If I cluster my membase instances, sa x1 and x2. When I connect to x1 from my java spymemcached client, will x1 distribute the load to x2? or do i have to use the smart java client to do this?
The smart client is the best approach, as it's just a wrapper around spymemcached anyway. Dustin Sallings, the author of spymemcached, is one of NorthScale's founders and our Chief Architect. Several of us contribute to and work on spy.
The smart client will automatically adjust to whatever nodes you add to the cluster. You can use spymemcached directly as well, but then you'll have to reconfigure it to add nodes whenever you cluster changes. The advantage of the smart client is that it maintains a connection to the cluster and will adjust nearly immediately.
Ahh Okay Interesting, so If i understand correctly there are two options.
1. Make a cluster in the admin console, then use spymemcached and define the list of all the clusters . Here the only reason for the admin console is for the analytics of the overall cluster and buckets
2. Make the cluster and use the smart client which auto adjusts the incoming connections to the connections.
It's more 'define the list of all nodes in the cluster' if using spymemcached directly, yes.
This is not an option for everyone, since there are only smart clients for C# and Java.
Correct. This is an option for you since you're using spymemcached already.
Cool! Thanks, you guys have an amazing product. Keep up the great work!!
© 2013 COUCHBASE All rights reserved.