Immich fails to migrate from Dragonfish to Electric Eel

I am running Scale and just upgraded from Dragonfish (24.04.2.5) to Electric Eel (24.10.0.2).

Three of my apps failed to migrate:

  1. Gitea
  2. Immich
  3. Jellyfin

I ran the following command, which fixed Gitea and Jellyfin, but not Immich:
midclt call -job k8s_to_docker.migrate POOL

For Immich, I’m getting the following output:

[{“name”: “immich”, “error”: “Failed to create app: [ENOENT] values.storage.library.host_path_config.path: This path does not exist.\n[ENOENT] values.storage.uploads.host_path_config.path: This path does not exist.\n[ENOENT] values.storage.additional_storage.0.storageEntry.host_path_config.path: This path does not exist.\n”, “successfully_migrated”: false}]

I don’t understand why it believes the paths don’t exist:

  1. They do exist
  2. They are in an unlocked dataset (SMB attached)
  3. They are the same as they were before upgrading to Electric Eel and Immich had no issue with these paths on Dragonfish*
  4. The 3 paths it’s complaining about are in the same dataset that Jellyfin is using. How could the paths exist for Jellyfin, but not exist for Immich?

*One thing I should mention: Just a few days ago, I moved these paths to their current location. In order to get that working, I had to change the ACL Mode for this dataset from “Restricted” to “Passthrough”. Perhaps this has something to do with this issue, but:

  1. It is still set to Passthrough
  2. I reiterate that Jellyfin isn’t having issues with it
  3. It was working fine for a few days on Dragonfish

Anyone have any ideas? I don’t want to screw this up because I’ve spent many hours curating my Immich content.

A few of other things worth noting:

  1. Immich was originally installed through the TrueNAS catalog, NOT TrueCharts
  2. I use Host Path for all paths configured in all apps, including Immich
  3. I did read the “Host Path ACL” section of the release notes. However, I didn’t believe this applied in the case of Immich because, although I was using host paths, I did not apply any ACLs.

I’m in the same boat. Immich and PassBolt, both gave an alert that they failed migration. Both were installed this year from the TrueNAS catalog and were working great before I migrated to EE.

Let me know if you figure anything out. Getting pretty desperate to fix this.

I can’t even figure out which file is storing the paths that the migration script is claiming don’t exist. Do you know where this is? Maybe I can make some temporary patches just to get the app migrated, then fix later.

I’m not even sure where to begin. But does anyone know of a way to attempt the migration again for the two that failed without rolling back to Dragonfish then upgrading back to EE?

Both of you might want to attempt TrueNAS Bug Ticket submission. It’s worth a try since nobody really replied to this thread.

Same boat here (Immich no longer starting up after migration), however the reason for rejection is a bit different:

Error response from daemon: Range of CPUs is from 0.01 to 1.00, as there are only 1 CPUs available

It makes no sense to me because I have plenty cores available and no one to use them, plus the CPU section of my Immich config only requires 1 “CPU”.

Anyone has the link to submit a TrueNAS Bug Ticket?
I read somewhere that it is a user-supported app though, not sure that will work. It’s worth a try anyway. If I find a link, I’ll definitely submit.

Directly from TrueNAS GUI, top right. SMILE is feedback / Report a Bug, or top right of this forum is Report A Bug

In GUI, you can attach debug dump there or on forum link, come back and attach

Thanks! Ticket now submitted (with my specific conditions though).
https://ixsystems.atlassian.net/browse/NAS-132700

1 Like

Is this a virtual or bare metal install? According to your debug, the entire system has only one CPU core.

Planning to follow this as I’d love to upgrade but do not want to run into similar issues. Thanks for reporting.

This is bare metal, an old z800 repurposed. Single Intel(R) Xeon(R) CPU X5650 @ 2.67GHz which is 6 core with HT enabled so it should reach 12 cores.
That is weird, thanks for catching that! It was not an issue in DF.

I think there is something to do with “The pgData dataset must be owned by the user netdata (UID 999) for postgres to start.”

Source:
https:/ /immich.app/docs/install/truenas/

Just to close the loop on my issue, a few days ago, the app indicated that it had an update available. I triggered the update and it starting working again. No data lost, everything is working as expected again.