Re: Open Server 5.0.7 versions and features
From: Don Williams (dwilli10@san.rr.com)
Date: 04/24/03
- Next message: Jean-Pierre Radley: "Re: Mozilla profile"
- Previous message: nope@none.com: "Re: cron will not run on its own"
- In reply to: Bela Lubkin: "Re: Open Server 5.0.7 versions and features"
- Next in thread: Don Williams: "Re: Open Server 5.0.7 versions and features"
- Messages sorted by: [ date ] [ thread ] [ subject ] [ author ] [ attachment ]
From: "Don Williams" <dwilli10@san.rr.com> Date: Thu, 24 Apr 2003 19:38:13 GMT
Bela did a super job in working through this problem and it's another
indication
of the support provided to Open Server users. Intel refused, and still
refuses,
to recognize that the problem exists.
After upgrading to a "super" motherboard I couldn't see anything after the
initial
boot process. It took some typing in the blind to add a patch but it
worked.
Since my livelihood depends on having a working OSR machine I will probably
just stay with 5.0.6 for the time being, since it works. It would be nice
to have
some of the 5.07 features (USB being one I think) but for me it's not worth
the risk.
Thanks to SCO!
DAW
"Bela Lubkin" <belal@sco.com> wrote in message
news:20030424012637.A12603@sco.com...
> Don Williams wrote:
>
> > On a related subject there has been a problem with some Intel boards
which
> > improperly reported the type of display adapter. I had the
understanding
> > that the new release handled that. There has been a recent bios update
for
> > the Intel D845PEBT2. The current bios is P08. Wonder whether that
fixed
> > the problem?
>
> I can't answer the question about specific Intel motherboards and BIOS
> revisions. I spent some time (a couple months ago) looking at their
> BIOS update site; I learned that there are over 20 different D845-series
> motherboards; that some of them share BIOSes; that there were over 10
> different BIOSes among them; and that BIOS updates were not released for
> all models at the same time. I've seen reports that at least one of the
> BIOSes has been fixed to store proper values into the problematic CMOS
> bits.
>
> Meanwhile, I can give more substantive information about OSR5. OSR507
> was supposed to completely fix this problem, but it does not. I did not
> have access to a problem machine while developing the fix, so I was only
> able to put in a speculative fix which turned out to be insufficient.
> However, I also put in a user override which _does_ work.
>
> The user override is a bootstring: "cn.primary=vga". So if you have
> this problem on an OSR507 system, boot with:
>
> Boot
> : defbootstr cn.primary=vga
>
> A supplement has also been released, oss653a, which can be found in:
>
> ftp://ftp.sco.com/pub/openserver5/
>
> The supplement applies to OpenServer releases 5.0.4 through 5.0.7.
>
> Your description above summarizes the underlying cause of the problem,
> but not the symptoms, so I'll describe them here so other readers know.
>
> This problem is due to a BIOS bug, so it should only be seen on new
> systems or systems whose BIOS has been flash-updated. On new systems it
> would typically be seen at the beginning of OpenServer installation.
> After hitting return at the installation Boot: prompt, various dots are
> displayed ("Loading kernel .text" etc.) After that, normally the screen
> would clear and you would see OpenServer copyrights and hardware
> detection messages.
>
> On problem machines, the screen clears, but the copyrights and hardware
> messages aren't readable. The exact form of unreadability varies with
> different hardware. Some machines give a completely blank screen. Some
> give a blank screen in which a blinking cursor is visible, and moves
> around in an appropriate manner (if you had two machines side by side,
> both starting an install, you could operate the blind machine by
> reference to the normal machine and by comparing cursor positions).
> Finally, some machines display the correct text using a damaged font.
> The font looks like the top or bottom half of a normal font. For
> instance, the word "BELL" might look like a crude rendition of "DC||"
> because the bottom half of each letter was missing.
>
> Since the problem can also appear after a BIOS update, it might be seen
> on an already-installed system. (Another possibility would be a
> motherboard transplant. Replacing a motherboard, even with a supposedly
> identical model, might bring in a different BIOS revision.) Such a
> machine will likely boot up and be accessible over its network
> connections or serial ports, but not on the console. If you log in and
> run `hwconfig -h`, you would normally see:
>
> console - - - unit=vga type=0 12 screens=68k
>
> A problem system reports "unit=cga" or "unit=mono". This may also be
> visible on a system which displays the half-font symptom.
>
> For OSR507, use "cn.primary=vga". For earlier releases, use oss653a.
> oss653a can also be used on OSR507, with the potential advantage that
> you can boot arbitrary kernels that contain it without having to
> remember a bootstring.
>
> >Bela<
- Next message: Jean-Pierre Radley: "Re: Mozilla profile"
- Previous message: nope@none.com: "Re: cron will not run on its own"
- In reply to: Bela Lubkin: "Re: Open Server 5.0.7 versions and features"
- Next in thread: Don Williams: "Re: Open Server 5.0.7 versions and features"
- Messages sorted by: [ date ] [ thread ] [ subject ] [ author ] [ attachment ]
Relevant Pages
|
