on 11-03-2011 5:53 PM
Long runtime for "Fill Internal Header Table for VBUK via Doc. Flow Predeces." activity
we are at TDMS 3.0 - SP18, DMIS SP18, DMIS_CNT SP18, DMIS_EXT 12
We are running the time slice data extract (4 months) and the data selection is in progress.
The background job "CNV_MBT_CNV_TDMS_05_TI_CO_VBUK_2" for the sub- activity "TD05X_FILL_VBUK_2" is running in sender system for 11 hours and the SQL statement which is running for this process is using up the transaction log space at the DB level.
[SQL statement: SELECT "VBELV", "POSNV", "VBELN" FROM "VBFA" WHERE "MANDT" = ? AND "VBELN" = ? WITH UR -- OPTLEVEL( 5 ) -- QUERY_DEGREE( 1 ) -- LOCATION( CNV_TDMS_05_TI_CO_VBUK_2_N , 1504 ) --
SYSTEM( QR5 , SAPR3A )]
We implemeted the "SAP Note 1620101 - Performance issue in VBUK 2 Fill header" in our receiver and sender before creating the current data transfer package 90001.
We also refferred the SAP Note 1256679, but this note is not applicable to our TDMS release.
Request your help in resolving this issue.
Thanks in advance,
Prajeevan
Hello Prajeevan,
First of all, Note 1256679 is always a good start for a long runtime on vbuk_2.
If possible please create the index refered in that note on the Sender System.
Another option (recommended) is to activate the cluster read method for all vbuk activities.
Problem in your case is that to have a consistent dataload the cluster method neerds to be switched for all three vbuk activities.
For the next run i would really recommend to use the troubleshooter for the vbuk_0 activitiy and set P_CLU to yes in the activity parameters. This normally will automatically also set the P_CLU Parameter for the other two VBUK activities.
To really get to the core of the issue i think the same like me preposter that a more detailed analysis needs to be done.
(How many records do you have in VBUK - VBFA?)
Best Regards
Joerg
You must be a registered user to add a comment. If you've already registered, sign in. Otherwise, register and sign in.
Hi,
one other comment. The other note 1620101 seems only to be including changes for the cluster read method.
So as long as you have not switched to P_CLU you will not see any differences!
If you want to activate P_CLU please retrigger also the first two VBUK activities again.
Best Regards
Joerg
Hi Joerg,
We checked the OSS note 1054584 and Implement the corrections specified in it.
Inserted parameter P_ROLL with the value 'X' (upper case) for activities TD05X_FILL_VBUK_2 and TD05X_FILL_VBUK_2_0
in table CNVMBTACTPARAMS and restarted the jobs.
The job that run for 24 hours without completing got completed now in 2 hour 15 mins.
Next time before we start our run, we will go for cluster read method
Thanks for your help and suggestions.
Regards,
Prajeevan
This particular issue will require a detailed analysis in your system. I suggest you to create an OSS ticket for the same.
You must be a registered user to add a comment. If you've already registered, sign in. Otherwise, register and sign in.
User | Count |
---|---|
85 | |
10 | |
10 | |
9 | |
7 | |
6 | |
5 | |
5 | |
4 | |
3 |
You must be a registered user to add a comment. If you've already registered, sign in. Otherwise, register and sign in.