On May 29, 5:25 pm, "Dereck L. Dietz" <diet... (AT) ameritech (DOT) net> wrote:
Varies. You can search the bug database or release content lists for
memory leaks. There have been arguments over the years as to whether
Windows needs to be rebooted.

Of course, nothing says that all bugs have been found.

It also depends on what exactly you are running. Bug 5600676, for
example, is PIPELINED FUNCTION WITH IN PARAMETER LEAKS UGA AND PGA
MEMORY. If you are running packaged software, how would you even know
if you are doing that?

Now, many people believe that rebooting as a prophylactic is stupid,
and the root cause needs to be analyzed and found. Some vendors
certainly have "funny" recommendations that only serve to keep their
support departments less busy. Do you have the tools and manpower to
dispute it? Do you believe in magic?

Personally, I can't stand not knowing what is really going on. But if
someone else makes the judgement that it is too expensive to do a root
cause analysis, I grin and bear it. Unless I'm a paying consumer and
I know what is wrong and I'm being shined on. That blows the bolts
out of my neck.

jg
--
@home.com is bogus.
http://www.mccartyphotoworks.com/

On May 29, 8:25 pm, "Dereck L. Dietz" <diet... (AT) ameritech (DOT) net> wrote:
I am running Oracle 10.2.0.2 on 64 bit Windows. Bouncing an Oracle
database on Windows twice a year might be reasonable, twice a month
points to a problem, but twice a week points to a severe
administration problem. You might want to check the value for
CURSOR_SHARING. When set to FORCE (and possibly SIMILAR), it might be
necssary to bounce an Oracle 10.2.0.2 database instance with any patch
applied a couple times a week, or even as often as every couple hours
- this is not a Windows only problem.

Try to obtain a better explanation of the problem. Flush junk may
mean many things (hope that they do not mean flushing a twice weekly
cold backup to tape).

Charles Hooper
IT Manager/Oracle DBA
K&M Machine-Fabricating, Inc.

"Dereck L. Dietz" <dietzdl (AT) ameritech (DOT) net> wrote in
news:0E27i.6149$C96.2183 (AT) newssvr23 (DOT) news.prodigy.net:

I have one Oracle database & a system that has been up a while

uptime
7:25pm up 1714 day(s), 21:24, 3 users, load average: 0.11, 0.16, 0.18

Dereck L. Dietz wrote:
Fire the admins.

If there is a real problem this isn't diagnosing and fixing it.
If there isn't a problem they are incompetent and/or dangerous.

My instinct is the later ... not the former but getting someone
that knows what they are doing is the correct solution either way.
--
Daniel A. Morgan
University of Washington
damorgan@x.washington.edu
(replace x with u to respond)
Puget Sound Oracle Users Group
www.psoug.org