statistics seem to be wrong
Tue, 04/12/2011 - 07:19
Hi all,
We recently had an event where we needed to failover and then re-add a membase cluster node. After doing so, the statistics on our one bucket seem to be wrong. The disk write queue is quite large (1.4M) whereas before it would stay at zero. Other stats are increased as well and a flush does not bring the unique item count to zero as it did before. Otherwise, the cluster seems to be functioning fine.
Any ideas? Is this an issue of the stats being wrong? If so, anyway to reset them.
Thanks.
Hey Brandon...you can reset stats with: /opt/membase/bin/ep_engine/management/stats :11210 reset [bucket_name] [bucket_password]
The disk write queue is likely due to the rebalancing and replication rematerializing...if it's not going down, you should check the disk space available and permissions to make sure that it is able to write to disk.
Perry
Forum support is great for free but sometimes you need a guaranteed response time and dedicated resources for your questions or issues.
Consider purchasing enterprise-level support from Couchbase: http://www.couchbase.com/products-and-services/overview
Call or email "sales -at- couchbase-dot- com" today!