He1p Required - Please??? (fwd)

iang at iafrica.com iang at iafrica.com
Fri Aug 13 15:58:49 UTC 1999


Lloyd Rennie wrote:
> The negative numbers are not a problem.  Last time this happened, we ruled
> it out as a possibility.  Can't remember why though - it was late at
> night... (ain't it always?)  I see negative available space quite often...
> Only in BSD though.

Doh! Can't believe I missed this one. FreeBSD's df will start
counting negative disk space once the file system is full.
Apporximately 10% of the file system is reserved for root once the
file system has filled up.

Contrived example:
If you have a 100M file system, df of the empty file system will show:
Filesystem   1k-blocks     Used    Avail Capacity  Mounted on
/dev/da0s1a     102400        0    92160     0%    /

When you fill the file system, to 100%:
Filesystem   1k-blocks     Used    Avail Capacity  Mounted on
/dev/da0s1a     102400    92160        0   100%    /

Now ordinary users may not create files on this file system, however,
root may continue to append to and create files. The following may
happen:

Filesystem   1k-blocks     Used    Avail Capacity  Mounted on
/dev/da0s1a     102400   101376    -9216   110%    /

You can use tunefs to adjust the amount of spare space in the file
system. 'man tunefs' for more info:
     -m minfree
	     Specify the percentage of space held back from normal
	     users; the minimum free space threshold.  The default
	     value used is 8%.  This value can be set to zero,
	     however up to a factor of three in throughput will be
	     lost over the performance obtained at a 10% threshold.
	     Settings of 5% and less force space optimization to
	     always be used which will greatly increase the overhead
	     for file writes.  Note that if the value is raised
	     above the current usage level, users will be unable
	     to allocate files until enough files have been deleted
	     to get under the higher threshold.

If you want to change this, you will have to shutdown the news
server and perform the tunefs on the unmounted file systems.

> Server throttled No space left on device writing SMstore file -- throttling
> Allowing remote connections
> Parameters c 14 i 50 (0) l 1000000 o 96 t 300 H 2 T 60 X 0 normal specified
> Not reserved
> Readers separate disabled No space left on device writing SMstore file --
> throttling
> Perl filtering enabled

That was innwatch throttling the server. "ctlinnd go ''" will let
it resume. See ~news/etc/innwatch.ctl for the values and configurable
options.

Hope this information helps.

-- 
Ian Freislich
UUNET S.A. (Pty) Ltd


More information about the inn-workers mailing list