[6.12.3] Docker container update loop

the same happens on the older versions, 6.11.5 for example. bit odd but you can just leave/close the page after you know its done the job. 

Ah ok. I was on 6.9.3 for many years and finally recently upgraded and started running into this. So something introduced in 6.10 or later.

yeah,might be. Ive strong hunch, at least in my case, thats its cause by containers that use other containers as network. But i have no clue because theres absolutely nothing to find in the logs for any of this.

I think this might be related to another issue that I only occasionally had but in recent days has gotten significantly worse. Docker images that need to pull/extract larger components (Nextcloud/Immich/Calibre all have components >200MB, and seem to be the worst offenders for me). The extraction process times out and the updater seems to kill the update mid-extraction, then just reverts to whatever the previous installed version was. I did some reading about how updating the repository to "lscr.io" or "ghcr.io" would fix this, but didn't make a change for me, which makes sense because the stall is happening during the extraction process, not the download. 

 

I'm attaching a screenshot of the output after trying to update Nextcloud.

 

@aspiziri have you tried a manual update of one of the single containers that doesn't get updated from the loop? If so, are you experiencing something similar to my issue above?

Edited October 29, 2023 by ndauphin583

Its intermittent for me so will see if I can spot a trend next time I see it. The last time it happened there were 3 containers in the update list: InfluxDB, Redis and a random other one. The InfluxDB and random other one are both decent in size (150MB+) so its possible it's related?

 

This is what I see after the loop starts:

On 10/29/2023 at 3:49 AM, Mainfrezzer said:

the same happens on the older versions, 6.11.5 for example. bit odd but you can just leave/close the page after you know its done the job. 

 

I also have this problem and leaving the page does not help, the loop continues to run in the background (6.12.4 here). I had to stop the array to break the loop.

 

to be very clear: this is a very serious problem. If you have containers that run transactions on databses in other containers and the update restarts wither of them, the database most oftenly will not come back up clean, you will need to take a manual intervention to roll back the broken transactions.

 

Due to this this bug is actually urgent, not minor.

Edited November 24, 2023 by murkus

On 11/24/2023 at 11:51 AM, murkus said:

Due to this this bug is actually urgent, not minor.

 

1 - We always need diagnostics (preferably grabbed while this is happening) 

2 - If anyone in a position to fix the problem could ever replicate it, then it would be fixed.

Haven't seen it happen the past few days but will circle back with diagnostics if I run into it again