HP just did announce support on the 3000 for the latest Oracle
version (9?).  Now, hopefully, HP will do more to effortlessly
port UNIX applications and the 3000 will really take off.

Nick D,


Shah, Roshan (Windsor, HQ) wrote:
>
> Ppl say HP3K is dying.. not because it is not a good system. All blame
> on HP for  the sloppy work.
> Only good thing is the CSY team which is fighting the battle all alone.
>
> What is more important is Selling Boxes and Support Applications to run
> on them...There has to be
> something really really dramatic to turn this around... HP Can do it,
> but direction & resources has to come from the Top Management.
>
> Roshan
> Opinions my Own.
>
> >----------
> >From:  Joe Smith[SMTP:[log in to unmask]]
> >Sent:  Monday, November 16, 1998 11:10 AM
> >To:    [log in to unmask]
> >Subject:       update on Oracle and system aborts
> >
> >Last month I posted a message regarding Oracle and multiple system aborts.
> >We had recently updated to PP5 and also installed Oracle 7.3, with the new
> >listener process.  In all, we experienced 7 system aborts and 3 listener
> >'hangs', with a bad pin requiring a reboot of the system to clear.  This
> >problem was referred to the HP escalation center where HP's and Oracle's
> >labs in India worked to determine the cause of the problem.  I was very
> >impressed by the dedication of resources to this problem by  HP's
> >Escalation Center.
> >
> >The point of failure was isolated to a piece of HP's networking code
> >working on behalf of a call from an Oracle process.  The call was basically
> >initiated to kill a socket, and that socket did not exist so the system
> >failed.  HP and Oracle are looking at why this scenario exists, but in the
> >mean time we turned off an Oracle parameter, 'sqlnet_expire_time', to
> >prevent the code from ever taking the fatal path.  This parameter, aka dead
> >connection detection,  has been off for 3 weeks now and there have been no
> >more problems.  The loss of that parameter has not proven detrimental in
> >any other regard...
> >
> >On another note, the issue about the future of Oracle on the HP3000 was
> >recently discussed on this list.  In the corporate world, this issue has
> >already been decided.  Most new applications that we see come in require
> >Oracle8, or at least 7.3.3.  The HP3000 is seen as lagging far behind NT
> >and Unix  as HP and Oracle's preferred platforms.  Thus, the NT side of the
> >house has gained a firm foothold in our Oracle area.  All new apps will
> >certainly (and justifiably) go there.  I proposed a single high-end HP9000
> >to handle the upcoming load of Oracle based applications, including data
> >warehousing, but our Oracle reps have said the NT box can handle 100,000
> >Oracle transactions a day and is dirt cheap and more reliable than a Unix
> >based solution.  We'll soon see...
> >
> >The sad note in this is that the HP3000 proved to be an extremely reliable
> >platform for Oracle for 4 years, other than our 7.3 related issues.  Now,
> >even if the powers that be get Oracle8 on the HP3000, it will still be seen
> >as a secondary platform for Oracle since it is so far behind.  Just one
> >less reason for us to keep our HP3000...
> >