DSXchange

Posted: **Mon Aug 29, 2005 11:58 am**

Hi,

I am running a master sequence and it calls 5 more sequences at a time.

Out of the 5 couple of them run fine but remaining of them abort giving the following error.

One of the job this sequence calls aborted with the following error

Code: Select all

Controller problem: Error calling DSRunJob(LoadCalendarMonthDim), code=-99
[General repository interface 'other error']

And the sequence also has the following message in job log

Code: Select all

Controller problem: Error calling DSWaitForJob(), code=-99
[General repository interface 'other error']

The sequence runs fine when run independently instead of calling from the master sequence.

Some of the jobs this child sequence calls aborted with a time out problem.

Code: Select all

Controller problem: Error calling DSRunJob(ExtractCalendarQtrDim), code=-14
[Timed out while waiting for an event]

Some of the jobs the child sequence calls ran fine but has the following warning in job log.

Code: Select all

Unable to allocate Type 30 descriptor, table is full.
DataStage Job 1197 Phantom 8589
DataStage Phantom Finished

Is there any limit on the no. of sequences or jobs that a sequencer can call at a time?

what does code -99 stand for?

Sorry for asking so many questions in one post. But all of them comes when i run my main sequencer.

Thanks,
Srikanth

Posted: **Mon Aug 29, 2005 5:20 pm**

Code -99 is "other" errors; those for which there is no specific diagnostic handler within DataStage code.

If you are getting T30 table full errors you have reached the maximum number of hashed files that can be open at any one time. (Remember that all the DataStage Repository tables are hashed files, and there are up to seven of these per job.)

Sometime very soon you are going to need to increase the value of your T30FILE configuration parameter in the uvconfig file. If it's currently 200, make it 1000. Then you need to stop DataStage, run uvregen, and re-start DataStage.

Check that the change has taken effect by executing the command

Code: Select all

$DSHOME/bin/analyze.shm -t | grep T30FILE

Posted: **Tue Aug 30, 2005 8:57 am**

Thank you Ray for the response.

We will change the T30FILE configuration parameter and rerun the sequence again.

And why is this time out problem coming?

Code: Select all

Controller problem: Error calling DSRunJob(ExtractCalendarQtrDim), code=-14 
[Timed out while waiting for an event]

Is it that the server capacity is not good enough to call so many processes at a time or that datastage has a limitation on the no. of jobs or sequences it can trigger at a time?

Thanks,
Srikanth

Posted: **Tue Aug 30, 2005 4:35 pm**

Server overloaded.

I've run quite a few more than five simultaneous jobs, but on a 32 CPU machine with lots of memory.