mimmus

do I need to: - stop containerd service - start from cli in debug mode ?

Here is debug output: [containerd-debug.log](https://github.com/containerd/containerd/files/8728077/containerd-debug.log) I think that containerd never ended its startup process, I was unable to interrupt with CTRL-C, node was in NotReady status, then I rebooted it....

I understand. I saw that similar issues are solved in more recent versions of "containerd" but I'm locked to this version by CentOS and by vendor.

problem is that our k8s distro’s management script (a mix of Terraform+Ansible) sometimes decides to restart containerd and dies. in addition, I suspect that forcibly deleting pods in Terminating state...

> > 1. figurę out the initial problem, fix that... It seems that our nodes are missing this kernel parameter: `fs.may_detach_mounts` Setting it to 1: `echo 1 > /proc/sys/fs/may_detach_mounts` released...