machine-learning pod CrashLoopBackOff
I'm running immich 1.58.0 in my k3s cluster and see the machine learning pod constantly crashing....is this due to the time it takes to download the models?
I can´t tell the reason for the crash from the log messages...is there any way to verify my theory?
What is the expected size of all the downloaded models so that i can compare it with the combined size of the folders in the machine learning cache?
1 Reply
i was able to get the pod running with overwriting the defaults for the for the machine-learning probes (liveness, readiness, startup) in the helm chart values
also it seems that a clean download of all the models is currently around 1.3G