Jens
This is a good idea.. Currently the "fault" sequence is invoked when
Synapse encounters such an error - and it defaults to just logging the
message. So one could use this to store the message into a persistent
store, and maybe use an admin console or JMX to later pull out messages
from this and retry
asankha
Ruwan Linton wrote:
Hi Jens,
The Dead Letter Channel makes much sense to me :) but I think the
notification through IM and all the rest has to e done through the
configuration by the admin.
It is nice to support all these through a console but at least for the
moment we don't have a console inside Synapse, may be you are talking about
the WSO2 ESB console for synapse, in which case yes of course we can do
that.
You could go ahead and add a JIRA for the Dead Letter Channel
implementation. I think we can implement this as a mediator.
Thanks,
Ruwan
On Wed, Mar 12, 2008 at 11:31 PM, Jens Goldhammer <
[EMAIL PROTECTED]> wrote:
Hello mailinglist,
I am working with two Synapse/WSO2 ESB which work as a proxy for two
bpel engines.
bpel 1<->synapse 1<->synapse 2<->bpel 2
If the bpel engine invokes an other process in bpel 2, Synapse is called
and it tries to connect to the other synapse (2).
If the synapse instance 2 is down, I will have the problem that Synapse
1 only tries one time to send the request. BPEL engine 1 is believing
that everything is ok (long running process), so that I have no chance
to retry the request to the synapse instance 2. As long as ws-rm does
not work correctly, what can I do?
It would be really cool to have a dead-letter-channel (see gregor hohpe
eip) where all lost messages will be make persistent with the date of
the last try for sending and to have the ability to resend the message
manually from the admin console. I would prefer a list of messages to
see their content, the date of last try for sending and the value of
the to-header. Then you can select several messages, press the button
retry and all selected messages will be send to the destination. It
would be nice to have a mechanism to determine if a request of a proxy
service to another service should be make persistent in the
dead-letter-channel, maybe by specifiying it in the send-mediator!
Furthermore it would be great to notify the admin after a message was
inserted in the dead letter channel by taking the im-mediator...
What´s your opinion on this? Should I raise a JIRA or does this makes no
sense?
Thanks
Jens
