R
Runpod3mo ago
elle

Container not running error

I was working on ComfyUI normally when it suddenly stopped working, a restart of the pod did not fix the problem. Now when creating new pods the container would not start running, any suggestions/fixes? Im on EU-RO Region running 4090rtx, cuda 12.7 on Runpod Pytorch 2.4 template
21 Replies
kamran2608
kamran26083mo ago
I'm getting the same issue - I cant get any containers to start running (in any regions, on any GPU and any template). Logs remain empty and I see "Container is not running"
julian
julian3mo ago
same issue ig
No description
james
james3mo ago
same issue its not related to your pod, nor to mine then came looking for this
Mahdi GWB
Mahdi GWB3mo ago
same issue here any solution ?
james
james3mo ago
im on EU-CZ-1 RTX 5090 cuda 12.8 custom image
Mahdi GWB
Mahdi GWB3mo ago
i tried diffrent ones all didnt work , just saying ( No logs found ) i even keep waiting and i lost even some funds :/ still nothing
james
james3mo ago
I'm not on ComfyUI, but it did stop suddenly and for me too; a new pod or even new network drive and pod wont even start the pod now.
Mahdi GWB
Mahdi GWB3mo ago
i just sent an email to support i think we should all do so they can nottice there is an issue (https://contact.runpod.io/hc/en-us/requests/new )
Aasif Khan
Aasif Khan3mo ago
same issue for me. not starting or running
No description
Mahdi GWB
Mahdi GWB3mo ago
same thing here
james
james3mo ago
did it
Aasif Khan
Aasif Khan3mo ago
working now guys. newly created pod started up
james
james3mo ago
confirm, mine works too again *new pod, had already terminated the other
mwon
mwon3mo ago
I'm having the same issue. I start a pod but not container is started. System logs the following:
error starting container: Error response from daemon: failed to create task for container: failed to create shim task: OCI runtime create failed: runc create failed: unable to start container process: error during container init: error running hook #0: error running hook: exit status 1, stdout: , stderr: Auto-detected mode as 'legacy'
nvidia-container-cli: requirement error: unsatisfied condition: cuda>=12.8, please update your driver to a newer version, or use an earlier cuda container: unknown
error starting container: Error response from daemon: failed to create task for container: failed to create shim task: OCI runtime create failed: runc create failed: unable to start container process: error during container init: error running hook #0: error running hook: exit status 1, stdout: , stderr: Auto-detected mode as 'legacy'
nvidia-container-cli: requirement error: unsatisfied condition: cuda>=12.8, please update your driver to a newer version, or use an earlier cuda container: unknown
What's worst is that although no pod I'm being charged anyway...
james
james3mo ago
this is not the same issue, ours was related to a temporary network outage, yours is a codebase problem
Madiator2011
Madiator20113mo ago
the error tells you what is wrong your docker image is based on CUDA 12.8 and you did not used filter to pick machines that supports it
mwon
mwon3mo ago
It is a new/recent filter? I have launched many pods in the past and never had to select machines based in CUDA version.
Madiator2011
Madiator20113mo ago
nope it always been there
Henky!!
Henky!!3mo ago
When will the other machines finally have 12.8?
God Of Destruction Kwijiboe
Same issues Loras failing to load slow startup slow generation
MD
MD3mo ago
So some RTX 4090s have CUDA 12.8 and others don't? I use 4090s a lot and usually work with my image and then sometimes I get the unsatisfied condition error.

Did you find this page helpful?