[jboss-jira] [JBoss JIRA] (WFWIP-119) [Artemis 2.x upgrade] Lost message in replicated HA scenario with MDB

[Erich Duda (JIRA)]

    [ https://issues.jboss.org/browse/WFWIP-119?page=com.atlassian.jira.plugin.system.issuetabpanels:comment-tabpanel&focusedCommentId=13623018#comment-13623018 ] 

Erich Duda commented on WFWIP-119:
----------------------------------

bq. Also: the test stopped the replica before live when I ran manually. Do you know in what circunstance that happens? I am playing with some possible options on clearing up tokens, but I need to understand this first.

Yes the test stops at first backup, then it stops live and then it stops server 4. At this point, servers 1, 2 and 4 are stopped. Then the servers are started in order 1, 4, 2.

> [Artemis 2.x upgrade] Lost message in replicated HA scenario with MDB
> ---------------------------------------------------------------------
>
>                 Key: WFWIP-119
>                 URL: https://issues.jboss.org/browse/WFWIP-119
>             Project: WildFly WIP
>          Issue Type: Bug
>          Components: Artemis
>            Reporter: Erich Duda
>            Assignee: Martyn Taylor
>            Priority: Blocker
>
> *Scenario*
> * Start live/backup pair in replicated dedicated topology and start one more server which is in Artemis cluster with live.
> * Send messages to queue InQueue.
> * Start another server with MDB connected to cluster with live server which resend messages from InQueue to OutQueue.
> * Stop backup server and then live server.
> * Stop server which was in cluster with live server.
> * Start live server and server in cluster.
> * Start backup server.
> The test fails because one message get lost. It wasn't sent to OutQueue.
> Logs from failed test including traces you can find in \[1\].
> \[1\] https://eap-qe-jenkins.rhev-ci-vms.eng.rdu2.redhat.com/job/eap-7.x-messaging-run-one-test-until-fail/201/



--
This message was sent by Atlassian JIRA
(v7.5.0#75005)