machine-learning container crashes on Truenas when uploading videos
🗄️Server❓Question
Setup:
R7 3700x
RTX 1060 6gb
256gb boot ssd
4 10tb hdd in raid 5
Truenas Scale ElectricEel-24.10.2.2
Immich v2.5.0 (Installed through truenas apps page)
The problem:
Uploading videos to immich causes "machine-learning" container to get stuck on "starting." Manifests physically as my cpu usage pinned at 20-40% (usually less than 6%) and gpu fans pinned at max speed.
What I have learned so far:
-Trying to stop the app results in a "failed down action for immich" error
-Rebooting truenas stops the fan noise, but app crashes on startup
-After app crashes on startup, can be started again manually without issue and will work until next video upload
-Problem has occurred through multiple major and minor versions of Truenas and Immich since I built the server 1 year ago
-Changing the machine learning model in immich settings does not change anything
-Issue persisted through a fresh install (since I did that when migrating from old file structure to new some number of months ago)
Forgot to dl the logs from machine-learning last time it happened before restarting, but will try to trigger the issue again to collect some more data.
R7 3700x
RTX 1060 6gb
256gb boot ssd
4 10tb hdd in raid 5
Truenas Scale ElectricEel-24.10.2.2
Immich v2.5.0 (Installed through truenas apps page)
The problem:
Uploading videos to immich causes "machine-learning" container to get stuck on "starting." Manifests physically as my cpu usage pinned at 20-40% (usually less than 6%) and gpu fans pinned at max speed.
What I have learned so far:
-Trying to stop the app results in a "failed down action for immich" error
-Rebooting truenas stops the fan noise, but app crashes on startup
-After app crashes on startup, can be started again manually without issue and will work until next video upload
-Problem has occurred through multiple major and minor versions of Truenas and Immich since I built the server 1 year ago
-Changing the machine learning model in immich settings does not change anything
-Issue persisted through a fresh install (since I did that when migrating from old file structure to new some number of months ago)
Forgot to dl the logs from machine-learning last time it happened before restarting, but will try to trigger the issue again to collect some more data.
