In the data I posted above, I just noticed that the timestamps are weird.
The task there says it ran from 5:10:39 to 5:15:39, which is exactly 5m. On vSphere, the matching task says it ran from 5:10:40 to 5:15:53. So the PowerCLI task claims that it finished earlier than vCenter says it finished, by 14 seconds.
So there's a 5 minute timeout that causes the task to report "Success"? You'd think a timeout there would at least report a failure.
But if the results from Get-Task are wrong, Wait-Task should still just block until the task is completed. So it feels like there's a timeout that's shared between Get-Task and Wait-Task.
I'm seeing some old bug reports about a shared HTTP handler under PowerShell in Linux. I wonder if that could be relevant.
