DSXchange

DS Version : 3.6.3
OS : Digital

I have a job with three hashed stages, these are checked "pre-load file to memory".
After I ran the job, I listed temp directory.
The result,
cbcdw:/tmp]ls -al dstage*
--w------- 1 cbcdw system 16 3?10? 10:45
dstage.hash.61452aa
--w------- 1 cbcdw system 33554432 3?10? 10:45
dstage.hash.61452ab
--w------- 1 cbcdw system 16 3?10? 10:45
dstage.hash.61452ac
--w------- 1 cbcdw system 33554432 3?10? 10:45
dstage.hash.61452ad
--w------- 1 cbcdw system 16 3?10? 10:45
dstage.hash.61452ae
--w------- 1 cbcdw system 33554432 3?10? 10:45
dstage.hash.61452af
--w------- 1 cbcdw system 33554432 3?10? 10:45
dstage.heap.61452aa
--w------- 1 cbcdw system 33554432 3?10? 10:45
dstage.heap.61452ab
--w------- 1 cbcdw system 33554432 3?10? 10:45
dstage.heap.61452ac

Is that nomal?
Why the file was made 32M?
Let me know the reason.

The file size is 32M.
I chmod 600 the file.(chmod 600 /tmp/dstage*)
I checked disk usage.(du -k /tmp/dstage.hash.61452ab)
I dont know disk usage is not 32M but very small size.

On SOLARIS, my friend(?) look this case. He upgrade. The problem solved.
On DIGITAL, What can I do?

Help me.

The initial size of hashed files is set by the environment variable DS_MMAPSIZE. Yours is probably set to 32M. When you updgrade, the default may have been reset (to 16).

The location for these files (which is used for memory-mapped hashed files rather than using conventional swap space) is set by the environment variable DS_MMAPPATH. Its default value is /tmp. It is wise to change this so that /tmp does not become full.

The permissions you have chosen for the hash files is not wise. No-one else but the owner can run the job. chmod them to 660.

-----Original Message-----
From: Jang,JungHee [mailto:jhjang@lexken.co.kr]
Sent: Wednesday, 14 March 2001 07:02
To: informix-datastage@oliver.com
Subject: pre-load file to memory


DS Version : 3.6.3
OS : Digital

I have a job with three hashed stages, these are checked "pre-load file to memory". After I ran the job, I listed temp directory. The result, cbcdw:/tmp]ls -al dstage*
--w------- 1 cbcdw system 16 3?10? 10:45
dstage.hash.61452aa
--w------- 1 cbcdw system 33554432 3?10? 10:45
dstage.hash.61452ab
--w------- 1 cbcdw system 16 3?10? 10:45
dstage.hash.61452ac
--w------- 1 cbcdw system 33554432 3?10? 10:45
dstage.hash.61452ad
--w------- 1 cbcdw system 16 3?10? 10:45
dstage.hash.61452ae
--w------- 1 cbcdw system 33554432 3?10? 10:45
dstage.hash.61452af
--w------- 1 cbcdw system 33554432 3?10? 10:45
dstage.heap.61452aa
--w------- 1 cbcdw system 33554432 3?10? 10:45
dstage.heap.61452ab
--w------- 1 cbcdw system 33554432 3?10? 10:45
dstage.heap.61452ac

Is that nomal?
Why the file was made 32M?
Let me know the reason.

The file size is 32M.
I chmod 600 the file.(chmod 600 /tmp/dstage*)
I checked disk usage.(du -k /tmp/dstage.hash.61452ab)
I dont know disk usage is not 32M but very small size.

On SOLARIS, my friend(?) look this case. He upgrade. The problem solved. On DIGITAL, What can I do?

Help me.