Conversation
|
Dear Sidney An, Thank you for your interest in StarlingX and submitting your pull request. The StarlingX repos on GitHub are mirrored from the original ones on opendev.org and therefore we do not accept code contributions here in the form of PRs. You can find the original repositories here: https://opendev.org/starlingx You can find information about how to contribute to the StarlingX project here: https://docs.starlingx.io/contributor/index.html You can reach out to the community on the starlingx-discuss mailing list: http://lists.starlingx.io/cgi-bin/mailman/listinfo/starlingx-discuss You can also join the weekly community calls: https://wiki.openstack.org/wiki/Starlingx/Meetings#7am_Pacific_-_Technical_Steering_Committee_.26_Community_Call Thanks and Best Regards, |
| run_immediately=True) | ||
| # @periodic_task.periodic_task(spacing=CONF.conductor.audit_interval, | ||
| # run_immediately=True) | ||
| def _k8s_application_audit(self, context): |
During node replacement, topology can change for identical CPU models. An example, CPU #1 has core 2 disabled in HW, but CPU #2 has core 3 disabled. This will cause a divergence in the core ID enumeration posted by firmware. Even though technically the CPUs are functionally equivalent. To mitigate this, during restore, sysinv will now update core_ids to the new values, but only if the rest of the topology matches. TEST PLAN PASS: AIO-SX deploy * No regressions PASS: AIO-DX deploy * No regressions PASS: AIO-SX optimized B&R * No regressions PASS: AIO-SX optimized B&R on affected system * Before backup, change some cores to application-isolated * CPU configuration successfully restored Closes-Bug: 2120725 Change-Id: Ie35f2acb4faa9c5973fead31d58533b40d6aef73 Signed-off-by: Joshua Kraitberg <joshua.kraitberg@windriver.com>
3 participants
No description provided.