On Fri, Jun 27, 2014 at 12:37:33PM -0700, Scott Vanderbilt wrote: > On 6/27/2014 12:10 PM, Stuart Henderson wrote: > >On 2014-06-26, Scott Vanderbilt <li...@datagenic.com> wrote: > >>Having done a little man page reading on boot-time configuration, I > >>learned about the existence of ukc. I'm wondering whether something like > >> > >> ukc> disable acpi0 > >> > >>might circumvent the kernel panic and allow the boot to successfully > >>complete. I'm hoping that since this is a server, ACPI is non-essential. > >>Just grasping at straws in an effort to get this machine up and running > >>again. > > > >I think you should consider ACPI essential on pretty much any x86 > >machine from the last 4-5 years or so - servers, laptops, standard PCs. > > > >In an emergency such as this you might get away with it briefly, but > >some devices are likely not to work, and it's not recommended leaving > >it like that for any length of time, ACPI is involved in a lot of > >system controls (thermal controls, power etc) and most modern machines > >are just not designed/tested to work without it. > > > > Thanks for clarifying. > > Disabling acpi was only meant to be a stopgap measure so I could get > around the assertion in the kernel that caused a panic on boot. Once > I was able to boot the machine, I upgraded to a later snapshot in > which the assertion was removed. I never intended to permanently > disable acpi. As the machine was at a remote co-lo, I felt had no > other choice. > > Is there some better way that I should have handled this situation? >
Keep another kernel in / that is known working. -ml