Btrfs devel - Nov 2009 - fsck.btrfs assertion failure with large number of disks in fs

Hello all

We are experimenting with btrfs and we''ve run into some problems.

We are running on two Sun Storage J4400 Arrays containing a total of
48 1 TB disks.

With 24 disks in the btrfs:

# mkfs.btrfs /dev/sd[b-y]

WARNING! - Btrfs Btrfs v0.19 IS EXPERIMENTAL
WARNING! - see http://btrfs.wiki.kernel.org before using

adding device /dev/sdc id 2
...
adding device /dev/sdy id 24
fs created label (null) on /dev/sdb
        nodesize 4096 leafsize 4096 sectorsize 4096 size 21.83TB
Btrfs Btrfs v0.19

fsck.btrfs runs fine:

# sudo fsck.btrfs /dev/sdb
found 36864 bytes used err is 0
total csum bytes: 0
total tree bytes: 36864
total fs tree bytes: 4096
btree space waste bytes: 26174
file data blocks allocated: 0
 referenced 0
Btrfs Btrfs v0.19

Now with one more disk:

# mkfs.btrfs /dev/sd[b-z]

WARNING! - Btrfs Btrfs v0.19 IS EXPERIMENTAL
WARNING! - see http://btrfs.wiki.kernel.org before using

adding device /dev/sdc id 2
...
adding device /dev/sdz id 25
fs created label (null) on /dev/sdb
        nodesize 4096 leafsize 4096 sectorsize 4096 size 22.74TB
Btrfs Btrfs v0.19

# fsck.btrfs /dev/sdb
fsck.btrfs: volumes.c:1367: btrfs_read_sys_array: Assertion `!(ret)''
failed.
Aborted

mkfs.btrfs compiled from the Git repository fails with the same error message.

Other pertinent information:

# uname -a
Linux blade1 2.6.31-14-generic #48-Ubuntu SMP Fri Oct 16 14:05:01 UTC
2009 x86_64 GNU/Linux

We are investigating btrfs for a long term project, so we would be
happy to provide any debug information or test any kernel or
btrfs-progs patches.

Regards,

Albert
--
To unsubscribe from this list: send the line "unsubscribe linux-btrfs"
in
the body of a message to majordomo@vger.kernel.org
More majordomo info at  http://vger.kernel.org/majordomo-info.html

On Fri, Nov 20, 2009 at 8:31 AM, Albert Strasheim <fullung@gmail.com>
wrote:
> We are experimenting with btrfs and we''ve run into some problems.
> We are running on two Sun Storage J4400 Arrays containing a total of
> 48 1 TB disks.
> With 24 disks in the btrfs:
> Now with one more disk:
False alarm. Turns out the server was seeing the same JBOD twice,
instead of seeing both JBODs, so trying to use more than 24 disks of
course led to a corrupted file system.

Regards,

Albert
--
To unsubscribe from this list: send the line "unsubscribe linux-btrfs"
in
the body of a message to majordomo@vger.kernel.org
More majordomo info at  http://vger.kernel.org/majordomo-info.html