Details
-
Bug
-
Resolution: Fixed
-
Major
-
2.0-beta-2
-
None
-
Security Level: Public
-
None
-
CentOS release 5.7 (Final)
Couchbase Server 2.0 beta through build 1950
Description
A testcase I'm running starts XDCR between 2 clusters (each 1 node, 4 core, 4g).
It then runs a mix of create/update/delete operations (items=10000).
At some point, the rate of XDCR replication drops, and it takes a long time (varying, but over still over 20 minutes in build 1950) to finish replicating all the changes. Even though the cpu/disk are idle on both nodes the entire time.