Description of problem: In the case where a slow or new member receives a snapshot from leader and attempts to applySnapshot meanwhile cluster runs compaction a deadlock can occur.
Version-Release number of selected component (if applicable): etcd 3.x
How reproducible: conditions would need to exist which would be relatively low probability. But the result would be very hard for an already struggling cluster to overcome.
Steps to Reproduce:
2.perform a lot of writes to cluster
3.unpartition member right before compaction
Actual results: cluster can experience deadlock and hang causing possible disruption.
Expected results: cluster remains available
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 (OpenShift Container Platform 4.6 GA Images), and where to find the updated
files, follow the link below.
If the solution does not work for you, open a new bug report.