fsck fail

[Yarok]

Hello
When trying to run the fsck it finally bring the message: cannot red blk 212080.
It seems that Vol 19 is crushed.

The OS is up but without the normal GUI (like init 3).

Please see the attached rc.log file.

Please advice.

[Shibin_2]

You need to unmount the file system on /dev/vg00/lvol9 then do fsck

I think, it is /usr file system.

If any application running, please stop and unmount /usr . Otherwise, boot into single user mode ( this is safe option ) and do fscsk on /usr

fsck -F vxfs -o full

[Yarok]

Hello,

Thanks for the reply,

I rebooted the system into Singel mode.
When I tried to umount /dev/vg00/lvol19, it says that /dev/vg00/lvol19 cannot find in /etc/mnttab.

Should I run the fsck -F vxfs -o anyway ?

BR,

Yaron

[Shibin_2]

Single user mode ( in maintenance ), it won't mount all file system. You can start straight away to do fsck.

You may have to give full path.

/sbin/fsck -F vxfs -o full

[Yarok]

sorry, but what is the 'vxfs' and what vol should I check? (the problematic one - 19) ?

[Ismail Azad]

Hi,

Vxfs there is the type of filesystem that is being mentioned at the time of execution of fsck and an lvol19 will not be mounted in runlevel s. Yes the logical volume that is corrupted should be mentioned as the argument on the execution of fsck.

Regards
Ismail Azad

[Yarok]

Sorry, but I still cannot run it correctly.
I tried:

/sbin/fsck -F vxfs -o full /dev/vg00/lvol19

received:

fsck vxfs cannot stat /dev/vg00/lvol19

What does it mean?

Thanks,

Yaron

[Dennis Handly]

>fsck vxfs cannot stat /dev/vg00/lvol19

Are you using the raw/character device?

>Shibin: I think, it is /usr file system.

It seems to be /var/?

[Matti_Kurkela · Accepted solution]

Not "Vol 19" but LVOL 9:

---from the attachment:---
mount: /dev/vg00/lvol9 needs to be fsck'ed before mounting
----end----

Depending on the font used on your display, the second lowercase L in default-style LV names (pattern "lvolN", N = number) can be easy to confuse with number 1.

You aren't the first to make this mistake, and probably won't be the last one either.

Based on the error messages in your attached rc.log, looks like your /dev/vg00/lvol9 probably contains your /var. As a result of the failed mount, several system daemons have failed to start.

You also have an exportfs error message about /scc: is this symlinked to some sub-directory within your /var? If not, this might be a separate issue.

If /scc does not exist at all, it might be a remnant of some old configuration: while the /scc directory may have been removed, /etc/exports seems to still refer to it.

MK

[Shibin_2]

// Have you checked that this is mounted ?

bdf /dev/vg00/lvol19 ?

If mounted, then need to unmount it first, then do fsck.

[Yarok]

Hello,

Thank you for the correction. Indeed, it is

LVOL9 and not LVO19.

I could not find /scc and /dev/vg00/lvol9 as well.
I checked the mount to /dev/vg00/lvol9 and it says that its not in /etc/mnttab

bdf check olso indicates that /dev/vg00/lvol9 does not exist.

/etc/exports shows:

/scc -annon = 0
/disk2/scc = 65534

I compared with Other HP machine and it shows that /dev/vg00/lvol9 is mounted to /scc and
/etc/exports shows

/scc -annon = 65534
/disk2/scc = 65534

Please note that /scc is necessary for work.

Please advice.

[Matti_Kurkela]

Your previous fsck command was directed to LVOL 19, which apparently does not exist. No wonder it failed.

This command will try to fix LVOL 9, accessing it in raw mode. For this command, the VG must be activated, but the LVOL must be unmounted. If I understood correctly, that's exactly the situation you have now.

/sbin/fsck -F vxfs -o full /dev/vg00/rlvol9

If the filesystem check is successful, you should be able to mount the LVOL 9 filesystem after that:

mount /dev/vg00/lvol9

If this is successful too, it should be possible to reboot (with the command "shutdown -r now") and have your system start up normally again.

The original error message (cannot read block ...) suggests your disk may have already failed partially and may soon fail completely. If you don't have a backup of the system, make one as soon as possible.
(I hope you *do* have a backup: some files may already have been lost or damaged)

If you have a support agreement with HP, open a support call for disk replacement, let the HP engineer fix it and ask him/her how to proceed.

If you don't have a support agreement, please read this HP document called "When Good Disks go Bad: dealing with disk failures under LVM":
http://bizsupport2.austin.hp.com/bc/docs/support/SupportManual/c01911837/c01911837.pdf

MK

[Yarok]

Sorry, but the fsck cannot be start. I think, again ,it concerns to the reason that there is no LVOL9.

Any idea how to search for the vol?

[Dennis Handly]

>the fsck cannot be start.

What error did you get?

>Any idea how to search for the vol?

What do these show:
ls -l /dev/vg00/lvol9
/sbin/lssf /dev/vg00/lvol9

[Yarok]

The error I get is:

vxfs fsck: cannot stat /dev/vg00/rlvol9

ls -l /dev/vg00/lvol9 shows:

/dev/vg00/lvol9 not found


/sbin/lssf /dev/vg00/lvol9 shows:
lssf: no shuch file or directory /dev/vg00/lvol9

[Shibin_2]

Please provide the output of

vgdisplay -v vg00

If it is too long, please DO NOT copy paste here; but attach.

[Yarok]

Hello again,

Sorry for the delay but I was out of the office.

I attached 2 files:

LVOL_check - is the vgdisplay -v vg00 resault of the "demaged" computer

and

LVOL_check_OK - is the vgdisplay -v vg00 command resault of other "good" computer(an HP similar machine that does not suffer from disk problem).

[Yarok]

Ho,

I can see that only one file canbe attached at a time. so here is the second one (LVOL_check_OK)

[Dennis Handly]

>only one file can be attached at a time.

Yes, unless you tar and/or zip the files.
(If you zip, make sure the suffix is .Zip.)

[Yarok]

Can someone give me a solution to this problem?
In my poor opinion, I should recreate the lost Volume - (Vol19).

1. How can I do that?
2. How can I restore the lost information? (I think the information is still there, somewhere, I need to check for it).

[Yarok]

Hello,

I am still waitting for help regarding my problem. I will be glad to have a solution.

[Shibin_2]

I can see your lvol9 is in available status.

LV Name /dev/vg00/lvol9
LV Status available/syncd
LV Size (Mbytes) 4500
Current LE 1125
Allocated PE 1125
Used PV 1


Please post the output of bdf /dev/vg00/lvol9
and lvdisplay /dev/vg00/lvol9

Have you tried to do fsck on this LV in single user mode ?

First try to do fsck and fix the issue. Otherwise, restore from ignite. If you have good ignite backup, you may restore from that.

[Yarok]

Thank you for your reply

# lvdispay /dev/vg00/lvol9

Filesystem kbytes used avail %used Mounted on
/dev/vg00/lvol9 960725 328509 536143 38%

# lvdisplay /dev/vg00/lvol9
--- Logical volumes ---
LV Name /dev/vg00/lvol9
VG Name /dev/vg00
LV Permission read/write
LV Status available/syncd
Mirror copies 0
Consistency Recovery MWC
Schedule parallel
LV Size (Mbytes) 1000
Current LE 250
Allocated PE 250
Stripes 0
Stripe Size (Kbytes) 0
Bad block on
Allocation strict
IO Timeout (Seconds) default
How do I make this check - fsck on this LV in single user mode ? Please specify exactly what command should I ran.

[Shibin_2]

You need to reboot the server, interrupt the booting and use following command at ISL prompt

ISL> hpux -is

Once booted into single user mode, mount /usr then do fsck to your lvol9. You may have to give full path of fsck.

/sbin/fsck -F vxfs -o full -y /dev/vg00/lvol9

[sarfaraj ahmad]

hi,

please put the bdf output.

have tried with full path of fsck (/usr/sbin/fsck) command?