I'd be interested in the output of iostat and vmstat while the apache benchmark is running. Also, I question your '90% faster claim'.
> RHEL 5: Requests per second: 9426.71 (10.608151 seconds) > RHEL 6: Requests per second: 5595.64 (17.871 seconds) 5595 / 9426 = 0.593 (59.3%) CC PS: If the above is proved wrong then obviously the email was corrupted in transit. On Tue, Nov 9, 2010 at 10:58 AM, Ashley Penney <apen...@gmail.com> wrote: > One thing to consider is the 'nobarrier' option for ext4 - apparently makes > a big performance difference. > Thanks, > > On Mon, Nov 8, 2010 at 9:21 PM, Marco Shaw <marco.s...@gmail.com> wrote: >> >> I'd love to try it and break it down using SystemTap... I just don't >> have the free time. >> >> "RHEL 5.5 + updates" versus "RHEL 6 Beta 2 July refresh"? >> >> On Mon, Nov 8, 2010 at 10:12 PM, Morten P.D. Stevens >> <mstev...@imt-systems.com> wrote: >> > 2010/11/9 Marco Shaw <marco.s...@gmail.com>: >> > >> >> Well, right off, when you listed your hardware, I never assumed it was >> >> virtualized. At the very least, I would want to see this test in a >> >> hypervisor virtualized environment, and even better directly on >> >> hardware. >> >> >> >> Adding an virtual application layer adds so much to the equation that >> >> your RHEL5 being faster than RHEL6 is no longer valid, in my mind. >> > >> > A few days ago I had the same tests on real hardware and VMWare ESXi 4.1 >> > with exactly the same results. >> > >> > If you want you can like to try it yourself. >> >> _______________________________________________ >> rhelv6-beta-list mailing list >> rhelv6-beta-list@redhat.com >> https://www.redhat.com/mailman/listinfo/rhelv6-beta-list > > > _______________________________________________ > rhelv6-beta-list mailing list > rhelv6-beta-list@redhat.com > https://www.redhat.com/mailman/listinfo/rhelv6-beta-list > -- RHCE#805007969328369 _______________________________________________ rhelv6-beta-list mailing list rhelv6-beta-list@redhat.com https://www.redhat.com/mailman/listinfo/rhelv6-beta-list
