Loading...

XML

Word

Printable

Details

Type: Bug
Resolution: Done
Priority: Major
Fix Version/s: EAP_EWP 5.1.1
Affects Version/s: EAP 5.0.1, EAP_EWP 5.1.0
Component/s: mod_cluster
Labels:
None
Environment:

RHEL 4, EWS 1

Release Note Text:

Hide
Because of problems with the <code>pause()</code> method of the connector, a race condition could occur when applications were undeployed. Sessions that were forwarded to the application server by mod_cluster before receiving an undeploy notification could result in an <code>error 503 - This application is not currently available</code>. The method has been updated, addressing these issues.

Show
Because of problems with the <code>pause()</code> method of the connector, a race condition could occur when applications were undeployed. Sessions that were forwarded to the application server by mod_cluster before receiving an undeploy notification could result in an <code>error 503 - This application is not currently available</code>. The method has been updated, addressing these issues.
Release Note Status:
Documented as Resolved Issue

Description

Simple scenario - sessions are active while undeploy of a test war takes place.

It looks like this is a race condition, the STOP-APP doesn't get to the server on time, it forwards requests. Which seems fine to me, however I would expect that the server needs to retry if this happened. Is there a setting for this?

If I understand correctly how mod_jk handles this, 503 is setup as one of the error codes and there is a retry setup on a different node where it gets through.

Response
========

[ org.jboss.smartfrog.loaddriver.RequestProcessingException: Invalid response code: 503 Content:]

HTTP Status 503 - This application is not currently available

type Status report

message This application is not currently available

description The requested service (This application is not currently available) is not currently available.
JBoss Web/2.1.3

JBAS logging
=========
2010/01/19 12:08:46:771 EST [INFO ][RMI TCP Connection(11)-10.16.88.187] HOST perf09.qa.atl2.redhat.com:rootProcess:sf - Undeploying application (/tmp/hudson/jboss-eap-5.0/jboss-as/server/production/deploy/SessionTest.war) on perf03 (on component starting).
[JBoss] 12:08:47,010 INFO [TomcatDeployment] undeploy, ctxPath=/st
[JBoss] 12:08:47,097 INFO [SessionTestServlet] Invalidating session - 3rDyHbo-IYQLwhJ5bkvTfg__.node3.
[JBoss] 12:08:47,097 INFO [SessionTestServlet] Invalidating session - MECYJTPe+TtjS7YauqTvfQ__.node3.
[JBoss] 12:08:47,099 INFO [SessionTestServlet] Invalidating session - +Sk1uQMpOAhUCC+nRp8Opg__.node3.
[JBoss] 12:08:47,328 INFO [RPCManagerImpl] Disconnecting and closing the Channel
[JBoss] 12:08:47,329 WARN [ComponentRegistry] Received a remote call but the cache is not in STARTED state - ignoring call.
[JBoss] 12:08:47,375 INFO [RPCManagerImpl] Stopping the RpcDispatcher