[Gllug] Odd (but apparently correct) behaviour from du

John Winters john at sinodun.org.uk
Wed Jul 16 14:22:12 UTC 2008

I've just come across something slightly odd which I thought I should share.

I back up my systems using dirvish run nightly on a dedicated machine
and the other night the job took over twice as long as expected.  I was
just looking into why, and to see whether the disc usage in one of my
backed up partitions had suddenly increased a lot I went to the relevant
vault and typed:

du -sk *

The results which came back were startlingly variable.  Instead of a
consistent 30G or so (slowly increasing) in each snapshot I got wildly
varying values.  One night's snapshot would be 31G and the next night's
apparently 2G.  Looking at them manually however they all seemed to be

Worrying thoughts about a corrupt filesystem started flitting across my
mind, but a bit of research eventually gave the answer (and to me it's a
surprising answer).  Quote from an online man page which I found "Files
with multiple links shall be counted and written for only one entry.
The directory entry that is selected in the report is unspecified."  As
far as I can see, this information is missing from the man page provided
in Linux.

So - if you're using dirvish you have to be very careful about how you
use du.  By changing my invocation to:

for $root in *; do; du -sk $root; done

I got the right answer.  Now hands up all those who knew that du behaved
this way.

Gllug mailing list  -  Gllug at gllug.org.uk

More information about the GLLUG mailing list