Details
-
Bug
-
Resolution: Fixed
-
Critical
-
6.5.0
-
Centos 7.6
-
Untriaged
-
Centos 64-bit
-
Yes
-
KV-Engine Mad-Hatter GA, KV Sprint 2019-12
Description
Steps to reproduce
- Install Couchbase server 5.0.0 on 3 centos 7.6 servers and create cluster with services of each node (kv,index-kv,n1ql -kv,fts).
- Create default bucket and load data to default bucket.
- Graceful failover 1st node and upgrade node from 5.0.0 to 6.5.0-4908 (use rpm -U cbserver.rpm)
- Set full recovery and rebalance. Rebalance success as expected.
- Graceful failover 2nd node and upgrade node from 5.0.0 to 6.5.0-4908 (use rpm -U cbserver.rpm)
- Set full recovery and rebalance. Rebalance success as expected.
- Graceful failover 3rd node (last 5.0.0 one with ip 224) and upgrade node from 5.0.0 to 6.5.0-4908 (use rpm -U cbserver.rpm)
- Set full recovery and rebalance. Rebalance failed instantly with error " 'memcached' exited with status 134."
- Rebalance again, rebalance success
Attachments
Issue Links
- causes
-
MB-37204 [Jepsen] Couchbase crash during hard-reboot scenario due to HCS > persistedCompletedSeqno
- Closed
- is duplicated by
-
MB-37103 [System test]: Disk Checkpoint does not have an initialised HCS
- Closed
- relates to
-
MB-37188 Processing DCP_SEQNO_ACKNOWLEDGEMENT before the topology is setup
- Closed
-
MB-37187 [Test code] Expand cluster test suite for Upgrade tests
- Closed
- mentioned in
-
Page Loading...