Unable to mount XFS partition on Sun Ultra/Linux 2.6.17 - Function not implemented

Unable to mount XFS partition on Sun Ultra/Linux 2.6.17 - Function not implemented

[Dennis Varouxis]

I've got an ailing SGI which currently doesn't boot, so I plucked the disk out
and slotted it into my Sun box to snoop around. Unfortunately despite my
best technical and google skills I haven't been able to resolve this.

Attempting to mount the main disk partition with a simple "mount -t xfs
/dev/sda1 /mnt/sgidisk" yields only "mount: Function not implemented".

That's it, that's all I've got. Except that I know from debugging mount that the mount
syscall raises errno 90, which goes into strerror() which returns the "not
implemented" message.

It's a SCSI disk formatted to XFS by IRIX. IRIX and linux XFS are reportedly disk
compatible so I don't see a problem there. There's a boot partition (1), a swap
partition (2), and the 9 and 11 partitions used for the SGI disk label.

The host system is a Sun Ultra 5 (sparc64 sun4u) with an Adaptec aix7xxx
controller, running Gentoo Linux kernel 2.6.17-r8. The kernel has been built
with SGI partition map support. I've got the aic7xxx and xfs drivers compiled as modules 
and set to autoload (I've tried with built-in versions as well and get the same result).

The kernel at this point has been compiled with debug enabled and with CONFIG_XFS_DEBUG=y. 
The XFS driver confirms this in dmesg:

"SGI XFS with ACLs, security attributes, large block/inode numbers, debug
enabled
SGI XFS Quota Management subsystem"

However when I mount there's nothing emitted in /var/log/messages.

Note the partition type says "SGI efs". However trying to mount -t efs fails
with the expected "mount: wrong fs type". (EFS compiled as a module and
modprobed).

The output from xfs_db which admittedly doesn't mean much, to me at least,  is:

azrael ~ # xfs_db -c sb -c p /dev/sda1
magicnum = 0x58465342
blocksize = 4096
dblocks = 491395
rblocks = 0
rextents = 0
uuid = dbfdfd61-f556-1021-8e74-0800690ac5d3
logstart = 262148
rootino = 128
rbmino = 129
rsumino = 130
rextsize = 16
agblocks = 61425
agcount = 8
rbmblocks = 0
logblocks = 1000
versionnum = 0x1094
sectsize = 512
inodesize = 256
inopblock = 16
fname = "\000\000\000\000\000\000\000\000\000\000\000\000"
blocklog = 12
sectlog = 9
inodelog = 8
inopblog = 4
agblklog = 16
rextslog = 0
inprogress = 0
imax_pct = 25
icount = 35648
ifree = 1636
fdblocks = 164789
frextents = 0
uquotino = 0
gquotino = 0
qflags = 0
flags = 0
shared_vn = 0
inoalignmt = 2
unit = 0
width = 0
dirblklog = 0
logsectlog = 0
logsectsize = 0
logsunit = 0

Also interesting is the output from xfs_check:

"ERROR: The filesystem has valuable metadata changes in a log which needs to
be replayed.  Mount the filesystem to replay the log, and unmount it before
re-running xfs_check."

Although this indicates that the filesystems should at least be mountable.

Any help would be appreciated - what am I missing? Is there somewhere else
that the kernel and/or the xfs driver would log their debug messages?


thanks

dennis


features2 = 0

_________________________________________________________________
Be one of the first to try Windows Live Mail.
http://ideas.live.com/programpage.aspx?versionId=5d21c51a-b161-4314-9b0e-4911fb2b2e6d

Unable to mount XFS partition on Sun Ultra/Linux 2.6.17 - Function not implemented

[f1sh]

I've got an ailing SGI which currently doesn't boot, so I plucked the disk
out
and slotted it into my Sun box to snoop around. Unfortunately despite my
best technical and google skills I haven't been able to resolve this.

Attempting to mount the main disk partition with a simple "mount -t xfs
/dev/sda1 /mnt/sgidisk" yields only "mount: Function not implemented".

That's it, that's all I've got. Except that I know from debugging mount that
the mount
syscall raises errno 90, which goes into strerror() which returns the "not
implemented" message.

It's a SCSI disk formatted to XFS by IRIX. IRIX and linux XFS are reportedly
disk
compatible so I don't see a problem there. There's a boot partition (1), a
swap
partition (2), and the 9 and 11 partitions used for the SGI disk label.

The host system is a Sun Ultra 5 (sparc64 sun4u) with an Adaptec aix7xxx
controller, running Gentoo Linux kernel 2.6.17-r8. The kernel has been built
with SGI partition map support. I've got the aic7xxx and xfs drivers
compiled as modules and set to autoload (I've tried with built-in versions
as well and get the same result).

The kernel at this point has been compiled with debug enabled and with
CONFIG_XFS_DEBUG=y. The XFS driver confirms this in dmesg:

"SGI XFS with ACLs, security attributes, large block/inode numbers, debug
enabled
SGI XFS Quota Management subsystem"

However when I attempt the mount there's nothing emitted in
/var/log/messages.

Note the partition type says "SGI efs". However trying to mount -t efs fails
with the expected "mount: wrong fs type". (EFS compiled as a module and
modprobed).

The output from xfs_db which admittedly doesn't mean much, to me at least, 
is:

azrael ~ # xfs_db -c sb -c p /dev/sda1
magicnum = 0x58465342
blocksize = 4096
dblocks = 491395
rblocks = 0
rextents = 0
uuid = dbfdfd61-f556-1021-8e74-0800690ac5d3
logstart = 262148
rootino = 128
rbmino = 129
rsumino = 130
rextsize = 16
agblocks = 61425
agcount = 8
rbmblocks = 0
logblocks = 1000
versionnum = 0x1094
sectsize = 512
inodesize = 256
inopblock = 16
fname = "\000\000\000\000\000\000\000\000\000\000\000\000"
blocklog = 12
sectlog = 9
inodelog = 8
inopblog = 4
agblklog = 16
rextslog = 0
inprogress = 0
imax_pct = 25
icount = 35648
ifree = 1636
fdblocks = 164789
frextents = 0
uquotino = 0
gquotino = 0
qflags = 0
flags = 0
shared_vn = 0
inoalignmt = 2
unit = 0
width = 0
dirblklog = 0
logsectlog = 0
logsectsize = 0
logsunit = 0
features2 = 0

Also interesting is the output from xfs_check:

"ERROR: The filesystem has valuable metadata changes in a log which needs to
be replayed.  Mount the filesystem to replay the log, and unmount it before
re-running xfs_check."

Although this indicates that the filesystems should at least be mountable.

Any help would be appreciated - what am I missing? Is there somewhere else
that the kernel and/or the xfs driver would log their debug messages?


thanks

dennis

-- 
View this message in context: http://www.nabble.com/Unable-to-mount-XFS-partition-on-Sun-Ultra-Linux-2.6.17---Function-not-implemented-tf2781318.html#a7759642
Sent from the linux-xfs mailing list archive at Nabble.com.

Re: Unable to mount XFS partition on Sun Ultra/Linux 2.6.17 - Function not implemented

[Eric Sandeen]

Dennis Varouxis wrote:

> The output from xfs_db which admittedly doesn't mean much, to me at least,  is:
> 
> azrael ~ # xfs_db -c sb -c p /dev/sda1
...

> logblocks = 1000
> versionnum = 0x1094

Looks like version one directories, recently ripped out of Linux...

Try an older kernel... can't recall exactly when this changed, but you 
can probably find it in the archives, search "dirv1" I think.

> Also interesting is the output from xfs_check:
> 
> "ERROR: The filesystem has valuable metadata changes in a log which needs to
> be replayed.  Mount the filesystem to replay the log, and unmount it before
> re-running xfs_check."

Linux can't replay irix logs, you could zero out the log with xfs_repair 
-L, or, less destructively, mount with -o ro,norecovery and see what you 
can get.

-Eric

Re: Unable to mount XFS partition on Sun Ultra/Linux 2.6.17 - Function not implemented

[Chris Wedgwood]

On Fri, Dec 08, 2006 at 10:50:46AM -0600, Eric Sandeen wrote:

> Looks like version one directories, recently ripped out of Linux...

there should be a prinkt to this effect for these cases

Re: Unable to mount XFS partition on Sun Ultra/Linux 2.6.17 - Function not implemented

[Eric Sandeen]

Chris Wedgwood wrote:
> On Fri, Dec 08, 2006 at 10:50:46AM -0600, Eric Sandeen wrote:
> 
>> Looks like version one directories, recently ripped out of Linux...
> 
> there should be a prinkt to this effect for these cases
> 
There was some mount error message logic that was backwards, that might 
be why there was none.

see:

TAKE 958469 - Quiet mounts on XFS aren't

The XFS quiet mount logic was inverted making quiet
mounts noisy and vice versa. Fix it.



Haven't looked too closely but that may well be the problem.

-Eric