[Top][All Lists]
[Date Prev][Date Next][Thread Prev][Thread Next][Date Index][Thread Index]
Re: address@hidden: Bug#108231: BSD label FS type wrong]
From: |
Andrew Clausen |
Subject: |
Re: address@hidden: Bug#108231: BSD label FS type wrong] |
Date: |
Mon, 20 Aug 2001 20:47:36 +1000 |
User-agent: |
Mutt/1.2.5i |
Hi Timshel,
On Mon, Aug 20, 2001 at 08:13:57PM +1000, Timshel Knoll wrote:
> Seems that the disklabel type is being incorrectly listed as bsd, I
> don't know why.
He didn't complain about that. Doesn't Digital Unix use BSD labels?
> Also, partition 2 is really ext2, but is being detected
> as linux-swap - I believe others have also experienced this problem. Is
> it fixed in 1.4.18?
There haven't been any changes wrt ext2 detection in 1.4.18. There
HAVE been for ext3 getting mixed up with linux-swap. Maybe he has an
ext3 fs, with an ext2 code in the table?
The root of the bug is in mke2fs not zeroing out blocks 4-7 for 4k blocks.
I've told Ted (e2fsprogs maintainer), and he said he will/has fixed it
(couple of days ago).
> Also, partition 4 (which is part of a md device) is being detected as
> ext2. I assume that this is either the first device of a raid 0
> (striping) md volume, or part of a raid 1 (mirroring) md device, which
> would explain an ext2 superblock being at the start of the disk.
I assume so too.
> There's
> probably not much that can be done to address this for raid 1 stuff, as
> it should look just like a single partition AFAIK, but there would be
> issues in the former case (ie. raid 0 - the superblock may be on the
> first disk, but the filesystem is striped across 2 or more disks) ... is
> there any way to address this "issue"?
In the long term, we can support RAID / LVM, etc., although this is
a big effort, although one that seems to be in demand. I'm interested
in doing it, but it's really hard :/
We could also hack up a "RAID-fs". But, the problem is purely cosmetic,
right?
Thanks!
Andrew