JDBC to PROXY Sync Scenario

[former_member340212]

Hi All,

I am facing an issue during JDBC to Proxy sync scenario. In Sandbox, JDBC to RFC sync scenario is working fine, but in Dev (JDBC to Proxy sync scenario without BPM), the JDBC Sender CC is trying to pick data from DB but no processing happens. Instead, the JDBC Snd CC is locking that particular record. If we try to modify that particular record from db level, we are getting ORA-00054 error as the table is getting locked. To solve this, the DBA is killing the sessions from db level.

But still, even after killing the sessions, the JDBC Snd CC is not picking up the data and throwing the same error.

In JDBC Snd CC, QoS used is BE and both OB & IB Service Interface's were Sync.

Referred the below links while configuring the scenario:

http://wiki.sdn.sap.com/wiki/display/XI/Using+Request+Response+Bean+Module+in+FILE+Adapter

http://scn.sap.com/thread/3207257

http://scn.sap.com/thread/1749229

Can someone help me in fixing the above mentioned issue.

Br,

Praneeth.

[former_member193393 · Accepted solution]

Hi Praneeth,

The issue should be wid the config of ur Sender JDBC Comm Channel. Make sure tat the option 'Disconnect from DB' under Advanced tab is chked as shown below and test again. This will surely work.

If the above option is not chked, the Snd CC while picking up the record tries to update the same using the update statement and since ur SI is a sync one, it expects a response and waits for the same. Meanwhile ur Rcv CC fetches the response and tries to update/insert and finds tat the record is locked and finally ends up in a long transaction. This results in a lock at the DB level.

Hope this helps.

Cheers,

SaiGanesh.

[former_member184681]

Hi Praneeth,

Did you try to test the proxy itself? My suspiction is that your proxy receiver is working for a very long time (performance issues), which you misinterpret as a lock on the table. I mean: PI locks the DB table entries while processing the data and sending it to proxy, and since the latter takes a lot of time, the locks stay there for so long. Can you confirm if my suspiction is true?

Regards,

Greg

[former_member340212]

Hi Greg,

There were no time delays during proxy execution, as we are testing with only one record (9 fields) and also there are no transformations in PI mapping logic. The scenario behaves as below, after the locked sessions were killed at DB level.

1. JDBC Snd CC tries to pick data from DB, but doesn't change the STATUS_FLAG from 0 to 1 in the db table

2. No log is seen in Comm Channel monitoring, but we can see the message flow in the IE (both request & response)

3. But the response is not getting updated in the DB and again the DB table is locked

Br,

Praneeth Rao

[baskar_gopalakrishnan2]

I believe you are using two jdbc channels one for sender and the second one for receiver. This interface pattern is async sync async. So I would recommend to use Exactly once as Quality of service for the sender jdbc channel.  Please remember jdbc sender is async based.

You might also want to see this link for design and configuration

http://wiki.sdn.sap.com/wiki/display/XI/File-RFC-File(Without%20BPM)

[former_member193393]

Hi Praneeth,

Also don't forget to chk the option 'Database Auto-Commit-Enabled (No Transaction Handling). Revert if the issue still persists.

Cheers,

SaiGanesh.