Re: Long Sessions

On Aug 18, 5:49 pm, DA Morgan <damor..._at_psoug.org> wrote:
> pcarr01 wrote:
> > Hi,
> > We have an issue with the amount of time that a firewall will hold a
> > session open.
> > Currently we have 72 hours, this drives the security and network
> > people crazy but we have a requirement for stuff like very long
> > database queries, launch an ssh or sqlplus session onto an oracle box
> > and the session expires before the query finishes and we lose the
> > result, very bad when you are working on a problem at 3 in the
> > morning
> > and the Damages for not getting the database back up by 5 are in 6
> > figures.
> > How does the industry in general deal with this sort of issue?
>
> A number of thoughts.
>
> The first parallels Sybrand's ... a 72 hour query is outrageous.
> Get someone in to tune it.
>
> The second is that this query should not be running through the
> firewall and be, in any manner, dependent upon the survival of
> a session.
>
> Have the query report its output to a table and then log on later
> to retrieve the resultset, then SELECT * FROM report_table, or use
> a materialized view to produce the result set in the first place.
> --
> Daniel A. Morgan
> University of Washington
> damor..._at_x.washington.edu (replace x with u to respond)
> Puget Sound Oracle Users Groupwww.psoug.org

Appreciate the comments on performance but this isnt an oracle issue its a network issue that affects oracle. I was working on a problem that was exposed because Oracle 8 had known issues with bitmapped indexes and that was what ruined the performance, we were running queries and doing updates in a diagnostic and recovery context that were not returning within an acceptable timescale & if some network device had detected no activity then .. bang .. query lost. There is no way I could have done this without some sort of client session on the target platform be it ssh, plus80 or whatever.

I did ask the question on the firewall group.

The question is Received on Sun Aug 19 2007 - 04:54:27 CDT