centos-8-scenario002-standalone failing tempest - SSH timed out

https://review.opendev.org/762865 added to the skiplist while debug is in progress

I'm digging here but can't find something useful yet.

rlandy merged an addition to the skip at https://review.opendev.org/#/c/763078/ Add periodic sc002 tests to skip for bug 1904223

Main issue I've seen thus far is the one from the description, i.e. compute-log has

        * https://80137ce53930819135d8-42d904af0faa486c8226703976d821a0.ssl.cf2.rackcdn.com/762497/4/check/tripleo-ci-centos-8-scenario002-standalone/175ce9a/logs/undercloud/var/log/containers/nova/nova-compute.log
        * 2020-11-12 20:53:08.963 7 WARNING os_brick.encryptors.luks [req-02224c40-831a-4539-bc17-229de4975b5e 724335db83504df59d7bc8cbd4461b80 17940204a4f64c4dab4937478c3585eb - default default] isLuks exited abnormally (status 1): : oslo_concurrency.processutils.ProcessExecutionError: Unexpected error while running command.

and it seems to line up with the tempest fail time

ft1.2: barbican_tempest_plugin.tests.scenario.test_volume_encryption.VolumeEncryptionTest.test_encrypted_cinder_volumes_luks[compute,id-89165fb4-5534-4b9d-8429-97ccffb8f86f,image,volume]testtools.testresult.real._StringException: pythonlogging:'': {{{
2020-11-12 20:52:44,912 253297 WARNING [barbican_tempest_plugin.tests.scenario.barbican_manager] Starting Tempest 25.0.0 release, CONF.scenario.img_file need a full path for the image. CONF.scenario.img_dir was deprecated and will be removed in the next release. Till Tempest 25.0.0, old behavior is maintained and keep working but starting Tempest 26.0.0, you need to specify the full path in CONF.scenario.img_file config option.

since we are now skipping this in scen2 on check/gate and periodics it will be difficult to debug

So to make debug easier I posted a revert at https://review.opendev.org/763359 (includes https://review.opendev.org/762865 and https://review.opendev.org/#/c/763078/) and testing with https://review.rdoproject.org/r/31154
We can use this to run on demand with the test enabled or to test a fix with added depends-on in r/31154

Confirmed we are still seeing that in the test https://review.rdoproject.org/r/31154

        * https://logserver.rdoproject.org/54/31154/1/check/tripleo-ci-centos-8-scenario002-standalone/665b8bb/logs/undercloud/var/log/tempest/stestr_results.html.gz
        * 2020-11-19 13:13:14,598 321649 ERROR [barbican_tempest_plugin.tests.scenario.manager] (VolumeEncryptionTest:test_encrypted_cinder_volumes_luks) Initializing SSH connection to 192.168.24.112 failed. Error: Connection to the 192.168.24.112 via SSH timed out.
        * 2020-11-19 13:13:14.598 321649 ERROR barbican_tempest_plugin.tests.scenario.manager paramiko.ssh_exception.AuthenticationException: Authentication failed.

so you can use https://review.rdoproject.org/r/31154 (+depends-on https://review.opendev.org/#/c/763359/) for debug and re-run when we think we have a fix

Looking at logs in failed job it seems for me that this is the same issue like in https://bugs.launchpad.net/tripleo/+bug/1906769
Can we mark it as duplicate of 1906769?

@Slawek let's do that once we confirm? it might well be duplicate but we won't know for sure until we see if whatever fixes the other bug also closes this one?

no scenario002 in skiplist - closing this out