[jboss-jira] [JBoss JIRA] (WFCORE-1528) EAP 7 with secured management become unavailable after ~8 days

[Brian Stansberry (JIRA)]

    [ https://issues.jboss.org/browse/WFCORE-1528?page=com.atlassian.jira.plugin.system.issuetabpanels:comment-tabpanel&focusedCommentId=13201728#comment-13201728 ] 

Brian Stansberry commented on WFCORE-1528:
------------------------------------------

I believe WFCORE-1428 is cause of WFCORE-1528, as it brings a non-fixed-size thread pool into Undertow's request handling, and Undertow's internal caching mechanisms are not designed for use with a non-fixed pool. As the pool churns threads, pooled ByteBuffers are not getting freed.

It's also better to use a fixed pool for the WFCORE-1428 work as this call path is for end-user requests and we want to limit concurrency of those a la what we do for native clients at https://github.com/wildfly/wildfly-core/blob/master/controller/src/main/java/org/jboss/as/controller/remote/AbstractModelControllerOperationHandlerFactoryService.java#L98.

> EAP 7 with secured management become unavailable after ~8 days
> --------------------------------------------------------------
>
>                 Key: WFCORE-1528
>                 URL: https://issues.jboss.org/browse/WFCORE-1528
>             Project: WildFly Core
>          Issue Type: Bug
>          Components: Domain Management, Security
>    Affects Versions: 2.1.0.Final
>         Environment: EAP 7.0.0.CR01 2-way ssl over remote https communication
>            Reporter: Simeon Pinder
>            Assignee: Brian Stansberry
>            Priority: Critical
>             Fix For: 3.0.0.Alpha1
>
>
> Appears there may be an EAP 7.0.0.CR01 2-way ssl issue in domain mode.  
> Issue discovered by JON QE after testing for 8 days with SSL.
> JON QE did not see the same issue when running identical test against EAP 6. 
> Restarting the remote client(JON Agent) did not repair/fix the issue, and only was repaired after EAP instance restarted.
> See BZ https://bugzilla.redhat.com/show_bug.cgi?id=1330180#c0 for more details.
> JON team would like some evaluation done to determine:
> i)if this is EAP 7 issue or
> ii)somehow just a JON issue
> iii)suggestions on what to do when we try to reproduce properly or ways to shorten reproduce cycle.
> This issue could affect GA release schedules.



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