SUMMARY: filesystem full

From: Ulla Fischer (ulla@dmi.min.dk)
Date: Thu Jan 26 1995 - 03:33:45 CST


Dear managers.

I solved the problem myself after a good night sleep.

I tried to fsck /, and I got some errors. It was lp that had left some
files (3Mb) on the disk. This is the third time in half a year I
experience that with lp.

Yours,

-- 
Ulla Fischer
Danmarks Meteorologiske Institut      Email: ulla@dmi.min.dk
Edb-afdelingen                        Phone: + 45 39 15 75 00 
Lyngbyvej 100                         Phone: + 45 39 15 75 54
2100 Koebenhavn 0                     Fax:   + 45 39 15 75 98
Denmark

My original question:

>From ulla Wed Jan 25 14:01 GMT 1995 Return-Path: <ulla> From: ulla (Ulla Fischer) Date: Wed, 25 Jan 95 14:01:51 GMT To: sun-managers@eecs.nwu.edu Cc: ulla, jab Subject: filesystem full Content-Type: text Content-Length: 4130

Dear managers.

I have asked my local Sun dealer without any result and checked the Solaris2/FAQ made by casper@fwi.uva.nl (ftp.fwi.uva.nl:/pub/solaris) without luck.

Problem: df and the operating system thinks that the filesystem is full, ls and du does not.

/ is 20 Mb. /var located on /. The machine is an enduser, so there is not added any special software. Our experience tells us, that 20 Mb /, which works just fine in this configuration. This machine have also been working fine until yesterday.

Now df says: # df -k / Filesystem kbytes used avail capacity Mounted on /dev/dsk/c0t0d0s0 20223 19261 0 100% /

In /var/adm/messages it says: Jan 25 09:07:24 hobbes unix: NOTICE: alloc: /: file system full Jan 24 13:09:37 hobbes unix: NOTICE: realloccg /: file system full

du says: du -dk /: 15335 Kb

and find says # find / -mount \! -type d -print | xargs ls -ls | awk '{ s += $1 } END { print s }' 30438 blocks Which is about 14Mb

/var uses 4046 Kb of the root partition du -dk /var: 4046 Kb

Why do du -dk / and df -k / not agree on the used space? I have not changed vfstab in about half a year, so I do not mount on top of some files. I have rebooted, without any change. I havn't fsck'd / because fsck didn't ask me to during the boot. Do I have a file with holes? If so, how do I locate such one? (I do not have any core-files)!

If I compare with another sane comparable machine (opr22), I get the following results: # find / -mount \! -type d -print | xargs ls -ls | awk '{ s += $1 } END { print s }' 27768 blocks Which is about 13Mb du -dk /: 13789 Mb du -dk /var: 4378 Mb

opr22# df -k / Filesystem kbytes used avail capacity Mounted on /dev/dsk/c0t3d0s0 22927 13799 6838 67% /

Facts: Machine: IPC, Solaris 2.3 with patches 101317-09 101318-54 101327-06 101331-03 101344-08 101347-02 101615-02 101739-02 101362-17 101593-02 101495-01 # df -k Filesystem kbytes used avail capacity Mounted on /dev/dsk/c0t0d0s0 20223 19261 0 100% / /dev/dsk/c0t0d0s6 213717 158748 33599 83% /usr /proc 0 0 0 0% /proc fd 0 0 0 0% /dev/fd swap 33236 32 33204 0% /tmp /dev/dsk/c0t0d0s7 621213 19462 539631 3% /export/home /dev/dsk/c0t0d0s5 59893 44239 9674 82% /opt edb:/export/home/jab 962054 761895 103959 88% /home/jab mailserver:/priv_3/mail 481823 3988 429655 1% /var/mail butler:/priv_2/emacs/SunOS5 4880572 666749 3725773 15% /pack/emacs

vfstab: /dev/dsk/c0t0d0s0 /dev/rdsk/c0t0d0s0 / ufs 1 no - /dev/dsk/c0t0d0s6 /dev/rdsk/c0t0d0s6 /usr ufs 2 no - /dev/dsk/c0t0d0s7 /dev/rdsk/c0t0d0s7 /export/home ufs 5 yes - /dev/dsk/c0t0d0s5 /dev/rdsk/c0t0d0s5 /opt ufs 6 yes - /dev/dsk/c0t0d0s1 - - swap - no - /proc - /proc proc - no - fd - /dev/fd fd - no - swap - /tmp tmpfs - yes -

Current partition table (original): Part Tag Flag Cylinders Size Blocks 0 root wm 0 - 81 21.62MB (82/0/0) 1 swap wu 82 - 264 48.25MB (183/0/0) 2 backup wm 0 - 3802 1002.74MB (3803/0/0) 3 unassigned wm 0 0 (0/0/0) 4 unassigned wm 0 0 (0/0/0) 5 - wm 265 - 501 62.49MB (237/0/0) 6 usr wm 502 - 1346 222.80MB (845/0/0) 7 home wm 1347 - 3802 647.58MB (2456/0/0)

Any help/hints are appreciated.



This archive was generated by hypermail 2.1.2 : Fri Sep 28 2001 - 23:10:14 CDT