on 02-14-2005 7:12 AM
Hi,
I have a situation where in we are loading Sales Data for the Year. The loading is going fine, full of green light per data package. Until a yellow one comes along and was skipped since successive packets turns green again.
Question is how could I reprocess this yellow data packages to retry and reload? The only option when you right click is "HELP FOR NODE".
--Jkyle
Hi Jkyle,
There are 2 possibilties
Reload the request from the PSA back to the cube OR
You could go to the Monitor of the load and make the request red. Then go to Details -> Processing (right click) and select the manual update.
Hope this works.
You must be a registered user to add a comment. If you've already registered, sign in. Otherwise, register and sign in.
Hi Jkyle,
Another possibility is to split the data by giving the data selection in the infopackage, say 4 infopackages for 3 months for that year like Jan-Mar, Apr-June and so on..
You must be a registered user to add a comment. If you've already registered, sign in. Otherwise, register and sign in.
Hi Dinesh,
I have a question on splitting. What if I have an existing INIT DELTA Request, can I request a <b>Full Load</b> with parameters in the Info Package as what you have mentioned? Won't it ruin my INIT DELTA Request? What about my DELTA? Will it remain intact? My worry is that I might redo a <b>REINIT</b> if I do this the wrong way...
Can I execute them in Parallel? (I dont know if it can be since I only have one Info Package.)
Thanks for your asnwers...
--Jkyle
Jkyle,
if you have an existing init, you can perform full loads (but only if you are using a cube as destination provider, not an ODS; in this latter case you have to perform your full from another source than the init one) and no problem for your delta queue....you can execute several loading in parallel (and you can easily use the same infopackage or you can create different ones with no problem), but you have to be sure that no system bottleneck will be caused...
Bye,
Roberto
Hi Jkyle,
Since the issue is table space issue and you are not using PSA. You can deactivate the aggregates for some of the cubes which would provide you extra tablespace. After the cube is loaded you can then activate them.
Hope it helps.
You must be a registered user to add a comment. If you've already registered, sign in. Otherwise, register and sign in.
This depends a bit on the error. If this is a delta package you could request a repeat from your source system. If it was a full upload you could schedule it again. If the data is completely in the PSA you can delete the request from the cube, fix the error (either in PSA or your update/transfer rules) and reconstruct the cube. Sometimes you'll find a button under the status in your monitor that says Manually post the package (or similar).
What error did you get?
Best regards
Dirk
You must be a registered user to add a comment. If you've already registered, sign in. Otherwise, register and sign in.
Hi Dirk and Dinesh,
Thanks for the reply... I''l give your further info:
1.) The error was - say tables space issue. Its easy, BASIS guys can re-adjust tables spaces...
2.) We dont use PSA in our data loads to increase performance and save space.
3.) We cannot afford a reload of 50,000,000.00 ++ records. It took us days to load this (Sales data for 1 year).
4.) The request was INIT Delta Update. We used the "How to Minimize Downtime During INIT" PDF document so we are extracting from our Flash Copy so that normal R/3 operations will not be affected...
Based from your feedback, theres no option but to reload???
Thanks and hoping for your help...
--Jkyle
Hi Roberto,
Right you are. I have not used PSA since it consumes space as well as it decreases the efficiency of the extraction process. So that means, this is the trade off? I cannot repost errors if I did not use PSA?
Is that the same also if I wanted to resend a request using the same request number? It wont work without a PSA?
Thanks,
--Jkyle
User | Count |
---|---|
83 | |
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.