Comment 31 for bug 1696125

While debugging the LiveMigrationTest.test_iscsi_volume tempest test, I am seeing this in my local env as one of the causes of intermittent failures with this test. I can get this to fail maybe 2-3x a day. This is what I run to provoke it:

$ tempest run --regex test_live_migration --concurrency=4

I haven't seen it fail on a single threaded test, but it's hard to say since it fails so infrequently.

The failure in n-cpu.log is:

2017-09-20 14:01:50.192 ^[[01;31mERROR oslo_messaging.rpc.server [^[[01;36mreq-6db9664a-4c6e-4d47-bffc-224b45b34eb0 ^[[00;36mtempest-LiveMigrationTest-1489892041 tempest-LiveMigrationTest-1489892041^[[01;31m] ^[[01;35m^[[01;31mException during message handling^[[00m: DeviceDetachFailed: Device detach failed for vdb: Unable to detach from guest transient domain.

The tempest error is:

    tempest.lib.exceptions.TimeoutException: Request timed out
    Details: volume 3194b6d9-cb70-4e67-8c28-ebf9ea094297 failed to reach ['available'] status (current in-use) within the required time (196 s).

I am adding more debug code to help debug this.