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: Memory Leak Oracle 8.0

Re: Memory Leak Oracle 8.0

From: Jari Fredriksson <jarif_at_bitwell.com>
Date: Fri, 11 Dec 1998 14:30:56 GMT
Message-ID: <36762c91.19724935@news.eunet.fi>

Hello!

Well, I tried this too, now with Server Manager, which I somehow had not tried before..

My current Oracle version is 8.0.4 in NT 4 SP3.

Yes, no leak in svrmgr80, none whatsoever. I even extended that test script to look like

connect abba/babba_at_myserv
select 1 from dual;
disconnect
...

This is because SQL*Plus does some invisible SQL and opens cursors 'behind curtains' when connected, and I though that Server Manager maybe does not. No joy, svrmgr30 still stable.

I think they all should use the same OCI. Programs are linked to OCIW32.DLL, which in turn calls to whatever runtime is installed (so there's no need to relink apps when the runtime environment changes).

Maybe it is so, that SQL*Plus code has a leak in it, it's not in OCI then.

BUT, I remember that when I discovered the leak in my apps (it's maybe 1 year ago, I tried a simple program just connecting/disconnecting (just like that SQL-script, only with OCI calls orlon/ologof), the leak was there. Made that SQL*Plus test then, to prove the leak is not in my code..

Have to test that again, maybe it's gone with 8.0.4!

>
>Hi,
>
>I tried your test with Oracle Client and Server 8.0.3. Yes,
>there is a memory leak in plus80.exe and plus80w.exe
>(Windowed version). But I did not find any memory leak
>with your test in svrmgr30.exe (Server Manager). Maybe
>it depends on OCI version? I do not know what OCI 7 or
>OCI 8 each of them uses.
>
>And there is no any memory leak in Oracle 8.0.3 Server
>for NT during this test.
>
Received on Fri Dec 11 1998 - 08:30:56 CST

Original text of this message

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