qpid-dev mailing list archives

Site index · List index
Message view « Date » · « Thread »
Top « Date » · « Thread »
From "Andy Goldstein (JIRA)" <j...@apache.org>
Subject [jira] [Commented] (QPID-4223) Unable to run qpid-cpp-benchmark successfully after shutting down a backup
Date Thu, 16 Aug 2012 02:41:38 GMT

    [ https://issues.apache.org/jira/browse/QPID-4223?page=com.atlassian.jira.plugin.system.issuetabpanels:comment-tabpanel&focusedCommentId=13435730#comment-13435730
] 

Andy Goldstein commented on QPID-4223:
--------------------------------------

I've been able to narrow this down to a simpler test case, and have an explanation for what
is going on (although I haven't traced it all the way to come up with a fix yet).  Here are
my steps to reproduce:

Start broker 1
Start broker 2
Promote broker 1 to primary
Stop broker 2
Open connection to broker 1
Create session 1
Create session 2
Create sender using session 1 (ready; {create:always}) and send a message
Create receiver using session 2 (ready; {create:receiver, delete:receiver}) and call fetch()
Close the receiver (important - WITHOUT acking the received message)
Call sync() on session 1 (the sending session)

Session 1's sync() call will hang because it is waiting for the completion of the message.transfer
command it sent, but the QueueGuard for the (now dead) backup is keeping the command from
completing.
                
> Unable to run qpid-cpp-benchmark successfully after shutting down a backup
> --------------------------------------------------------------------------
>
>                 Key: QPID-4223
>                 URL: https://issues.apache.org/jira/browse/QPID-4223
>             Project: Qpid
>          Issue Type: Bug
>          Components: C++ Broker, C++ Clustering
>    Affects Versions: 0.18
>            Reporter: Andy Goldstein
>            Assignee: Alan Conway
>
> When testing the new HA in 0.18 (from the svn branch, revision 1370693), I've run into
a problem where I can't run qpid-cpp-benchmark after I shut down one of the backups.  Here's
the scenario:
> Start 3 brokers (in my test on 3 different hosts) using the following command line:
> qpidd --auth no --load-module ha.so --ha-cluster yes --ha-brokers-url <insert url
here> --ha-replicate all
> Promote 1 of the brokers to primary.  Run qpid-cpp-benchmark with the default options.
 The test should succeed and provide results.  Next, quit 1 of the 2 backups.  Finally, run
qpid-cpp-benchmark again.  This time, it hangs.  Looking at the output of "qpid-stat -q,"
I see there are some unacked messages that have been delivered to the running qpid-receive
process.

--
This message is automatically generated by JIRA.
If you think it was sent incorrectly, please contact your JIRA administrators: https://issues.apache.org/jira/secure/ContactAdministrators!default.jspa
For more information on JIRA, see: http://www.atlassian.com/software/jira

        

---------------------------------------------------------------------
To unsubscribe, e-mail: dev-unsubscribe@qpid.apache.org
For additional commands, e-mail: dev-help@qpid.apache.org


Mime
View raw message