SUMMARY: Variables not staying set under 4.0f?

From: Riggs, Joan <Joan.Riggs_at_kls.usaka.smdc.army.mil>
Date: Fri, 12 May 2000 15:41:05 +1200

Thanks so much for those who responded and especially:
Alex Gorbachev, Steve VanDevender, and Kevin Partin, each of whom took the
time to examine my scripts.

My original message:
On my 4.0d systems, I use scripts to set certain variables for our Oracles
users and the process works fine. If I try to run the same scripts under
4.0f, they do not - here is what happens:

User A sets certain variable in his .profile.
The .profile runs Script A which sets more variables.
        They can be verified within the script as being set correctly.
After Script A completes, User A tries to invoke one of the variables set in
Script A BUT the variable is no longer set!

Has anyone seen this anomaly before? These are std. ksh scripts that run
fine under 4.0d.... I can provide the scripts if necessary.

Responses:
Most people who responded had excellent suggestions, most of them said:
* source the internal script so the original script knows what it did
* export and variables set int the scripts
Excellent suggestions, but I was already following those rules, so that was
not the problem.

The solution:
It turns out that my scripts were functioning perfectly fine and that there
was a setting in Oracle that
was changing the environment behind the scenes. This was turned off and the
scripts run fine now.
The puzzling thing is that this setting is turned ON in our 4.0d system and
the scripts run fine there!?!
The setting was TWO_TASK, for those inquiring minds.

        Thanks again to all~

Joan Riggs
Raytheon RSE / KLS
Information Technology
805.355.9877
joan.riggs_at_kls.usaka.smdc.army.mil
Received on Fri May 12 2000 - 03:45:24 NZST

This archive was generated by hypermail 2.4.0 : Wed Nov 08 2023 - 11:53:40 NZDT