CentOS - Feb 2016 - "upstream testing"??

On 02/07/2016 04:09 PM, Alexander Dalloz wrote:
> Am 07.02.2016 um 22:00 schrieb Bear Tooth:
>>     [Follow-ups set to gmane.linux.centos.general]
>>
>>       My wife had been running CentOS 6.4 almost since
>>     its inception; then her PC broke down.
>>
>>          We got a PC from System76, and Ubuntu turned out
>>     utterly unsuitable for us, as expected -- as bad
>>       for us as Gnome3. (I had previously bought a System76
>>     net book (starling iirc), and immediately installed the
>>     then current Fedora; all has been well with that.
>>
>>          This time, alas!, I thought I should let her try Ubuntu;
>>     so I tried running it myself for an houror two
>>     to get it set up and tweaked.
>>
>>          I couldn't even find any of the apps I wanted to tweak!
>>     So I put in an install disk for CentOS, and rebooted.
>>
>>          It never came near finishing the reboot. Up popped the
>>     following:
>>
>>
>>                  Detected CPU family 6 model 94.
>>
>>                  Warning: Intel CPU model -- this hardware has not
>>                  undergone upstream testing. Please see
>>
>>                  http://wiki.centos.org/FAQ for more information.
>>
>>                  tsc: Fast TSC calibration failed.
>>
>>          I have consulted that FAQ and more, and also System76's.
>>     I've consulted and tried more other things than most of you
>>     likely want to hear about. No joy.
>>
>>          I've also tried rebooting without any install disk, with a
>>     Fedora install disk, with various helps such as super grub disk,
>>     and finally even with DBAN.
>>
>>          The machine doesn't even find any of those. On any reboot,
it
>>     just goes to that CentOS error message, and stops.
>>
>>     I've also googled for '"upstream testing"
hardware'
>>
>>     Any thoughts or experience??.
>>
>>
>
>
> Did you try adding to the kernel line the parameter
"clocksource=tsm"
> or "clocksource=acpi_pm"?
>
> Alexander
>
> _______________________________________________
> CentOS mailing list
> CentOS at centos.org
> https://lists.centos.org/mailman/listinfo/centos
Just a thought, but maybe try doing the "Unetbootin/.ISO file build"
on
another pc / laptop and attempt booting from the USB instead of a CD? 
Just my thoughts on the matter. Its something I would do just to get the 
OS installed, then I'd worry about the upstream stuff 
afterwards...perhaps after the install and a tremendous system-wide 
upgrade, things might look a little better? PLUS she'd at least have the 
OS on her machine....I'm just sayin' LoL!


EGO II

Which System76 model? How is the install media created? Presumably it's a
USB stick, but how is it being created?

The easiest and most reliable is to use dd. Livecd-tools is also reliable
but has a number of options required to boot UEFi systems. LiveUSB Creator
should work. Everything else is prone to failure.

CentOS 6.4 is kinda old for new hardware. You're better off looking at
CentOS 7.1.


Chris Murphy

On Mon, Feb 8, 2016 at 1:35 AM, Chris Murphy <lists at colorremedies.com>
wrote:
> Everything else is prone to failure.
Specifically, but not limited to, unetbootin. Really, people need to
just purge unetbootin from memory and stop recommending it. I've never
had it work on any (U)EFI system. And more often than not it would
fail to create media even for BIOS systems. Also, it's not supported
at all on Fedora, I seriously doubt it's supported by CentOS or Red
Hat. Just use dd, and accept the obliteration of the USB stick. That's
easy and safe.

Future talk. There's a rewrite happening on the Fedora side for
LiveUSB Creator that will initially use dd on the backend. I'll guess
that it'll still accept being pointed to a local ISO. There's some
talk about it hopefully being more modular so it can be "branded" by
different distros and hence more widely used, maintained, and
reliable.


-- 
Chris Murphy