kvm-83-183.el5 is not released? Regards, Kirby Zhou from SOHU-RD +86-10-6272-8261
-----Original Message----- From: [email protected] [mailto:[email protected]] On Behalf Of nodata Sent: Tuesday, August 10, 2010 12:06 AM To: Red Hat Enterprise Linux 5 (Tikanga) discussion mailing-list Subject: Re: [rhelv5-list] [rhelv6-beta-list] KVM performance is very poor? Fixed in version: kvm-83-183.el5 You have kvm-83-164.el5_5.15 On 09/08/10 15:34, Kirby Zhou wrote: > I think I have already applied this update: > > ]# rpm -q kvm kmod-kvm kvm-qemu-img kvm-tools kernel > kvm-83-164.el5_5.15 > kmod-kvm-83-164.el5_5.15 > kvm-qemu-img-83-164.el5_5.15 > kvm-tools-83-164.el5_5.15 > kernel-2.6.18-194.8.1.el5 > > /dev/vgext/kvm-rhel5guest on host == /dev/vda on guest > > host]# ./randio /dev/vgext/kvm-rhel5guest 10 100000 -j crc32 -j md5 -t seq -Q -d > total time=13.217346 io=100000 (7565.82/s) iosz=3200000k (242106.09k/s) rd=100000 (7565.82/s) rdsz=3200000k (242106.09k/s) wr=0 (0.00/s) wrsz=0k (0.00k/s) > > guest]# ./randio /dev/vda 10 100000 -j crc32 -j md5 -t seq -Q -d > total time=23.319551 io=100000 (4288.25/s) iosz=3200000k (137223.91k/s) rd=100000 (4288.25/s) rdsz=3200000k (137223.91k/s) wr=0 (0.00/s) wrsz=0k (0.00k/s) > > Regards > Kirby Zhou > > -----Original Message----- > From: [email protected] [mailto:[email protected]] On Behalf Of Alexander Dalloz > Sent: Monday, August 09, 2010 8:45 PM > To: 'Red Hat Enterprise Linux 5 discussion mailing-list' > Subject: Re: [rhelv5-list] [rhelv6-beta-list] KVM performance is very poor? > >> KVM performance is very poor? >> >> I have tested a simple IO/CPU mixture testcase under both RHEL5/Xen and >> RHEL6/KVM guest. > >> Under the naked machine, both RHEL5 and RHEL6 can get 250M/s, >> XEN guest can get 240M/s >> KVM guest can get only 120M/s > >> Regards >> Kirby Zhou > > Do not know the KVM version you tested against. But on RHEL 5 there was a > performance hitting bug > > https://rhn.redhat.com/errata/RHBA-2010-0536.html > > * the virtio-blk back end for qemu-kvm performed unnecessary zeroing of > memory on every I/O request, which reduced virtual guest performance. With > this update,t he virtio-blk back end avoids spurious zeroing of request > structures on each I/O request, with the result that guest performance is > increased. (BZ#604159) > > Alexander > > > > _______________________________________________ > rhelv5-list mailing list > [email protected] > https://www.redhat.com/mailman/listinfo/rhelv5-list > > > _______________________________________________ > rhelv5-list mailing list > [email protected] > https://www.redhat.com/mailman/listinfo/rhelv5-list > _______________________________________________ rhelv5-list mailing list [email protected] https://www.redhat.com/mailman/listinfo/rhelv5-list _______________________________________________ rhelv5-list mailing list [email protected] https://www.redhat.com/mailman/listinfo/rhelv5-list
