R
Runpod•2d ago
bghira

pod not running

why does this happen so much? 
-- RUNPOD.IO --
Enjoy your Pod #2hcudh6igw3bgb ^_^

Error response from daemon: container 4081368d927e2905cc411823cc920bda3c553d17f4c819f0f9f64aa680e0fc30 is not running
Connection to 100.65.14.18 closed.
Connection to ssh.runpod.io closed.
-- RUNPOD.IO --
Enjoy your Pod #2hcudh6igw3bgb ^_^

Error response from daemon: container 4081368d927e2905cc411823cc920bda3c553d17f4c819f0f9f64aa680e0fc30 is not running
Connection to 100.65.14.18 closed.
Connection to ssh.runpod.io closed.
it seems to waste credits for startup and then it idles thinking it's running, but nothing can connect.
6 Replies
bghira
bghiraOP•2d ago
No description
bghira
bghiraOP•2d ago
like, shouldn't this kind of thing be detected by middleware? 
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 prestart hook #0: exit status 1, stdout: , stderr: Using requested mode 'cdi' invoking the NVIDIA Container Runtime Hook directly (e.g. specifying the docker --gpus flag) is not supported. Please use the NVIDIA Container Runtime (e.g. specify the --runtime=nvidia flag) instead.: 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 prestart hook #0: exit status 1, stdout: , stderr: Using requested mode 'cdi' invoking the NVIDIA Container Runtime Hook directly (e.g. specifying the docker --gpus flag) is not supported. Please use the NVIDIA Container Runtime (e.g. specify the --runtime=nvidia flag) instead.: unknown
Dj
Dj•12h ago
Looking at the logs for this Pod, this error can happen on misconfigured Community Cloud hosts, generally the platform makes an assumption that dhis error sort of doesn't happen. For about 200ms at a time docker tries to start the pod, sees an error, then repeats.
No description
Dj
Dj•12h ago
I'll flag this machine with the host who runs it, and I've delisted it in the meantime.
bghira
bghiraOP•12h ago
it should do a bit more than that for community hosts since it's RunPod that's doing the billing (in my opinion) it's only so much blame that can go to the end-hosts
Dj
Dj•12h ago
The error described is a pretty basic config error, where you're entirely right at the same time we'd expect for the machines to be properly configured in the first place. I'll make sure we build something for it in our project to identify errors with GPUs.

Did you find this page helpful?