I feel there can be something done using the ratelimit module here: <meta http-equiv="content-type" content="text/html; charset=utf-8"><a href="http://www.opensips.org/html/docs/modules/1.6.x/ratelimit.html">http://www.opensips.org/html/docs/modules/1.6.x/ratelimit.html</a>, but I don't see that happening without a way to do a "select next" with the loadbalancer module. Currently, the implementation calculates load based on resources available, but is there a way to forcibly skip a resource and pick the "next best" resource, like a lb_do_next() function?<div>
<br></div><div>Thanks,</div><div><br></div><div>Bobby Smith<br><br><div class="gmail_quote">On Sat, Nov 27, 2010 at 3:25 PM, Bobby Smith <span dir="ltr"><<a href="mailto:bobby.smith@gmail.com">bobby.smith@gmail.com</a>></span> wrote:<br>
<blockquote class="gmail_quote" style="margin:0 0 0 .8ex;border-left:1px #ccc solid;padding-left:1ex;">Greetings,<div><br>Currently with our load balancing implementation, we're encountering a problem where when a server is added to the active pool during load, all calls immediately hit that server until all the resources are reached. While in a perfect situation (restarting call load) this is not a problem, in a situation where we're under call load with many servers and need to dynamically add more servers, this causes a problem as the resource added can't handle the sudden burst of calls per second. This is also causing a problem when a resource is removed from the pool for troubleshooting purposes and needs to be added back without a restart of opensips.</div>
<div><br>Is there any way from a scripting perspective to throttle how this works? If not, I'd like to request it as a feature (something like maximum CPS per resource or something). Also, resetting the load balancing metrics is not an option; we need to be able to keep an accurate track of resources load balanced to each server.</div>
<div><br></div><div>Thanks,</div><div><br></div><div>BobbyS</div>
</blockquote></div><br></div>