On Tue, Nov 18, 2008 at 11:51 AM, Gary Gendel <[EMAIL PROTECTED]> wrote: > This past weekend I set up a MacBook Pro to time machine via CIFS to my > OpenSolaris server. > > It took over 7 hours to do the initial population using Samba as the CIFS > server. During that time, I noticed that the server was showing a relatively > high load... About 45% cpu (basically, it used 1 of my 2 cpus) and a ton of > filesystem activity. A quick Dtrace showed me that it was almost all Samba, > which was doing about 2000 llseeks per second. > > I shut down Samba and fired up the kernel cifs stuff. I ran the initial > population again. This time it took ~2 hours and the CPU was well under 1% > the whole time. Looks like it's network bandwidth limited now. > > Thanks for the hard work and the impressive results. > > Gary > -- > This message posted from opensolaris.org > _______________________________________________ > cifs-discuss mailing list > [email protected] > http://mail.opensolaris.org/mailman/listinfo/cifs-discuss >
I'd like to say that the in-kernel Solaris CIFS server blows almost all over implementations out of the water for performance. I love the Samba project, it does a great deal of work and for compatibility, nothing can touch it. I hope Samba can use some of the technologies of the in-kernel CIFS server, and visa versa, they could both benefit from eachother -- Brent Jones [EMAIL PROTECTED] _______________________________________________ cifs-discuss mailing list [email protected] http://mail.opensolaris.org/mailman/listinfo/cifs-discuss
