Page 1 of 1

execution of a job hangs after a few seconds

Posted: Tue Jun 19, 2007 11:09 am
by tbrosi
On execution of a job hangs after a few seconds.

The job design is as follows:

________________________________> T3______ SF_1
OCI1----> T1 ^ ^
__> HF_1------------>T2--------HF_3--------| |
^ ^ HF_4
OCI2_______________________| |
|
OCI3 ______> HF_2-------------------|

There are three OCI stages.
OCI stage 1 writes a hashed file HF_1 and goes to a transformer T1.
OCI stage 2 is the stream for transformer T2.
OCI stage 3 writes a hasehd file HF_2.
HF_1 and HF_2 are used as lookups in T2.
T2 writes hashed file HF_3.
There is another hashed file HF_4, which is created in another job.

On execution of the job, OCI3 writes the rows to HF_2 but all other
stages are waiting and nothing happens.

Can anybody tell me the reason for this behaviour and how can i force
the job to read the other oci stages?

Thanks in advance.

Tom

Posted: Tue Jun 19, 2007 11:16 am
by DSguru2B
Welcome Aboard :)
Enclose your ascii art in

Code: Select all

 tags. Its hard to tell the design otherwise.
The other oci stages might be waiting to finish the query. Are your queries just simple selects or complex joins and other functions?

Posted: Tue Jun 19, 2007 2:48 pm
by ray.wurlod
Passive stage types (such as Hashed File stages) can not open their output link(s) until their input link(s) are closed. Until the hashed file is fully populated and the write cache flushed to disk, the remainder of the job will wait.