Oracle FAQ Your Portal to the Oracle Knowledge Grid
HOME | ASK QUESTION | ADD INFO | SEARCH | E-MAIL US
 

Home -> Community -> Usenet -> c.d.o.server -> Re: Monitoring questionable long running sessions

Re: Monitoring questionable long running sessions

From: Rick Denoire <100.17706_at_germanynet.de>
Date: Tue, 13 Jul 2004 20:58:31 +0200
Message-ID: <6ib8f0dti4qmakbcomf2va4dbqqrljhal5@4ax.com>


"VC" <boston103_at_hotmail.com> wrote:

>OK.
>
>Could please trace the OS process which corresponds to SID 112 ?
>
>You can use truss -f -o a.log -p process_id in order to obtain the trace.
>You'll see what system calls are executed by the PID and will know for sure
>there are /are not any read operations.

No I can't anymore. I killed the sessions already - again, I had to kill the OS processes. But guess, I tried the truss command last weekend, after the problem appeared the first time on Friday. Unfortunately, no such command in RHAS 2.1.

Anyway, today I gathered a lot of information before terminating the sessions. Among other things, I produced trace files that the Oracle support will inspect.

Funny enough, trying to compute the execution plan from the SQL text found in some v$sql* views, I get a syntax error about missing parenthesis! (Not true). The execution plan would not reveal anything exciting and does not seem to be relevant to the problem itself. This is not a long running session (which can be monitored using the view v$session_longops, by the way), but a session spinning on the same spot - forever.

Thanks for your hints.

Bye
Rick Denoire Received on Tue Jul 13 2004 - 13:58:31 CDT

Original text of this message

HOME | ASK QUESTION | ADD INFO | SEARCH | E-MAIL US