►
From YouTube: KubeVirt Community Meeting 2021-09-15
Description
Meeting notes: https://docs.google.com/document/d/1kyhpWlEPzZtQJSjJlAqhPcn3t0Mt_o0amhpuNPGs1Ls/edit#heading=h.8fnfqz957tgn
A
Okay,
hello:
everyone
welcome
to
another
weekly
edition
of
cooper
community
meeting.
This
is
your
chance
to
talk
to
your
developers
and
community
about
anything.
Github
issues,
pull
requests
using
kubevert.
A
So
posting
meeting
notes
into
chat.
If
you
could,
please
enter
your
attendance.
A
And
we
usually
start
the
meeting
off
by
giving
a
few
minutes
for
new
people
to
introduce
themselves.
So
anybody
want
to
say
hi.
A
A
Okay,
let's
get
started
then
daniel.
You
have
the
first,
the
first
agenda
item.
B
Is
it
better,
though?
Yes
much
better?
Thank
you.
Okay.
Great.
Thank
you
for
the
heads
up
yeah.
I
was
wanting
to
talk
about
that.
We
have
good
news
that
finally,
cuber
ci
has
got
an
integration
of
1.21
kubernetes
providers,
and
these
have
already
been
merged
through
the
latest
scuba
ti
and
the
they
are
going
to
land
in
cuba.
Soon.
B
I've
attached
the
pr
where
this
bump
pr,
which,
which
integrates
the
latest
qpc
item
keyboard,
is
going
on
all
1.22
test
lanes
for
all
six
have
been
positive,
which
is
great
to
see
and
yeah.
We
hope
that
it
will
merge
today
or
tomorrow.
B
Yeah,
I
think
I
think
it.
Finally,
it
took
a
while
that
that
one
more
than
more
than
one
month
after
release
of
kubernetes
1.22
but
yeah,
I
think
we're.
Finally,
there.
A
I
just
got
ownership
rights
to
coobert,
which
have
been
populated
to
the
cncf,
and
now
I
can't
have
access
to
their
jarrow
and
I
can
create
a
ticket
for
them
to
actually
do
it.
So
hopefully
I
can
get
that
for
you
today.
B
Yeah
yeah
yeah
yeah.
Another
thing
I
wanted
to
talk
about
is
that
we
are
currently
working
on
integrating
or
bundling
cdi
with
cupid's
eye,
so
you
don't
have
the
hassle
to
install
cdi
all
by
yourself.
Instead,
you
can
just
set
a
environment
variable
to
true
and
then
keep
it
will
after
cluster
up,
it
will
deploy
cdi
in
the
qpti
cluster,
which
in
turn
makes
us
get
rid
of
this
manifest
in
the
cupola
code,
which
has
this
bundled
manifest
of
the
cdi.
B
So
we
can
just
update
cdi
in
one
place
and
have
everything
in
one
place
and
get
rid.
Finally
of
the
keyword
cdi,
install
installation.
B
C
B
A
How
come
cdi
doesn't
install
by
hco
by
def
by
default?
Why
is
it
a
two-step
process.
C
Yes,
for
keyboard
development,
we
are
not
using
hco,
because
hco
is
there
to
install
components
like
keyboard
and
cdi
and
so
on.
So
this
is
achievement
x
thing.
This
is
really
just
for
keyword
development,
where
we
are
deploying
the
keyboard
source
code,
but
it
also
needs
cdi
for
testing
features
where
the
two
work
together.
A
A
All
right,
roman
go
ahead
with
the
thank
you,
daniel
sorry
for
the
the
brain
failure
there.
So,
let's
move
on
to
you.
C
Okay,
so
some
of
you,
even
in
the
call,
probably
have
tried
to
to
make
use
of
cuber
dot.
Io
slash
client,
go
at
one
point
and
they've
probably
seen
that
we
have
huge
issues
there
with
our
dependencies
which
we're
pulling
in
from
various
others.
C
And
I
guess
it's
really
time
for
us
to
to
work
towards
a
more
clean,
client-chain
approach
where
we
potentially
have
our
own
api
packages
for
components
like
cdi
and
kubert,
and
people
would
just
take
these
api
packages
and
use
the
incline
chain
to
generate
the
clients
or
use
a
dynamic
client,
and
we
stay
out
of
the
business
of
providing
a
rest
client
ourselves,
because
right
now,
you
it's
almost
impossible
to
render
the
client
code
package
in
there
michael.
What
are
your
thoughts
on
that
mechanics.
C
But
yeah,
I
guess
it
would
make
sense
if
we
make
progress
on
that
and
then
the
goal
of
this
would
be
that
we
basically
almost
only
depend
on
core
kubernetes
dependencies
like
kubernetes.api,
kubernetes,
audio,
slash
api
machinery
and
that's
actually
pretty
much
it
already.
There
should
be
much
more.
A
Thanks
roman,
I
just
had
to
deal
with
our
dependencies
filling
out
the
cncf
due
diligence
and
notice.
How
many
there
were
we
can
clean
up
that
list.
A
Okay,
how
about
we
talk
about
some
events,
then.
A
I
think
she
has
birds
of
the
feather
this
week.
A
Okay,
she
was
there.
Hopefully
she
had
success,
I'll,
try
and
get
her
on
email
and
see
how
oh
he
said.
Everything
went
good
great
yeah
that
was
kvm
forum,
birdsville
feather
just
informal
q,
a
and
introduction
to
cooper.
E
A
Great
glad
to
hear
and
congratulations
on
your
first
presentation
with
cooper.
A
Okay,
I
volunteered
david
vassell
and
myself
to
talk
about
coobert
at
kubecon
la
so
we're
going
to
do
the
same
thing:
we're
going
to
do
the
intro
to
cooper
along
with
the
demo
it's
office
hours.
So
again,
it's
it's!
It's
informal!
A
The
the
red
hat
ospo
appreciates
our
are
volunteering
for
taking
on
office
hours.
They
have
nine
slots,
they
have
to
fill
and
so
they're
they're
reaching
out
to
all
cncf
communities
to
find
find
people
to
fill
those
slots.
D
A
Stu
stu
is
not
able
to
make
the
meeting
this
morning.
I'm
really
gonna
be
hounding
you
guys
to
to
get
on
this.
One.
D
Yeah,
so
if
I
can
say
I
can
have
a
demo
working
probably
by
tomorrow
as
far
as
having
a
local
cluster
going.
If
you
want
me
to
pair,
I
can
pair
with
my
another
site
at
the
location,
or
I
can
give
instructions
on
another
user
here
as
far
as
how
to
pair
with
them.
But
I
did
some
research
with
wireguard
primer.
I
used.
I
kila.
D
Sorry
yeah,
then
super
super
right
that
works
yeah.
A
Okay,
why
do
I
have
a
feeling
that
I'm
missing
an
event.
A
A
Okay,
do
we
have
any
pull
requests
that
anybody
would
like
to
talk
about.
A
Okay,
I
actually
have
some
full
requests
that
I
found
over
in
the
uber
community.
We
have
another
a
number
of
design
proposals
that
have
been
in
progress.
A
If
you
guys
can
take
a
look
at
the
at
these
proposals,
one
of
them
has
been
in
progress
since
april
29th.
We
have
one
from
july
6..
A
Let's
make
sure
we
get
some
some
attention
on
these
proposals,
and
let
me
know
when,
when
we're
good
for
good
for
merging.
A
And
we'll
get
some
communication
out
to
the
email
list
as
well.
A
Okay,
let's
see
mr
this
item
in
the
open
floor.
Just
a
quick
note
about
the
cncf
pull
request
for
graduation
to
incubation.
A
Please
comment
on
it.
If
you
have
anything
that
you
want
to
add.
A
A
A
Are
huge
wins
for
us
guys
massive
companies
to
that
are
using
cooper
and
hopefully,
along
with
that
eleno
will
authorize
us
for.
C
This
is
the
the
pr
already
open
for
in
cnc
everybody.
A
It's
out
of
our
hands
now
this
is
they're
gonna.
Take
this
document
and
they're
going
to
have
their
own
discussion
with
the
board
and
that's
something
that
is
it's
private,
but
then
they
they
publish
a
video
and
of
course
they
they
interface
with
us
regarding
the
results
of
their
discussion.
A
The
toc
has
a
youtube
channel
and
if
you
want
to
look
at
some
of
their
other
board
meetings,
you
can
there
are
there's
no
code
or
anything.
It's
just
all
paperwork
discussion.
A
But
I'm
I'm
really
interested
in
being
able
to
get
get
an
apple
logo
advertised
here.
I
think
that
would
would
be
amazing.
A
A
Mailing
list,
I
did
a
mailing
list
review.
I
think.
A
Notes,
we
talked
about
vert
handler
restarts
already,
so
I
think
we
have
everything
covered.
A
C
A
C
C
Okay,
then
stop
sharing.
Maybe
someone
can
share.
E
A
C
C
A
C
A
Yeah
david's
been
having
trouble
with
it
for
quite
some
time.
Yeah
like
I'm,
I'm
not
fond
of
my
mac,
but
it
works.
I
don't
I
don't
fight
it.
C
Yeah
we
can
start
there
yeah
and-
and
I
already
respond
there
and
I
so
the
person
is
trying
to
run
a
mac
on
keyboard
but
seems
to
have
issues
I'll.
Let
the
triage
need
more
info
flag.
I
asked
for
some
state
of
the
virtual
machine
down
there.
If
you
could,
there
is
no
further
context,
given.
A
C
C
A
A
C
6397,
you
are
yeah,
so
we
have
what
janusz,
as
far
as
it
goes
from
his
github
account.
That's
it.
His
name
is
talking
about.
Is
that
when
you
start
a
virtual
machine
with
the
word
launcher
starts,
and
it
actually
focus
itself
itself
and
monitors
itself
so,
and
this
is
there
to
ensure
that
we
have
a
proper
exit.
If,
if
we,
for
instance,
have
a
communication
bug
with
liver
or
cream
or
whatever
and
root
launcher
crashes,
we
would
we
would
lose
our
outmost
process
of
the
shell
and
the
van
would
crash
with
it.
C
That's
not
necessarily
what
we
want.
We
want
to
monitor
the
process
potentially
restart
it
ensure
that
we
at
least
give
the
process
enough
time
to
properly
shut
down
without
crashing
the
vm,
and
therefore
we
have
a
f4
a
a
4k
monitor
model
here,
but
this
requires,
I
don't
know
50
or
16
megabytes
of
ram
in
addition
for
each
vm,
which
is
an
issue
in
scale,
and
so
the
question
is
legitimate
to
remove
the
memory
footprint
a
little
bit
like
we
could
have
a
static
compiled
very
small
c
fork
and
monitor
loop
or
whatever.
F
Yeah
rowan
we're
gonna.
I
was
hoping
we
could
talk
about
this
one
that
tomorrow
and
in
the
scale
meeting
because,
like
you
said
like
with
the
amount
of
memory
per
vm,
it
does
add
up
significantly
and,
and
one
of
the
things
is
like
one
of
the
things
he
tried
here,
like
the
no
fork
flag
was
didn't
quite
like
it
sort
of.
We
didn't
really
like
find
what
what
and
the
use
for
it
was
because
we
kind
of
wanted
something
in
between
like
we
don't
want.
F
We
don't
want
to
use
the
memory,
but
we
also
don't
real
like
we
want
to
get
like.
We
we're
okay
with
forking.
As
long
as
like
the
memory
is
like
tiny
like
we
don't
like
you
know,
essentially
that's
so
we
were
kind
of
caught
in
between
like
like
so
we
one
of
the
things
we
looked
at
is
trying
to
find
something
in
the
middle,
but
we
didn't
really
know
what
the
no
fork
was
for.
So
we
were
kind
of
one
of
the
things
we
were
thinking
about.
F
B
C
F
Yeah,
I'm
hoping
the
other
guy
that
commented
there,
the
ex
pivark
there
was
also
going
to
be
in
the
discussion
because
he
was
looking
at
this
he's
got
another
issue.
He
was
kind
of
curious
about
that's
kind
of
been
related
to
this.
So
maybe
we'll
have
a
discussion
in
this
area
tomorrow
or
we
have
everyone:
okay,
cool.
A
Okay,
thank
you
for
that
ryan
client.
Go.
A
C
C
B
A
B
A
Okay,
I'm
going
once
going
twice
all
right.
That
concludes
this
week's
meeting.
Thank
you,
everybody
for
joining
us
and
we'll
talk
to
you
again
next
week.