Can't restart a not responding app while backups are running

[marcusquinn]

I can't restart a not responding app while backups are running, even though the backups stage is currently uploading another app.

Since our backups are likely to take hours, the choice is to stop backups to allow for the restart and repair attempt, or wait a long time. Not ideal for production.

Thinking this might need to be a Feature Request?

[humptydumpty]

You got my upvote for this. I ran into this exact thing yesterday. I just set up a B2 bucket w/ rsync and it took over 16 hours before I had to kill it. Tgz took less than 10 minutes today. My entire server is around 28GB.

Edit: Just found out how to upvote lol.

[girish]

@marcusquinn Good idea. I will move this to feature requests.

In case anyone is wondering why this is, we have a "global lock" for app task & backups. app tasks can run with a concurrency of 3 but will wait for backups to finish. I think restart can definitely be somehow "whitelisted" to be allowed without a lock.

[jimcavoli]

@girish This has just started to become a noticeable/frequent problem for me as well; a smarter locking scheme with respect to backups would be much appreciated!

[girish]

@jimcavoli A hack for the moment of course is to simply docker restart <container>. Not ideal, agreed.

[marcusquinn]

Is this still not solved?

Pretty fundamental flaw, given the choice between; losing a backup (because they take half a day to run now), or not being able to recover a stalled app until forever.

[girish]

@marcusquinn not yet, but it is scheduled for next release - https://forum.cloudron.io/topic/9975/what-s-coming-in-9-0-was-8-0 (the "remove global lock for backups" bullet point)