DSD.stagerun active stage finishing

[Xanadu]

Hello all,
My job takes such a long time to finish.

I have this in the log :
DSD.stagerun active stage finishing

This is my job setup:

DRS Stage - IPC Stage - hashfile.
(Writing from the table into the hashfile.)
(Single CPU , 2 GB RAM server)

The table size is 10.2 million.
All the 10.2 million are written to the IPC stage at ~ 22000 rows/sec.
THe second link of IPC stage - hashfile finishes writing all the 10.2 million but the job just hangs there for about 8 minutes till it finishes. (~9600 rows/sec)

I cleared the &PH& directory on the server (except the NLS file.)

Any other thoughts ?
Thanks
-Xanadu

[ds_developer]

Are you using the "Allow stage write cache" option on the hash file? You will see a delay in the job finishing as it writes the file.

John

[Xanadu]

You think it would take that long ?
Just to write the memory cache to the disk ? 384MB...
Hmm..I would try turning it off and trying it..but I think it would delay the rest of the process...

Thanks for the reply though
-Xanadu

[ray.wurlod]

Do you have any long-running after-stage or after-job subroutines? These, too, can lead to a long time between stage finish and job finish.

[Xanadu]

No Ray.
I donot have any after-stage or after-job subroutines. Does the type and properties of the hash file cause this ? I have a dynamic type 30 hash file with the default properties and min modulus of 1. Do you think that matters ? I have a pretty decent speed of reading from teh database into IPC stage - 22000 rows/sec.
The IPC stage writes to the file at the same speed but is unable to finish writing at the same speed.
Any otehr thoughts ?
Thanks
-Xanadu