In article <[log in to unmask]>,
  "Michel Gauthier" <[log in to unmask]> wrote:
> We do have Oracle 7 (7.2.3 at first, 7.3.2 now) on a 969-120 and a
979-300.
> You may experience troubles connecting in C/S mode to Oracle if you do
not
> have enough contiguous disk space available. Also, adding memory would
help,
> but if you tell that when access to Oracle seems to be slow but you
cannot
> put the finger on it with glance, there might be something else. When
a
> Oracle server process takes a lot of ressources (I/O, CPU, etc), you
should
> see it in glancexl as the TNSLSNR process. In fact, the server process
is
> ORACLE.PUB.something, but glancexl sees it as TNSLSNR.
<snip>

Thank you for your feedback Michel - what I have been trying lately is
to setup a batch job that creates a session to run glancexl; the output
of glancexl is then posted to a logfile.

A good way of doing this is described by Steve Barrett in his post to
this list back on April 14th this year, his message title is -
     Re: Performance Monitoring

The benefit of this approach for me is that there is not an easy way of
determining when key processes aRE being started by the users.

As yet, I have not encountered the TNSLSNR process you refer to. But,
the main Oracle process that glancexl does capture on this system is
ORAPROC. On this particular 969/120 this program is being run by a batch
job - ORASRV7,MGR.ORACLE7

Using glancexl on-line, it is possible for me to see all of the child
processes under the ORASRV7 batch job by viewing the details of the
associated Job No; i.e. on the GLOBAL screen, simply pressing the 'J'
key and then supplying the appropriate job number.

Thanks again!

Robert.


Sent via Deja.com http://www.deja.com/
Share what you know. Learn what you don't.