[
https://issues.apache.org/jira/browse/QPID-5982?page=com.atlassian.jira.plugin.system.issuetabpanels:comment-tabpanel&focusedCommentId=14092198#comment-14092198
]
ASF subversion and git services commented on QPID-5982:
-------------------------------------------------------
Commit 1617159 from [~k-wall] in branch 'qpid/trunk'
[ https://svn.apache.org/r1617159 ]
QPID-5982: [Java Broker] BDB HA: Reset replication transaction id when contact
with remote node is lost
* Change BDB HA VHN tab so that replication transaction id -1 is shown as N/A
* Adjust width of VH table on Broker tab to avoid state UNAVAIABLE wrapping
over two lines.
> Stopped node viewed remotely reports stale replication transaction id
> ---------------------------------------------------------------------
>
> Key: QPID-5982
> URL: https://issues.apache.org/jira/browse/QPID-5982
> Project: Qpid
> Issue Type: Bug
> Components: Java Broker
> Affects Versions: 0.29
> Reporter: Keith Wall
> Assignee: Keith Wall
> Priority: Minor
> Labels: bdbstore, ha
> Fix For: 0.31
>
>
> I have a HA group formed of three nodes. I stop one the nodes using the Stop
> button on its VHN tab. The node stops correctly and temporarily detaches
> from the group leaving the two remaining nodes providing service.
> If I then observe the group from the VHN tab of either of the two remaining
> nodes, I see the role of the stopped node as UNKNOWN (correct behaviour), but
> the replication transaction id is still populated (with the last known
> value). This information is stale and might mislead the user. In the case
> where the group has suffered a network partition the value could be wrong.
> A node should discard the remote replication transaction id when it looses
> contact with the remote node and should display N/A.
--
This message was sent by Atlassian JIRA
(v6.2#6252)
---------------------------------------------------------------------
To unsubscribe, e-mail: [email protected]
For additional commands, e-mail: [email protected]
