Node Down, cannot failover
DPV3: I have a node down, but am unable to fail it over.
Because the Web Console reports that it Lost Connection...
My experience of the last couple of weeks is that if a node goes down, the Web Console becomes next to useful for resolving the issue. Sometimes I can click to Remove the node. The rebalance will then fail. So, I cancel the rebalance/remove. Then I try to Failover the node. Won't succeed.
Couchbase is certainly not at the "it just works" stage.
I'm in a 'deathmarch' development cycle, and don't have time to do the alpha testing this product clearly still needs.
So, I'm going back to the 'just works' solution of Northscale for Caching, and Sql Server for database behind the cached data, for cache misses.
I understand there are some big names using Couchbase Server (e.g., Draw Something) successfully. But, they maybe have a developer who can make a career out of managing Couchbase, or a team dedicated to keep it healthy. I don't have that capacity nor team, so have to drop back to the tried and true (we have another product that just works, running on NorthScale. That product is paying for the development of the new product).
Darn. I was so excited to use Couchbase with promises of Reliability with CouchDB behind it, and replication. But, the constant firefighting just to keep the cluster healthy is draining development cycles that should be used on the actual product.
May have to give Couchbase a try again once it's out of Developer Preview stage and more at the 'real product' stage.
(Sure, I could try CBS1.8, but at this point I don't want to 'try' anything more; I need to simplify away experiements and go with working tech.)