From: Jan-Olaf Droese (jan-olaf.droese@lahmeyer.de)
Date: Thu Sep 19 2002 - 08:27:23 EDT
Dear sun-managers,
this is sort of a late summary for my question regarding df displaying
wrong numbers (used + avail >> total space in filesystem). The original
question is attached below.
Although I received a couple of answers the problem is still
unresolved. The answers I received fell into the following categories:
- not enough inodes (we have plenty of free inodes, I checked this
before posting)
- the problem has something to do with the free space reserved when
creating the filesystem (I don't think this is the problem here,
because I was doing my tests as root who is allowed to use up
the reserved space AFAIK. Also, I have two other identical
machines which show used == total space and avail = 0k when their
filesystem is full.)
- df does not count the size of open files, so these might be
eating up the diskspace. I have restarted all relevant processes,
but df still displays the same inconsistency. Also, we are missing
some 15GB and the machine is only used as a samba file- and
printserver for PC clients so this would mean 1000s of open files
which are eating up diskspace.
Currently I'm not sure what to do. The next thing to try is a
reboot of the machine, but this is difficult to organise because
there are lots of users on it.
The df-output of the filesystem in question still displays
Filesystem kbytes used avail capacity Mounted on
/dev/dsk/c2t0d0s6 275823048 268663257 22527159 93% /export/raid
Error messages logged when the fs filled up the last time were
Aug 30 10:00:58 host unix: NOTICE: realloccg /export/raid: file system full
Aug 30 10:01:02 host unix: NOTICE: alloc: /export/raid: file system full
Thanks to all who offered their help,
Jan-Olaf
Original question:
===========================================================================
Dear sun-managers,
after one of our servers started reporting "file system full" errors
whereas all our monitoring tools told us that there were still some
15Gb free I noticed some strange problem:
For that file system the df-output seems to be wrong. Used+avail
capacity is about 15Gb more than the total space in that file system
(see df output below). So when "used" equaled "kbytes" the machine
started reporting "file system full". "avail" still showed 15Gb free.
Does anyone know what might be the cause of this problem? I have
two other identical machines, where the df output is correct (the only
difference there is that the file systems on those machines only
have 150Gb capacity).
The machine is a Ultra-60 running Solaris 7 with kernel patch 106541-11
and a patch cluster which contained that kernel patch. The file
system in question is mounted with the "logging" option.
Thanks a lot for any help to fix the problem. As usual I will post
a summary.
Jan-Olaf
df-output:
bash# df -k /export/raid
Filesystem kbytes used avail capacity Mounted on
/dev/dsk/c2t0d0s6 275823048 261513716 29676700 90% /export/raid
Format shows the following:
partition> print
Current partition table (original):
Total disk cylinders available: 35773 + 2 (reserved cylinders)
Part Tag Flag Cylinders Size Blocks
0 unassigned wm 0 0 (0/0/0) 0
1 unassigned wu 0 0 (0/0/0) 0
2 backup wu 0 - 35772 279.48GB (35773/0/0) 586104832
3 unassigned wm 0 0 (0/0/0) 0
4 unassigned wm 0 0 (0/0/0) 0
5 unassigned wm 0 0 (0/0/0) 0
6 usr wm 0 - 35772 279.48GB (35773/0/0) 586104832
7 unassigned wm 0 0 (0/0/0) 0
===========================================================================
-- Jan-Olaf Droese | Lahmeyer International GmbH email: jan-olaf.droese@lahmeyer.de | Friedberger Stra_e 173 phone: +49-6101-55-1611 | 61118 Bad Vilbel fax: +49-6101-55-1623 | Germany _______________________________________________ sunmanagers mailing list sunmanagers@sunmanagers.org http://www.sunmanagers.org/mailman/listinfo/sunmanagers
This archive was generated by hypermail 2.1.7 : Wed Apr 09 2008 - 23:24:57 EDT