[Date Prev][Date Next][Thread Prev][Thread Next][Date Index][Thread Index]

Re: [Xen-devel] Phoronix Xen vs KVM vs Virtualbox benchmark with Ubuntu 11.10



I was able to load Oneiric as PV on HVM at Xen 4.1.2 Oneiric Dom0 (3.1.0-030100-generic)  after rebuilding the recent 3.1.0-3 Ubuntu's kernel with
CONFIG_XEN_PLATFORM_PCI=y , uploading debian packages  :-

linux-headers-3.1.0-2_3.1.0-2.3_all.deb
linux-headers-3.1.0-2-i7_3.1.0-2.3_amd64.deb
linux-image-3.1.0-2-i7_3.1.0-2.3_amd64.deb

to regular Oneiric HVM and new kernel install.
Then same HVM was restarted with
xen_platform_pci=1
and this time it gets loaded ( vs hanging with regular kernel)

I got in dmesg log :-

[    0.000000] Initializing cgroup subsys cpuset
[    0.000000] Initializing cgroup subsys cpu
[    0.000000] Linux version 3.1.0-2-i7 (root@boris-System-P5Q3) (gcc version 4.6.1 (Ubuntu/Linaro 4.6.1-9ubuntu3) ) #3 SMP Wed Nov 2 15:15:13 MSK 2011 (Ubuntu 3.1.0-2.3-i7 3.1.0)
[    0.000000] Command line: BOOT_IMAGE=/boot/vmlinuz-3.1.0-2-i7 root=UUID=bff48eed-20df-4abf-b788-935c75ea6226 ro quiet splash vt.handoff=7
[    0.000000] KERNEL supported cpus:
[    0.000000]   Intel GenuineIntel
[    0.000000]   AMD AuthenticAMD
[    0.000000]   Centaur CentaurHauls
[    0.000000] BIOS-provided physical RAM map:
[    0.000000]  BIOS-e820: 0000000000000000 - 000000000009e000 (usable)
[    0.000000]  BIOS-e820: 000000000009e000 - 00000000000a0000 (reserved)
[    0.000000]  BIOS-e820: 00000000000e0000 - 0000000000100000 (reserved)
[    0.000000]  BIOS-e820: 0000000000100000 - 000000003f800000 (usable)
[    0.000000]  BIOS-e820: 00000000fc000000 - 0000000100000000 (reserved)
[    0.000000] NX (Execute Disable) protection: active
[    0.000000] DMI 2.4 present.
[    0.000000] DMI: Xen HVM domU, BIOS 4.1.2 10/20/2011
[    0.000000] Hypervisor detected: Xen HVM
[    0.000000] Xen version 4.1.
[    0.000000] Xen Platform PCI: I/O protocol version 1
[    0.000000] Netfront and the Xen platform PCI driver have been compiled for this kernel: unplug emulated NICs.
[    0.000000] Blkfront and the Xen platform PCI driver have been compiled for this kernel: unplug emulated disks.
[    0.000000] You might have to change the root device
[    0.000000] from /dev/hd[a-d] to /dev/xvd[a-d]
[    0.000000] in your root= kernel command line option
[    0.000000] e820 update range: 0000000000000000 - 0000000000010000 (usable) ==> (reserved)
[    0.000000] e820 remove range: 00000000000a0000 - 0000000000100000 (usable)
[    0.000000] No AGP bridge found
[    0.000000] last_pfn = 0x3f800 max_arch_pfn = 0x400000000
[    0.000000] MTRR default type: write-back
[    0.000000] MTRR fixed ranges enabled:
[    0.000000]   00000-9FFFF write-back
[    0.000000]   A0000-BFFFF write-combining
[    0.000000]   C0000-FFFFF write-back
. . . . .

Boris.

--- On Tue, 11/1/11, Ian Campbell <Ian.Campbell@xxxxxxxxxx> wrote:

From: Ian Campbell <Ian.Campbell@xxxxxxxxxx>
Subject: Re: [Xen-devel] Phoronix Xen vs KVM vs Virtualbox benchmark with Ubuntu 11.10
To: "Pasi Kärkkäinen" <pasik@xxxxxx>
Cc: "xen-devel@xxxxxxxxxxxxxxxxxxx" <xen-devel@xxxxxxxxxxxxxxxxxxx>, "dokter@xxxxxxxxxxxxx" <dokter@xxxxxxxxxxxxx>
Date: Tuesday, November 1, 2011, 7:00 PM

On Tue, 2011-11-01 at 09:26 -0400, Pasi Kärkkäinen wrote:
> On Tue, Nov 01, 2011 at 08:57:14PM +0900, Ian Campbell wrote:
> > On Tue, 2011-11-01 at 07:31 -0400, Mark Dokter wrote:
> > > In that test, the KVM machine uses virtio disk and nic devices, while
> > > the Xen machine uses the qemu emulated devices. Could that be the cause
> > > for the poor performance?
> >
> > Very much so. Without using the PVHVM drivers in the Xen case this is
> > very much an apples to oranges comparison.
> >
>
> Actually the first page of the article says:
> "The only Xen issue encountered when testing it with an Ubuntu 11.10
> guest and host was the need for manually loading the xen-blkfront
> driver for disk support."
>
> So it sounds like they actually did use PVHVM drivers..

Perhaps, the table of configuration details doesn't make it particularly
obvious (and as Alex points out they may not have been running what they
thought). Also as Alex mentioned it's not obvious which backend they
used in each case, e.g. the qemu based disk backend is not known to be
all that great.

It'd definitely be worth someone having a go at repeating even one of
the benchmarks where we did badly, to rule out these sorts of slip ups.

Ian.
>
> -- Pasi
>



_______________________________________________
Xen-devel mailing list
Xen-devel@xxxxxxxxxxxxxxxxxxx
http://lists.xensource.com/xen-devel

Attachment: dmesg-PV-on-HVM-oneiric.log
Description: Text Data

_______________________________________________
Xen-devel mailing list
Xen-devel@xxxxxxxxxxxxxxxxxxx
http://lists.xensource.com/xen-devel

 


Rackspace

Lists.xenproject.org is hosted with RackSpace, monitoring our
servers 24x7x365 and backed by RackSpace's Fanatical Support®.