Red Hat Bugzilla – Bug 1535766
VIFs not always detached from ironic nodes after stack delete
Last modified: 2018-06-27 09:43:13 EDT
Description of problem: Sometimes when a baremetal instance is terminated, some VIFs are not detached from the node. This can lead to the node becoming unusable, with subsequent attempts to provision it fail during VIF attachment due to there being insufficient free ironic ports to attach the VIF to. Version-Release number of selected component (if applicable): How reproducible: after multiple attemt of stack create/delete to test multiple scenarios it faild with VIF attachment Steps to Reproduce: 1. 2. 3. Actual results: Expected results: instance should boot successfully Additional info: Looks similar to upstream bug https://bugs.launchpad.net/nova/+bug/1733861
This seems related to https://bugs.launchpad.net/ironic/+bug/1743652 which has a patch in progress to update the upstream conductor behavor.
Changing state on this as the fix in https://review.openstack.org/#/c/534441/ was reverted. Upstream patch https://review.openstack.org/#/c/556748/ will probably be fix after backport.
Per DFG discussion on 5/11 - this problem is being seen more often as the cause of failed deployments and as such we are considering it a blocker. A patch has landed in stable/queens that will be backported.
Installed latest osp13 5.18.01 We don't have same exact setup but make every effort to try and re-produce. In RHV env since it was available ... 3 controller, 1 compute deployment. 1. delete deployment 2. check vif exist on nodes (openstack baremetal node vif list <uuid>) reported bug showing this has been seen leftover from previous delete 3. re-deploy deployment 4. check vif exisit 5. delete deployment 6. check vif exist repeated above a two times ^^^ -- In another virtual environment (3 controller, 1 compute, 3 ceph nodes) followed steps 1..7 above repeated this procedure two times. In total (4) attempts to reproduce the issue. It appears now we can verify. Please re-open or file new bz if seen again.
Since the problem described in this bug report should be resolved in a recent advisory, it has been closed with a resolution of ERRATA. For information on the advisory, and where to find the updated files, follow the link below. If the solution does not work for you, open a new bug report. https://access.redhat.com/errata/RHEA-2018:2086