qpid-users mailing list archives

Site index · List index
Message view « Date » · « Thread »
Top « Date » · « Thread »
From Gordon Sim <g...@redhat.com>
Subject Re: qpid 0.22 reconnect fails on centos 5
Date Wed, 26 Feb 2014 11:25:27 GMT
On 02/25/2014 08:46 PM, bruce.parr wrote:
> I have a two node cluster on a lan where I've installed qpid v 0.22.  I have
> a centos 6 and a centos 5 client machine (not same machines as brokers) with
> same qpid version installed (built from source).  The spout/drain test apps
> handle reconnects just fine on centos 6.  On centos 5, messages are sent and
> received with no problems, through a simple direct exchange.  When I kill
> the qpidd-primary daemon on the broker, I see something on the centos 5
> client like:
>
> essage(properties={spout-id:f840202c-fc7b-49cf-bbad-5690cffb33d2:0,
> x-amqp-0-10.routing-key:bruce, x-amqp-0-10.timestamp:1393356961},
> content='More content 7')
> Message(properties={spout-id:12ef8897-b0ba-4634-9171-7d9dc336e490:0,
> x-amqp-0-10.routing-key:bruce, x-amqp-0-10.timestamp:1393356965},
> content='More content 8')
> Message(properties={spout-id:7b8d40d8-3c23-4fcf-bef8-e98eaba73a3a:0,
> x-amqp-0-10.routing-key:bruce, x-amqp-0-10.timestamp:1393356969},
> content='More content 9')
> 2014-02-25 13:36:13 [Client] warning Connection [10.20.34.6:43352-cirm:5672]
> closed
> Connection [10.20.34.6:43352-cirm:5672] closed
>
> It just disconnects.

Is an exception thrown?

> The cluster behaves as it should, ie. the primary moves to the other node,
> and eventually I can reconnect manually by kicking off drain again.  Am
> using drain with:
>
>    ./drain -b host:5672 -f -c 99999 queuename
>
> I modified the source to used hard coded connection options:
>
>        qpid::messaging::Connection connection("cirm:5672");
>        connection.setOption("reconnect", true);
>        connection.setOption("heartbeat", 3);
>
> No other reconnect options are used.  I've set logging to trace+, found
> nothing untoward.
> Any ideas why its not working on centos 5?  Thanks.

It might be due to a problem with the exception types which would be 
fixed by http://svn.apache.org/viewvc?view=revision&revision=1464930. 
The reconnect logic works by handling a TransportException, however 
prior to the change above the actual types of the exceptions thrown by 
the messaging library were being lost (on RHEL5, so likely also on centos5).


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


Mime
View raw message