dansmith | gmann: yep, sorry, didn't see this | 00:06 |
---|---|---|
gmann | thanks | 00:32 |
opendevreview | Merged openstack/grenade stable/zed: Fix setting the tempest virtual env constraints env var https://review.opendev.org/c/openstack/grenade/+/871947 | 01:28 |
gmann | whoami-rajat: it seems like volume_type deletion is failing consistently in glance-multistore-cinder-import job https://review.opendev.org/c/openstack/tempest/+/875372/4/tempest/api/compute/base.py#567 | 02:22 |
gmann | order of cleanup - volume first and then volume_type seems fine in tests but I have not looked into the logs to know why volume_type is still in use | 02:23 |
gmann | I am now in internal meeting and may have another meeting. If I get time I will check today otherwise will check tomorrow | 02:23 |
*** jpena|off is now known as jpena | 08:39 | |
whoami-rajat | gmann, thanks for looking into it, I've rechecked the patch to look for longs (not sure why the logs expired in a day) | 10:00 |
*** elodilles is now known as elodilles_afk | 13:20 | |
*** elodilles_afk is now known as elodilles | 14:58 | |
gmann | whoami-rajat: thanks.will check you comment | 15:58 |
whoami-rajat | gmann, I've added a comment explaining the issue and another comment regarding the solutions for it, in summary the image_volume_cache_enabled is the culprint, I'm not sure why devstack enables it by default, it does provide significant improvement in creating bootable volumes but isn't necessary for any test to run successfully | 16:03 |
whoami-rajat | s/culprint/culprit | 16:03 |
*** jpena is now known as jpena|off | 17:33 | |
gmann | whoami-rajat: thanks for the detailed debugging in this. I think 2nd solution seems more feasible which is close to what actual users can try | 22:05 |
gmann | I think we should not make test depends on CINDER_IMG_CACHE_ENABLED setting as that can enabled in local/production run and then tempest will fail | 22:07 |
Generated by irclog2html.py 2.17.3 by Marius Gedminas - find it at https://mg.pov.lt/irclog2html/!