Subject: Re: Great Success .... dmesg!
To: None <email@example.com>
From: Havard Eidnes <firstname.lastname@example.org>
Date: 07/06/2004 11:26:29
> below, the dmesg dump from my Server running 2.0 Kernel.
> I think 2.0 is getting to become one of best Releases ever.
That's good to hear! However, I have to point out that you are
running -current, not a beta release of 2.0. Admittedly, the
changes between the two may not be huge at this moment. For further
information about how 2.0E and 2.0_BETA (which will become 2.0
proper when it's ready) see
It's nice to hear your benchmark test results as well, of course!
> If you select "whole disk" during Installation, NetBSD is not able to=
> because it installs itself in the 3rd Hard Disk Partition.
> If you select "part of disk" and then choose the 1st Partition 0 then=
> everything works fine.
That sounds like a bug to me. Can you please report this formally
by submitting a bug report with send-pr? Of interest would be the
result of "fdisk" on the boot disk, as well as an "od -c" dump of
the first 512 bytes of the physical boot disk (/dev/rld0d) as well
as the NetBSD part (/dev/rld0c) when sysinst has done it's thing.
That is, if you can afford to reinstall...
> Frag Size & Block Size:
> This was the first time, I saw, that NetBSD could not newfs. I had to=
> the frag & blocksize. (this should be done automatically)
Yes, it should. Again, could you please submit a bug report with
the exact error message, and what you had to adjust the fragment and
block sizes to?
> There were two 120GB running RAID 0 (mirror) recognized as ld0.
RAID 0 is commonly referred to as striping, but since your ld0
reports 111 (binary) GB, you must be doing RAID 1 (mirroring). I
don't know, but it's conceivable that the drive geometry in the BSD
disklabel makes a difference which might explain why you didn't
observe the problem on a single-120GB-disk system. Since ld0 is
only addressed by block number, I guess you could tweak the drive
geometry in the BSD disklabel and thereby possibly avoid having to
adjust the block and fragment sizes. I don't know what it would
take to produce more reasonable defaults; I'm guessing the drive
geometry comes from the BIOS on the twe controller.