on 02-26-2014 3:49 PM
Interface Summary –
Domain Student Marks Processing in SLCM.
Interface - Async Web- service to Proxy scenario.
Issue –
Marks are getting processed in ECC through Web-service to Proxy
scenario.
Issue occurs when many marks messages are triggered to ECC some of the
messages are failing in SXMB_MONI with the below error message.
Usually Marks are processed in ECC takes on an average 6-15 min to
complete the entire process.
ER_USER used is used to call ECC and configured in PI
communication
channel.
Error Message of MONI -
<SAP:Category>XIServer</SAP:Category>
<SAP:Code area="INTERNAL">QRFC_SYSTEM_DUMP</SAP:Code>
<SAP:Stack>Short dump occurred when executing message in qRFC queue
XBTR0001: Date/time 25.02.2014 18:16:34, user ER_USER, runtime
error: Object requested is currently locked by user
ER_USER</SAP:Stack>
Analysis Done so far -
ST22 doesn’t show any short dump as well as which objects are locked by user. Messages are getting assigned to XBTR* queues some messages are getting processed successfully others are getting into Locking issue. Our Program is cable enough to handle the lock in the program and if there is an lock during the processing it will just capture the error message in Error table and continue processing other messages within the payload.
Has any one come across above issue any time and kindly provide your inputs.
Thanks
C G
The Solution for releasing a queue early is not working fully as some of the messages are lost in background processing.
Has any body has any other solution.
Thanks
Chirag
You must be a registered user to add a comment. If you've already registered, sign in. Otherwise, register and sign in.
HI Chirag,
We had something similar some time back in our system and we checked everything. This error is very misleading and points to QRFC queses which is not an issue because you can see other messages are flowing successfully.
This error is because of data. Rectify data issue and it will flow successfully.
Regards
Aashish Sinha
You must be a registered user to add a comment. If you've already registered, sign in. Otherwise, register and sign in.
Thanks Aashish,
I don't see its a data issue because when I manually restart the message from MONI - message is getting processed successfully in ECC and MONI also shows successful flag.
Test performed but Not worked fully
Based on the above test conclusion is –
Thanks
Chirag.
Hi Chirag,
Indeed it looks like a different issue, i don't see this as a locking issue.
Do you mind checking this sap note?
1625236 - Restart for XI messages after dump in qRFC does not work
In the meantime, i will look out for some other solution for you
Regards
Aashish Sinha
We have implemented workaround solution which is working but it has some drawbacks.
we have change the proxy code with ‘Starting New Task’ – this should release the queue early to avoid the locking issue and message is hand over to internal work process for further processing.
Only drawback is proxy moni message will be successful after queue is released early, where as process is still in flight at the background, if there is some proxy dump happen message still be in successful status and can not be restarted again.
Thanks
chirag
Hi Chirag - Did you check if the locks are getting created in sm12 during the heavy load processing?
May be your proxy code is being executed in parallel and the objects are getting locked for that user
I'm not an ABAP expert but how about keeping a wait time for couple of second for each execution?
You must be a registered user to add a comment. If you've already registered, sign in. Otherwise, register and sign in.
Hi Hareesh,
there are no lock in sm12,
Our Program is capble enough to handle the lock in the program and if there is any lock during the processing it will just capture the error message in Error table and continue processing other messages within the same payload.
Error happened before the program starts execution at queue level.
thanks
chirag
Hello Chirag ,
Seems like the the entry is failing if there is already one entry being updated.
This can not be solved by putting a wait in th eproxy code. As if there are 10 messages then 10 concurrent processing will be started by proxy.
I think the messages should come one by one instead all at the same time.There should be some time difference between subsequent messages.
Thanks
You must be a registered user to add a comment. If you've already registered, sign in. Otherwise, register and sign in.
User | Count |
---|---|
81 | |
9 | |
9 | |
7 | |
7 | |
7 | |
6 | |
5 | |
5 | |
5 |
You must be a registered user to add a comment. If you've already registered, sign in. Otherwise, register and sign in.