[CentOS] dual head on centos 5 and ancient Nvidia Quadro NVS280SD card

Tue May 7 19:59:53 UTC 2013
Fred Smith <fredex at fcshome.stoneham.ma.us>

On Tue, May 07, 2013 at 03:09:13PM -0400, Ross Walker wrote:
> On Tue, May 7, 2013 at 2:34 PM, Ross Walker <rswwalker at gmail.com> wrote:
> 
> > On Mon, May 6, 2013 at 4:47 PM, Fred Smith <fredex at fcshome.stoneham.ma.us>wrote:
> >
> >> I was just handed a 2nd monitor for my system at work, and using Centos 5
> >> (latest)
> >> can't make dual head work. a good bit of googling isn't being particularly
> >> helpful either.
> >>
> >
> > Dual head or dual monitor?
> >
> > Dual head typically means running two instances of X, one on each video
> > out, which can only be done with two or more graphic cards from what I can
> > gather.

hmm. good point.

> >> It's a HP workstation xw4100, with Nvidia Quadro NVS280SD graphics card.
> >> Enabling dual head in the "display" app simply configures X so that it
> >> (i.e., X) won't start. I haven't found the x log file in /var/log to be
> >> helpful, either. This is using the legacy 96.x.x driver from Nvidia.
> >>
> >> the NvidiaDetect app (from epel) says it should be using a newer driver
> >> than the 96.x.x (forgot which one, exactly) but when attempting to install
> >> it I get a msg that the card requires a 96.x.x driver.
> >>
> >> browsing to nvidia.com and entering the model numbers into their driver
> >> finder app gives another newer version that also gives the same result.
> >> So it looks like I'm stuck with the 96.x.x driver.
> >>
> >> some googling indicates a few people have made it work, but none of their
> >> methods are working for me.
> >>
> >> I was beginning to wonder if the hardware even supported dual head, so I
> >> booted up a Fedora 17 LIVE CD. it initializes both monitors with no action
> >> from me at all, with a desktop spanning the two screens, just fine. It
> >> must
> >> be using the Nouveau driver (which, AFAIK, can't be used on Centos 5),
> >> so still the issue could either be spanning/dual head doesn't work with
> >> the
> >> ancient nvidia driver, or we (neither me, nor the tools on Centos) knows
> >> how to configure it.
> >>
> >> Clues would be appreciated. thanks in advance!
> >>
> >
> > In my experience I was able to drive both the DisplayPort and DVI (or HDMI
> > and VGA) interfaces off my card to give me dual monitor support.
> >
> > To setup the monitor preference I just created a monitors.conf file in
> > /etc/X11/xorg.conf.d as such:
> >
> > Section "Monitor"
> >         Identifier      "HDMI1"
> >         Option          "Primary"       "true"
> > EndSection
> >
> > Section "Monitor"
> >         Identifier      "VGA1"
> >         Option          "RightOf"       "HDMI1"
> > EndSection
> >
> > Substitute the Identifier for whatever 'xrandr' gives you, and use LeftOf
> > if you secondary monitor is left of your primary.
> >
> 
> Nevermind, I am not paying attention here, my setup is C6 with Intel.

This is strange...

xrandr (on C5) only shows one monitor even though two are connected

Booting up a Fedora 17 live CD, automatically configures for a single
desktop spread across both monitors, AND xrandr shows both monitors. This
makes me think that the hardware supports what I was seeking but C5
doesn't.

So, some more poking around, and I discover that Nvidia supports "Twinview"
which actually works on C5, giving me a single desktop spread across both
displays. According to the Nvidia readme, it fools X into thinking it has only
one monitor, and it does seem to work.

While I think I might like two separate desktops, it doesn't look like
it's likely to be possible with the system I have, so I'll manage with
what I've got, I think.

thanks for the responses!

Fred
-- 
---- Fred Smith -- fredex at fcshome.stoneham.ma.us -----------------------------
                    The Lord detests the way of the wicked 
                  but he loves those who pursue righteousness.
----------------------------- Proverbs 15:9 (niv) -----------------------------