I have been dealing with this for a while now.

Whenever I update a docker app (primarily those with larger images), it fails, repeatedly.

Will often fail to remove old image or will have orphaned image

 

Because it was affecting larger images thought maybe it was a read/write problem.  So changed out SATA cables and even tried a different USB boot drive. (Grasping at straws)

 

Reviewed the forums and tried what others had suggested:

Have tired changing from single img file to directory (under settings/docker/ )

I have tired removing container and reinstalling.  Sometimes it works, most times, even after a re-install, it will still say update available.

Tried updating with Portainer.

 

I know it's not a internet issue because I have another unraid server right next to it that updates just fine.

 

The included diagnostic file was created just after I tried to update paperless-ngx

 

TIA

unRaid-diagnostics-20240729-1246.zip

See if this helps and please report back:

https://forums.unraid.net/topic/146853-docker-containers-not-updating/?do=findComment&comment=1443361

 

I increased timeout from 10 to 120.

Seemed to be stuck at removing old image

Competed with update failed and still shows "update ready"

Page refresh now show paperless-ngx gone and there is now orphaned image

Edited July 30, 20241 yr by uglytuna
added more information

Reread the linked post and tried updating with timeout set to 60.

Updating a different container yielded:

Removing container:

Error:

 

The command finished successfully!

 

But container still shows as needing update

And does this help:

https://forums.unraid.net/topic/146853-docker-containers-not-updating/?do=findComment&comment=1444073

 

Tried the --init parameter and it did update successfully.  Will try with other containers as they become available and report back.

 

Thanks for all your help so far

1 hour ago, uglytuna said:

and report back

Please do, I've tracking this issue for some time, but I cannot reproduce it myself, so trying to find out if any of these options really helps.

Hello,

 

Got the same exact issue since at least a couple weeks on my side.

 

I thought it may have been linked to a faulty / slow array, but even since migrating to Docker Directory, as well as a full speed SSD ZFS mirror cache, it still exhibits the same symptoms than uglytuna above.

 

Setting Docker Stop Timeout: to 60s instead of 10s didn't change anything.

 

Would it be more efficient to create my own thread, or pile up on this one ?

 

Thanks,

unraid-804-diagnostics-20240731-1322.zip

2 hours ago, triskell34 said:

Setting Docker Stop Timeout: to 60s instead of 10s didn't change anything.

And does this help?

https://forums.unraid.net/topic/146853-docker-containers-not-updating/?do=findComment&comment=1444073

 

I tried adding --init to the last container that has the problem at the moment, but it actually doesn't start at all with it : ( lscr.io/linuxserver/homeassistant )

s6-overlay-suexec: fatal: can only run as pid 1

** Press ANY KEY to close this window ** 

 

Anyway, the docker is stopped right now, with a pending known update :

 

 

That hangs at this stage... :

 

And woah, it actually worked this time. Without the --init line just for confirmation.

 

Is there any way to help any further ? I tried updating this specific container about a dozen times, I have no more pending update at the moment.

49 minutes ago, triskell34 said:

it actually worked this time.

Yeah, that's one of the difficulties with this issue, even for the affected users, it can be intermittent.

Adding the --init to Netdata container seemed to work.

Adding to Immich did not.  Althought the Immich container had a lot of other things listed in "additional parameters"

--init <br>PostgreSQL 14/15/16 with (https://github.com/tensorchord/pgvecto.rs), Redis<br>This container requires an external application to be run separately.<br>You can also use a docker mods to start redis within the container,<br>Set `DOCKER_MODS=imagegenius/mods:universal-redis`, and `REDIS_HOSTNAME` to `localhost`.<br>&nbsp;&nbsp;&nbsp;&nbsp;

Ran update twice and still shows as update available

When I tried starting it, it said execution error, Immich container now gone and there is an orphaned container now listed.

Try deleting that orphan container and then restoring from Apps.

yep, did that and restored just fine (have had to do numerous times in the past), but still showing as update available

 

I did do a diagnostic download

unraid-diagnostics-20240731-1609.zip

OK, I guess it's not possible to try just the --init as an extra option, just to see if it updates?

did you mean remove everything from extra parameters except --init?

Tried again with everything as before with same result - still saying needs updating

19 minutes ago, uglytuna said:

did you mean remove everything from extra parameters except --init?

Yep, not sure that would work, but just as a test

Same result, "finished successfully", but still saying needs updating.

OK, thanks for testing.

19 hours ago, uglytuna said:

Same result, "finished successfully", but still saying needs updating.

Please try updating this container with the Docker Stop Timeout set to 300 secs, according to another report lower values didn't work, but 300 did.

Will do

Hello all,

 

Got the same issue with immich as well since a couple days and the latest immich update,

 

I did :

 - put docker timeout at 300s

 - add --init as Extra parameter on immich docker, despite said parameter not enabling docker to start successfully

 

So, with these settings and immich docker stopped & waiting for upgrade, i did push the upgrade, which stayed stuck as following for a couple minuts :

After which it tried to start the docker, ending in the following situation :

 

My docker setup has no Extra parameter or post arguments, only --init for this test.

Edited August 4, 20241 yr by triskell34

2 hours ago, triskell34 said:

put docker timeout at 300s

So, just to confirm, this didn't help right?

Correct indeed, the immich docker was still in the old version, the new update didn't correctly apply.