on 11-29-2011 10:22 AM
Hi All,
I am using PI 7.1 SP 07.
For the outbound interafces whenever MQ system is down PI messages are stuck in adapter engine and once the MQ is up manually we need to stop and start communication channels and resend the messages.
To avoid this we have used one additional parameter in JMS communication channel called "CONNECTION_TEST_POLL".
But this is creating lot of performance issue in MQ side as i belive it keep polling the JMS provider and creating a new TCP/IP connection every time if there is any connection failure. Result in large number of TCP/IP connection which make Queue manager down at some point of time.
We are using same Qeue manager for all outbound interface( 40 interafces).
While serching for solution i found the weblog
[http://www.sdn.sap.com/irj/scn/weblogs?blog=/pub/wlg/17842.] [original link is broken] [original link is broken];
There it is mentioned about the new feature available to JMS adapter u201CConnection on Demandu201D.
But this feature is available for PI 7.1 SP 8 or PI 7.11(EPH 1)
There is no plan to upgrade the PI system .
Any idea.. how we can solve the issue in Pi 7.1 SP 07.
Hello,
To start and stop JMS Communication channel automatically you can use the Availability Time planning (ATP)as per you convenience.
I hope this will work for you.
Thanks
Monica
You must be a registered user to add a comment. If you've already registered, sign in. Otherwise, register and sign in.
One of the simple solutions that we had done before the SP upgrade was to schedule the adapter to STOP and START every hour for a duration of 30 sec.
Thus the connection was reestablished and the message delivered.
You must be a registered user to add a comment. If you've already registered, sign in. Otherwise, register and sign in.
Hi Shabarish,
This seems to be a good workaround.Thanks for this.
Are you suggesting to use the paramtere "CONNECTION_TEST_POLL".
and then use ATP on top of it?
Considering the fact that in production it is rare that system goes down, do u think ATP will be useful and also we are having many time critical interface so the 30 seconds stop period will cause issue for few interfaces.
Regards,
Biplab Das
Hi Shabarish,
>
> This seems to be a good workaround.Thanks for this.
>
> Are you suggesting to use the paramtere "CONNECTION_TEST_POLL".
> and then use ATP on top of it?
>
you can use ATP without the parameter
> Considering the fact that in production it is rare that system goes down, do u think ATP will be useful and also we are having many time critical interface so the 30 seconds stop period will cause issue for few interfaces.
>
>
> Regards,
> Biplab Das
you could reduce the timings from 30 sec to 5-10 sec. this can be a work around until you patch the system
User | Count |
---|---|
84 | |
10 | |
10 | |
10 | |
7 | |
6 | |
6 | |
5 | |
4 | |
4 |
You must be a registered user to add a comment. If you've already registered, sign in. Otherwise, register and sign in.