Disk Failure?



Hi everyone,

Yesterday I had a disk fail on a SunFire 280R Solaris 8.
The questions first:
Can the experts confirm my opinion that this is a hardware problem -
disk failure?
Have you any suggestions for recovering data from this drive (I do have
backups, but I would still lose some important data)?

Thanks for any help
Richard Butler


Symptoms:
This machine had two 72G disks (not mirrored) and during reboot after
installing the latest recommended patches I get the warning:
...
Jun 14 12:06:30 ed pcisch: [ID 370704 kern.info] PCI-device: SUNW,qlc@4,
qlc0
Jun 14 12:06:30 ed genunix: [ID 936769 kern.info] qlc0 is
/pci@8,600000/SUNW,qlc@4
Jun 14 12:06:30 ed genunix: [ID 936769 kern.info] fp0 is
/pci@8,600000/SUNW,qlc@4/fp@0,0
Jun 14 12:06:31 ed genunix: [ID 405830 kern.warning] WARNING: Device
ssd0 failed to power up.
Jun 14 12:06:32 ed genunix: [ID 749148 kern.warning] WARNING: Please see
your system administrator or reboot.
Jun 14 12:06:32 ed scsi: [ID 799468 kern.info] ssd0 at fp0: name
w21000004cf8e7591,0, bus address e8
Jun 14 12:06:32 ed genunix: [ID 936769 kern.info] ssd0 is
/pci@8,600000/SUNW,qlc@4/fp@0,0/ssd@w21000004cf8e7591,0
Jun 14 12:06:32 ed scsi: [ID 365881 kern.info] Vendor 'SEAGATE',
product 'ST373405FSUN72G', (unknown capacity)
Jun 14 12:06:32 ed genunix: [ID 408114 kern.info]
/pci@8,600000/SUNW,qlc@4/fp@0,0/ssd@w21000004cf8e7591,0 (ssd0) online
Jun 14 12:06:32 ed scsi: [ID 799468 kern.info] ssd1 at fp0: name
w21000004cf8e7555,0, bus address ef
Jun 14 12:06:32 ed genunix: [ID 936769 kern.info] ssd1 is
/pci@8,600000/SUNW,qlc@4/fp@0,0/ssd@w21000004cf8e7555,0
Jun 14 12:06:32 ed scsi: [ID 365881 kern.info] <SUN72G cyl 14087 alt 2
hd 24 sec 424>
Jun 14 12:06:32 ed genunix: [ID 408114 kern.info]
/pci@8,600000/SUNW,qlc@4/fp@0,0/ssd@w21000004cf8e7555,0 (ssd1) online
Jun 14 12:06:32 ed swapgeneric: [ID 308332 kern.info] root on
/pci@8,600000/SUNW,qlc@4/fp@0,0/disk@w21000004cf8e7555,0:a fstype ufs
...
And of course all the filesystems on this disk failed to fsck or mount.

Using format I can see the bad disk as c1t1d0 (although searching for
disks... seems to take longer than normal)
0. c1t0d0 <SUN72G cyl 14087 alt 2 hd 24 sec 424>
/pci@8,600000/SUNW,qlc@4/fp@0,0/ssd@w21000004cf8e7555,0
1. c1t1d0 <drive type unknown>
/pci@8,600000/SUNW,qlc@4/fp@0,0/ssd@w21000004cf8e7591,0

If I go ahead and give the correct geometry I can then see the partition
table as I had it before the crash.
Part Tag Flag Cylinders Size Blocks
0 root wm 0 - 412 2.00GB (413/0/0) 4202688
1 swap wu 413 - 1237 4.00GB (825/0/0) 8395200
2 backup wm 0 - 14086 68.35GB (14087/0/0) 143349312
3 unassigned wm 1238 - 1240 14.91MB (3/0/0) 30528
4 var wm 1241 - 2065 4.00GB (825/0/0) 8395200
5 unassigned wm 2066 - 6187 20.00GB (4122/0/0) 41945472
6 usr wm 6188 - 8248 10.00GB (2061/0/0) 20972736
7 home wm 8249 - 14086 28.33GB (5838/0/0) 59407488

I cannot however mount any partition:
mount /dev/dsk/c1t1d0s5 /mnt
mount: I/O error
mount: cannot mount /dev/dsk/c1t1d0s5
_______________________________________________
sunmanagers mailing list
sunmanagers@xxxxxxxxxxxxxxx
http://www.sunmanagers.org/mailman/listinfo/sunmanagers



Relevant Pages

  • Re: Help Failing Disk Problem
    ... and the FreeBSD OS on a hard disk that is failing. ... Either use sysinstall to slice and partition ... fdisk, bsdlabel and newfs. ... 512 MB I mount as /tmp ...
    (freebsd-questions)
  • Re: Relabel partition didnt work
    ... > You're confusing disk lables with mount points. ... > When installing etch, at some point partman asked for a mount point for ... > that partition and I entered /xyz. ... There's nothing to do to the second hard disk. ...
    (Debian-User)
  • Re: Relabel partition didnt work
    ... > mount: mount point /store does not exist ... You're confusing disk lables with mount points. ... /xyz is the 4th primary partition on my 2nd hd. ... Debian will, by default, make labels for disks based on the mount point given during installation. ...
    (Debian-User)
  • Re: Put /usr on a different drive
    ... moving the entire /usr partition to a second hard disk, ... I would just mount that space previously used for /usr as ... once upon a time /usr was thought to be a place for user accounts - ...
    (freebsd-questions)
  • Re: Copying A Disk
    ... A good way to see what device the system is loading your USB drive as ... When you insert your USB disk, you will see some lines of text appear ... sees my USB disk as device sdb and the partition on it as sdb1. ... and it will mount the partition as read-write by default if possible. ...
    (Ubuntu)