More than 100% disk usage is possible !!!

shreckbull

Hi !

With an installation made on June 2007 of 1.2 beta 2, i havec more than 100% of disk usage, with no reboot and no problems …


# cd /
# df
Filesystem  1K-blocks    Used   Avail Capacity  Mounted on
/dev/ad0s1a   1484678 1478042 -112138   108%    /
devfs               1       1       0   100%    /dev
/dev/md0         1710      40    1534     3%    /var/run
devfs               1       1       0   100%    /var/dhcpd/dev

I have reboot it, uninstall Squid, delete its directory, reboot again …
I don't use any package, just Squid ...

When i reboot, this installation is fully fonctionnal ...

What do you think about this problem ... ?

Cry Havok

There is no problem. See the FreeBSD newfs man page for details on space reserved from normal users, which is by default 8%. This is the cause of the figures you're seeing having almost filled your root partition (you've got about 6 MB left).

shreckbull

I haven't say an important thing … yesterday, i was connected on this firewall for know IP Address of one printer and the disk usage by GUI (index.php) graphic waq about 55-57% ...

cmb

Your root partition is only 1.5 GB? That's fine with no packages, but I'd imagine squid will fill that up pretty quickly. It seems it goes away when you uninstall squid, so I think the problem is you don't have enough drive space to run squid.

shreckbull

I use 2GB hard drive …
configured squid for use only 100MB of cache ...

in two others sites, i use the same config, same version, same hardware ... without this problem ... and i had never use more than 45/50% space on my hard drive.

cmb

Next time it happens, run the following commands from a shell and paste the output here.

df -h
cd / && du -hd1

shreckbull

df -h

Filesystem Size Used Avail Capacity Mounted on
/dev/ad0s1a 1.4G 1.4G -111M 108% /
devfs 1.0K 1.0K 0B 100% /dev
/dev/md0 1.7M 40K 1.5M 3% /var/run
devfs 1.0K 1.0K 0B 100% /var/dhcpd/dev

cd / && du -hd1

2.0K ./.snap
21M ./boot
1.2M ./bin
1.5K ./dev
16K ./conf.default
4.6M ./etc
158K ./libexec
3.5M ./lib
42K ./root
3.7M ./sbin
966M ./usr
9.2M ./var
431M ./tmp
2.0K ./mnt
982K ./cf
3.2M ./rescue
1.4G .

It's my /usr/local/www the bigger directory with "just" 650MB …

cd /usr/local/www && du -hd1

102K ./m0n0
106M ./usr
60K ./code-syntax-highlighter
14K ./includes
1.2M ./javascript
4.0K ./sajax
4.2M ./themes
38K ./tree-images
110K ./wizards
16K ./wlan_strong_key_generator
102K ./widgets
2.0K ./packages
316K ./pfSense
1.8M ./phpSysInfo
80K ./bin
16M ./boot
4.0K ./dev
410K ./etc
78K ./lib
3.2M ./rescue
4.0K ./root
1.5M ./sbin
2.0K ./java
650M .

Cry Havok

I don't know if it's unusual or not (I only have my own host to compare with), but your /usr/local/www/usr is 106 MB, mine is closer to 200 KB. Your /sbin is also about twice the size of mine (1.2 Beta2).

As a reference, I have Squid, FreeRadius, IMspector, NUT, PFFlowed, Snort and LightSquid all installed with 438 MB of disk space in use. Squid is set to use 128 MB of disk (and memory) for cache.

shreckbull

This morning, i perform a succefull test with MAxtor utilities, this evening … the same test return errors ... !

I think hard drive is death ... but Pfsense can always boot and be used but some service aren't fully fonctionnal such as DHCP server ...

After reinstall with new HDD ... all is OK, i will compare directory size's for see.

cmb

400+ MB in your /tmp is odd. What's in there? Should be more like a few KB.

having a /usr/local/www/usr folder is also strange, I'm guessing the squid package must have put that there as that shouldn't be there.

werneman

Hello all,I have got problem with "more than 100% disk usage is possible" -Jan 22 10:04:39 kernel: pid 824 (dhcpd), uid 1002 inumber 1719482 on /: filesystem full, I can not install any package ntop etc. , RRD Graphs are not running - there is system log
I use 1.2-RC4
built on Wed Jan 16 23:13:03 EST 2008
Jan 22 10:06:40 php: /status_rrd_graph_img.php: Failed to create graph with error code 1, the error is: /libexec/ld-elf.so.1: Shared object "libpng.so.5" not found, required by "rrdtool"
Jan 22 10:06:40 php: /status_rrd_graph_img.php: Failed to create graph with error code 1, the error is: /libexec/ld-elf.so.1: Shared object "libpng.so.5" not found, required by "rrdtool"
Jan 22 10:06:39 php: /status_rrd_graph_img.php: Failed to create graph with error code 1, the error is: /libexec/ld-elf.so.1: Shared object "libpng.so.5" not found, required by "rrdtool"
Jan 22 10:06:39 php: /status_rrd_graph_img.php: Failed to create graph with error code 1, the error is: /libexec/ld-elf.so.1: Shared object "libpng.so.5" not found, required by "rrdtool"
Jan 22 10:06:39 php: /status_rrd_graph_img.php: Failed to create graph with error code 1, the error is: /libexec/ld-elf.so.1: Shared object "libpng.so.5" not found, required by "rrdtool"
Jan 22 10:06:38 php: /status_rrd_graph_img.php: Failed to create graph with error code 1, the error is: /libexec/ld-elf.so.1: Shared object "libpng.so.5" not found, required by "rrdtool"
Jan 22 10:04:39 kernel: pid 824 (dhcpd), uid 1002 inumber 1719482 on /: filesystem full
Jan 22 10:01:38 kernel: pid 824 (dhcpd), uid 1002 inumber 1719482 on /: filesystem full

df

Filesystem Size Used Avail Capacity Mounted on
/dev/ad0s1a 18G 18G -1.4G 109% /
devfs 1.0k 1.0k 0B 100% /dev
/dev/md0 1.8M 29k 1.6M 2% /var/run
devfs 1.0k 1.0k 0B 100% /var/dhcpd/dev

# cd / && du -hd1
2.0K ./.snap
5.3M ./boot
780K ./bin
1.5K ./dev
16K ./conf.default
4.3M ./etc
158K ./libexec
3.4M ./lib
14K ./root
1.9M ./sbin
91M ./usr
17G ./var
9.1M ./tmp
2.0K ./mnt
458K ./cf
2.0K ./media
2.0K ./proc
2.0K ./rescue
2.0K ./kernels
12K ./scripts
17G .

Do you know solution this problem? Thanks Ladislav

Cry Havok

Your /var is 17 GB. Now repeat the du command inside /var.