I am trying to work out whether it is practical to propose Dom0 xen support as a feature for Fedora 15.
The kernel situation is that Domain 0 has been accepted upstream for 2.6.37. Assuming a 3 month kernel release cycle, F15 will most likely ship with a 2.6.37.x kernel, with 2.6.38 coming out either after the F15 release or just before but too late to be included. If the plan to get key xen drivers into 2.6.38 succeeds, then F15 may be become usable as a Domain 0 system at some point during its lifetime as the kernel package in a Fedora version typically has one major update.
If the kernel team accept backported patches then it might just be possible to ship F15 with usable Domain 0 support but the timescale for that would be very tight.
The other thing we would need to consider is what needs to be done to make xen friendly enough to be usable by an ordinary user. The page https://fedoraproject.org/wiki/Features/XenPvopsDom0 contains plans from when dom0 xen support was expected to make a quick return to Fedora, but they are a couple of years old now so probably need updating.
I think as a minimum we would need a way to add a dom0 enabled grub entry for a kernel, rather than requiring the user to hand edit the grub file. We should also make sure that xen works with the other Fedora virtualisation tools.
What do others think about this? For example is it achievable as a feature, is it too early and better to wait for F16, and what else should we aim to do to make xen usable in Fedora?
Michael Young
On 10-11-07 03:55 PM, M A Young wrote:
I am trying to work out whether it is practical to propose Dom0 xen support as a feature for Fedora 15.
The kernel situation is that Domain 0 has been accepted upstream for 2.6.37. Assuming a 3 month kernel release cycle, F15 will most likely ship with a 2.6.37.x kernel, with 2.6.38 coming out either after the F15 release or just before but too late to be included. If the plan to get key xen drivers into 2.6.38 succeeds, then F15 may be become usable as a Domain 0 system at some point during its lifetime as the kernel package in a Fedora version typically has one major update.
If the kernel team accept backported patches then it might just be possible to ship F15 with usable Domain 0 support but the timescale for that would be very tight.
The other thing we would need to consider is what needs to be done to make xen friendly enough to be usable by an ordinary user. The page https://fedoraproject.org/wiki/Features/XenPvopsDom0 contains plans from when dom0 xen support was expected to make a quick return to Fedora, but they are a couple of years old now so probably need updating.
I think as a minimum we would need a way to add a dom0 enabled grub entry for a kernel, rather than requiring the user to hand edit the grub file. We should also make sure that xen works with the other Fedora virtualisation tools.
What do others think about this? For example is it achievable as a feature, is it too early and better to wait for F16, and what else should we aim to do to make xen usable in Fedora?
Michael Young
I wish I was more technically adept to make an argument in favour of dom0 in F15 natively. As a user though, I would be absolutely thrilled to see this happen.
I've been working on tutorials meant to introduce Xen's use in Red Hat clusters, and I would be more that happy to donate or adapt it to help the cause.
I am trying to work out whether it is practical to propose Dom0 xen support as a feature for Fedora 15.
The kernel situation is that Domain 0 has been accepted upstream for 2.6.37. Assuming a 3 month kernel release cycle, F15 will most likely ship with a 2.6.37.x kernel, with 2.6.38 coming out either after the F15 release or just before but too late to be included. If the plan to get key xen drivers into 2.6.38 succeeds, then F15 may be become usable as a Domain 0 system at some point during its lifetime as the kernel package in a Fedora version typically has one major update.
If the kernel team accept backported patches then it might just be possible to ship F15 with usable Domain 0 support but the timescale for that would be very tight.
The other thing we would need to consider is what needs to be done to make xen friendly enough to be usable by an ordinary user. The page https://fedoraproject.org/wiki/Features/XenPvopsDom0 contains plans from when dom0 xen support was expected to make a quick return to Fedora, but they are a couple of years old now so probably need updating.
I think as a minimum we would need a way to add a dom0 enabled grub entry for a kernel, rather than requiring the user to hand edit the grub file. We should also make sure that xen works with the other Fedora virtualisation tools.
What do others think about this? For example is it achievable as a feature, is it too early and better to wait for F16, and what else should we aim to do to make xen usable in Fedora?
My personal opinion is that we should make an aggressive attempt to get Domain 0 in Fedora 15. As Michael Young points out, there is presently some uncertainty as to when the Xen drivers will be upstream. They may make the 2.6.37 series and therefore Fedora 15 or they may not. Regardless, we will have to touch a lot of subsystems (e.g. grubby, Red Hat Bugzilla #640486), so it makes sense to start early.
If the Xen drivers don't make Fedora 15, then little is lost. We can continue to test the Domain 0 kernel (without Domain U) and work on the supporting infrastructure. It would be useful for our effort to have a working Domain 0, grubby, etc in Fedora 15 even if other features do not make it. In the worst case, we may delay the announcement of the Domain 0 feature to Fedora 16, but can include all progress completed in Fedora 15. There is no shame in a delayed feature -- this happened to systemd during the preparation for Fedora 14.
Hi,
If the kernel team accept backported patches then it might just be possible to ship F15 with usable Domain 0 support but the timescale for that would be very tight.
Talk to the kernel people. I think the chances to get the backend drivers into the rawhide kernel even before the 2.6.38 merge window are not that bad. Are they in linux-next already?
The situation is quite different from the 2.6.2x times when dom0 support was dropped from the fedora kernels. The patches are pretty self-contained now, i.e. they touch xen bits only instead of sprinkling all over arch/x86/. Also the Xen people have reasonable and realistic plan to get the stuff upstream soon.
cheers, Gerd
On Mon, 8 Nov 2010, Gerd Hoffmann wrote:
Talk to the kernel people. I think the chances to get the backend drivers into the rawhide kernel even before the 2.6.38 merge window are not that bad. Are they in linux-next already?
Not yet. The patches I have seen so far are mostly bug fixes and updates for things already in 2.6.37 though I think work is about to start on porting the drivers to 2.6.37.
Michael Young
On Sun, Nov 07, 2010 at 08:55:13PM +0000, M A Young wrote:
I am trying to work out whether it is practical to propose Dom0 xen support as a feature for Fedora 15.
The kernel situation is that Domain 0 has been accepted upstream for 2.6.37. Assuming a 3 month kernel release cycle, F15 will most likely ship with a 2.6.37.x kernel, with 2.6.38 coming out either after the F15 release or just before but too late to be included. If the plan to get key xen drivers into 2.6.38 succeeds, then F15 may be become usable as a Domain 0 system at some point during its lifetime as the kernel package in a Fedora version typically has one major update.
If the kernel team accept backported patches then it might just be possible to ship F15 with usable Domain 0 support but the timescale for that would be very tight.
The other thing we would need to consider is what needs to be done to make xen friendly enough to be usable by an ordinary user. The page https://fedoraproject.org/wiki/Features/XenPvopsDom0 contains plans from when dom0 xen support was expected to make a quick return to Fedora, but they are a couple of years old now so probably need updating.
I think as a minimum we would need a way to add a dom0 enabled grub entry for a kernel, rather than requiring the user to hand edit the grub file. We should also make sure that xen works with the other Fedora virtualisation tools.
Regardless of whether the kernel gets into F15, the work to improve userspace integration, in particular grub, would be useful sooner or later, so there's no need to delay that. Likewise testing of all the virt tools can be done any time against your custom kernel builds in anticipation of it being officially supported in fedora kernels.
Regards, Daniel
On Sun, Nov 07, 2010 at 08:55:13PM +0000, M A Young wrote:
I think as a minimum we would need a way to add a dom0 enabled grub entry for a kernel, rather than requiring the user to hand edit the grub file. We should also make sure that xen works with the other Fedora virtualisation tools.
We should be pretty good here. And if there have been bugs or regressions, then it would be a priority to fix them. The whole aim of libvirt and virt-tools is to be hypervisor independent after all.
Rich.
M A Young wrote:
I am trying to work out whether it is practical to propose Dom0 xen support as a feature for Fedora 15.
The kernel situation is that Domain 0 has been accepted upstream for 2.6.37. Assuming a 3 month kernel release cycle, F15 will most likely ship with a 2.6.37.x kernel, with 2.6.38 coming out either after the F15 release or just before but too late to be included. If the plan to get key xen drivers into 2.6.38 succeeds, then F15 may be become usable as a Domain 0 system at some point during its lifetime as the kernel package in a Fedora version typically has one major update.
If the kernel team accept backported patches then it might just be possible to ship F15 with usable Domain 0 support but the timescale for that would be very tight.
The other thing we would need to consider is what needs to be done to make xen friendly enough to be usable by an ordinary user. The page https://fedoraproject.org/wiki/Features/XenPvopsDom0 contains plans from when dom0 xen support was expected to make a quick return to Fedora, but they are a couple of years old now so probably need updating.
I think as a minimum we would need a way to add a dom0 enabled grub entry for a kernel, rather than requiring the user to hand edit the grub file. We should also make sure that xen works with the other Fedora virtualisation tools.
What do others think about this? For example is it achievable as a feature, is it too early and better to wait for F16, and what else should we aim to do to make xen usable in Fedora?
Michael Young _______________________________________________ virt mailing list virt@lists.fedoraproject.org https://admin.fedoraproject.org/mailman/listinfo/virt
IMO, the pieces that the Xen folks have to whip into shape are the tools; the missing kernel pieces are fairly well known (backend drivers), although Xen folks are considering blkback in userspace instead of kernel space, so that may be a new twist.
So, getting Xen support in the following 2 userspace tools is critical: qemu & libvirt; the latter means a new interface to xen libxc (dropping xend), so xen fixes can be tracked to latest xen upstream. I've seen a fair amt of activity to add xen support to upstream qemu; I haven't been following libvirt devel to see if xen support to libxc has been active or not (but would like to hear if so).
- Don
On Mon, 8 Nov 2010, Don Dutile wrote:
IMO, the pieces that the Xen folks have to whip into shape are the tools; the missing kernel pieces are fairly well known (backend drivers), although Xen folks are considering blkback in userspace instead of kernel space, so that may be a new twist.
I did see a post which mentioned that qemu had blkback and netback and wondering if they would work with xen, but I don't know any more than that.
So, getting Xen support in the following 2 userspace tools is critical: qemu & libvirt; the latter means a new interface to xen libxc (dropping xend), so xen fixes can be tracked to latest xen upstream. I've seen a fair amt of activity to add xen support to upstream qemu; I haven't been following libvirt devel to see if xen support to libxc has been active or not (but would like to hear if so).
I think the line now is that xl/libxenlight is the way forward, eg. see http://lists.xensource.com/archives/html/xen-devel/2010-11/msg00352.html Both the current xen (4.0.1) and the next (4.1, which might be the current version when F15 is released) will still have xend but it will deprecated in 4.1.
Michael Young
On 11/09/10 20:44, M A Young wrote:
On Mon, 8 Nov 2010, Don Dutile wrote:
IMO, the pieces that the Xen folks have to whip into shape are the tools; the missing kernel pieces are fairly well known (backend drivers), although Xen folks are considering blkback in userspace instead of kernel space, so that may be a new twist.
I did see a post which mentioned that qemu had blkback and netback and wondering if they would work with xen, but I don't know any more than that.
Last time I tested they worked fine. blkback should even provide reasonable performance. netback most likely will be outperformed by far by the kernel's netback implementation.
cheers, Gerd
On 11/08/2010 04:55 AM, M A Young wrote:
I am trying to work out whether it is practical to propose Dom0 xen support as a feature for Fedora 15.
The kernel situation is that Domain 0 has been accepted upstream for 2.6.37. Assuming a 3 month kernel release cycle, F15 will most likely ship with a 2.6.37.x kernel, with 2.6.38 coming out either after the F15 release or just before but too late to be included. If the plan to get key xen drivers into 2.6.38 succeeds, then F15 may be become usable as a Domain 0 system at some point during its lifetime as the kernel package in a Fedora version typically has one major update.
If the kernel team accept backported patches then it might just be possible to ship F15 with usable Domain 0 support but the timescale for that would be very tight.
The other thing we would need to consider is what needs to be done to make xen friendly enough to be usable by an ordinary user. The page https://fedoraproject.org/wiki/Features/XenPvopsDom0 contains plans from when dom0 xen support was expected to make a quick return to Fedora, but they are a couple of years old now so probably need updating.
I think as a minimum we would need a way to add a dom0 enabled grub entry for a kernel, rather than requiring the user to hand edit the grub file. We should also make sure that xen works with the other Fedora virtualisation tools.
What do others think about this? For example is it achievable as a feature, is it too early and better to wait for F16, and what else should we aim to do to make xen usable in Fedora?
Have you consider kvm? it's upstream since 2.6.20 and now its more ready than ever.
Michael Young _______________________________________________ virt mailing list virt@lists.fedoraproject.org https://admin.fedoraproject.org/mailman/listinfo/virt
Dor Laor wrote:
On 11/08/2010 04:55 AM, M A Young wrote:
I am trying to work out whether it is practical to propose Dom0 xen support as a feature for Fedora 15.
The kernel situation is that Domain 0 has been accepted upstream for 2.6.37. Assuming a 3 month kernel release cycle, F15 will most likely ship with a 2.6.37.x kernel, with 2.6.38 coming out either after the F15 release or just before but too late to be included. If the plan to get key xen drivers into 2.6.38 succeeds, then F15 may be become usable as a Domain 0 system at some point during its lifetime as the kernel package in a Fedora version typically has one major update.
If the kernel team accept backported patches then it might just be possible to ship F15 with usable Domain 0 support but the timescale for that would be very tight.
The other thing we would need to consider is what needs to be done to make xen friendly enough to be usable by an ordinary user. The page https://fedoraproject.org/wiki/Features/XenPvopsDom0 contains plans from when dom0 xen support was expected to make a quick return to Fedora, but they are a couple of years old now so probably need updating.
I think as a minimum we would need a way to add a dom0 enabled grub entry for a kernel, rather than requiring the user to hand edit the grub file. We should also make sure that xen works with the other Fedora virtualisation tools.
What do others think about this? For example is it achievable as a feature, is it too early and better to wait for F16, and what else should we aim to do to make xen usable in Fedora?
Have you consider kvm? it's upstream since 2.6.20 and now its more ready than ever.
There are some good tutorials which should explain the difference between xen and kvm, particularly the performance and hardware requirements of each.
----- Original Message -----
From: "Bill Davidsen" davidsen@tmr.com To: dlaor@redhat.com Cc: xen@lists.fedoraproject.org, virt@lists.fedoraproject.org, "M A Young" m.a.young@durham.ac.uk Sent: Monday, November 8, 2010 11:52:08 AM Subject: Re: [fedora-virt] [Fedora-xen] Dom0 xen support in Fedora 15? Dor Laor wrote:
On 11/08/2010 04:55 AM, M A Young wrote:
I am trying to work out whether it is practical to propose Dom0 xen support as a feature for Fedora 15.
The kernel situation is that Domain 0 has been accepted upstream for 2.6.37. Assuming a 3 month kernel release cycle, F15 will most likely ship with a 2.6.37.x kernel, with 2.6.38 coming out either after the F15 release or just before but too late to be included. If the plan to get key xen drivers into 2.6.38 succeeds, then F15 may be become usable as a Domain 0 system at some point during its lifetime as the kernel package in a Fedora version typically has one major update.
If the kernel team accept backported patches then it might just be possible to ship F15 with usable Domain 0 support but the timescale for that would be very tight.
The other thing we would need to consider is what needs to be done to make xen friendly enough to be usable by an ordinary user. The page https://fedoraproject.org/wiki/Features/XenPvopsDom0 contains plans from when dom0 xen support was expected to make a quick return to Fedora, but they are a couple of years old now so probably need updating.
I think as a minimum we would need a way to add a dom0 enabled grub entry for a kernel, rather than requiring the user to hand edit the grub file. We should also make sure that xen works with the other Fedora virtualisation tools.
What do others think about this? For example is it achievable as a feature, is it too early and better to wait for F16, and what else should we aim to do to make xen usable in Fedora?
Have you consider kvm? it's upstream since 2.6.20 and now its more ready than ever.
There are some good tutorials which should explain the difference between xen and kvm, particularly the performance and hardware requirements of each.
re: hardware requirements, KVM's requirement for VT-X/AMD-V extensions certainly used to be a concern 2-3 years ago but today even laptops come with this support. And regarding performance they days of Xen outperforming KVM have long-since passed.
-- Bill Davidsendavidsen@tmr.com "We can't solve today's problems by using the same thinking we used in creating them." - Einstein
virt mailing list virt@lists.fedoraproject.org https://admin.fedoraproject.org/mailman/listinfo/virt
On 11/08/2010 06:02 PM, Andrew Cathrow wrote:
----- Original Message -----
From: "Bill Davidsen"davidsen@tmr.com To: dlaor@redhat.com Cc: xen@lists.fedoraproject.org, virt@lists.fedoraproject.org, "M A Young"m.a.young@durham.ac.uk Sent: Monday, November 8, 2010 11:52:08 AM Subject: Re: [fedora-virt] [Fedora-xen] Dom0 xen support in Fedora 15? Dor Laor wrote:
On 11/08/2010 04:55 AM, M A Young wrote:
I am trying to work out whether it is practical to propose Dom0 xen support as a feature for Fedora 15.
The kernel situation is that Domain 0 has been accepted upstream for 2.6.37. Assuming a 3 month kernel release cycle, F15 will most likely ship with a 2.6.37.x kernel, with 2.6.38 coming out either after the F15 release or just before but too late to be included. If the plan to get key xen drivers into 2.6.38 succeeds, then F15 may be become usable as a Domain 0 system at some point during its lifetime as the kernel package in a Fedora version typically has one major update.
If the kernel team accept backported patches then it might just be possible to ship F15 with usable Domain 0 support but the timescale for that would be very tight.
The other thing we would need to consider is what needs to be done to make xen friendly enough to be usable by an ordinary user. The page https://fedoraproject.org/wiki/Features/XenPvopsDom0 contains plans from when dom0 xen support was expected to make a quick return to Fedora, but they are a couple of years old now so probably need updating.
I think as a minimum we would need a way to add a dom0 enabled grub entry for a kernel, rather than requiring the user to hand edit the grub file. We should also make sure that xen works with the other Fedora virtualisation tools.
What do others think about this? For example is it achievable as a feature, is it too early and better to wait for F16, and what else should we aim to do to make xen usable in Fedora?
Have you consider kvm? it's upstream since 2.6.20 and now its more ready than ever.
There are some good tutorials which should explain the difference between xen and kvm, particularly the performance and hardware requirements of each.
re: hardware requirements, KVM's requirement for VT-X/AMD-V extensions certainly used to be a concern 2-3 years ago but today even laptops come with this support. And regarding performance they days of Xen outperforming KVM have long-since passed.
Citations needed. I'm not saying what you claim isn't true but without data this opinion doesn't carry much weight.
Regards, Dennis
----- Original Message -----
From: "Dennis Jacobfeuerborn" dennisml@conversis.de To: "Andrew Cathrow" acathrow@redhat.com Cc: "Bill Davidsen" davidsen@tmr.com, xen@lists.fedoraproject.org, virt@lists.fedoraproject.org, "M A Young" m.a.young@durham.ac.uk Sent: Monday, November 8, 2010 1:59:54 PM Subject: Re: [fedora-virt] [Fedora-xen] Dom0 xen support in Fedora 15? On 11/08/2010 06:02 PM, Andrew Cathrow wrote:
----- Original Message -----
From: "Bill Davidsen"davidsen@tmr.com To: dlaor@redhat.com Cc: xen@lists.fedoraproject.org, virt@lists.fedoraproject.org, "M A Young"m.a.young@durham.ac.uk Sent: Monday, November 8, 2010 11:52:08 AM Subject: Re: [fedora-virt] [Fedora-xen] Dom0 xen support in Fedora 15? Dor Laor wrote:
On 11/08/2010 04:55 AM, M A Young wrote:
I am trying to work out whether it is practical to propose Dom0 xen support as a feature for Fedora 15.
The kernel situation is that Domain 0 has been accepted upstream for 2.6.37. Assuming a 3 month kernel release cycle, F15 will most likely ship with a 2.6.37.x kernel, with 2.6.38 coming out either after the F15 release or just before but too late to be included. If the plan to get key xen drivers into 2.6.38 succeeds, then F15 may be become usable as a Domain 0 system at some point during its lifetime as the kernel package in a Fedora version typically has one major update.
If the kernel team accept backported patches then it might just be possible to ship F15 with usable Domain 0 support but the timescale for that would be very tight.
The other thing we would need to consider is what needs to be done to make xen friendly enough to be usable by an ordinary user. The page https://fedoraproject.org/wiki/Features/XenPvopsDom0 contains plans from when dom0 xen support was expected to make a quick return to Fedora, but they are a couple of years old now so probably need updating.
I think as a minimum we would need a way to add a dom0 enabled grub entry for a kernel, rather than requiring the user to hand edit the grub file. We should also make sure that xen works with the other Fedora virtualisation tools.
What do others think about this? For example is it achievable as a feature, is it too early and better to wait for F16, and what else should we aim to do to make xen usable in Fedora?
Have you consider kvm? it's upstream since 2.6.20 and now its more ready than ever.
There are some good tutorials which should explain the difference between xen and kvm, particularly the performance and hardware requirements of each.
re: hardware requirements, KVM's requirement for VT-X/AMD-V extensions certainly used to be a concern 2-3 years ago but today even laptops come with this support. And regarding performance they days of Xen outperforming KVM have long-since passed.
Citations needed. I'm not saying what you claim isn't true but without data this opinion doesn't carry much weight.
Citations are really needed on both sides of the debate, 2 or 3 year old metrics no longer apply. Vendors published benchmarks are typically questionable, they focus on their products strengths and their competitors weakness. The only hope for a fair comparison is a vendor neutral set of benchmarks such as SPECvirt http://www.spec.org/virt_sc2010/ But obviously this isn't a simple test to run.
Regards, Dennis
Not to mention it costs $3000 ....
On Mon, Nov 8, 2010 at 2:29 PM, Andrew Cathrow acathrow@redhat.com wrote:
----- Original Message -----
From: "Dennis Jacobfeuerborn" dennisml@conversis.de To: "Andrew Cathrow" acathrow@redhat.com Cc: "Bill Davidsen" davidsen@tmr.com, xen@lists.fedoraproject.org,
virt@lists.fedoraproject.org, "M A Young"
m.a.young@durham.ac.uk Sent: Monday, November 8, 2010 1:59:54 PM Subject: Re: [fedora-virt] [Fedora-xen] Dom0 xen support in Fedora 15? On 11/08/2010 06:02 PM, Andrew Cathrow wrote:
----- Original Message -----
From: "Bill Davidsen"davidsen@tmr.com To: dlaor@redhat.com Cc: xen@lists.fedoraproject.org, virt@lists.fedoraproject.org, "M A Young"m.a.young@durham.ac.uk Sent: Monday, November 8, 2010 11:52:08 AM Subject: Re: [fedora-virt] [Fedora-xen] Dom0 xen support in Fedora 15? Dor Laor wrote:
On 11/08/2010 04:55 AM, M A Young wrote:
I am trying to work out whether it is practical to propose Dom0 xen support as a feature for Fedora 15.
The kernel situation is that Domain 0 has been accepted upstream for 2.6.37. Assuming a 3 month kernel release cycle, F15 will most likely ship with a 2.6.37.x kernel, with 2.6.38 coming out either after the F15 release or just before but too late to be included. If the plan to get key xen drivers into 2.6.38 succeeds, then F15 may be become usable as a Domain 0 system at some point during its lifetime as the kernel package in a Fedora version typically has one major update.
If the kernel team accept backported patches then it might just be possible to ship F15 with usable Domain 0 support but the timescale for that would be very tight.
The other thing we would need to consider is what needs to be done to make xen friendly enough to be usable by an ordinary user. The page https://fedoraproject.org/wiki/Features/XenPvopsDom0 contains plans from when dom0 xen support was expected to make a quick return to Fedora, but they are a couple of years old now so probably need updating.
I think as a minimum we would need a way to add a dom0 enabled grub entry for a kernel, rather than requiring the user to hand edit the grub file. We should also make sure that xen works with the other Fedora virtualisation tools.
What do others think about this? For example is it achievable as a feature, is it too early and better to wait for F16, and what else should we aim to do to make xen usable in Fedora?
Have you consider kvm? it's upstream since 2.6.20 and now its more ready than ever.
There are some good tutorials which should explain the difference between xen and kvm, particularly the performance and hardware requirements of each.
re: hardware requirements, KVM's requirement for VT-X/AMD-V extensions certainly used to be a concern 2-3 years ago but today even laptops come with this support. And regarding performance they days of Xen outperforming KVM have long-since passed.
Citations needed. I'm not saying what you claim isn't true but without data this opinion doesn't carry much weight.
Citations are really needed on both sides of the debate, 2 or 3 year old metrics no longer apply. Vendors published benchmarks are typically questionable, they focus on their products strengths and their competitors weakness. The only hope for a fair comparison is a vendor neutral set of benchmarks such as SPECvirt http://www.spec.org/virt_sc2010/ But obviously this isn't a simple test to run.
Regards, Dennis
-- xen mailing list xen@lists.fedoraproject.org https://admin.fedoraproject.org/mailman/listinfo/xen
On 11/08/2010 08:29 PM, Andrew Cathrow wrote:
----- Original Message -----
From: "Dennis Jacobfeuerborn"dennisml@conversis.de To: "Andrew Cathrow"acathrow@redhat.com Cc: "Bill Davidsen"davidsen@tmr.com, xen@lists.fedoraproject.org, virt@lists.fedoraproject.org, "M A Young" m.a.young@durham.ac.uk Sent: Monday, November 8, 2010 1:59:54 PM Subject: Re: [fedora-virt] [Fedora-xen] Dom0 xen support in Fedora 15? On 11/08/2010 06:02 PM, Andrew Cathrow wrote:
----- Original Message -----
From: "Bill Davidsen"davidsen@tmr.com To: dlaor@redhat.com Cc: xen@lists.fedoraproject.org, virt@lists.fedoraproject.org, "M A Young"m.a.young@durham.ac.uk Sent: Monday, November 8, 2010 11:52:08 AM Subject: Re: [fedora-virt] [Fedora-xen] Dom0 xen support in Fedora 15? Dor Laor wrote:
On 11/08/2010 04:55 AM, M A Young wrote:
I am trying to work out whether it is practical to propose Dom0 xen support as a feature for Fedora 15.
The kernel situation is that Domain 0 has been accepted upstream for 2.6.37. Assuming a 3 month kernel release cycle, F15 will most likely ship with a 2.6.37.x kernel, with 2.6.38 coming out either after the F15 release or just before but too late to be included. If the plan to get key xen drivers into 2.6.38 succeeds, then F15 may be become usable as a Domain 0 system at some point during its lifetime as the kernel package in a Fedora version typically has one major update.
If the kernel team accept backported patches then it might just be possible to ship F15 with usable Domain 0 support but the timescale for that would be very tight.
The other thing we would need to consider is what needs to be done to make xen friendly enough to be usable by an ordinary user. The page https://fedoraproject.org/wiki/Features/XenPvopsDom0 contains plans from when dom0 xen support was expected to make a quick return to Fedora, but they are a couple of years old now so probably need updating.
I think as a minimum we would need a way to add a dom0 enabled grub entry for a kernel, rather than requiring the user to hand edit the grub file. We should also make sure that xen works with the other Fedora virtualisation tools.
What do others think about this? For example is it achievable as a feature, is it too early and better to wait for F16, and what else should we aim to do to make xen usable in Fedora?
Have you consider kvm? it's upstream since 2.6.20 and now its more ready than ever.
There are some good tutorials which should explain the difference between xen and kvm, particularly the performance and hardware requirements of each.
re: hardware requirements, KVM's requirement for VT-X/AMD-V extensions certainly used to be a concern 2-3 years ago but today even laptops come with this support. And regarding performance they days of Xen outperforming KVM have long-since passed.
Citations needed. I'm not saying what you claim isn't true but without data this opinion doesn't carry much weight.
Citations are really needed on both sides of the debate, 2 or 3 year old metrics no longer apply.
Which is why I hoped I could nudge you into providing some data if you had any. :) There seems to be very little recent and hard data on the subject out there.
Vendors published benchmarks are typically questionable, they focus on their products strengths and their competitors weakness. The only hope for a fair comparison is a vendor neutral set of benchmarks such as SPECvirt http://www.spec.org/virt_sc2010/ But obviously this isn't a simple test to run.
AFAIK there are some issues with that benchmark not being free.
What I'm basically looking for is a simple comparison of a few key metrics between common setups. For example a simple bonnie++ run on XenPV and KVM+virtio_blk could already give people at least some baseline numbers.
Regards, Dennis
On 10-11-09 11:49 AM, Dennis Jacobfeuerborn wrote:
AFAIK there are some issues with that benchmark not being free.
What I'm basically looking for is a simple comparison of a few key metrics between common setups. For example a simple bonnie++ run on XenPV and KVM+virtio_blk could already give people at least some baseline numbers.
Regards, Dennis
I'd be happy to run some benchmarks. I've got some identical (if humble) machines... Would you want just a Xen vs. KVM? If so, what parameters or setup options would provide the best apples to apples comparison?
The machines are; - quad core athlon II x4 - 4gb ddr3 - 1x 7200rom 500GB drive - Fedora 14, stripped (but identical) installs
If these are too low end for useful results, let me know. Otherwise, I'll fire up Xen under Michael's 2.6.32-25-172 kernel with the stock 4.0.1 hypervisor on one machine, and can use the stock KVM/QEMU setup for another.
Assuming all is fine, would it matter whether the VMs were F14 vs CentOS 5.5?
I was thinking 3x runs each of;
- bonnie++ - Recompile Michael's kernel - ?? Please make suggestions
Quoting Digimer linux@alteeve.com:
On 10-11-09 11:49 AM, Dennis Jacobfeuerborn wrote:
AFAIK there are some issues with that benchmark not being free.
What I'm basically looking for is a simple comparison of a few key metrics between common setups. For example a simple bonnie++ run on XenPV and KVM+virtio_blk could already give people at least some baseline numbers.
Regards, Dennis
I'd be happy to run some benchmarks. I've got some identical (if humble) machines... Would you want just a Xen vs. KVM? If so, what parameters or setup options would provide the best apples to apples comparison?
The machines are;
- quad core athlon II x4
- 4gb ddr3
- 1x 7200rom 500GB drive
- Fedora 14, stripped (but identical) installs
If these are too low end for useful results, let me know. Otherwise, I'll fire up Xen under Michael's 2.6.32-25-172 kernel with the stock 4.0.1 hypervisor on one machine, and can use the stock KVM/QEMU setup for another.
Assuming all is fine, would it matter whether the VMs were F14 vs CentOS 5.5?
I was thinking 3x runs each of;
- bonnie++
- Recompile Michael's kernel
- ?? Please make suggestions
Digimer,
I would think the latest xen vs the latest kvm would be best. Fully virtualized vs Para-virtualized, is that possible with kvm? I don't believe your specs are too low end, they actually match pretty close to what I have and I don't think the OS matters that much as long as they are identical setups.
Thanks for offering to do the comparison, I am very interested in your results.
Jon
On 10-11-09 12:54 PM, jonr@destar.net wrote:
Digimer,
I would think the latest xen vs the latest kvm would be best. Fully virtualized vs Para-virtualized, is that possible with kvm? I don't believe your specs are too low end, they actually match pretty close to what I have and I don't think the OS matters that much as long as they are identical setups.
Thanks for offering to do the comparison, I am very interested in your results.
Jon
I will look into whether fully virtualized is an option for KVM. If not, I probably won't worry as all "modern" operating systems support paravirtualized setups.
I've got no way of knowing how to run tests on non-linux VMs. If anyone can give advice on benchmarking non-Linux OS', I'd be grateful and will give it a go.
Can you (or anyone) suggests tests to run beyond bonnie++ and a kernel compile that would be "real-world" and useful benchmarks? I think one test that will help would be to run tests concurrently on two VMs to see if there is a difference in how Xen or KVM handle very random disk I/O.
PS - I will likely need a week or so to get these tests done.
Quoting Digimer linux@alteeve.com:
On 10-11-09 12:54 PM, jonr@destar.net wrote:
Digimer,
I would think the latest xen vs the latest kvm would be best. Fully virtualized vs Para-virtualized, is that possible with kvm? I don't believe your specs are too low end, they actually match pretty close to what I have and I don't think the OS matters that much as long as they are identical setups.
Thanks for offering to do the comparison, I am very interested in your results.
Jon
I will look into whether fully virtualized is an option for KVM. If not, I probably won't worry as all "modern" operating systems support paravirtualized setups.
I've got no way of knowing how to run tests on non-linux VMs. If anyone can give advice on benchmarking non-Linux OS', I'd be grateful and will give it a go.
Can you (or anyone) suggests tests to run beyond bonnie++ and a kernel compile that would be "real-world" and useful benchmarks? I think one test that will help would be to run tests concurrently on two VMs to see if there is a difference in how Xen or KVM handle very random disk I/O.
PS - I will likely need a week or so to get these tests done.
-- Digimer
Does bonnie++ do network tests? I would also be interested in how well the network performs. I know this would depend on the NIC's but it would be "good enough" for me. :)
Thanks again and I am willing to wait a week or two or more if you need it.
Jon
On Tue, Nov 9, 2010 at 4:01 PM, Digimer linux@alteeve.com wrote:
On 10-11-09 12:54 PM, jonr@destar.net wrote:
Digimer,
I would think the latest xen vs the latest kvm would be best. Fully virtualized vs Para-virtualized, is that possible with kvm? I don't believe your specs are too low end, they actually match pretty close to what I have and I don't think the OS matters that much as long as they are identical setups.
Thanks for offering to do the comparison, I am very interested in your results.
Jon
I will look into whether fully virtualized is an option for KVM. If not, I probably won't worry as all "modern" operating systems support paravirtualized setups.
It is the other way around. KVM only runs on vt/svm hardware, so what we don't have, is a paravirtualization offering (apart from the drivers, clock, etc)
I've got no way of knowing how to run tests on non-linux VMs. If anyone can give advice on benchmarking non-Linux OS', I'd be grateful and will give it a go.
Can you (or anyone) suggests tests to run beyond bonnie++ and a kernel compile that would be "real-world" and useful benchmarks? I think one test that will help would be to run tests concurrently on two VMs to see if there is a difference in how Xen or KVM handle very random disk I/O.
PS - I will likely need a week or so to get these tests done.
-- Digimer E-Mail: digimer@alteeve.com AN!Whitepapers: http://alteeve.com Node Assassin: http://nodeassassin.org -- xen mailing list xen@lists.fedoraproject.org https://admin.fedoraproject.org/mailman/listinfo/xen
Digimer wrote:
On 10-11-09 11:49 AM, Dennis Jacobfeuerborn wrote:
AFAIK there are some issues with that benchmark not being free.
What I'm basically looking for is a simple comparison of a few key metrics between common setups. For example a simple bonnie++ run on XenPV and KVM+virtio_blk could already give people at least some baseline numbers.
Regards, Dennis
I'd be happy to run some benchmarks. I've got some identical (if humble) machines... Would you want just a Xen vs. KVM? If so, what parameters or setup options would provide the best apples to apples comparison?
The machines are;
- quad core athlon II x4
- 4gb ddr3
- 1x 7200rom 500GB drive
- Fedora 14, stripped (but identical) installs
If these are too low end for useful results, let me know. Otherwise, I'll fire up Xen under Michael's 2.6.32-25-172 kernel with the stock 4.0.1 hypervisor on one machine, and can use the stock KVM/QEMU setup for another.
Assuming all is fine, would it matter whether the VMs were F14 vs CentOS 5.5?
I was thinking 3x runs each of;
- bonnie++
- Recompile Michael's kernel
- ?? Please make suggestions
Thanks for the offer. The services which have in the past shown significant differences (my experience) are small, with lots of connects and io. Examples are NNTP (ak usenet news) and POP mail servers. I suspect that DHCP and DNS have too little io to show the effect, but I never had enough load to know. Unless web service has a lot of dynamic content, I doubt it would have enough load to show the difference unless someone has server image and load generator to use. I could probably dig up a simulated load for NNTP, but I don't have an easy way to preload a server. My days of generating test loads were when I ran ISP NNTP servers and only had to provide test loading against the servers.
Performance aside, There is still a lot of old or small hardware that just can't do KVM at all, and which are easier to administer with a VM which can be moved.
On 11/09/2010 10:59 PM, Bill Davidsen wrote:
Digimer wrote:
On 10-11-09 11:49 AM, Dennis Jacobfeuerborn wrote:
AFAIK there are some issues with that benchmark not being free.
What I'm basically looking for is a simple comparison of a few key metrics between common setups. For example a simple bonnie++ run on XenPV and KVM+virtio_blk could already give people at least some baseline numbers.
Regards, Dennis
I'd be happy to run some benchmarks. I've got some identical (if humble) machines... Would you want just a Xen vs. KVM? If so, what parameters or setup options would provide the best apples to apples comparison?
The machines are;
- quad core athlon II x4
- 4gb ddr3
- 1x 7200rom 500GB drive
- Fedora 14, stripped (but identical) installs
If these are too low end for useful results, let me know. Otherwise, I'll fire up Xen under Michael's 2.6.32-25-172 kernel with the stock 4.0.1 hypervisor on one machine, and can use the stock KVM/QEMU setup for another.
Assuming all is fine, would it matter whether the VMs were F14 vs CentOS 5.5?
I was thinking 3x runs each of;
- bonnie++
- Recompile Michael's kernel
- ?? Please make suggestions
Thanks for the offer. The services which have in the past shown significant differences (my experience) are small, with lots of connects and io. Examples are NNTP (ak usenet news) and POP mail servers. I suspect that DHCP and DNS have too little io to show the effect, but I never had enough load to know. Unless web service has a lot of dynamic content, I doubt it would have enough load to show the difference unless someone has server image and load generator to use. I could probably dig up a simulated load for NNTP, but I don't have an easy way to preload a server. My days of generating test loads were when I ran ISP NNTP servers and only had to provide test loading against the servers.
I think these application level benchmarks should be considered second tier/optional because it's difficult to get realistic numbers for these.
When I'm looking for benchmarks I'm usually mostly interested in low-level ones because they are more easy to apply to my current situation. If I run the mysql benchmarking suite then the results may or may not be indicative of what my app is actually doing. In comparison seeing that Hypervisor A has only 50% of the random-i/o performance of Hypervisor B then that can give you a good ballpark figure for all your applications depending on how dependent they are on random-i/o access.
In the networking case I would rather opt for a plain TCP/UDP benchmark that measures throughput and latency (and maybe connections per second) as these numbers are useful for people running any networking service.
As for OS and hardware configuration I think it's more important to get at least one run on each Hypervisor on the same os/hardware combination to make them comparable. I'd obviously prefer RHEL/Centos since that's what a lot of people run on their servers (including me) but I think Fedora would work just as well to at least get performance comparisons of the hypervisors relative to each other.
One important bit would be the parameters for the virtual machines as things like for example the cache configuration can have a great impact on the outcome i/o tests for example.
Performance aside, There is still a lot of old or small hardware that just can't do KVM at all, and which are easier to administer with a VM which can be moved.
But how important is a benchmark in this case? If you want to move them to new hardware then you'd buy a machine with hardware virtualization support anyway and if you can't run the VM on new hardware then you probably should just run it on xen as you don't have a lot of options anyway.
I hope I'll eventually get my hand on a server that doesn't have to go into production the same day it is delivered so I'm certainly interested in participating but right now I could only deliver comparative benchmarks from desktop type systems and my new dual core intel 4gb 7200rpm drive notebook. Only really useful to pitch the hypervisors against each other in relative terms but the more data the better right?
Regards, Dennis
On Tue, Nov 9, 2010 at 12:46 PM, Digimer linux@alteeve.com wrote:
On 10-11-09 11:49 AM, Dennis Jacobfeuerborn wrote:
AFAIK there are some issues with that benchmark not being free.
What I'm basically looking for is a simple comparison of a few key metrics between common setups. For example a simple bonnie++ run on XenPV and KVM+virtio_blk could already give people at least some baseline numbers.
Regards, Dennis
I'd be happy to run some benchmarks. I've got some identical (if humble) machines... Would you want just a Xen vs. KVM? If so, what parameters or setup options would provide the best apples to apples comparison?
The machines are;
- quad core athlon II x4
- 4gb ddr3
- 1x 7200rom 500GB drive
- Fedora 14, stripped (but identical) installs
If these are too low end for useful results, let me know. Otherwise, I'll fire up Xen under Michael's 2.6.32-25-172 kernel with the stock 4.0.1 hypervisor on one machine, and can use the stock KVM/QEMU setup for another.
Assuming all is fine, would it matter whether the VMs were F14 vs CentOS 5.5?
I was thinking 3x runs each of;
- bonnie++
- Recompile Michael's kernel
- ?? Please make suggestions
There is also an ongoing discussion about Xen vs. KVM performance on the CentOS virt list
http://lists.centos.org/pipermail/centos-virt/2010-October/002074.html
On 10-11-12 09:57 AM, Todd Deshane wrote:
I'd be happy to run some benchmarks. I've got some identical (if humble) machines... Would you want just a Xen vs. KVM? If so, what parameters or setup options would provide the best apples to apples comparison?
The machines are;
- quad core athlon II x4
- 4gb ddr3
- 1x 7200rom 500GB drive
- Fedora 14, stripped (but identical) installs
If these are too low end for useful results, let me know. Otherwise, I'll fire up Xen under Michael's 2.6.32-25-172 kernel with the stock 4.0.1 hypervisor on one machine, and can use the stock KVM/QEMU setup for another.
Assuming all is fine, would it matter whether the VMs were F14 vs CentOS 5.5?
I was thinking 3x runs each of;
- bonnie++
- Recompile Michael's kernel
- ?? Please make suggestions
There is also an ongoing discussion about Xen vs. KVM performance on the CentOS virt list
http://lists.centos.org/pipermail/centos-virt/2010-October/002074.html
One thing I can say already, I've been unable to get Fedora 14 to start as a KVM VM with more than one CPU. I'm having to rework my test bed to use RHEL 6.0 now that it's out and hope that it solves the problem. :)
On 10-11-12 09:57 AM, Todd Deshane wrote:
There is also an ongoing discussion about Xen vs. KVM performance on the CentOS virt list
http://lists.centos.org/pipermail/centos-virt/2010-October/002074.html
There are some interesting ideas and insightful comments there. I am tempted to abort my own benchmarking, as it seems others are doing the same. I think I will proceed though, until and unless someone there releases numbers first.
As I mentioned before, I've hit a snag with my in-progress benchmarks. I've filed a bug but I doubt I will wait for resolution. I'll need a bit more time now to setup a RHEL 6 based platform (host and guests). I'm still hoping to get most done this weekend, but I make no promises. If not, I'll aim for the weekend after to finish.
Cheers
On 11/12/2010 10:58 AM, Digimer wrote:
On 10-11-12 09:57 AM, Todd Deshane wrote:
There is also an ongoing discussion about Xen vs. KVM performance on the CentOS virt list
http://lists.centos.org/pipermail/centos-virt/2010-October/002074.html
There are some interesting ideas and insightful comments there. I am tempted to abort my own benchmarking, as it seems others are doing the same. I think I will proceed though, until and unless someone there releases numbers first.
As I mentioned before, I've hit a snag with my in-progress benchmarks. I've filed a bug but I doubt I will wait for resolution. I'll need a bit more time now to setup a RHEL 6 based platform (host and guests). I'm still hoping to get most done this weekend, but I make no promises. If not, I'll aim for the weekend after to finish.
Cheers
I wanted to post a follow-up, as I had hoped to have results well before now.
I've been working on getting Xen 4.0, dom0 2.6.32 and DRBD 8 running on RHEL 6 as part of the cluster testing I want to do. I've *just* finished getting these RPMs rolled, thanks to the great work of Pasi and Boris.
I am now fairly pressed for time with the holidays coming, so I do not expect to have a chance to revisit the benchmarking until the new year. I won't bother giving a date, instead opting for "when it's done, I'll post it" statement.
Sorry for the delays. With luck, anyone interested should be able to duplicate the setup with a simple repo addition, if all goes well. :)
On Mon, Nov 08, 2010 at 11:52:08AM -0500, Bill Davidsen wrote:
Dor Laor wrote:
On 11/08/2010 04:55 AM, M A Young wrote:
What do others think about this? For example is it achievable as a feature, is it too early and better to wait for F16, and what else should we aim to do to make xen usable in Fedora?
Have you consider kvm? it's upstream since 2.6.20 and now its more ready than ever.
There are some good tutorials which should explain the difference between xen and kvm, particularly the performance and hardware requirements of each.
In any case, the question of whether KVM or Xen is best, is not really relevant to whether Xen Dom0 has a place in F15. Fedora will welcome any software that meets the packaging & licensing guidelines, and has someone who is willing to maintain it. So if people want to maintain Xen as an alternative virtualization option in Fedora, they're welcome todo so. KVM will of course remain the default virt host setup offered in the installer
Regards, Daniel
Daniel P. Berrange wrote:
On Mon, Nov 08, 2010 at 11:52:08AM -0500, Bill Davidsen wrote:
Dor Laor wrote:
On 11/08/2010 04:55 AM, M A Young wrote:
What do others think about this? For example is it achievable as a feature, is it too early and better to wait for F16, and what else should we aim to do to make xen usable in Fedora?
Have you consider kvm? it's upstream since 2.6.20 and now its more ready than ever.
There are some good tutorials which should explain the difference between xen and kvm, particularly the performance and hardware requirements of each.
In any case, the question of whether KVM or Xen is best, is not really relevant to whether Xen Dom0 has a place in F15. Fedora will welcome any software that meets the packaging& licensing guidelines, and has someone who is willing to maintain it. So if people want to maintain Xen as an alternative virtualization option in Fedora, they're welcome todo so. KVM will of course remain the default virt host setup offered in the installer
Fine. The point I was making is that a non-trivial user base has hardware which does not support KVM, both legacy and recent low end CPUs like ATOM (and Celeron, I believe). And there is a fair amount of old hardware which does support KVM, but not all that well, like Q6600, which benefit from using xen. So while KVM may be a choice for recent hardware, there is a user base which would benefit from having xen.
And I would hope that the installer would be clever enough to see if KVM is supported and offer xen if not. Doesn't need to be default, available is fine.
And I would hope that the installer would be clever enough to see if KVM is supported and offer xen if not. Doesn't need to be default, available is fine.
What if i don't want KVM install no matter of hardware ? I will want Xen on icore7 + ASUS P6TDT+ 16 GB RAM , Dell PowerEdge and so on ...
Boris
--- On Tue, 11/9/10, Bill Davidsen davidsen@tmr.com wrote:
From: Bill Davidsen davidsen@tmr.com Subject: Re: [fedora-virt] [Fedora-xen] Dom0 xen support in Fedora 15? To: "Daniel P. Berrange" berrange@redhat.com Cc: xen@lists.fedoraproject.org, virt@lists.fedoraproject.org Date: Tuesday, November 9, 2010, 11:21 AM
Daniel P. Berrange wrote:
On Mon, Nov 08, 2010 at 11:52:08AM -0500, Bill Davidsen wrote:
Dor Laor wrote:
On 11/08/2010 04:55 AM, M A Young wrote:
What do others think about this? For example is it achievable as a feature, is it too early and better to wait for F16, and what else should we aim to do to make xen usable in Fedora?
Have you consider kvm? it's upstream since 2.6.20 and now its more ready than ever.
There are some good tutorials which should explain the difference between xen and kvm, particularly the performance and hardware requirements of each.
In any case, the question of whether KVM or Xen is best, is not really relevant to whether Xen Dom0 has a place in F15. Fedora will welcome any software that meets the packaging& licensing guidelines, and has someone who is willing to maintain it. So if people want to maintain Xen as an alternative virtualization option in Fedora, they're welcome todo so. KVM will of course remain the default virt host setup offered in the installer
Fine. The point I was making is that a non-trivial user base has hardware which does not support KVM, both legacy and recent low end CPUs like ATOM (and Celeron, I believe). And there is a fair amount of old hardware which does support KVM, but not all that well, like Q6600, which benefit from using xen. So while KVM may be a choice for recent hardware, there is a user base which would benefit from having xen.
And I would hope that the installer would be clever enough to see if KVM is supported and offer xen if not. Doesn't need to be default, available is fine.
On Tue, Nov 09, 2010 at 10:40:10AM -0800, Boris Derzhavets wrote:
And I would hope that the installer would be clever enough to see if KVM is supported and offer xen if not. Doesn't need to be default, available is fine.
What if i don't want KVM install no matter of hardware ? I will want Xen on icore7 + ASUS P6TDT+ 16 GB RAM , Dell PowerEdge and so on ...
I guess you will have to manually yum install kernel-xen. We can't magically know your preference, but it makes sense to default to KVM (by which I am really saying "default to a plain Linux kernel") because that is far less intrusive for users.
Rich.
On Wed, Nov 10, 2010 at 09:55:19AM +0000, Richard W.M. Jones wrote:
On Tue, Nov 09, 2010 at 10:40:10AM -0800, Boris Derzhavets wrote:
And I would hope that the installer would be clever enough to see if KVM is supported and offer xen if not. Doesn't need to be default, available is fine.
What if i don't want KVM install no matter of hardware ? I will want Xen on icore7 + ASUS P6TDT+ 16 GB RAM , Dell PowerEdge and so on ...
I guess you will have to manually yum install kernel-xen. We can't magically know your preference, but it makes sense to default to KVM (by which I am really saying "default to a plain Linux kernel") because that is far less intrusive for users.
NB, there is no kernel-xen anymore with pvops dom0. To enable xen you just need to yum install xen-hypervisor and update grub to boot the kernel under the hypervisor
Regards, Daniel
Config (.config) might be different for pvops kernel running as vanilla or under xen.
Boris.
--- On Wed, 11/10/10, Daniel P. Berrange berrange@redhat.com wrote:
From: Daniel P. Berrange berrange@redhat.com Subject: Re: [fedora-virt] [Fedora-xen] Dom0 xen support in Fedora 15? To: "Richard W.M. Jones" rjones@redhat.com Cc: "Boris Derzhavets" bderzhavets@yahoo.com, "Bill Davidsen" davidsen@tmr.com, xen@lists.fedoraproject.org, virt@lists.fedoraproject.org Date: Wednesday, November 10, 2010, 5:30 AM
On Wed, Nov 10, 2010 at 09:55:19AM +0000, Richard W.M. Jones wrote:
On Tue, Nov 09, 2010 at 10:40:10AM -0800, Boris Derzhavets wrote:
And I would hope that the installer would be clever enough to see if KVM is supported and offer xen if not. Doesn't need to be default, available is fine.
What if i don't want KVM install no matter of hardware ? I will want Xen on icore7 + ASUS P6TDT+ 16 GB RAM , Dell PowerEdge and so on ...
I guess you will have to manually yum install kernel-xen. We can't magically know your preference, but it makes sense to default to KVM (by which I am really saying "default to a plain Linux kernel") because that is far less intrusive for users.
NB, there is no kernel-xen anymore with pvops dom0. To enable xen you just need to yum install xen-hypervisor and update grub to boot the kernel under the hypervisor
Regards, Daniel
On Wed, Nov 10, 2010 at 02:42:56AM -0800, Boris Derzhavets wrote:
Config (.config) might be different for pvops kernel running as vanilla or under xen.
The config is a build-time thing. When Dom0 pvops Xen is fully merged into Fedora there will only be one single binary kernel image & RPM, that supports both Xen and Native. ie There is to be no separate kernel-xen at that time. Thus by definition there is no different .config for Xen vs native.
Regards, Daniel
Exactly, i only need on Fedora :) # yum install kernel-xen
Boris.
--- On Wed, 11/10/10, Richard W.M. Jones rjones@redhat.com wrote:
From: Richard W.M. Jones rjones@redhat.com Subject: Re: [fedora-virt] [Fedora-xen] Dom0 xen support in Fedora 15? To: "Boris Derzhavets" bderzhavets@yahoo.com Cc: "Daniel P. Berrange" berrange@redhat.com, "Bill Davidsen" davidsen@tmr.com, xen@lists.fedoraproject.org, virt@lists.fedoraproject.org Date: Wednesday, November 10, 2010, 4:55 AM
On Tue, Nov 09, 2010 at 10:40:10AM -0800, Boris Derzhavets wrote:
And I would hope that the installer would be clever enough to see if KVM is supported and offer xen if not. Doesn't need to be default, available is fine.
What if i don't want KVM install no matter of hardware ? I will want Xen on icore7 + ASUS P6TDT+ 16 GB RAM , Dell PowerEdge and so on ...
I guess you will have to manually yum install kernel-xen. We can't magically know your preference, but it makes sense to default to KVM (by which I am really saying "default to a plain Linux kernel") because that is far less intrusive for users.
Rich.
Dear Santa,
I don't have much, but I am willing to give up any and all of my Christmas presents if you can see your way to bring Christmas to a boy in the UK. You see, they have even less there in the UK - surveillance cameras everywhere to make sure they keep in line and boiled gruel twice a day if they are lucky. There is a boy, Michael, he works ever so hard - first shift at a coal mine and then second shift at a textile mill - and then after 2 10 hour shifts of backbreaking labor, he only has a couple hours per day by candle light while eating his nightly gruel to work on Xen for Fedora.
Mr. Scrooge and the other fat cats at Scrooge, Inc. tell Michael and the rest of us that we should be happy with their KVM. This, while dining on the finest roasted meats and full cream puddings in front of roaring fires fueled by wood gathered by orphans and widows paid in crust of week-old bread while young, robust, well-fed hamsters on whale oiled wheels run their giant gleaming servers - you see Santa, our hamsters are gaunt, long-of-tooth and have only a portion of Michael's nightly gruel that he is able to share with them and alas we have no monies for whale blubber with which to oil their wheels.
The rich seem to believe that we press for Xen only through laziness or recalcitrance - I can assure you this is not the case, we have simple needs and wants, we ask not for full bodied HVM capable X5680s but only that our humble, yet proud and hardworking Core, Pentium and Opterons are not dumped on the ground and doused in kerosene that they might never be used again. We are not asking for luxury and riches, just humble tools that a man might fashion an honest day's work from what he has. These hopes ride on young Michael's work. They are not the sole hopes of Michael or me but the hopes of thousands, living lives of quiet desperation, hoping beyond hope that one call goes their way, of having one branch to grab onto.
Dearest Santa - please don't leave these desperate souls a lump of KVM in their stocking.
Thank you and god bless us every one!
On 11/8/10 8:00 AM, Dor Laor wrote:
On 11/08/2010 04:55 AM, M A Young wrote:
I am trying to work out whether it is practical to propose Dom0 xen support as a feature for Fedora 15.
The kernel situation is that Domain 0 has been accepted upstream for 2.6.37. Assuming a 3 month kernel release cycle, F15 will most likely ship with a 2.6.37.x kernel, with 2.6.38 coming out either after the F15 release or just before but too late to be included. If the plan to get key xen drivers into 2.6.38 succeeds, then F15 may be become usable as a Domain 0 system at some point during its lifetime as the kernel package in a Fedora version typically has one major update.
If the kernel team accept backported patches then it might just be possible to ship F15 with usable Domain 0 support but the timescale for that would be very tight.
The other thing we would need to consider is what needs to be done to make xen friendly enough to be usable by an ordinary user. The page https://fedoraproject.org/wiki/Features/XenPvopsDom0 contains plans from when dom0 xen support was expected to make a quick return to Fedora, but they are a couple of years old now so probably need updating.
I think as a minimum we would need a way to add a dom0 enabled grub entry for a kernel, rather than requiring the user to hand edit the grub file. We should also make sure that xen works with the other Fedora virtualisation tools.
What do others think about this? For example is it achievable as a feature, is it too early and better to wait for F16, and what else should we aim to do to make xen usable in Fedora?
Have you consider kvm? it's upstream since 2.6.20 and now its more ready than ever.
Michael Young
virt mailing list virt@lists.fedoraproject.org https://admin.fedoraproject.org/mailman/listinfo/virt
-- xen mailing list xen@lists.fedoraproject.org https://admin.fedoraproject.org/mailman/listinfo/xen
On 10-11-12 08:22 AM, fcxen user wrote:
Dear Santa,
I don't have much, but I am willing to give up any and all of my Christmas presents if you can see your way to bring Christmas to a boy in the UK. You see, they have even less there in the UK - surveillance cameras everywhere to make sure they keep in line and boiled gruel twice a day if they are lucky. There is a boy, Michael, he works ever so hard - first shift at a coal mine and then second shift at a textile mill - and then after 2 10 hour shifts of backbreaking labor, he only has a couple hours per day by candle light while eating his nightly gruel to work on Xen for Fedora.
Mr. Scrooge and the other fat cats at Scrooge, Inc. tell Michael and the rest of us that we should be happy with their KVM. This, while dining on the finest roasted meats and full cream puddings in front of roaring fires fueled by wood gathered by orphans and widows paid in crust of week-old bread while young, robust, well-fed hamsters on whale oiled wheels run their giant gleaming servers - you see Santa, our hamsters are gaunt, long-of-tooth and have only a portion of Michael's nightly gruel that he is able to share with them and alas we have no monies for whale blubber with which to oil their wheels.
The rich seem to believe that we press for Xen only through laziness or recalcitrance - I can assure you this is not the case, we have simple needs and wants, we ask not for full bodied HVM capable X5680s but only that our humble, yet proud and hardworking Core, Pentium and Opterons are not dumped on the ground and doused in kerosene that they might never be used again. We are not asking for luxury and riches, just humble tools that a man might fashion an honest day's work from what he has. These hopes ride on young Michael's work. They are not the sole hopes of Michael or me but the hopes of thousands, living lives of quiet desperation, hoping beyond hope that one call goes their way, of having one branch to grab onto.
Dearest Santa - please don't leave these desperate souls a lump of KVM in their stocking.
Thank you and god bless us every one!
This made me laugh. Bravo! Someone obviously had their left brain firing on all cylinders. :)
M A Young wrote:
I am trying to work out whether it is practical to propose Dom0 xen support as a feature for Fedora 15.
The kernel situation is that Domain 0 has been accepted upstream for 2.6.37. Assuming a 3 month kernel release cycle, F15 will most likely ship with a 2.6.37.x kernel, with 2.6.38 coming out either after the F15 release or just before but too late to be included. If the plan to get key xen drivers into 2.6.38 succeeds, then F15 may be become usable as a Domain 0 system at some point during its lifetime as the kernel package in a Fedora version typically has one major update.
If the kernel team accept backported patches then it might just be possible to ship F15 with usable Domain 0 support but the timescale for that would be very tight.
The other thing we would need to consider is what needs to be done to make xen friendly enough to be usable by an ordinary user. The page https://fedoraproject.org/wiki/Features/XenPvopsDom0 contains plans from when dom0 xen support was expected to make a quick return to Fedora, but they are a couple of years old now so probably need updating.
I think as a minimum we would need a way to add a dom0 enabled grub entry for a kernel, rather than requiring the user to hand edit the grub file. We should also make sure that xen works with the other Fedora virtualisation tools.
What do others think about this? For example is it achievable as a feature, is it too early and better to wait for F16, and what else should we aim to do to make xen usable in Fedora?
Michael Young
It would be gret if it could be made to happen, dare I say that the documentation hurdles seem more challenging than the technical issues? I believe the pieces will all be in mainstream by 2.6.38, and hopefully the last bits can be included as patches to 2.6.37 (in addition to what's there). It's a bit rough for medium experienced users at the moment.
On 11/07/2010 03:55 PM, M A Young wrote:
If the plan to get key xen drivers into 2.6.38 succeeds
With the caveat that I don't really know what I'm looking for, I didn't see these make 2.6.38-rc1 (plenty of other Xen work, though). I saw a proposed patch for netback hit xen-devel yesterday.
So, did we miss Fedora 15?
-Bill
On Thu, Jan 20, 2011 at 09:34:50PM -0500, Bill McGonigle wrote:
On 11/07/2010 03:55 PM, M A Young wrote:
If the plan to get key xen drivers into 2.6.38 succeeds
With the caveat that I don't really know what I'm looking for, I didn't see these make 2.6.38-rc1 (plenty of other Xen work, though). I saw a proposed patch for netback hit xen-devel yesterday.
So, did we miss Fedora 15?
Yep, netback most probably won't be in 2.6.38.
Upstreaming of the various dom0 bits needs to be done step-by-step, the previous "everything at once" clearly didn't work.
netback could always be added as an out-of-tree patch, or even some dkms module/rpm ?
-- Pasi
On Fri, Jan 21, 2011 at 09:47:43AM +0200, Pasi Kärkkäinen wrote:
On Thu, Jan 20, 2011 at 09:34:50PM -0500, Bill McGonigle wrote:
On 11/07/2010 03:55 PM, M A Young wrote:
If the plan to get key xen drivers into 2.6.38 succeeds
With the caveat that I don't really know what I'm looking for, I didn't see these make 2.6.38-rc1 (plenty of other Xen work, though). I saw a proposed patch for netback hit xen-devel yesterday.
So, did we miss Fedora 15?
Yep, netback most probably won't be in 2.6.38.
Upstreaming of the various dom0 bits needs to be done step-by-step, the previous "everything at once" clearly didn't work.
netback could always be added as an out-of-tree patch, or even some dkms module/rpm ?
If what is left of netback is a purely self-contained driver module which doesn't touch any central kernel/arch code, then I think you'd stand a good chance of convincing the Fedora kernel maintainers to take it as a add-on patch to the official kernels.
Daniel
On Thu, 20 Jan 2011, Bill McGonigle wrote:
On 11/07/2010 03:55 PM, M A Young wrote:
If the plan to get key xen drivers into 2.6.38 succeeds
With the caveat that I don't really know what I'm looking for, I didn't see these make 2.6.38-rc1 (plenty of other Xen work, though). I saw a proposed patch for netback hit xen-devel yesterday.
There was talk of userland netback and blkback in xen 4.1 which use the gntdev interface which did make it into 2.6.38 (userland blkback apparently gives the similar performance as the kernel one, netback is not as good). I am not sure of the details though, nor how you might use them if they are available.
Michael Young
On Fri, Jan 21, 2011 at 08:39:46AM +0000, M A Young wrote:
On Thu, 20 Jan 2011, Bill McGonigle wrote:
On 11/07/2010 03:55 PM, M A Young wrote:
If the plan to get key xen drivers into 2.6.38 succeeds
With the caveat that I don't really know what I'm looking for, I didn't see these make 2.6.38-rc1 (plenty of other Xen work, though). I saw a proposed patch for netback hit xen-devel yesterday.
There was talk of userland netback and blkback in xen 4.1 which use the gntdev interface which did make it into 2.6.38 (userland blkback apparently gives the similar performance as the kernel one, netback is not as good). I am not sure of the details though, nor how you might use them if they are available.
Xen folks worked to give QEMU a userland netback/blkback implementation, but I don't believe this functionality is integrated into the mgmt stack at this time, so likely not much use yet.
Daniel
On Fri, Jan 21, 2011 at 11:19:44AM +0000, Daniel P. Berrange wrote:
On Fri, Jan 21, 2011 at 08:39:46AM +0000, M A Young wrote:
On Thu, 20 Jan 2011, Bill McGonigle wrote:
On 11/07/2010 03:55 PM, M A Young wrote:
If the plan to get key xen drivers into 2.6.38 succeeds
With the caveat that I don't really know what I'm looking for, I didn't see these make 2.6.38-rc1 (plenty of other Xen work, though). I saw a proposed patch for netback hit xen-devel yesterday.
There was talk of userland netback and blkback in xen 4.1 which use the gntdev interface which did make it into 2.6.38 (userland blkback apparently gives the similar performance as the kernel one, netback is not as good). I am not sure of the details though, nor how you might use them if they are available.
Xen folks worked to give QEMU a userland netback/blkback implementation, but I don't believe this functionality is integrated into the mgmt stack at this time, so likely not much use yet.
userspace blkback is included in xen-unstable (4.1) !
Afaik xen-unstable automatically 'falls back' to userspace blkback if kernel blkback is not available.
I'm not sure if xen-unstable also has userspace netback..
-- Pasi
xen@lists.stg.fedoraproject.org