fsck with "invalid device config entry found" response

[CARLOS ROY GOLET]

Dear friends,
when I try to run fsck, I get an error:
.............
# fsck -F vxfs -y -o full /dev/vg00/rlvol8
vxfs fsck: invalid device config entry found
log replay in progress
............
so, according to some help in the forum, I have tried a different option, but the results are not good:
............
fsck -F vxfs -o full,nolog -y /dev/vg00/rlvol8
vxfs fsck: invalid device config entry found
pass0 - checking structural files
fileset 1 primary-ilist inode 33 (SuperBlock)
failed validation clear? (ynq)y
fileset 1 primary-ilist inode 40 (DEV 1)
failed validation clear? (ynq)y
fileset 1 primary-ilist inode 32 (State)
failed validation clear? (ynq)y
fileset 1 primary-ilist inode 34 (Sumino)
failed validation clear? (ynq)y
fileset 1 primary-ilist inode 42 (Mapino 1)
failed validation clear? (ynq)y
pass1 - checking inode sanity and blocks
vxfs fsck: fsck read failure bno = 2912, off = 0, len = 8192
file system check failure, aborting ...
.........

Any suggestion ?

Thanks

[Deoncia Grayson_1]

have you tried running a dd to make sure it comes back clean?

tried to check with dd if=/dev/vg00/rlvol8 of=/dev/null bs=1024k (blocks read test

[Mridul Shrivastava]

Try this

fsck -F vxfs -o full /dev/vg00/lvol8

It will resolve the issue.
Mount lvol8 to /var, and check it with bdf.
You also want to ensure that you have the latest vxfs/fsck patches installed.

[CARLOS ROY GOLET]

Running dd:

# dd if=/dev/vg00/rlvol8 of=/dev/null bs=1024k
dd read error: I/O error
22+0 records in
22+0 records out


---------

Running the other fsck and mount commands:

# fsck -F vxfs -o full /dev/vg00/lvol8
vxfs fsck: invalid device config entry found
log replay in progress
vxfs fsck: fsck read failure bno = 7553, off = 0, len = 8192
full file system check required, exiting ...
# mount /dev/vg00/lvol8 /var
vxfs mount: /dev/vg00/lvol8 is corrupted. needs checking

[Mridul Shrivastava]

could you please check the lvol8, it looks like that lvol8 got corrupted or some issue with media as dd also returned read error.

[CARLOS ROY GOLET]

I am sorry, but I do not understand: what do you mean with "check lvol8" apart from running fsck ? how can we verify there is a media problem ?

[A. Clay Stephenson]

# dd if=/dev/vg00/rlvol8 of=/dev/null bs=1024k
dd read error: I/O error
22+0 records in
22+0 records out

This really says it all (well, almost). Almost certainly this is a disk hardware problem. This is a small probability that this could actually be a problem with the LVM metadata pointing to a bogus device.

Do an lvdisplay -v /dev/vg00/lvol8 and note the PV(s) associated with this LVOL (e.g. /dev/dsk/c1t5d0). I suspect there is only one. Now run your dd against this raw disk device (e.g. /dev/rdsk/c1t5d0).

e.g. dd if=/dev/rdsk/c1t5d0 bs=1024k of=/dev/null

If this reports errors then you have a bad disk.

[Mridul Shrivastava]

I mean to say execute lvdisplay command and then check ioscan for the device belongs to this lv.
check syslog.log and event.log for any errors as dd cleary says there is some hardware issue.

[Albert_31]

just a thought..

# ll /dev/vg00/

check the minor /major number for lvol8..

[Srikanth Arunachalam]

Hi ,

Did you try fsck without any specific option. like. Please try this and let know the result.

umount ..
fsck -F vxfs /dev/vg00/rlvol8
mount ...

Thanks,
Srikanth

[CARLOS ROY GOLET]

Hi,

the disk is c2t0d0 and

dd if=/dev/rdsk/c2t0d0 of=/dev/null bs=1024k

does not show any result after 2 hours.
A similar disk ends in 10 minutes without error.

So, i suppose you are right and the disk has hw problem. I am going to open a hw issue with HP.

My biggest problem is that the original /var directory has been lost and I have had to create an alternative /var in a new volume but some features do not run properly (cron, glance, swlist, etc.)