${error_file}"
# test if the last line of the log file is "200"
tail -n 1 "${error_file}" | grep --quiet "^200$" "${error_file}"
local last_rc=$?
# shellcheck disable=SC2086
if [ ${last_rc} -ne 0 ]; then
log_error "LOCAL_TASKS - ${FUNCNAME[0]}: curl returned an error ${last_rc}" "${error_file}"
GLOBAL_RC=${E_DUMPFAILED}
else
rm -f "${error_file}"
fi
fi
log "LOCAL_TASKS - ${FUNCNAME[0]}: stop ${option_snapshot}"
}

View file

@ -0,0 +1,559 @@
#!/usr/bin/env bash
# shellcheck disable=SC2034,SC2317,SC2155
#######################################################################
# Dump LDAP files (config, data, all)
#
# Arguments: <none>
#######################################################################
dump_ldap() {
## OpenLDAP : example with slapcat
local dump_dir="${LOCAL_BACKUP_DIR}/ldap"
rm -rf "${dump_dir}"
mkdir -p "${dump_dir}"
chmod 700 "${dump_dir}"
log "LOCAL_TASKS - ${FUNCNAME[0]}: start ${FUNCNAME[0]} to ${dump_dir}"
dump_cmd="slapcat -n 0 -l ${dump_dir}/config.bak"
log "LOCAL_TASKS - ${FUNCNAME[0]}: ${dump_cmd}"
${dump_cmd}
dump_cmd="slapcat -n 1 -l ${dump_dir}/data.bak"
log "LOCAL_TASKS - ${FUNCNAME[0]}: ${dump_cmd}"
${dump_cmd}
dump_cmd="slapcat -l ${dump_dir}/all.bak"
log "LOCAL_TASKS - ${FUNCNAME[0]}: ${dump_cmd}"
${dump_cmd}
log "LOCAL_TASKS - ${FUNCNAME[0]}: stop ${FUNCNAME[0]}"
}
#######################################################################
# Copy dump file of Redis instances
#
# Arguments:
# --instances=[Integer] (default: all)
#######################################################################
dump_redis() {
all_instances=$(find /var/lib/ -mindepth 1 -maxdepth 1 '(' -type d -o -type l ')' -name 'redis*')
local option_instances=""
# Parse options, based on https://gist.github.com/deshion/10d3cb5f88a21671e17a
while :; do
case ${1:-''} in
--instances)
# instances options, with key and value separated by space
if [ -n "$2" ]; then
if [ "${2}" == "all" ]; then
read -a option_instances <<< "${all_instances}"
else
IFS="," read -a option_instances <<< "${2}"
fi
shift
else
log_error "LOCAL_TASKS - ${FUNCNAME[0]}: '--instances' requires a non-empty option argument."
exit 1
fi
;;
--instances=?*)
# instances options, with key and value separated by =
if [ "${1#*=}" == "all" ]; then
read -a option_instances <<< "${all_instances}"
else
IFS="," read -a option_instances <<< "${1#*=}"
fi
;;
--instances=)
# instances options, without value
log_error "LOCAL_TASKS - ${FUNCNAME[0]}: '--instances' requires a non-empty option argument."
exit 1
;;
--)
# End of all options.
shift
break
;;
-?*|[[:alnum:]]*)
# ignore unknown options
log_error "LOCAL_TASKS - ${FUNCNAME[0]}: unknown option '${1}' (ignored)"
;;
*)
# Default case: If no more options then break out of the loop.
break
;;
esac
shift
done
for instance in "${option_instances[@]}"; do
name=$(basename "${instance}")
local dump_dir="${LOCAL_BACKUP_DIR}/${name}"
local errors_dir=$(errors_dir_from_dump_dir "${dump_dir}")
rm -rf "${dump_dir}" "${errors_dir}"
mkdir -p "${dump_dir}" "${errors_dir}"
# No need to change recursively, the top directory is enough
chmod 700 "${dump_dir}" "${errors_dir}"
if [ -f "${instance}/dump.rdb" ]; then
local error_file="${errors_dir}/${name}.err"
log "LOCAL_TASKS - ${FUNCNAME[0]}: start ${dump_dir}"
# Copy the Redis database
dump_cmd="cp -a ${instance}/dump.rdb ${dump_dir}/dump.rdb"
log "LOCAL_TASKS - ${FUNCNAME[0]}: ${dump_cmd}"
${dump_cmd} 2> "${error_file}"
local last_rc=$?
# shellcheck disable=SC2086
if [ ${last_rc} -ne 0 ]; then
log_error "LOCAL_TASKS - ${FUNCNAME[0]}: cp ${instance}/dump.rdb to ${dump_dir} returned an error ${last_rc}" "${error_file}"
GLOBAL_RC=${E_DUMPFAILED}
else
rm -f "${error_file}"
fi
# Compress the Redis database
dump_cmd="gzip ${dump_dir}/dump.rdb"
log "LOCAL_TASKS - ${FUNCNAME[0]}: ${dump_cmd}"
${dump_cmd}
local last_rc=$?
# shellcheck disable=SC2086
if [ ${last_rc} -ne 0 ]; then
log_error "LOCAL_TASKS - ${FUNCNAME[0]}: gzip ${dump_dir}/dump.rdb returned an error ${last_rc}" "${error_file}"
GLOBAL_RC=${E_DUMPFAILED}
else
rm -f "${error_file}"
fi
log "LOCAL_TASKS - ${FUNCNAME[0]}: stop ${dump_dir}"
else
log_error "LOCAL_TASKS - ${FUNCNAME[0]}: '${instance}/dump.rdb' not found."
fi
done
}
#######################################################################
# Dump all collections of a MongoDB database
# using a custom authentication, instead of /etc/mysql/debian.cnf
#
# Arguments:
# --port=[String] (default: <blank>)
# --user=[String] (default: <blank>)
# --password=[String] (default: <blank>)
# --dump-label=[String] (default: "default")
# used as suffix of the dump dir to differenciate multiple instances
# Other options after -- are passed as-is to mongodump
#
# don't forget to create use with read-only access
# > use admin
# > db.createUser( { user: "mongobackup", pwd: "PASS", roles: [ "backup", ] } )
#######################################################################
dump_mongodb() {
local option_port=""
local option_user=""
local option_password=""
local option_dump_label=""
local option_others=""
# Parse options, based on https://gist.github.com/deshion/10d3cb5f88a21671e17a
while :; do
case ${1:-''} in
--port)
# port options, with value separated by space
if [ -n "$2" ]; then
option_port="${2}"
shift
else
log_error "LOCAL_TASKS - ${FUNCNAME[0]}: '--port' requires a non-empty option argument."
exit 1
fi
;;
--port=?*)
# port options, with value separated by =
option_port="${1#*=}"
;;
--port=)
# port options, without value
log_error "LOCAL_TASKS - ${FUNCNAME[0]}: '--port' requires a non-empty option argument."
exit 1
;;
--user)
# user options, with value separated by space
if [ -n "$2" ]; then
option_user="${2}"
shift
else
log_error "LOCAL_TASKS - ${FUNCNAME[0]}: '--user' requires a non-empty option argument."
exit 1
fi
;;
--user=?*)
# user options, with value separated by =
option_user="${1#*=}"
;;
--user=)
# user options, without value
log_error "LOCAL_TASKS - ${FUNCNAME[0]}: '--user' requires a non-empty option argument."
exit 1
;;
--password)
# password options, with value separated by space
if [ -n "$2" ]; then
option_password="${2}"
shift
else
log_error "LOCAL_TASKS - ${FUNCNAME[0]}: '--password' requires a non-empty option argument."
exit 1
fi
;;
--password=?*)
# password options, with value separated by =
option_password="${1#*=}"
;;
--password=)
# password options, without value
log_error "LOCAL_TASKS - ${FUNCNAME[0]}: '--password' requires a non-empty option argument."
exit 1
;;
--dump-label)
# dump-label options, with value separated by space
if [ -n "$2" ]; then
option_dump_label="${2}"
shift
else
log_error "LOCAL_TASKS - ${FUNCNAME[0]}: '--dump-label' requires a non-empty option argument."
exit 1
fi
;;
--dump-label=?*)
# dump-label options, with value separated by =
option_dump_label="${1#*=}"
;;
--dump-label=)
# dump-label options, without value
log_error "LOCAL_TASKS - ${FUNCNAME[0]}: '--dump-label' requires a non-empty option argument."
exit 1
;;
--)
# End of all options.
shift
option_others=${*}
break
;;
-?*|[[:alnum:]]*)
# ignore unknown options
log_error "LOCAL_TASKS - ${FUNCNAME[0]}: unknown option '${1}' (ignored)"
;;
*)
# Default case: If no more options then break out of the loop.
break
;;
esac
shift
done
if [ -z "${option_dump_label}" ]; then
if [ -n "${option_port}" ]; then
option_dump_label="${option_port}"
else
option_dump_label="default"
fi
fi
local dump_dir="${LOCAL_BACKUP_DIR}/mongodb-${option_dump_label}"
local errors_dir=$(errors_dir_from_dump_dir "${dump_dir}")
rm -rf "${dump_dir}" "${errors_dir}"
mkdir -p "${dump_dir}" "${errors_dir}"
# No need to change recursively, the top directory is enough
chmod 700 "${dump_dir}" "${errors_dir}"
local error_file="${errors_dir}.err"
log "LOCAL_TASKS - ${FUNCNAME[0]}: start ${dump_dir}"
declare -a dump_options
dump_options=()
if [ -n "${option_port}" ]; then
dump_options+=(--port="${option_port}")
fi
if [ -n "${option_user}" ]; then
dump_options+=(--username="${option_user}")
fi
if [ -n "${option_password}" ]; then
dump_options+=(--password="${option_password}")
fi
dump_options+=(--out="${dump_dir}/")
if [ -n "${option_others}" ]; then
# word splitting is deliberate here
# shellcheck disable=SC2206
dump_options+=(${option_others})
fi
dump_cmd="mongodump ${dump_options[*]}"
log "LOCAL_TASKS - ${FUNCNAME[0]}: ${dump_cmd} > /dev/null"
${dump_cmd} 2> "${error_file}" > /dev/null
local last_rc=$?
# shellcheck disable=SC2086
if [ ${last_rc} -ne 0 ]; then
log_error "LOCAL_TASKS - ${FUNCNAME[0]}: mongodump to ${dump_dir} returned an error ${last_rc}" "${error_file}"
GLOBAL_RC=${E_DUMPFAILED}
else
rm -f "${error_file}"
fi
log "LOCAL_TASKS - stop ${FUNCNAME[0]}: ${dump_dir}"
}
#######################################################################
# Dump RAID configuration
#
# Arguments: <none>
#######################################################################
dump_raid_config() {
local dump_dir="${LOCAL_BACKUP_DIR}/raid"
local errors_dir=$(errors_dir_from_dump_dir "${dump_dir}")
rm -rf "${dump_dir}" "${errors_dir}"
mkdir -p "${dump_dir}" "${errors_dir}"
# No need to change recursively, the top directory is enough
chmod 700 "${dump_dir}" "${errors_dir}"
if command -v megacli > /dev/null; then
local error_file="${errors_dir}/megacli.cfg"
local dump_file="${dump_dir}/megacli.err"
log "LOCAL_TASKS - ${FUNCNAME[0]}: start ${dump_file}"
dump_cmd="megacli -CfgSave -f ${dump_file} -a0"
log "LOCAL_TASKS - ${FUNCNAME[0]}: ${dump_cmd}"
${dump_cmd} 2> "${error_file}"
local last_rc=$?
# shellcheck disable=SC2086
if [ ${last_rc} -ne 0 ]; then
log_error "LOCAL_TASKS - ${FUNCNAME[0]}: megacli to ${dump_file} returned an error ${last_rc}" "${error_file}"
GLOBAL_RC=${E_DUMPFAILED}
else
rm -f "${error_file}"
fi
log "LOCAL_TASKS - ${FUNCNAME[0]}: stop ${dump_file}"
elif command -v perccli > /dev/null; then
local error_file="${errors_dir}/perccli.cfg"
local dump_file="${dump_dir}/perccli.err"
# log "LOCAL_TASKS - ${FUNCNAME[0]}: start ${dump_file}"
# TODO: find out what the correct command is
# dump_cmd="perccli XXXX"
# log "LOCAL_TASKS - ${FUNCNAME[0]}: ${dump_cmd}"
# ${dump_cmd} 2> ${error_file}
# local last_rc=$?
# # shellcheck disable=SC2086
# if [ ${last_rc} -ne 0 ]; then
# log_error "LOCAL_TASKS - ${FUNCNAME[0]}: perccli to ${dump_file} returned an error ${last_rc}" "${error_file}"
# GLOBAL_RC=${E_DUMPFAILED}
# else
# rm -f "${error_file}"
# fi
# log "LOCAL_TASKS - ${FUNCNAME[0]}: stop ${dump_file}"
else
log "LOCAL_TASKS - ${FUNCNAME[0]}: 'megacli' and 'perccli' not found, unable to dump RAID configuration"
fi
}
#######################################################################
# Save some traceroute/mtr results
#
# Arguments:
# --targets=[IP,HOST] (default: <none>)
#######################################################################
dump_traceroute() {
local option_targets=""
# Parse options, based on https://gist.github.com/deshion/10d3cb5f88a21671e17a
while :; do
case ${1:-''} in
--targets)
# targets options, with key and value separated by space
if [ -n "$2" ]; then
IFS="," read -a option_targets <<< "${2}"
shift
else
log_error "LOCAL_TASKS - ${FUNCNAME[0]}: '--targets' requires a non-empty option argument."
exit 1
fi
;;
--targets=?*)
# targets options, with key and value separated by =
IFS="," read -a option_targets <<< "${1#*=}"
;;
--targets=)
# targets options, without value
log_error "LOCAL_TASKS - ${FUNCNAME[0]}: '--targets' requires a non-empty option argument."
exit 1
;;
--)
# End of all options.
shift
break
;;
-?*|[[:alnum:]]*)
# ignore unknown options
log_error "LOCAL_TASKS - ${FUNCNAME[0]}: unknown option '${1}' (ignored)"
;;
*)
# Default case: If no more options then break out of the loop.
break
;;
esac
shift
done
local dump_dir="${LOCAL_BACKUP_DIR}/traceroute"
local errors_dir=$(errors_dir_from_dump_dir "${dump_dir}")
rm -rf "${dump_dir}" "${errors_dir}"
mkdir -p "${dump_dir}" "${errors_dir}"
# No need to change recursively, the top directory is enough
chmod 700 "${dump_dir}" "${errors_dir}"
mtr_bin=$(command -v mtr)
if [ -n "${mtr_bin}" ]; then
for target in "${option_targets[@]}"; do
local dump_file="${dump_dir}/mtr-${target}"
log "LOCAL_TASKS - ${FUNCNAME[0]}: start ${dump_file}"
${mtr_bin} -r "${target}" > "${dump_file}"
log "LOCAL_TASKS - ${FUNCNAME[0]}: stop ${dump_file}"
done
fi
traceroute_bin=$(command -v traceroute)
if [ -n "${traceroute_bin}" ]; then
for target in "${option_targets[@]}"; do
local dump_file="${dump_dir}/traceroute-${target}"
log "LOCAL_TASKS - ${FUNCNAME[0]}: start ${dump_file}"
${traceroute_bin} -n "${target}" > "${dump_file}" 2>&1
log "LOCAL_TASKS - ${FUNCNAME[0]}: stop ${dump_file}"
done
fi
}
#######################################################################
# Save many system information, using dump_server_state
#
# Arguments:
# any option for dump-server-state (except --dump-dir) is usable
# (default: --all)
#######################################################################
dump_server_state() {
local dump_dir="${LOCAL_BACKUP_DIR}/server-state"
rm -rf "${dump_dir}"
# Do not create the directory
# mkdir -p -m 700 "${dump_dir}"
log "LOCAL_TASKS - ${FUNCNAME[0]}: start ${dump_dir}"
# pass all options
read -a options <<< "${@}"
# if no option is given, use "--all" as fallback
if [ ${#options[@]} -le 0 ]; then
options=(--all)
fi
# add "--dump-dir" in case it is missing (as it should)
options+=(--dump-dir "${dump_dir}")
dump_server_state_bin=$(command -v dump-server-state)
if [ -z "${dump_server_state_bin}" ]; then
log_error "LOCAL_TASKS - ${FUNCNAME[0]}: dump-server-state is missing"
rc=1
else
dump_cmd="${dump_server_state_bin} ${options[*]}"
log "LOCAL_TASKS - ${FUNCNAME[0]}: ${dump_cmd}"
${dump_cmd}
local last_rc=$?
# shellcheck disable=SC2086
if [ ${last_rc} -ne 0 ]; then
log_error "LOCAL_TASKS - ${FUNCNAME[0]}: dump-server-state returned an error ${last_rc}, check ${dump_dir}"
GLOBAL_RC=${E_DUMPFAILED}
fi
fi
log "LOCAL_TASKS - ${FUNCNAME[0]}: stop ${dump_dir}"
}
#######################################################################
# Save RabbitMQ data
#
# Arguments: <none>
#
# Warning: This has been poorly tested
#######################################################################
dump_rabbitmq() {
local dump_dir="${LOCAL_BACKUP_DIR}/rabbitmq"
local errors_dir=$(errors_dir_from_dump_dir "${dump_dir}")
rm -rf "${dump_dir}" "${errors_dir}"
mkdir -p "${dump_dir}" "${errors_dir}"
# No need to change recursively, the top directory is enough
chmod 700 "${dump_dir}" "${errors_dir}"
local error_file="${errors_dir}.err"
local dump_file="${dump_dir}/config"
log "LOCAL_TASKS - ${FUNCNAME[0]}: start ${dump_file}"
dump_cmd="rabbitmqadmin export ${dump_file}"
log "LOCAL_TASKS - ${FUNCNAME[0]}: ${dump_cmd}"
${dump_cmd} 2> "${error_file}"
local last_rc=$?
# shellcheck disable=SC2086
if [ ${last_rc} -ne 0 ]; then
log_error "LOCAL_TASKS - ${FUNCNAME[0]}: pg_dump to ${dump_file} returned an error ${last_rc}" "${error_file}"
GLOBAL_RC=${E_DUMPFAILED}
else
rm -f "${error_file}"
fi
log "LOCAL_TASKS - ${FUNCNAME[0]}: stop ${dump_file}"
}
#######################################################################
# Save Files ACL on various partitions.
#
# Arguments: <none>
#######################################################################
dump_facl() {
local dump_dir="${LOCAL_BACKUP_DIR}/facl"
local errors_dir=$(errors_dir_from_dump_dir "${dump_dir}")
rm -rf "${dump_dir}" "${errors_dir}"
mkdir -p "${dump_dir}" "${errors_dir}"
# No need to change recursively, the top directory is enough
chmod 700 "${dump_dir}" "${errors_dir}"
log "LOCAL_TASKS - ${FUNCNAME[0]}: start ${dump_dir}"
dump_cmd="getfacl -R /etc > ${dump_dir}/etc.txt"
log "LOCAL_TASKS - ${FUNCNAME[0]}: ${dump_cmd}"
${dump_cmd}
dump_cmd="getfacl -R /home > ${dump_dir}/home.txt"
log "LOCAL_TASKS - ${FUNCNAME[0]}: ${dump_cmd}"
${dump_cmd}
dump_cmd="getfacl -R /usr > ${dump_dir}/usr.txt"
log "LOCAL_TASKS - ${FUNCNAME[0]}: ${dump_cmd}"
${dump_cmd}
dump_cmd="getfacl -R /var > ${dump_dir}/var.txt"
log "LOCAL_TASKS - ${FUNCNAME[0]}: ${dump_cmd}"
${dump_cmd}
log "LOCAL_TASKS - ${FUNCNAME[0]}: stop ${dump_dir}"
}

View file

@ -0,0 +1,1551 @@
#!/usr/bin/env bash
# shellcheck disable=SC2034,SC2317,SC2155
#######################################################################
# Dump complete summary of an instance (using pt-mysql-summary)
#
# Arguments:
# --port=[Integer] (default: <blank>)
# --socket=[String] (default: <blank>)
# --user=[String] (default: <blank>)
# --password=[String] (default: <blank>)
# --defaults-file=[String] (default: <blank>)
# --defaults-extra-file=[String] (default: <blank>)
# --defaults-group-suffix=[String] (default: <blank>)
# --dump-label=[String] (default: "default")
# used as suffix of the dump dir to differenciate multiple instances
#######################################################################
dump_mysql_summary() {
local option_port=""
local option_socket=""
local option_defaults_file=""
local option_defaults_extra_file=""
local option_defaults_group_suffix=""
local option_user=""
local option_password=""
local option_dump_label=""
# Parse options, based on https://gist.github.com/deshion/10d3cb5f88a21671e17a
while :; do
case ${1:-''} in
--defaults-file)
# defaults-file options, with value separated by space
if [ -n "$2" ]; then
option_defaults_file="${2}"
shift
else
log_error "LOCAL_TASKS - ${FUNCNAME[0]}: '--defaults-file' requires a non-empty option argument."
exit 1
fi
;;
--defaults-file=?*)
# defaults-file options, with value separated by =
option_defaults_file="${1#*=}"
;;
--defaults-file=)
# defaults-file options, without value
log_error "LOCAL_TASKS - ${FUNCNAME[0]}: '--defaults-file' requires a non-empty option argument."
exit 1
;;
--defaults-extra-file)
# defaults-file options, with value separated by space
if [ -n "$2" ]; then
option_defaults_extra_file="${2}"
shift
else
log_error "LOCAL_TASKS - ${FUNCNAME[0]}: '--defaults-file' requires a non-empty option argument."
exit 1
fi
;;
--defaults-extra-file=?*)
# defaults-extra-file options, with value separated by =
option_defaults_extra_file="${1#*=}"
;;
--defaults-extra-file=)
# defaults-extra-file options, without value
log_error "LOCAL_TASKS - ${FUNCNAME[0]}: '--defaults-extra-file' requires a non-empty option argument."
exit 1
;;
--defaults-group-suffix)
# defaults-group-suffix options, with value separated by space
if [ -n "$2" ]; then
option_defaults_group_suffix="${2}"
shift
else
log_error "LOCAL_TASKS - ${FUNCNAME[0]}: '--defaults-group-suffix' requires a non-empty option argument."
exit 1
fi
;;
--defaults-group-suffix=?*)
# defaults-group-suffix options, with value separated by =
option_defaults_group_suffix="${1#*=}"
;;
--defaults-group-suffix=)
# defaults-group-suffix options, without value
log_error "LOCAL_TASKS - ${FUNCNAME[0]}: '--defaults-group-suffix' requires a non-empty option argument."
exit 1
;;
--port)
# port options, with value separated by space
if [ -n "$2" ]; then
option_port="${2}"
shift
else
log_error "LOCAL_TASKS - ${FUNCNAME[0]}: '--port' requires a non-empty option argument."
exit 1
fi
;;
--port=?*)
# port options, with value separated by =
option_port="${1#*=}"
;;
--port=)
# port options, without value
log_error "LOCAL_TASKS - ${FUNCNAME[0]}: '--port' requires a non-empty option argument."
exit 1
;;
--socket)
# socket options, with value separated by space
if [ -n "$2" ]; then
option_socket="${2}"
shift
else
log_error "LOCAL_TASKS - ${FUNCNAME[0]}: '--socket' requires a non-empty option argument."
exit 1
fi
;;
--socket=?*)
# socket options, with value separated by =
option_socket="${1#*=}"
;;
--socket=)
# socket options, without value
log_error "LOCAL_TASKS - ${FUNCNAME[0]}: '--socket' requires a non-empty option argument."
exit 1
;;
--user)
# user options, with value separated by space
if [ -n "$2" ]; then
option_user="${2}"
shift
else
log_error "LOCAL_TASKS - ${FUNCNAME[0]}: '--user' requires a non-empty option argument."
exit 1
fi
;;
--user=?*)
# user options, with value separated by =
option_user="${1#*=}"
;;
--user=)
# user options, without value
log_error "LOCAL_TASKS - ${FUNCNAME[0]}: '--user' requires a non-empty option argument."
exit 1
;;
--password)
# password options, with value separated by space
if [ -n "$2" ]; then
option_password="${2}"
shift
else
log_error "LOCAL_TASKS - ${FUNCNAME[0]}: '--password' requires a non-empty option argument."
exit 1
fi
;;
--password=?*)
# password options, with value separated by =
option_password="${1#*=}"
;;
--password=)
# password options, without value
log_error "LOCAL_TASKS - ${FUNCNAME[0]}: '--password' requires a non-empty option argument."
exit 1
;;
--dump-label)
# dump-label options, with value separated by space
if [ -n "$2" ]; then
option_dump_label="${2}"
shift
else
log_error "LOCAL_TASKS - ${FUNCNAME[0]}: '--dump-label' requires a non-empty option argument."
exit 1
fi
;;
--dump-label=?*)
# dump-label options, with value separated by =
option_dump_label="${1#*=}"
;;
--dump-label=)
# dump-label options, without value
log_error "LOCAL_TASKS - ${FUNCNAME[0]}: '--dump-label' requires a non-empty option argument."
exit 1
;;
--)
# End of all options.
shift
break
;;
-?*|[[:alnum:]]*)
# ignore unknown options
log_error "LOCAL_TASKS - ${FUNCNAME[0]}: unkwnown option (ignored): '${1}'"
;;
*)
# Default case: If no more options then break out of the loop.
break
;;
esac
shift
done
if [ -z "${option_dump_label}" ]; then
if [ -n "${option_defaults_group_suffix}" ]; then
option_dump_label="${option_defaults_group_suffix}"
elif [ -n "${option_port}" ]; then
option_dump_label="${option_port}"
elif [ -n "${option_socket}" ]; then
option_dump_label=$(path_to_str "${option_socket}")
else
option_dump_label="default"
fi
fi
local dump_dir="${LOCAL_BACKUP_DIR}/mysql-${option_dump_label}-summary"
local errors_dir=$(errors_dir_from_dump_dir "${dump_dir}")
rm -rf "${dump_dir}" "${errors_dir}"
mkdir -p "${dump_dir}" "${errors_dir}"
# No need to change recursively, the top directory is enough
chmod 700 "${dump_dir}" "${errors_dir}"
## Dump all grants (requires 'percona-toolkit' package)
if command -v pt-mysql-summary > /dev/null; then
local error_file="${errors_dir}/mysql-summary.err"
local dump_file="${dump_dir}/mysql-summary.out"
log "LOCAL_TASKS - ${FUNCNAME[0]}: start ${dump_file}"
## Connection options
declare -a connect_options
connect_options=()
if [ -n "${option_defaults_file}" ]; then
connect_options+=(--defaults-file="${option_defaults_file}")
fi
if [ -n "${option_defaults_extra_file}" ]; then
connect_options+=(--defaults-extra-file="${option_defaults_extra_file}")
fi
if [ -n "${option_defaults_group_suffix}" ]; then
connect_options+=(--defaults-group-suffix="${option_defaults_group_suffix}")
fi
if [ -n "${option_port}" ]; then
connect_options+=(--protocol=tcp)
connect_options+=(--port="${option_port}")
fi
if [ -n "${option_socket}" ]; then
connect_options+=(--protocol=socket)
connect_options+=(--socket="${option_socket}")
fi
if [ -n "${option_user}" ]; then
connect_options+=(--user="${option_user}")
fi
if [ -n "${option_password}" ]; then
connect_options+=(--password="${option_password}")
fi
declare -a options
options=()
options+=(--sleep=0)
dump_cmd="pt-mysql-summary ${options[*]} -- ${connect_options[*]}"
log "LOCAL_TASKS - ${FUNCNAME[0]}: ${dump_cmd}"
${dump_cmd} 2> "${error_file}" > "${dump_file}"
local last_rc=$?
# shellcheck disable=SC2086
if [ ${last_rc} -ne 0 ]; then
log_error "LOCAL_TASKS - ${FUNCNAME[0]}: pt-mysql-summary to ${dump_file} returned an error ${last_rc}" "${error_file}"
GLOBAL_RC=${E_DUMPFAILED}
else
rm -f "${error_file}"
fi
log "LOCAL_TASKS - ${FUNCNAME[0]}: stop ${dump_file}"
else
log "LOCAL_TASKS - ${FUNCNAME[0]}: 'pt-mysql-summary' not found, unable to dump summary"
fi
}
#######################################################################
# Dump grants of an instance
#
# Arguments:
# --port=[Integer] (default: <blank>)
# --socket=[String] (default: <blank>)
# --user=[String] (default: <blank>)
# --password=[String] (default: <blank>)
# --defaults-file=[String] (default: <blank>)
# --dump-label=[String] (default: "default")
# used as suffix of the dump dir to differenciate multiple instances
#######################################################################
dump_mysql_grants() {
local option_port=""
local option_socket=""
local option_defaults_file=""
local option_user=""
local option_password=""
local option_dump_label=""
# Parse options, based on https://gist.github.com/deshion/10d3cb5f88a21671e17a
while :; do
case ${1:-''} in
--defaults-file)
# defaults-file options, with value separated by space
if [ -n "$2" ]; then
option_defaults_file="${2}"
shift
else
log_error "LOCAL_TASKS - ${FUNCNAME[0]}: '--defaults-file' requires a non-empty option argument."
exit 1
fi
;;
--defaults-file=?*)
# defaults-file options, with value separated by =
option_defaults_file="${1#*=}"
;;
--defaults-file=)
# defaults-file options, without value
log_error "LOCAL_TASKS - ${FUNCNAME[0]}: '--defaults-file' requires a non-empty option argument."
exit 1
;;
--port)
# port options, with value separated by space
if [ -n "$2" ]; then
option_port="${2}"
shift
else
log_error "LOCAL_TASKS - ${FUNCNAME[0]}: '--port' requires a non-empty option argument."
exit 1
fi
;;
--port=?*)
# port options, with value separated by =
option_port="${1#*=}"
;;
--port=)
# port options, without value
log_error "LOCAL_TASKS - ${FUNCNAME[0]}: '--port' requires a non-empty option argument."
exit 1
;;
--socket)
# socket options, with value separated by space
if [ -n "$2" ]; then
option_socket="${2}"
shift
else
log_error "LOCAL_TASKS - ${FUNCNAME[0]}: '--socket' requires a non-empty option argument."
exit 1
fi
;;
--socket=?*)
# socket options, with value separated by =
option_socket="${1#*=}"
;;
--socket=)
# socket options, without value
log_error "LOCAL_TASKS - ${FUNCNAME[0]}: '--socket' requires a non-empty option argument."
exit 1
;;
--user)
# user options, with value separated by space
if [ -n "$2" ]; then
option_user="${2}"
shift
else
log_error "LOCAL_TASKS - ${FUNCNAME[0]}: '--user' requires a non-empty option argument."
exit 1
fi
;;
--user=?*)
# user options, with value separated by =
option_user="${1#*=}"
;;
--user=)
# user options, without value
log_error "LOCAL_TASKS - ${FUNCNAME[0]}: '--user' requires a non-empty option argument."
exit 1
;;
--password)
# password options, with value separated by space
if [ -n "$2" ]; then
option_password="${2}"
shift
else
log_error "LOCAL_TASKS - ${FUNCNAME[0]}: '--password' requires a non-empty option argument."
exit 1
fi
;;
--password=?*)
# password options, with value separated by =
option_password="${1#*=}"
;;
--password=)
# password options, without value
log_error "LOCAL_TASKS - ${FUNCNAME[0]}: '--password' requires a non-empty option argument."
exit 1
;;
--dump-label)
# dump-label options, with value separated by space
if [ -n "$2" ]; then
option_dump_label="${2}"
shift
else
log_error "LOCAL_TASKS - ${FUNCNAME[0]}: '--dump-label' requires a non-empty option argument."
exit 1
fi
;;
--dump-label=?*)
# dump-label options, with value separated by =
option_dump_label="${1#*=}"
;;
--dump-label=)
# dump-label options, without value
log_error "LOCAL_TASKS - ${FUNCNAME[0]}: '--dump-label' requires a non-empty option argument."
exit 1
;;
--)
# End of all options.
shift
break
;;
-?*|[[:alnum:]]*)
# ignore unknown options
log_error "LOCAL_TASKS - ${FUNCNAME[0]}: unknown option '${1}' (ignored)"
;;
*)
# Default case: If no more options then break out of the loop.
break
;;
esac
shift
done
if [ -z "${option_dump_label}" ]; then
if [ -n "${option_port}" ]; then
option_dump_label="${option_port}"
elif [ -n "${option_socket}" ]; then
option_dump_label=$(path_to_str "${option_socket}")
else
option_dump_label="default"
fi
fi
local dump_dir="${LOCAL_BACKUP_DIR}/mysql-${option_dump_label}-grants"
local errors_dir=$(errors_dir_from_dump_dir "${dump_dir}")
rm -rf "${dump_dir}" "${errors_dir}"
mkdir -p "${dump_dir}" "${errors_dir}"
# No need to change recursively, the top directory is enough
chmod 700 "${dump_dir}" "${errors_dir}"
## Dump all grants (requires 'percona-toolkit' package)
if command -v pt-show-grants > /dev/null; then
local error_file="${errors_dir}/all_grants.err"
local dump_file="${dump_dir}/all_grants.sql"
log "LOCAL_TASKS - ${FUNCNAME[0]}: start ${dump_file}"
declare -a options
options=()
if [ -n "${option_defaults_file}" ]; then
options+=(--defaults-file="${option_defaults_file}")
fi
if [ -n "${option_port}" ]; then
options+=(--port="${option_port}")
fi
if [ -n "${option_socket}" ]; then
options+=(--socket="${option_socket}")
fi
if [ -n "${option_user}" ]; then
options+=(--user="${option_user}")
fi
if [ -n "${option_password}" ]; then
options+=(--password="${option_password}")
fi
options+=(--flush)
options+=(--no-header)
dump_cmd="pt-show-grants ${options[*]}"
log "LOCAL_TASKS - ${FUNCNAME[0]}: ${dump_cmd}"
${dump_cmd} 2> "${error_file}" > "${dump_file}"
local last_rc=$?
# shellcheck disable=SC2086
if [ ${last_rc} -ne 0 ]; then
log_error "LOCAL_TASKS - ${FUNCNAME[0]}: pt-show-grants to ${dump_file} returned an error ${last_rc}" "${error_file}"
GLOBAL_RC=${E_DUMPFAILED}
else
rm -f "${error_file}"
fi
log "LOCAL_TASKS - ${FUNCNAME[0]}: stop ${dump_file}"
else
log "LOCAL_TASKS - ${FUNCNAME[0]}: 'pt-show-grants' not found, unable to dump grants"
fi
}
#######################################################################
# Dump a single compressed file of all databases of an instance
# and a file containing only the schema.
#
# Arguments:
# --masterdata (default: <absent>)
# --port=[Integer] (default: <blank>)
# --socket=[String] (default: <blank>)
# --user=[String] (default: <blank>)
# --password=[String] (default: <blank>)
# --defaults-file=[String] (default: <blank>)
# --defaults-extra-file=[String] (default: <blank>)
# --defaults-group-suffix=[String] (default: <blank>)
# --dump-label=[String] (default: "default")
# used as suffix of the dump dir to differenciate multiple instances
# --compress=<gzip|pigz|bzip2|xz|none> (default: "gzip")
# Other options after -- are passed as-is to mysqldump
#######################################################################
dump_mysql_global() {
local option_masterdata=""
local option_port=""
local option_socket=""
local option_defaults_file=""
local option_defaults_extra_file=""
local option_defaults_group_suffix=""
local option_user=""
local option_password=""
local option_dump_label=""
local option_compress=""
local option_others=""
# Parse options, based on https://gist.github.com/deshion/10d3cb5f88a21671e17a
while :; do
case ${1:-''} in
--masterdata)
option_masterdata="--masterdata"
;;
--defaults-file)
# defaults-file options, with value separated by space
if [ -n "$2" ]; then
option_defaults_file="${2}"
shift
else
log_error "LOCAL_TASKS - ${FUNCNAME[0]}: '--defaults-file' requires a non-empty option argument."
exit 1
fi
;;
--defaults-file=?*)
# defaults-file options, with value separated by =
option_defaults_file="${1#*=}"
;;
--defaults-file=)
# defaults-file options, without value
log_error "LOCAL_TASKS - ${FUNCNAME[0]}: '--defaults-file' requires a non-empty option argument."
exit 1
;;
--defaults-extra-file)
# defaults-file options, with value separated by space
if [ -n "$2" ]; then
option_defaults_extra_file="${2}"
shift
else
log_error "LOCAL_TASKS - ${FUNCNAME[0]}: '--defaults-file' requires a non-empty option argument."
exit 1
fi
;;
--defaults-extra-file=?*)
# defaults-extra-file options, with value separated by =
option_defaults_extra_file="${1#*=}"
;;
--defaults-extra-file=)
# defaults-extra-file options, without value
log_error "LOCAL_TASKS - ${FUNCNAME[0]}: '--defaults-extra-file' requires a non-empty option argument."
exit 1
;;
--defaults-group-suffix)
# defaults-group-suffix options, with value separated by space
if [ -n "$2" ]; then
option_defaults_group_suffix="${2}"
shift
else
log_error "LOCAL_TASKS - ${FUNCNAME[0]}: '--defaults-group-suffix' requires a non-empty option argument."
exit 1
fi
;;
--defaults-group-suffix=?*)
# defaults-group-suffix options, with value separated by =
option_defaults_group_suffix="${1#*=}"
;;
--defaults-group-suffix=)
# defaults-group-suffix options, without value
log_error "LOCAL_TASKS - ${FUNCNAME[0]}: '--defaults-group-suffix' requires a non-empty option argument."
exit 1
;;
--port)
# port options, with value separated by space
if [ -n "$2" ]; then
option_port="${2}"
shift
else
log_error "LOCAL_TASKS - ${FUNCNAME[0]}: '--port' requires a non-empty option argument."
exit 1
fi
;;
--port=?*)
# port options, with value separated by =
option_port="${1#*=}"
;;
--port=)
# port options, without value
log_error "LOCAL_TASKS - ${FUNCNAME[0]}: '--port' requires a non-empty option argument."
exit 1
;;
--socket)
# socket options, with value separated by space
if [ -n "$2" ]; then
option_socket="${2}"
shift
else
log_error "LOCAL_TASKS - ${FUNCNAME[0]}: '--socket' requires a non-empty option argument."
exit 1
fi
;;
--socket=?*)
# socket options, with value separated by =
option_socket="${1#*=}"
;;
--socket=)
# socket options, without value
log_error "LOCAL_TASKS - ${FUNCNAME[0]}: '--socket' requires a non-empty option argument."
exit 1
;;
--user)
# user options, with value separated by space
if [ -n "$2" ]; then
option_user="${2}"
shift
else
log_error "LOCAL_TASKS - ${FUNCNAME[0]}: '--user' requires a non-empty option argument."
exit 1
fi
;;
--user=?*)
# user options, with value separated by =
option_user="${1#*=}"
;;
--user=)
# user options, without value
log_error "LOCAL_TASKS - ${FUNCNAME[0]}: '--user' requires a non-empty option argument."
exit 1
;;
--password)
# password options, with value separated by space
if [ -n "$2" ]; then
option_password="${2}"
shift
else
log_error "LOCAL_TASKS - ${FUNCNAME[0]}: '--password' requires a non-empty option argument."
exit 1
fi
;;
--password=?*)
# password options, with value separated by =
option_password="${1#*=}"
;;
--password=)
# password options, without value
log_error "LOCAL_TASKS - ${FUNCNAME[0]}: '--password' requires a non-empty option argument."
exit 1
;;
--dump-label)
# dump-label options, with value separated by space
if [ -n "$2" ]; then
option_dump_label="${2}"
shift
else
log_error "LOCAL_TASKS - ${FUNCNAME[0]}: '--dump-label' requires a non-empty option argument."
exit 1
fi
;;
--dump-label=?*)
# dump-label options, with value separated by =
option_dump_label="${1#*=}"
;;
--dump-label=)
# dump-label options, without value
log_error "LOCAL_TASKS - ${FUNCNAME[0]}: '--dump-label' requires a non-empty option argument."
exit 1
;;
--compress)
# compress options, with value separated by space
if [ -n "$2" ]; then
option_compress="${2}"
shift
else
log_error "LOCAL_TASKS - ${FUNCNAME[0]}: '--compress' requires a non-empty option argument."
exit 1
fi
;;
--compress=?*)
# compress options, with value separated by =
option_compress="${1#*=}"
;;
--compress=)
# compress options, without value
log_error "LOCAL_TASKS - ${FUNCNAME[0]}: '--compress' requires a non-empty option argument."
exit 1
;;
--)
# End of all options.
shift
option_others=${*}
break
;;
-?*|[[:alnum:]]*)
# ignore unknown options
log_error "LOCAL_TASKS - ${FUNCNAME[0]}: unknown option '${1}' (ignored)"
;;
*)
# Default case: If no more options then break out of the loop.
break
;;
esac
shift
done
case "${option_compress}" in
none)
compress_cmd="cat"
dump_ext=""
;;
bzip2|bz|bz2)
compress_cmd="bzip2 --best"
dump_ext=".bz"
;;
xz)
compress_cmd="xz --best"
dump_ext=".xz"
;;
pigz)
compress_cmd="pigz --best"
dump_ext=".gz"
;;
gz|gzip|*)
compress_cmd="gzip --best"
dump_ext=".gz"
;;
esac
if [ -z "${option_dump_label}" ]; then
if [ -n "${option_defaults_group_suffix}" ]; then
option_dump_label="${option_defaults_group_suffix}"
elif [ -n "${option_port}" ]; then
option_dump_label="${option_port}"
elif [ -n "${option_socket}" ]; then
option_dump_label=$(path_to_str "${option_socket}")
else
option_dump_label="default"
fi
fi
## Connection options
declare -a connect_options
connect_options=()
if [ -n "${option_defaults_file}" ]; then
connect_options+=(--defaults-file="${option_defaults_file}")
fi
if [ -n "${option_defaults_extra_file}" ]; then
connect_options+=(--defaults-extra-file="${option_defaults_extra_file}")
fi
if [ -n "${option_defaults_group_suffix}" ]; then
connect_options+=(--defaults-group-suffix="${option_defaults_group_suffix}")
fi
if [ -n "${option_port}" ]; then
connect_options+=(--protocol=tcp)
connect_options+=(--port="${option_port}")
fi
if [ -n "${option_socket}" ]; then
connect_options+=(--protocol=socket)
connect_options+=(--socket="${option_socket}")
fi
if [ -n "${option_user}" ]; then
connect_options+=(--user="${option_user}")
fi
if [ -n "${option_password}" ]; then
connect_options+=(--password="${option_password}")
fi
## Global all databases in one file
local dump_dir="${LOCAL_BACKUP_DIR}/mysql-${option_dump_label}"
local errors_dir=$(errors_dir_from_dump_dir "${dump_dir}")
rm -rf "${dump_dir}" "${errors_dir}"
mkdir -p "${dump_dir}" "${errors_dir}"
# No need to change recursively, the top directory is enough
chmod 700 "${dump_dir}" "${errors_dir}"
local error_file="${errors_dir}/mysqldump.err"
local dump_file="${dump_dir}/mysqldump.sql${dump_ext}"
log "LOCAL_TASKS - ${FUNCNAME[0]}: start ${dump_file}"
declare -a dump_options
dump_options=()
dump_options+=(--opt)
dump_options+=(--force)
dump_options+=(--events)
dump_options+=(--hex-blob)
dump_options+=(--all-databases)
if [ -n "${option_masterdata}" ]; then
dump_options+=("${option_masterdata}")
fi
if [ -n "${option_others}" ]; then
# word splitting is deliberate here
# shellcheck disable=SC2206
dump_options+=(${option_others})
fi
## WARNING : logging and executing the command must be separate
## because otherwise Bash would interpret | and > as strings and not syntax.
dump_cmd="mysqldump ${connect_options[*]} ${dump_options[*]} 2> ${error_file} | ${compress_cmd} > ${dump_file}"
log "LOCAL_TASKS - ${FUNCNAME[0]}: ${dump_cmd}"
mysqldump "${connect_options[@]}" "${dump_options[@]}" 2> "${error_file}" | ${compress_cmd} > "${dump_file}"
local last_rc=$?
# shellcheck disable=SC2086
if [ ${last_rc} -ne 0 ]; then
log_error "LOCAL_TASKS - ${FUNCNAME[0]}: mysqldump returned an error ${last_rc}" "${error_file}"
GLOBAL_RC=${E_DUMPFAILED}
else
rm -f "${error_file}"
fi
log "LOCAL_TASKS - ${FUNCNAME[0]}: stop ${dump_file}"
## Schema only (no data) for each databases
local error_file="${errors_dir}/mysqldump.schema.err"
local dump_file="${dump_dir}/mysqldump.schema.sql"
log "LOCAL_TASKS - ${FUNCNAME[0]}: start ${dump_file}"
declare -a dump_options
dump_options=()
dump_options+=(--force)
dump_options+=(--no-data)
dump_options+=(--all-databases)
if [ -n "${option_others}" ]; then
# word splitting is deliberate here
# shellcheck disable=SC2206
dump_options+=(${option_others})
fi
dump_cmd="mysqldump ${connect_options[*]} ${dump_options[*]}"
log "LOCAL_TASKS - ${FUNCNAME[0]}: ${dump_cmd}"
${dump_cmd} 2> "${error_file}" > "${dump_file}"
local last_rc=$?
# shellcheck disable=SC2086
if [ ${last_rc} -ne 0 ]; then
log_error "LOCAL_TASKS - ${FUNCNAME[0]}: mysqldump returned an error ${last_rc}" "${error_file}"
GLOBAL_RC=${E_DUMPFAILED}
else
rm -f "${error_file}"
fi
log "LOCAL_TASKS - ${FUNCNAME[0]}: stop ${dump_file}"
}
#######################################################################
# Dump a file of each databases of an instance
# and a file containing only the schema.
#
# Arguments:
# --port=[Integer] (default: <blank>)
# --socket=[String] (default: <blank>)
# --user=[String] (default: <blank>)
# --password=[String] (default: <blank>)
# --defaults-file=[String] (default: <blank>)
# --defaults-extra-file=[String] (default: <blank>)
# --defaults-group-suffix=[String] (default: <blank>)
# --dump-label=[String] (default: "default")
# used as suffix of the dump dir to differenciate multiple instances
# --compress=<gzip|pigz|bzip2|xz|none> (default: "gzip")
# Other options after -- are passed as-is to mysqldump
#######################################################################
dump_mysql_per_base() {
local option_port=""
local option_socket=""
local option_defaults_file=""
local option_defaults_extra_file=""
local option_defaults_group_suffix=""
local option_user=""
local option_password=""
local option_dump_label=""
local option_compress=""
local option_others=""
# Parse options, based on https://gist.github.com/deshion/10d3cb5f88a21671e17a
while :; do
case ${1:-''} in
--defaults-file)
# defaults-file options, with value separated by space
if [ -n "$2" ]; then
option_defaults_file="${2}"
shift
else
log_error "LOCAL_TASKS - ${FUNCNAME[0]}: '--defaults-file' requires a non-empty option argument."
exit 1
fi
;;
--defaults-file=?*)
# defaults-file options, with value separated by =
option_defaults_file="${1#*=}"
;;
--defaults-file=)
# defaults-file options, without value
log_error "LOCAL_TASKS - ${FUNCNAME[0]}: '--defaults-file' requires a non-empty option argument."
exit 1
;;
--defaults-extra-file)
# defaults-file options, with value separated by space
if [ -n "$2" ]; then
option_defaults_extra_file="${2}"
shift
else
log_error "LOCAL_TASKS - ${FUNCNAME[0]}: '--defaults-file' requires a non-empty option argument."
exit 1
fi
;;
--defaults-extra-file=?*)
# defaults-extra-file options, with value separated by =
option_defaults_extra_file="${1#*=}"
;;
--defaults-extra-file=)
# defaults-extra-file options, without value
log_error "LOCAL_TASKS - ${FUNCNAME[0]}: '--defaults-extra-file' requires a non-empty option argument."
exit 1
;;
--defaults-group-suffix)
# defaults-group-suffix options, with value separated by space
if [ -n "$2" ]; then
option_defaults_group_suffix="${2}"
shift
else
log_error "LOCAL_TASKS - ${FUNCNAME[0]}: '--defaults-group-suffix' requires a non-empty option argument."
exit 1
fi
;;
--defaults-group-suffix=?*)
# defaults-group-suffix options, with value separated by =
option_defaults_group_suffix="${1#*=}"
;;
--defaults-group-suffix=)
# defaults-group-suffix options, without value
log_error "LOCAL_TASKS - ${FUNCNAME[0]}: '--defaults-group-suffix' requires a non-empty option argument."
exit 1
;;
--port)
# port options, with value separated by space
if [ -n "$2" ]; then
option_port="${2}"
shift
else
log_error "LOCAL_TASKS - ${FUNCNAME[0]}: '--port' requires a non-empty option argument."
exit 1
fi
;;
--port=?*)
# port options, with value separated by =
option_port="${1#*=}"
;;
--port=)
# port options, without value
log_error "LOCAL_TASKS - ${FUNCNAME[0]}: '--port' requires a non-empty option argument."
exit 1
;;
--socket)
# socket options, with value separated by space
if [ -n "$2" ]; then
option_socket="${2}"
shift
else
log_error "LOCAL_TASKS - ${FUNCNAME[0]}: '--socket' requires a non-empty option argument."
exit 1
fi
;;
--socket=?*)
# socket options, with value separated by =
option_socket="${1#*=}"
;;
--socket=)
# socket options, without value
log_error "LOCAL_TASKS - ${FUNCNAME[0]}: '--socket' requires a non-empty option argument."
exit 1
;;
--user)
# user options, with value separated by space
if [ -n "$2" ]; then
option_user="${2}"
shift
else
log_error "LOCAL_TASKS - ${FUNCNAME[0]}: '--user' requires a non-empty option argument."
exit 1
fi
;;
--user=?*)
# user options, with value separated by =
option_user="${1#*=}"
;;
--user=)
# user options, without value
log_error "LOCAL_TASKS - ${FUNCNAME[0]}: '--user' requires a non-empty option argument."
exit 1
;;
--password)
# password options, with value separated by space
if [ -n "$2" ]; then
option_password="${2}"
shift
else
log_error "LOCAL_TASKS - ${FUNCNAME[0]}: '--password' requires a non-empty option argument."
exit 1
fi
;;
--password=?*)
# password options, with value separated by =
option_password="${1#*=}"
;;
--password=)
# password options, without value
log_error "LOCAL_TASKS - ${FUNCNAME[0]}: '--password' requires a non-empty option argument."
exit 1
;;
--dump-label)
# dump-label options, with value separated by space
if [ -n "$2" ]; then
option_dump_label="${2}"
shift
else
log_error "LOCAL_TASKS - ${FUNCNAME[0]}: '--dump-label' requires a non-empty option argument."
exit 1
fi
;;
--dump-label=?*)
# dump-label options, with value separated by =
option_dump_label="${1#*=}"
;;
--dump-label=)
# dump-label options, without value
log_error "LOCAL_TASKS - ${FUNCNAME[0]}: '--dump-label' requires a non-empty option argument."
exit 1
;;
--compress)
# compress options, with value separated by space
if [ -n "$2" ]; then
option_compress="${2}"
shift
else
log_error "LOCAL_TASKS - ${FUNCNAME[0]}: '--compress' requires a non-empty option argument."
exit 1
fi
;;
--compress=?*)
# compress options, with value separated by =
option_compress="${1#*=}"
;;
--compress=)
# compress options, without value
log_error "LOCAL_TASKS - ${FUNCNAME[0]}: '--compress' requires a non-empty option argument."
exit 1
;;
--)
# End of all options.
shift
option_others=${*}
break
;;
-?*|[[:alnum:]]*)
# ignore unknown options
log_error "LOCAL_TASKS - ${FUNCNAME[0]}: unknown option '${1}' (ignored)"
;;
*)
# Default case: If no more options then break out of the loop.
break
;;
esac
shift
done
case "${option_compress}" in
none)
compress_cmd="cat"
dump_ext=""
;;
bzip2|bz|bz2)
compress_cmd="bzip2 --best"
dump_ext=".bz"
;;
xz)
compress_cmd="xz --best"
dump_ext=".xz"
;;
pigz)
compress_cmd="pigz --best"
dump_ext=".gz"
;;
gz|gzip|*)
compress_cmd="gzip --best"
dump_ext=".gz"
;;
esac
if [ -z "${option_dump_label}" ]; then
if [ -n "${option_defaults_group_suffix}" ]; then
option_dump_label="${option_defaults_group_suffix}"
elif [ -n "${option_port}" ]; then
option_dump_label="${option_port}"
elif [ -n "${option_socket}" ]; then
option_dump_label=$(path_to_str "${option_socket}")
else
option_dump_label="default"
fi
fi
## Connection options
declare -a connect_options
connect_options=()
if [ -n "${option_defaults_file}" ]; then
connect_options+=(--defaults-file="${option_defaults_file}")
fi
if [ -n "${option_defaults_extra_file}" ]; then
connect_options+=(--defaults-extra-file="${option_defaults_extra_file}")
fi
if [ -n "${option_defaults_group_suffix}" ]; then
connect_options+=(--defaults-group-suffix="${option_defaults_group_suffix}")
fi
if [ -n "${option_port}" ]; then
connect_options+=(--protocol=tcp)
connect_options+=(--port="${option_port}")
fi
if [ -n "${option_socket}" ]; then
connect_options+=(--protocol=socket)
connect_options+=(--socket="${option_socket}")
fi
if [ -n "${option_user}" ]; then
connect_options+=(--user="${option_user}")
fi
if [ -n "${option_password}" ]; then
connect_options+=(--password="${option_password}")
fi
local dump_dir="${LOCAL_BACKUP_DIR}/mysql-${option_dump_label}-per-base"
local errors_dir=$(errors_dir_from_dump_dir "${dump_dir}")
rm -rf "${dump_dir}" "${errors_dir}"
mkdir -p "${dump_dir}" "${errors_dir}"
# No need to change recursively, the top directory is enough
chmod 700 "${dump_dir}" "${errors_dir}"
databases=$(mysql "${connect_options[@]}" --execute="show databases" --silent --skip-column-names \
| grep --extended-regexp --invert-match "^(Database|information_schema|performance_schema|sys)")
for database in ${databases}; do
local error_file="${errors_dir}/${database}.err"
local dump_file="${dump_dir}/${database}.sql${dump_ext}"
log "LOCAL_TASKS - ${FUNCNAME[0]}: start ${dump_file}"
declare -a dump_options
dump_options=()
dump_options+=(--opt)
dump_options+=(--force)
dump_options+=(--events)
dump_options+=(--hex-blob)
dump_options+=(--databases "${database}")
if [ -n "${option_others}" ]; then
# word splitting is deliberate here
# shellcheck disable=SC2206
dump_options+=(${option_others})
fi
## WARNING : logging and executing the command must be separate
## because otherwise Bash would interpret | and > as strings and not syntax.
log "LOCAL_TASKS - ${FUNCNAME[0]}: mysqldump ${connect_options[*]} ${dump_options[*]} | ${compress_cmd} > ${dump_file}"
mysqldump "${connect_options[@]}" "${dump_options[@]}" 2> "${error_file}" | ${compress_cmd} > "${dump_file}"
local last_rc=$?
# shellcheck disable=SC2086
if [ ${last_rc} -ne 0 ]; then
log_error "LOCAL_TASKS - ${FUNCNAME[0]}: mysqldump returned an error ${last_rc}" "${error_file}"
GLOBAL_RC=${E_DUMPFAILED}
else
rm -f "${error_file}"
fi
log "LOCAL_TASKS - ${FUNCNAME[0]}: stop ${dump_file}"
## Schema only (no data) for each databases
local error_file="${errors_dir}/${database}.schema.err"
local dump_file="${dump_dir}/${database}.schema.sql"
log "LOCAL_TASKS - ${FUNCNAME[0]}: start ${dump_file}"
declare -a dump_options
dump_options=()
dump_options+=(--force)
dump_options+=(--no-data)
dump_options+=(--databases "${database}")
if [ -n "${option_others}" ]; then
# word splitting is deliberate here
# shellcheck disable=SC2206
dump_options+=(${option_others})
fi
dump_cmd="mysqldump ${connect_options[*]} ${dump_options[*]}"
log "LOCAL_TASKS - ${FUNCNAME[0]}: ${dump_cmd}"
${dump_cmd} 2> "${error_file}" > "${dump_file}"
local last_rc=$?
# shellcheck disable=SC2086
if [ ${last_rc} -ne 0 ]; then
log_error "LOCAL_TASKS - ${FUNCNAME[0]}: mysqldump returned an error ${last_rc}" "${error_file}"
GLOBAL_RC=${E_DUMPFAILED}
else
rm -f "${error_file}"
fi
log "LOCAL_TASKS - ${FUNCNAME[0]}: stop ${dump_file}"
done
}
#######################################################################
# Dump "tabs style" separate schema/data for each database of an instance
#
# Arguments:
# --port=[Integer] (default: <blank>)
# --socket=[String] (default: <blank>)
# --user=[String] (default: <blank>)
# --password=[String] (default: <blank>)
# --defaults-file=[String] (default: <blank>)
# --defaults-extra-file=[String] (default: <blank>)
# --defaults-group-suffix=[String] (default: <blank>)
# --dump-label=[String] (default: "default")
# used as suffix of the dump dir to differenciate multiple instances
# --compress=<gzip|pigz|bzip2|xz|none> (default: "gzip")
# Other options after -- are passed as-is to mysqldump
#######################################################################
dump_mysql_tabs() {
local option_port=""
local option_socket=""
local option_defaults_file=""
local option_defaults_extra_file=""
local option_defaults_group_suffix=""
local option_user=""
local option_password=""
local option_dump_label=""
local option_compress=""
local option_others=""
# Parse options, based on https://gist.github.com/deshion/10d3cb5f88a21671e17a
while :; do
case ${1:-''} in
--defaults-file)
# defaults-file options, with value separated by space
if [ -n "$2" ]; then
option_defaults_file="${2}"
shift
else
log_error "LOCAL_TASKS - ${FUNCNAME[0]}: '--defaults-file' requires a non-empty option argument."
exit 1
fi
;;
--defaults-file=?*)
# defaults-file options, with value separated by =
option_defaults_file="${1#*=}"
;;
--defaults-file=)
# defaults-file options, without value
log_error "LOCAL_TASKS - ${FUNCNAME[0]}: '--defaults-file' requires a non-empty option argument."
exit 1
;;
--defaults-extra-file)
# defaults-file options, with value separated by space
if [ -n "$2" ]; then
option_defaults_extra_file="${2}"
shift
else
log_error "LOCAL_TASKS - ${FUNCNAME[0]}: '--defaults-file' requires a non-empty option argument."
exit 1
fi
;;
--defaults-extra-file=?*)
# defaults-extra-file options, with value separated by =
option_defaults_extra_file="${1#*=}"
;;
--defaults-extra-file=)
# defaults-extra-file options, without value
log_error "LOCAL_TASKS - ${FUNCNAME[0]}: '--defaults-extra-file' requires a non-empty option argument."
exit 1
;;
--defaults-group-suffix)
# defaults-group-suffix options, with value separated by space
if [ -n "$2" ]; then
option_defaults_group_suffix="${2}"
shift
else
log_error "LOCAL_TASKS - ${FUNCNAME[0]}: '--defaults-group-suffix' requires a non-empty option argument."
exit 1
fi
;;
--defaults-group-suffix=?*)
# defaults-group-suffix options, with value separated by =
option_defaults_group_suffix="${1#*=}"
;;
--defaults-group-suffix=)
# defaults-group-suffix options, without value
log_error "LOCAL_TASKS - ${FUNCNAME[0]}: '--defaults-group-suffix' requires a non-empty option argument."
exit 1
;;
--port)
# port options, with value separated by space
if [ -n "$2" ]; then
option_port="${2}"
shift
else
log_error "LOCAL_TASKS - ${FUNCNAME[0]}: '--port' requires a non-empty option argument."
exit 1
fi
;;
--port=?*)
# port options, with value separated by =
option_port="${1#*=}"
;;
--port=)
# port options, without value
log_error "LOCAL_TASKS - ${FUNCNAME[0]}: '--port' requires a non-empty option argument."
exit 1
;;
--socket)
# socket options, with value separated by space
if [ -n "$2" ]; then
option_socket="${2}"
shift
else
log_error "LOCAL_TASKS - ${FUNCNAME[0]}: '--socket' requires a non-empty option argument."
exit 1
fi
;;
--socket=?*)
# socket options, with value separated by =
option_socket="${1#*=}"
;;
--socket=)
# socket options, without value
log_error "LOCAL_TASKS - ${FUNCNAME[0]}: '--socket' requires a non-empty option argument."
exit 1
;;
--user)
# user options, with value separated by space
if [ -n "$2" ]; then
option_user="${2}"
shift
else
log_error "LOCAL_TASKS - ${FUNCNAME[0]}: '--user' requires a non-empty option argument."
exit 1
fi
;;
--user=?*)
# user options, with value separated by =
option_user="${1#*=}"
;;
--user=)
# user options, without value
log_error "LOCAL_TASKS - ${FUNCNAME[0]}: '--user' requires a non-empty option argument."
exit 1
;;
--password)
# password options, with value separated by space
if [ -n "$2" ]; then
option_password="${2}"
shift
else
log_error "LOCAL_TASKS - ${FUNCNAME[0]}: '--password' requires a non-empty option argument."
exit 1
fi
;;
--password=?*)
# password options, with value separated by =
option_password="${1#*=}"
;;
--password=)
# password options, without value
log_error "LOCAL_TASKS - ${FUNCNAME[0]}: '--password' requires a non-empty option argument."
exit 1
;;
--dump-label)
# dump-label options, with value separated by space
if [ -n "$2" ]; then
option_dump_label="${2}"
shift
else
log_error "LOCAL_TASKS - ${FUNCNAME[0]}: '--dump-label' requires a non-empty option argument."
exit 1
fi
;;
--dump-label=?*)
# dump-label options, with value separated by =
option_dump_label="${1#*=}"
;;
--dump-label=)
# dump-label options, without value
log_error "LOCAL_TASKS - ${FUNCNAME[0]}: '--dump-label' requires a non-empty option argument."
exit 1
;;
--compress)
# compress options, with value separated by space
if [ -n "$2" ]; then
option_compress="${2}"
shift
else
log_error "LOCAL_TASKS - ${FUNCNAME[0]}: '--compress' requires a non-empty option argument."
exit 1
fi
;;
--compress=?*)
# compress options, with value separated by =
option_compress="${1#*=}"
;;
--compress=)
# compress options, without value
log_error "LOCAL_TASKS - ${FUNCNAME[0]}: '--compress' requires a non-empty option argument."
exit 1
;;
--)
# End of all options.
shift
option_others=${*}
break
;;
-?*|[[:alnum:]]*)
# ignore unknown options
log_error "LOCAL_TASKS - ${FUNCNAME[0]}: unknown option '${1}' (ignored)"
;;
*)
# Default case: If no more options then break out of the loop.
break
;;
esac
shift
done
case "${option_compress}" in
none)
compress_cmd="cat"
dump_ext=""
;;
bzip2|bz|bz2)
compress_cmd="bzip2 --best"
dump_ext=".bz"
;;
xz)
compress_cmd="xz --best"
dump_ext=".xz"
;;
pigz)
compress_cmd="pigz --best"
dump_ext=".gz"
;;
gz|gzip|*)
compress_cmd="gzip --best"
dump_ext=".gz"
;;
esac
if [ -z "${option_dump_label}" ]; then
if [ -n "${option_defaults_group_suffix}" ]; then
option_dump_label="${option_defaults_group_suffix}"
elif [ -n "${option_port}" ]; then
option_dump_label="${option_port}"
elif [ -n "${option_socket}" ]; then
option_dump_label=$(path_to_str "${option_socket}")
else
option_dump_label="default"
fi
fi
## Connection options
declare -a connect_options
connect_options=()
if [ -n "${option_defaults_file}" ]; then
connect_options+=(--defaults-file="${option_defaults_file}")
fi
if [ -n "${option_defaults_extra_file}" ]; then
connect_options+=(--defaults-extra-file="${option_defaults_extra_file}")
fi
if [ -n "${option_defaults_group_suffix}" ]; then
connect_options+=(--defaults-group-suffix="${option_defaults_group_suffix}")
fi
if [ -n "${option_port}" ]; then
connect_options+=(--protocol=tcp)
connect_options+=(--port="${option_port}")
fi
if [ -n "${option_socket}" ]; then
connect_options+=(--protocol=socket)
connect_options+=(--socket="${option_socket}")
fi
if [ -n "${option_user}" ]; then
connect_options+=(--user="${option_user}")
fi
if [ -n "${option_password}" ]; then
connect_options+=(--password="${option_password}")
fi
databases=$(mysql "${connect_options[@]}" --execute="show databases" --silent --skip-column-names \
| grep --extended-regexp --invert-match "^(Database|information_schema|performance_schema|sys)")
for database in ${databases}; do
local dump_dir="${LOCAL_BACKUP_DIR}/mysql-${option_dump_label}-tabs/${database}"
local errors_dir=$(errors_dir_from_dump_dir "${dump_dir}")
rm -rf "${dump_dir}" "${errors_dir}"
mkdir -p "${dump_dir}" "${errors_dir}"
# No need to change recursively, the top directory is enough
chmod 700 "${dump_dir}" "${errors_dir}"
chown -RL mysql "${dump_dir}"
local error_file="${errors_dir}.err"
log "LOCAL_TASKS - ${FUNCNAME[0]}: start ${dump_dir}"
declare -a dump_options
dump_options=()
dump_options+=(--force)
dump_options+=(--quote-names)
dump_options+=(--opt)
dump_options+=(--events)
dump_options+=(--hex-blob)
dump_options+=(--skip-comments)
dump_options+=(--fields-enclosed-by='\"')
dump_options+=(--fields-terminated-by=',')
dump_options+=(--tab="${dump_dir}")
if [ -n "${option_others}" ]; then
# word splitting is deliberate here
# shellcheck disable=SC2206
dump_options+=(${option_others})
fi
dump_options+=("${database}")
dump_cmd="mysqldump ${connect_options[*]} ${dump_options[*]}"
log "LOCAL_TASKS - ${FUNCNAME[0]}: ${dump_cmd}"
${dump_cmd} 2> "${error_file}"
local last_rc=$?
# shellcheck disable=SC2086
if [ ${last_rc} -ne 0 ]; then
log_error "LOCAL_TASKS - ${FUNCNAME[0]}: mysqldump to ${dump_dir} returned an error ${last_rc}" "${error_file}"
GLOBAL_RC=${E_DUMPFAILED}
else
rm -f "${error_file}"
fi
log "LOCAL_TASKS - ${FUNCNAME[0]}: stop ${dump_dir}"
done
}

View file

@ -0,0 +1,343 @@
#!/usr/bin/env bash
# shellcheck disable=SC2034,SC2317,SC2155
#######################################################################
# Dump a single file of all PostgreSQL databases
#
# Arguments:
# --dump-label=[String] (default: "default")
# used as suffix of the dump dir to differenciate multiple instances
# --compress=<gzip|pigz|bzip2|xz|none> (default: "gzip")
# Other options after -- are passed as-is to pg_dump
#######################################################################
dump_postgresql_global() {
local option_dump_label=""
local option_compress=""
local option_others=""
# Parse options, based on https://gist.github.com/deshion/10d3cb5f88a21671e17a
while :; do
case ${1:-''} in
--dump-label)
# dump-label options, with value separated by space
if [ -n "$2" ]; then
option_dump_label="${2}"
shift
else
log_error "LOCAL_TASKS - ${FUNCNAME[0]}: '--dump-label' requires a non-empty option argument."
exit 1
fi
;;
--dump-label=?*)
# dump-label options, with value separated by =
option_dump_label="${1#*=}"
;;
--dump-label=)
# dump-label options, without value
log_error "LOCAL_TASKS - ${FUNCNAME[0]}: '--dump-label' requires a non-empty option argument."
exit 1
;;
--compress)
# compress options, with value separated by space
if [ -n "$2" ]; then
option_compress="${2}"
shift
else
log_error "LOCAL_TASKS - ${FUNCNAME[0]}: '--compress' requires a non-empty option argument."
exit 1
fi
;;
--compress=?*)
# compress options, with value separated by =
option_compress="${1#*=}"
;;
--compress=)
# compress options, without value
log_error "LOCAL_TASKS - ${FUNCNAME[0]}: '--compress' requires a non-empty option argument."
exit 1
;;
--)
# End of all options.
shift
option_others=${*}
break
;;
-?*|[[:alnum:]]*)
# ignore unknown options
log_error "LOCAL_TASKS - ${FUNCNAME[0]}: unknown option '${1}' (ignored)"
;;
*)
# Default case: If no more options then break out of the loop.
break
;;
esac
shift
done
case "${option_compress}" in
none)
compress_cmd="cat"
dump_ext=""
;;
bzip2|bz|bz2)
compress_cmd="bzip2 --best"
dump_ext=".bz"
;;
xz)
compress_cmd="xz --best"
dump_ext=".xz"
;;
pigz)
compress_cmd="pigz --best"
dump_ext=".gz"
;;
gz|gzip|*)
compress_cmd="gzip --best"
dump_ext=".gz"
;;
esac
if [ -z "${option_dump_label}" ]; then
if [ -n "${option_defaults_group_suffix}" ]; then
option_dump_label="${option_defaults_group_suffix}"
elif [ -n "${option_port}" ]; then
option_dump_label="${option_port}"
elif [ -n "${option_socket}" ]; then
option_dump_label=$(path_to_str "${option_socket}")
else
option_dump_label="default"
fi
fi
local dump_dir="${LOCAL_BACKUP_DIR}/postgresql-${option_dump_label}-global"
local errors_dir=$(errors_dir_from_dump_dir "${dump_dir}")
rm -rf "${dump_dir}" "${errors_dir}"
mkdir -p "${dump_dir}" "${errors_dir}"
# No need to change recursively, the top directory is enough
chmod 700 "${dump_dir}" "${errors_dir}"
## example with pg_dumpall and with compression
local error_file="${errors_dir}/pg_dumpall.err"
local dump_file="${dump_dir}/pg_dumpall.sql${dump_ext}"
log "LOCAL_TASKS - ${FUNCNAME[0]}: start ${dump_file}"
declare -a dump_options
dump_options=()
if [ -n "${option_others}" ]; then
# word splitting is deliberate here
# shellcheck disable=SC2206
dump_options+=(${option_others})
fi
dump_cmd="(sudo -u postgres pg_dumpall ${dump_options[*]}) 2> ${error_file} | ${compress_cmd} > ${dump_file}"
log "LOCAL_TASKS - ${FUNCNAME[0]}: ${dump_cmd}"
${dump_cmd}
local last_rc=$?
# shellcheck disable=SC2086
if [ ${last_rc} -ne 0 ]; then
log_error "LOCAL_TASKS - ${FUNCNAME[0]}: pg_dumpall to ${dump_file} returned an error ${last_rc}" "${error_file}"
GLOBAL_RC=${E_DUMPFAILED}
else
rm -f "${error_file}"
fi
log "LOCAL_TASKS - ${FUNCNAME[0]}: stop ${dump_file}"
## example with pg_dumpall and without compression
## WARNING: you need space in ~postgres
# local error_file="${errors_dir}/pg_dumpall.err"
# local dump_file="${dump_dir}/pg_dumpall.sql"
# log "LOCAL_TASKS - ${FUNCNAME[0]}: start ${dump_file}"
#
# (su - postgres -c "pg_dumpall > ~/pg.dump.bak") 2> "${error_file}"
# mv ~postgres/pg.dump.bak "${dump_file}"
#
# log "LOCAL_TASKS - ${FUNCNAME[0]}: stop ${dump_file}"
}
#######################################################################
# Dump a compressed file per database
#
# Arguments: <none>
#######################################################################
dump_postgresql_per_base() {
local option_dump_label=""
local option_compress=""
local option_others=""
# Parse options, based on https://gist.github.com/deshion/10d3cb5f88a21671e17a
while :; do
case ${1:-''} in
--dump-label)
# dump-label options, with value separated by space
if [ -n "$2" ]; then
option_dump_label="${2}"
shift
else
log_error "LOCAL_TASKS - ${FUNCNAME[0]}: '--dump-label' requires a non-empty option argument."
exit 1
fi
;;
--dump-label=?*)
# dump-label options, with value separated by =
option_dump_label="${1#*=}"
;;
--dump-label=)
# dump-label options, without value
log_error "LOCAL_TASKS - ${FUNCNAME[0]}: '--dump-label' requires a non-empty option argument."
exit 1
;;
--compress)
# compress options, with value separated by space
if [ -n "$2" ]; then
option_compress="${2}"
shift
else
log_error "LOCAL_TASKS - ${FUNCNAME[0]}: '--compress' requires a non-empty option argument."
exit 1
fi
;;
--compress=?*)
# compress options, with value separated by =
option_compress="${1#*=}"
;;
--compress=)
# compress options, without value
log_error "LOCAL_TASKS - ${FUNCNAME[0]}: '--compress' requires a non-empty option argument."
exit 1
;;
--)
# End of all options.
shift
option_others=${*}
break
;;
-?*|[[:alnum:]]*)
# ignore unknown options
log_error "LOCAL_TASKS - ${FUNCNAME[0]}: unknown option '${1}' (ignored)"
;;
*)
# Default case: If no more options then break out of the loop.
break
;;
esac
shift
done
case "${option_compress}" in
none)
compress_cmd="cat"
dump_ext=""
;;
bzip2|bz|bz2)
compress_cmd="bzip2 --best"
dump_ext=".bz"
;;
xz)
compress_cmd="xz --best"
dump_ext=".xz"
;;
pigz)
compress_cmd="pigz --best"
dump_ext=".gz"
;;
gz|gzip|*)
compress_cmd="gzip --best"
dump_ext=".gz"
;;
esac
if [ -z "${option_dump_label}" ]; then
if [ -n "${option_defaults_group_suffix}" ]; then
option_dump_label="${option_defaults_group_suffix}"
elif [ -n "${option_port}" ]; then
option_dump_label="${option_port}"
elif [ -n "${option_socket}" ]; then
option_dump_label=$(path_to_str "${option_socket}")
else
option_dump_label="default"
fi
fi
local dump_dir="${LOCAL_BACKUP_DIR}/postgresql-${option_dump_label}-per-base"
local errors_dir=$(errors_dir_from_dump_dir "${dump_dir}")
rm -rf "${dump_dir}" "${errors_dir}"
mkdir -p "${dump_dir}" "${errors_dir}"
# No need to change recursively, the top directory is enough
chmod 700 "${dump_dir}" "${errors_dir}"
(
# shellcheck disable=SC2164
cd /var/lib/postgresql
databases=$(sudo -u postgres psql -U postgres -lt | awk -F \| '{print $1}' | grep -v "template.*")
for database in ${databases} ; do
local error_file="${errors_dir}/${database}.err"
local dump_file="${dump_dir}/${database}.sql${dump_ext}"
log "LOCAL_TASKS - ${FUNCNAME[0]}: start ${dump_file}"
declare -a dump_options
dump_options=()
dump_options+=(--create)
dump_options+=(-U postgres)
dump_options+=(-d "${database}")
if [ -n "${option_others}" ]; then
# word splitting is deliberate here
# shellcheck disable=SC2206
dump_options+=(${option_others})
fi
dump_cmd="(sudo -u postgres /usr/bin/pg_dump ${dump_options[*]}) 2> ${error_file} | ${compress_cmd} > ${dump_file}"
log "LOCAL_TASKS - ${FUNCNAME[0]}: ${dump_cmd}"
${dump_cmd}
local last_rc=$?
# shellcheck disable=SC2086
if [ ${last_rc} -ne 0 ]; then
log_error "LOCAL_TASKS - ${FUNCNAME[0]}: pg_dump to ${dump_file} returned an error ${last_rc}" "${error_file}"
GLOBAL_RC=${E_DUMPFAILED}
else
rm -f "${error_file}"
fi
log "LOCAL_TASKS - ${FUNCNAME[0]}: stop ${dump_file}"
done
)
}
#######################################################################
# Dump a compressed file per database
#
# Arguments: <none>
#
# TODO: add arguments to include/exclude tables
#######################################################################
dump_postgresql_filtered() {
local dump_dir="${LOCAL_BACKUP_DIR}/postgresql-filtered"
local errors_dir=$(errors_dir_from_dump_dir "${dump_dir}")
rm -rf "${dump_dir}" "${errors_dir}"
mkdir -p "${dump_dir}" "${errors_dir}"
# No need to change recursively, the top directory is enough
chmod 700 "${dump_dir}" "${errors_dir}"
local error_file="${errors_dir}/pg-backup.err"
local dump_file="${dump_dir}/pg-backup.tar"
log "LOCAL_TASKS - ${FUNCNAME[0]}: start ${dump_file}"
## example with all tables from MYBASE excepts TABLE1 and TABLE2
# pg_dump -p 5432 -h 127.0.0.1 -U USER --clean -F t --inserts -f "${dump_file}" -t 'TABLE1' -t 'TABLE2' MYBASE 2> "${error_file}"
## example with only TABLE1 and TABLE2 from MYBASE
# pg_dump -p 5432 -h 127.0.0.1 -U USER --clean -F t --inserts -f "${dump_file}" -T 'TABLE1' -T 'TABLE2' MYBASE 2> "${error_file}"
local last_rc=$?
# shellcheck disable=SC2086
if [ ${last_rc} -ne 0 ]; then
log_error "LOCAL_TASKS - ${FUNCNAME[0]}: pg_dump to ${dump_file} returned an error ${last_rc}" "${error_file}"
GLOBAL_RC=${E_DUMPFAILED}
else
rm -f "${error_file}"
fi
log "LOCAL_TASKS - ${FUNCNAME[0]}: stop ${dump_file}"
}

View file

@ -0,0 +1,466 @@
#!/usr/bin/env bash
# shellcheck disable=SC2034,SC2317
readonly VERSION="24.04.1"
# set all programs to C language (english)
export LC_ALL=C
# If expansion is attempted on an unset variable or parameter, the shell prints an
# error message, and, if not interactive, exits with a non-zero status.
set -o nounset
# The pipeline's return status is the value of the last (rightmost) command
# to exit with a non-zero status, or zero if all commands exit successfully.
set -o pipefail
# Enable trace mode if called with environment variable TRACE=1
if [[ "${TRACE-0}" == "1" ]]; then
set -o xtrace
fi
source "${LIBDIR}/utilities.sh"
source "${LIBDIR}/dump/elasticsearch.sh"
source "${LIBDIR}/dump/mysql.sh"
source "${LIBDIR}/dump/postgresql.sh"
source "${LIBDIR}/dump/misc.sh"
# Called from main, it is wrapping the local_tasks function defined in the real script
local_tasks_wrapper() {
log "START LOCAL_TASKS"
# Remove old log directories (recursively)
find "${LOCAL_BACKUP_DIR}/" -type d -name "${PROGNAME}.errors-*" -ctime +30 -exec rm -rf \;
local_tasks_type="$(type -t local_tasks)"
if [ "${local_tasks_type}" = "function" ]; then
local_tasks
else
log_error "There is no 'local_tasks' function to execute"
fi
# TODO: check if this is still needed
# print_error_files_content
log "STOP LOCAL_TASKS"
}
# Called from main, it is wrapping the sync_tasks function defined in the real script
sync_tasks_wrapper() {
declare -a SERVERS # Indexed array for server/port values
declare -a RSYNC_INCLUDES # Indexed array for includes
declare -a RSYNC_EXCLUDES # Indexed array for excludes
case "${SYSTEM}" in
linux)
# NOTE: remember to single-quote paths if they contain globs (*)
# and you want to defer expansion
declare -a rsync_default_includes=(
/bin
/boot
/lib
/opt
/sbin
/usr
)
;;
*bsd)
# NOTE: remember to single-quote paths if they contain globs (*)
# and you want to defer expansion
declare -a rsync_default_includes=(
/bin
/bsd
/sbin
/usr
)
;;
*)
echo "Unknown system '${SYSTEM}'" >&2
exit 1
;;
esac
if [ -f "${CANARY_FILE}" ]; then
rsync_default_includes+=("${CANARY_FILE}")
fi
readonly rsync_default_includes
# NOTE: remember to single-quote paths if they contain globs (*)
# and you want to defer expansion
declare -a rsync_default_excludes=(
/dev
/proc
/run
/sys
/tmp
/usr/doc
/usr/obj
/usr/share/doc
/usr/src
/var/apt
/var/cache
'/var/db/munin/*.tmp'
/var/lib/amavis/amavisd.sock
/var/lib/amavis/tmp
/var/lib/amavis/virusmails
'/var/lib/clamav/*.tmp'
/var/lib/elasticsearch
/var/lib/metche
/var/lib/mongodb
'/var/lib/munin/*tmp*'
/var/lib/mysql
/var/lib/php/sessions
/var/lib/php5
/var/lib/postgres
/var/lib/postgresql
/var/lib/sympa
/var/lock
/var/run
/var/spool/postfix
/var/spool/smtpd
/var/spool/squid
/var/state
/var/tmp
lost+found
'.nfs.*'
'lxc/*/rootfs/tmp'
'lxc/*/rootfs/usr/doc'
'lxc/*/rootfs/usr/obj'
'lxc/*/rootfs/usr/share/doc'
'lxc/*/rootfs/usr/src'
'lxc/*/rootfs/var/apt'
'lxc/*/rootfs/var/cache'
'lxc/*/rootfs/var/lib/php5'
'lxc/*/rootfs/var/lib/php/sessions'
'lxc/*/rootfs/var/lock'
'lxc/*/rootfs/var/run'
'lxc/*/rootfs/var/state'
'lxc/*/rootfs/var/tmp'
/home/mysqltmp
)
readonly rsync_default_excludes
sync_tasks_type="$(type -t sync_tasks)"
if [ "${sync_tasks_type}" = "function" ]; then
sync_tasks
else
log_error "There is no 'sync_tasks' function to execute"
fi
}
sync() {
local sync_name=${1}
local -a rsync_servers=("${!2}")
local -a rsync_includes=("${!3}")
local -a rsync_excludes=("${!4}")
## Initialize variable to store SSH connection errors
declare -a SSH_ERRORS=()
log "START SYNC_TASKS - sync=${sync_name}"
# echo "### sync ###"
# for server in "${rsync_servers[@]}"; do
# echo "server: ${server}"
# done
# for include in "${rsync_includes[@]}"; do
# echo "include: ${include}"
# done
# for exclude in "${rsync_excludes[@]}"; do
# echo "exclude: ${exclude}"
# done
local -i n=0
local server=""
if [ "${SERVERS_FALLBACK}" = "1" ]; then
# We try to find a suitable server
while :; do
server=$(pick_server ${n} "${sync_name}")
rc=$?
if [ ${rc} != 0 ]; then
GLOBAL_RC=${E_NOSRVAVAIL}
log "STOP SYNC_TASKS - sync=${sync_name}'"
return
fi
if test_server "${server}"; then
break
else
server=""
n=$(( n + 1 ))
fi
done
else
# we force the server
server=$(pick_server "${n}" "${sync_name}")
fi
rsync_server=$(echo "${server}" | cut -d':' -f1)
rsync_port=$(echo "${server}" | cut -d':' -f2)
log "SYNC_TASKS - sync=${sync_name}: use ${server}"
# Rsync complete log file for the current run
RSYNC_LOGFILE="/var/log/${PROGNAME}.${sync_name}.rsync.log"
# Rsync stats for the current run
RSYNC_STATSFILE="/var/log/${PROGNAME}.${sync_name}.rsync-stats.log"
# reset Rsync log file
if [ -n "$(command -v truncate)" ]; then
truncate -s 0 "${RSYNC_LOGFILE}"
truncate -s 0 "${RSYNC_STATSFILE}"
else
printf "" > "${RSYNC_LOGFILE}"
printf "" > "${RSYNC_STATSFILE}"
fi
# Initialize variable here, we need it later
local -a mtree_files=()
if [ "${MTREE_ENABLED}" = "1" ]; then
mtree_bin=$(command -v mtree)
if [ -n "${mtree_bin}" ]; then
# Dump filesystem stats with mtree
log "SYNC_TASKS - sync=${sync_name}: start mtree"
# Loop over Rsync includes
for i in "${!rsync_includes[@]}"; do
include="${rsync_includes[i]}"
if [ -d "${include}" ]; then
# … but exclude for mtree what will be excluded by Rsync
mtree_excludes_file="$(mktemp --tmpdir "${PROGNAME}.${sync_name}.mtree-excludes.XXXXXX")"
add_to_temp_files "${mtree_excludes_file}"
for j in "${!rsync_excludes[@]}"; do
echo "${rsync_excludes[j]}" | grep -E "^([^/]|${include})" | sed -e "s|^${include}|.|" >> "${mtree_excludes_file}"
done
mtree_file="/var/log/evobackup.$(basename "${include}").mtree"
add_to_temp_files "${mtree_file}"
${mtree_bin} -x -c -p "${include}" -X "${mtree_excludes_file}" > "${mtree_file}"
mtree_files+=("${mtree_file}")
fi
done
if [ "${#mtree_files[@]}" -le 0 ]; then
log_error "SYNC_TASKS - ${sync_name}: ERROR: mtree didn't produce any file"
fi
log "SYNC_TASKS - sync=${sync_name}: stop mtree (files: ${mtree_files[*]})"
else
log "SYNC_TASKS - sync=${sync_name}: skip mtree (missing)"
fi
else
log "SYNC_TASKS - sync=${sync_name}: skip mtree (disabled)"
fi
rsync_bin=$(command -v rsync)
# Build the final Rsync command
# Rsync main options
rsync_main_args=()
rsync_main_args+=(--archive)
rsync_main_args+=(--itemize-changes)
rsync_main_args+=(--quiet)
rsync_main_args+=(--stats)
rsync_main_args+=(--human-readable)
rsync_main_args+=(--relative)
rsync_main_args+=(--partial)
rsync_main_args+=(--delete)
rsync_main_args+=(--delete-excluded)
rsync_main_args+=(--force)
rsync_main_args+=(--ignore-errors)
rsync_main_args+=(--log-file "${RSYNC_LOGFILE}")
rsync_main_args+=(--rsh "ssh -p ${rsync_port} -o 'ConnectTimeout ${SSH_CONNECT_TIMEOUT}'")
# Rsync excludes
for i in "${!rsync_excludes[@]}"; do
rsync_main_args+=(--exclude "${rsync_excludes[i]}")
done
# Rsync local sources
rsync_main_args+=("${rsync_includes[@]}")
# Rsync remote destination
rsync_main_args+=("root@${rsync_server}:${REMOTE_BACKUP_DIR}/")
# … log it
log "SYNC_TASKS - sync=${sync_name}: Rsync main command : ${rsync_bin} ${rsync_main_args[*]}"
# … execute it
${rsync_bin} "${rsync_main_args[@]}"
rsync_main_rc=$?
# Copy last lines of rsync log to the main log
tail -n 30 "${RSYNC_LOGFILE}" >> "${LOGFILE}"
# Copy Rsync stats to special file
tail -n 30 "${RSYNC_LOGFILE}" | grep --invert-match --extended-regexp " [\<\>ch\.\*]\S{10} " > "${RSYNC_STATSFILE}"
# We ignore rc=24 (vanished files)
if [ ${rsync_main_rc} -ne 0 ] && [ ${rsync_main_rc} -ne 24 ]; then
log_error "SYNC_TASKS - sync=${sync_name}: Rsync main command returned an error ${rsync_main_rc}" "${LOGFILE}"
GLOBAL_RC=${E_SYNCFAILED}
else
# Build the report Rsync command
local -a rsync_report_args
rsync_report_args=()
# Rsync options
rsync_report_args+=(--rsh "ssh -p ${rsync_port} -o 'ConnectTimeout ${SSH_CONNECT_TIMEOUT}'")
# Rsync local sources
if [ "${#mtree_files[@]}" -gt 0 ]; then
# send mtree files if there is any
rsync_report_args+=("${mtree_files[@]}")
fi
if [ -f "${RSYNC_LOGFILE}" ]; then
# send rsync full log file if it exists
rsync_report_args+=("${RSYNC_LOGFILE}")
fi
if [ -f "${RSYNC_STATSFILE}" ]; then
# send rsync stats log file if it exists
rsync_report_args+=("${RSYNC_STATSFILE}")
fi
# Rsync remote destination
rsync_report_args+=("root@${rsync_server}:${REMOTE_LOG_DIR}/")
# … log it
log "SYNC_TASKS - sync=${sync_name}: Rsync report command : ${rsync_bin} ${rsync_report_args[*]}"
# … execute it
${rsync_bin} "${rsync_report_args[@]}"
fi
log "STOP SYNC_TASKS - sync=${sync_name}"
}
setup() {
# Default return-code (0 == succes)
GLOBAL_RC=0
# Possible error codes
readonly E_NOSRVAVAIL=21 # No server is available
readonly E_SYNCFAILED=20 # Failed sync task
readonly E_DUMPFAILED=10 # Failed dump task
# explicit PATH
PATH=/sbin:/usr/sbin:/bin:/usr/bin:/usr/local/sbin:/usr/local/bin
# System name (linux, openbsd…)
: "${SYSTEM:=$(uname | tr '[:upper:]' '[:lower:]')}"
# Hostname (for logs and notifications)
: "${HOSTNAME:=$(hostname)}"
# Store pid in a file named after this program's name
: "${PROGNAME:=$(basename "$0")}"
: "${PIDFILE:="/var/run/${PROGNAME}.pid"}"
# Customize the log path if you want multiple scripts to have separate log files
: "${LOGFILE:="/var/log/evobackup.log"}"
# Canary file to update before executing tasks
: "${CANARY_FILE:="/zzz_evobackup_canary"}"
# Date format for log messages
: "${DATE_FORMAT:="%Y-%m-%d %H:%M:%S"}"
# Should we fallback on other servers when the first one is unreachable?
: "${SERVERS_FALLBACK:=1}"
# timeout (in seconds) for SSH connections
: "${SSH_CONNECT_TIMEOUT:=90}"
: "${LOCAL_BACKUP_DIR:="/home/backup"}"
# shellcheck disable=SC2174
mkdir -p -m 700 "${LOCAL_BACKUP_DIR}"
: "${ERRORS_DIR:="${LOCAL_BACKUP_DIR}/${PROGNAME}.errors-${START_TIME}"}"
# shellcheck disable=SC2174
mkdir -p -m 700 "${ERRORS_DIR}"
# Backup directory on remote server
: "${REMOTE_BACKUP_DIR:="/var/backup"}"
# Log directory in remote server
: "${REMOTE_LOG_DIR:="/var/log"}"
# Email address for notifications
: "${MAIL:="root"}"
# Email subject for notifications
: "${MAIL_SUBJECT:="[info] EvoBackup - Client ${HOSTNAME}"}"
# Enable/disable local tasks (default: enabled)
: "${LOCAL_TASKS:=1}"
# Enable/disable sync tasks (default: enabled)
: "${SYNC_TASKS:=1}"
# Enable/disable mtree (default: enabled)
: "${MTREE_ENABLED:=1}"
# If "setup_custom" exists and is a function, let's call it
setup_custom_type="$(type -t setup_custom)"
if [ "${setup_custom_type}" = "function" ]; then
setup_custom
fi
## Force umask
umask 077
# Initialize a list of temporary files
declare -a TEMP_FILES=()
# Any file in this list will be deleted when the program exits
trap "cleanup" EXIT
}
run_evobackup() {
# Start timer
START_EPOCH=$(/bin/date +%s)
START_TIME=$(/bin/date +"%Y%m%d%H%M%S")
# Configure variables and environment
setup
log "START GLOBAL - VERSION=${VERSION} LOCAL_TASKS=${LOCAL_TASKS} SYNC_TASKS=${SYNC_TASKS}"
# /!\ Only one backup processus can run at the sametime /!\
# Based on PID file, kill any running process before continuing
enforce_single_process "${PIDFILE}"
# Update canary to keep track of each run
update-evobackup-canary --who "${PROGNAME}" --file "${CANARY_FILE}"
if [ "${LOCAL_TASKS}" = "1" ]; then
local_tasks_wrapper
fi
if [ "${SYNC_TASKS}" = "1" ]; then
sync_tasks_wrapper
fi
STOP_EPOCH=$(/bin/date +%s)
case "${SYSTEM}" in
*bsd)
start_time=$(/bin/date -f "%s" -j "${START_EPOCH}" +"${DATE_FORMAT}")
stop_time=$(/bin/date -f "%s" -j "${STOP_EPOCH}" +"${DATE_FORMAT}")
;;
*)
start_time=$(/bin/date --date="@${START_EPOCH}" +"${DATE_FORMAT}")
stop_time=$(/bin/date --date="@${STOP_EPOCH}" +"${DATE_FORMAT}")
;;
esac
duration=$(( STOP_EPOCH - START_EPOCH ))
log "STOP GLOBAL - start='${start_time}' stop='${stop_time}' duration=${duration}s"
send_mail
exit ${GLOBAL_RC}
}

View file

@ -0,0 +1,143 @@
#!/usr/bin/env bash
# Output a message to the log file
log() {
local msg="${1:-$(cat /dev/stdin)}"
local pid=$$
printf "[%s] %s[%s]: %s\\n" \
"$(/bin/date +"${DATE_FORMAT}")" "${PROGNAME}" "${pid}" "${msg}" \
>> "${LOGFILE}"
}
log_error() {
local error_msg=${1}
local error_file=${2:-""}
if [ -n "${error_file}" ] && [ -f "${error_file}" ]; then
printf "\n### %s\n" "${error_msg}" >&2
# shellcheck disable=SC2046
if [ $(wc -l "${error_file}" | cut -d " " -f 1) -gt 30 ]; then
printf "~~~{%s (tail -30)}\n" "${error_file}" >&2
tail -n 30 "${error_file}" >&2
else
printf "~~~{%s}\n" "${error_file}" >&2
cat "${error_file}" >&2
fi
printf "~~~\n" >&2
log "${error_msg}, check ${error_file}"
else
printf "\n### %s\n" "${error_msg}" >&2
log "${error_msg}"
fi
}
add_to_temp_files() {
TEMP_FILES+=("${1}")
}
# Remove all temporary file created during the execution
cleanup() {
# shellcheck disable=SC2086
rm -f "${TEMP_FILES[@]}"
find "${ERRORS_DIR}" -type d -empty -delete
}
enforce_single_process() {
local pidfile=$1
if [ -e "${pidfile}" ]; then
pid=$(cat "${pidfile}")
# Does process still exist?
if kill -0 "${pid}" 2> /dev/null; then
# Killing the childs of evobackup.
for ppid in $(pgrep -P "${pid}"); do
kill -9 "${ppid}";
done
# Then kill the main PID.
kill -9 "${pid}"
printf "%s is still running (PID %s). Process has been killed" "$0" "${pid}\\n" >&2
else
rm -f "${pidfile}"
fi
fi
add_to_temp_files "${pidfile}"
echo "$$" > "${pidfile}"
}
# Build the error directory (inside ERRORS_DIR) based on the dump directory path
errors_dir_from_dump_dir() {
local dump_dir=$1
local relative_path=$(realpath --relative-to="${LOCAL_BACKUP_DIR}" "${dump_dir}")
# return absolute path
realpath --canonicalize-missing "${ERRORS_DIR}/${relative_path}"
}
# Call test_server with "HOST:PORT" string
# It will return with 0 if the server is reachable.
# It will return with 1 and a message on stderr if not.
test_server() {
local item=$1
# split HOST and PORT from the input string
local host=$(echo "${item}" | cut -d':' -f1)
local port=$(echo "${item}" | cut -d':' -f2)
local new_error
# Test if the server is accepting connections
ssh -q -o "ConnectTimeout ${SSH_CONNECT_TIMEOUT}" "${host}" -p "${port}" -t "exit"
# shellcheck disable=SC2181
if [ $? = 0 ]; then
# SSH connection is OK
return 0
else
# SSH connection failed
new_error=$(printf "Failed to connect to \`%s' within %s seconds" "${item}" "${SSH_CONNECT_TIMEOUT}")
log "${new_error}"
SSH_ERRORS+=("${new_error}")
return 1
fi
}
# Call pick_server with an optional positive integer to get the nth server in the list.
pick_server() {
local -i increment=${1:-0}
local -i list_length=${#SERVERS[@]}
local sync_name=${2:""}
if (( increment >= list_length )); then
# We've reached the end of the list
new_error="No more server available"
new_error="${new_error} for sync '${sync_name}'"
log "${new_error}"
SSH_ERRORS+=("${new_error}")
# Log errors to stderr
for i in "${!SSH_ERRORS[@]}"; do
printf "%s\n" "${SSH_ERRORS[i]}" >&2
done
return 1
fi
# Extract the day of month, without leading 0 (which would give an octal based number)
today=$(/bin/date +%e)
# A salt is useful to randomize the starting point in the list
# but stay identical each time it's called for a server (based on hostname).
salt=$(hostname | cksum | cut -d' ' -f1)
# Pick an integer between 0 and the length of the SERVERS list
# It changes each day
n=$(( (today + salt + increment) % list_length ))
echo "${SERVERS[n]}"
}
send_mail() {
tail -20 "${LOGFILE}" | mail -s "${MAIL_SUBJECT}" "${MAIL}"
}
path_to_str() {
echo "${1}" | sed -e 's|^/||; s|/$||; s|/|:|g'
}

View file

@ -0,0 +1,326 @@
#!/usr/bin/env bash
#
# Evobackup client
# See https://gitea.evolix.org/evolix/evobackup
#
# This is a generated backup script made by:
# command: @COMMAND@
# version: @VERSION@
# date: @DATE@
#######################################################################
#
# You must configure the MAIL variable to receive notifications.
#
# There is some optional configuration that you can do
# at the end of this script.
#
#######################################################################
# Email adress for notifications
MAIL=__NOTIFICATION_MAIL__
#######################################################################
#
# The "sync_tasks" function will be called by the "run_evobackup" function.
#
# You can customize the variables:
# * "SYNC_NAME" (String)
# * "SERVERS" (Array of HOST:PORT)
# * "RSYNC_INCLUDES" (Array of paths to include)
# * "RSYNC_EXCLUDES" (Array of paths to exclude)
#
# WARNING: remember to single-quote paths if they contain globs (*)
# and you want to pass them as-is to Rsync.
#
# The "sync" function can be called multiple times
# with a different set of variables.
# That way you can to sync to various destinations.
#
# Default includes/excludes are defined in the "main" library,
# referenced at this end of this file.
#
#######################################################################
# shellcheck disable=SC2034
sync_tasks() {
########## System-only backup (to Evolix servers) #################
SYNC_NAME="evolix-system"
SERVERS=(
__SRV0_HOST__:__SRV0_PORT__
__SRV1_HOST__:__SRV1_PORT__
)
RSYNC_INCLUDES=(
"${rsync_default_includes[@]}"
/etc
/root
/var
)
RSYNC_EXCLUDES=(
"${rsync_default_excludes[@]}"
)
sync "${SYNC_NAME}" "SERVERS[@]" "RSYNC_INCLUDES[@]" "RSYNC_EXCLUDES[@]"
########## Full backup (to client servers) ########################
### SYNC_NAME="client-full"
### SERVERS=(
### client-backup00.evolix.net:2221
### client-backup01.evolix.net:2221
### )
### RSYNC_INCLUDES=(
### "${rsync_default_includes[@]}"
### /etc
### /root
### /var
### /home
### /srv
### )
### RSYNC_EXCLUDES=(
### "${rsync_default_excludes[@]}"
### )
### sync "${SYNC_NAME}" "SERVERS[@]" "RSYNC_INCLUDES[@]" "RSYNC_EXCLUDES[@]"
}
#######################################################################
#
# The "local_tasks" function will be called by the "run_evobackup" function.
#
# You can call any available "dump_xxx" function
# (usually installed at /usr/local/lib/evobackup/dump-*.sh)
#
# You can also write some custom functions and call them.
# A "dump_custom" example is available further down.
#
#######################################################################
local_tasks() {
########## Server state ###########
# Run dump-server-state to extract system information
#
# Options : any dump-server-state supported option
# (except --dump-dir that will be overwritten)
# See 'dump-server-state -h' for details.
#
dump_server_state
########## MySQL ##################
# Very common strategy for a single instance server with default configuration :
#
### dump_mysql_global; dump_mysql_grants; dump_mysql_summary
#
# See below for details regarding dump functions for MySQL/MariaDB
# Dump all databases in a single compressed file
#
# Options :
# --masterdata (default: <absent>)
# --port=[Integer] (default: <blank>)
# --socket=[String] (default: <blank>)
# --user=[String] (default: <blank>)
# --password=[String] (default: <blank>)
# --defaults-file=[String] (default: <blank>)
# --defaults-extra-file=[String] (default: <blank>)
# --defaults-group-suffix=[String] (default: <blank>)
# --dump-label=[String] (default: "default")
# used as suffix of the dump dir to differenciate multiple instances
#
### dump_mysql_global
# Dump each database separately, in a compressed file
#
# Options :
# --port=[Integer] (default: <blank>)
# --socket=[String] (default: <blank>)
# --user=[String] (default: <blank>)
# --password=[String] (default: <blank>)
# --defaults-file=[String] (default: <blank>)
# --defaults-extra-file=[String] (default: <blank>)
# --defaults-group-suffix=[String] (default: <blank>)
# --dump-label=[String] (default: "default")
# used as suffix of the dump dir to differenciate multiple instances
#
### dump_mysql_per_base
# Dump permissions of an instance (using pt-show-grants)
#
# Options :
# --port=[Integer] (default: <blank>)
# --socket=[String] (default: <blank>)
# --user=[String] (default: <blank>)
# --password=[String] (default: <blank>)
# --defaults-file=[String] (default: <blank>)
# --dump-label=[String] (default: "default")
# used as suffix of the dump dir to differenciate multiple instances
#
# WARNING - unsupported options :
# --defaults-extra-file
# --defaults-group-suffix
# You have to provide credentials manually
#
### dump_mysql_grants
# Dump complete summary of an instance (using pt-mysql-summary)
#
# Options :
# --port=[Integer] (default: <blank>)
# --socket=[String] (default: <blank>)
# --user=[String] (default: <blank>)
# --password=[String] (default: <blank>)
# --defaults-file=[String] (default: <blank>)
# --defaults-extra-file=[String] (default: <blank>)
# --defaults-group-suffix=[String] (default: <blank>)
# --dump-label=[String] (default: "default")
# used as suffix of the dump dir to differenciate multiple instances
#
### dump_mysql_summary
# Dump each table in separate schema/data files
#
# Options :
# --port=[Integer] (default: <blank>)
# --socket=[String] (default: <blank>)
# --user=[String] (default: <blank>)
# --password=[String] (default: <blank>)
# --defaults-file=[String] (default: <blank>)
# --defaults-extra-file=[String] (default: <blank>)
# --defaults-group-suffix=[String] (default: <blank>)
# --dump-label=[String] (default: "default")
# used as suffix of the dump dir to differenciate multiple instances
#
### dump_mysql_tabs
########## PostgreSQL #############
# Dump all databases in a single file (compressed or not)
#
### dump_postgresql_global
# Dump a specific databse with only some tables, or all but some tables (must be configured)
#
### dump_postgresql_filtered
# Dump each database separately, in a compressed file
#
### dump_postgresql_per_base
########## MongoDB ################
### dump_mongodb [--user=foo] [--password=123456789]
########## Redis ##################
# Copy data file for all instances
#
### dump_redis [--instances=<all|instance1|instance2>]
########## Elasticsearch ##########
# Snapshot data for a single-node cluster
#
### dump_elasticsearch_snapshot_singlenode [--protocol=http] [--host=localhost] [--port=9200] [--user=foo] [--password=123456789] [--repository=snaprepo] [--snapshot=snapshot.daily]
# Snapshot data for a multi-node cluster
#
### dump_elasticsearch_snapshot_multinode [--protocol=http] [--host=localhost] [--port=9200] [--user=foo] [--password=123456789] [--repository=snaprepo] [--snapshot=snapshot.daily] [--nfs-server=192.168.2.1]
########## RabbitMQ ###############
### dump_rabbitmq
########## MegaCli ################
# Copy RAID config
#
### dump_megacli_config
# Dump file access control lists
#
### dump_facl
########## OpenLDAP ###############
### dump_ldap
########## Network ################
# Dump network routes with mtr and traceroute (warning: could be long with aggressive firewalls)
#
### dump_traceroute --targets=host_or_ip[,host_or_ip]
dump_traceroute --targets=8.8.8.8,www.evolix.fr,travaux.evolix.net
# No-op, in case nothing is enabled
:
}
# This is an example for a custom dump function
# Uncomment, customize and call it from the "local_tasks" function
### dump_custom() {
### # Set dump and errors directories and files
### local dump_dir="${LOCAL_BACKUP_DIR}/custom"
### local dump_file="${dump_dir}/dump.gz"
### local errors_dir=$(errors_dir_from_dump_dir "${dump_dir}")
### local error_file="${errors_dir}/dump.err"
###
### # Reset dump and errors directories
### rm -rf "${dump_dir}" "${errors_dir}"
### # shellcheck disable=SC2174
### mkdir -p -m 700 "${dump_dir}" "${errors_dir}"
###
### # Log the start of the function
### log "LOCAL_TASKS - ${FUNCNAME[0]}: start ${dump_file}"
###
### # Prepare the dump command (errors go to the error file and the data to the dump file)
### dump_cmd="my-dump-command 2> ${error_file} > ${dump_file}"
### log "LOCAL_TASKS - ${FUNCNAME[0]}: ${dump_cmd}"
###
### # Execute the dump command
### ${dump_cmd}
###
### # Check result and deal with potential errors
### local last_rc=$?
### # shellcheck disable=SC2086
### if [ ${last_rc} -ne 0 ]; then
### log_error "LOCAL_TASKS - ${FUNCNAME[0]}: my-dump-command to ${dump_file} returned an error ${last_rc}" "${error_file}"
### GLOBAL_RC=${E_DUMPFAILED}
### else
### rm -f "${error_file}"
### fi
###
### # Log the end of the function
### log "LOCAL_TASKS - ${FUNCNAME[0]}: stop ${dump_file}"
### }
########## Optional configuration #####################################
setup_custom() {
# System name ("linux" and "openbsd" currently supported)
### SYSTEM="$(uname)"
# Host name for logs and notifications
### HOSTNAME="$(hostname)"
# Email subject for notifications
### MAIL_SUBJECT="[info] EvoBackup - Client ${HOSTNAME}"
# No-op in case nothing is executed
:
}
########## Libraries ##################################################
# Change this to wherever you install the libraries
LIBDIR="/usr/local/lib/evobackup"
source "${LIBDIR}/main.sh"
########## Let's go! ##################################################
run_evobackup

View file

@ -0,0 +1,50 @@
---
- name: Dependencies are present
ansible.builtin.apt:
name:
- rsync
- mtree-netbsd
state: present
- name: "Remount /usr if needed"
include_role:
name: remount-usr
when: evobackup_client__lib_dir is search("/usr") or evobackup_client__bin_dir is search("/usr")
- name: copy evobackup libs
ansible.builtin.copy:
src: upstream/lib
dest: "{{ evobackup_client__lib_dir }}/"
force: True
mode: "0644"
owner: root
group: root
- name: copy evobackupctl script
ansible.builtin.copy:
src: upstream/bin/evobackupctl
dest: "{{ evobackup_client__bin_dir }}/evobackupctl"
force: True
mode: "0755"
owner: root
group: root
- name: LIBDIR is customized in evobackupctl
ansible.builtin.replace:
path: "{{ evobackup_client__bin_dir }}/evobackupctl"
regexp: "^LIBDIR=.+"
replace: "LIBDIR=\"{{ evobackup_client__lib_dir }}\""
- name: Evobackup canary cron is present
ansible.builtin.template:
src: update-evobackup-canary.sh.j2
dest: "{{ evobackup_client__update_canary_path }}"
mode: "0700"
when: evobackup_client__update_canary_enable | bool
- name: Evobackup canary cron is absent
ansible.builtin.file:
path: "{{ evobackup_client__update_canary_path }}"
state: absent
when: not ( evobackup_client__update_canary_enable | bool)

View file

@ -1,26 +1,31 @@
---
- ansible.builtin.include: "ssh_key.yml"
tags:
- evobackup_client
- evobackup_client_backup_ssh_key
- name: Install evobackup client components
ansible.builtin.include: "install.yml"
- ansible.builtin.include: "jail.yml"
tags:
- evobackup_client
- evobackup_client_jail
### This is commented because supposedly non-functionnal
- ansible.builtin.include: "upload_scripts.yml"
tags:
- evobackup_client
- evobackup_client_backup_scripts
# - ansible.builtin.include: "ssh_key.yml"
# tags:
# - evobackup_client
# - evobackup_client_backup_ssh_key
- ansible.builtin.include: "open_ssh_ports.yml"
tags:
- evobackup_client
- evobackup_client_backup_firewall
# - ansible.builtin.include: "jail.yml"
# tags:
# - evobackup_client
# - evobackup_client_jail
- ansible.builtin.include: "verify_ssh.yml"
tags:
- evobackup_client
- evobackup_client_backup_hosts
# - ansible.builtin.include: "upload_scripts.yml"
# tags:
# - evobackup_client
# - evobackup_client_backup_scripts
# - ansible.builtin.include: "open_ssh_ports.yml"
# tags:
# - evobackup_client
# - evobackup_client_backup_firewall
# - ansible.builtin.include: "verify_ssh.yml"
# tags:
# - evobackup_client
# - evobackup_client_backup_hosts

View file

@ -0,0 +1,3 @@
#!/bin/sh
update-evobackup-canary --who {{ evobackup_client__update_canary_who | mandatory }}

View file

@ -1,305 +0,0 @@
#!/bin/sh
# Careful, the zzz_evobackup template was last updated on 2020/06/08
#
# Script Evobackup client
# See https://gitea.evolix.org/evolix/evobackup
#
# Author: Gregory Colpart <reg@evolix.fr>
# Contributors:
# Romain Dessort <rdessort@evolix.fr>
# Benoît Série <bserie@evolix.fr>
# Tristan Pilat <tpilat@evolix.fr>
# Victor Laborie <vlaborie@evolix.fr>
# Jérémy Lecour <jlecour@evolix.fr>
#
# Licence: AGPLv3
#
# /!\ DON'T FORGET TO SET "MAIL" and "SERVERS" VARIABLES
# Fail on unassigned variables
set -u
##### Configuration ###################################################
# email adress for notifications
MAIL={{ evobackup_client__mail }}
# list of hosts (hostname or IP) and SSH port for Rsync
SERVERS="{% for host in evobackup_client__hosts %}{{ host.name }}:{{ host.port }}{% if loop.index != loop.length %} {% endif %}{% endfor %}"
# Should we fallback on servers when the first is unreachable ?
SERVERS_FALLBACK={{ evobackup_client__servers_fallback }}
# timeout (in seconds) for SSH connections
SSH_CONNECT_TIMEOUT=${SSH_CONNECT_TIMEOUT:-30}
## We use /home/backup : feel free to use your own dir
LOCAL_BACKUP_DIR="{{ evobackup_client__backup_path }}"
# You can set "linux" or "bsd" manually or let it choose automatically
SYSTEM=$(uname | tr '[:upper:]' '[:lower:]')
# Change these 2 variables if you have more than one backup cron
PIDFILE="{{ evobackup_client__pid_path }}"
LOGFILE="{{ evobackup_client__log_path }}"
## Enable/Disable tasks
LOCAL_TASKS=${LOCAL_TASKS:-1}
SYNC_TASKS=${SYNC_TASKS:-1}
##### SETUP AND FUNCTIONS #############################################
BEGINNING=$(/bin/date +"%d-%m-%Y ; %H:%M")
# shellcheck disable=SC2174
mkdir -p -m 700 ${LOCAL_BACKUP_DIR}
PATH=/sbin:/usr/sbin:/bin:/usr/bin:/usr/local/sbin:/usr/local/bin
## lang = C for english outputs
export LANGUAGE=C
export LANG=C
## Force umask
umask 077
## Initialize variable to store SSH connection errors
SERVERS_SSH_ERRORS=""
# Call test_server with "HOST:PORT" string
# It will return with 0 if the server is reachable.
# It will return with 1 and a message on stderr if not.
test_server() {
item=$1
# split HOST and PORT from the input string
host=$(echo "${item}" | cut -d':' -f1)
port=$(echo "${item}" | cut -d':' -f2)
# Test if the server is accepting connections
ssh -q -o "ConnectTimeout ${SSH_CONNECT_TIMEOUT}" -i {{ evobackup_client__root_key_path }} "${host}" -p "${port}" -t "exit"
# shellcheck disable=SC2181
if [ $? = 0 ]; then
# SSH connection is OK
return 0
else
# SSH connection failed
new_error=$(printf "Failed to connect to \`%s' within %s seconds" "${item}" "${SSH_CONNECT_TIMEOUT}")
SERVERS_SSH_ERRORS=$(printf "%s\\n%s" "${SERVERS_SSH_ERRORS}" "${new_error}" | sed -e '/^$/d')
return 1
fi
}
# Call pick_server with an optional positive integer to get the nth server in the list.
pick_server() {
increment=${1:-0}
list_length=$(echo "${SERVERS}" | wc -w)
if [ "${increment}" -ge "${list_length}" ]; then
# We've reached the end of the list
new_error="No more server available"
SERVERS_SSH_ERRORS=$(printf "%s\\n%s" "${SERVERS_SSH_ERRORS}" "${new_error}" | sed -e '/^$/d')
# Log errors to stderr
printf "%s\\n" "${SERVERS_SSH_ERRORS}" >&2
# Log errors to logfile
printf "%s\\n" "${SERVERS_SSH_ERRORS}" >> $LOGFILE
return 1
fi
# Extract the day of month, without leading 0 (which would give an octal based number)
today=$(date +%e)
# A salt is useful to randomize the starting point in the list
# but stay identical each time it's called for a server (based on hostname).
salt=$(hostname | cksum | cut -d' ' -f1)
# Pick an integer between 0 and the length of the SERVERS list
# It changes each day
item=$(( (today + salt + increment) % list_length ))
# cut starts counting fields at 1, not 0.
field=$(( item + 1 ))
echo "${SERVERS}" | cut -d' ' -f${field}
}
## Verify other evobackup process and kill if needed
if [ -e "${PIDFILE}" ]; then
pid=$(cat "${PIDFILE}")
# Does process still exist ?
if kill -0 "${pid}" 2>/dev/null; then
# Killing the childs of evobackup.
for ppid in $(pgrep -P "${pid}"); do
kill -9 "${ppid}";
done
# Then kill the main PID.
kill -9 "${pid}"
printf "%s is still running (PID %s). Process has been killed" "$0" "${pid}\\n" >&2
else
rm -f ${PIDFILE}
fi
fi
echo "$$" > ${PIDFILE}
# shellcheck disable=SC2064
trap "rm -f ${PIDFILE}" EXIT
##### LOCAL BACKUP ####################################################
if [ "${LOCAL_TASKS}" = "1" ]; then
## Dump system and kernel versions
uname -a > ${LOCAL_BACKUP_DIR}/uname
## Dump network routes with mtr and traceroute (warning: could be long with aggressive firewalls)
for addr in 8.8.8.8 www.evolix.fr travaux.evolix.net; do
mtr -r ${addr} > ${LOCAL_BACKUP_DIR}/mtr-${addr}
traceroute -n ${addr} > ${LOCAL_BACKUP_DIR}/traceroute-${addr} 2>&1
done
## Dump process with ps
ps auwwx >${LOCAL_BACKUP_DIR}/ps.out
if [ "${SYSTEM}" = "linux" ]; then
## Dump network connections with ss
ss -taupen > ${LOCAL_BACKUP_DIR}/netstat.out
## List Debian packages
dpkg -l > ${LOCAL_BACKUP_DIR}/packages
dpkg --get-selections > ${LOCAL_BACKUP_DIR}/packages.getselections
apt-cache dumpavail > ${LOCAL_BACKUP_DIR}/packages.available
## Dump MBR / table partitions
disks=$(lsblk -l | grep disk | grep -v -E '(drbd|fd[0-9]+)' | awk '{print $1}')
for disk in ${disks}; do
dd if="/dev/${disk}" of="${LOCAL_BACKUP_DIR}/MBR-${disk}" bs=512 count=1 2>&1 | grep -Ev "(records in|records out|512 bytes)"
fdisk -l "/dev/${disk}" > "${LOCAL_BACKUP_DIR}/partitions-${disk}" 2>&1
done
cat ${LOCAL_BACKUP_DIR}/partitions-* > ${LOCAL_BACKUP_DIR}/partitions
## Dump iptables
if [ -x /sbin/iptables ]; then
{ /sbin/iptables -L -n -v; /sbin/iptables -t filter -L -n -v; } > ${LOCAL_BACKUP_DIR}/iptables.txt
fi
## Dump findmnt(8) output
FINDMNT_BIN=$(command -v findmnt)
if [ -x "${FINDMNT_BIN}" ]; then
${FINDMNT_BIN} > ${LOCAL_BACKUP_DIR}/findmnt.txt
fi
else
## Dump network connections with netstat
netstat -finet -atn > ${LOCAL_BACKUP_DIR}/netstat.out
## List OpenBSD packages
pkg_info -m > ${LOCAL_BACKUP_DIR}/packages
## Dump MBR / table partitions
disklabel sd0 > ${LOCAL_BACKUP_DIR}/partitions
## Dump pf infos
pfctl -sa > ${LOCAL_BACKUP_DIR}/pfctl-sa.txt
fi
## Dump rights
#getfacl -R /var > ${LOCAL_BACKUP_DIR}/rights-var.txt
#getfacl -R /etc > ${LOCAL_BACKUP_DIR}/rights-etc.txt
#getfacl -R /usr > ${LOCAL_BACKUP_DIR}/rights-usr.txt
#getfacl -R /home > ${LOCAL_BACKUP_DIR}/rights-home.txt
fi
##### REMOTE BACKUP ###################################################
n=0
server=""
if [ "${SERVERS_FALLBACK}" = "1" ]; then
# We try to find a suitable server
while :; do
server=$(pick_server "${n}")
test $? = 0 || exit 2
if test_server "${server}"; then
break
else
server=""
n=$(( n + 1 ))
fi
done
else
# we force the server
server=$(pick_server "${n}")
fi
SSH_SERVER=$(echo "${server}" | cut -d':' -f1)
SSH_PORT=$(echo "${server}" | cut -d':' -f2)
HOSTNAME=$(hostname)
if [ "${SYSTEM}" = "linux" ]; then
rep="/bin /boot /lib /opt /sbin /usr /srv"
else
rep="/bsd /bin /sbin /usr"
fi
if [ "${SYNC_TASKS}" = "1" ]; then
# /!\ DO NOT USE COMMENTS in the rsync command /!\
# It breaks the command and destroys data, simply remove (or add) lines.
# Remote shell command
RSH_COMMAND="ssh -i {{ evobackup_client__root_key_path }} -p ${SSH_PORT} -o 'ConnectTimeout ${SSH_CONNECT_TIMEOUT}'"
# ignore check because we want it to split the different arguments to $rep
# shellcheck disable=SC2086
rsync -avzh --stats --delete --delete-excluded --force --ignore-errors --partial \
--exclude "lost+found" \
--exclude ".nfs.*" \
--exclude "/var/log" \
--exclude "/var/log/evobackup*" \
--exclude "/var/lib/mysql" \
--exclude "/var/lib/postgres" \
--exclude "/var/lib/postgresql" \
--exclude "/var/lib/sympa" \
--exclude "/var/lib/metche" \
--exclude "/var/run" \
--exclude "/var/lock" \
--exclude "/var/state" \
--exclude "/var/apt" \
--exclude "/var/cache" \
--exclude "/usr/src" \
--exclude "/usr/doc" \
--exclude "/usr/share/doc" \
--exclude "/usr/obj" \
--exclude "dev" \
--exclude "/var/spool/postfix" \
--exclude "/var/lib/amavis/amavisd.sock" \
--exclude "/var/lib/munin/*tmp*" \
--exclude "/var/lib/php5" \
--exclude "/var/spool/squid" \
--exclude "/var/lib/elasticsearch" \
--exclude "/var/lib/amavis/tmp" \
--exclude "/var/lib/clamav/*.tmp" \
--exclude "/home/mysqltmp" \
--exclude "/var/lib/php/sessions" \
${rep} \
/etc \
/root \
/var \
-e "${RSH_COMMAND}" \
"root@${SSH_SERVER}:/var/backup/" \
| tail -30 >> $LOGFILE
fi
##### REPORTING #######################################################
END=$(/bin/date +"%d-%m-%Y ; %H:%M")
printf "EvoBackup - %s - START %s ON %s (LOCAL_TASKS=%s SYNC_TASKS=%s)\\n" \
"${HOSTNAME}" "${BEGINNING}" "${SSH_SERVER}" "${LOCAL_TASKS}" "${SYNC_TASKS}" \
>> $LOGFILE
printf "EvoBackup - %s - STOP %s ON %s (LOCAL_TASKS=%s SYNC_TASKS=%s)\\n" \
"${HOSTNAME}" "${END}" "${SSH_SERVER}" "${LOCAL_TASKS}" "${SYNC_TASKS}" \
>> $LOGFILE
tail -10 $LOGFILE | \
mail -s "[info] EvoBackup - Client ${HOSTNAME}" \
${MAIL}

View file

@ -142,6 +142,9 @@ evolinux_ssh_group: "evolinux-ssh"
#
# evolinux_users_include: True
# bash
evolinux_bash_config_include: true
# root
evolinux_root_include: True

View file

@ -64,12 +64,19 @@
when: evolinux_logs_default_dateext | bool
# Logcheck
- name: Check if journald.logfiles exists
stat:
path: /etc/logcheck/logcheck.logfiles.d/journal.logfiles
register: _logcheck_journald_logfiles
- name: Disable logcheck monitoring of journald
ansible.builtin.lineinfile:
dest: /etc/logrotate.conf
dest: /etc/logcheck/logcheck.logfiles.d/journal.logfiles
line: "#journal"
regexp: "^journal"
when: evolinux_logs_disable_logcheck_journald | bool
when:
- _logcheck_journald_logfiles.stat.exists
- evolinux_logs_disable_logcheck_journald | bool
# Journald
- name: /etc/systemd/journald.conf.d/ is present

View file

@ -76,6 +76,7 @@
- name: Bash configuration
ansible.builtin.import_tasks: bash.yml
when: evolinux_bash_config_include | bool
- name: Root user configuration
ansible.builtin.import_tasks: root.yml
@ -147,9 +148,14 @@
tags:
- postfix
- name: Autosysadmin
- name: Autosysadmin (agent)
ansible.builtin.include_role:
name: 'evolix/autosysadmin'
name: 'evolix/autosysadmin-agent'
when: evolinux_autosysadmin_include | bool
- name: Autosysadmin (restart_nrpe)
ansible.builtin.include_role:
name: 'evolix/autosysadmin-restart_nrpe'
when: evolinux_autosysadmin_include | bool
- name: fail2ban

View file

@ -24,6 +24,8 @@ nagios ALL = NOPASSWD: /sbin/megacli -LdInfo -Lall -aALL -NoLog
nagios ALL = NOPASSWD: /sbin/megacli -AdpBbuCmd -GetBbuStatus -aALL -NoLog
nagios ALL = NOPASSWD: /sbin/ssacli controller all show status
nagios ALL = NOPASSWD: /sbin/ssacli controller slot=0 logicaldrive all show
nagios ALL = NOPASSWD: /usr/local/bin/mvcli info -o blk
nagios ALL = NOPASSWD: /usr/local/bin/mvcli info -o vd
nagios ALL = NOPASSWD: /usr/local/lib/nagios/plugins/check_gluster.rb
nagios ALL = (clamav) NOPASSWD: /usr/bin/clamscan /tmp/safe.txt

View file

@ -115,7 +115,7 @@
- name: Script unban_ip is installed
ansible.builtin.copy:
src: unban_ip.sh
dst: /usr/local/sbin/unban_ip
dest: /usr/local/sbin/unban_ip
mode: '0700'
tags:
- fail2ban

View file

@ -2,12 +2,15 @@
# Juin - Decembre 2022 : #64088
# Purge pour Stretch et Buster
/usr/bin/ionice -c3 /usr/bin/sqlite3 /var/lib/fail2ban/fail2ban.sqlite3 "DELETE FROM bans WHERE datetime('now', '-{{ fail2ban_dbpurgeage_default }}') > datetime(timeofban, 'unixepoch');"
/usr/bin/ionice -c3 /usr/bin/sqlite3 /var/lib/fail2ban/fail2ban.sqlite3 ".timeout 5000; DELETE FROM bans WHERE datetime('now', '-{{ fail2ban_dbpurgeage_default }}') > datetime(timeofban, 'unixepoch');"
place_dispo=$( df -h /var/lib/fail2ban/fail2ban.sqlite3 --output="avail" -h --block-size=1 |tail -n1 )
place_pris=$( echo $(("$(stat --format %s /var/lib/fail2ban/fail2ban.sqlite3 ) * 2" )) )
place_dispo="$(df /var/lib/fail2ban/fail2ban.sqlite3 --output="avail" --block-size=1 | tail -n1)"
place_pris="$(stat --format %s /var/lib/fail2ban/fail2ban.sqlite3)"
if [ $place_pris -lt $place_dispo ]
if [ "$place_pris" -lt "$place_dispo" ]
then
/usr/bin/ionice -c3 /usr/bin/sqlite3 /var/lib/fail2ban/fail2ban.sqlite3 "VACUUM;"
else
>&2 echo "Fail2ban SQLite VACUUM not done because /var lacks of space (VACUUM may use twice the database size)."
exit 1
fi

View file

@ -59,7 +59,7 @@ bantime = {{ fail2ban_wordpress_soft_bantime }}
enabled = {{ fail2ban_roundcube }}
port = http, https
filter = roundcube
logpath = /var/lib/roundcube/logs/errors
logpath = /var/log/roundcube/errors
maxretry = {{ fail2ban_roundcube_maxretry }}
findtime = {{ fail2ban_roundcube_findtime }}
bantime = {{ fail2ban_roundcube_bantime }}

View file

@ -35,5 +35,6 @@ haproxy_deny_ips: []
haproxy_backports_packages_stretch: haproxy libssl1.0.0
haproxy_backports_packages_buster: haproxy
haproxy_backports_packages_bullseye: haproxy
haproxy_backports_packages_bookworm: haproxy
haproxy_allow_ip_nonlocal_bind: Null
haproxy_allow_ip_nonlocal_bind: Null

Some files were not shown because too many files have changed in this diff Show more