Backup is failing for UpTime Kuma on 9.0.15

[potemkin_ai]

@james said in Backup is failing for UpTime Kuma on 9.0.15:

Or do you mean this setting is getting overwritten again and again?

yeah - I always have my settings setup in a way, that I would be getting an e-mail for the threads I'm participating and from time to time I'm finding myself not getting those updates - going to check the settings and they are again in 'Notification only'.

It's sporadic, but I'm quite sure I didn't change them manually (too lazy / busy for that - I don't even remember where they are in the profile).

[potemkin_ai]

@james said in Backup is failing for UpTime Kuma on 9.0.15:

How many hosts do you monitor with Uptime Kuma?

not much - like 20 max; UpTime Kuma on itself works just fine.

[james]

Hello @potemkin_ai
Is this issue still persistent? If so, can you please try to stop the Uptime Kuma app once and start it again and attempt a backup?
Want to make sure this is not a file handling issue with the SQLITE.
In the sense of, while trying to back the SQLITE file, it gets changed and thus messes up the process in some way.

[potemkin_ai]

Yep - it does.
Stop & start didn't help, unfortunately.

[nebulon]

If you still get the same docker error about an already existing container with the name sqlite-<appid>, can you check with the docker cli if some dead container exists and if so, if you just remove that, it starts working?

[potemkin_ai]

@nebulon there seems to be none of the dead ones:

ubuntu@cloudron:~$ sudo docker ps -a | grep sqlite | wc -l
1
ubuntu@cloudron:~$ sudo docker ps -a | grep sqlite
bbb52ef9e686   cloudron/louislam.uptimekuma.app:202511081423470000       "sqlite3 /app/data/d…"   2 weeks ago   Created                                                                                                                                                                sqlite-be0be218-57bf-427b-abb6-b7660943eaf6

There are not that much non-running containers - only 2 nextcloud workers with exit(0) and another service, that I have stopped explicitely.

[nebulon]

So yeah looks like there is a dangling container with the conflicting name sqlite-be0be218-57bf-427b-abb6-b7660943eaf6, not sure why that wasn't cleanly exiting, but it exists but stopped, hence no new one for another backup can be created.

The fix here would probably be to purge that with docker rm bbb52ef9e686 and then the backup should work again. Before you do this, maybe check with docker logs bbb52ef9e686 if there is anything useful, why it possibly failed.

[potemkin_ai]

Thank you, that fixed the issue.
There been no logs in the container.

Any ideas what is the nature of the sqlite container for uptime kuma? I've been under impression that sqlite requires no server side...

[nebulon]

That is true, no server is required for sqlite, however to make a correct and consistent backup, one cannot just copy the sqlite file on disk, as data might not have been flushed fully to disk. This is why the localstorage addon in Cloudron has a way to signal the system, which file is an sqlite database and the backup will spin up a container to make a proper database dump. See https://docs.cloudron.io/packaging/addons#localstorage

[potemkin_ai]

@nebulon thank you. And yes - I understand why the work-around required. If you don't mind sharing the source code to read the logic exactly?

I've been wondering on the best way to achieve that, but never seen any good practical approach - would love to see how you are approaching that, without a container shutdown, if you don't mind sharing of course.

If that's too complicated - never mind and please, feel free to close the issue - the question now is purely to satisfy my curiosity.