Migration Guide: Immich Bookworm to Immich Trixie for existing Immich LXCs 🙌🎉 🤯

[vhsdream]

Hey. It's vhsdream with another long-winded Immich screed.

So the Immich Base Images repo has moved to Debian Trixie already, and it wasn't a big deal for any Docker users, because well I guess that's a big benefit to Docker. Anyway, we've just pushed the Debian Trixie update to Immich and there are breaking changes for sure. I wouldn't blame you if you wanted to start fresh with a new Immich installation.

This guide is for those intrepid (and maybe foolish) souls who wish to make the transition to Immich Trixie (13), from their existing Immich Bookworm (12) installation. Perhaps some of you were brought here by your Immich instance refusing to update, because you are on Bookworm and if we allowed that to go through all hell would break loose. Read on!

Massive warning ⚠️⚠️⚠️

I am still using Proxmox Virtual Environment 8.4. While I am pretty sure this is going to work for those of you who are on Proxmox VE 9, I have no first-hand experience with it. That being said, TAKE A SNAPSHOT AND A BACKUP before proceeding! Luckily the upgrade to Debian 13 on the part of Immich does not come with any mandatory DB/VectorChord changes (yet), so your photos/videos will be there and accessible afterwards. But still. Take 1000 backups and caution with your precious memories.

Upgrade your LXC to Trixie

Follow these steps/copy and paste these commands into your LXC console. In this order.

• systemctl disable --now immich-web immich-ml
• apt update && apt dist-upgrade -y
• apt install screen -y
• sed -i 's/bookworm/trixie/' /etc/apt/sources.list
• sed -i 's/bookworm/trixie/' /etc/apt/sources.list.d/jellyfin.sources
• sed -i 's/bookworm/trixie/' /etc/apt/sources.list.d/pgdg.list
• rm -f /etc/apt/sources.list.d/immich.list
• For OpenVINO users - apt-mark hold libigdgmm12
• apt update. Any errors? DO NOT PROCEED
• screen apt dist-upgrade. Follow along with the upgrade and make your own choices regarding overwriting configs. You might see ones for crontab, e2scrub_all, or sshd_config. What you decide is up to you.
• Any errors during the upgrade? DO NOT PROCEED
• Remove and clean: apt autoremove -y && apt autoclean -y
• apt modernize-sources, then approve the changes with a y.
• OK you are going to see a warning, about the postgres repo. No worries, lets fix it!
• sed -i 's|Signed-By:|& /etc/apt/trusted.gpg.d/postgresql.gpg|' /etc/apt/sources.list.d/pgdg.sources
• apt update. You'll see a warning about NodeJS. Just ignore it.
• Any other errors? DO NOT PROCEED
• For those who elected to enable VAAPI for hardware transcoding, before rebooting you may need to change the GID of your /dev/dri/renderD12x device. Check if you need to do this like so: ls -l /dev/dri. If in place of the group name render you see some number (992 perhaps?), you will need to change the GID in the 'Resources' tab for the LXC in the ProxmoxVE UI:

• Then again, you might not. I think the GID change is for brand new installs that use the Debian 13 LXC template, but I thought I'd mention it just in case.
• OK, time to restart the LXC, so do that, and reattach.

Update Immich

First let's reset Immich to start on boot. Run systemctl enable immich-ml immich-web.

Next, let's force an update to Immich. There's a chance your version is already up-to-date but we want to migrate.

Open the /root/.immich file in a terminal text editor (nano,vi,etc) and change the version number to something below what it is. For example, if the version is 1.141.1 then change it to 1.141.0. Save and exit the editor.

Now the update should begin once you run the update command. I suggest choosing 'verbose' this time. You'll likely see Imagemagick recompiling so this update will take a bit longer than normal. And with any luck it will complete. As almost always with any update to Immich, you should make a habit of tailing the logs which will let you know if anything is broken.

• Machine-learning: tail -f /var/log/immich/ml.log
• Immich Web: tail -f /var/log/immich/web.log

You made it? Nice work 😉

Problems? I sure hope you have snapshots/backups! And let me know in this discussion what happened. I expect logs and informative screenshots!

[mokhos]

Hi @vhsdream,
thanks for the guide. I did went through it in my pve-9.0.9 and everything went smoothly to the best of my knowledge but then at the end when updating the immich itself(I changed the version in .immich file from 1.141.2 to 1.141.1), I saw this error:

...
-- Installing: /usr/local/lib/cmake/libheif/libheif-config-version.cmake
-- Installing: /usr/local/lib/x86_64-linux-gnu/gdk-pixbuf-2.0/2.10.0/loaders/libpixbufloader-heif.so
-- Set non-toolchain portion of runtime path of "/usr/local/lib/x86_64-linux-gnu/gdk-pixbuf-2.0/2.10.0/loaders/libpixbufloader-heif.so" to ""
  ✔️   Recompiled libheif
  ✔️   Image-processing libraries up to date
  ✔️   Stopped immich
cp: cannot stat '/opt/immich/app/machine-learning/ml_start.sh': No such file or directory

[ERROR] in line 103: exit code 0: while executing command cp "$ML_DIR"/ml_start.sh "$INSTALL_DIR"

Looks like the script is not there anymore!

From the logs point of view, they look like this:

root@immich:~# tail -f /var/log/immich/web.log
[Nest] 8651  - 09/18/2025, 2:00:00 AM     LOG [Microservices:BackupService] Database Backup Starting. Database Version: 16
[Nest] 8651  - 09/18/2025, 2:00:13 AM     LOG [Microservices:BackupService] Database Backup Success
[Nest] 8651  - 09/18/2025, 3:21:08 AM     LOG [Microservices:VersionService] Found v1.142.1, released at 9/16/2025, 4:51:53 PM
[Nest] 8651  - 09/18/2025, 5:21:08 AM     LOG [Microservices:VersionService] Found v1.142.1, released at 9/16/2025, 4:51:53 PM
[Nest] 8651  - 09/18/2025, 7:21:08 AM     LOG [Microservices:VersionService] Found v1.142.1, released at 9/16/2025, 4:51:53 PM
[Nest] 8651  - 09/18/2025, 9:21:08 AM     LOG [Microservices:VersionService] Found v1.142.1, released at 9/16/2025, 4:51:53 PM
[Nest] 8651  - 09/18/2025, 11:21:08 AM     LOG [Microservices:VersionService] Found v1.142.1, released at 9/16/2025, 4:51:53 PM
[Nest] 8651  - 09/18/2025, 1:21:08 PM     LOG [Microservices:VersionService] Found v1.142.1, released at 9/16/2025, 4:51:53 PM
[Nest] 8651  - 09/18/2025, 3:21:08 PM     LOG [Microservices:VersionService] Found v1.142.1, released at 9/16/2025, 4:51:53 PM
[Nest] 8651  - 09/18/2025, 5:21:08 PM     LOG [Microservices:VersionService] Found v1.142.1, released at 9/16/2025, 4:51:53 PM

root@immich:~# tail -f /var/log/immich/ml.log
                             of inactivity.                                     
[09/14/25 11:21:07] INFO     Initialized request thread pool with 4 threads.    
[09/14/25 11:21:07] INFO     Application startup complete.                      
[09/18/25 18:57:40] INFO     Handling signal: term                              
[09/18/25 18:57:40] INFO     Shutting down                                      
[09/18/25 18:57:40] INFO     Waiting for application shutdown.                  
[09/18/25 18:57:41] INFO     Application shutdown complete.                     
[09/18/25 18:57:41] INFO     Finished server process [8663]                     
[09/18/25 18:57:41] ERROR    Worker (pid:8663) was sent SIGTERM!                
[09/18/25 18:57:41] INFO     Shutting down: Master

Could you be able to give me a hint here?
Thanks in advance.

P.S: I already setup daily backups. So I have some backups available!

[mokhos]

I attempted first to update the immich, then I got the error that I must update the lxc to trixie and then ended up here!

[MickLesk]

Which Version before?

[fedottt]

Y'all, that's weird, because machine-learning wouldn't even start if there is no ml_start.sh in /opt/immich/app/machine-learning, so it was there at some point. Did either of you attempt to update Immich before upgrading to Trixie? Maybe the check I added is failing on PVE9 and the script tries to keep going?

pve8 now, host machine on bookworm
i update lxc to trixie, next reboot.
before update to trixie i have 1.142.0 version
copy folder, run update. it fails.
copy .sh and run update in extended mode
and it done ok. now i have 1.142.1
ml_start.sh after update gone from /opt/immich

[mokhos]

ahha, i fixed it

i found folder machine-learning in /opt/immich/source/ and ml_start.sh in /opt/immich

next cp /opt/immich/source/machine-learning/ /opt/immich/app/machine-learning/ cp /opt/immich/ml_start.sh /opt/immich/app/machine-learning/ml_start.sh

and run update

I went ahead with this solution and it fixed my problem.
Hope it doesn't mess with later updates.
Thanks anyway :)

[hyzteric]

Just to report that I just updated without any issue, updating from pve 9.0.9, immich 1.42.0.
I did not have to touch the .immich file since it updated to 1.43.0

[ozgencaslan]

I had no errors up to the step with GID. After trying to reboot i get the following error :

"run_buffer: 571 Script exited with status 25 lxc_init: 845 Failed to run lxc.hook.pre-start for container "103" __lxc_start: 2034 Failed to initialize container "103" TASK ERROR: startup for container '103' failed"

[vhsdream]

Is this the "13.1" not supported bug?

[MickLesk]

Looks so. Proxmox Tell its Not an bug, but with latest pve-container its fixed as Bug 😂

[deThommy]

I am curently on proxmox 8.3.5 and i got the same exact error (different lyx id ofc).
unfortunately i'm not too deep into all of this. ChatGPT tells me the following after a little bit of trial and error back and forth:

Script exec ... lxc-pve-prestart-hook ... produced output: unsupported debian version '13.1'

upgrading all my packages (seemed to be long over due) solved the problem! CHEERS!

[gpuma]

I had the same issue. I think it could be added a line at the beginning of this guide: "Make sure to update the packages in your PVE instance: apt update && apt upgrade -y". This fixed the container not being able to start.

I'm on PVE 9 btw

[cavinpabua]

Just in case anybody needs, if apt update && apt upgrade -y doesnt solve your problem then you need to edit
/usr/share/perl5/PVE/LXC/Setup/Debian.pm
on line 39
and change

from

`die "unsupported debian version '$version'\n" if !($version >= 4 && $version <= 13);`

to

`die "unsupported debian version '$version'\n" if !($version >= 4 && $version <= 14);`

[lukeski14]

Everything went smoothly for me...thanks for the extremely detailed guides as always!

[GjMan78]

The upgrade went smoothly. Thanks!

[finkerle]

That worked like a charm, thank you!

[piyush97]

Smooth as butter!

[lnoormohamed]

When i went through the upgrade steps. it seems. to be all good untill the reboot. could not get it to start up again.

[MickLesk]

Read above and Update your pve-container

[lnoormohamed]

Thanks. I ended up deleting it and reinstalling a fresh copy, but I had updated the containers before installing a fresh copy

[O-beron]

Upgrade went smoothly, seems to be working so far. Thanks so much for the detailed walk-through !

Fresh Immich 1.142.0 install from last week... that's just my luck that I start using a service and a mandatory OS upgrade pops up a few days later haha

Anyway, didn't run into the error mentioned by @mokhos and @fedottt so didn't have to copy the files.

[CLKRUN]

✔️ Updated immich web and microservices
⏳ Updating machine-learningcpython-3.11.13-linux-x86_64-gnu (download) ------------------------------ 0 B/28.76 MiB
Using CPython 3.11.13
Creating virtual environment at: /opt/immich/app/machine-learning/ml-venv
Resolved 140 packages in 4ms
× Failed to build immich-ml @ file:///opt/immich/source/machine-learning
├─▶ Failed to install requirements from build-system.requires
├─▶ Failed to download pluggy==1.6.0
├─▶ Failed to extract archive: pluggy-1.6.0-py3-none-any.whl
├─▶ I/O operation failed during extraction
╰─▶ Failed to download distribution due to network timeout. Try increasing UV_HTTP_TIMEOUT (current value: 30s).

[ERROR] in line 169: exit code 0: while executing command $STD sudo --preserve-env=VIRTUAL_ENV -nu immich uv sync --extra cpu --active -n -p python3.11 --managed-python

[vhsdream]

Restore from a backup, then before running update, do export UV_HTTP_TIMEOUT=60 or whatever.

[CLKRUN]

Failed to download distribution due to network timeout. Try increasing UV_HTTP_TIMEOUT (current value: 30s).

Unfortunately, it doesn't work even though I set it to 90 and check with echo $UV_HTTP_TIMEOUT. Oh well, I'll just create a new LXC container.

[vsa-d]

Worked perfectly with your step-by-step tutorial. I came from 1.142.1, changed .immich to 1.142.0 and updated. No errors. Just the warning you described. Thank you very much!!!

[madcar86]

This was great, thank you for putting it together.

Will there be a script add to the site to update to Trixie? That way we dont have to do step by step for each lxc

[MickLesk]

No, because the most LXC break with dep Issues and it Takes 2-5min per lxc to upgrade

[s09931]

I just wanted to leave a commented to say this worked flawlessly for me on Proxmox 8.4.14. I was on Immich 1.142.0, changed to 1.141.0 and upgraded from there. Thank you for the detailed walkthrough.

[rh8056]

Also reporting success. Running Proxmox 8.2.2 and had Immich 1.41.1.

I suppose these types of things are avoided with VM images like Home Assistant? I'm very happy for the LXC support here, but it would be great for a more turnkey solution to evolve once Immich hits stable.

[twillux]

Also wanted to echo success with no issues upgrading Immich 1.142.0 on Proxmox 8.4.14. I then upgraded my host to Proxmox 9.0.10, and also no issues with the LXC after the host upgrade.

[xaabi6]

Upgraded without issues using Proxmox 8.4.14 and Immich 1.142.1. Thanks for this great guide!

[jjpmdejong]

Like a charm...

Thanks!

[maf2we]

Update to trixie and update of immich went well.

apt-mark hold libigdgmm12

keeps version of libigdgmm12 - shouldn't i unhold this after upgrading.

[vhsdream]

Nope, pinning it is intentional. The Immich devs haven't updated it in their base image so I'm not going to either 😄

[Strana-Mechty]

This helped a lot! Thank you!

[leojei]

This is a solid guide, I was able to follow the steps and get immich LXC updated after upgrading the LXC to Trixie. Kudos and much appreciated!

[PaulePils]

I wanted to run the "update" command and got this error. The host is Proxmox 9 and the LXC already running Trixie

   ✖️   Wrong Debian version detected!
   ✖️   You must upgrade your LXC to Debian Trixie before updating.
   ✖️   Please visit https://github.com/community-scripts/ProxmoxVE/discussions/7726 for details.
        If you have upgraded your LXC to Trixie and you still see this message, please open an Issue in the Community-Scripts repo.

 Static hostname: immich
       Icon name: computer-container
         Chassis: container ☐
      Machine ID: 76d2b6a83c92432db3b640952e2d6643
         Boot ID: 6f2c46a6e9ce4542af9fce7636a41fff
  Virtualization: lxc
Operating System: Debian GNU/Linux 13 (trixie)    
          Kernel: Linux 6.14.11-2-pve
    Architecture: x86-64

[MickLesk]

Then you dont have followed the Guide above. You have an immich source left. So the Script Work fine, it Checks for the Guide. If you dont follow the good Guide of vhs, then we cannot Help you. The URL and the guide are not linked for fun, but to be strictly adhered to.

[vhsdream]

@PaulePils In your Immich LXC, check for /etc/apt/sources.list.d/immich.list. If it exists, delete it, then try to update again. If it doesn't exist, then I can't help you without more info about your system.

[PaulePils]

It works now :-)
I thought when I first used the script it automatically was Trixie but it was not...
After following the script the "update" command works perfectly
Sorry for the disturbance

[qweargss]

worked just as well as this guide was written! thank you very much :)

[Spanakopitas]

Worked flawlessly with Proxmox 9.0.10. Thank you so much!

[kamisama824]

Just thought i may add something here, but i noticed that after i conducted the process of upgrading (worked perfectly well without any issues on Proxmox version: 8.4.14), i saw something quite strange where my space utilization doubled e.g. 80GB of active storage of photos and videos before the upgrade to after the upgrade jumped to 160GB.

Did anyone else experience this? (I have yet to investigate the exact cause).

[kamisama824]

Actually, after digging into the directory. I've discovered that there were temp files created during the upgrade. After a reboot the temporary files were removed.

[fiveiron]

I followed your guide and thank you that seemed to work fine. After updating Debian I then updated Immich successfully to 1.143.1. Now however, it says version 1.144.1 is available but won't let me update to it.

[fiveiron]

Thanks. My confusion is centered around the fact that immediately after updating to 1.143.1, the pop-up came up saying I needed to update to 1.144.1 but the update script won't let me. I just assumed something went wrong with the upgrade.

[MickLesk]

Yeah because every Version will be checked by VHS before Upgrade, because some Upgrades in the past failed massive. And it was precisely for this case that I had built the pinning function. If you watch our releases or Changelog you see every Pin to next Version (Like some hours ago the merge to 144)

[vsa-d]

1.144.1 update was smooth without errors (came from1.43.1) Thank you very much @vhsdream

[vsa-d]

And sucessfully to 2.0.0. - thanks!!!!!

[tomatensaus]

Upgraded from 1.37??? all the way to 2.0.1. Thank you for the guide. Still on proxmox 8 and also had to enable trixie for lxc. After this the sync from the app started working again

[oerix]

Hey @vhsdream, I upgraded to trixie using your guide above, except from changing .immich file as that file didn't exist in my lxc, however once I upgraded to 2.0.0 using the script the web ui fails to load.

ML seems fine, here's the log for Immich Web:

lease use REINDEX to rebuild the index.
      at ErrorResponse (/opt/immich/app/node_modules/.pnpm/[email protected]/node_modules/postgres/cjs/src/connection.js:794:26)
      at handle (/opt/immich/app/node_modules/.pnpm/[email protected]/node_modules/postgres/cjs/src/connection.js:480:6)
      at Socket.data (/opt/immich/app/node_modules/.pnpm/[email protected]/node_modules/postgres/cjs/src/connection.js:315:9)
      at Socket.emit (node:events:519:28)
      at addChunk (node:internal/streams/readable:561:12)
      at readableAddChunkPushByteMode (node:internal/streams/readable:512:3)
      at Readable.push (node:internal/streams/readable:392:5)
      at TCP.onStreamRead (node:internal/stream_base_commons:189:23) {
    severity_local: 'ERROR',
    severity: 'ERROR',
    code: 'XX000',
    file: 'tuples.rs',
    line: '145'
  },
  sql: 'SELECT vchordrq_prewarm($1)',
  params: [ 'face_index' ]
}
Query failed : {
  durationMs: 8.857088999999178,
  error: PostgresError: deserialization: bad version number; after upgrading VectorChord, please use REINDEX to rebuild the index.
      at ErrorResponse (/opt/immich/app/node_modules/.pnpm/[email protected]/node_modules/postgres/cjs/src/connection.js:794:26)
      at handle (/opt/immich/app/node_modules/.pnpm/[email protected]/node_modules/postgres/cjs/src/connection.js:480:6)
      at Socket.data (/opt/immich/app/node_modules/.pnpm/[email protected]/node_modules/postgres/cjs/src/connection.js:315:9)
      at Socket.emit (node:events:519:28)
      at addChunk (node:internal/streams/readable:561:12)
      at readableAddChunkPushByteMode (node:internal/streams/readable:512:3)
      at Readable.push (node:internal/streams/readable:392:5)
      at TCP.onStreamRead (node:internal/stream_base_commons:189:23) {
    severity_local: 'ERROR',
    severity: 'ERROR',
    code: 'XX000',
    file: 'tuples.rs',
    line: '145'
  },
  sql: 'SELECT vchordrq_prewarm($1)',
  params: [ 'clip_index' ]
}
PostgresError: deserialization: bad version number; after upgrading VectorChord, please use REINDEX to rebuild the index.
    at ErrorResponse (/opt/immich/app/node_modules/.pnpm/[email protected]/node_modules/postgres/cjs/src/connection.js:794:26)
    at handle (/opt/immich/app/node_modules/.pnpm/[email protected]/node_modules/postgres/cjs/src/connection.js:480:6)
    at Socket.data (/opt/immich/app/node_modules/.pnpm/[email protected]/node_modules/postgres/cjs/src/connection.js:315:9)
    at Socket.emit (node:events:519:28)
    at addChunk (node:internal/streams/readable:561:12)
    at readableAddChunkPushByteMode (node:internal/streams/readable:512:3)
    at Readable.push (node:internal/streams/readable:392:5)
    at TCP.onStreamRead (node:internal/stream_base_commons:189:23) {
  severity_local: 'ERROR',
  severity: 'ERROR',
  code: 'XX000',
  file: 'tuples.rs',
  line: '145'
}
microservices worker error: PostgresError: deserialization: bad version number; after upgrading VectorChord, please use REINDEX to rebuild the index., stack: PostgresError: deserialization: bad version number; after upgrading VectorChord, please use REINDEX to rebuild the index.
    at ErrorResponse (/opt/immich/app/node_modules/.pnpm/[email protected]/node_modules/postgres/cjs/src/connection.js:794:26)
    at handle (/opt/immich/app/node_modules/.pnpm/[email protected]/node_modules/postgres/cjs/src/connection.js:480:6)
    at Socket.data (/opt/immich/app/node_modules/.pnpm/[email protected]/node_modules/postgres/cjs/src/connection.js:315:9)
    at Socket.emit (node:events:519:28)
    at addChunk (node:internal/streams/readable:561:12)
    at readableAddChunkPushByteMode (node:internal/streams/readable:512:3)
    at Readable.push (node:internal/streams/readable:392:5)
    at TCP.onStreamRead (node:internal/stream_base_commons:189:23)
microservices worker exited with code 1
Killing api process
Starting api worker
Starting microservices worker
[Nest] 58468  - 10/02/2025, 2:43:01 PM     LOG [Microservices:EventRepository] Initialized websocket server
Postgres notice: {
  severity_local: 'WARNING',
  severity: 'WARNING',
  code: '42602',
  message: 'invalid configuration parameter name "vchordrq.prewarm_dim"',
  detail: '"vchordrq" is a reserved prefix.',
  file: 'guc.c',
  line: '1225',
  routine: 'find_option'
}
Postgres notice: {
  severity_local: 'WARNING',
  severity: 'WARNING',
  code: '42602',
  message: 'invalid configuration parameter name "vchordrq.prewarm_dim"',
  detail: '"vchordrq" is a reserved prefix.',
  file: 'guc.c',
  line: '1225',
  routine: 'find_option'
  }
  

Any help would be appreciated, thanks

[MickLesk]

To be honest, I'm wondering how the .immich file could be gone. We always write it and check for it. .

[vhsdream]

That and the Vectorchord thing leads me to believe that Immich hadn't been updated in a while.

[oerix]

That and the Vectorchord thing leads me to believe that Immich hadn't been updated in a while.

Yes that's right, unfortunately I didn't check what version before upgrading, but I believe it's likely to be (v.1.137-) or 3-4 releases out, not sure sorry.

[wmoreland]

@oerix I was having the same issues but found the solution here which was to log in to the immich postgres (found the log-in info in /opt/immich/.env) and run the following two commands:

REINDEX INDEX face_index;
REINDEX INDEX clip_index;

After exiting the database, reloading the immich web page, voila it worked.

[oerix]

@wmoreland Great that fixed it, thank you! :)

Although note, that it did break 'SYMLINKS' again, so I had to run these useful commands again.

[M1SJE]

[vhsdream]

@M1SJE Just choose yes and keep going

[M1SJE]

Not sure what I did differently, but it seems to have worked this time.

1st time it just errored out and screwed it all up. Yay for snapshots.

Thanks for your help.

[Oly77]

SOLVED

Version: 1.142.1
Hi guys,
my debian is 13.0 so Trixie
And when i ask for an upgrade it still says "Wrong Debian version detected!"
Thanks for you help

[MickLesk]

Then you dont have followed the Guide above. Special with the Sources file

[Oly77]

The guide is to upgrade to Debian 12 to Debian 13 and upgrade Immich.
My Debian is already in 13 but Immich upgrade says it is not.

[vhsdream]

Just delete the /etc/apt/sources.list.d/immich.list file.

[MickLesk]

I didn't want to reveal it because apparently neither the source code is checked nor the instructions followed 😅

[putos11]

Nice job,,
Thank you very much..
Hat super funktioniert...

[orgebar]

Goddamn,
following your steps 1:1, this upgrade went smooth like a warm knife through butter! :D

THANK YOU!!!

[seratg]

At the end of the update process, i get qsql errors:

Installing symlink pointing to libvips.so.42 to /usr/local/lib/libvips.so
Installing symlink pointing to libvips-cpp.so.42.19.2 to /usr/local/lib/libvips-cpp.so .42
Installing symlink pointing to libvips-cpp.so.42 to /usr/local/lib/libvips-cpp.so
  ✔️   Recompiled libvips
  ✔️   Image-processing libraries up to date
  ✔️   Update available: immich 1.125.0 → 2.3.1
  ✔️   Stopped Services
Note, selecting 'postgresql-16-vchord' instead of './vchord.deb'
Upgrading:
  postgresql-16-vchord

Summary:
  Upgrading: 1, Installing: 0, Removing: 0, Not Upgrading: 1
  Download size: 0 B / 1,999 kB
  Space needed: 0 B / 7,773 MB available

Get:1 /opt/staging/vchord.deb postgresql-16-vchord amd64 0.5.3-1 [1,999 kB]
apt-listchanges: Reading changelogs...
(Reading database ... 52326 files and directories currently installed.)
Preparing to unpack /opt/staging/vchord.deb ...
Unpacking postgresql-16-vchord (0.5.3-1) over (0.4.3-1) ...
Setting up postgresql-16-vchord (0.5.3-1) ...
Processing triggers for postgresql-common (287.pgdg13+1) ...
Building PostgreSQL dictionaries from installed myspell/hunspell packages...
Removing obsolete dictionary files:
psql: error: connection to server on socket "/var/run/postgresql/.s.PGSQL.5432" failed : No such file or directory
        Is the server running locally and accepting connections on that socket?

[ERROR] in line 117: exit code 0: while executing command $STD sudo -u postgres psql - d immich -c "ALTER EXTENSION vector UPDATE;"

tail -f /var/log/immich/ml.log

[11/23/25 14:41:58] INFO     Initialized request thread pool with 4 threads.
[11/23/25 14:41:58] INFO     Application startup complete.
[11/23/25 14:42:10] INFO     Shutting down
[11/23/25 14:42:10] INFO     Handling signal: term
[11/23/25 14:42:11] INFO     Waiting for application shutdown.
[11/23/25 14:42:11] INFO     Application shutdown complete.
[11/23/25 14:42:11] INFO     Finished server process [316]
[11/23/25 14:42:11] ERROR    Worker (pid:316) was sent SIGTERM!
[11/23/25 14:42:11] INFO     Shutting down: Master

tail -f /var/log/immich/web.log

  syscall: 'connect',
  address: '127.0.0.1',
  port: 5432
}
microservices worker error: Error: connect ECONNREFUSED 127.0.0.1:5432, stack: Error: connect ECONNREFUSED 127.0.0.1:5432
    at TCPConnectWrap.afterConnect [as oncomplete] (node:net:1637:16)
microservices worker exited with code 1
Killing api process
Starting api worker
Starting microservices worker

[vhsdream]

Your Postgres service is not running. Make sure it's able to start - systemctl restart postgresql - and check systemctl status postgresql/journalctl -xeu postgresql in case there are any issues with it.

Your Immich version is pretty out of date, so there's a nonzero chance this 'upgrade' might require more manual intervention on your part.