fsck problem on a filesystem

[Shaamil]

Hi
I get the following error when I try to do an fsck on my filesystem
#> fsck -y /dev/vg01/opt
log replay in progress
checking structural files
pass1 - checking inode sanity and blocks
vxfs fsck: getblk read fail bno 1162
cannot read inode 2 in fileset 999

When I try to do the fsck on the disk it give me the following.
fsck -y /dev/rdsk/c1t5d0
fsck: /etc/default/fs is used for determining the file system type
** /dev/rdsk/c1t5d0
BAD SUPER BLOCK: MAGIC NUMBER WRONG
USE -b OPTION TO FSCK TO SPECIFY LOCATION OF AN ALTERNATE
SUPER-BLOCK TO SUPPLY NEEDED INFORMATION; SEE fsck(1M).

This is the disk my vg01 is on.
Also this is a very old machine.
Model = E55 running Hp-ux 10.10.
Really old hardware.
I also checked the ioscan output & all my disks is ClAIMED.

[Balaji N]

hi


dont do an fsck on the disk. since it is using lvm, do it only on the lv.

first do a vgdisplay on the vg and see if the hard disks under it.

then do a dd to see if all the blocks on the hard disk. me feels that there is some read problems on one disk.

dd if=/dev/dsk/xxxxx of=/dev/null bs=2048k

and make sure you have the latest backups.

hth
-balaji

[massimo poli_1]

Hi,
check if /etc/default/fs match with the fs type of your filesystem.
If your filesytem is an hfs one a list of alternate superblocks is recorded in /etc/sbtab

If the filesystem is a vxfs type you can try with
fsck -o full -y
(see man fsck_vxfs)

Good luck

[Cheryl Griffin]

Shaamil,
You might try a pvdisplay on the disk to see if it's readable. This type of corruption is usually associated with a disk failure and if so the filesystem may not be recoverable.

You might try a full fsck:
# fsck -F vxfs -o full /dev/vg01/lvol#

[Michael Steele_2]

This problem occurrs when the /etc/default file indicates a file system type other then /dev/vg01/opt'. Use 'fstyp' and the '-F' option to get around this.

fstyp /dev/vg01/opt

fsck -F vxfs -y -o full /dev/vg01/opt (* for vxfs file systme *)

fsck -F hfs -y /dev/vg01/opt (* for hfs file system *)

If the problem continues and 'HFS' is indicated then refer to the /etc/sbtab file.

If 'raw' is indicated then this is neither hfs nor vxfs and the application controls the data in the lv, so refer to the application and ignore using 'fsck'.

[Shaamil]

HI
The one disk seems to be faulty.
vg01 has 3 disks.
I am going to remove all 3 these disks & put in some newer disks.
Thanks for all the help.
Shaamil

[T G Manikandan]

This is a super block error.

Run the fsck using an alternate block number from the /etc/sbtab file.

The next alternate block is 16.
so just use

#fsck -F vxfs -b 16 /dev/vg01/opt

It should also repair the problem.


After this fixes the problem run fsck again.


Revert