vx_nospace

[j773303]

What's the mean of the below message, how to trouble it?

vmunix: vxfs: mesg 001: vx_nospace - /dev/vg00/lvol8 fil
e system full (1 block extent)
Jun 12 23:29:02 ERP-ES1 vmunix: NFS server n4000b.dlink.com.tw not responding st
ill trying

[Bill Hassell · Accepted solution]

The filesystem mounted from: /dev/vg00/lvol8 is full. You'll have to look at the files and directories on that filesystem to see what you can remove, move to another mountpoint or perhaps increase the size of lvol8.

The second message indicates that n4000b is dead, or off the network, or NFS services were turned off on that server. Unfortunately, this causes all clients to have big problems since NFS is a filesystem.

[V.Tamilvanan]

Hi,
The first error message means /dev/vg00/lvol8 normally /var FS is full. So you need clean up some of the logfiles to get more space on /var/adm, /var/mail. Check any big unwanted file and remove it.

The second error means taht your system is unable to use NFS resource of n4000b.dlink.com.tw due to some reasons. It may be a network congestion or n4000b.dlink.com.tw may be down.

HTH

[John Dvorchak]

Your Logical volume /dev/vg00/lvol8 is out of disk space. I would suggest that the NFS server n4000b.dlink.com.tw is also experiencing problems, or your network connection to it.

[John Dvorchak]

Since lvol8 is usually the /var file system you will have to remove or move some files. The system will not function very long with the /var filesystem full. That is where most logging is done. A tip I got from Bill Hassel is to do the following to determine where the big directorys are and then you can look for old log files etc

cd /var
du -k . |sort -rn | more

Then you can look for the biggest disk users. One caution, don't just remove log files, they may be in use and that will cause you problems. If you have old files like in /var/adm/sa (old sar files) go ahead and delete them. If you absolutely MUST reduce the size of /var/adm/syslog/syslog.log then DO NOT delete it. Instead use:

> /var/adm/syslog/syslog.log to zero it out. Better yet see if you can copy it somewhere else, not in /var before you zero it out.

Good luck

[j773303]

I fount that the PHKL_28632 shoule be installed. But I'm not sure. Does anyone has the experience about this message.

note:PHKL_28632 VxFS returns vx_nospace on an NFS write or when writing
to a sparse file on a heavily fragmented filesystem, even
though the filesystem has sufficient disk space. Please install
the patch for fix this problem.

[Steven E. Protter]

What the last message is saying is that you are trying to write to a nfs mounted filesystem and getting a failure.

The patch you mention may fix that if installed on the remote server.

The trick with nfs is that if you mount serverb and serverb gets booted the nfs connection gets stale and is useless until you remount the volume.

This is supposedly fixed in later patches/NFS releases. Some people believe its a feature. I'm on the fence.

You should read the patch notes on that patch carefully, because it requires a system boot.

SEP