[
https://issues.apache.org/jira/browse/QPID-8006?page=com.atlassian.jira.plugin.system.issuetabpanels:comment-tabpanel&focusedCommentId=16237494#comment-16237494
]
Lorenz Quack commented on QPID-8006:
------------------------------------
I tried to reporduce this but was not able to.
I think this might be a race condition.
Looking through the logs it seems like I preformed these actions:
# Initially we have to VHNs {{sdgfsfsdfsfsdf}} (MASTER) and {{dfsdf}} (REPLICA)
forming the group {{g1}}
# at 14:57:14,901 I delete the VH {{g1}} on the {{sdgfsfsdfsfsdf}}
# at 14:57:42,864 I create a new VH {{sdgfsfsdfsfsdf}} on the VHN
{{sdgfsfsdfsfsdf}}
# at 14:57:57,782 I stop the VHN {{sdgfsfsdfsfsdf}}
# at 14:58:07,891 I set some attributes on VHN {{dfsdf}}. potentially
{{designatedPrimary=true}}
# at 14:58:16,648 {{dfsdf}} tries to switch to MASTER
# There is an Exception at 14:58:16,659 but the node seems to switch to MASTER
regardless.
# at 14:58:30,790 I try to create a VH {{dfsdf}} on VHN {{dfsdf}}.
# This results in a
[{{com.sleepycat.je.rep.DatabasePreemptedException}}|https://docs.oracle.com/cd/E17277_02/html/java/com/sleepycat/je/rep/DatabasePreemptedException.html]
when trying to open the database {{AMQP_1_0_LINKS_VERSION}} at 14:58:30,830
# This exception is converted to a {{StoreException}} and brings down the
broker.
I think when the VH {{g1}} on {{sdgfsfsdfsfsdf}} gets deleted it deletes the
LinkRegistry including the AMQP_1_0_LINKS(_VERSION) databases. This db deletion
is replicated to {{dfsdf}} and causes the {{DatabasePreemptedException}} when
we try to open the new {{dfsdf}} VH.
> [Broker-J, HA] Deleting VH, stoping/starting VH(N)s, and switching mastership
> can result in broker crash
> --------------------------------------------------------------------------------------------------------
>
> Key: QPID-8006
> URL: https://issues.apache.org/jira/browse/QPID-8006
> Project: Qpid
> Issue Type: Bug
> Components: Java Broker
> Affects Versions: qpid-java-broker-7.0.0
> Reporter: Lorenz Quack
> Attachments: qpid_.log
>
>
> Not sure which steps I took to cause this but it was some order of deleting
> VirtualHosts, stoping/starting VirtualHosts/VirtualHostNodes, and transfering
> MASTERship between nodes. But the end result was a crashing broker.
> I attach the complete (verbose) log so we can figure out what exactly I did
> and what exactly happened.
--
This message was sent by Atlassian JIRA
(v6.4.14#64029)
---------------------------------------------------------------------
To unsubscribe, e-mail: dev-unsubscr...@qpid.apache.org
For additional commands, e-mail: dev-h...@qpid.apache.org
