[6011] in Release_7.7_team
Re: Beta Installer Bug?
daemon@ATHENA.MIT.EDU (andrew m. boardman)
Mon Jun 2 16:34:39 2008
Message-Id: <200806022034.m52KYRnS011300@pothole.mit.edu>
To: Jonathan Reed <jdreed@MIT.EDU>
cc: release-team@MIT.EDU
In-Reply-To: Your message of "Tue, 27 May 2008 09:59:10 EDT."
<A1572593-8857-4F67-BF07-8C3C1A44A02A@mit.edu>
Date: Mon, 02 Jun 2008 16:34:27 -0400
From: "andrew m. boardman" <amb@MIT.EDU>
X-Spam-Flag: NO
X-Spam-Score: 0.00
The same logic is actually used in both the regular and beta installers;
it's in the bit that lives in AFS. (If you're curious, see
/afs/dev/system/rhlinux/installer/phase2.backend, line 1160 and
subsequent.)
> The installer correctly determined that something was installed on
> sda1 and asked me to create a boot menu entry for it, but it also
> insisted something was installed on sda7, even though that's
> impossible. Does it automatically assume any Linux partitions that
> aren't being used for the install have something on them? Or have I
> encountered a bug?
The former, basically. It assumes that any partition that's not extended
and not one of the four specified partitions (boot/root/swap/cache) could
be another operating system, thus erring on the side of overprompting.
It doesn't look at filesystem contents at all, just the partition table.
This seems like a reasonably conservative approach, and indeed it's
possible you could want a grub.conf entry pointing at the linux install
you were about to put on sda7.
That said, if I'm missing something with the above logic, please let me
know. Now that the new media is basically done I'm collecting a queue of
minor improvements for the AFS side of the installer, and if there are
reasonable heuristics for limiting the boot options I'd be happy to add
them later this summer.
thanks for the report,
andrew
Andrew Boardman/ZIPS/ZINC/ISDA/IS&T/MIT +1.617.253.4040 amb@mit.edu
