Incorrect issue with JDBC Adapter

[former_member5086]

HI,

In OUR production system, we have 38 Sender communication channel with JDBC Adapter and com.microsoft.sqlserver.jdbc.SQLServerDriver to access to multiple databases with 5 minutes polling.

All these channels work correctly since this night, when five of this channels seems to became inconsistent. They don't execute any store procedure and don't return any error. Monitoring SQL sessions, we don't find any access of our XI system.

In channel monitor I can see only:

10/2/09 10:56:31 AM Processing started

10/2/09 10:46:25 AM Processing started

10/2/09 10:36:34 AM Processing started

10/2/09 10:26:34 AM Processing started

10/2/09 10:16:39 AM Processing started

10/2/09 10:06:21 AM Processing started

10/2/09 9:57:08 AM Processing started

10/2/09 9:46:21 AM Processing started

10/2/09 9:36:23 AM Processing started

10/2/09 9:26:28 AM Processing started

I stop and restart this five channels but nothing has changed.

Any input, please? Thanks,

Carme

[former_member187339 · Accepted solution]

Hi Carme,

We have faced this type of issue in our JDBC landscape and when we set the option Disconnect from Database After Processing Each Message in communication channel the problem was solved

Check this help.sap page http://help.sap.com/saphelp_nw04/helpdata/EN/7e/5df96381ec72468a00815dd80f8b63/content.htm

Also check note Note 831162 - FAQ: XI 3.0 / PI 7.0 / PI 7.1 JDBC Adapter

point 7

Regards

suraj

[Former Member · Accepted solution]

Yes, this option solved my problem also. It works good.

Apparently, connection to database becomes stale after some time due to this this problem occurs. I tried using a JDBC client, left it inactive for some time, same problem reappeared.

[former_member5086 · Accepted solution]

thanks for your replies.

Suraj, i have this parameter (Disconnect from Database after processing each message) checked in all this channels.

I think this parameter can minimize this issue but as i can see, it doesn't solve.

Ninad, the restart our production system with a great volum of message and channels is not a good solution of this problem because i need to avoid it could be reproduced.

Any other input?

Camre.

[former_member207622 · Accepted solution]

Hi

Can you tell us whether we are having decentralized AE or centralized one and also there is one note mentioning that JDBC doesnt pick the processed data

Please have a look at these links

SAP note : Note 800787

Thnx

ninad

[former_member5086 · Accepted solution]

Ninad.

We have centralized adapter engine.

Sap note 800787 doesn't apply because of in our case, sourde data is not processed but the update statement is neither executed . In fact, we have checked that any access from XI system is executed in SQL server.

Carme.

[Former Member · Accepted solution]

Please add Pool waiting time and Task Time Out parameter in advanced mode parameters of JDBC Channel.

Check is there any lock in Adapter Locking Services in Visual Admin when the issue is occouring.

Is your production system HA? If yes, is this issue happening for any particular server node? Have u used cluster synch mode = lock?

What is the driver version that you are using and corresponding DB version for which problem occours?

Warm Regards,

Gouri

[former_member5086]

thanks for your replies.

To avoid this issue from happening again i hve set the tasktimeout as described in note

1083488 - XI FTP/JDBC sender channel stop polling indefinitely

Carme

[former_member207622]

Hi

Please check whether the sender system port is open or not that might be an issue and

you might need to restart J2EE engine for this

check a similar discussion here

Thnx

ninad