On Wed, 2006-02-15 at 21:24 -0500, Ming Zhang wrote:
> On Mon, 2006-02-06 at 12:05 -0800, Brad Dameron wrote:
> > > > I first did cross-over cables from the initiators to the target
> > ports so
> > > > each one had a 1Gb connection. This was definitely the faster way.
> > I saw
> > > > max of about 70MB/sec reads/writes with both doing simultaneous
> > test's.
>
> what u mean "simultaneous test's"? my feeling is that you could get
> better performance with your nice HW.
>
> Ming
>
Simultaneous as in reading/writing at the same time to the same iSCSI
target. I also tried tests where one read and the other wrote. I did
some additional testing to try to see if GFS added any latency and could
not see it causing anything major since all it does is control locking
between machines.
My current test I am working on is using 4 machines all with targets and
initiators on them striping all the machines together in a RAID10 and
using GFS. This way I can leverage the speed of all machines with drives
and offer redundancy. I do believe the proper name for this is "piSCSI".
Or parallel iSCSI.
Brad Dameron
SeaTab Software
www.seatab.com
-------------------------------------------------------
This SF.net email is sponsored by: Splunk Inc. Do you grep through log files
for problems? Stop! Download the new AJAX search engine that makes
searching your log files as easy as surfing the web. DOWNLOAD SPLUNK!
http://sel.as-us.falkag.net/sel?cmd=lnk&kid=103432&bid=230486&dat=121642
|
