Topic: Strange! Out of space error!

[tymiles]

I can not figure out what is wrong with my server. For some reason I can not get into the server manager! When I try to start it or I reboot my server I get an out of space error but when I check the space this is what I get:

Filesystem       1k-blocks        Used         Available   Use%   Mounted on
/dev/hda6      38223820     10359092   25952596 29%       /
/dev/hda1           46636              8774         35454 20%      /boot
none                 256584                    0        256584  0%      /dev/shm

Can someone tell me what is wrong with this picture. I am thinking I got hacked!  (I have plenty of drive space)

(I also can't get my mail to work etc. )

Also if I try to run midnight commander I get this error:
bash: /root/.mc/tmp/mc-2663: No space left on device  

(Could my tmp folder be full)

Help......

[tymiles]

I did a df -i and see that the Inodes on / are at 100%

Is there an easy way to fix this?

[Brave Dave]

your problem will be there are too many files

for mine i'd be checking the log folder, mgetty files - but only because i've been caught there before, the logrotate scripts used to go ballistic on renaming files

other candidates could be the squid cache if you fiddled with the default settings

you will likely only have to delete a few files (1 or 2) to get the likes of mc running

maybe;
rm /var/log/messages.4
rm /var/log/maillog.4
rm /var/log/secure.4

then fire mc and have alook around (if there are 50,000 files in your /var/log or somewhere you might need a cup of tea while mc lists them.)

[tymiles]

Thank you David for your help! I will check this out!

[smeghead]

I'm using 2% of the inodes with 27% of the disk full (80GB).

Why have you used up all the inodes?  Check your / drive carefully to see what has consumed all the space.

If you have a good backup (just in case) do a single user boot and run fsck / and see if you have accumulated a lot of orphaned stuff; make sure your system is well documented as well!!

In normal operation you would need to accumulate a lot of files to chew up the inodes (poss 100,000 plus).

HTH

[mbachmann]

Flushing the Squid cache might help also as he likes to accumulate lotsa files.

[buknoy]

Hi, I had the same problem and since I could not figure this out why, I already re-installed my server thrice this week. This is my network:

-Intel Pentium III-500 with 512MB RAM and 10GB harddisk
-SME 6.01 updated after fresh install
-15 workstations
-no mail service
-no web service
-been using this setup since December 2004

What I could not understand is how could my /dev/hda3 be filled up after one day of running? I tried to erase old cache and even changed logrotation to 2 and fixed the size of access.log to 1M, store.log and cache.log to 500k. I also observed that leaving the fresh install and configuration "as-is" results to slower browsing and downloading.

What should I check or configure? Is my server attacked by corrupting my /dev/hda3? I don't think increasing my harddisk to 40GB will do any good if my /dev/hda3 kept filling up.

Don

[brad]

Does this help?

Title: "Count Files under multiple directories"
Link: http://www.linuxforums.org/forum/post-174743.html

Title: Counting Files in the Current Directory
Link: http://www.linux.org/docs/ldp/howto/Bash-Prompt-HOWTO/x700.html

These are just some links in my tool box for counting files.

Good Luck!