Open mhagen-vmware opened 7 years ago
Mike should this be in "New Issues"? Don't want to question the master but...
I'm going to start on #6039 first, @cgtexmex you want to take this one?
In general, VIC triggering a TaskInProgress "error" is not going to cause any harm. We could check the target resource for any task in progress via the recentTask
property, but I think it'll always be a race since any VM resource could be transitioned to the RECONFIGURING state outside of VIC (such as via the UI). But we may be able to reduce the chance of this case from the VIC perspective.
based on the above info and after conversations with @mhagen-vmware, pulling this from 1.2
It's likely that this will result in a new condition in the pkg/vsphere/tasks for immediate retry, along with the VM_STATE_GUEST_OPERATION (not necessarily the actual state name!) and maybe VM_MIGRATION.
@mhagen-vmware I will look through the logs for this as I am working on #6370 as well. Looks like this should get handled by the task
package and if it is not we will need to wrap this in the task retry which is designed for all of the intermittent failures due to VC state/config fluctuations.
@mhagen-vmware did this occur on a specific CI run? I am curious if you also have the VCH logs.
I am grabbing this for Sprint 18. I intend to atleast confirm that the tasks
package is handling this properly. Or figure out where the task package is not being used where it should be.
Generated from: docker %{VCH-PARAMS} create ${busybox} /bin/top
This is from 1-10-Docker-PS test.
It actually causes an exception that the server ends up catching from the looks of it, but I don't think we should be causing exceptions in the server in our standard day to day ops.