Root File System Full

[Ali Imran Abbas]

My syslog is giving root file system error message:

Jan 9 04:44:34 usunx12 vmunix: msgcnt 19 vxfs: mesg 001: vx_nospace - /dev/root file system full (1 block extent)
Jan 9 05:44:41 usunx12 vmunix: msgcnt 21 vxfs: mesg 001: vx_nospace - /dev/root file system full (1 block extent)
Jan 9 06:44:47 usunx12 vmunix: msgcnt 23 vxfs: mesg 001: vx_nospace - /dev/root file system full (1 block extent)
Jan 9 07:44:54 usunx12 vmunix: msgcnt 25 vxfs: mesg 001: vx_nospace - /dev/root file system full (1 block extent)
Jan 9 08:45:00 usunx12 vmunix: msgcnt 27 vxfs: mesg 001: vx_nospace - /dev/root file system full (1 block extent)
Jan 9 09:45:06 usunx12 vmunix: msgcnt 29 vxfs: mesg 001: vx_nospace - /dev/root file system full (1 block extent)
Jan 9 10:45:13 usunx12 vmunix: msgcnt 31 vxfs: mesg 001: vx_nospace - /dev/root file system full (1 block extent)
Jan 9 11:21:33 usunx12 vmunix: msgcnt 33 vxfs: mesg 001: vx_nospace - /dev/root file system full (1 block extent)
Jan 9 11:22:08 usunx12 vmunix: msgcnt 35 vxfs: mesg 001: vx_nospace - /dev/root file system full (1 block extent)
Jan 9 11:45:21 usunx12 vmunix: msgcnt 37 vxfs: mesg 001: vx_nospace - /dev/root file system full (1 block extent)

Can some body tell me how to resolve this situation?

[MarkSyder]

The most common problem with root filesystem is someone mistyping a tape device name and creating a large text file in /dev. Assuming /dev/root is a filesystem in its own right, that won't apply in your case.

Look for the most recent files:

cd /dev/root
ll -tr

See what has changed recently. Do you have any processes writing to /dev/root that should be writing to a dedicated filesystem instead?

Mark Syder (like the drink but spelt different)

[Ali Imran Abbas]

Do I need to go to single user mode for that?

[Steven E. Protter]

Shalom,

cd /
du -k | sort -rn | more

Identify the folders that are full and should not have files.

Besides a few configuration files in /etc the root fs should have nothing in it but mount points for other filestems.

Good Luck Ali

SEP

[Ali Imran Abbas]

What are the steps u need to perform if your root file system is full?
I have the following questions:
Can root login to the machine whose root file system is full.
Whats the name of core file and where its location
where does the file created by wrong tape device reside?
Can some one detail one by one step in this regard pls?

[Roberto Arias]

hi all:

First check that your FS is mounted and the mount point hasn't permissions for all users (755 only)

Second check where is files higest with du -kx |sort -n and remove or move the files.

Third check that you don't have any proccess hang that he could use /

For end, i see you FS is /dev/root when this must be /dev/vg00/lvol3, please reboot system for resolve this

thanks

[Rita C Workman]

What are the steps u need to perform if your root file system is full?
I have the following questions:
Can root login to the machine whose root file system is full.
Whats the name of core file and where its location
where does the file created by wrong tape device reside?
Can some one detail one by one step in this regard pls?
=========================================
#1 - Folks have given you that information above.
#2 - Yes
#3 - /var/adm/crash
- But you can have core files anywhere on your box. You need to look for them and clean them up. This statement isn't selective - so beware:
find / -name core -print | xargs rm {} \;

#4 - Nobody knows your boxes but you, so the best folks can do is give you enough to start your process. The rest is up to you.

A couple other hints:
Look under /etc for unnecessary copies of host,group,lvmtab
Look under /dev for bogues files, especially where someone created file while trying to write to tape (my personal favorite) /dev/rmt/
If you run MC/SG, then do some logfile trimming on your packages. /etc/cmcluster/packages//


Rgrds,
Rita

...ps....0 out of 68. Assign points earned to the folks who have helped you.

[Rajesh SB]

Hi,

You look for following and try to delete to get the free spaces.

1. Mail log file, there may be chances of flooding emails from different sources.
/var/adm/syslog/mail.log
Also lookd for any application logs are growing rappidly (if trace level is on).

2. Look for crash dumps if any in the directory /var/adm/crash and delete.

3. Find for any core dump in any directory under root filesystem and delete it.

Thanks & Regards,
Rajesh SB

[IT Response]

Find /filesystems -size +1000c -exec ll {} \; --- to find 1000 bytes files and those sizes
find / -type f -mtime 0 -exec ls -l {} \; --- lists all files modified in last 24 hrs
find / -mtime + -exec ls -l {} \; --- to list files which are older than particular number of days
find / -mtime + -exec rm {} \; ---- to remove the files which are older than particular number of days
du -k -x / --- to find the size of the files only in '/'
du -kx / | sort -rn -k1 | head -n 10 ---- 10 largest directories in the root filesystem:
find / -type f -xdev -print | xargs -e ll | sort -rn -k5 | head -n 10 ---- 10 largest files in the root filesystem
find / -type f -xdev -mtime -1 -print | xargs ll | sort -rn -k5 ---- Recently modified files in the root filesystem

[Prashanth Waugh]

Hi ,

#find / -name core

Check for any core file

Regards
Prashant

[Prabhu Hosmani]

Hi,

Try this one, it helped me to resolve the same issue

#cd /dev/rmt
#>0m

regards,
Prabhu H

[Deepak Kr]

Imaran,

Can you tell which unix are you using ?

is it linux box?

[Torsten.]

Nice to see more solutions - but does somebody think the problem still persists since january? Unlikely ...

;-)