Date: Sun, 21 Jan 2007 20:12:24 -0500 From: "Ahnjoan Amous" <ahnjoan@gmail.com> To: "Scott Long" <scottl@samsco.org> Cc: freebsd-fs@freebsd.org Subject: Re: aac scsi raid driver performance Message-ID: <5e575c8a0701211712t4ea42316we02b03221a901841@mail.gmail.com> In-Reply-To: <5e575c8a0701201003j4de54712h6913f0edd62c4ffd@mail.gmail.com> References: <5e575c8a0701111155l859d0ecif617dbda43cef842@mail.gmail.com> <45A6A138.8090208@samsco.org> <5e575c8a0701111307s5a839b82ra4ba5c45d554d3b9@mail.gmail.com> <45A7551C.5030006@samsco.org> <5e575c8a0701130703y6c528cecoed9019472a4956c8@mail.gmail.com> <5e575c8a0701201003j4de54712h6913f0edd62c4ffd@mail.gmail.com>
next in thread | previous in thread | raw e-mail | index | archive | help
I can not say I understand why, but a CPU upgrade seems to have solved my I/O problems. If you read this thread you will see a fairly consistent test of two concurrent sequential writes to different SCSI devices on the same SCSI channel. This test was to figure out why GEOM mirror writes were so slow, and GEOM mirror tests were to figure out why ggatec/ggated duplication was so slow. I have re-installed my slower CPUs and see the poor IO performance return so I am fairly confident that my conclusion can be supported by testing. FYI =96 Hardware PowerEdge 2650, 2.4G/533/512L2/0L3, 1G ram, PERC 3/Di, 2x36G & 3x300G drive= s I have two identical 2650s in this configuration for testing. The CPU upgrade is a 3.066G/533/512L2/1ML3 For more detail in the difference=85 Old CPU (slow concurrent IO) model number is SL6VL New CPU (fast concurrent IO) model number is SL72G Now on to e-bay to find a couple more of these! Results: My IO speed almost tripled=85 I'm able to get over 55MB/sec to each drive in the concurrent write test on the new CPU, much better than the 20MB/sec to each drive in the same test on the old CPU. Thanks again for all the help.
Want to link to this message? Use this URL: <https://mail-archive.FreeBSD.org/cgi/mid.cgi?5e575c8a0701211712t4ea42316we02b03221a901841>