Nicholas Turner
Nicholas Turner
Sorry, I was only just able to start on this issue. Attempting to repro this, I created a kops cluster using the CCM, with one master node and one worker...
The simple case of terminating both instances at the same time also did not end up with the behavior your described, and the kops mechanism to restore etcd state is...
By staggering the instance deletion times slightly, I thought I had reproduced the issue. The load balancer did not have the new instance for a few minutes after the CCM...
Can you share the CCM logs from the new CCM, from startup to startup time + 6 or 7 minutes (without restarting it)? Additionally the instance ids, private DNS names,...
/kind bug /triage accepted
I spoke with @jaypipes and he is going to create a repository under the aws organization to put the throttle package and any other widely shared utils. @M00nF1sh WDYT about...
/test pull-cloud-provider-aws-e2e
/test pull-cloud-provider-aws-e2e
/test pull-cloud-provider-aws-e2e
/kind bug /triage accepted /release-note-edit ```release-note Fixed a bug which reduces the number of unnecessary calls to STS in the event of assume role failures```