Delays in sm37

Hello BW Experts,
I am observing lot many delays in various jobs in our BW Production system.
Around 300 to 400 jobs are getting delayed by 900 to 2600 seconds over past week.
Out of which around 100-200 jobs are delayed over 2000seconds.
And on some odd days these delays went up as high as 7500 seconds. Like 80 - 100 jobs getting delayed by 5000 to 7500 seconds.
These delays some or the other way hit our SLA with client.
I just wanna ask is this a serious issue to be looked upon on priority. If yes please give your suggestions and if possible any reference to OSS notes.

Hi,
I am not sure by upgrading patch to over come this issue. As your jobs system jobs and data load jobs get getting delay 1000s of seconds means it is due to Unavailability of the work processes and load balancing ( it can be done if you have multiple App.servers), you can check in SM50 for availability of the processes to run your job or try by scheduling the jobs on sequence and based on priority. it may reduce some how but it will not incease overall performance of the system.

Similar Messages

SM37 Job Delay Seconds

Hi, When my scheduled LIS Jobs run on R3 I am noticing that they all show some kind of Delay in SM37 before they run. The Jobs run every 4 hours and during most of the runs there isn't any real amount of system activity going on so I'm guessing this shouldn't be happening.
Can anyone tell me what would cause this delay and what/should I do anything about it?
Thanks!

Ken,
How much is the delay? Is it < 60 seconds or more than that? The job scheduler wakes intermittently to run the background jobs (parameter rdisp/btctime in the instance profile). If it is more than that, then it usually there are no available jobs on the designated server or in the system.
-regards

How to change the Timestamp In a program

Hi,
We faced a problem today, Some of the critical jobs from a source system are going to short Dump. Unfortunately we couldn't find any lockwaits, deadlocks and any delay in SM37. After analyzing, we came to know that, SAP trace indicates failure occured during program /GLB/RGTCSR_RC_INV execution :
B *** ERROR => DISTRIBUTED_TA_FAILED:
The timestamp for the execution of the program has been set as "20.070.801.000.000", that does mean it will try to retrieve all values after the date 01.08.2007 and the time 00:00:00. Now checking on the code we can see there is a logic written which retrieves the values from the table /GLB/RGTT_CS_RC, checking on the Goods Movement status (wbsta) , timestamp and document number (VBELN_SO) if any. For this case VBELN_SO is blank and vbsta is equal to space (Not Relevant). This retrieval helped to logic to retrieve 706706 document and items from the table. This amount of data is huge and we would like to request BW team to run the program with a smaller timestamp to avoid any inconsistencies and also to avoid the dumps which can occur when we try to play with this huge entries.
Checking on the code, two similar selects have been written on the same table /GLB/RGTT_CS_RC, firstly to select all the document numbers relevant to the timestamp and wbsta value and secondly to check "Get records that may have already been used", depending on the document number and item numbers. But as we are doing a select * it will automatically select VBELN_SO as well as POSNR_SO. Hence, the second select from the same table is irrelevant and it will only effect the performance of the system, in the worst case it will terminate the program with this huge amount of entries. Hence, this logic has to be combined to one single select for which the first select statement is enough. Playing with a huge amount of data 706706, and retrieving from the same table twice can surely result in a short dump. Also there are couple of loops has been done on the same table which reduces the performance too.
Possible solution looks like we need to reduce the time stamp and run it,
can some one tell me how to change the time stamp in a program???

First check the connection with the source system. If its ok then replicate the datasource and activate. Then try with running job.
Time stamp error usually happens when source system in BI and r/3 side differs.

Background job Delay In ( sec)

Hi Experts ,
I submitted one job in background Via SM37 here i have one question.
Job is still in Active but it shows Delay ( sec ) 2,181.
Duration ( Sec) 1,999
what it means ?
Thnx .

Hi
refer the below Link:
http://help.sap.com/saphelp_nw70/helpdata/en/1d/ab3207b610e3408fff44d6b1de15e6/content.htm

How to remove spool generation setting for SM37 job.

Hi All,
please help ASAP as this is production issue.
There is a job in SM37 which is running daily and it is creating a spool with nearly 10000 pages.
we dont need this spool as it is now not useful to us.We want to run this job without generating the spool.....
Im stuck here...pls help me...what should I do?
As user are not priting these spool requests so the option "delete after print" is also not working....
so due to huge spool,all jobs are getting delayed..
please help ASAP....

Hi,
Is this related to Business Objects WebIntelligence reporting ontop of BW ?
Perhaps this thread need to be moved to BW warehouse administration?
Regards,
H

Need clarification on one running job in SM37

Dear friends,
right now am working for one client , so we are the second vendor for them and we have been taking care of the project
now i can able to see my client's previous vendor user names are still availbae in SU01..
so instead of deleting their user names i just wanted to see whether is there any scheduled jobs are there for that particular user id and then using SM37 transaction i found one job got scheduled and it is running daily... i.e.
job job created status start date start time duration delay
STAUS-EMAIL RAM FINISHED 21.02.2011 00.00.39 1 39
PLEASE HELP ME FRIENDS ..WHAT DOES IT MEAN ? CAN I LOCK THAT PARTICULAR USER??? OR CAN I DELETE THIS JOB??
CAN U PLEASE PRIOVIDE SOME INFORAMTION ....
THANK U...
REGARDS
YUVA
Edited by: ganandkumar on Feb 21, 2011 9:27 AM
Edited by: ganandkumar on Feb 21, 2011 9:45 AM
Edited by: ganandkumar on Feb 21, 2011 9:45 AM

For this JOB, you can check what is the Program and Program variant it is executing...
From the log you gave, it seems its a custom job scheduled on some program.
Select the job, and click on steps.. it will take you to step list overview...
then menu goto -> variant.
this will tell you what program was running..
based on the program you can decide if the SM37 job should be deleted or not..
If you found that you still need the job, then you can just change the executing user of that JOB,,, replacing the older ID with any new ID

Rollup Getting Delayed

Hi All,
I am doing the Rollup for ZIC_C03 cube manually,the problem is after execute the Rollup it is getting delayed to complete. I have started one job on 02/01/06 that particular job is still in active mode in SM37 and while trying to cancel the job it's not getting canceled.
I have found some errors in Details tab for the particular load at monitor level the error says that <b>System error: CONDENSE_AGGREGATES RSCDS_CONDENSE_CUBE</b>
Any suggestions let me know
Thanks in Advance

Hi Kris,
Refer OSS Note - 642819 and 834829.
Bye
Dinesh

Delay in loads

Hi All,
can you pleas tell me what might be the problme when the data is pulling from r3 to bw i am pasting the log from sm37 there is a delay of job from 7:42 and 9:52 am
Date Time Message text
05.12.2008 07:41:28 Job started
05.12.2008 07:41:28 Step 001 started (program SBIE0001, variant &0000000737525, user
05.12.2008 07:41:29 Asynchronous transmission of info IDoc 2 in task 0001 (0 parallel
05.12.2008 07:41:29 05.12.2008 07:41:29 *****************************************************************
05.12.2008 07:41:29 * Current Values for Selected Profile Parameters
05.12.2008 07:41:29 *****************************************************************
05.12.2008 07:41:29 * abap/heap_area_nondia......... 2000683008
05.12.2008 07:41:29 * abap/heap_area_total.......... 4000000000
05.12.2008 07:41:29 * abap/heaplimit................ 40894464
05.12.2008 07:41:29 * zcsa/installed_languages...... DEJMN13
05.12.2008 07:41:29 * zcsa/system_language.......... E
05.12.2008 07:41:29 * ztta/max_memreq_MB............ 2047
05.12.2008 07:41:29 * ztta/roll_area................ 3000320
05.12.2008 07:41:29 * ztta/roll_extension........... 1048576000
*05.12.2008 07:41:29 ******************************************************************
05.12.2008 09:52:18 BEGIN BW_BTE_CALL_BW204020_E 3.226
05.12.2008 09:52:18 END BW_BTE_CALL_BW204020_E 3.22605.12.2008 09:52:18 BEGIN EXIT_SAPLRSAP_002 3.226
05.12.2008 09:52:18 END EXIT_SAPLRSAP_002 3.226
05.12.2008 09:52:18 Asynchronous send of data package 1 in task 0002 (1 parallel task
Thanks
Dp

Hi......
Due to this ur log is showing in that way..........in cse of Direct update first Extraction will happen........at that time log will not change............it will stop there only............at the end of the Extraction.......when first data packet will be picked..........again log will start changing...........and it seems that job is not doing anything.........we also sometimes get confused with this issue...........but nothing to worry about.....
Regards,
Debjani.......

Process Chain delayed to load

hi Experts,
am using BI 7.0(BW 7.0) my process chain is scheduled to run daily morning, it works fine, sometimes it gets delay in loading the data for even a longtime or stucks in the early start itself.Please help me in having a solution so that the process chain doesn't stuck or how to reduce the loadtime.
thanks
Sankaresh S

One reason could be the system resources are maximum utilized when this process chain was about to start. If it's the case you can change the schedule time.
You can check the Monitor screen - details tab for further information when it gets hanged and also check the job log / status of job in SM37 / any dumps recorded in ST22 ?

Reducing the delay time.

hi sapgurus,
   I am facing problem with the process chains.
1.The job BI_PROCESS_TRIGGER is taking a delay time of 1445 secs.
   The chain is getting after 15 minutes of scheduled time,
I checked in sm37,its showing delay.
2. BI_PROCESS_LOADING taking a delay of 1203 secs.
   The infopackages are getting started after about 10 minutes,from previous job.
Can anybody help me to solve this problem.
thanks in advance.

Hi,
If the chain start, and the job is in delay mode, check in SM50 that you have a BGD/BCT Work Process to take the request from the process chain.
May be you dont have free Work Process to take the request.After waiting for 10 or 15 minutes, the system get a free work process which starts processing the process chain.

Tables related to SM37

Hi All,
I need the table name holding the field values of Duration and Delay viewed via SM37.
I did check on the below tables, but no luck.
TBTCO
TBTCP
TBTCS
Thanks,
Karthik

Hi Karthik,
Try using the function mudule BP_JOB_EDITOR
You can also check - In se11 for table D010TAB enter the program name for SM37,ie SAPLBTCH,you'll get all the tables used for this transaction.
But i think if you use this FM you'll get values in the export structure TBTCSTRT from which you'll have to caluculate.
LASTSTRTTM - SDLSTRTTM (Latest Execution Time for Background Job - Planned start time for background Job)
I hope it helps.
Regards,
Shraddha

SPAM IMPORT_PROPER Phase delay

Hi all,
I have SOLMAN 7 EHP1 SP 24. in a windows host, with Oracle database.
I just update my spam to latest version 7.01/0041
then i need to updatethe component "ST-ICO 150_700" my actual SP is 24, i need get updated to patch 0025.
I add to the queue only this ST-ICO component.
In sm37 I see the job running without errors.
Job Spool Job Doc Job CreatedB Status Start date Start time Duration(sec.) Delay (sec.)
OCS_QUEUE_IMPORT 71360900 Active 19.01.2011 16:21:02 55.578 0
In spam in action log see that all previous phases was successfully finished, for example the latest phase was Import phase 'DDIC_ACTIVATION' .
but in IMPORT_PROPER Phase is taking more than 16 hours right now.
If is just a ST-iCO component why this take so long? Is normal this behaviour?
Where can i see detailed log of actions, or how can i ensure me that all is running ok?
Thanks a lot
update: in tp system log i see so many lines with this:
WARNING: \\solman\sapmnt\trans\tmp\SAPKK-1507PINSTPL.SSM is already in use (17060), I'm waiting 2 sec (20110120064823). My nam
WARNING: \\solman\sapmnt\trans\tmp\SAPKK-1507PINSTPL.SSM is already in use (17070), I'm waiting 2 sec (20110120064853). My nam
WARNING: \\solman\sapmnt\trans\tmp\SAPKK-1507PINSTPL.SSM is already in use (17080), I'm waiting 4 sec (20110120064921). My nam
WARNING: \\solman\sapmnt\trans\tmp\SAPKK-1507PINSTPL.SSM is already in use (17090), I'm waiting 2 sec (20110120064952). My nam
WARNING: \\solman\sapmnt\trans\tmp\SAPKK-1507PINSTPL.SSM is already in use (17100), I'm waiting 1 sec (20110120065010). My nam
WARNING: \\solman\sapmnt\trans\tmp\SAPKK-1507PINSTPL.SSM is already in use (17110), I'm waiting 3 sec (20110120065038). My nam
WARNING: \\solman\sapmnt\trans\tmp\SAPKK-1507PINSTPL.SSM is already in use (17120), I'm waiting 5 sec (20110120065117). My nam
WARNING: \\solman\sapmnt\trans\tmp\SAPKK-1507PINSTPL.SSM is already in use (17130), I'm waiting 1 sec (20110120065153). My nam
WARNING: \\solman\sapmnt\trans\tmp\SAPKK-1507PINSTPL.SSM is already in use (17140), I'm waiting 4 sec (20110120065222). My nam
WARNING: \\solman\sapmnt\trans\tmp\SAPKK-1507PINSTPL.SSM is already in use (17150), I'm waiting 1 sec (20110120065259). My nam
WARNING: \\solman\sapmnt\trans\tmp\SAPKK-1507PINSTPL.SSM is already in use (17160), I'm waiting 2 sec (20110120065332). My nam
WARNING: \\solman\sapmnt\trans\tmp\SAPKK-1507PINSTPL.SSM is already in use (17170), I'm waiting 3 sec (20110120065404). My nam
I wiil try restart the import .
Edited by: Manfred Kaufmann on Jan 20, 2011 1:55 PM

Thanks for the tip. This seemed to resolve my issue.

Backghround process queuing - and delays

Hi,
I'm hoping someone could spare a little time to explain a couple of things to me please.
Each of our servers has a number of work processes, including, for example, 5 BKD sessions.
When a user submits a job to the background it sometimes sits there for ages in the queue with the delay (sec) amount steadily increasing - sometimes up to 4000 - 5000 seconds, even though there are completely free BKD sessions. The users often ask me what is happening to their job and I can't explain it.
As far as I know the submitted jobs are independent of any of the other runing jobs.
So I have 3 main questions please.
1. Can you see if a particular job is submitted to a particular work process - and if you can , can you change this to a free one?
2. Why do these delays happen? As far as I know there is no coding or planned scheduling like this.
3. Why do some jobs get immediately set to a BKD work processes whereas others just sit there with the time delay building up?
I've tried using SAP Help to understand and I can't find any detailed explanation of how this works.
Any help/explanation, much appreciated!!!
Many thanks
Mark Jeffery

Hello Mark,
Most of the questions have been very well answered already however here are a few inputs from my side:
1. Can you see if a particular job is submitted to a particular work process - and if you can , can you change this to a free one?
A job or a program will be assigned to a particular work process only when it gets started not when it is in delayed status. However there is a catch to it. Suppose you create a job will class B or C and it is going in delay and there are processes available for class A jobs then you can change the priority of the job and then when back ground job scheduler runs next time it will consider this job with the new job class and and if at that point of time there are free processes reserved for class A jobs then it will start the job immediately without any further delays.
2. Why do these delays happen? As far as I know there is no coding or planned scheduling like this.
The reasons for the delay have been explained already by Kunal. But there is plan to the scheduling. You see there is some thing called background job scheduler that runs every minute (depends on profile parameter rdisp/btctime) which scans all the jobs waiting in the queue and then triggers the job on various parameters like available background work processes, job class, planned job run time etc.
3. Why do some jobs get immediately set to a BKD work processes whereas others just sit there with the time delay building up?
As explained already scheduler works on certain paramerters.Basically delays will happen when the queue is very long.
To solve this you can either temporarily increase the number off BG work processes or reduce number of class A processes. it depends on situation at hand. From my personal experience I can say the best approach for avoiding these delays can be based on finding out what is causing the delays. Look at at entire queue of released and activve jobs in Sm37 and figure out what is causing the issue.
Additionally if you want to optimize back ground processing I would recommend implementing the suggestions of OSS note 923228. I had done it last year and it really helped us somewhat in achieving better performance. Actually delays can happen even if all the back ground processes are free since next lot of jobs won't get triggered unless batch scheduler does not run again. This OSS note will help you in averting such situations. very handy if you a long list of jobs which have a run time of few seconds.
With Regards.
Ruchit Khushu

Job RDDIMPDP Delayed

Hello !!!!
I have a little ploblem in my ECC 6.0 ehp 4 systems in hp-ux
my system consist of "2 dbs, 2 ascss, 5aps"
ploblem is rddimpdp Delayed about 40~60 seconds
when I saw the job logs in SM37, rddimpdp running in random servers every time
so, I shutting down 4ap servers, and retry cts & using tp command
It works normally, (it means no delayed rddimpdp)
but, when I start 2 servers, rddimpdp delayed.
I checked permission of filesystem, timezone, scheduled in client 000
can you tell me, what is the reason?
what can i do for fixing job delayed?
please help me !!
sleep does not come out of curiosity !!
yours truly.

Hi,
can you tell me, what is the reason?
RDDIMPDP is actually starts all the transport steps comes from tp via different separate batch work processes. it all depends on the parameter rdisp/wp_no_btc, default you should have atleast 2 batch processes to start. if you have more work processes then obvious RDDIMPDP finish very fast.
If your system has more application instances, we can distribute to different work processes from all the application servers via message server. check the note
I could say, this is the reason, when you run on 4app servers, the job quite fast than the 2 app servers.
other work arround i could suggest to check the batch work process count or set the operation modes to get more batch work processes.
or check out other suggestions here 386646 - Event-controlled jobs start with a delay
Thanks
Jansi

Data loading delay

Hi Friends.,
Shall i have an answer for one error,
The Issue is: Every day i load to one info cube, whatever the cube it is, it takes 2 Hours for every load, but once it has taken 5 Hours, what might be the reason? just confusing with that, can anybody let me clarify !!!!
Regards.,
Balaji Reddy K.

Reddy,
1. Is the time taken for loading to PSA or to load from PSA to cube ? if it is to oad to PSA then uaually the problem lies at the extractor
2. If it is loading to the cube.. then check out if statistics are being maintained for the cube and they would give an accurate picture of where the dataload is taking up most time.
Do an SQL trace during the data load and if you find a lot of aster Data Lookups .. make sure that master data is loaded and if there are a lot of looups to Table NRIV check if number range buffering is on so that dim IDs get generated faster
Check if the data load happens fast if you drop any existing indexes...
Are you loading any agregates after the data load ? check fi th aggregates are necessary or if they have been enabled for delta loads..
If you have indexes active and there is a huge data loa , depending on the index , the data load can get delayed..
If the cube is not compressd , some times the data load can get delayed..
Also when the data load is going on check in SM50 and SM37 to see if the jobs are active - this means that the data load is active from both sides...
Always update the statistics for the cube before the load and ater the load , this helps in deciphering the time it takes for the data load... after activating the statistics .. check table RSDDSTAT or the standard reports available as part of BW tecnical content..
Hope it helps..
Arun
Assgn points if helpful

Delays in sm37

Similar Messages

Maybe you are looking for