Strange DataStage problem

LNarayan · Post by **LNarayan** » Tue Jun 30, 2009 4:57 pm

Hi all,

I have come across a strange problem while using DataStage and I have mentioned strange becuase I have almost dug deep into many posts but could not get any solution. Hence I am relying on anyone of you for the solution before i take forward with support which will again drag me longer.

Coming to the problem...

I have a job called 'DiRDMActivityDimn' which exists in DS Manager window but not in director or designer. I have tried all sort of things one by one given below

1) DS.REINDEX ALL (reindexed project)
2) Ran DS.CHECKER
3) rebuilt repository indices through DS.TOOLS

when run the below command in ds admin window i get this:

SELECT * FROM DS_JOBS WHERE UPPER(NAME)='DIRDMACTIVITYDIMN';

Job name.... DiRDMActivityDimn
Description.
No.......... 1
Category.... 15000 - AtomicExport
\15070-EDMR\15070b -
DataMartImport

1 records listed.

but with command i get this error message

LIST DS_JOBS WITH JOBNO EQ 1

0 records listed.
"DiRDMActivityDimn" not found.

I tried deleting through manager it shows the below error

"job DiRDMActivityDimn doesnot exists in this project." I also tried deleting the category but all jobs got deleted except this job and category with same error message.

when i rename/delete the job in DS_JOBS through DS admin i get the below error message

UPDATE DS_JOBS SET NAME='DIRDMACTIVITYDIMN_OLD' WHERE UPPER(NAME)='DIRDMACTIVITYDIMN';

Lock error: invalid lock requested.
Unable to continue. Aborting!

please let me know if you have any idea. Thanks all for your repsonse.

Thanks
Narayan

ray.wurlod · Post by **ray.wurlod** » Tue Jun 30, 2009 6:56 pm

Verify the internal structure of DS_JOBS, using UVFIXFILE or fixtool (neither running in "fix" mode).

Do the following exist? RT_BP1, RT_CONFIG1, RT_LOG1, RT_STATUS1

LNarayan · Post by **LNarayan** » Wed Jul 01, 2009 3:06 am

Hi Ray,

Thanks for your suggestion.

Could you please tell me how do we run this UVFIXFILE utility i.e. command line to run this utility and from where it has to be run.

and these RT files existed but got deleted while running the DS.CHECKER tool.

Thanks
Narayan

LNarayan · Post by **LNarayan** » Wed Jul 01, 2009 3:27 am

This is the output of the command UVFIXFILE DS_JOBS

>UVFIXFILE DS_JOBS

Beginning TRACE of DS_JOBS.
ID "\\\15000 - AtomicExport" doesn't hash to primary group 1.
ID "DiRDMActivityDimn" doesn't hash to primary group 1.
ID "\\\15000 - AtomicExport\15070-EDMR" doesn't hash to primary group 1.
ID "\\\15000 - AtomicExport\15070-EDMR\15070b - DataMartImport" doesn't hash to primary group 1.
ID "\\NextNumber" doesn't hash to primary group 24.
TRACE of DS_JOBS completed.

Warning: file load discrepancy detected.

Scanning overflow buffers.
Scan complete.

306 group(s) processed.
447 group buffer(s) processed.
5190 record(s) processed.
Number of data bytes = 468164.

ArndW · Post by **ArndW** » Wed Jul 01, 2009 4:36 am

You have corruptioni in your DS_JOBS hashed file. You can ensure that no users are logged, then make a backup of the DS_JOBS file from UNIX and try the UVFIXFILE again, this time with the "FIX" option to see if the file can be repaired. Run it two times (the 2nd run should show no more errors) and post the results here.

LNarayan · Post by **LNarayan** » Wed Jul 01, 2009 4:55 am

Hi ArndW,

Thanks for your response.

I ran the UVFIXFILE utility in DSADMIN window but thrown a message saying "User does not have superuser privileges! Disabling the FIX option and continuing" even when I run it with dsadm user which I guess is the administrator. Please let me know if my understanding is wrong.

Thanks
Narayan

ArndW · Post by **ArndW** » Wed Jul 01, 2009 5:54 am

In that case you probably didn't install DS in impersonation mode, so you'll have to run that as root.

LNarayan · Post by **LNarayan** » Wed Jul 01, 2009 6:55 am

Hi Arndw,

I have requested the admin (having root prvillege) to run the below command even then issue with DS_JOBS was not sorted out. find the log below.

./fixtool -file DS_JOBS -level 5 -fix

3 Errors found in file groups.
0 Other errors found in file.

5190 records processed.

3 Errors on file.

Start Fixing.
0 Errors fixed.

Fixtool Completed.

Will this be fixed using UVFIXFILE utility if yes I can request the admin to run this utility?

Thanks
Narayan

ArndW · Post by **ArndW** » Wed Jul 01, 2009 7:12 am

It is hard to answer that question, so I'll just "maybe it will". Another option is to try the "RESIZE DS_JOBS * * *" so see if that might force a fix. Please note that it looks like any repair of this file will probably involve losing records in the file if the fixtool cannot repair it.

LNarayan · Post by **LNarayan** » Wed Jul 01, 2009 8:29 am

RESIZE DS_JOBS * * * has thrown an error saying "Unable to gain exclusive use necessary to continue. File is currently in use by other processes. Please try again at a later time when the file is not in use"

The error is strange becuase no one has logged into the system and none of the jobs are running.

ArndW · Post by **ArndW** » Wed Jul 01, 2009 8:50 am

Any running clients (director or designer) will have this file open.

chulett · Post by **chulett** » Wed Jul 01, 2009 9:22 am

Or defunct / zombie processes or connections as well, I would think.

LNarayan · Post by **LNarayan** » Wed Jul 01, 2009 9:24 am

Is there a way to identify the running clients in a particular project? I have the below command ps -ef |grep dsapi since we have several projects, this command lists out the opened client connection of all the projects in the server.

ray.wurlod · Post by **ray.wurlod** » Wed Jul 01, 2009 4:42 pm

list_readu (or LIST.READU from TCL) will show an "admin" lock on the project to which the client's agent process is connected.

LNarayan · Post by **LNarayan** » Mon Jul 06, 2009 2:57 am

ArndW wrote:Any running clients (director or designer) will have this file open.

RESIZE DS_JOBS * * * has thrown an error saying "Unable to gain exclusive use necessary to continue. File is currently in use by other processes. Please try again at a later time when the file is not in use"

This has occured even after closing all the client connections. Is there any solution to fix the problem.