On 05/11/2012 09:25 AM, Aaron Davies wrote:
> `ls -l' (v. 8.12) is often brutally slow on our networked filesystems (nfs & 
> afs) at work (~40s for a 1500 file directory). stracing reveals that almost 
> all the time (~4.5ms per call) is spent in lgetxattr, which doesn't even 
> appear to be supported on these fs's. (all the calls return -1 and report 
> EOPNOTSUPP (as do all the calls to getxattr, but they only seem to take ~43µs 
> each, slightly less than the lstat's which are doing the actual work (~39µs), 
> so that isn't such a big deal.)
> 
> is this expected behavior?

Upgrade.  From the NEWS for 8.16 (and 8.17 is the current release), we have:

  ls can be much more efficient, especially with large directories on file
  systems for which getfilecon-, ACL-check- and XATTR-check-induced syscalls
  fail with ENOTSUP or similar.

-- 
Eric Blake   [email protected]    +1-919-301-3266
Libvirt virtualization library http://libvirt.org

Attachment: signature.asc
Description: OpenPGP digital signature

Reply via email to