SUMMARY Re: weird Sun Blade 100 lockup problem

Thanks to the following people and their suggestions.

I have removed the Belkin KVM switch.  I am not running xntpd, the machine
is not overheating (as per prtdiag -v), and I don't run a screen saver in
either Windows or Solaris.

If anyone has any clue about the latest patch bundle, though, I'd
appreciate a little more help.  I grabbed the Solaris 8 patch cluster
(9/26/2001) from Sunsolve, but *none* of the patches will install.  Some
are code 25, but most are code 2 or code 5.  Code 2 says patch already
installed, and code 5 is a pkgadd error.  In the detailed logs, almost all
the patches list as already installed, but showrev -p shows that they are
*not*.  I tried installing after booting to single user mode, so I don't
know where the problem could be.

Thanks,
-Adam

Responses follow:
Alex Stade <alex@trdlnk.com>
Belkin switch lockups

Dan Astoorian <djast@cs.toronto.edu>
xntpd causes problems -- read BugID 4461525 smbus_intr_timeout() priority
inversion induces wedge
use ntpdate instead.
Also, change xntpd's priority priocntl -s -c TS -p 30 -m 30 `pgrep xntpd`
Add set smbus:intr_timeout = 100000 to /etc/system
Also, to fix clock issue, add set sys_tick_freq = 5549020 to /etc/system

Olivier Masse <olivier@prosys.ca>
Check if machine is overheating

Marcelino Mata <mmata@multiinc.com>
Don't run a screen saver with the SunPCI windows installation

Chan C <ctcao@hotmail.com>
Apply the latest patch bundle.

Lars Hecking <lhecking@nmrc.ie>
xntpd

David A Stewart <David.Stewart@ee.ed.ac.uk>
Is the Blade connected to an HP 4108gl Procurve switch?  Problems when
switch reboots.

Original question follows:

On Fri, 28 Sep 2001, Adam and Christine Levin wrote:
>
> I have a Sun Blade 100 running Solaris 8 04/01 with a patchset from around
> the beginning of August.
>
> Twice in the last couple of months, the system's just *stopped*.  It's my
> desktop workstation.
>
> It's got 512MB memory, plus a SunPCI II card with 384MB memory running
> Windows 2000 server, but not doing much (sits idle most of the time, but
> booted).
>
> I run dozens of xterms, plus star office and netscape, all in CDE.  The
> xterms are all ssh sessions to remote hosts on our LAN/WAN.  I also run
> xclock, and an xbiff window fired back from our mail server.
>
> The first time it happened, I was on the phone, and when I turned back to
> the machine and started typing, I got nothing.  The keyboard was frozen,
> the mouse was frozen.  I tried telnetting in from another machine,
> nothing.  I tried pinging -- nothing.  No network activity on the switch.
> It was just frozen.
>
> The second time, I was in the middle of typing something (in a vi session
> on a remote Sun box -- an E250), and again, it just froze up.  Same thing
> as before -- no connectivity, no response.
>
> My keyboard is a lexmark selectease, so I don't have a stop-a available to
> me, but the keyboard is completely dead (the lights don't toggle on the
> lock keys), so I don't think a stop-a would do anything.
>
> The keyboard and mouse are connected through a Belkin KVM switch and a
> ps2->USB converter.
>
> It could be a million things.  Has anybody seen this before?  My first
> step is going to be applying the latest patch sets and getting rid of the
> KVM.
>
> -Adam
>
>
>