►
From YouTube: KubeVirt Community Meeting 2022-03-09
Description
Meeting Notes: https://docs.google.com/document/d/1kyhpWlEPzZtQJSjJlAqhPcn3t0Mt_o0amhpuNPGs1Ls/
A
Okay,
everyone:
I
think
we
should
start
welcome
everyone
to
the
cubic
community
meeting,
it's
march,
9th
9th
of
2022.
First
of
all,
we
have
the
introduction
round.
A
A
A
Then
I
think
we
can
already
go
to
the
next
topic,
which
is
the
agenda.
The
first
is
just
a
heads
up
that
recently
cupid
0.51
was
released.
I'm
not
sure
if
anyone
wants
to
say
anything
about
that.
B
Yeah,
I
think
that
vasily
already
mentioned
it
and
he's
also
on
the
call
mentioned
it
already
in
the
mailing
list
that
new
discussion
line
popped
up
around
amd
suv
just
want
to
encourage
people
which
are
interested
in
confidential
computing
to
have
a
look
there,
we're
trying
to
figure
out
a
little
bit
on
how
to
make
in
amd
suv,
accessible
for
users
or
confidential
computers
a
whole
and
happy
about
any
additional
opinions.
There.
A
Okay,
yeah
short
agenda
today,
I'll
give
the
community
probably
another
minute.
Maybe
someone
wants
to
fill
something
in
or
add
something
to
the
open
floor.
B
A
As
I
don't
see
any
activity,
I
think
we
should
move
on,
so
we
also
don't
have
any
pull
requests
that
need
attention
at
the
moment.
So
I
think
we
should
just
go
to
the
mailing
list
preview,
but
let
me
share
my
window
one
second,
so
let
me
see.
C
A
A
Okay,
so,
okay,
so
what
I'm
seeing
at
the
moment
is
that
there
is
a
mail
from
the
regarding
the
renaming
of
the
eviction
strategy.
None
option,
I'm
not
sure
if
antonio
is
in
the
call.
A
D
So,
basically,
the
the
pr
that
you
see
in
the
email
got
merged.
I
think
one
two
weeks
ago.
Basically
it's
about
something
else
but
introduced
as
the
side
effect
introduced
this
non-option
for
the
edition
strategy
and
yeah,
I'm
just
having
doubts.
If
the
name
is
actually
you
know
spot
on
or
it's
it's
misleading.
D
What
none
means
is.
Basically,
there
is
no
actual
strategy
for
eviction.
So
when
the
bmi
gets
evicted,
the
pod
gets
deleted,
which
means
the
bmi
gets,
I
mean
shuts
down
and
then,
according
to
the
run
strategy
of
dvm,
the
the
vm
itself
can
be
restarted
or
not
restarted
anymore,
because
maybe
the
run
strategy,
it's
once,
which
means
like
run
to
completion,
and
that's
why
I
chose
the
name
non,
but
yeah.
B
D
But
as
I
as
I
say
in
the
email,
shutdown
might
be
incorrect
in
some
cases,
because
I
mean
it
depends
on
how
you
look
at
it,
because
the
vmi
will
actually
get
shut
down
because,
of
course
the
pod
gets
evicted.
But
if
the
run
strategy
is
always
for
example,
then
you
will
get
recreated
turning
the
whole
thing
into
a
reboot
rather
than
a
shutdown.
B
This
is,
this
is
fine.
I
mean
it
depends
on
what
you
have
specified
on
the
vm
when
you
specified
all
in
a
run
strategy
of
always
and
have
an
eviction
strategy
of
power
off.
I
think
it's
clear
that
it
will
come
back.
B
C
B
D
Yeah
yeah,
that's
why
I
had
the
dogs.
I
mean
internally
with
dogs,
because
I
mean
it's
not
really.
It
doesn't
really
tell
you
what's
going
to
happen.
I
guess
I
mean
having
a
you
know
a
developer
bias.
Then
you
know
what's
the
default
behavior,
but
someone
who
is
just
none,
it's
not
very
clear,
so
yeah.
B
D
You
thank
you
for
the
feedback.
Maybe
I
don't
know
if
someone
else
has
any
other
opinion.
If
not
now,
just
you
know
reply
to
the
email
so
that
we
can
reach
some
kind
of
consensus.
A
Probably
put
the
link
to
the
message
again
or
about
the
non-eviction
strategy.
It
is
a
document.
D
Yeah
but
definitely
shut
down,
I
mean
it's
definitely
more
suited
from
a
ux.
B
A
Okay,
so
then
we
have
a
look
whether
we
have
something
else
here:
okay,
I'm
not
sure
there
is
also
something
a
question
about
difference
in
disk
clone
and
vm
snapshot
using
volume
snapshot,
I'm
not
sure
if
anyone
has
already
had
an
idea
on
that
or
has
already
had
a
look
at
that.
A
Is
it
big
enough,
can
you
read
it
there.
B
B
D
B
Various
vm
cuber
driven
vm
snapshot
is
cooperating
with
the
guest
operating
system
and
you
get
a
crash.
Consistent
snapshot
of
the
disks,
still
not
an
application
crash,
consistent
snapshot,
but
an
operating
system
crash
consistent
snapshot.
That's
basically.
This
means,
when
you
think
about
linux,
the
linux
kernel
is
going
is
going
to
not
crash
when
you're
unpausing
or
restoring
that
vm.
It
can
just
continue
where
it
left
but
left
off,
because
the
disks
aren't
from
its
perspective,
but
applications
may
not
like
this
and
be
out
of
synth
or
something
still
so.
B
This
is
independent
of
applications,
whereas
when
you're
just
doing
a
csi
snapshot-
and
you
would
do
that
on
an
online
vm-
you
just
it-
basically,
just
I
don't
know
potentially
streams
through
your
disk
or
something
and-
and
this
is
of
course
not
necessarily
the
actual
list
actually
stating
the
guest,
because
the
linux
kernel
has
a
cache.
So
you
don't
know
if
the
disk
state
makes
sense
from
a
from
a
kernel
which
you're
snapshotting.
B
A
B
B
A
A
Just
from
brian
okay,
launching
erg
looks
also
good,
so
I
don't
think
we
have
any
other
industry,
so
I
think
we
can
move
over
to
the
next
topic.
That
would
already
be
a
bug.
Scrub.
A
A
A
B
A
C
C
B
Yeah,
so
I've
been
in
contact
with
a
person
on
with
tyler
ipr,
not
on
this
issue,
where
he
also
mentioned
it,
and
I
guess
this
is
about
having
a
transparent
reconnect.
So
when
he's
when
the
person,
when
it's
created,
when
the
user
is
creating
a
a
proxy,
it
connects
with
vnc
right
now,
the
behavior
would
be
that
if
the
connection
glow
the
web
socket
connection
closes,
the
proxy
just
goes
down
and
which
means
the
dnc
application
normally
closes
too.
D
On
which
pnc
server
you're
using
because,
for
example,
I
used
terminal
exactly
how
it's
called
but
like
x11,
being
see
something
it
kind
of
keeps
the
the
session.
Like
always
open.
Like.
B
B
B
It
makes
sense
the
request
we
can
probably
edit
try
to
accept
it.
I
guess
and-
and
you
can
write
something
like
at
least
having
an
option
to
transparently
reconnect-
sounds
reasonable.
A
A
I
I
guess
that
is
like
I
just
liked
it,
but
maybe
this
went
away
after
another
one
or
something
but
yeah.
I
guess
we
don't
have
any
validation
for
the
goal
target
so
that
could
be
yeah.
B
C
B
E
B
B
D
A
So
lubu
can
I
can
I
just
attack
you
on
that.
Do
you
want
to
fall
off
yeah.
A
A
A
E
It
seems
to
me
that
the
user
is
using
some
local
storage,
the
the
bit
flip
flop
pot,
which
is
the
wait
for
first
consumer,
gets
shuttled.
Then
we
try
to
shuttle
the
underlying
launcher,
but
maybe
memory
or
something
is
missing
there,
so
we
can't
schedule
it
anymore.
There.
A
B
You
know
what
I
guess
is
going
on
here.
This
is
like
if
there,
if
there
is
wait
first
customer,
then
we
have
this
kind
of
work
around
in
place
that
we
first
start
kind
of
a
dummy
part
which
binds
the
volume
to
a
node
which
is
most
likely
suitable
later
on
for
the
real
part,
so
that
we
kind
of
ensure
that
we
land
on
a.
A
I
I
I
if
I
understood
that
there
are
vmis
that
are
coming
up
successfully
and
others.
B
B
B
A
B
E
C
E
D
So
opera
is
more
interesting
the
status
then
I
don't
know
exactly
what's
information
it's
in
in
the
status
of
the
data
volume.
Maybe
it
has
something
insightful,
but
probably
we
could
see
the
same
thing
from
the
elements.
I
don't
know.
A
E
E
Yes,
that's
exactly
what
I
was
trying
to
suggest:
okay
to
mention
their
power
or
something.
E
A
A
A
A
But
did
anyone
just
curious
just
anyone
hear
about
truthstock
somehow,
is
that
some.
B
A
A
B
E
A
E
B
A
A
I
think
this
is
a
valid
issue
right
so,
but
we
could
of
course
suggest
that
if
he
has
the
time
he
could
probably
update
the
code
so
that
it
also
builds.
B
E
Troubles
here
do
we
need
to
test
the
binary,
though.
B
A
Yeah,
maybe
maybe
if
he
has
time,
I'm
not
sure
if
I
should
ask
for
this,
but
in
general
of
course,
if
you,
if
you
want
something
like
that,
you
could
probably
fiddle
around
and
try
whether
he
finds
something
out
himself
and
and
can
build
it
somehow.
A
Okay,
so
what
working
on
make
that
was
the
last
one,
this
one
is
from
petter,
okay,
this
one
is
already
accepted,
so
I
get.
I
guess
that
we
are
through
with
our
box
drop
great,
I'm
just
going
to
refresh
the
list
once
again,
so
that
we
see
if
every
she
has
had
a
label
somehow
that
we
handled
this
looks
good.
A
So,
okay,
I'm
going
to
find
the
button,
of
course,
stop
sharing
my
screen
and
leave
this
opportunity
for
anyone.
If
it
wants
to
mention
something
still
or.
B
A
Happy
to
give
everyone
10
minutes
back
from
your
precious
time,
and
I
guess
we'll
see
us
next
week
then.