Job geting hanged

[ramank1981]

Hi ,
I am facing a strange problem.
I am running one job with sequential file as source and Db2 as target with some transformations in between.

Problem is that sometimes the job runs fine and populates data in target table, but on other day the same job gets hanged and keeps on runing until i go and stop from the director.

We are facing this isse since long time.

[Ultramundane]

Hi ,
I am facing a strange problem.
I am running one job with sequential file as source and Db2 as target with some transformations in between.

Problem is that sometimes the job runs fine and populates data in target table, but on other day the same job gets hanged and keeps on runing until i go and stop from the director.

We are facing this isse since long time.

What OS and version? Did you make sure that nothing was hung in DB2? What step was the job on when it hung?

[ray.wurlod]

Did you make sure that nothing was locked in DB2?

[ramank1981]

I checked the DB2 stage, there is no hung process of DB2.
Stange part is that we ran the same job with same source , same target 2 days back and the job was completed in few seconds.
But now the job is hanging ...

[ray.wurlod]

Wrap your job design in

Code: Select all

 tags so we can see it properly.

Add Peek stages on relevant links so you can truly verify where data are and are not flowing.

And you need to check the DB2 database (not the DB2 stage) to determine whether anything is locked.  Are you, for example, updating one or both of the tables that provide input to the Join stage?

[Ultramundane]

I checked the DB2 stage, there is no hung process of DB2.
Stange part is that we ran the same job with same source , same target 2 days back and the job was completed in few seconds.
But now the job is hanging ...

What OS and version? Just curious, because AIX has a patch to fix fifo thread intermittent hangs caused by a timing issue. This caused us big problems.

[rajendharreddy]

One thing you can try is, drop the table and recreate the table with proper table space. Thanks..

[dspxguy]

I checked the DB2 stage, there is no hung process of DB2.
Stange part is that we ran the same job with same source , same target 2 days back and the job was completed in few seconds.
But now the job is hanging ...

What OS and version? Just curious, because AIX has a patch to fix fifo thread intermittent hangs caused by a timing issue. This caused us big problems.

Hi, can you please explain the problems you faced and how did you resolve them, what version and o/s you have?
Please let me know, Thanks

[ramank1981]

we are using AIX 5.3. it still seems like a resource problem, no further clue.

[Raghavendra]

can you please check if your /tmp directory is filled.
And also check for &PH& directory.

[Ultramundane]

we are using AIX 5.3. it still seems like a resource problem, no further clue.

You could be running into the fifo_open problem. This caused us a lot of job hangs. There is patch for the AIX kernel to address this issue. The patch is: IY86253

I think you can google the patch.

[ramank1981]

I had earlier cleared the &PH& , but the job was still hanging.
I have asked AIX admin for the patch IY86253, waiting for him to respond.

[ramank1981]

It seems we finally found the root cause of the problem. One of the routine was creating the bottle neck. After we implemented the same thing with differnent logic, the jobs are running fine. :D

Thanks to all of you for giving inputs.