Pods keep failing to start due to Error 'lstat /proc/?/ns/ipc : no such file or directory: unknown'

Issue description

Since yesterday (2022/09/06 KST), all pods in the namespaces are failling to start up because of following error:

2022-09-07 14:12:21.00 UTCgymwork-django-dev-db-0[pod-event]Pulling image "docker.io/bitnami/minideb:stretch"
2022-09-07 14:12:21.00 UTCgymwork-django-dev-db-0[pod-event]Successfully pulled image "docker.io/bitnami/minideb:stretch" in 94.625212ms
2022-09-07 14:12:21.00 UTCgymwork-django-dev-db-0[pod-event]Created container init-chmod-data
2022-09-07 14:14:02.00 UTCgymwork-django-dev-db-0[pod-event]Error: failed to create containerd task: start failed: dial /run/containerd/s/ef4ee4b11e9b5fa9ef7fecf2085189f1cfb387a54111ad404a39f57fee36314a: timeout: unknown
2022-09-07 14:14:03.00 UTCgymwork-django-dev-db-0[pod-event]Pod sandbox changed, it will be killed and re-created.
2022-09-07 14:14:11.00 UTCgymwork-django-dev-db-0[pod-event]Successfully pulled image "docker.io/bitnami/minideb:stretch" in 116.5723ms
2022-09-07 14:14:13.00 UTCgymwork-django-dev-db-0[pod-event]Error: failed to create containerd task: failed to create shim: OCI runtime create failed: container_linux.go:364: creating new parent process caused: container_linux.go:2005: running lstat on namespace path "/proc/1082682/ns/ipc" caused: lstat /proc/1082682/ns/ipc: no such file or directory: unknown

I checked that the same error occur when I deploy new dev environments in a new namespace as well.
I found that error showing up after I woke up pods from sleep mode.
It is weird that I’ve been using the same chart with exactly same setting in 100+ days.

Steps to reproduce the issue

  1. Redeploy (any existing) charts including postgres, minio (okteto helm), and my own helm chart.

What’s the expected result?

  • Pods (init-container, containers) are starting and raising no errors.

What’s the actual result?

  • Pods are failing and raising the error above.
1 Like

Same here

2022-09-08 22:00:11.00 UTCdeployment-demo-reset-27711240-4chpk[pod-event]Successfully pulled image "bitnami/kubectl" in 83.186364ms
2022-09-08 22:00:12.00 UTCdeployment-demo-reset-27711240-4chpk[pod-event]Created container kubectl
2022-09-08 22:00:13.00 UTCdeployment-demo-reset-27711240-4chpk[pod-event]Error: failed to create containerd task: failed to create shim: OCI runtime create failed: container_linux.go:364: creating new parent process caused: container_linux.go:2005: running lstat on namespace path "/proc/2723689/ns/ipc" caused: lstat /proc/2723689/ns/ipc: no such file or directory: unknown

Is this still happening to you? Also, is this in cloud.okteto.com or your own infrastructure?

I had similar errors but the issue seems to be resolved since Friday evening in my case… :ballot_box_with_check:

1 Like

Nope. I just found it’s not happening right now.
It kept occurring in cloud.okteto.com dev environment from 2022/9/6 KST ~ 2022/09/11 KST.
Is there any way to resolve this issue when the same issue comes up?