[jboss-jira] [JBoss JIRA] (WFLY-5310) WARN HttpContextLifecycle guard leak detected. The Servlet container is not fully compliant during application deployment

[Stuart Douglas (JIRA)]

    [ https://issues.jboss.org/browse/WFLY-5310?page=com.atlassian.jira.plugin.system.issuetabpanels:comment-tabpanel&focusedCommentId=13108485#comment-13108485 ] 

Stuart Douglas commented on WFLY-5310:
--------------------------------------

This is a problematic one. Basically when Undertow is told to shutdown it will destroy the CDI listener, this means that if there are any active requests that finish after this has happened the requestDestroyed listener will not be invoked. 

The simple solution to this is to use graceful shutdown, so the requests are allowed to finish before the listener is destroyed. I don't think there is much that can be done about this otherwise, this problem is just one of many that can occur if the deployment is stopped while requests are active.

> WARN HttpContextLifecycle guard leak detected. The Servlet container is not fully compliant during application deployment
> -------------------------------------------------------------------------------------------------------------------------
>
>                 Key: WFLY-5310
>                 URL: https://issues.jboss.org/browse/WFLY-5310
>             Project: WildFly
>          Issue Type: Bug
>          Components: CDI / Weld
>    Affects Versions: 10.0.0.Beta1
>            Reporter: Michal Vinkler
>            Assignee: Stuart Douglas
>            Priority: Minor
>
> Setup: 4-node cluster, one node at time undeploys application (clusterbench-ee7.ear), while standalone clients keep calling the application.
> This error was seen during clustering failover testing in the following scenario:
> ejb-ejbservlet-repl-async (failover of HTTP traffic accessing a servlet that locally invokes a stateful clustered EJB, with asynchronously replicated cache)
> During application re-deployment (after previous undeployment) on any of the nodes, this WARN message is being logged approx. 60 times - stacktrace:
> {code}
> [JBossINF] [0m[0m19:07:15,617 INFO  [org.jboss.as.server] (management-handler-thread - 2) WFLYSRV0010: Deployed "clusterbench-ee7.ear" (runtime-name : "clusterbench-ee7.ear")
> 2015/08/28 19:07:15:621 EDT [DEBUG][RMI TCP Connection(16)-10.16.90.52] HOST perf17.mw.lab.eng.bos.redhat.com:rootProcess:c - Got result: {"outcome" : "success"}
> [JBossINF] [0m[0m19:07:20,563 INFO  [org.jboss.as.server.deployment.scanner] (DeploymentScanner-threads - 1) WFLYDS0033: Deployment clusterbench-ee7.ear was previously undeployed by this scanner but has been redeployed by another management tool. Marker file /tmp/hudson/jboss-eap-7.0/standalone/deployments/clusterbench-ee7.ear.undeployed is being removed to record this fact.
> [JBossINF] [0m[33m19:08:01,353 WARN  [org.jboss.weld.Servlet] (default task-2) WELD-000714: HttpContextLifecycle guard leak detected. The Servlet container is not fully compliant. The value was 1
> {code}
> Server log:
> http://jenkins.mw.lab.eng.bos.redhat.com/hudson/job/eap-7x-failover-ejb-ejbservlet-undeploy-repl-async/6/console-perf18/



--
This message was sent by Atlassian JIRA
(v6.4.11#64026)