Closed InAnimaTe closed 4 years ago
Hi @InAnimaTe, thanks for reporting this. What happens if you run the update command again? (In this case it should be safe since it won't upgrade to 1.16).
Looks like that seemed to work. (I'd share the output but its massive and I don't feel like filtering out private stuff) It ran successfully, did some things, and I see an updated Cloudformation stack:
2020-04-28 15:14:24 UTC-0400 eksctl-arryn-cluster UPDATE_COMPLETE -
2020-04-28 15:14:23 UTC-0400 eksctl-arryn-cluster UPDATE_COMPLETE_CLEANUP_IN_PROGRESS -
2020-04-28 15:14:20 UTC-0400 IngressDefaultClusterToNodeSG CREATE_COMPLETE -
2020-04-28 15:14:20 UTC-0400 IngressNodeToDefaultClusterSG CREATE_COMPLETE -
2020-04-28 15:14:20 UTC-0400 IngressDefaultClusterToNodeSG CREATE_IN_PROGRESS Resource creation Initiated
2020-04-28 15:14:19 UTC-0400 IngressDefaultClusterToNodeSG CREATE_IN_PROGRESS -
2020-04-28 15:14:19 UTC-0400 IngressNodeToDefaultClusterSG CREATE_IN_PROGRESS Resource creation Initiated
2020-04-28 15:14:19 UTC-0400 IngressNodeToDefaultClusterSG CREATE_IN_PROGRESS -
2020-04-28 15:14:13 UTC-0400 eksctl-arryn-cluster UPDATE_IN_PROGRESS User Initiated
2020-02-20 18:33:40 UTC-0500 eksctl-arryn-cluster UPDATE_COMPLETE
@martina-if could you clarify what things are actually being done here post-upgrade (or point me to the relevant documentation/functions)? Seems like some security group tuning for Ingress to function properly.
So basically, if this step times out, just re-run the command and I should be good?
@InAnimaTe eksctl performs a few operations but mainly upgrading the control plane and the stacks so I think rerunning it would perform the stack upgrade. This is not something to rely on because there are other things happening too and as I said, there is no continuation implemented.
I'm glad that it worked :+1:
What happened?
A few minutes after this, the cluster does appear to be healthy and shows version 1.15.
What you expected to happen?
I expected the cluster to be upgraded within the ~25 minute timeframe, as the other 3 clusters I did this upgrade on completed within time. Example:
Additionally, I can't see anything in Cloudformation (the cluster stack doesn't have any new events, as I'd expect) about the upgrade as I presume some of eksctl's post upgrade process didn't run? Am I now in a bad state? How do I properly "finish" this upgrade with eksctl?
How to reproduce it?
Anything else we need to know?
Versions
Post upgrade:
Logs https://gist.github.com/InAnimaTe/814af03d2a247b1667bbc10a8b05837e