Cisco Bug: CSCvt39916 - Zookeeper out-of-sync situation during quick keepalived flaps
Oct 07, 2020
- Cisco Ultra Services Framework
Known Affected Releases
Symptom: Symptoms are varied. One possible symptom is the removal of functioning VMs from the system leading to potential service outages. On the EMs, examination of the zookeeper logs under /var/log/em/zookeeper show a continuous stream of warning messages complaining that the accepted epoch is greater than the leader epoch. Here is an example: 2020-02-18 08:43:26,900 [myid:51] - WARN [QuorumPeer[myid=51]/192.168.46.51:2181:Follower@87] - Exception when following the leader java.io.IOException: Leaders epoch, 6 is less than accepted epoch, 9 at org.apache.zookeeper.server.quorum.Learner.registerWithLeader(Learner.java:293) at org.apache.zookeeper.server.quorum.Follower.followLeader(Follower.java:70) at org.apache.zookeeper.server.quorum.QuorumPeer.run(QuorumPeer.java:846) Conditions: 2 EM deployments where connectivity is impacted simultaneously on management and orchestration networks between the 2 EM instances.
Bug details contain sensitive information and therefore require a Cisco.com account to be viewed.
Bug Details Include
- Full Description (including symptoms, conditions and workarounds)
- Known Fixed Releases
- Related Community Discussions
- Number of Related Support Cases