Long running Jobs

DeepakMohanty · Post by **DeepakMohanty** » Wed May 07, 2008 4:57 am

We are using the db2 enterprise edition stage, to load data from a sequential file into the database. Write meathod is LOAD
The issue is when the source file is empty the jobs do not terminte. The sequential file does import zero records. But after that it just hangs. This happens to all jobs with empty source files. If anyone could help me out i would really appreciate it..
Thanks

ArndW · Post by **ArndW** » Wed May 07, 2008 5:05 am

That sounds like a bug, I have never seen anything like that. If you change the load method does the job still hang? Are these actual data sequential files or are they perhaps named pipes?

DeepakMohanty · Post by **DeepakMohanty** » Wed May 07, 2008 5:13 am

Hi,
We have sequential file.
When we change the write meathod from LOAD to Write.
Still its not working. Jobs are keeps on running.

Thanks

ArndW · Post by **ArndW** » Wed May 07, 2008 5:15 am

That is even odder. If you replace the DB2 output with a peek stage does the job still hang?

DeepakMohanty · Post by **DeepakMohanty** » Wed May 07, 2008 5:57 am

if we use peek or file as a target, job finishing, but if we use data base table its keep on running.

"we would like to talk on phone, if you are intrest please send your number to kunnal.deepak@gmail.com"

ArndW · Post by **ArndW** » Wed May 07, 2008 6:10 am

One of the powerful aspects of DSXchange is that anyone can search the collected posts to see if their error has already been encountered by others and what the solution might have been. By going off-line (be it by phone, e-mail or private messaging) this seamless history is broken and others cannot benefit.

I think that if your jobs run and complete correctly when the source file contains data but hangs when it does not is sufficient information for you to contact your support provider for assitance as it really does look like a bug.

DeepakMohanty · Post by **DeepakMohanty** » Wed May 07, 2008 7:09 am

Hi,
Problem we have some jobs having no records also running successfully,
But some having no records failing.

Can you let me know do u have ever face such situation if u had faced how u resolved that.

our environment is like that our DB is in one AIX box and Data stage in another box.

regards

ArndW · Post by **ArndW** » Wed May 07, 2008 7:56 am

No idea what could be causing your problem - unless your source file is a named pipe.

Minhajuddin · Post by **Minhajuddin** » Wed May 07, 2008 10:17 am

If it's a bug. You can stop the job from running whenever the file has no lines. You would have to do this check before you trigger the job, But, it's just a workaround.

DeepakMohanty · Post by **DeepakMohanty** » Wed May 07, 2008 10:23 am

How to do that

ray.wurlod · Post by **ray.wurlod** » Wed May 07, 2008 3:52 pm

test -z filename

DeepakMohanty · Post by **DeepakMohanty** » Thu May 08, 2008 1:53 am

Hi,

What for this command?

How to use it in my problem?

Regards
Deepak

ArndW · Post by **ArndW** » Thu May 08, 2008 2:42 am

Ray anwsered your question regarding how to find out if the source file is a named pipe. "ls -al" will also show a "p" in the attributes for pipes.

Minhajuddin · Post by **Minhajuddin** » Thu May 08, 2008 10:34 am

In the sequence which controls the job which has your join, Do this:

>> Add an execute command stage and run the command test -s <file_name> in that command by pointing it to the file which is expected to be empty.

>>Add a Nested Condition activity and use a custom trigger which triggers the job with the join only when the output from the above command is zero.

DeepakMohanty · Post by **DeepakMohanty** » Thu May 08, 2008 10:46 am

The file having one header column only.