On Thu, 2007-11-15 at 04:31 -0800, James Brook wrote: > I have an application that is very dependent on HTTPClient for back-end > communication with another system, which is a farm of Web servers. The > infrastructure team have not supplied a load balancing/fail-over solution on > our internal network, so I am wondering about the feasibility of > implementing fail-over in the software. > > Does HTTPClient have any out-of-the-box support for this? I have seen the > 'HttpMethodRetryHandler' class. Is that a good place to attempt to implement > fail-over, or should I do it at a higher level? > > I think I saw a post discussing something like this on the dev list, but I > am not sure whether it has been implemented yet. I would very much > appreciate any suggestions or code snippets. > > -- > James
James, Reasonable fail-over strategies tend to be application / environment / context specific, therefore fail-over is not provided out of the box. If all you want is a trivial DNS round-robin, you can implement it using a custom socket factory: https://issues.apache.org/jira/browse/HTTPCLIENT-643 Oleg --------------------------------------------------------------------- To unsubscribe, e-mail: [EMAIL PROTECTED] For additional commands, e-mail: [EMAIL PROTECTED]
