You are browsing a read-only backup copy of Wikitech. The primary site can be found at wikitech.wikimedia.org

GitLab: Difference between revisions

From Wikitech-static
Jump to navigation Jump to search
imported>Jelto
No edit summary
imported>Jelto
No edit summary
Line 1: Line 1:
{{Sidebar
| style = background: white; padding:10px; padding-{{dir|{{pagelang}}|left|right}}:13px; margin:{{dir|{{pagelang}}|5px 12px 5px 0|5px 0 5px 12px}}; width: 350px;
| name = GitLab
| title = GitLab
| image = [[File:Gitlab-logo.svg.svg|center|250px]]
| headingstyle = font-size: 130%; padding: .5em;
| contentstyle = text-align: {{dir|{{pagelang}}|right|left}}; font-size: 14px; padding: .5em; line-height: 1.5;
| abovestyle = text-align: {{dir|{{pagelang}}|right|left}};
| content1 =
{{Special:PrefixIndex/{{FULLPAGENAME}}/ |hideredirects=1 |stripprefix=1}}
* External resources:
** [https://gitlab.wikimedia.org/ Production GitLab]
** [https://docs.gitlab.com GitLab docs]
** [https://www.mediawiki.org/wiki/GitLab GitLab in MediaWiki]
** [https://phabricator.wikimedia.org/project/view/5057/ GitLab in Phabricator]
}}
This page contains SRE related topics for GitLab. For GitLab application specific information, please see https://www.mediawiki.org/wiki/GitLab (under Implementation).
This page contains SRE related topics for GitLab. For GitLab application specific information, please see https://www.mediawiki.org/wiki/GitLab (under Implementation).


GitLab is reachable at https://gitlab.wikimedia.org/. We run multiple instances of GitLab:
GitLab is reachable at https://gitlab.wikimedia.org/. We run multiple instances of GitLab:
Line 9: Line 26:
* gitlab in WMCS gitlab-test project
* gitlab in WMCS gitlab-test project


gitlab1001 and gitlab2001 are setup using puppet and Ansible. The base configuration currently lives in [[gerrit:plugins/gitiles/operations/puppet/+/refs/heads/production/modules/profile/manifests/gitlab.pp|profile::gitlab]]. Additional GitLab specific configuration comes from [[gerrit:plugins/gitiles/operations/gitlab-ansible|gitlab-ansible]]. Its planned to migrate all of the logic to puppet and drop Ansible (see [[phab:T283076|T283076]]).
gitlab1001 and gitlab2001 are setup using puppet and Ansible. The base configuration currently lives in [[gerrit:plugins/gitiles/operations/puppet/+/refs/heads/production/modules/profile/manifests/gitlab.pp|profile::gitlab]]. Additional GitLab specific configuration comes from [[gerrit:plugins/gitiles/operations/gitlab-ansible|gitlab-ansible]]. Its planned to migrate all of the logic to puppet and drop Ansible (see [[phab:T283076|T283076]]). GitLab is installed as a [https://docs.gitlab.com/omnibus/ Omnibus installation] on all instances. So all GitLab components are installed using the official packages and are executed on a single host. The reasons for this setup can be found in the [[mw:GitLab/Initialization|Initialization docs in Mediawiki]].
 
==Backup and restore==
This section describes backup configuration and restore procedure for GitLab instance.
 
===Backups===
To backup application data GitLabs build in [https://docs.gitlab.com/ee/raketasks/backup_restore.html#back-up-gitlab backup functionality] is used. Application data backups are created by calling the <code>/usr/bin/gitlab-backup create</code> command. Configuration backups are created by calling <code>/usr/bin/gitlab-ctl backup-etc</code>. The commands are executed once a day in cronjobs [[gerrit:plugins/gitiles/operations/gitlab-ansible/+/refs/heads/master/roles/gitlab_server/defaults/main.yml#14|created with Ansible]] and will create full backups. To configure the backups please refer to all [[gerrit:plugins/gitiles/operations/gitlab-ansible/+/refs/heads/master/roles/gitlab_server/templates/gitlab-crontab.j2|backup related variables]] in Ansible.
 
So GitLab will create two new .tar archives every day:
 
*full data backup in <code><nowiki>{{gitlab_backup_path}}</nowiki></code>
*full config backup in <code>/etc/gitlab/config_backup</code>
 
Partial backups are [[gerrit:plugins/gitiles/operations/gitlab-ansible/+/refs/heads/master/roles/gitlab_server/defaults/main.yml#22|disabled currently]]. For the initialization phase daily full backups are used. In the future we may start implementing partial and incremental backups.
 
====Backup retention====
Data backups and config backups will be deleted after three days on the production instance (see [[phab:T274463#7147179|T274463#7147179]]). Release Engineering wanted to have three days of local retention for fast troubleshooting and restores. Deletion of the data backups is handled by GitLab (using the <code>gitlab_backup_keep_time</code> variable). Deletion of the config backup is implemented in the [[gerrit:plugins/gitiles/operations/gitlab-ansible/+/refs/heads/master/roles/gitlab_server/defaults/main.yml#14|backup cronjob]] (using the  <code>gitlab_backup_config_keep_num</code> variable).
 
===Storing backups in bacula===
For enhanced reliability backups are also stored in [[Bacula]]. Bacula is the standard for secure, encrypted backup storage in the WMF.
 
For the initialization phase we decided to only backup the most recent .tar file with the data backup and the most recent .tar file with the configuration backup. Furthermore these .tar files will be shipped to Bacula once a day as a full backup (see [[Bacula#Backup Strategy|backup strategy daily]]). This backup strategy is '''not the default''' used by most services. The following concerns and advantages were discovered during our discussion when comparing daily full backups instead of weekly full backups and daily incremental backups (see [[phab:T274463|T274463]] and comments in [[gerrit:c/operations/puppet/+/697850|/puppet/+/697850]]):
 
* Incremental backups of GitLab's self-contained full backups would introduce an artificial technical dependency between revisions without having an actual dependency. To restore a backup Bacula would have to merge and diff all recent incremental backups and combine them with the last full backup. However, the latest backup should be enough to restore GitLab to the previous state.
* The default backup policy would conflict with the requirement of Release Engineering to have three days of local backup retention on the GitLab host. This conflict would cause up to three times of additional disk usage in Bacula in comparison to a non-default backup policy.
* Incremental-only backups would solve the problem of additional disk usage but can't be used long term due to technical limitations of Bacula according to Data Persistence. The restore process with a lot of incremental revisions would need a long time and computing resources. Furthermore we would introduce a dependency between revisions (see above).
Because of the reasons above we decided against the default strategy and instead use Daily Full Backups. For this decisions it was necessary to implement two changes:
 
* Add a new Daily Full policy to Bacula (see [[gerrit:c/operations/puppet/+/700183|/puppet/+/700183]])
* Create dedicated folder structure for GitLab latest backup (see [[gerrit:c/operations/gitlab-ansible/+/700084|/gitlab-ansible/+/700084]] and below)
 
==== "Latest" backup ====
To implement the strategy of daily full backups, a dedicated folder structure is needed for Bacula. We have to make sure that Bacula will not save the last three backups available on the GitLab host. Bacula must only backup the directory with the most recent files. For this purpose we created a additional <code>./latest</code> directory inside each of the backup directories ([[gerrit:c/operations/gitlab-ansible/+/700084|using Ansible]]). Since our goal is to replace the Ansible code with puppet eventually, we also ensured the "latest" backup dirs exist using Puppet. We did this in 2 places, the profile class currently used in production ([[gerrit:700622]]) and the backup class from the gitlab module currently used only in cloud ([[gerrit:700595]]). Ideally we want to get to a situation where both production and cloud machines are setup automatically by the same puppet role, both using the module. The backup scripts on the GitLab machine will update the <code>latest.tar</code> file. <syntaxhighlight lang="bash">
/srv/gitlab-backup/
├── 1624752267_2021_06_27_13.11.5_gitlab_backup.tar
├── 1624838667_2021_06_28_13.11.5_gitlab_backup.tar
├── 1624925067_2021_06_29_13.11.5_gitlab_backup.tar
└── latest
    └── latest.tar
</syntaxhighlight>
Bacula is then configured to just use the <code>/latest</code> folder and save the most recent backup. Here is the fileset used in bacula:<syntaxhighlight lang="bash">
    bacula::director::fileset { 'gitlab':
        includes => [ '/srv/gitlab-backup/latest', '/etc/gitlab/config_backup/latest' ]
    }
</syntaxhighlight>
 
===Restore===
WIP
 
=== Failover ===
WIP


== Monitoring ==
GitLab login is implemented with SSO using the [[CAS-SSO|CAS/SSO]]. So users will be redirected to idp.wikimedia.org to login to the SSO portal. Groups <code>ops</code>, <code>wmf</code> or <code>nda</code> are needed to access GitLab currently.<ref>https://gerrit.wikimedia.org/r/c/operations/puppet/+/709383/2/hieradata/role/common/idp.yaml#175</ref>
WIP

Revision as of 13:13, 3 August 2021

GitLab
File:Gitlab-logo.svg.svg


This page contains SRE related topics for GitLab. For GitLab application specific information, please see https://www.mediawiki.org/wiki/GitLab (under Implementation).

GitLab is reachable at https://gitlab.wikimedia.org/. We run multiple instances of GitLab:

gitlab1001 and gitlab2001 are setup using puppet and Ansible. The base configuration currently lives in profile::gitlab. Additional GitLab specific configuration comes from gitlab-ansible. Its planned to migrate all of the logic to puppet and drop Ansible (see T283076). GitLab is installed as a Omnibus installation on all instances. So all GitLab components are installed using the official packages and are executed on a single host. The reasons for this setup can be found in the Initialization docs in Mediawiki.

GitLab login is implemented with SSO using the CAS/SSO. So users will be redirected to idp.wikimedia.org to login to the SSO portal. Groups ops, wmf or nda are needed to access GitLab currently.[1]

  1. ↑ https://gerrit.wikimedia.org/r/c/operations/puppet/+/709383/2/hieradata/role/common/idp.yaml#175
Retrieved from "https://wikitech-static.wikimedia.org/w/index.php?title=GitLab&oldid=582415"