please help./...fsck with fulllog failed

[hpuxhelp]

fsck -F vxfs -o full,nolog /dev/vg03/lvol1
vxfs fsck: file system had I/O error(s) on meta-data.
pass0 - checking structural files
pass1 - checking inode sanity and blocks
pass2 - checking directory linkage
warning - no lost+found directory for fileset 999
pass3 - checking reference counts
pass4 - checking resource maps
vxfs fsck: fsck read failure bno = 290, off = 0, len = 8192
file system check failure, aborting ...


what can I do next

[G. Vrijhoeven]

Hi,

Can you do a :

dd if=/dev/dsk/c?t?d? of=/dev/null bs=4048
of the disk that contains the lvolinfo. To see the disk do lvdisplay -v /dev/vg03/lvol1.

Just to make sur ethe disk is not dead.

Gideon

[Sridhar Bhaskarla]

Hi,

Most likely there is a physical problem with the disk holding this logical volume.

Find out the PV by doing an 'lvdisplay -v /dev/vg03/lvol1' and do a dd on it to see if there are any problems.

dd if=/dev/rdsk/cxtydz of=/dev/null bs=64k

See if you get any IO errors. If so, then it is a bad disk and you will have to restore data from your backups after replacing the disk.

Also look at /var/adm/syslog/syslog.log for more errors.

-Sri

[Jeff Schussele]

You can run another pass of fsck - it's not uncommon for several passes to be required to fix an FS, BUT in this case it appears you may have a bad spot on the disk that cannot be realloctaed - block # 290 in this case.

Try another pass or that dd command to verify whether the disk needs to be replaced.

Rgds,
Jeff

[hpuxhelp]

1493+0 records in
1493+0 records out

that is the result...so i don't think the disk the dead

[hpuxhelp]

if that is the case , if theere is a bad block on 290 how can I bypass this or fix this ??

[G. Vrijhoeven]

Hi,

What you can do is run the fsck again on the raw lvol. If it works you might want to run mklost+found after you mounted it.
If this does not work i see no other option but to create a new filesystem and restore a backup.

DO you hava any idea what coused this ( a crask?)

Gideon

[Sridhar Bhaskarla]

Hi,

Is this logical volume only one one disk?. Can you post the output of

'lvdisplay -v /dev/vg03/lvol1|awk '/dsk/ && /current/ {print $2}'|sort|uniq

If it is only one disk and if the dd is successful, then there is a good chance that the filesystem got corrupted and you may have to end up doing a 'newfs' on it and restore the data.

-Sri

[hpuxhelp]

Can you show me how to run fsck on the raw logical volume and what to do next?
I need to try all the options..

the disk also belongs to another logical volume ...this volume has 2 physical volumes and 2 out of 3 mount ok...that is lovl2 and 3 except lovl1 ...

thanks again for your help

[G. Vrijhoeven]

Sorrie it took a while to answer but i needed a good nights sleep.

To fsck the rlvol just change the lvol into rlvol so:

fsck -F vxfs -o full,nolog /dev/vg03/rlvol1

The dd indicated that there were no problems on the disk so the other lvols are not effevted if you need to issue an newfs and a restore on lvol1.

HTH,

Gideon