►
From YouTube: KubeVirt Community Meeting 2022-09-07
Description
Meeting Notes: https://docs.google.com/document/d/1kyhpWlEPzZtQJSjJlAqhPcn3t0Mt_o0amhpuNPGs1Ls/
A
B
Good
day,
everyone
we're
everywhere
in
the
world
at
the
moment
this
is
the
Cuba
community
meeting,
I'll
rehearsing
it
instead
of
cat
today,
so
you'll
have
to
bear
with
me
yeah
once
again,
I
feel
good
at
the
Nancy
attendees
list
and
while
I
do
that,
I've
got
someone
new
here
today,
Larry
is
it
Dewey.
B
That's
fine
say
a
little
bit
about
yourself
or
if
you'd
like
yeah,.
C
Of
course,
yeah
and
it'd
be
my
pleasure,
so
my
name
is
Larry
Dewey
I
work
for
Advanced,
Micro,
Devices
or
AMD
and
I
am
just
attending
today,
because
we
heard
through
the
grapevine
that
you
guys
might
be
potentially
roadblocked
on
some
hardware
for
CI,
so
I
I
work
on
working
AMD
on
the
engineering
side,
I
work
on
the
confidential
containers
projects,
as
well
as
I'm
on
the
steering
committee
for
that
and
I
also
work
on
a
lot
of
Open
Source
stuff
as
the
open
source
ecosystem
team.
C
So
joining
me
today
is
going
to
be
Diego
Gonzalez
Villalobos.
He
is
just
on
my
team
and
another
person.
Who's
going
to
be
joining
me
today
is
Kim
narrow,
I'm,
not
sure
if
he
is
on
the
call
yet
or
not.
B
D
Hey
Larry
how'd,
you
go
good.
D
I
will
ping
you
with
Brian,
who
is
on
CI
team,
and
we
will
discuss
what
is
needed
from
our
side.
That
sounds
good.
A
B
Okay,
cool
doesn't
actually
tell
me
yeah
and
we
met
I
believe
two
of
your
colleagues
yesterday,
and
perhaps
someone
can
help
with
the
names
because
I'm
currently
sharing
my
screen.
We.
B
C
Yeah,
so
suddenly
it's
my
manager-
and
she
is
over
over
our
teams
and
and
unmashes-
is
over
the
CI.
So.
B
Excellent
excellent
cool
all
right.
It
sounds
like
you're
gonna,
sync
up
with
a
little
bit
about
this
meeting
all
right,
so
we
have
a
very
light
agenda
today.
If
anyone
wants
to
add
something
to
the
open
floor
of
the
agenda,
we
can
do
that.
But
while
we
just
in
case
not
we'll
go
through
the
pull
requests,
there
was
a
couple
of
pull
requests
that
I
didn't
have
too
many
comments,
but
already
have
an
lgtm
tag
on
them.
B
B
We
have
I
believe
blue
goes
on.
This
call
see
if
we
tagged
on
this
good
have
a
quick
look
at
that
that'd
be
great.
D
B
A
B
B
It's
back
on
all
right
cool.
It's
eight.
B
Nine
this
one
for
me
so
I'll
leave
that
in
your
capable
hands
on
the
mailing
list
review
we've
got
something
from
Daniel,
Heller
I
think
it
was
about
six
days
ago,
mentioning
thick
compute
migration.
We
don't
have
a
c
compute
migrations,
Lane,
yeah
gone
unanswered,
at
least
on
this
thread.
B
And
this
one
I
think
a
lot
of
people
have
seen,
but
I
thought
I'd
put
it
here.
Just
for
awareness,
it's
about
the
multi-architecture
cluster
proposal.
B
I'll
bring
up
the
Google
Doc
I've
got
the
Google
account
the
pl
link
here.
If
anyone's
interested,
it
seems
like
something
I
think
we've
been
talking
about
quite
a
long
time
and.
B
Welcome
to
looks
at
that
and
digest
it
at
your
leisure,
and
we
just
have
two
bugs.
E
Hello
yeah
hypo
he
built
kubernetes
wave,
1.19
class,
her
from
source
code
and
install
Calico
3.18
QB
words
0.35,
when
it's
used
qbivert
to
start
a
VM.
It
gets
following
errors
and
it's
English.
So
he
want
to
ask
how
to
solve
this
Arrow.
E
B
Is
there
anyone
on
this
call
that
is
able
to
take
this
and
I
can
assign
him.
D
A
D
Yeah,
that's
great,
it's
highly
possible.
We
can
accept
it.
Can
you
ask
if
the
author
can
or
wants
to
work
on
it.
B
B
Excellent
alrighty,
oh
good
I've,
got
some
stuff
appearing
on
this
all
right.
F
F
Yeah
just
wondering
why
expert
server
and
expert
proxy
developed
as
part
of
Cubit
project
and
not
as
part
of
CDI
project,
which
is
probably
intended
more
for
that.
Oh.
H
The
export
yes,
so
it's
because
the
export
CDI
doesn't
have
any
knowledge
of
virtual
machines
or
virtual
machine
snapshots,
and
the
export
API
supports
virtual
machines
and
virtual
machines
snapshots.
So
you
would
export
you
could
refer
to
a
virtual
machine
and
then
the
export
code
knows
what
disks
belong
to
that.
B
Everyone's
welcome
to
review
my
response.
There
and
Andre
you've
also
got
a
question
about
version.
1.0.
I
Yeah,
let
me
ask
he
he's
more
clearly:
did
that's
gonna
be
released
this
month
and
we
are
releasing
completely
based
on
cook
vert,
but
our
investors
ask
how
when
we
gonna
have
a
coupon
third
version,
one
and
I
was
nothing
to
answer
it
right
now.
I
would
like
to
know
who
decides
to
have
the
version
1.1
I
saw
that
we
have
some
documents
regarding
that
with
missing
something.
I
But
can
we
talk
here?
How
far
away
we
are
for
from
version
1.1
release.
J
You
asking
about
you,
you
said
1.1.
Are
you
looking
for
the
1.0.
J
The
ga
release-
okay,
yeah
so
I,
guess
the
so.
The
the
ga
release
of
Qbert
10,
okay,
so
like
this-
is
something
that's
been
in
discussion.
J
So
I
can
tell
you
that
that
that
I
was
gonna
volunteer
to
help
push
this
over
the
Finish
Line
I
can
tell
you
that,
like
the
dates
for
me
to
start
looking
at
this
is
going
to
be
in
September
and
but
I
don't
have
a
date
exactly
when
it's
going
to
be
released.
There's
a
document
that
has
the
outline
of
her
main
work,
but
there
are
some
things
at
least
that
I'm
interested
in
doing
before.
J
We
say
that
when
I
was
released
and
one
of
them
is
like
the
release
Cadence
and
changing
the
release
Cadence
instead
of
being
monthly
to
be
three
times
a
year
to
follow
kubernetes
and
there's
there's
already
a
discussion.
That's
that's
ongoing.
I
J
So,
to
give
you
a
perspective,
I
I
would
like
to
close
on
that
and
then
go
through
and
sort
of
adjust,
look
at
our
remaining
work
and
adjust
sort
of
what
what
we
need
to
do
to
align
around
that
release
and
then
I
think
we
can
discuss
some
dates
for
a
ga1o.
J
Yeah,
so
I
would
all
well
do.
Andre
is
when
you
know,
after
we
close
on
some
on
this
release.
Cadence
and
I
do
some
due
diligence
on
what's
remaining
I'll
I'll
I'll.
What
I'll
likely
do
is
start
a
meeting
on
this,
probably
something
weekly
or
either.
Maybe
it's
here
or
something
else,
and
then
also
the
mailing
list
started
going
with
the
list
of
remaining
items
just
so
that
everyone's
in
the
loop,
so
I
can
include
you
in
both
of
those
those.
I
Sorry
I
was
hearing
you
well.
Thank
you
so
much
if
we
can
contribute
with
something,
please
let
us
know.
Okay,
are
you
planning.
J
To
this
point,
so
at
this
point
I
would
say
the
only
thing
that
you
can
at
least
look
at
is
that
the
there's
a
document
laying
the
release
Cadence?
If
you
have
any
opinions
on
it,
you
know
review
it
and
and
say
your
piece
and
then
we
can
and
then
we'll
see
when
we
get
that
through
and
then
I
think
after
that,
it's
just
going
to
be
a
discussion
in
the
meeting
format.
J
So
I
haven't
I
I
think
it's
still
yet
to
be
discussed,
I
mean
because
the
keywords
had
technically
like
there's
been
a
Wonder
beliefs
but
I
guess
what
we're
saying
is
like
if
we
want
the
top
of
the
ga
release
for
it
right
so
I
I,
don't
we
haven't
really
discussed
what
exactly
this
will
look
like
when
when
it's
done
so
but
I
think
like
that's,
why
I
keep
saying
like
it's
important
that
we
discuss
any
other
concerns
and
like
at
least
in
the
release
Cadence,
because
this
is
a
big.
J
This
has
a
big.
This
would
affect
how
we
how
we
go
forward
and
releases
and
then,
if
we
have
like
you,
know,
I
think
as
a
follow-up
we
should
discuss.
You
know
what
we
want
to
do
is
like
announcing
this
and
so
there's
a
lot
of
conversation
still
to
be
had
as
this,
but
I
just
want
to
focus
on
the
pr
for
now.
J
B
Foreign
that
covers
everything
that
we
have
on
the
minutes.
Does
anyone
have
anything
last
minute
they
want
to
bring
up
while
we're
all
here.
I
We
are
working
to
increase
to
to
to
make
better
documentations
regarding
windows
and
convert,
because
the
solutions
I
saw
so
far
is
to
have
Windows
Vamps
without
activity
directory
domain
join,
and
we
have
several
difficulties
to
make
it
happen.
We
plan
to
contribute
on
that
when
it's
working,
we
are
still
fixing
some
things
we
find
so
far.
Okay,.
F
I
I
Have
to
upload
the
device,
do
it
even
the
the
golden
image
we
do
it
in
on
ramdisk.
Okay,
if
you
send
me
an
email,
I
can
show
you
what
how
we
have
done
that
then
we
have
don't
have
the
issues
we
have
before.
When
we
are,
we
was
using
what
is
called
Rock
and
Seth.
F
I
Thank
you
with
your
team
help
I
I
need
to
say
that
Ilya
help
us
with
some
ideas.
Okay,.
B
Oh
absolutely
sorry,
I
I
thought
someone
said
that
I'm
gonna
talk
to
you
offline
about
that.
But
absolutely
yes,
oh.
C
I
mean
we
definitely
can
I
had
invited
Kim
narrow
on.
He
helps
a
lot
with
that,
but
we
wanted
to
kind
of
get
a
good
understanding
of
what
was
needed
and
where
you
guys
are
I
know
you
spoke
with
sonly,
but
we
want
to
make
sure
everyone
was
kind
of
in
line
with
with
what's
going
on
in
Hubert
and
What
needs.
You
guys
have
so
make
sure
there's
no
there's
no
hold
UPS.
C
B
I
think
so,
who
have
we
got
from
represent
CI
here.
A
C
Yeah
I
guess:
there's
a
little
bit
of
a
there's,
there's
kind
of
been
some
discussion
from
different
teams,
as
well
as
some
partners
that
we
have
about
people
who
are
working
on
Cube
for,
and
they
had
mentioned
that
you
guys
have
some
needs.
So
would
you
mind
kind
of
outlining
what
that
need
is
and
what
we
can
do
to
help
foreign.
G
Yeah
sure
so,
basically
from
for
how
we
consume
other
kind
of
external
resources,
so
we
have
a
number
of
different
clusters,
like
an
iron
cluster,
that
we
extern
that
we
consume
externally
as
well.
We
generally
look
for
to
be
provided
with
kubernetes
installed
so
that
we
just
get
provided
with
a
cube
config,
and
then
we
can
interact
with
cluster
in
that
way.
G
That's
generally
how
we
look
for
external
hardware.
Does
that
sound
like
something
that
will
be
possible
to
me
before
I
just.
C
G
G
So
in
in
Cuba
CI,
we
is
proud
and
proud
can
communicate
with
external
clusters,
but
we
also
have
our
main
CI
cluster,
which
would
be
our
main
workloads
cluster,
and
so
when
I
refer
to
external
cluster,
I
mean
one
that's
outside
of
that
workloads
cluster,
which
we
can
just
provide
a
cubeconfig
to
prow,
and
then
Pro
will
be
able
to
communicate
with
that
cluster
and
schedule
workloads
on
there.
Gotcha.
C
Okay,
well,
let
me
again,
let
me
kind
of
introduce
Kim
narrow.
He
works
a
lot
with
with
it
with
enabling
Partners,
as
well
as
individuals
who
have
hibernating
Kim.
What
do
you?
What
do
you
think
about?
What
do
we
have
that
we
might
be
able
to
help
them
out
with
here.
K
So
I
I
guess
it
sounds
to
me
like
that.
This
is
more
of
a
generic
request.
I'm
just
wondering
is
there
specific
problems?
Of
course
you
know
working
for
AMD
I'm
concerned.
If
there
are
any
Hang-Ups
on
our
side
and
I
would
like
to
help
unblock
them
in
regards
to
clusters.
K
If
you
can
Define
the
number
of
systems
that
you
that
would
help
I
think
the
my
concern,
of
course,
is
you
know:
there's
some
Ambi
specific
functionality
I.E
under
Sev
SNP,
which
would
need
AMD
Hardware,
so
I'm
curious
how
the
Clusters
would
fall
into
that
and
help
the
system
getting
past
development
challenges
in
that
space.
G
G
D
G
G
K
D
Yeah
I
think
that's
sustainable
for
us,
because
there
are
just
a
couple
of
er's
a
couple
of
tests.
We
need
to
run
okay,
one
one
old
cluster
would
be
enough
for
us.
So.
K
Then
the
follow-up
question
is
I
mean
I'm.
I
would
be
thinking
Milan,
which
is
currently
available.
You
know,
processor,
in
the
market,
a
milan-based
system.
Are
there
any
memory,
requirements,
storage
requirements,
number
of
cores
requirements,
there's
an
eight
core
system.
Fine,
or
do
you
need
a
64
core
or
128
I
mean?
Does
that
matter.
G
So
are
generally
two
jobs:
they're
generally
made
up
of
about
29
gigs
of
RAM,
and
we
say
yes,
I'd
say
a
course
minimum
outside
of
that,
like
we
would
probably
be
with
with
this
special
we'll,
probably
be
scheduling
one
job
at
a
time
which,
realistically,
with
the
amount
of
testing
I,
think
that
is
there.
Okay
it'd,
be
more
it'll,
be
more
targeted
testing.
We
won't
be
running
everything
against
us,
so
I
think
yeah
the
the
requirements
there
would
be
reasonably
kind
of
like
a
a
mid-spec
system,
I
suppose,
okay.
G
So
our
cluster
is
is
in
IBM
is
in
it's
in
a
cloud:
it's
provided
by
a
cloud
provider,
so
our
kind
of
our
clusters
are
kind
of
spread
out.
So
we
don't
have
a
specific
address.
So
I
think
what
we
discussed
with
your
colleagues
was
that
an
endpoint
would
be
provided
to
us
that
are
maybe
in
AMD
labs.
K
G
Yeah
yeah
again,
we
I
think
this.
That's
the
same
conclusion
we
had
with
Richard
colleagues
a
couple
of
days
ago
on
the
call.
A
C
Gotcha
yeah,
we
had
had
a
discussion
with
one
of
our
partners
and
they
I
guess.
There
was
a
misunderstanding
where
they
thought
you
guys
actually
needed
physical,
hard
work,
and
so
we
were.
We
were
going
to
make
it
different
to
help
out
where
we
could
there
so
we're
we're
something
that
they're
I'll
coordinate
more
with
son
Mali,
then,
where
it's
just
a
remote
access
node
we
can.
We
can
get
that
sorted
out
for
you.
A
B
Excellent
wow,
thank
you.
All
righty
and
I
I
think
that
now
brings
us
to
the
conclusion
of
this
meeting.
Unless
anyone
has
anything
extra
to
add.
I
I
didn't
find
on
the
questions.
What
is
the
AMD
CPU
that
is
compatible
with
that
part
of
security.
I
No,
what
you're
planning
to
test
here.
C
It
has
an
onboard
security
chip
which
allows
for
for
Memory
full
memory
encryption
as
well
as
and
that's
that's
kind
of
the
Legacy
first
gen
SUV
work
and
then
Sev
es
allows
for
CPU
register
encryption
as
well,
and
so
it
provides
complete
confidentiality
and
security
for
VMS
or
pods,
and
so
that's
one
of
the
things
that
we're
looking
to
have
tested.
C
Need
yeah
so
yeah,
absolutely
so
the
and
that
that's
I
guess
is
in
terms
of
of
code
names,
so
Naples
forward,
so
it
would
be
Naples,
Rome
and
Milan
would
be
the
the
three
that
would
have
that.
I
can
there's.
C
C
And
SUV
SMP,
and
so
these
are
the
different
versions
that
are
available
for
that
and
what
you're
going
to
be
looking
at
for
it.
So
SUV
again
is
kind
of
first
gen.
What
you'll
likely
be
getting
access
to
is
is
the
the
ROM
box,
but
we
also
can
work
if
it's
necessary
at
the
time
to
provide
Milan
access
so
yeah.
So
those
are
the
different
versions
in
the
security
mechanisms.