Description
As of now xdcr does target topology change detection once every 10 seconds, and checks if target bucket is valid once every 15 seconds. This can result in a lot of connections to target, because:
- the number of connections are proportional to the number of source nodes.
- all connections are made to one particular node on target, which is the host name in remote cluster ref
The problem will get even worse when there are multiple replications running.
The fix is:
1.
Attachments
Issue Links
- blocks
-
MB-22188 4.6.2 Minor Release
- Closed