Re: OWS problem

On Thu, 01 May 1997 11:27:53 +0100, Brent Smith <smithb_at_x500.bt.co.uk> wrote:

>Unfortunately, this is simply a bug in OWS 2.1 and I understand there
>is no work-around for it. The answer is either to wait for 2.11 (2.1.1?)
>- (Oracle support know about the bug and claim that it is fixed), or for
>OWS 3 (available as a beta release, but rumour has it that it is not
>worth trying out), or to go back to 2.0 which does not have the problem
>- at least not as bad - I would still recommend stopping and starting it
>overnight.

On our system 2.0 was worse. If you ever filled all the connections (not zombies, actvie connections), then the server would just lock up and stop responding. With 2.1, the server does not lock up and the active connections will continue to work. On our busy days, we have had 30K queries. We limit the sqlnet connections to 16, because we have 350+ active users on our database machine also. If we set the connections any higher, then our sequent machine goes off the chart and no one gets any work done.. :(

Thanks for the information. Now I know it's not just me! :)

Dan

>
>We're using OWS 2.0, Solaris 2.5, Oracle 7.2 on a SUN SparcServer 20
>that handles over 60K queries per day and it generally works alright. We
>tried OWS 2.1 on our test machine and had the same problems, so it looks
>like a cross-platform bug.
>
>Good luck,
>
>Brent Smith
>
Received on Thu May 01 1997 - 00:00:00 CDT