Stu Charlton

Is this to be scheduled for 1.10?

The workaround for this is to cordon+drain the node on failure after X minutes. This can be automated by a simple controller script loop that watches for node conditions.

For what it’s worth , I am curious @marratj what you mean by the PVs getting broken? The PVs are FCDs under the covers and vcenter maintains the link to...

@SandeepPissay speaking from what I’ve seen in the past, other situations with our BOSH experience with TAS and TKGI - different datastores on different compute clusters, different datacenters (ideally some...

Finally looking to include this, will mull over how I'd like to see it merged if there are any tweak's I'll add it to a branch.

There are three ways to fix this one @grenzr 1. Customer enables DDNS registration on their DNS server for their K8s clusters (or deploys an intermediate DNS server) so that...

I’ve tried that, but I don’t believe kubelets can self-register themselves as a master with the `—node-labels` option, there’s only a small set of labels allowed: https://github.com/kubernetes/kubernetes/blob/master/pkg/kubelet/apis/well_known_labels.go This is why...

This can be mitigated by using the etcdctl command to remove the member and re-adding it. It has been resolved in CFCR etcd release 1.12.4 by using an older version...

Try it in a WSL2 instance & terminal window? This looks like the underlying NVIDIA library is panicking. Docker Desktop might not be able to reach the NVIDIA card unless...