on 12-23-2008 5:07 AM
Hi Guru's,
I have doubt's about process chains can any body clarify my doubts please,
1. How we monitor process chains.
2. How we know that whether it is a Master data chain or transaction data chain.
3. If team lead ask to monitor some chains, what are the steps i have to take to monitor
those chains.
4. If you have any screen shots regarding to this please forward.
regards
sanath
Hi ,
1.You can monitor the process chains through RSPC1 ,RSPC,Se38 .. /ssa/bwt ..
and for monitoring the Ip .. check the job overview of Ip and the packet processing .. timings ..
Check the short dumps in st22 ,TRFC in sm58 .. if trfc are there then push it through F6 ..
iyou can check the job in sm37 ...>>> double click on job details >> take the PID and application server ID .. and check wether the insert ,delet and all is progessing or not ..
like this you can monitor the otehr stetp in chain .like rollup ..
2. you can distinguish between the chain s .. as Masterdata chains .. willl be having name end with _MD .. trasaction data chains will be having _TD ...
3. keep a sm37 session open to check the jobs ..status .. for perfromance checking you can check st04 .. there should be no lock waits .. st22 not may short dumps .. sm66 ..
check this link to find all kind of error ..
hope this helps ..
Regards ,
Shikha..
You must be a registered user to add a comment. If you've already registered, sign in. Otherwise, register and sign in.
Hi,
ere are simple and straight answers
1. How we monitor process chains.
Ans: SA38->/ssa/bwt->Process chains. Here you can se all chains with their status in Green, yellow and red.Here you cal also see their start time, end time, runtime, Source system, datssource, targets etc.
2. How we know that whether it is a Master data chain or transaction data chain.
Ans: You can see the targets. if it's an infooject only then it's a master data load and if it's cube, DSO then it's transaction data.
3. If team lead ask to monitor some chains, what are the steps i have to take to monitor
those chains.
Ans: In SA38->/ssa/bwt->Process chains give the chain names in the selection and monitor only these chains instead of all the chains. Also the background jobs executed by these chains can be viewed in SM37.
You must be a registered user to add a comment. If you've already registered, sign in. Otherwise, register and sign in.
Hi Sanath,
After executing Program /ssa/bwt in SE38 select Process chain analysis in the list to monitor process chians and then execute you can give from date and to date also .
In BI 7.0 When the list of process chains displays it will show all the local and process chains together. The chians which have tick mark are Main chains.Remaining are local chians of its main chian. You can filter only main chains by giving x in Set Filter with Main tab selected.
To Find a particular chain give the technical name in Find .Keep refreshing the cahins if any thing fails it turns to red.If an Infopackage is running in the process chian check wheteher IP is picking the records or not.
Hope it helps you...
Thanks,
J.Silpa
You must be a registered user to add a comment. If you've already registered, sign in. Otherwise, register and sign in.
Hi..........
Another way to check whether a chain is master data chain or transaction data chain.................U will only find attribute change run in Master data chain........
Regards,
Debjani......
You must be a registered user to add a comment. If you've already registered, sign in. Otherwise, register and sign in.
Hi,
1. How we monitor process chains. We can monitor our process chains in the process chain Log view. Here you can see them in green color if they run without any errors. If any error occurs then you can onserve that respective node in Red color.
2. How we know that whether it is a Master data chain or transaction data chain.
In Meta chains you can see the local process chians with the node name LOCAL i.e.., if any of the process chains node name appears as Local then it is known as Meta chain
3. If team lead ask to monitor some chains, what are the steps i have to take to monitor those chains.
To monitor the Process chains, folllow the below steps,
1. Start your PC in planning view, after starting it sucessfully,
2. Goto Log view of your PC and the clickon refresh.
3. Keeping refreshing the PC so that you will come to know its status
4. IF everything is green then your chain seems running without any issues.
5. Your failed node will appear in Red color and to see the error right clcik on it goto Maintain variant, there you can see the propable errors for failure of your PC.
I hope this helps you.
You must be a registered user to add a comment. If you've already registered, sign in. Otherwise, register and sign in.
Hi Sanath,
I will give the answers according to your Question number....
1. We monitor Process chains by using Transaction code SE38 give the Program name as /SSA/BWT. Then execute. There you can find the chains.If you click the chian name.You can find the chain.
We can also use Transaction code RSPC .Here we find the chains in Planning view.
2. Mater Data chains will be generally named by using MD in the chain name .For Ex: LSA_MD_DAILY_HR_00. For Transaction Data chains will be generally named by using TD in the chain name .
When you see the Info Packages in Master Data chain you can find the Master data symbol. In the same way when you check the Info Packages in Transaction Data chain you can find the Transaction data symbol.
3.When team lead ask to monitor some chains Use SE38,RSMO,SM37,
SM37----Transaction Code for monitoring the jobs
RSMO--Transaction Code for monitoring for monitoring the Info Packages
When the chain fails it will turn to Red.Then we have to check the error and solve it.
When the chain is running it will be in Yellow .These chains we have to keep monitoring.
When the chains are completed successfully it will be in green
If any further queries please reply
Please give points if you are satisfied with the answer.
Thanks & Regards,
J.Silpa
You must be a registered user to add a comment. If you've already registered, sign in. Otherwise, register and sign in.
Hi
Discussed so many times.
How we monitor process chains.
***
Monitor the process chains through RSPC1 ,RSPC,RSPCM, Se38 .. /ssa/bwt ..*
*Check the short dumps in st22 ,TRFC in sm58 .. if trfc are there then push it through F6 ..*
*Check the job in sm37 .
How we know that whether it is a Master data chain or transaction data chain.
You will know it by seeing its dependent processes. If it contains all Info Objects like 0material, 0customer, 0 plant with attributes, text then its a MASTER data Chain. If it is loading the 0amount, 0debit, Keyfigs then its a Tran data chain ...
If team lead ask to monitor some chains, what are the steps i have to take to monitor
those chains.
SM37 to check the jobs/status --ST 04 perfromance checking --ST21/22 Short dumps --SM58/58/66 FOR QRFC
If you have any screen shots regarding to this please forward.
for the errors you have many posts discussed already
Process chain Monitoring Activities:
Q1 In case a chain or load fails what is the reporting stucture you follow
before correction,general steps during correction(from a process point of view)
and post correction what is the reporting matrix(please probide names and mail ids of people you inform to)?
A1 The BW on-call person generally checks the system as scheduled time.. Additionally, if the chain fails at certain points as defined in the Batch Monitor, whoever is on-call for BW is paged. Emails are also sent to them and anyone else specified in the Batch Monitor. We may also (but not necessarily) recieve a failure email message from the chain in BWP. At that point, we assess the problem and are normally able to get everything going before any deadlines (see A4 below). The reporting of the error largely depends upon the type of error. For example, if it was "Sending packages from OLTP to BW lead to errors", we just follow the directions (you have already been given) and resume the chain. If it fails due to system-type errors, we inform Basis, our management, the rest of the BW team and usually Rob. If it fails due to a master data problem, we ususally inform our management and the appropriate functional person so that they can check on it and aid in resolving.
Q2 Is load and chain corrections done in the production system?Do you inform anybody before solving the problem?
A2 Generally speaking, we do not need to change anything like infopkgs or chains to correct problems in BWP. We usually only need to run something again or to run something manually, but not change anything that you would normally transport. Se A1 and A4 regarding informing others.
Q3 Do you do corrections after all the loads have finished if not what is the processs followed?
A3 See A1.
Q4 Is there a reporting structure in case all loads are successful on a daily or weekly basis?If yes whom do you inform?
A4 As mentioned in our meeting, we have success/failure messages set up in the chains. You should receive messages Mon-Fri from BWP upon successful completion for master data, GL, AP, PCA, COPA, AR, Inventory, HR, and CWT. Once we see that everything is complete, we send an email to our management and copy the BW team stating that everything is complete. Or, if there was some problem that delayed the chain in any area, we report that the chains are not complete, ususally indicate the area and why the delay and also indicate about when we estimate that everything will be complete. We generally send this message prior to 7:30 am CDT so the information is known for the morning meeting. All loads are generally complete before 7 am CDT. Our current deadline is noon CDT. If there is a significant delay, we will also add a message to SM02 stating the area and estimated availability so that users will be aware.
Q5 How do you re-schedule a process chain in case of failure?By maintinig date and time in variant or is any other process followed?
A5 You need to be careful on this. You will more than likely need to resume a failed chain within it at the point of the failure or just beyond the failure, depending on what the failure was and what you did to correct it. If you reschedule the chain at that point, you will mess things up. Now, there are times when the start of the job failed in SM37. This means that the chain did not even start (See E1 below). In that case, you copy the failed job (in sm37) and start it immediately. All process chains and mat_plant_full are scheduled to start via jobs in SM37 - not directly from the Start process in the chains.
Q6 Do you send out reports on process chain performance on a predefined time line?like monthly or quarterly?
A6 No, there are no reports sent out. We maintain the document "BWP Load Times and Chain Information.xls". When a new area (like HR, COPA, etc.) is added to the daily chains in production, we generally let them run a week or so, then we update the document with the appropriate timing information/changes. Also, if we make performance improvements to any of the chains or if we notice that the chains are consistently completing earlier/later than the previous "normal", we will update that information as well.
Q7 Can you please out line the steps followed in the ticketing system with regards to process chains only?
A7 Because we are paged and emailed directly from BWP for problems and, additionally, since the on-call person generally checks the system time, we are not notified by the help desk nor are any tickets entered. As described above, we correct the problem or report it to the appropriate parties if it is not in our domain. If Basis assistance is needed, we will request a Basis ticket through basis team If functional assistance is required, we do not generally open a ticket.
Note: Again, please remember that if you "look" at the process chain while they are running using RSPC (and some times RSMO), an error may be reported falsely (i.e. report up red) or the chain could just stop. Would recommend checking the process of chains, if need be, by looking at sm37 on BWP and on P01 until chains complete.
Some other errors that might help:
E1 We have been getting the "NO_MORE_SPACE: Can not create MTE: no more space: no more MESSAGE_CONTAINER slots available" available message sometimes when our jobs starting jobs fire in SM37. In this case, simply copy the failed job and start it immediately. I believe we reported this to Basis a while back. I have asked for a new basis ticket for this.
E2 If you get the error "Missing data packages for PSA Table.", follow the same instructions for correcting as for the error "Sending packages from OLTP to BW lead to errors".
E3 If an ODS activation step fails, check the errors first, but generally first try to repeat (restart) the step in the chain. We have not had this happen in quite a while though.
E4 This only happened once - 0HRPOSITION_ATTR invalid time interval and 0HRPOSITION_TEXT time overlapping error -
The first failure occurred when loading HR position attribute data. Position 30011244 had an invalid time interval. The valid from date = 08/16/2006 and the valid to date = 08/15/2006. The valid to date was back dated. In order to resume processing, I changed the valid to date to 12/31/9999 (an open end date). I left the valid from date as it was (08/16/2006). The remaining data loaded successfully.
The second failure occurred when loading HR position text data. The same position, 30011244 had two records with a time overlapping error. The first record had valid from date = 08/01/2003 and valid to date = 08/15/2006 with a description of SSA (Sales and Service Associate). The second record for the same position had valid from date = 07/01/2005 and valid to date = 08/15/2006 with same description. Both records had the same valid to date = 08/15/2006, causing the error. I changed the dates on the second record as follows: valid from date = 08/16/2006 and valid to date = 12/31/9999 (again an open end date) to resume processing. The load completed successfully.
Resumed process in master data chain with repeat
These failures were holding up the completion of our master data load and our transaction data loads, so I made these changes manually within BW to continue our load processing. I did not change any data in P01, just the data in BW.
Both of these HR master data loads are full loads, so the dates for the position need to be corrected today on P01 or the BW loads will fail again tomorrow with the same errors. Once the corrections are made in P01, then tomorrow's full loads in BWP will overwrite the changes I made with the correct dates.
Check these blogs also for the better understanding:
You must be a registered user to add a comment. If you've already registered, sign in. Otherwise, register and sign in.
| User | Count |
|---|---|
| 3 | |
| 3 | |
| 2 | |
| 1 | |
| 1 | |
| 1 | |
| 1 | |
| 1 | |
| 1 | |
| 1 |
You must be a registered user to add a comment. If you've already registered, sign in. Otherwise, register and sign in.