Server Fault is a question and answer site for system and network administrators. It's 100% free, no registration required.

Sign up
Here's how it works:
  1. Anybody can ask a question
  2. Anybody can answer
  3. The best answers are voted up and rise to the top

I am currently experiencing very bad performance using the following on an NFS network folder:

time find . | while read f; do test -L "$f" && f=$(readlink -m $f); grp="$(stat -c %G $f)"; perm="$(stat -c %A $f)"; done

Question 1) Within the loop permissions are checked using the variables grp and perm. Is there a way to lower the amount of disc I/O for these kind of checks over the network (e.g. read all meta data at once using find)?

Question 2) It seems like the NFS isn't tuned very well, the same operation on a similar network link via SSHFS take only one third of the time. All parameters are auto-negotiated. Any suggestions?

share|improve this question

Your line is performing three calls for each file; a single stat + parsing the output would be enough. For starters, re-design your script to call stat only once with stat -c "%n %G %A" ... if you need help with that, throw us a comment.

share|improve this answer
Seems like I didn't quite check the f** manuals. Sorry. I posted a version using only find for meta-info. – fungs Jun 5 '12 at 16:24

Fastest solution I found during the last hour was:

failed=$(find -L . -printf "%p %g %M\n" | awk '{ if ($2 != "XYZ"){ printf $1; exit 1 }; if ( substr( $3, 9, 1 ) != "-" ) { printf $1; exit 2 } }')
test ! $ret -eq 0 && echo "Error with file $failed"

which checks for some group owner and permission as an example. This version using only find but not stat and following symbolic links is at least by a factor of 100 faster.

share|improve this answer

Your Answer


By posting your answer, you agree to the privacy policy and terms of service.

Not the answer you're looking for? Browse other questions tagged or ask your own question.