nextcloud shows update available when is already up to date

[sbe-arg]

No clue where to log the issue let me know if needs to go into another github space

[stavros-k]

Hello, can you post the output of the bellow command? Run it on the TrueNAS Shell.

midclt call container.image.query

Thanks

[sbe-arg]

I assume the important bit is "update_available": true," ?

{
    "id": "sha256:ea0d219f8302b3fafca6bf69dacdf294d4e5fb9ef0a7ce517d8c271393da2cca",
    "repo_tags": [
        "docker.io/library/nextcloud:28.0.3"
    ],
    "repo_digests": [
        "docker.io/library/nextcloud@sha256:4aa8fdda6f280d3a3bc7786de33af3f27fdf00f158427672a962d010d4789064"
    ],
    "size": 422669625,
    "dangling": false,
    "update_available": true,
    "system_image": false
},

[sbe-arg]

I ran the "update" and the available notification went away... waited a few hours and it was there again.

[stavros-k]

Okay, I see whats happening.
Nextcloud overwrite's the 28.0.3 tag and the digest is changing.
So that's why it shows you that there is an update.

Both container version and chart version is the same, but the digest of the upstream container is newer.

[sbe-arg]

Not sure why was marked as completed XD
Not sure why the digest got updated then 🤔 a re-tag?

[sbe-arg]

@stavros-k how can I force a pull of the image from the cli so I get the correct digest?

[stavros-k]

Not sure why was marked as completed XD Not sure why the digest got updated then 🤔 a re-tag?

Yes nextcloud re-used the same tag with different underling image, which caused the digest to change.
I can' know the exact reason.

Updating from the Apps UI should re-pull the latest digest!

[sbe-arg]

That worked

[sbe-arg]

hey @stavros-k this keeps happening every 1-2 weeks after I update the images says there is an update and a sh mismatch.

there is something odd. can we reopen?

[sbe-arg]

what the?

seems like nextcloud is updating the existing tags?

[sbe-arg]

@stavros-k if you see here nextcloud/docker#2197 maybe an option is to change the pull strategy to get some alignment into how to avoid the mismatch

[stavros-k]

I don't think changing the strategy would help.
If you set it to Never then you would have to manually pull the image before you can even install the App.
If you set it to Always you will be downloading 400mb + every time you edit or stop and start the App.

[sbe-arg]

But it is a problem as it is now.

[sync-by-unito]

➤ Stavros Kois commented:

Re-opening this:

We need a way to either pin to specific digests for the containers on our catalog.
Or handle better the “update available” on the UI/middleware side.
(or both)

[stavros-k]

It shouldn't be an issue anymore on TrueNAS 24.10 and later.
If you encounter such issue again, feel free to reopen at truenas/apps repo.

Thanks