Bly, MJ (Martin) wrote:

>Hi,
>
>Yes, I had this problem (or similar).  Last thing on a Friday I was trying
>to install a dual *Opteron* system with x86_64 SL3.0.3 and encountered
>exactly the symptoms you describe.  There is a `fix' for it: add idle=poll
>to the kernel parameters.
>
>This problem is alleged to be caused by a bios bug - I updated the bios on
>the box concerned and lo the problem had gone.
>
>
>
Ok, thank. I will have a look to the current release of our BIOS to see
if I can
update that. I will also add this kernel option into the %post script in
order
to see if it solve the problem.
By the way we never met that problem with our opterons :o)

Thank.

>Now as to whether a Nocona would suffer a similar problem I don't know but
>you could try it and see.
>
>I'm trying a Nocona system but so far I've not managed to get it to install
>cleanly at all - all sorts problems too numerous to recount.
>
>
>
Good luck.

--
Yannick

>Martin.
>
>-----Original Message-----
>From: [log in to unmask]
>[mailto:[log in to unmask]] On Behalf Of Perret
>Yannick
>Sent: 17 November 2004 09:20
>To: [log in to unmask]
>Cc: Scientific Linux maillist
>Subject: SL, Nocona & 64 bits troubles
>
>
>Hello,
>
>we've got some new machines, which are Intel Xeon "Nocona".
>
>These processors are supposed to be similar to the Opteron ones,
>with the support for 64 and 32 bits instructions.
>By the way, the /proc/cpuinfo for these machines (installed with
>SL in 3.02) gives
>flags           : fpu vme de pse tsc msr pae mce cx8 apic sep mtrr pge
>mca cmov pat pse36 clflush dts acpi mmx fxsr sse sse2 ss ht tm lm
>The last one is for "long mode" with is supposed to be be the
>32+64 bits mode.
>
>I manage to install both a SL 3.02 and a SL 3.03 in 32 bits for
>these machines, but the 64 bits install failed.
>I use the x86_64 scheme for that, all all is fine (PXE boot
>on the x86_64 boot kernel, anaconda installation...) until the
>machine reboots.
>At this point the kernel starts to load and crashes after a
>'kernel panic'. The only message I can see is a dump of the
>execution stack...
>
>Does someone met the same problem ? Is it the good arch
>to use (x86_64) ?
>It is not really important as they are used in 32 bits at this
>time, but I don't like that :o)
>
>Thank.
>
>--
>Yannick Perret
>
>