HASH FILES AND SMAT

Post questions here relative to DataStage Enterprise/PX Edition for such areas as Parallel job design, Parallel datasets, BuildOps, Wrappers, etc.

Moderators: chulett, rschirm, roy

Post Reply
sendmk
Charter Member
Charter Member
Posts: 136
Joined: Mon Oct 03, 2005 5:02 am

HASH FILES AND SMAT

Post by sendmk »

Hi

We have an error message where we get "Failed to Open RT_SC or RT_CONFIG" is that anything related to smat -d what i do in the .vin of DSEngine. If so then if i do a smat -d then is that the number what we get is number of RT_files.

Thank you
ray.wurlod
Participant
Posts: 54607
Joined: Wed Oct 23, 2002 10:52 pm
Location: Sydney, Australia
Contact:

Post by ray.wurlod »

The maximum number of dynamic (Type 30) hashed files that can be open simultaneously is controlled by the T30FILE configuration parameter (in the uvconfig file). This determines the number of slots in the in-memory table that is reported by smat -d command. Most of the Repository tables (DS_* and RT_*) are dynamic hashed files; this is why the T30FILE table seems to fill with these, particularly since you are in a parallel environment and not using your own hashed files.

The default value for T30FILE is 200, which is not really enough. Increase it to 500 or even 1000. Each slot (row) in the T30FILE table is only about 128 bytes long so the impact on shared memory is negligible.

Each entry in the T30FILE table is identified by device and inode number, rather than by file name. It appears you've figured out how to do this mapping. My choice is to run ACCOUNT.FILE.STATS ALL LOCAL regularly in the project, then

Code: Select all

SELECT FILENAME.LONG, CAST(FILEINODE AS INTEGER) FMT '10R' FROM STAT.FILE ORDER BY 2;
to yield my map.
IBM Software Services Group
Any contribution to this forum is my own opinion and does not necessarily reflect any position that IBM may hold.
Post Reply