bdf command doesn't show filesystems!!

[Carles Viaplana]

Hello,

when I execute bdf command I haven't any output:

# bdf
#

But they're accessible!

Do you have any idea?
This system belongs to a cluster, does this issue related with?

Note that OS is HP-UX B.11.11.

Thanks in advance for your help.
Regards,

Carles

[A. Clay Stephenson]

Does df work? If it too fails, I suspect you have a missing/corrupt mnttab. Being a cluster member shouldn't matter. At the very least, vg00 is not under MC/SG and it should display.

[Geoff Wild]

What about mount - does that work?

Rgds...Geoff

[Steven E. Protter]

Shalom Carles,

If you have an nfs mount done before filesystem mount or high in the /etc/fstab file bdf will hang waiting for the nfs mount to return data if the nfs connection is cut off.

I suspect that the cluster and volume groups may be misconfigured.

SEP

[Carles Viaplana]

Hello Clay,

df fails too and /etc/mnttab is empty. I also checked if /usr/sbin/syncer is running:

# ps -ef | grep /usr/sbin/syncer
root 431 1 0 Nov 9 ? 10:33 /usr/sbin/syncer
root 970 29101 2 19:18:20 pts/ta 0:00 grep /usr/sbin/syncer
#

If I'm right, this process should update mnttab automatically. Can I kill process and restart again executing "/usr/sbin/syncer&"?

Thanks for your help.
Regards,

Carles

[CAS_2]

run

mount -p
mount -a
bdf

[Carles Viaplana]

Hello,

First at all, thanks for all your replies.

In fact, I haven't any nfs on this system, but problem could be due cluster issue.

I had some problems with a package (I couldn't switch it to other node). After halting and running again on same node all packages work fine again.

As a package mount an ummount filesystems it uses, problem could be caused when I tried to switch package to other node. Maybe it hungs mnttab.

Any idea about how can I fix it?
Regards,

Carles

[Carles Viaplana]

Hello again,

Here you're output from mount -p command:

# mount -p
mount: file system table may be corrupt
#

Regards,

Carles

[Jeff Schussele]

Hi Carles,

Well, your problem is your /etc/mnttab file.
It should never be empty when the system is up & running. It should always conatain *at least* the LVs from vg00.
I suspect you'll need to reboot the system to clear this up.

Rgds,
Jeff

[Carles Viaplana]

Hello again,

My apologies. I remembered we have 3 nfs filesystems mounted there and remote system was rebooted few days ago.

Could it cause filesystem issue?
Is there any way to resolve it without rebooting?

Regards,

Carles

[James R. Ferguson]

Hi Carles:

Try issuing :

# mount all

This should recreate your 'etc/mnttab'.

Regards!

...JRF...

[Geoff Wild · Accepted solution]

mv /etc/mnttab /etc/mnttab.bad

mount


Should fix it.


Rgds...Geoff

[Devender Khatana]

Hi,

The file is overwritten by syncer, mount and unmount. As syncer is not allready doing it you can try it through other two methods.

HTH,
Devender

[Yogeeraj_1]

hi Carles,

i will also suggest that you verify, check and backup your /etc/fstab

While troubleshooting, it will be also wise not to mount your remote file systems.

good luck!

kind regards
yogeeraj

[Carles Viaplana]

Hi all,

I tried to mount filesystems directly:

# mount -a
mount: file system table may be corrupt
#

On the other hand, filesystems mounted from packages aren't into /etc/fstab file. If I execute mount -a command they won't be mounted/updated.

Thanks for your help.
Regards,

Carles

[Arunvijai_4]

Hi Carles,

You can do # mount all, before doing that move your corrupted /etc/mnttab

-Arun

P.S Assign points to the people who answered.

[Muthukumar_5]

Do this:

# mv /etc/mnttab /etc/mnttab
# sync
# sleep 120
# bdf

It will work.

--
Muthu

[Carles Viaplana]

I just moved /etc/mnttab file and then I mount all filesystems. After this all filesystems (mounted from /etc/fstab file or from packages) appeares as usual.

Thanks to all for your help.

Regards,

Carles