Subject: Re: sysinst segfaults near end of installation
To: jon <TYR124840@tyler.net>
From: Manuel Bouyer <firstname.lastname@example.org>
Date: 04/07/2002 21:50:42
On Sun, Apr 07, 2002 at 01:56:17PM -0500, jon wrote:
> Don't think so - like I said the values are ALWAYS off by one, not just in
> some instances, and the exact same off-by-one situation occurs also with
OK, so it's an error in the rounding parameters I guess.
> > >
> > > by the way, i can duplicate the above for any size value, its ALWAYS
> > > increased by one when reported later
> > >
> > > then during configuration of the disklabel, the size of the netbsd part
> > > the disk (/dev/wd0c) is reported as 9998, instead of 9999
> > This may also be a rounding problem. What do the sectors number looks like
> Making my NetBSD partition, I choose 0 as the offset since it's the first on
> my disk, and that should automatically reserve track 1. But it reports the
> offset as 63, not 64. Then, for size, I enter -1, so that it should end at
63 is, OK, it starts counting at 0.
> the beginning of my second partition, which is of type 0C. But the size of
> the NetBSD partition is reported 20582812 and the end 20582875, which is the
> same sector as my next partition begins on.
I think this is OK too, end point to the start of the next partition.
> During disklabel config though, the size is still reported 20582812, but the
> end this time is reported 20582874, not ...75.
> > >
> > > again, the above i can duplicate
> > >
> > > then, when i then enter 0 as the offset for /dev/wd0a and 9999 as the
> > > i end up with 9998
> > You make wd0a the whole NetBSD partition, so you don't have any swap
> > partition ?
> > This may be because of this sysinst dumps core.
> Like I said, I've got a gig of ram, so I really don't think I have much need
> of swap. I'm thinking this would be a bug with sysinst...
> Possibly I could circumvent it in the short term by enlarging my ramdisk,
> but I don't know how to do that unfortunately.
No enlarging the ramdisk won't help. The "file system full" is a consequence
of sysinst dumping core, not the cause. The cause is likely to be a bug
in sysinst, which doesn't check if a swap partition exists or not.
The workaround in your case is probably to make a very small swap parition
(1 cylinder for example)
Manuel Bouyer <email@example.com>