activemq-dev mailing list archives

Site index · List index
Message view « Date » · « Thread »
Top « Date » · « Thread »
From "Trym A (JIRA)" <j...@apache.org>
Subject [jira] Reopened: (AMQ-492) Design error in MulticastDiscoveryAgent use of FailoverTransportConnector?
Date Wed, 20 Dec 2006 09:59:02 GMT
     [ https://issues.apache.org/activemq/browse/AMQ-492?page=all ]

Trym A reopened AMQ-492:
------------------------

    Regression: [Regression]
             
Hi,

We have been experiencing a strange issue with our setup of paired brokers (using incubator-activemq-4.1).
Running our brokers with the Rootlogger set to DEBUG level we experience the same issues with
the FailoverTransport going into a very tight endless loop of reconnection attempts upon shutdown
of one broker of a pair.

Our environment:

We have two servers (dev1 and dev2) where we have a broker running on each. They both share
the same Oracle datasource. Starting the broker on either server, this broker becomes the
Master-broker on the datasource, whereas the second broker will report "Attempting to acquire
the exclusive lock to become the Master broker" upon startup.

This is all fine. The problem occurs however when attempting to shutdown the Master-broker
where we expect the slave-broker to become the new Master-broker. Running under DEBUG-level,
the Master-broker hangs indefinitely.

Changing to log-level INFO fixes the problem.

Please advise if there is some bug connected to the FailoverTransport if the log-level happens
to be DEBUG.

> Design error in MulticastDiscoveryAgent use of FailoverTransportConnector?
> --------------------------------------------------------------------------
>
>                 Key: AMQ-492
>                 URL: https://issues.apache.org/activemq/browse/AMQ-492
>             Project: ActiveMQ
>          Issue Type: Bug
>          Components: Connector
>    Affects Versions: 4.0
>         Environment: w2k and linux w java 1.5
>            Reporter: Dennis Cook
>         Assigned To: Hiram Chirino
>             Fix For: 4.0
>
>
> I have been trying to get my head around the multicast discovery mechanism.  In particular
its use of the “FailoverTransport” class is bothering me.  The issue presented itself when
viewing the DEBUG level log entries that were generated after one broker of a pair, connected
using this discovery mechanism, was stopped.  The FailoverTransport went into a very tight
endless loop of attempts to reconnect.
> My first thought to “tweak” the Failover properties in the network connector URI. 
But alais, the uri was for the multicast not the failover.  No problem, just duplicate the
failover properties on the multicast agent.  After doing so, I started to think “why is failover
needed at all”  won’t the connection be reestablished when the broker returns and advertises
itself again?
> Can someone comment on this issue?  Is the bug that the failover properties where not
available or is the problem that failover was used at all?

-- 
This message is automatically generated by JIRA.
-
If you think it was sent incorrectly contact one of the administrators: https://issues.apache.org/activemq/secure/Administrators.jspa
-
For more information on JIRA, see: http://www.atlassian.com/software/jira

       

Mime
View raw message