Yes. The usage information is coming from v$sort_usage, and is NOT
the raw tempspace allocation for the instance.
On 4/19/05, Hollis, Les <Les.Hollis_at_ps.net> wrote:
> Are you sure that the transactions are still holding the space?
>=20
> Temp tablespace segments do NOT de-allocate when a session quits using
> it. It will remain allocated so the next user to need temp space does
> not have to go through the space allocation......
>=20
> PMON will de-allocate the segments on the next instance startup.
>=20
> -----Original Message-----
> From: oracle-l-bounce_at_freelists.org
> [mailto:oracle-l-bounce_at_freelists.org] On Behalf Of Greg Norris
> Sent: Tuesday, April 19, 2005 2:55 PM
> To: ORACLE-L
> Subject: tempspace usage
>=20
> Is there a way to determine (or trace) which individual statements are
> causing tempspace to be allocated to a session? I've got a databases
> where half a dozen sessions seem to gradually allocate more and more
> tempspace (as measured by v$sort_usage), and then never release it.=3D20
> Eventually they acquire almost all of the available space, which of
> course causes a flurry of ORA-1652 errors from other sessions.
>=20
> At the moment, I'm examining a trace from one of these sessions. So
> far, I don't see anything which could cause this behaviour... a few
> inserts, some relatively simple selects, nothing particularly complex.
> Any suggestions on how to troubleshoot this sort of issue?
>=20
> The DB in question is Oracle 8.1.7.4.0 (32-bit), running on Solaris 8.
> --
> http://www.freelists.org/webpage/oracle-l
>=20
>=20
--=20
"I'm too sexy for my code." - Awk Sed Fred.
--
http://www.freelists.org/webpage/oracle-l
Received on Wed Apr 20 2005 - 08:47:13 CDT
