Details
-
Bug
-
Resolution: Fixed
-
Major
-
7.2.1
-
OS - Debian 10
Initial version - 7.2.1-5861 community edition
Upgrade version - 7.2.1-5861 enterprise edition
-
Untriaged
-
Linux x86_64
-
-
0
-
Unknown
Description
Steps:
- Initialise a 3 node cluster, each node in 7.2.1-5861 community version and running only kv service.
- Create a couchstore bucket and load 2 million items into it.
- Install the enterprise edition of 7.2.1-5861 on a spare node.
- Upgrade one of the nodes through the swap rebalance method. The node that was swapped out now becomes the spare node for the upgrade of the next node in the cluster.
- The first swap rebalance went fine.
- After the upgrade of the first node, a few writes were performed.
- The same steps were carried out for the upgrade of the second community node in the cluster, but it was observed that the rebalance started but hung shortly after and there was no progress for the next 20 minutes.
- Rebalance was re-started and this time, it was observed that the rebalance went through fine.
Cb-collect log links have been attached.
Attachments
Issue Links
- duplicates
-
MB-39344 [checkpoint] Allocation in replica checkpoints can push the system to hard OOM
- Closed