How to automatic start scanning after nc restart?

[DenuxPlays]

Hey,
my nextcloud restarts every day (for backup purposes) and recognize does not restart scanning all photo.

How can I change this?
Because its very annoying to go into nextcloud and click on rescan every file ever day.

[DenuxPlays]

Still not working:

Recognize is not working after nextcloud restart...

[marcelklehr]

my nextcloud restarts every day (for backup purposes)

When a background job is killed while running it will not be executed for a minimum of 12h, afaicr. This is not up to recognize.

[ShapeShifter499]

I'm struggling to find the bit of code. Where does it dictate a minimum of 12h?

[marcelklehr]

https://github.com/nextcloud/server/blob/master/lib/private/BackgroundJob/JobList.php#L220

[marcelklehr]

click on rescan every file ever day.

This is not a good idea, as it will add all files to the queue again, regardless of whether they have been scanned before.

[marcelklehr]

In my opinion, when a) a container is killed ad hoc and thus b) the processes in it get killed, causing c) a background job to throw an error, in turn causing d) that background job to not be called for a while; if you don't want d) to happen you need to avoid a).

[marcelklehr]

Maybe you can enable maintenance-mode and wait for all cron.php processes to stop before killing the container. Maintenance mode will halt execution of new background jobs.

[szaimen]

Maybe you can enable maintenance-mode and wait for all cron.php processes to stop before killing the container. Maintenance mode will halt execution of new background jobs.

I tried to listen for the container shutdown process and wait for cron to finish but it didnt work. See nextcloud/all-in-one#3214. Maybe you have a good idea how to do that?

[szaimen]

Alao I guess we cannot really wait longer than 600s so it will not fix all issues as otherwise you might need to wait just too long until you can start a backup or start the containers again... (The instance will be down for that time)

[DenuxPlays]

@DenuxPlays does this still occur for you? I noticed in the last few days the issue went away despite having automatic updates enabled in AIO

Yes still the same...

[szaimen]

The fix nextcloud/all-in-one#3214 is now released with v7.8.0 Beta. Testing and feedback is welcome! See nextcloud/all-in-one#how-to-switch-the-channel

[DenuxPlays]

The fix nextcloud/all-in-one#3214 is now released with v7.8.0 Beta. Testing and feedback is welcome! See nextcloud/all-in-one#how-to-switch-the-channel

Nice
I am going to test this when I am home (in about 2,5h)

[DenuxPlays]

The fix nextcloud/all-in-one#3214 is now released with v7.8.0 Beta. Testing and feedback is welcome! See nextcloud/all-in-one#how-to-switch-the-channel

Update:
atm nothing seems to work:

Tried rescanning all files, recrawl everything via the terminal...
Nothing seems to work.

[DenuxPlays]

Info:
I found something interesting.
If you stop the nextcloud container and then start it again (manually) it invokes a new recognize scan:

If nextcloud restarts automatically (because it's doing an backup) nothing happens.

[szaimen]

You might run into this timeout then: https://github.com/nextcloud/all-in-one/blob/ad4b48c82e56b2d0769fa78912da9839bf290b99/php/containers.json#L218
But I fear we cannot increase this because it would increase the downtime. Also having to wait up to 10min for your containers to stop already feels too long imho...

[DenuxPlays]

Yeah I totally agree...
5 minutes should be enough.
But can we make it configurable via env-vars?
Then I could at least test it or for me personal it is okay if my nextcloud isn't available for a few more minutes as I usally sleep during the downtime, but this will not apply to everyone especially if more than a few users use it.

[szaimen]

Instead of having yet another config option would the more elegant way be this:

Shouldnt there be a way to differentiate between broken vs not ended background jobs (jobs that took less than 12h but were ended from the outside)? If there was one, it should be possible to add a logic to restart not ended background jobs earlier...

[DenuxPlays]

Instead of having yet another config option would the more elegant way be this:

Shouldnt there be a way to differentiate between broken vs not ended background jobs (jobs that took less than 12h but were ended from the outside)? If there was one, it should be possible to add a logic to restart not ended background jobs earlier...

Yes that sounds like a clean solution but it is realistic that this will be implemented any time soon?
I imagine this has to be implemented in nextcloud directly?