virt September 2015

virt@lists.centos.org

33 participants
33 discussions

kernel-3.18.21-13.el6 into virt6-xen-44-testing
by Johnny Hughes 24 Sep '15

24 Sep '15

The following have been added to the virt6-xen-44-testing repo: x86_64: kernel-3.18.21-13.el6.x86_64 kernel-devel-3.18.21-13.el6.x86_64 kernel-firmware-3.18.21-13.el6.noarch kernel-headers-3.18.21-13.el6.x86_64 perf-3.18.21-13.el6.x86_64 src: kernel-3.18.21-13.el6.x86_64 Please test these for a week or so and we will move them into production next week. I am also building this kernel into virt7-xen-44-testing and virt7-xen-46-testing, should be there in the next couple of hours. Thanks, Johnny Hughes

1 0

xen-4.4.3-1 packages released
by Johnny Hughes 24 Sep '15

24 Sep '15

I have signed and released the xen4centos CentOS-6 packages to our main mirrors, the following are released: x86_64: xen-4.4.3-1.el6.x86_64.rpm xen-devel-4.4.3-1.el6.x86_64.rpm xen-doc-4.4.3-1.el6.x86_64.rpm xen-hypervisor-4.4.3-1.el6.x86_64.rpm xen-libs-4.4.3-1.el6.x86_64.rpm xen-licenses-4.4.3-1.el6.x86_64.rpm xen-ocaml-4.4.3-1.el6.x86_64.rpm xen-ocaml-devel-4.4.3-1.el6.x86_64.rpm xen-runtime-4.4.3-1.el6.x86_64.rpm src: xen-4.4.3-1.el6.src.rpm

1 0

Re: [CentOS-virt] Trouble installing KVM
by paul.greene.va＠verizon.net 22 Sep '15

22 Sep '15

1 0

Trouble installing KVM
by paul.greene.va＠verizon.net 22 Sep '15

22 Sep '15

2 1

Re: [CentOS-virt] Official openvswitch package for CentOS7
by Sandro Bonazzola 21 Sep '15

21 Sep '15

On Mon, Sep 21, 2015 at 1:53 PM, Haïkel Guémar <hguemar(a)redhat.com> wrote: > On 21/09/15 12:58, Alan Pevec wrote: > >> +1 on my side. oVirt has several packages tagged in from cloud SIG, > >> maybe we can have them in this virt-common repo as well. Or maybe > >> already exists something like cloud-common we can depend on. > > > > We might end-up in circular dependecies, looks like we actually need > > cross-SIG-common for EPEL rebuilds and stuff like OVS. > > > > As for cloud-common: no we don't have it defined yet, OpenStack/RDO were > > the only active cloud-infra project in CentOS CloudSIG, and we only had > > cloud-openstack-common for OpenStack release-agnostic dependencies. > > > > Cheers, > > Alan > > > > +2 for a CBS common, it's not realistic to maintain common EPEL rebuilds > or keep retagging them from other SIGs. > > Alan suggested that we encourage EPEL to keep in fedora dist-git and > build in CBS. That could be a way to solve the issue around EPEL. > It solves the problem that EPEL is lacking contributors, and duplicating > work within CBS. > that would be great, +1 > > Regards, > H. > -- Sandro Bonazzola Better technology. Faster innovation. Powered by community collaboration. See how it works at redhat.com

1 0

Official openvswitch package for CentOS7
by C. L. Martinez 21 Sep '15

21 Sep '15

Hi all, Exists an official openvswitch package for CentoS7?? If I am not wrong, this: http://mirror.centos.org/centos/7/cloud/x86_64/openstack-kilo/openstack-neu… it can be used with openstack only. Do I need to recompile from source like is is explained here: https://n40lab.wordpress.com/2015/06/28/centos-7-installing-openvswitch-2-3… or exists another rpm package for CentOS7?? Thanks.

5 8

Re: [CentOS-virt] Official openvswitch package for CentOS7
by George Dunlap 21 Sep '15

21 Sep '15

On Fri, Sep 18, 2015 at 4:22 PM, Alan Pevec <alan.pevec(a)redhat.com> wrote: >>> Alan what do you think? Should we move openvswitch from cloud/openstack to >>> virt/kvm-common ? >> >> Since Xen users may want it too, it should probably have its own repo >> (so that both virt/kvm and virt/xen can use it). > > Yeah, is there virt/common ? If not, there definitely should be. :-) I'm actually not working today -- Sandro and Alan, let's sync up over e-mail on Monday (and/or at the Virt SIG meeting on IRC on Tuesday). -George

2 1

Re: [CentOS-virt] poor performance with dom0 on centos7
by Christoph 17 Sep '15

17 Sep '15

Am 2015-09-17 09:29, schrieb Pasi Kärkkäinen: > > Are you using nfs over UDP or TCP ? > TCP, but Network cant be the bottleneck, have tested it with iperf between bare metal/domU's and the nfs domU and it was perfectly fast... > > I don't think. > > > If you used NFS over UDP, try running it over TCP. no I use it over TCP... > > What does 'top' and/or 'iostat -x 1' say during the 'benchmark' ? top: top - 09:01:12 up 22:45, 1 user, load average: 1,97, 2,01, 1,99 Tasks: 210 total, 1 running, 209 sleeping, 0 stopped, 0 zombie %Cpu0 : 0,3 us, 1,0 sy, 0,0 ni, 91,4 id, 7,3 wa, 0,0 hi, 0,0 si, 0,0 st %Cpu1 : 0,0 us, 0,0 sy, 0,0 ni,100,0 id, 0,0 wa, 0,0 hi, 0,0 si, 0,0 st %Cpu2 : 0,0 us, 0,0 sy, 0,0 ni,100,0 id, 0,0 wa, 0,0 hi, 0,0 si, 0,0 st %Cpu3 : 0,0 us, 0,3 sy, 0,0 ni, 13,0 id, 86,7 wa, 0,0 hi, 0,0 si, 0,0 st KiB Mem : 1013016 total, 19548 free, 591456 used, 402012 buff/cache KiB Swap: 1048572 total, 990776 free, 57796 used. 353468 avail Mem iostat: avg-cpu: %user %nice %system %iowait %steal %idle 0,00 0,00 0,00 50,00 0,00 50,00 Device: rrqm/s wrqm/s r/s w/s rkB/s wkB/s avgrq-sz avgqu-sz await r_await w_await svctm %util xvda 0,00 0,00 0,00 0,00 0,00 0,00 0,00 0,00 0,00 0,00 0,00 0,00 0,00 xvdb 0,00 0,00 0,00 0,00 0,00 0,00 0,00 0,00 0,00 0,00 0,00 0,00 0,00 xvdc 0,00 0,00 0,00 0,00 0,00 0,00 0,00 0,00 0,00 0,00 0,00 0,00 0,00 xvdd 0,00 0,00 0,00 26,00 0,00 2336,00 179,69 69,31 1060,62 0,00 1060,62 38,46 100,00 xvde 0,00 0,00 0,00 0,00 0,00 0,00 0,00 0,00 0,00 0,00 0,00 0,00 0,00 xvdf 0,00 0,00 0,00 0,00 0,00 0,00 0,00 0,00 0,00 0,00 0,00 0,00 0,00 xvdg 0,00 0,00 0,00 0,00 0,00 0,00 0,00 0,00 0,00 0,00 0,00 0,00 0,00 xvdh 0,00 0,00 0,00 0,00 0,00 0,00 0,00 0,00 0,00 0,00 0,00 0,00 0,00 xvdi 0,00 0,00 0,00 0,00 0,00 0,00 0,00 0,00 0,00 0,00 0,00 0,00 0,00 xvdj 0,00 0,00 0,00 0,00 0,00 0,00 0,00 0,00 0,00 0,00 0,00 0,00 0,00 dm-0 0,00 0,00 0,00 0,00 0,00 0,00 0,00 0,00 0,00 0,00 0,00 0,00 0,00 dm-1 0,00 0,00 0,00 0,00 0,00 0,00 0,00 0,00 0,00 0,00 0,00 0,00 0,00 dm-2 0,00 0,00 0,00 0,00 0,00 0,00 0,00 0,00 0,00 0,00 0,00 0,00 0,00 dm-3 0,00 0,00 0,00 0,00 0,00 0,00 0,00 0,00 0,00 0,00 0,00 0,00 0,00 dm-4 0,00 0,00 0,00 0,00 0,00 0,00 0,00 0,00 0,00 0,00 0,00 0,00 0,00 dm-5 0,00 0,00 0,00 0,00 0,00 0,00 0,00 69,31 0,00 0,00 0,00 0,00 100,00 dm-6 0,00 0,00 0,00 0,00 0,00 0,00 0,00 0,00 0,00 0,00 0,00 0,00 0,00 so not really a problem... -- ------ Greetz

2 1

Re: [CentOS-virt] poor performance with dom0 on centos7
by Keith Roberts 17 Sep '15

17 Sep '15

On 2 Sep 2015, at 09:22, Pasi Kärkkäinen <pasik(a)iki.fi> wrote: > On Tue, Sep 01, 2015 at 06:47:18AM +0200, Christoph wrote: >> Hi All >> > > Hello, > >> it is possible to tune dom0/domU for better IO/network performance? >> Since I have changed to Cenots7 dom0, I have a really poor IO >> performance inside a PV VM. >> >> I have already done what is described on >> http://wiki.xenproject.org/wiki/Tuning_Xen_for_Performance >> It is better now but still significantly worse than with centos6 dom0 >> >> my settings: >> >> xen parameter: dom0_mem=1024M cpufreq=xen dom0_max_vcpus=2 >> dom0_vcpus_pin >> >> xl sched-credit >> Cpupool Pool-0: tslice=30ms ratelimit=1000us >> Name ID Weight Cap >> Domain-0 0 1024 0 >> samael 1 256 0 >> satan 2 512 0 >> amon 3 256 0 >> leviathan 4 512 0 >> >> echo 1048576 > /proc/sys/net/ipv4/tcp_limit_output_bytes on dom0 >> >> the both domU's satan and leviathan are very IO performance oriented >> (NFS server and downloading vm) >> >> Is there something more what I can do or try? >> >> could it be a selinux issue? I have it in permissive mode there, not >> disabled. But permissive means only to collect the info not enforcing >> the rules... >> >> (I use the xen45 pkgs) >> > > You forgot to mention the most important thing.. what kind of performance numbers are you seeing? What are you expecting? > > > Thanks, > > — > Pasi Here’s a useful disk i/o stress test program that I got good results with on openSUSE 13.1, by running it on several VM’s overnight: Using Bonnie++ for filesystem performance benchmarking http://archive09.linux.com/feature/139742 HTH Keith Roberts

1 0

poor performance with dom0 on centos7
by Christoph 17 Sep '15

17 Sep '15

Hi All it is possible to tune dom0/domU for better IO/network performance? Since I have changed to Cenots7 dom0, I have a really poor IO performance inside a PV VM. I have already done what is described on http://wiki.xenproject.org/wiki/Tuning_Xen_for_Performance It is better now but still significantly worse than with centos6 dom0 my settings: xen parameter: dom0_mem=1024M cpufreq=xen dom0_max_vcpus=2 dom0_vcpus_pin xl sched-credit Cpupool Pool-0: tslice=30ms ratelimit=1000us Name ID Weight Cap Domain-0 0 1024 0 samael 1 256 0 satan 2 512 0 amon 3 256 0 leviathan 4 512 0 echo 1048576 > /proc/sys/net/ipv4/tcp_limit_output_bytes on dom0 the both domU's satan and leviathan are very IO performance oriented (NFS server and downloading vm) Is there something more what I can do or try? could it be a selinux issue? I have it in permissive mode there, not disabled. But permissive means only to collect the info not enforcing the rules... (I use the xen45 pkgs) -- ------ Greetz

2 3

2025

2024

2023

2022

2021

2020

2019

2018

2017

2016

2015

2014

2013

2012

2011

2010

2009

2008

2007

virt September 2015