error in syslog mesg 001: vx_nospace

[m.huisman]

Hi I hare ar entry in the syslog:
vmunix: msgcnt 1 vxfs: mesg 001: vx_nospace - /dev/root file system full (1 block extent)
But i can find anything wrong.
Does somebody know this error and solution?

[Pete Randall]

Hi,

You probably had a crash that was writing a core dump into your root file system. Check your bdf output now to see if / is OK or if it's nearly full (like greater than 95%). Also make sure it's actually mount as / rather than /dev/root.


Pete

[T G Manikandan]

What does your bdf report.

Check for the timestamp.

Sometime or now your / file system has crossed 100%.

Also,

when you go into lvm maintenance mode the / file system gets mounted as /dev/root.


Just do a
#mv /etc/mnttab /etc/mnttab.old
#mount -a

Thanks

[m.huisman]

My bdf shows 69% full on the root file system.
So that not the problem.

[Patrick Wallek]

What is the time stamp on those entries in your syslog.log file?

There may not be anything wrong right now, but at some point there was.

[Michael Steele_2]

Did you reboot or kill a backup process or any process?

A runaway process could be the cause for / filling up.

A malformed backup, /devrmt/0mn, for example, often happens.

The time stamp will allow cross referencing into other logs. You should check the cron log, /var/mail/root and /etc/opt/resmon/log/*.

find / -xdev -ctime 0 (* files accessed today in root *)

find /dev -type f (* find files that don't belong in /dev *)

quot / (* find owners with files in / *)

du -x / | sort -rn | more (* big files in / *)

[T. M. Louah]

.: You should start by TG's suggestion, I have seen this after igniting a system. Even if vg00 is active run:
# vgchange -a y vg00
# vgdisplay vg00 (verify that Cur PV = Act PV, if not you have a ghost disk)
# lvdisplay -v /dev/vg00/lvol3 |more (verify the PEs)
# lvlnboot -v (to make sure root, boot, swap & dump are seated were they should be)
.: verfiy in fstab how lvol3 is supposed to be mounted
# mv mnttab to mnttab.old
# mount -a
# mount -p

Ur done! if the above doesnot help, call HP!
Cheers,
T?

[Helen French]

This messages in syslog normally comes when there is an issue with the file system. A possible solution given by HP is to reformat the file system. Since you have problem with the root file system, this is not very simple. I would try the following:

1) Check why root (/) is mounted on /dev/root instead of /dev/vg00/lvol3? This is normal after a system LVM maintenance mode boot or a system crash. The issue is root file system was unable to access /dev/vg00/lvol3 device file during it's last boot. One possible solution is to remove and recreate /etc/mnttab:
# mv /etc/mnttab /etc/mnttab.old
# mount -a
This will put root file system back to vg00/lvol3 if there is no other issue.
2) If that doesn't solve the problem, I would boot in single user mode OR lvm maintenance mode. Then check the root file system with fsck and see if that fix or solve the issue. Again, if it mounts on /dev/root, recreate mnttab file.
3) If that also didn't help, I woukd consider reloading the system from a make_tape_recovery tape. That will create new file system and fix any software errors on it.
4) If still problem exists, I would consider changing the root disk.

[Caesar_3]

Hello!

Looks like in some moment your /dev/root
was full it depend on when this entry shown.

Could be that some process write to and stoped/removed the data.

Maybe you run in cron remove of core files?

Caesar

[mango_1]

Hi all! I have a problem when I log in as root to my hp machine. I get the following log message

vxfs: mesg 001: vx_nospace - /dev/root file system full (1 block extent)

The reason I found out was that my / directory is 100%. And I was wondering if any of you guys know how would I know what filled up my / directory and how to clear up this problem?

thanks!

[Ken Hubnik_2]

Look for core files or in your dev directory perhaps there is a device file spelled wrong. A common problem it /dev/rmt/om and alpha o instead of a zero that caused a backup to create a tape archive in your /dev filesystem.

[Pete Randall]

First thing, "mango", is to open your own question so that you can assign points to those who help you, get notifications when someone replies, and not mess up anyone else's point assignments ratio (not that m.huisman seemed to care about his).


Pete

[Vasikaran Venkatesan]

At some point the /dev/root was full - assuming you were in "hpux -lm" mode.
otherwise have u mounted any LVOL on /dev/root ?.

Possibility ?.
Could it be that the core file was getting created by some application and could not complete writing as the FS was full - and the application removed the partially written core.

[vofsky]

How can I check for the timestamp?Who can give me a favor?