connection fail over

Brian Roach roach at
Tue Mar 27 11:29:06 EDT 2012


By using a cluster client (HTTPClusterClient or PBClusterClient) you would get not only fail-over on an operation failure but also round-robin balancing of operations across your nodes. 

When you create either a HTTPClusterClient or a PBClusterClient you provide a list of Raik nodes. When performing an operation there's a Retrier object that is used by the Bucket object. The default Retrier tries an operation 3 times before reporting a failure. On each try, a different node is selected from the list of Riak nodes (this same mechanism also gives you load balancing across the nodes in your list in a round-robin fashion). 
The only deficiency here is that a node is never failed completely out of the list or deprioritized in any way. We do have plans to add a pluggable load balancing feature at some point that would allow for the end user to implement whatever scheme they would like.    
I've recently started work on a Riak Java client "Cookbook" and have an example of how to create a cluster client here:

Brian Roach

On Mar 26, 2012, at 11:56 AM, Zheng Zhu wrote:

> Hello, I am new to Riak and want to know if Riak java client has the node connection feature to automatically fail over to another node after timed out for current node. If so, can you share links to doc, java doc or sample code?
> Thanks
> Zheng
> The information contained in this e-mail is confidential and/or proprietary to the 41st Parameter. The information transmitted herewith is intended only for use by the individual or entity to which it is addressed. If you are not the intended recipient, you should not copy, distribute, disclose or use the information it contains, please e-mail the sender immediately and delete this message from your system.
> _______________________________________________
> riak-users mailing list
> riak-users at

-------------- next part --------------
An HTML attachment was scrubbed...
URL: <>

More information about the riak-users mailing list