9:45 a.m.
--Apple-Mail=_6D569541-1CA3-4944-AFE7-86FD31122B7D
Content-Transfer-Encoding: quoted-printable
Content-Type: text/plain;
charset=us-ascii
cross-posting to the right list, by mistake a wrong ovirt list was used =
originally
Begin forwarded message:
=20
From: Martin Kletzander <mkletzan(a)redhat.com>
Subject: [kubevirt-dev] Project for profiles and defaults for libvirt =
domains
Date: 20 March 2018 at 15:20:31 CET
To: libvir-list(a)redhat.com, openstack-dev(a)lists.openstack.org, =
ovirt-devel(a)redhat.com, kubevirt-dev(a)googlegroups.com, =
virt-tools-list(a)redhat.com, cockpit-devel(a)lists.fedorahosted.org
=20
Hi everyone!
=20
First of all sorry for such wide distribution, but apparently that's =
the
best way to make sure we cooperate nicely. So please be considerate
=
as
this is a cross-post between huge amount of mailing lists.
=20
After some discussions with developers from different projects that =
work
with libvirt one cannot but notice some common patterns and =
workarounds.
So I set off to see how can we make all our lives better and our =
coding
more effective (and maybe more fun as well). If all goes well we
will
create a project that will accommodate most of the defaulting, =
policies,
workarounds and other common algorithms around libvirt domain
definitions. And since early design gets you half way, I would like =
to
know your feedback on several key points as well as on the general =
idea.
Also correct me brutally in case I'm wrong.
=20
In order to not get confused in the following descriptions, I will =
refer
to this project idea using the name `virtuned`, but there is really
no
name for it yet (although an abbreviation for "Virtualization
Abstraction Definition and Hypervisor Delegation" would suit well,
IMHO).
=20
Here are some common problems and use cases that virtuned could solve
(or help with). Don't take it as something that's impossible to solve
on your own, but rather something that could be de-duplicated from
multiple projects or "done right" instead of various hack-ish =
solutions.
=20
1) Default devices/values
=20
Libvirt itself must default to whatever values there were before any
particular element was introduced due to the fact that it strives to
keep the guest ABI stable. That means, for example, that it can't =
just
add -vmcoreinfo option (for KASLR support) or magically add the =
pvpanic
device to all QEMU machines, even though it would be useful, as that
would change the guest ABI.
=20
For default values this is even more obvious. Let's say someone =
figures
out some "pretty good" default values for various HyperV
enlightenment
feature tunables. Libvirt can't magically change them, but each one =
of
the projects building on top of it doesn't want to keep that
list
updated and take care of setting them in every new XML. Some projects
don't even expose those to the end user as a knob, while others might.
=20
One more thing could be automatically figuring out best values based =
on
libosinfo-provided data.
=20
2) Policies
=20
Lot of the time there are parts of the domain definition that need to =
be
added, but nobody really cares about them. Sometimes it's enough
to
have few templates, another time you might want to have a policy
per-scenario and want to combine them in various ways. For example =
with
the data provided by point 1).
=20
For example if you want PCI-Express, you need the q35 machine type, =
but
you don't really want to care about the machine type. Or you
want to
use SPICE, but you don't want to care about adding QXL.
=20
What if some of these policies could be specified once (using some DSL
for example), and used by virtuned to merge them in a unified and
predictable way?
=20
3) Abstracting the XML
=20
This is probably just usable for stateless apps, but it might happen
that some apps don't really want to care about the XML at all. They
just want an abstract view of the domain, possibly add/remove a device
and that's it. We could do that as well. I can't really tell how =
much
of a demand there is for it, though.
=20
4) Identifying devices properly
=20
In contrast to the previous point, stateful apps might have a problem
identifying devices after hotplug. For example, let's say you don't
care about the addresses and leave that up to libvirt. You hotplug a
device into the domain and dump the new XML of it. Depending on what
type of device it was, you might need to identify it based on =
different
values. It could be <target dev=3D''/> for disks,
<mac address=3D''/> =
for
interfaces etc. For some devices it might not even be possible and =
you
need to remember the addresses of all the previous devices and then
parse them just to identify that one device and then throw them away.
=20
With new enough libvirt you could use the user aliases for that, but
turns out it's not that easy to use them properly anyway. Also the
aliases won't help users identify that device inside the guest.
=20
<rant>
We really should've gone with new attribute for the user alias instead
of using an existing one, given how many problems that is causing.
</rant>
=20
5) Generating the right XML snippet for device hot-(un)plug
=20
This is kind of related to some previous points.
=20
When hot-plugging a device and creating an XML snippet for it, you =
want
to keep the defaults from point 1) and policies from 2) in mind. Or
something related to the already existing domain which you can =
describe
systematically. And adding something for identification (see
previous
point).
=20
Doing the hot-unplug is easy depending on how much information about
that device is saved by your application. The less you save about the
device (or show to the user in a GUI, if applicable) the harder it =
might
be to generate an XML that libvirt will accept. Again, some
problems
with this should be fixed in libvirt, some of them are easy to
workaround. But having a common ground that takes care of this should
help some projects.
=20
Hot-unplug could be implemented just based on the alias. This is
something that would fit into libvirt as well.
=20
=
=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=
=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=
=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D
=20
To mention some pre-existing solutions:
=20
- I understand OpenStack has some really sensible and wisely chosen
and/or tested default values.
=20
- I know KubeVirt has VirtualMachinePresets. That is something =
closely
related to points 1) and 2). Also their abstraction of the XML
might
be usable for point 3).
=20
- There was an effort on creating policy based configuration of =
libvirt
objects called libvirt-designer. This is closely related to points
2)
and 3). Unfortunately there was no much going on lately and part of
virt-manager repository has currently more features implemented with
the same ideas in mind, just not exported for public use.
=20
We could utilize some of the above to various extents.
=20
Let me know what you think and have a nice day.
Martin
=20
--=20
You received this message because you are subscribed to the Google =
Groups
"kubevirt-dev" group.
To unsubscribe from this group and stop receiving emails from it,
send =
an email to kubevirt-dev+unsubscribe(a)googlegroups.com.
To post to this group, send email to kubevirt-dev(a)googlegroups.com.
To view this discussion on the web visit =
https://groups.google.com/d/msgid/kubevirt-dev/20180320142031.GB23007%40wh=
eatley.
For more options, visit https://groups.google.com/d/optout.
--Apple-Mail=_6D569541-1CA3-4944-AFE7-86FD31122B7D
Content-Transfer-Encoding: quoted-printable
Content-Type: text/html;
charset=us-ascii
<html><head><meta http-equiv=3D"Content-Type"
content=3D"text/html; =
charset=3Dus-ascii"></head><body style=3D"word-wrap: break-word; =
-webkit-nbsp-mode: space; line-break: after-white-space;" =
class=3D"">cross-posting to the right list, by mistake a wrong ovirt =
list was used originally<div class=3D""><div><br
class=3D""><blockquote =
type=3D"cite" class=3D""><div class=3D"">Begin
forwarded =
message:</div><br class=3D"Apple-interchange-newline"><div =
style=3D"margin-top: 0px; margin-right: 0px; margin-bottom: 0px; =
margin-left: 0px;" class=3D""><span style=3D"font-family: =
-webkit-system-font, Helvetica Neue, Helvetica, sans-serif; =
color:rgba(0, 0, 0, 1.0);" class=3D""><b
class=3D"">From: =
</b></span><span style=3D"font-family: -webkit-system-font, Helvetica
=
Neue, Helvetica, sans-serif;" class=3D"">Martin Kletzander <<a
=
href=3D"mailto:mkletzan@redhat.com" =
class=3D"">mkletzan(a)redhat.com</a>&gt;<br
class=3D""></span></div><div =
style=3D"margin-top: 0px; margin-right: 0px; margin-bottom: 0px; =
margin-left: 0px;" class=3D""><span style=3D"font-family: =
-webkit-system-font, Helvetica Neue, Helvetica, sans-serif; =
color:rgba(0, 0, 0, 1.0);" class=3D""><b
class=3D"">Subject: =
</b></span><span style=3D"font-family: -webkit-system-font, Helvetica
=
Neue, Helvetica, sans-serif;" class=3D""><b
class=3D"">[kubevirt-dev] =
Project for profiles and defaults for libvirt domains</b><br =
class=3D""></span></div><div style=3D"margin-top: 0px;
margin-right: =
0px; margin-bottom: 0px; margin-left: 0px;" class=3D""><span =
style=3D"font-family: -webkit-system-font, Helvetica Neue, Helvetica, =
sans-serif; color:rgba(0, 0, 0, 1.0);" class=3D""><b
class=3D"">Date: =
</b></span><span style=3D"font-family: -webkit-system-font, Helvetica
=
Neue, Helvetica, sans-serif;" class=3D"">20 March 2018 at 15:20:31 =
CET<br class=3D""></span></div><div
style=3D"margin-top: 0px; =
margin-right: 0px; margin-bottom: 0px; margin-left: 0px;"
class=3D""><span=
style=3D"font-family: -webkit-system-font, Helvetica Neue, Helvetica, =
sans-serif; color:rgba(0, 0, 0, 1.0);" class=3D""><b
class=3D"">To: =
</b></span><span style=3D"font-family: -webkit-system-font, Helvetica
=
Neue, Helvetica, sans-serif;" class=3D""><a =
href=3D"mailto:libvir-list@redhat.com" =
class=3D"">libvir-list(a)redhat.com</a>, <a =
href=3D"mailto:openstack-dev@lists.openstack.org" =
class=3D"">openstack-dev(a)lists.openstack.org</a>, <a =
href=3D"mailto:ovirt-devel@redhat.com" =
class=3D"">ovirt-devel(a)redhat.com</a>, <a =
href=3D"mailto:kubevirt-dev@googlegroups.com" =
class=3D"">kubevirt-dev(a)googlegroups.com</a>, <a =
href=3D"mailto:virt-tools-list@redhat.com" =
class=3D"">virt-tools-list(a)redhat.com</a>, <a =
href=3D"mailto:cockpit-devel@lists.fedorahosted.org" =
class=3D"">cockpit-devel(a)lists.fedorahosted.org</a><br =
class=3D""></span></div><br class=3D""><div
class=3D""><div class=3D"">Hi =
everyone!<br class=3D""><br class=3D"">First of all sorry
for such wide =
distribution, but apparently that's the<br class=3D"">best way to make
=
sure we cooperate nicely. So please be considerate as<br =
class=3D"">this is a cross-post between huge amount of mailing lists.<br
=
class=3D""><br class=3D"">After some discussions with
developers from =
different projects that work<br class=3D"">with libvirt one cannot but =
notice some common patterns and workarounds.<br class=3D"">So I set off =
to see how can we make all our lives better and our coding<br =
class=3D"">more effective (and maybe more fun as well). If all =
goes well we will<br class=3D"">create a project that will accommodate =
most of the defaulting, policies,<br class=3D"">workarounds and other =
common algorithms around libvirt domain<br class=3D"">definitions. =
And since early design gets you half way, I would like to<br =
class=3D"">know your feedback on several key points as well as on the =
general idea.<br class=3D"">Also correct me brutally in case I'm =
wrong.<br class=3D""><br class=3D"">In order to not get
confused in the =
following descriptions, I will refer<br class=3D"">to this project idea =
using the name `virtuned`, but there is really no<br class=3D"">name for
=
it yet (although an abbreviation for "Virtualization<br =
class=3D"">Abstraction Definition and Hypervisor Delegation" would suit
=
well,<br class=3D"">IMHO).<br class=3D""><br
class=3D"">Here are some =
common problems and use cases that virtuned could solve<br class=3D"">(or
=
help with). Don't take it as something that's impossible to =
solve<br class=3D"">on your own, but rather something that could be =
de-duplicated from<br class=3D"">multiple projects or "done
right" =
instead of various hack-ish solutions.<br class=3D""><br
class=3D"">1) =
Default devices/values<br class=3D""><br
class=3D"">Libvirt itself must =
default to whatever values there were before any<br class=3D"">particular
=
element was introduced due to the fact that it strives to<br =
class=3D"">keep the guest ABI stable. That means, for example, =
that it can't just<br class=3D"">add -vmcoreinfo option (for KASLR =
support) or magically add the pvpanic<br class=3D"">device to all QEMU =
machines, even though it would be useful, as that<br class=3D"">would =
change the guest ABI.<br class=3D""><br class=3D"">For
default values =
this is even more obvious. Let's say someone figures<br =
class=3D"">out some "pretty good" default values for various HyperV
=
enlightenment<br class=3D"">feature tunables. Libvirt can't
=
magically change them, but each one of<br class=3D"">the projects =
building on top of it doesn't want to keep that list<br
class=3D"">updated=
and take care of setting them in every new XML. Some projects<br =
class=3D"">don't even expose those to the end user as a knob, while =
others might.<br class=3D""><br class=3D"">One more thing
could be =
automatically figuring out best values based on<br =
class=3D"">libosinfo-provided data.<br class=3D""><br
class=3D"">2) =
Policies<br class=3D""><br class=3D"">Lot of the time
there are parts of =
the domain definition that need to be<br class=3D"">added, but nobody =
really cares about them. Sometimes it's enough to<br
class=3D"">have=
few templates, another time you might want to have a policy<br =
class=3D"">per-scenario and want to combine them in various ways. =
For example with<br class=3D"">the data provided by point
1).<br =
class=3D""><br class=3D"">For example if you want
PCI-Express, you need =
the q35 machine type, but<br class=3D"">you don't really want to care
=
about the machine type. Or you want to<br class=3D"">use SPICE,
=
but you don't want to care about adding QXL.<br class=3D""><br =
class=3D"">What if some of these policies could be specified once (using =
some DSL<br class=3D"">for example), and used by virtuned to merge them =
in a unified and<br class=3D"">predictable way?<br
class=3D""><br =
class=3D"">3) Abstracting the XML<br class=3D""><br
class=3D"">This is =
probably just usable for stateless apps, but it might happen<br =
class=3D"">that some apps don't really want to care about the XML at =
all. They<br class=3D"">just want an abstract view of the
domain, =
possibly add/remove a device<br class=3D"">and that's it. We
could =
do that as well. I can't really tell how much<br
class=3D"">of a =
demand there is for it, though.<br class=3D""><br
class=3D"">4) =
Identifying devices properly<br class=3D""><br
class=3D"">In contrast to =
the previous point, stateful apps might have a problem<br =
class=3D"">identifying devices after hotplug. For example,
let's =
say you don't<br class=3D"">care about the addresses and leave that up
=
to libvirt. You hotplug a<br class=3D"">device into the domain
and =
dump the new XML of it. Depending on what<br class=3D"">type of
=
device it was, you might need to identify it based on different<br =
class=3D"">values. It could be <target
dev=3D''/> for disks, =
<mac address=3D''/> for<br class=3D"">interfaces
etc. For =
some devices it might not even be possible and you<br class=3D"">need to
=
remember the addresses of all the previous devices and then<br =
class=3D"">parse them just to identify that one device and then throw =
them away.<br class=3D""><br class=3D"">With new enough
libvirt you =
could use the user aliases for that, but<br class=3D"">turns out it's
=
not that easy to use them properly anyway. Also the<br =
class=3D"">aliases won't help users identify that device inside the =
guest.<br class=3D""><br
class=3D""><rant><br class=3D"">We really =
should've gone with new attribute for the user alias instead<br =
class=3D"">of using an existing one, given how many problems that is =
causing.<br class=3D""></rant><br
class=3D""><br class=3D"">5) =
Generating the right XML snippet for device hot-(un)plug<br
class=3D""><br=
class=3D"">This is kind of related to some previous points.<br =
class=3D""><br class=3D"">When hot-plugging a device and
creating an XML =
snippet for it, you want<br class=3D"">to keep the defaults from point =
1) and policies from 2) in mind. Or<br class=3D"">something =
related to the already existing domain which you can describe<br =
class=3D"">systematically. And adding something for identification
=
(see previous<br class=3D"">point).<br class=3D""><br
class=3D"">Doing =
the hot-unplug is easy depending on how much information about<br =
class=3D"">that device is saved by your application. The less you
=
save about the<br class=3D"">device (or show to the user in a GUI, if =
applicable) the harder it might<br class=3D"">be to generate an XML that
=
libvirt will accept. Again, some problems<br class=3D"">with
this =
should be fixed in libvirt, some of them are easy to<br =
class=3D"">workaround. But having a common ground that takes care
=
of this should<br class=3D"">help some projects.<br
class=3D""><br =
class=3D"">Hot-unplug could be implemented just based on the alias. =
This is<br class=3D"">something that would fit into libvirt as
=
well.<br class=3D""><br =
class=3D"">=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=
=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=
=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=
=3D<br class=3D""><br class=3D"">To mention some
pre-existing =
solutions:<br class=3D""><br class=3D"">- I understand
OpenStack has =
some really sensible and wisely chosen<br class=3D""> and/or tested =
default values.<br class=3D""><br class=3D"">- I know
KubeVirt has =
VirtualMachinePresets. That is something closely<br class=3D"">
=
related to points 1) and 2). Also their abstraction of the XML =
might<br class=3D""> be usable for point 3).<br
class=3D""><br =
class=3D"">- There was an effort on creating policy based configuration =
of libvirt<br class=3D""> objects called libvirt-designer. This
is =
closely related to points 2)<br class=3D""> and 3).
Unfortunately =
there was no much going on lately and part of<br class=3D""> =
virt-manager repository has currently more features implemented with<br =
class=3D""> the same ideas in mind, just not exported for public use.<br
=
class=3D""><br class=3D"">We could utilize some of the above
to various =
extents.<br class=3D""><br class=3D"">Let me know what you
think and =
have a nice day.<br class=3D"">Martin<br class=3D""><br
class=3D"">-- =
<br class=3D"">You received this message because you are subscribed to =
the Google Groups "kubevirt-dev" group.<br class=3D"">To
unsubscribe =
from this group and stop receiving emails from it, send an email to <a =
href=3D"mailto:kubevirt-dev+unsubscribe@googlegroups.com" =
class=3D"">kubevirt-dev+unsubscribe(a)googlegroups.com</a>.<br
class=3D"">To=
post to this group, send email to <a =
href=3D"mailto:kubevirt-dev@googlegroups.com" =
class=3D"">kubevirt-dev(a)googlegroups.com</a>.<br
class=3D"">To view this =
discussion on the web visit <a =
href=3D"https://groups.google.com/d/msgid/kubevirt-dev/2018032014203...
007%40wheatley" =
class=3D"">https://groups.google.com/d/msgid/kubevirt-dev/20...
B23007%40wheatley</a>.<br class=3D"">For more options, visit <a
=
href=3D"https://groups.google.com/d/optout" =
class=3D"">https://groups.google.com/d/optout</a>.<br =
class=3D""></div></div></blockquote></div><br =
class=3D""></div></body></html>=
--Apple-Mail=_6D569541-1CA3-4944-AFE7-86FD31122B7D--
2468
days inactive
2468
days old
0 comments
1 participants
participants (1)
-
Michal Skrivanek