Details
-
Bug
-
Resolution: Won't Fix
-
Major
-
1.8.1
-
Security Level: Public
-
None
-
14 Node cluster, 1 Bucket, 1024 vbuckets, Single Replica
Description
Seeing a large variance in memory stats/fragmentation from the nodes in the following setup.
1.Setup a 14 node cluster
2.Load 32M(58G) items into the cluster, in DGM. Resident ratios - (80% active and 59% replica) with
-key-size(0.5-1k)
- set=1
3.Continue mutations with larger key-sizes on existing items
-key-size(1k-1.3k) - set(0.15 to 0.2)
4.Rebalance -Out/Reboot of few nodes.
5.Continue mutations with larger key-sizes on existing items
-key-size(1k-1.3k)
- set(0.15 to 0.2)
6. Collecting memory stats to see any unexpected behaviours.
The following setup has been running for a few hours(6-7 hours).
Attaching the diags from the servers.