Details
-
Bug
-
Resolution: Won't Fix
-
Critical
-
2.2.0
-
Security Level: Public
-
None
-
2.2.0-764-rel, Centos
Description
- Setup a unixdcr from cluster1 to cluster2.
- Load data on source cluster @ a uniform rate ~ 20k ops/sec. (80:20 sets, gets), dataset includes deletes.
- Data is replicated at a rate of 5K ops/sec, but this is not even, it shows burst of activity @ 5k and then drops to half and so on.
The XDC replication rate should not fluctuate this high and should be more uniform.
Attaching logs.
Adding screenshots
- The compaction was switched off , to eliminate any compaction interference w/ this behavior.
Attachments
Issue Links
- relates to
-
MB-9706 Sudden spikes in XDCR replication queue
- Closed