Re: how to tell when biosboot partition is needed?

[Date Prev][Date Next][Thread Prev][Thread Next][Date Index][Thread Index]

 



On Thu, Aug 16, 2012 at 02:10:40PM +0200, Richard Neuboeck wrote:
> On 16.08.12 09:46, Vratislav Podzimek wrote:
> > On Wed, 2012-08-15 at 18:29 -0700, Steve Rikli wrote:
> >> On Tue, Aug 14, 2012 at 09:08:10AM +0200, Richard Neuboeck wrote:
> >>> I've got an answer from David Lehman how the decision process for the
> >>> usage of biosboot works in anaconda:
> >>>
> >>> Here is the entire decision tree:
> >>>
> >>> - did the system boot in EFI mode or BIOS mode?
> >>>   - EFI
> >>>     - use gpt and never make biosboot
> >>>   - BIOS
> >>>     - is the disk larger than the max for msdos (2TB)?
> >>>       - yes
> >>>         - use gpt and ensure there's a biosboot partition
> >>>       - no
> >>>         - use msdos
> >>
> >> Either I'm not following that logic or I think something isn't behaving
> >> as advertised.  Here's my situation ....
> >>
> >> I have a bog standard Dell PowerEdge 1850 -- same test machine I used in
> >> the initial Fedora 16 work that prompted me to start this thread -- which
> >> has no EFI, and boots from an old & small (20GB) SCSI disk.
> >>
> >> If I read the above correctly, I should _not_ need a biosboot partition,
> >> yet a Fedora 16 Kickstart install fails with the known partitioning error
> >> mentioned in bugzillas et al, unless I create one.
> >>
> >> What am I missing?
> >
> > I believe this logic applies only to Fedora 17. In Fedora 16, there was
> > an attempt to have all systems using GPT (and thus some of them biosboot
> > partition), unless you used 'nogpt' command line option. However it
> > appeared that some systems have problems with such setups and thus
> > aforementioned logic was introduced in Fedora 17.
> 
> I can confirm, with several installs on different systems, that the
> decision tree David sent is working in Fedora 17 the way it should.

Alright, the behavior changed -- *that* I can understand.  :-)

> Fedora 16 didn't get used in our institution but the documentation
> states that gpt is the default. May be you can solve your problem by
> checking the distribution type you are installing and applying
> appropriate partition options. So you don't have to depend on the
> storage log file.

We specifically want to avoid doing anything like a lookup table
(e.g. "if F16 do this, if F17 do this" etc.) in favor of more
programmatic solutions.  Unless you mean something else when you
say "distribution type"?

We're OK with creating a biosboot partition whether it's strictly
needed or not.  Even on the older systems we still run, 1MB is not
going to hurt us.  But what we can't afford to do is try to install
a biosboot partition on distributions where it isn't yet supported
(e.g. CentOS 6.3) and cause an installation failure.

For a time we were forced to use a different %pre script for disk
partitioning between biosboot and non-biosboot installs, but that's
not very different from the lookup table idea, so is sub-optimal
for us.

Right now, grep'ing through the storage.log file is functional
enough for us in F16 & F17, without interfering with older/other
distributions, so it's satisfactory.  But some of the other ideas
in this thread (/sys/firmware/ etc.) may be useful for us too.

Thanks,
sr.

_______________________________________________
Kickstart-list mailing list
Kickstart-list@xxxxxxxxxx
https://www.redhat.com/mailman/listinfo/kickstart-list


[Index of Archives]     [Red Hat General]     [CentOS Users]     [Fedora Users]     [Fedora Maintainers]     [Fedora Desktop]     [Fedora SELinux]     [Big List of Linux Books]     [Yosemite News]     [KDE Users]

  Powered by Linux