In the CI for PMEM-CSI, we bring up and tear down PMEM-CSI containers quite a lot. Recently we ran into this failure:
[2020-06-05T09:31:10.587Z] container_linux.go:349: starting container process caused "process_linux.go:449: container init caused \"rootfs_linux.go:58: mounting \\\"/sys\\\" to rootfs \\\"/var/lib/containers/storage/overlay/a12aa66fc67b2aff8ba5dc15537b03c6bedc64fa20ef79ebf4e69aa7e3b8d34d/merged\\\" at \\\"/var/lib/containers/storage/overlay/a12aa66fc67b2aff8ba5dc15537b03c6bedc64fa20ef79ebf4e69aa7e3b8d34d/merged/sys\\\" caused \\\"no space left on device\\\"\""
This affected all PMEM-CSI instances on all nodes:
I came into a similar error message and solve it by unmounting all subdirectories (seems like Docker failed to clean them up). I am not sure if this is the same issue but just posting for your information.
In the CI for PMEM-CSI, we bring up and tear down PMEM-CSI containers quite a lot. Recently we ran into this failure:
This affected all PMEM-CSI instances on all nodes:
But according to df, there is enough space:
This was on Clear Linux 32690, Kubernetes 1.17, with crio as CRI.
Obviously this report doesn't have enough information to reproduce or fix the issue. I'm just reporting it so that we have a place to discuss it.