/var inode problem

[Eileen Millen]

Can anything be done to fix the inode problem listed below? We have Online JFS.
Aug 24 10:54:56 hplogin1 vmunix: vxfs: mesg 016: vx_ilisterr - /var file system error reading inode 19670
Aug 24 10:54:56 hplogin1 vmunix: vxfs: mesg 016: vx_ilisterr - /var file system error reading inode 19670

[Mark Vollmers · Accepted solution]

Eileen-

I'd start with running fsck on the filesystem. normally, that will identify the inode problem and fix it. You might need to drop down to single user mode to run it (I'm not sure with Online JFS if you have to or not). If you reboot after running fsck (and it finds problems), remember to use the reboot -n command.

Mark

[Sridhar Bhaskarla]

This could be a hard disk problem or corrupted file system. Boot into Single user mode and run an fsck on the file system.

Before that try doing an ncheck /dev/vg00/var_lvol > /tmp/varnlist and see what file this inode correponds to.

-Sri

[Eileen Millen]

Sridhar:
Thanks, I didn't know about the ncheck command.
The ncheck gets
read failed for block 4288 / IO error

Does that mean the disk is no good?

[linuxfan]

Hi Eileen,

Boot into single user mode and run
fsck -F vxfs -o full,nolog -y /dev/vg00/lvol_var

But if you keep getting these errors repeatedly inspite of doing fsck, it could indicate disk going bad.

-HTH
Ramesh

[Sridhar Bhaskarla]

I would say it as the physical problem with the disk if my fsck fails. So, you need to fsck it while bringing down the server in single user mode. If it keeps saying cannot read blk#, then you may have to replace the disk.

And, before doing so, since you have the system up and running now, it will be a good idea to prepare a make_recovery tape, verify it and also take a full backup of the system.
So,do not bring down the system unless you have a good make_recovery tape with you.

-Sri