Failed disk erasing

Often, when releasing many machines, several get stuck in “failed disk erasing”.
Releasing machines one at the time solves the problem.

How to debug this randomness? The state is often in “loading ephemeral”.

Could it be that the maas server is overloaded? Is there any benefit to having an extra rack server on the same network to offload? In this case, there is one pure rack server and one that is both rack and region.

i could only guess without more information.

can you have a look at this guide, purely for the list of data to collect? (not sure it’s time to file a bug, yet, but don’t lose your data once you gather it).

most important is going to be the version info on MAAS and the log files. if you can get log files while trying to simultaneously release multiple machines, and while serially releasing machines, that would help us the most.

thanks in advance.

@bjornrun, did you figure this one out?