-
Bug
-
Resolution: Unresolved
-
Major
-
None
-
None
-
Quality / Stability / Reliability
-
0.42
-
False
-
-
False
-
None
-
-
Important
-
None
Description of problem:
Based on initial testing, localnet live migration downtime is much worse when live migration is performed between clusters, compare to live migration within a cluster. It should be comparable.
Version-Release number of selected component (if applicable):
4.19
How reproducible:
Tried only once in https://redhat-internal.slack.com/archives/C019X3PEF2B/p1750750969236329?thread_ts=1750750738.384119&cid=C019X3PEF2B
Steps to Reproduce:
1. Measure downtime during live-migration between clusters, use a quiet small VM, and colocated workers
Actual results:
Downtime of 12 seconds
Expected results:
Comparable downtime to live migration within a cluster, i.e. sub 1 second
Additional info:
We may want to run the same test with linux bridge CNI. Bridge CNI does not involve any cross-component synchronization and so it should perform the same way within and between clusters.