Description of problem:
Client is in the middle of an OSP13 z3 to z11.
They faced other (different) issues while live migrating instances from z3 compute to z11 but this time its a bit different.

The live migration was triggered on October 29 at 9-10am (logs will come in the next private comment).

It stayed in "migrating" until I reset the state of the instance.

But before that the instance was in paused state on the destination compute.
I don't recall exactly the state on the source but it might have been gone already.

After resetting the state, I cleaned up the "bad" reserved and dest compute node entries in cinder.volume_attachment. Did a hard reboot and was able to live migrated the instance without an issue on that same destination compute.

I have logs about that period. More details to come in next comment.

If you could help identify what happened it would be awesome.

Version-Release number of selected component (if applicable):
OpenStack platform 13 z3 and z11 compute nodes.

How reproducible:
Happened once.

Steps to Reproduce:
1. N/A
2.
3.

Actual results:
Instance live migration was stuck endlessly

Expected results:
instance live migration works fine

Additional info:
We have sosreport and logs on supportshell