Hi I was wondering if anyone has seen this problem or whether they know a fix. So environment
2 x W2k8r2 NLB IIS servers IIS 7.5 + AJP (1.2.36) connected to 2 x 2 Jboss Server.. (2 cluster of 2 environments) I have a separate worker pool setup for each environment (2 vhost each with their own ajpconnector and config directory) Say Prd Uat Previously we had the worker poll process limit set to 2 and allowed over lapping recycles. We ran some tests where we created 1000 connections from a test machine to IIS connecting on urls that were being services by the AJP connector (say /ajp/ ) The LB ajp workers were setup for 1000 connection on each worker to each jboss node. What we saw was an issue, connections were getting stuck. These test connections use keep alive, so not a simple open close request! What we say was 2 windows processes get started and we watched the number of threads being created ... This seemed to be limited to 264 sometimes it would hit 265... and thus out testing would fail as 2 x 265 would only give us 530 connections. So out simple solution is to bump up the number of worker processes, but I am trying to find what is imposing the 265 thread limit ! Thanks Alex --------------------------------------------------------------------- To unsubscribe, e-mail: users-unsubscr...@tomcat.apache.org For additional commands, e-mail: users-h...@tomcat.apache.org
