some feedback...
I created my own quick and dirty concurrent connection pressure test which is
not very scientific, but importantly I could still access resources fine on the
opensolaris server from Windows with CIFS while the test was running.
The test was run from both Mac OS10.5 and SLES 10 SP2 both connecting to single
CIFS shared zfs filesystem on opensolaris 2009/06 containing users home
directories.
I created a bash script to mount a total of 600 opensolaris home directories
with 600 users onto 600 mount points on the clients.So I then had 600
concurrent CIFS connections as viewed from netstat on opensolaris.
I spawned off simultaneous instances of fdtree -
https://computing.llnl.gov/?set=codepage=sio_downloads which is 'software (a
bash script) is used for testing the metadata performance of a file system' and
therefore created load. In total there were 300 fdtree actions creating complex
directory and file structures in each CIFS share mount point.
My opensolaris server is just a VM running in ESX server, allocated 1 CPU and
6GB RAM.
As I say, it is not hugely deterministic, but it goes some way to giving me the
confidence that CIFS on solaris can handle some anger.
--
This message posted from opensolaris.org
___
cifs-discuss mailing list
cifs-discuss@opensolaris.org
http://mail.opensolaris.org/mailman/listinfo/cifs-discuss
