Skip to content

Latest commit

 

History

History
82 lines (66 loc) · 4.05 KB

RUNBOOK.md

File metadata and controls

82 lines (66 loc) · 4.05 KB
Raw

The Runbook

The ONLY handbook a sysadmin needs to maintain the MetaWiki in a production environment.

Table of Recipes

Upgrading MediaWiki Version

Please follow below instructions in order to upgrade the mediawiki version. Refer to this Pull Request for an example.

  1. Update mediawiki/Dockerfile to download the targeted version.
  2. Update mediawiki/install_extensions.sh to download compatible mediawiki extensions.
  3. Read through the change log and see if any other updates need to be made.
  4. Try to build, and rectify errors.
  5. Follow instructions mentioned in "Deploying to prod" section.
  6. Run docker compose exec mediawiki /srv/mediawiki/maintenance/run.php update

Deploying to Production

  1. Run the deploy script
$ sudo SLACK_NOTIFICATIONS_URL=$(SLACK_NOTIFICATIONS_URL) /root/metakgp-wiki/scripts/deploy-latest.sh --go

Note: The --go flag indicates that the deployment will be completed. For a dry run, run the same command without that flag at the end.

To get the SLACK_NOTIFICATIONS_URL, go to Manage Apps -> Custom Integrations -> Incoming Webhooks -> Choose to edit "Posts to #server as Deployment Notifications" -> copy Webhook URL.

Deprecated

Note: Manual operation for regular configuration updates is now deprecated. Please use the deploy script: ./scripts/deploy-latest.sh.

  1. Take a backup! Run docker compose -f docker-compose.prod.yml exec backup ./run_backup.sh and make sure it succeeded. Also run git log and note down the current deployed sha, in case you need to roll back.
  2. git pull and docker compose build. This builds and caches the new images locally without interrupting the old server, which reduces downtime.
  3. docker compose -f docker-compose.prod.yml down shuts down the server and removes containers. Now downtime has started ticking.
  4. docker volume rm metakgp-wiki_mediawiki-volume, so that the mediawiki container can create a new volume with updates in the next step.
  5. docker compose up --build -d starts all the services using the newly built images. Server is back online, verify by going to the wiki in a browser. If there are database problems, it's often fixed by running docker compose -f docker-compose.prod.yml exec mediawiki /srv/mediawiki/maintenance/run.php update.

Retrieving the Latest backup

# ensure that the backup container is running and drop into the container's bash
$ docker compose -f docker-compose.prod.yml exec backup /bin/bash
# below command will create a new backup of the server
container $ ./run_backup.sh
# to ensure that the backup tar was created
container $ ls /root/backups
container $ exit
# now copy the created tar file into the host filesystem
# docker cp doesn't take wildcard paths (??)
$ docker cp metakgp-wiki_backup_1:/root/backups/metakgpwiki_2017_10_23_10_11_44.tar.gz .
$ pwd

Now, rsync takes over. You need to rsync or scp this file back to your computer. This step is out of the scope of this tutorial. Check out rsync's man page for instructions.

Restoring From a Backup

  1. Ensure that the wiki container is running.
  2. Run the following command.
$ ./scripts/restore-from-backup.sh <path-to-sql-file>

You have to rsync the tar.gz archive to the host filesystem and extract it. Inside the archive, you will find the sql file whose path you must supply to the restore-from-backup.sh script.

Purging all Docker Containers, Volumes, and Networks

NOTE This is a destructive operation. DO NOT do this in production unless necessary.

# shut down all the containers before doing this
$ docker compose down
$ docker system prune -a
$ docker volume prune