2001-10-09 22:13:15

by Daniel Freedman

[permalink] [raw]
Subject: 'set_blocksize' & 'nr_blocks changed' during Raid1 mount...


Hi,

I'm looking for clarification on the following kernel messages that I
received during a mount of a recently-created raid1 device. I'm using
a 2.2.19 kernel (close to stock, actually Debian's Potato kernel,
which is stock +~ BigPhysArea patch), which I've then patched with
Ingo Molnar's 0.90 RAID layer (I used the correct 2.2.19 patches).

Here's what I did to create the /dev/md0 device:

# partition /dev/sda5 and /dev/sdb5 to participate in raid1
# create appropriate /etc/raidtab (listed at end of email)
mkraid /dev/md0
# following commands were executing while partitions were
# still re-syncing (if that's relevant)
mke2fs /dev/md0
# /dev/md0 is configured as /home as listed in fstab (at end of email)
mount /home

I then received the following kernel messages on mounting /dev/md0:

Oct 9 15:56:49 newton kernel: set_blocksize: b_count 1, dev md(9,0), block 31058688, from 00000900
Oct 9 15:56:49 newton kernel: set_blocksize: b_count 1, dev md(9,0), block 31058689, from 00000900
Oct 9 15:56:49 newton kernel: set_blocksize: b_count 1, dev md(9,0), block 31058690, from 00000900
Oct 9 15:56:49 newton kernel: set_blocksize: b_count 1, dev md(9,0), block 31058691, from 00000900
Oct 9 15:56:49 newton kernel: set_blocksize: b_count 1, dev md(9,0), block 31058692, from 00000900
<snip>
Oct 9 15:56:49 newton kernel: set_blocksize: b_count 1, dev md(9,0), block 31058807, from 00000900
Oct 9 15:56:49 newton kernel: set_blocksize: b_count 1, dev md(9,0), block 31058808, from 00000900
Oct 9 15:56:49 newton kernel: set_blocksize: b_count 1, dev md(9,0), block 31058809, from 00000900
Oct 9 15:56:49 newton kernel: set_blocksize: b_count 1, dev md(9,0), block 31058810, from 00000900
Oct 9 15:56:49 newton kernel: set_blocksize: b_count 1, dev md(9,0), block 31058811, from 00000900
Oct 9 15:56:49 newton kernel: set_blocksize: b_count 1, dev md(9,0), block 31058812, from 00000900
Oct 9 15:56:49 newton kernel: set_blocksize: b_count 1, dev md(9,0), block 31058813, from 00000900
Oct 9 15:56:49 newton kernel: set_blocksize: b_count 1, dev md(9,0), block 31058814, from 00000900
Oct 9 15:56:49 newton kernel: set_blocksize: b_count 2, dev md(9,0), block 31058815, from 00000900
Oct 9 15:56:49 newton kernel: md0: blocksize changed during read
Oct 9 15:56:49 newton kernel: nr_blocks changed to 32 (blocksize 4096, j 7764672, max_blocks 15868160)

Interestingly enough, I now no longer get any kernel messages when I
mount or unmount the raid1 partition. Could this possibly be because
md0's no longer in the resync process? I did a linux-kernel and
linux-raid mailing list search as well as a google and usenet search
and found a few messages posting the same or related scenaria but no
replies to their questions. Casually examining the source code did
not provide any obvious clues to me. Any advice is greatly
appreciated.

Thanks,
Daniel

ps: kindly cc me on replies if possible.





Additional info:

newton:~# fdisk -l /dev/sda

Disk /dev/sda: 255 heads, 63 sectors, 8924 cylinders
Units = cylinders of 16065 * 512 bytes

Device Boot Start End Blocks Id System
/dev/sda1 1 8924 71681998+ 5 Extended
/dev/sda5 1 7902 63472752 fd Linux raid autodetect
/dev/sda6 7903 8924 8209183+ fd Linux raid autodetect
newton:~# fdisk -l /dev/sdb

Disk /dev/sdb: 255 heads, 63 sectors, 8924 cylinders
Units = cylinders of 16065 * 512 bytes

Device Boot Start End Blocks Id System
/dev/sdb1 1 8924 71681998+ 5 Extended
/dev/sdb5 1 7902 63472752 fd Linux raid autodetect
/dev/sdb6 7903 8924 8209183+ fd Linux raid autodetect
newton:~# grep /home /etc/fstab
/dev/md0 /home ext2 rw 0 2
newton:~# cat /proc/mdstat
Personalities : [raid1]
read_ahead 1024 sectors
md0 : active raid1 sdb5[1] sda5[0] 63472640 blocks [2/2] [UU]
unused devices: <none>
newton:~# cat /etc/raidtab
raiddev /dev/md0
raid-level 1
nr-raid-disks 2
nr-spare-disks 0
chunk-size 4
persistent-superblock 1
device /dev/sda5
raid-disk 0
device /dev/sdb5
raid-disk 1
newton:~# e2fsck -n -f /dev/md0
e2fsck 1.18, 11-Nov-1999 for EXT2 FS 0.5b, 95/08/09
Pass 1: Checking inodes, blocks, and sizes
Pass 2: Checking directory structure
Pass 3: Checking directory connectivity
Pass 4: Checking reference counts
Pass 5: Checking group summary information
/dev/md0: 20/7946240 files (0.0% non-contiguous), 249369/15868160 blocks





--
Daniel A. Freedman
Laboratory for Atomic and Solid State Physics
Department of Physics
Cornell University


2001-10-10 00:44:57

by Daniel Freedman

[permalink] [raw]
Subject: Re: 'set_blocksize' & 'nr_blocks changed' during Raid1 mount... (more info)

On Tue, Oct 09, 2001, Daniel Freedman wrote:
>
> Hi,
>
> I'm looking for clarification on the following kernel messages that I
> received during a mount of a recently-created raid1 device. I'm using
> a 2.2.19 kernel (close to stock, actually Debian's Potato kernel,
> which is stock +~ BigPhysArea patch), which I've then patched with
> Ingo Molnar's 0.90 RAID layer (I used the correct 2.2.19 patches).
>
> Here's what I did to create the /dev/md0 device:
>
> # partition /dev/sda5 and /dev/sdb5 to participate in raid1
> # create appropriate /etc/raidtab (listed at end of email)
> mkraid /dev/md0
> # following commands were executing while partitions were
> # still re-syncing (if that's relevant)
> mke2fs /dev/md0
> # /dev/md0 is configured as /home as listed in fstab (at end of email)
> mount /home
>
> I then received the following kernel messages on mounting /dev/md0:
>
> Oct 9 15:56:49 newton kernel: set_blocksize: b_count 1, dev md(9,0), block 31058688, from 00000900
> Oct 9 15:56:49 newton kernel: set_blocksize: b_count 1, dev md(9,0), block 31058692, from 00000900
> <snip>
> Oct 9 15:56:49 newton kernel: set_blocksize: b_count 2, dev md(9,0), block 31058815, from 00000900
> Oct 9 15:56:49 newton kernel: md0: blocksize changed during read
> Oct 9 15:56:49 newton kernel: nr_blocks changed to 32 (blocksize 4096, j 7764672, max_blocks 15868160)
>
> Interestingly enough, I now no longer get any kernel messages when I
> mount or unmount the raid1 partition. Could this possibly be because
> md0's no longer in the resync process? I did a linux-kernel and


(responding to my own message):
Hi,

I experimented further with mounting my raid1 partitions during the
resync process (this is allowed, according to the linuxdoc HOWTO) and
I receive the above kernel messages whenever I do so. I've not yet
ever seen these messages on mounting one of my raid1 partitions when
it is not resyncing.

I'd be happy to try further suggestions to give additional info.

Thanks so much,

Daniel



> Thanks,
> Daniel
>
> ps: kindly cc me on replies if possible.
>

--
Daniel A. Freedman
Laboratory for Atomic and Solid State Physics
Department of Physics
Cornell University