We are trying to upgrade our GKE cluster and the operations keeps timing out at the pulumi end with

10

mins timeout. We are stuck and not able to proceed.

Does re-running the

pulumi update

solve this? Or does

pulumi refresh

followed by re-running

pulumi update

work? Which change is it you are making that triggers this? Is it the addition of

networkPolicy

? Does the change eventually succeed in the GKE cluster? https://github.com/pulumi/pulumi/issues/2655 is tracking making timeouts configurable, but I would expect re-running

update

and/or

refresh

to address this if it is just a timeout issue and not a GKE issue.

@white-balloon-205 I had this issue occasionally when increasing the

version

property of a large gke nodepool. Since GKE updates each node in the pool one by one and it takes typically ~4-5 minutes per node to do so it can easily exceed the 10 minutes operation timeout. Rerunning pulumi update and refresh after the update completed solved the problem for me.

@white-balloon-205 Yes , it is working. But we would certainly like an option to override this issue. The problem is that we would like to run this from CD and we cannot have an

refresh

and

update

part of standard installation. So we end up running from our local box which is not desirable.