cancel
Showing results for 
Search instead for 
Did you mean: 

PI message creation issue SXMB_MONI

Former Member
0 Kudos

Dear PI Expert,

I have a issue for one of the PI interface from SQL Server to ECC. There are two interfaces one interface syncronise data correctly on a regular basis i.e. after every 30 minutes. But second interface update table record status in SQL Server as if data has been posted to PI->ECC but message does not appear in SXMB_MONI.

but when i refresh full cache using  http://<hostname>:<J2EEport>/CPACache/refresh?mode=full and also 'Integration builder' - > 'clear SLD data cache'

it start creating message in 'SXMB_MONI' and data gets pushed to ECC but old records for which message not created remain missing in ECC. and problem is repetative and not able to diagnose.

Regards

Vaman Kamat

Accepted Solutions (0)

Answers (2)

Answers (2)

former_member190624
Active Contributor
0 Kudos

Hi ,

Few cents, please check the runtime parameter LOGGING_SYNC on integration engine(sxmb_adm) configuration . If parameter (LOGGING_SYNC) was set to zero, change it to 1 and check you scenario again.

Thanks

Hari.

Former Member
0 Kudos

Hello Vaman,

do you see messages in Adapter Engine monitoring before? Who or what triggers your messages?

The "clear SLD cache" is purely design time related and should have no effect on runtime behaviour. The full cache refresh only should have indirect effects, maybe it restarts Communication Channels.

Regards,

Jörg

Former Member
0 Kudos

Hello sir,

I have checked 'Adapter Engine' monitoring in runtime workbench both inteface shows successful with no. of successful attempt in 'success' column but as far as 'Error' column shows '-' ( No error )

Message gets trigger when SQL Server realise a table entry with status 'N' and once PI pulls the data status automatically changes to 'X'. Currently status gets change to 'X' without message in 'SXMB_MONI'

Regards

Vaman Kamat

Former Member
0 Kudos

Hello Vaman,

after the message pull by PI the message will first appear in Adapter Engine. Can you see that? Is the Communication Channel ok?

Can you confirm that you see the messages in Adapter Engine, but not in SXMB_MONI?

Or do you see the messages in AE after cache refresh etc.?

Regards,

Jörg

Former Member
0 Kudos

Hello Sir,

No . I am not able to see any new message in Adapter Engine and as far as communication channel is concern shows Functional and polling every 60 sec successfully.

Yes. after refreshing cache using http://<hostname>:<J2EEport>/CPACache/refresh?mode=full message appears in AE.

Regards

Vaman Kamat

ambrish_mishra
Active Contributor
0 Kudos

Hi Vaman,

Are you doing a select and update in the JDBC adapter ? If yes, it might be a problem when the select condition returned nothing but the update happened on the table.

What is the config of your adapter ?

Ambrish

Former Member
0 Kudos

Hi Ambarish,

What you are telling may be correct but after refresh and updating complete table record with status 'N' , all records/data get pushed into ECC.

Regards

Vaman Kamat

ambrish_mishra
Active Contributor
0 Kudos

HI Vaman,

What is the config of your adapter ?

Select and Update statements ?

Ambrish

Former Member
0 Kudos

Adapter config = 'JDBC' ;Transport Protocol 'JDBC 2.0'

Select Statement

select * from DATA1_B2 where flag = 'N'

Update statement

Update DATA1_B2 set flag = 'X' where flag = 'N'

Regards

Vaman Kamat

Former Member
0 Kudos

Hello Vaman,

apparently, your channel polls "successfully", but doesn't create any message in the system. So something about the CC or the adapter does not work. Can you try to stop and start the CC instead of performing a full CPACache refresh? Or, alternatively, restart the JDBC adapter in NWA? Do they create the same behaviour? If yes, something may be wrong with your JDBC connection, maybe it closes prematurely or loses connection somehow.

Regards,

Jörg

ambrish_mishra
Active Contributor
0 Kudos

Hi Vaman,

Are you saying that when the table has a mix of N and X records, no records are picked?...and only when all records in the DB table are set to N, the the records are picked.

Ambrish

Bhargavakrishna
Active Contributor
0 Kudos

Hi Vaman,

As per your previous replies, you told that message is not creating when the table is having Mix N and X records, right? if that is the case there is some problem with the update..

How many records that you are reading at a time? if you are reading more records it may be one of the reason for this issue. you should limit the record reading.

Try to stop and start the communication channel.

Do the server restart if possible, sometimes it will helps.

Regards

Bhargava krishna

rodrigoalejandro_pertierr
Active Contributor
0 Kudos

take a look to this blog

http://scn.sap.com/community/pi-and-soa-middleware/blog/2010/04/13/pixi-quick-tip-how-to-get-some-mo...

by the way do you execute the query in the DB directly? does it work?

regards

Rodrigo

Former Member
0 Kudos

Dear Ambarish,

Sometime partial/mix data with both 'N'  also gets syncronised in ECC without refresh Cache. Today few messages got generated automaticallly but few are missing.

On sunday complete data is available in ECC without error.

Regards

Vaman Kamat

Former Member
0 Kudos

Dear Bhargava,

Yes, Maximum record reading will be 30-40.  How i can limit the record reading. ?

Regards

Vaman Kamat

ambrish_mishra
Active Contributor
0 Kudos

Hi Vaman,

I have worked extensively in JDBC scenarios and never experienced such issues, thats why it surprises me.

>>>>Today few messages got generated automaticallly but few are missing.

I suggest you look at some aspects:

  • Are you sure that there are no connectivity issues between PI and DB. For this, you need to monitor this scenario during the day keeping someone who has access to the database in the loop. Keep a track when the records are there to be picked up and track PI when the polling happens (you mentioned it runs every 30 mins).
  • If the records are not picked up, check if there is a tale locking happening through some job internal to the database.
  • Also, for the timebeing, uncheck Disconnect from DB after message processing.

Hope it helps you nail down the problem.

Ambrish