Persistent REMOVED nodes in mod_cluster-manager

[sbhargi]

Hi,

We are having a setup (apache/mod_cluster <-> tomcat) and recently we are seeing a lot of REMOVED nodes are still present in the mod_cluster-manager. These entries used to disappear in the past, but recently they are hanging around for sometime. I believe this also causes the apache to consume more CPU. This is our development environment and many deployment/restart of applciations happen throught the day.

We are using mod_cluster 1.2.0.FINAL with apache

Entry in the mod_cluster-manager page

node: [392:0],Balancer: xxx-ApplicationRequestRouter,JVMRoute: REMOVED,LBGroup: [],Host: xxx.xx.xxx.xxx,Port: 46718,Type: ajp,flushpackets: 0,flushwait: 600,ping: 600,smax: 31,ttl: 60,timeout: 0

node: [393:0],Balancer: xxx-ApplicationRequestRouter,JVMRoute: REMOVED,LBGroup: [],Host: xxx.xx.xxx.xxx,Port: 54842,Type: ajp,flushpackets: 0,flushwait: 600,ping: 600,smax: 31,ttl: 60,timeout: 0

node: [394:0],Balancer: xxx-ApplicationRequestRouter,JVMRoute: REMOVED,LBGroup: [],Host: xxx.xx.xxx.xxx,Port: 48272,Type: ajp,flushpackets: 0,flushwait: 600,ping: 600,smax: 31,ttl: 60,timeout: 0

node: [395:0],Balancer: xxx-ApplicationRequestRouter,JVMRoute: REMOVED,LBGroup: [],Host: xxx.xx.xxx.xxx,Port: 48272,Type: ajp,flushpackets: 0,flushwait: 600,ping: 600,smax: 31,ttl: 60,timeout: 0

I've even tried to shutdown apache on one leg for 10 mins but still no luck with it. Can someone shed some light on this?

[jfclere]

usually that happens when some nodes have the same JVMRoute. There should be some error messsages in the logs.

[sbhargi]

I'll check on this one and get back to you. A quick search showed only messages about workers in error state for some application but nothing about duplicate JVMRoute. I'm not sure if this woud be any concern, what I noticed is that I see a lot of duplicate contexts in the mod_cluster dump output. Could this be a reason for the current state?

[jfclere]

duplicate contexts? What is the output?