How to run a disaster recovery in a kubernetes environment?

[FSeidinger-XI]

Hi community,

what is the preferred way of running a disaster recovery on a self hosted Weblate instance running in kubernetes and deployed via helm?

In current 5.x versions of Weblate there is no possibility to set Weblate to a mainteance mode. This means if running a restore procedure in an active instance, it might corrupt the database and the /data partition.

So what is the proper procedure to:

1. Set Weblate to maintenance. That means user cannot login or work with Weblate during restore.
2. Disable problems with the database during restore.
3. Disable problems with the /app/data during restore.
4. Reenable Weblate.
5. Sync database and /app/data partitions with the restored database and files.

Kind regards
Frank

[nijel]

Stopping the Weblate container is the best approach. You can then perform any maintenance on the file system and database.