Details
-
Bug
-
Resolution: Cannot Reproduce
-
Blocker
-
2.0
-
Security Level: Public
-
None
-
Cluster Details:
Ubuntu 12.04 LTS ec2 clusters
RAM: 15GB
Build 1916: http://builds.hq.northscale.net/latestbuilds/couchbase-server-enterprise_x86_64_2.0.0-1916-rel.deb.manifest.xml
Live clusters:
C1: http://ec2-184-169-250-146.us-west-1.compute.amazonaws.com:8091/
C2: http://ec2-54-251-5-97.ap-southeast-1.compute.amazonaws.com:8091/Cluster Details: Ubuntu 12.04 LTS ec2 clusters RAM: 15GB Build 1916: http://builds.hq.northscale.net/latestbuilds/couchbase-server-enterprise_x86_64_2.0.0-1916-rel.deb.manifest.xml Live clusters: C1: http://ec2-184-169-250-146.us-west-1.compute.amazonaws.com:8091/ C2: http://ec2-54-251-5-97.ap-southeast-1.compute.amazonaws.com:8091/
Description
Setup:
- C1 : C2 (each 10 nodes)
- Bidirectional replication for biXDCR_bucket between C1 <--> C2
- Unidirectional replication for uniXDCR_src on C1 to uniXDCR_dest on C2.
- Front end loads and replication set to run overnight on 10-30-2012.
Currently:
- Memcached Crashes seen on 5 nodes on each cluster (no cores seen though) causing the nodes to enter "PEND" state.
- On the 5 nodes on the west cluster (C1), that have gone down: Outgoing XDCR Operations section is missing.
- However on the 5 nodes on the southeast cluster (C2), that seem to have gone down, the section isn't missing.
Now as the stats are supposed to be an aggregate over all nodes in the cluster, why is the OUTGOING XDCR OPERATIONS tab missing on just the select nodes that are down on C1.
