Re: what's causing session to timeout?

On Mon, 08 Dec 1997 21:56:48 GMT, boulkenospam_at_globalnet.co.uk (Keith Boulton) wrote:

>On Fri, 05 Dec 1997 14:46:56 GMT, chuckh_at_dvol.com (Chuck Hamilton)
>wrote:
>
>>I have a sqlplus job as part of a korn shell script that looks like
>>this...
>>
>>sqlplus user/passwd @sqlscript > logfile
>>
>>The sqlscript has no EXIT at the end. When I ran the shell script
>>manually yesterday, it hung for about 20 minutes until I killed it.
>>It's then that I discovered the missing EXIT.
>>
>>My question is this. Every night for the last 3 months this script has
>>run from a job scheduler (maestro), but it gets past this line and the
>>shell script completes normally. What's causing it to complete so that
>>the rest of the shell script runs? The listener's timeout parameter is
>>set to 0, the user's profile has no limits (i.e. default), and there's
>>no sqlnet timeout parameter set on the client side either.
>>--
>>Chuck Hamilton
>>chuckh_at_dvol.com
>>
>>"Therefore the Lord Himself will give you a sign:
>>Behold, a virgin will be with child, and bear a son,
>>and she will call His name Immanuel." (Isa 7:14 NASB)
>If I remember aright, without the EXIT statement, SQLPlus will wait
>for terminal input. Attempting to read from a terminal ( ie getting to
>the end of redirected stdin) automatically kills a unix process which
>was started as a batch job, allowing the calling process (the shell
>script) to continue execution.
>

But when I ran the batch job from the shell prompt, it hung. When it runs from a job scheduler, it doesn't. Why?

I'm not redirecting input with either cat | sqlplus, or a "here document" (sqlplus <<EOF). The sql commands are in a sql script supplied on the command line (sqlplus @script.sql).

--
Chuck Hamilton
chuckh_at_dvol.com

"Therefore the Lord Himself will give you a sign:
Behold, a virgin will be with child, and bear a son,
and she will call His name Immanuel." (Isa 7:14 NASB)