Skip to contentSkip to conversation

Accesing data on (old) XFS devices

DISCUSSION Martin Mielke 2 participants 2 messages Feb 9, 2009

lore
Share & tools
lore
#0 lore raw

Hi all,

I have 2 SCSI disks which were used on an (old) Silicon Graphics running on IRIX 6.5.x I need to recover the data from those disks, dump them somewhere else and forget about the SGI disks...

OpenSUSE 11 sees the devices correctly and with fdisk I can see the partitions:

--

# fdisk -l /dev/sda

Disk /dev/sda (SGI disk label): 255 heads, 63 sectors, 4865 cylinders Units = cylinders of 16065 * 512 bytes

----- partitions -----

Pt#    Device  Info     Start       End   Sectors  Id  System
1: /dev/sda1  boot        66      2213  34510368   a  SGI xfs
2: /dev/sda2  swap         1        65   1048576   3  SGI raw
9: /dev/sda3               0         0      4096   0  SGI volhdr

11: /dev/sda4 0 2213 35563040 6 SGI volume
----- Bootinfo -----
Bootfile: /unix
----- Directory Entries -----

0: sash       sector    2 size  273408
1: ide        sector  536 size  249856
2: symmon     sector 1024 size 1113600
# fdisk -l /dev/sdd

Disk /dev/sdd (SGI disk label): 255 heads, 63 sectors, 9729 cylinders Units = cylinders of 16065 * 512 bytes

----- partitions -----

Pt#    Device  Info     Start       End   Sectors  Id  System
3: /dev/sdd1               1        16    262144   3  SGI raw
4: /dev/sdd2              17        32    262144   3  SGI raw
5: /dev/sdd3              33       293   4194304   3  SGI raw
6: /dev/sdd4             294       555   4194304   3  SGI raw
7: /dev/sdd5             556       816   4194304   3  SGI raw
8: /dev/sdd6             817      1077   4194304   3  SGI raw
9: /dev/sdd7               0         0      4096   0  SGI volhdr

11: /dev/sdd8 0 2213 35563040 6 SGI volume
----- Bootinfo -----
Bootfile: /unix
----- Directory Entries -----

--

I added the following to /etc/raw:
--
raw1:sdd1
raw2:sdd2
raw3:sdd3
raw4:sdd4
raw5:sdd5
raw6:sdd6
--

...but then I get this error:
--

# /etc/init.d/raw  start
bind /dev/raw/raw1 to /dev/sdd1...                                    failed
bind /dev/raw/raw2 to /dev/sdd2...                                    failed
bind /dev/raw/raw3 to /dev/sdd3...                                    failed
bind /dev/raw/raw4 to /dev/sdd4...                                    failed
bind /dev/raw/raw5 to /dev/sdd5...                                    failed
bind /dev/raw/raw6 to /dev/sdd6...                                    failed
--

Later on I found this thread:

http://oss.sgi.com/archives/xfs/2005-11/msg00376.html

and this FAQ:

http://linux.math.tifr.res.in/programming-doc/xfs/faq.html#xfsmountfail

when I run xfs_repair -n /dev/sda or /dev/sdd (-n = no modify mode) it yields the following message:
--
Sorry, could not find valid secondary superblock
Exiting now.
--

So...
my questions are: how destructive can a xfs_repair be in terms of destroying all the data (disk image is already made) be? and is there any way to access those devices??

I first asked this questions on the SuSE lists and I've been redirected here ;-)

TIA,
Martin

_______________________________________________
xfs mailing list
[email protected]
http://oss.sgi.com/mailman/listinfo/xfs

#1 lore raw
Quoted 48 lines · Martin Mielke

Hi all,

I have 2 SCSI disks which were used on an (old) Silicon Graphics running on IRIX 6.5.x I need to recover the data from those disks, dump them somewhere else and forget about the SGI disks...

OpenSUSE 11 sees the devices correctly and with fdisk I can see the partitions:

--

# fdisk -l /dev/sda

Disk /dev/sda (SGI disk label): 255 heads, 63 sectors, 4865 cylinders Units = cylinders of 16065 * 512 bytes

----- partitions -----

Pt#    Device  Info     Start       End   Sectors  Id  System
1: /dev/sda1  boot        66      2213  34510368   a  SGI xfs
2: /dev/sda2  swap         1        65   1048576   3  SGI raw
9: /dev/sda3               0         0      4096   0  SGI volhdr

11: /dev/sda4 0 2213 35563040 6 SGI volume
----- Bootinfo -----
Bootfile: /unix
----- Directory Entries -----

0: sash       sector    2 size  273408
1: ide        sector  536 size  249856
2: symmon     sector 1024 size 1113600
# fdisk -l /dev/sdd

Disk /dev/sdd (SGI disk label): 255 heads, 63 sectors, 9729 cylinders Units = cylinders of 16065 * 512 bytes

----- partitions -----

Pt#    Device  Info     Start       End   Sectors  Id  System
3: /dev/sdd1               1        16    262144   3  SGI raw
4: /dev/sdd2              17        32    262144   3  SGI raw
5: /dev/sdd3              33       293   4194304   3  SGI raw
6: /dev/sdd4             294       555   4194304   3  SGI raw
7: /dev/sdd5             556       816   4194304   3  SGI raw
8: /dev/sdd6             817      1077   4194304   3  SGI raw
9: /dev/sdd7               0         0      4096   0  SGI volhdr

11: /dev/sdd8 0 2213 35563040 6 SGI volume
----- Bootinfo -----
Bootfile: /unix
----- Directory Entries -----

--

Quoted 9 lines

I added the following to /etc/raw:
--
raw1:sdd1
raw2:sdd2
raw3:sdd3
raw4:sdd4
raw5:sdd5
raw6:sdd6
--

You shouldn't need to play games w/ raw devices, I think.

I'd try:

# file -s /dev/sdd?
# file -s /dev/sda?

and see which partitions claim to hold xfs filesystems (I can't recall offhand how the sgi partitioning works...)

Those *should* just mount right up. If they have dirty irix-format logs, you'll need to either mount (to replay the log) & unmount cleanly under irix, or zap it with xfs_repair -L

-Eric

_______________________________________________
xfs mailing list
[email protected]
http://oss.sgi.com/mailman/listinfo/xfs