Apps Stuck Updating - Cleaning up old install - even after stopping/restarting task

[girish]

@jdaviescoates What's happenning is that some container is getting stuck and locking the image from getting removed. Actually, I want to investigate what that app might be. When this happens, can you please drop me a note? Maybe it's always the same app which is causing the problem (some cron job related thing maybe).

[jdaviescoates]

@girish will do: best place to drop you that note? in reply to the support email thread? on this thread? in chat on here? on matrix?

(my guess: WordPress Developer app)

[girish]

@jdaviescoates support@ is best

[jdaviescoates]

@girish fyi (added this bit with an edit so not sure you saw it?)

@jdaviescoates said in Apps Stuck Updating - Cleaning up old install - even after stopping/restarting task:

(my guess: WordPress Developer app)

[humptydumpty]

@jdaviescoates I'm running WP dev too. Here's the full list of apps I'm using:

umami
ghost
wordpress dev
surfer
chatwoot
noco
element
vaultwarden
wekan
invoiceninja
etherpad
roundcube
snappymail
matrix
listmonk
paperless
freescout
alltube
kutt
snappymail

[avatar1024]

I also have a WordPress dev app on the cloudron that gets stuck and don't have it on the other instances. The only other app I have on it and not on the other ones is mattermost

PS: the instance I restarted earlier today and that was working now hangs again on updates. Restarted again and now working fine...for some time...

[subven]

@jdaviescoates said in Apps Stuck Updating - Cleaning up old install - even after stopping/restarting task:

Yeah mine is VIRTIO too

I use SCSI on all my Netcup hosts. Not hitting the problem like I said before.

VPS 200 G8 v7.3.2 (clean) Ubuntu 20.04.2 LTS
RS 4000 G9 SE v7.3.2 (clean) Ubuntu 20.04.2 LTS

Stuff I have installed: Wordpress (managed), Matrix, Element, Bookstack, Freescout, LAMP (new and old ones), Gitea, Lychee, Roundcoube, Nextcloud, Statping, Collabora, Paperless, Surfer, Kimai, Jirafeau, Vaultwarden

[avatar1024]

So I guess it could well have to do with the WP unmanaged package as it seems like a common factor so far across all instances (of everyone in the thread) experiencing the issue and the singularity on my buggy instance instance compared to the others

[humptydumpty]

@girish the server is running fine after upgrading to 22.04. I manually updated one app yesterday and it updated just fine. A bunch of apps auto updated last night without any issues and I manually updated Wekan minutes ago. I hope I’m not jumping the gun here fingers crossed!

[p44]

@girish hello Girish, I'm trying to update a LAMP instance and problem still there, any ideas on how to proceed? Thank's

[girish]

@p44 can you write to support@ . I am afraid to give a suggestion without knowing the current state of the server.

[p44]

@girish Ok I'll do ASAP. thank's a lot

[avatar1024]

@p44 Have you tried to reboot your server and then do the update?

@p44 out of interest, do you have a WP unmanaged app installed on your instance?

[girish]

@avatar1024 this is not unmanaged WP related. We have been investigating this a lot last few days but it appears to happen very often in netcup servers at least.

The main issue is that runc is getting stuck (you can also do ps auxf on the server):

root       84188  0.0  0.0 712448  9880 ?        Sl   11:09   0:01 /usr/bin/containerd-shim-runc-v2 -namespace moby -id afb25df337753f9944749bdba5bcb577ae58630d4529da3293ea3730c3164be4 -address /run/containerd/c
root       84200  0.0  0.0 1230148 10704 ?       Sl   11:09   0:00  \_ runc --root /var/run/docker/runtime-runc/moby --log /run/containerd/io.containerd.runtime.v2.task/moby/afb25df337753f9944749bdba5bcb577ae586
root       84210  0.0  0.0 1083864 10968 ?       Ssl  11:09   0:00      \_ runc init

Those runc init above is just hanging there waiting for something (you can strace and see it's waiting on some futex). It's unclear what the issue is so far.

[robi]

@girish I wonder if sysbox would have the same issue.

[p44]

@avatar1024 Yes I did reboot but nothing. No have any WP unmanaged on instance. Only two LAMP applications.

[subven]

@girish said in Apps Stuck Updating - Cleaning up old install - even after stopping/restarting task:

happen very often in netcup servers

I know you (or @nebulon) have an account at Netcups Forum. Maybe ask them for help?

[p44]

@subven I think is not Netcup related issue.

[nebulon]

So we had an instance which seems to not hit this issue anymore by making sure the kvm version is updated. On netcup this apparently means to shutdown and boot the instance through their control panel. A simple reboot/restart will not solve it. Maybe anyone here facing this and being on netcup could attempt this so we get more info.

[jdaviescoates]

@nebulon said in Apps Stuck Updating - Cleaning up old install - even after stopping/restarting task:

Maybe anyone here facing this and being on netcup could attempt this so we get more info.

I just did that (a power cycle via their scp server control panel), let's see if it resolves the issue for me...