RE: onconfig recommendations
> mosserp@wellsfargo.com wrote:
> >>> RESIDENT 0 # Forced residency flag> >>>
> >> (Yes = 1, No =
> >>
> >>>
> >>>
> >> Don't know HP model numbers, but it this is a PA-RISC machine, you
> >> almost have to set RESIDENT to -1 to prevent poor performance
> >> due to the
> >> architectures restrictions of only four active shared
> memory segments
> >> per process. On HP PA-RISC platforms, besides marking all
> >> shared memory
> >> lockable, it combines the 'resident' segment with the initial
> >> 'virtual'
> >> segment to reduce the count. VERY important on PA-RISC if
> >> you are using
> >> shared memory connections. On Itanium machines, no problem,
> >> though I'd
> >> still recommend using RESIDENT 1 or -1 to prevent the
> shared segments
> >> from swapping.
> >>
> >>
> >>> 0)
> >>> NETTYPE soctcp,4,150,NET> >>>
> >>>
> >> The above line is no-op with no TCP connection names in
> >> DBSERVERNAME or> >> DBSERVERALIASES.
> >>
> >>
> >>> NETTYPE ipcshm,1,25,NET> >>>
> >>>
> >> It's very bad for performance and responsiveness to have
> >> shared memory
> >> poll threads in NET VPs. It wastes CPU cycles for no reason.
> >> Configure
> >> this in CPU VPs and as many as there are NUMCPUVPs (see below
> >> on that).
> >>
> >
> > For HPUX, I'll have to disagree with you on this one, Art. When you
> > configure multiple shared memory poll threads on HPUX, you
> will get that
> > many separate Message-type memory segments, very quickly
> eating into /
> > exceeding that optimal maximum of 4 segments. I do
> wholeheartedly agree
> > on assigning ipcshm connections to CPU, though.
> >
> Hey Paul,
>
> You're the platform expert! Though I had thought that the message
> segments were less of a problem than the virtual segments since each
> listener only attaches to one of those segments. Good to know.
>
> Art S. Kagel
> Oninit
>
However, if Mark is correct below, then it looks like we are both a bit
behind in HPUX peculiarities. :-O
> Mark Jamison [majp51@yahoo.com] wrote:
Unless you are still on HP-UX 10.00 the 4 segments is no longer a worry.
HP-UX 11.0 the cutoff on performance is around 11 segments, and I
haven't tested beyond 11.0