samba trunk slower than SAMBA_3

Jeremy Allison jra at
Mon Jun 20 18:49:44 GMT 2005

On Mon, Jun 20, 2005 at 01:40:34PM -0500, Steve French wrote:
> I ran the connectathon test suite in timing (-t) mode and see that the
> Samba trunk branch was much, much slower than SAMBA_3 branch on
> file/directory create tests (0.45 vs. 0.14 seconds creating 155 files,
> 62 directories).  For more files (test1 -t 4 70 10 ie four levels deep,
> 70 files, 10 directories each) the difference, as expected, is much more
> dramatic: 41.76 seconds for trunk while only 2.82 seconds for SAMBA_3
> Perhaps the case sensitivity on path based calls is not being detected
> in trunk - but this is probably what made dbench run so much slower in
> trunk and makes it harder to tell the effects of the aio

No, there's no difference in the case-sensitivity detection between
3.0 and HEAD. Must be something else.... I'll do a dirdiff and look
for differences.


More information about the samba-technical mailing list