cancel
Turn on suggestions
Auto-suggest helps you quickly narrow down your search results by suggesting possible matches as you type.
Showing results for 
Search instead for 
Did you mean: 

Info Package execution very slow.

Former Member

on ‎10-09-2014 12:21 PM

0 Kudos

Hello,

It's a random problem that we are facing but it now it is happening quite frequently.

In our process chains, serial execution of 10 Info-Packages takes places and each one of them takes on an average 20 minutes max. to fetch data.

Lately, what has happened that one of them took 5 hours to fetch the data, after that two more took 1 hour 20 and 1 hour 35 minutes respectively.

It has happened one or two times previous weeks also.


Please let me know what steps/sequence to follow/monitor with details during IP getting stuck with 0 of 0 records. What can I suggest to the basis team for the same ? Where can I check IP logs/Job Overview ?

Regards,

Max

Show replies
Show replies
Answer

Accepted Solutions (0)

Answers (4)

Answers (4)

Former Member
‎10-14-2014 6:17 AM
0 Kudos

Good question! We had a similar issue in the past, most probable reason we found out is due to system backup running in parallel or it might be scheduled to run on specific days, ask Basis team to check system and organize backup schedule days.

You may also check BW system backup trigger times within past few weeks. Tcode DB02 will provide more info.

Hope this helps.

Show replies
Show replies
Former Member
‎10-14-2014 6:26 AM
0 Kudos

Hi WIDYL,

I am trying to get in touch with Basis team but there availability is also an issue. Data source is direct access , accessing the operational data, we have 10 info packages for each month and at around 10:20:00 pm, any info package ready to run during that point in time, goes on waiting from 2 - 5 hours . I have checked this datasource in ECC system and it doesn't exists and it's origin is BI system only,

Through customer exit it is getting the data. I have some difficulty understanding the system as I am new to this support project and documentation etc. is not there.

Regards,

Max

Former Member
‎10-28-2014 3:11 AM
0 Kudos

Hi Max,

First of all don't panic its a silly system where we have to give instruction in  order to perform something. You need to create documentation if possible to understand the system. If its a support environment its a must to create these.

Try to replicate the same scenario in Quality system if possible, Basis needs involvement as the data was retrieved directly from source, they need to trace if there are timing issues or background job processing issue.

Check info-package wait times as well.

Is this info packages are executed sequentially or in parallel?

Kind regards.

Former Member
‎10-29-2014 6:55 AM
0 Kudos

Hi WIDYL,

I will try to replicate the scenario in quality system, Should I replicate and run the process chain same time as it run in production system ?, They are running sequentially.

thanks and regards,

Max

Former Member
‎10-09-2014 1:03 PM
0 Kudos

Here is the partial log from SM37,

Please note after 10 : 22 : 44 PM , processes started again after 3:18:XX AM..???? It went idle ?

10/8/201410:22:44 PMJob started00516S
10/8/201410:22:44 PMStep 001 started (program SBIE0001, variant &0000000007558, user ID ALEREMOTE)00550S
10/8/201410:22:44 PMAsynchronous transmission of info IDoc 2 in task 0001 (0 parallel tasks)R3413S
10/8/201410:22:44 PMDATASOURCE = ZBW*******R3299S
10/8/201410:22:44 PMRLOGSYS    = ****************R3299S
10/8/201410:22:44 PMREQUNR     = RE***********************EJNVC0RR3299S
10/8/201410:22:44 PMUPDMODE    = FR3299S
10/8/201410:22:44 PMLANGUAGES  = *R3299S
10/8/201410:22:44 PM*************************************************************************R848S
10/8/201410:22:44 PM*          Current Values for Selected Profile Parameters               *R849S
10/8/201410:22:44 PM*************************************************************************R848S
10/8/201410:22:44 PM* abap/heap_area_nondia......... 7000000000                              *R850S
10/8/201410:22:44 PM* abap/heap_area_total.......... 10000000000                             *R850S
10/8/201410:22:44 PM* abap/heaplimit................ 40000000                                *R850S
10/8/201410:22:44 PM* zcsa/installed_languages...... DEF                                     *R850S
10/8/201410:22:44 PM* zcsa/system_language.......... E                                       *R850S
10/8/201410:22:44 PM* ztta/max_memreq_MB............ 2047                                    *R850S
10/8/201410:22:44 PM* ztta/roll_area................ 15000000                                *R850S
10/8/201410:22:44 PM* ztta/roll_extension........... 2000000000                              *R850S
10/8/201410:22:44 PM*************************************************************************R848S
10/9/20143:18:07 AMCall customer enhancement EXIT_SAPLRSAP_001 (CMOD) with 12.987 recordsR3407S
10/9/20143:18:07 AMResult of customer enhancement: 12.987 recordsR3408S
10/9/20143:18:07 AMPSA=1 USING SMQS SCHEDULER / IF [tRFC=ON] STARTING qRFC ELSE STARTING SAPIR3299S
10/9/20143:18:07 AMAsynchronous send of data package 1 in task 0002 (1 parallel tasks)R3409S
10/9/20143:18:07 AMtRFC: Data Package = 0, TID = , Duration = 00:00:00, ARFCSTATE =R338S
10/9/20143:18:07 AMtRFC: Start = 08.10.2014 22:22:44, End = 08.10.2014 22:22:44R339S
10/9/20143:18:08 AMAsynchronous transmission of info IDoc 3 in task 0003 (1 parallel tasks)R3413S
10/9/20143:18:08 AMCall customer enhancement EXIT_SAPLRSAP_001 (CMOD) with 12.987 recordsR3407S
10/9/20143:18:08 AMResult of customer enhancement: 12.987 recordsR3408S
10/9/20143:18:09 AMtRFC: Data Package = 0, TID = , Duration = 00:00:01, ARFCSTATE =R338S
10/9/20143:18:09 AMtRFC: Start = 09.10.2014 03:18:08, End = 09.10.2014 03:18:09R339S
10/9/20143:18:09 AMAsynchronous send of data package 2 in task 0004 (1 parallel tasks)R3409S
10/9/20143:18:09 AMtRFC: Data Package = 1, TID = 995FD19B3A80543636AF0632, Duration = 00:00:02, ARFCSTATE = SENDEDR338S
10/9/20143:18:09 AMtRFC: Start = 09.10.2014 03:18:07, End = 09.10.2014 03:18:09R339S
10/9/20143:18:10 AMtRFC: Data Package = 2, TID = 995FD19B18D0543636B1045F, Duration = 00:00:01, ARFCSTATE = RECORDEDR338S
10/9/20143:18:10 AMtRFC: Start = 09.10.2014 03:18:09, End = 09.10.2014 03:18:10R339S
10/9/20143:18:10 AMAsynchronous transmission of info IDoc 4 in task 0005 (0 parallel tasks)R3413S
10/9/20143:18:10 AMCall customer enhancement EXIT_SAPLRSAP_001 (CMOD) with 12.987 recordsR3407S
10/9/20143:18:10 AMResult of customer enhancement: 12.987 recordsR3408S
10/9/20143:18:10 AMAsynchronous send of data package 3 in task 0006 (1 parallel tasks)R3409S
10/9/20143:18:10 AMtRFC: Data Package = 0, TID = , Duration = 00:00:00, ARFCSTATE =R338S
10/9/20143:18:10 AMtRFC: Start = 09.10.2014 03:18:10, End = 09.10.2014 03:18:10R339S
10/9/20143:18:11 AMtRFC: Data Package = 3, TID = 995FD19B18D0543636B20460, Duration = 00:00:01, ARFCSTATE = RECORDEDR338S
10/9/20143:18:11 AMtRFC: Start = 09.10.2014 03:18:10, End = 09.10.2014 03:18:11R339S
10/9/20143:18:11 AMAsynchronous transmission of info IDoc 5 in task 0007 (0 parallel tasks)R3413S
Show replies
Show replies
karthik_vasudevan
Active Contributor
‎10-09-2014 1:20 PM
0 Kudos

Hi Max

This has to be notified to Basis team as well when you are analysing further.

It seems to be a full load and it hits CMOD for some logics. Hope you haven't changed anything in CMOD code to have the logic very complex.

Try reducing the data packet size if the code is complex by the below path. IP -> Schedule(Menu) -> Data transfer for source system -> reduce the data packet size and increase the maximum number of dialog processes of sending data.

Try this change the load the data again. Meanwhile keep basis informed and let them also analyse if there is no change in code part

Regards

Karthik

Former Member
‎10-09-2014 1:33 PM
0 Kudos

Hi Karthik

I haven't changed any code, I am new to this support project so I was just analyzing this loading that took time . I will monitor this for 2-3 days more and if this trend continues, I will ask the concerned basis team for the same.

Thanks a lot.

Regards,

Max

Former Member
‎10-09-2014 1:47 PM
0 Kudos

maybe lack of BGD process responsible for this ? Where can it be, in the Source of BI system itself ?

RamanKorrapati
Active Contributor
‎10-09-2014 1:54 PM
0 Kudos

Better to change your info pack trigger time and see the result.

it may run without much delay.

If you see your logs you will find below info.

Before starting of CMOD code, it tooks 4 to 5hrs to call CMOD logic.

So in that gap your ECC system will be overloaded.

so ask basis team and trace the ECC load in that time period.

You may get some idea about.

About BGDs, you can see at SM50 or SM51.

if you have very less bgd then ask basis team to increase more as per the needs.

Try to schedule your full info pack load where less burden on ecc/bw servers.

karthik_vasudevan
Active Contributor
‎10-09-2014 1:58 PM
0 Kudos

Hi Max

When the infopackage is triggered, it needs a background process in BI side and only when that job is completed, the extraction job will be initiated in source system.

In source system, a background process starts referring to your infopackage and in your case, it has started around 10 pm. Later when the data is passed from ECC to BI as TRFCs, it required dialog process for each TRFC

But in your case, the job has started and it got stopped before extracting the data. So this not related to lack of process, I believe.

As you have mentioned, monitor this for few days when the job is actually running. Inform Basis team to track the job when its actually running in the system. They will know how to check what is the job actually doing. meaning it can try to hit a database or something like that

Regards

Karthik

Former Member
‎10-10-2014 4:56 AM
0 Kudos

Hi Karthik,

One of the IPs loading again got stuck and it hasn't call the CMOD code yet and it's still waiting, But on checking on SM50 on BI system I can see some activity in DIA .

Please let me know what further things I can do like if manual process of IDOCs etc I can try or any other thing ?

Thanks and regards,

Max

10/9/201410:28:20 PMJob started00516S
10/9/201410:28:20 PMStep 001 started (program SBIE0001, variant &0000000007570, user ID ALEREMOTE)00550S
10/9/201410:28:20 PMAsynchronous transmission of info IDoc 2 in task 0001 (0 parallel tasks)R3413S
10/9/201410:28:20 PMDATASOURCE = Z**_TRANSR3299S
10/9/201410:28:20 PMRLOGSYS    = *****************R3299S
10/9/201410:28:20 PMREQUNR     = R***********************R3299S
10/9/201410:28:20 PMUPDMODE    = FR3299S
10/9/201410:28:20 PMLANGUAGES  = *R3299S
10/9/201410:28:20 PM*************************************************************************R848S
10/9/201410:28:20 PM*          Current Values for Selected Profile Parameters               *R849S
10/9/201410:28:20 PM*************************************************************************R848S
10/9/201410:28:20 PM* abap/heap_area_nondia......... 7000000000                              *R850S
10/9/201410:28:20 PM* abap/heap_area_total.......... 10000000000                             *R850S
10/9/201410:28:20 PM* abap/heaplimit................ 40000000                                *R850S
10/9/201410:28:20 PM* zcsa/installed_languages...... DEF                                     *R850S
10/9/201410:28:20 PM* zcsa/system_language.......... E                                       *R850S
10/9/201410:28:20 PM* ztta/max_memreq_MB............ 2047                                    *R850S
10/9/201410:28:20 PM* ztta/roll_area................ 15000000                                *R850S
10/9/201410:28:20 PM* ztta/roll_extension........... 2000000000                              *R850S
10/9/201410:28:20 PM*************************************************************************R848S
Former Member
‎10-10-2014 4:57 AM
0 Kudos

Hi Raman,

No use of scheduling to different time because there are like 10 IPs and this happens to any one of them at any point in time.

Please see my latest reply and if you can find anything it.

You said I should get the BGDs checked for in SM50. I have checked in SM50 on BI side and there are 12..with 2 running and rest on waiting so I guess that's not an issue from BI side.

But on the ECC side, there are only 5 BGDs and one is in waiting and rest are full loaded. So should I get the BGDs checked for ECC first ?

Thanks a lot

Regards,

Max

Former Member
‎10-10-2014 6:51 AM
0 Kudos

I have checked in SM50 on BI side and there are 12..with 2 running and rest on waiting so I guess that's not an issue from BI side.

But on the ECC side, there are only 5 BGDs and one is in waiting and rest are full loaded. So should I get the BGDs checked for ECC first ?

karthik_vasudevan
Active Contributor
‎10-10-2014 7:46 AM
0 Kudos

Hi Max

Please check what are those jobs which are currently active in ECC..That may be related to your other infopackages. but yes,you should check this and make sure you have the BGD processes free.

Regards

Karthik

Former Member
‎10-10-2014 8:02 AM
0 Kudos

Hi Karthik,

I have done some checking of logs and almost around after 10:20:00 - 10:50:00 PM this problem starts with any IP that is ready to start in this time period. I guess it's the overloading of ECC system or something like that which Basis team should check for BGD processes etc.

Thanks and regards,

Max

karthik_vasudevan
Active Contributor
‎10-10-2014 8:16 AM
0 Kudos

Yes Max. You are absolutely correct. Please inform basis team to check these job logs and investigate further.

You may also recommend them to increase the background processes in ECC system. To confirm its an issue with unavailablity of resource,you can also try running only one infopackage at a time to see if thats working fine. Just to make a point that it works when there are free resources.

Hope this helps

Regards

Karthik

Former Member
‎10-13-2014 12:40 PM
0 Kudos

Hi Karthik,

I have been trying to get some help from Basis team but they have some other high priority tasks to do and so I am left all by myself for it.


Please let me know how I can trace in source system about the workload on these periods/time.

karthik_vasudevan
Active Contributor
‎10-13-2014 1:29 PM
0 Kudos

Hi Max

This is not an easy task to trace and identify the reason.

We can get the tcode to do tracing and we can do that as well. But we wont be able to understand it properly. Unless we understand what is happening in the background, how could we resolve the issue.

ST12 is the tcode to initiate the trace, but I am not how useful it is when we (BI consultant) do it. I am not sure how to use this effectively as well.

Did you try running only one infopackage during different timing?

Regards

Karthik

Former Member
‎10-14-2014 5:01 AM
0 Kudos

Hi Karthik

I have checked using SM50 in ECC, BGD processes is not an issue.

I have done step by step analysis in IP, Posting some screenshots, if something can be extracted from that . Please see :

Job Overview is same as the above logs I had pasted earlier .

Checking in SM58, it's empty

Source system is pointing to BI system only.

I have checked the datasource and Info Package, In extraction the adapter is access SAP data through service API.

Where should the data be coming from ? I guess it's a real time data access.

karthik_vasudevan
Active Contributor
‎10-09-2014 1:00 PM
0 Kudos

Hi Max

There are various parameters that we need to check for the performance of infopackage. On the other hand, check the data volume as well.

The problem might be with tRFCs getting piled up in source system or the BGD process availability in BW system or the dialog processes available in source system to transact RFCs

When you go to the process monitor when the infopackage is running, click on Environment in the menu bar --> Job Overview in data warehouse and source system; Transact RFC in source system.

Compare the job logs when it ran fast and current log. you will get an idea if its an issue with jobs

If its a RFC issue, you can just go to SM58 in source system and execute the TRFCs which are in recorded mode by clicking F6.

Regards

Karthik

Show replies
Show replies
Former Member
‎10-09-2014 1:04 PM
0 Kudos

Please check the updated log

RamanKorrapati
Active Contributor
‎10-09-2014 1:00 PM
0 Kudos

Hi,

info pack job log you can see at ECC side if your loading data from ECC.

check at info pack monitor screen and details tab.

IP getting 0 of 0 records - mean source don't have data. in such case we need to turns that status to green by using below option.

Go to your info pack, Menu scheduler-->Color of traffic light for empty request, here yellow may be selected, select green and save it.

Next time onwards whenever IP get 0 records it turns to green without much delay.

Thanks

Show replies
Show replies
Former Member
‎10-09-2014 1:04 PM
0 Kudos

Please check the updated log

Ask a Question