Production loads..

[#1]

Hi all,

THis is production issue.. Please help me on urgent basis.

I have daily delta loads for 2lis_02_scl for 2 cubes

Some one has changed option from data target to PSA only .

We found today and from 04.07.2007 to 11.07.2007 data records are in PSA and not uploaded to cube. I have started one by one delta load(means 4th, 5th, 6th etc..) from PSA to Cube using Process Manually. I found one red request also there in my cube. Will it causes for slow.?

To check the status - I got a dialog with YES/NO/CANCEL

02223 IDocs searched through for date 04.07.2007.

00005 of the IDocs belonged to this request.

00005 IDocs found for this request up until now.

Should the search be continued for a further day ?

– what is this mean?

--IF psa (rsa15) is in red , can I run "PROCESS MANUALLY" again, does it allow?

OR

Shall I delete all delta loads from 04.07.2007 to 11.07.2007 and do delta with 0calday selections? Please advise.

THanks in advance.

Cheers,

Sri.

[Former Member · Accepted solution]

Hello Sri,

Dont delete the PSA data, If u have also it wont effort ur load ,

plz check tht the any other load had happened after the red request...

if the next load happened and tht was sucess plz check the type of load in <b>RSMO</b>with date selection ....

if it is repeat delta the no need to worry abt the data, every may be fine ....

Regards

Sandy

[Former Member]

Hi All,

I have a doubt..

In daily deltas, for eaxmple take dates 15/16/17/18/19/20 loads.

If 15/16/17 are failed and only 19/20 are success. How can we get the data of 15/16/17? what is the best procedure?

My view is:

I delate all 15 to 20th data and run delta load, Will it pick all 5 days delta records?

advise..

Cheers,

Sri.

[Former Member]

Your 19 th delta soul get all deltas, I don't think you will get success deltas..?

If your 5 deltas got failed then only you can get..?

Have you find the reason why these deltas got failed..?

MSR

[Former Member]

Hi again,

it all depends WHERE the load failed!

- if your PSA request has all the records and the error arose during the PSA to target load, then you can fix the issue, delete the request from your target(s).

At this stage (the request is only posted in the PSA you can set the QM status to green. Finally perform a load again from PSA to target.

- if you PSA request is red because of a problem between your source system and your BW PSA, and it is incomplete then you have to rmove it form all datatargets, KEEP the QM status to RED and then request a repeat.

If you delete the 15th to the 20th as you said, your delta data will be gone and the next delta won't pick them up!!! Then you are entering into big troubles because you'll have to perform s SETUP on these datasources; indeed LIS datasources cannot be loaded via full load and with selections.....

It is your call.....

hope this helps...

Olivier.

[Former Member]

HI Srini,

Thanks for your info.

Load failed because of Invalied date. I can fix it.

<b>present status is in SYS :</b>

04.07.2007 to 11.07..2007 are in PSA.

Now, 04.07.2007 is running, Once this will finish , I can start 5th 6th..etc from PSA to CUBE.

Finally I delete 11.07.2007 th request from target. I can fix date error. and do repeat delta from PSA to CUBE.

Please correct if any thing is wrong in the above steps.

<b>And I have one more doubt:</b>

To check the status - I got a dialog with <b>YES/NO/CANCEL</b>

02223 IDocs searched through for date 04.07.2007.

00005 of the IDocs belonged to this request.

00005 IDocs found for this request up until now.

Should the search be continued for a further day ?

– what is this mean?

====

<b>To speed up loads, u said:</b>

delete indexes means -

Shall I do like this: 1,2,3.

1. click that delete indexes button(now) and

2.load data from PSA to CUbe

3. REpair Indexes (now)

Please advise

Cheers,

Srini.

[Former Member]

Hi

you can answer NO to the IDOC search prompt...

The INDEX deletion will only speed up the process if the number of records your are loading is very high respectively to your F fact table....

If your F fact table is 500'000 records and you're loading request of around 50'000 then deleting the indexes and reconstructing will take more time....

By the way, you can schedule your PSA loads to target in paralel in order to speed this process up...

hope this helps...

Olivier.

[Former Member]

Hi,

You are doing good in loading your PSA requests to the cubes; however check that you aren't loading a DELTA request having a NEXT request of DELTA REPEAT; indeed in that case, the DELTA REPEAT will suffice otherwise you'll get duplicated data posted in your targets.

Before troubleshooting IDOCs (I don't think you have an issue of that kind since your data is loaded in the PSA; except if the number requested don't agree with the number of records posted in your PSA); what is the reason of having a RED status??

Before doing anything (DO NOT DELETE ANY PSA REQUEST; DO NOT REQUEST DELTAS ANYMORE) you should first have a clear picture of the situation

only after, we'll advise

hope this helps...

Olivier.

[Former Member]

looks like this red request..is older than this 4th-11th date range..??

if yes..first try to repair this failed request..(if this isnt done..u might find requests are not available for reporting)

and then process manually the 4th-11th requests that are sitting in the PSA..

if psa is red..process manually is allowed..

i think u cant place new selections on the delta load..u might have to do a full load with selections..

cheers,

Vishvesh

[Former Member]

HI All,

Thanks for your answers.

1. All requests in PSA are from 4th to 11th only.

2. BUt requests are running very slow from PSA to cube. Why?

3. THis is 2.x , no repair request concept.

4. If I get load error, the PSA will be in red. I can run again from this PSA to CUBE , Please advise.

Cheers,

Sri.

[Former Member]

BUt requests are running very slow from PSA to cube. Why?

Delete the indexes before loading the data to cube..!

If I get load error, the PSA will be in red. I can run again from this PSA to CUBE ..?

Delete the bad request in the cube, Update manually before updating find out why the load got failed..?

Hope this will help you..?

MSR

[Former Member]

Hello Sri,

U can delete the failed request from the Data target and you can load again from the PSA ,

But plz check one why it have failed to load into ODS/Cube and correct to load again

Regards

Sandy