Web applications failing if a node of the cluster is down?!
Hi
We use Memcache Server 1.0.3 on our production environment. Today one of our two cluster nodes was shut down, then there were some IIS resets. This lead to many exceptions of clients yelling about not to be able to connect. As soon as the cluster node was up again, everything went smooth.
We use windows 2008 R2 server, enyim client (from the date of memcached server 1.0.3). Since this is production we do not follow every update. My thinking is, that the client takes the first server url in the config entry and connects, even if it is unavailable. This brings the failure. One would assume that this is not possible, since the client shoult register that it can't connect and try the next server? This happens only on the creation of the client object. As soon as this is established (or for objects established before the downtime), there is no problem since the other node gets the downtime and reacts accordingly.
Any ideas on this broad shot? I will be able to supply some more details as reply to questions. Maybe this is more to seek in the client?
Thanks for any help,
Charmaster
Charmaster, thanks for posting here, sorry that you ran into some issues.
There have been a number of improvements both in the client and in the server to handle these failure cases. I think it would be well worth your while to upgrade the Enyim client as well as upgrade to the new Membase Server (using a memcached bucket if that's all you want).
It would certainly be worth testing out this exact scenario with your application to make sure that everything behaves properly. If it doesn't, please let us know so that we can diagnose and resolve it.
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!