Membase cluster server failure
Wed, 12/15/2010 - 12:35
Hi,
I have a question regarding the expected behavior of a membase cluster when a server fail. Do I have to manually "fail over" the server in the cluster server list?
I'm asking because one of our server went down and the memcached clients kept getting errors until I manually fail over the server. Then the client were working correctly again.
I would have expected membase to automatically remove a down node from the cluster.
Hi there!
You are correct, the software currently requires administrative action to "fail over" a server. The process can be scripted, but we made a concious decision to not provide automatic failover. It's fairly easy to imagine a scenario where it would wrong for the cluster to automatically fail over a node (network partition for example). To read more, check out this wiki page I wrote: http://wiki.membase.org/display/membase/Failover+with+Membase
Let me know if you have any other questions, thanks!
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!