►
From YouTube: KubeVirt Community Meeting 2021-07-28
Description
Meeting Notes: https://docs.google.com/document/d/1kyhpWlEPzZtQJSjJlAqhPcn3t0Mt_o0amhpuNPGs1Ls/edit#heading=h.y4kzkq4flc0x
A
A
Hello,
you
won't
be
able
to
join
today,
so
I
will
try
to
host
instead
of
him.
Let's
wait
one
more
minute
and
then
we
can
start
with
the
agenda,
see
if
somebody
else
joins.
A
Okay,
there's
a
tiny
forum
today,
so
maybe
quick.
If
you
want
to
please
record
your
name
in
the
meetings,
I
will
share
them
in
the
chat.
B
A
Thanks
all
right
now,
it
feels
better.
Now
anybody
knew
on
the
call
who
would
like
to
introduce
theirs
themselves.
A
A
If
you
want
to,
please
add
your
topics
to
the
agenda,
I
edit
just
one
and
I'm
like
feeling
that
it
was
resolved
a
while
back,
but
keep
me
honest
here,
so
we
have
an
automatic
closing
mechanism
for
issues.
So
if
there
are
rotting
there
for
two
months,
they
get
closed,
but
we
don't
have
the
same
thing
for
prs,
so
they
are
piloting
which
is
shameful
if
they
don't
get
any
reaction.
But
it's
I
mean
there's
no
point
in
having
them.
A
If
there
is
nobody
pushing
them
forward,
so
any
takers
on
having
the
same
stale
and
frozen
and
closed
mechanics
for
brs
as
we
do
for
issues.
C
Yeah,
I
think
it
makes
sense
even
my
own
prs.
I
forget
to
close
them
at
times,
so
I
wouldn't
mind
having
something
go
up
and
clean
up
behind
me,
especially
like
six
months
or
a
year
afterwards,
and
it
gives
a
warning
so
I
mean
it
puts
this
is
still
and
then
it's
gonna
come
up
in
my
github
notifications.
I'm
gonna
have
a
chance
to
be
like
oh
yeah.
I
forgot
about
that
and
then
I
can.
You
know
reopen
or
you
know
clean
it
up
and
try
to
get
merged
again
or
just
let
it
rot.
A
Yeah
yeah,
I
hear
you
I
agree
like
some
may
argue
that
we
should
solve
the
issue
of
the
pr's
not
getting
addressed,
but
it's,
I
guess
different
conversation
yeah,
but
I
guess
it
should
be
encouraged.
If
somebody
is
just
sitting
there
and
waiting
for
his
pr
or
hers
to
get
reviewed,
they
should
make
their
sum
appear
either
here
or
mailing
cluster.
D
Hi
everybody,
so
this
is
basically
a
first
sketch
for
live
migration
policies,
and
some
people
have
already
seen
it
and
an
initial
discussion
has
already
started
so
everybody
that
would
like
to
take
a
look
and
comment
and
have
questions
or
ideas
or
anything
else
is
invited
to
do
so
or
to
do
so
yeah
and
that's
it.
A
D
Yeah
sure
so
would
you
like
me
to
share
the
screen
or
something.
D
Yeah,
so
basically,
the
goal
is
to
design
and
after
afterwards
implement
lab
migration
policies
that
can
be
defined
globally.
D
Basically,
the
idea
is
to
give
to
give
the
system
administrator
control
to
fine-tune
the
migration
configurations
and
policies
according
to
the
vm's
workload
or
according
to
different
priorities
that
he
has
one
one
example
could
be
to
update
the
cluster.
So
if
you
wanna
drain
a
node,
maybe
you
want
to
give
those
migrations
a
higher
priority
yeah-
and
there
are,
as
I
said
it's
very
initial.
So
there
are
several
dilemmas
and
approaches
and
nothing
is
determined
yet
so
so
yeah
at
the
beginning.
D
I
I
basically
talk
about
the
goal:
the
possible
stories
that
system
admin
could
have
and
use
cases
design
guidelines
and
then
some
sketches
of
various
ideas
that
popped
into
into
our
heads
yeah.
This
is
very
initial
work,
so
it's
a
bit
messy
and
you
all
are
welcome
to
read
it
and
take
part
in
the
discussion
yeah.
Nothing
too
special
to
to
show
here.
A
Looks
interesting
thanks.
Anybody
has
any
question
for
itamar.
E
A
A
Three:
two
one:
okay,
the
next
section
is
pull
requests
that
need
attention
and
there's
one
item:
does
anybody
want
to
promote
that?
One.
F
Currently
it's
a
bit
distributed
over
the
code,
so
there
is
no
one
place
that
implements
it
and
there
are
several
several
ways
to
do
the
same
thing,
so
this
should
be
supposed
to
unify
them
all
and
have
one
way
to
do
it
and
to
be
clear,
I
hope
if
it's
not,
I
prefer
please
comment.
F
But
in
the
end,
in
the
end,
it
also
should
allow
us
to
use
in
srv
we
have
the
ability
to
migrate
with
those
devices.
So
what
we
do
there,
we
unplug
and
then
plug
on
the
other
side,
so
which
we
could
now
use
this
the
same
mechanism
with
all
right.
I
think
that
that's
the
following
pr
to
extract
this
unplug
plug
part
from
the
srov
specific
to
be
generic
for
all
of
them.
G
Sorry,
I
I
was
just
wondering
if
the
design
of
the
pr
was
discussed
with
with
vladik
before
he
went
on
vacation,
because
he
I
mean
that's
his
code.
Originally,
I
was
wondering
if
he
had
an
opinion
on
that
pr.
H
Sorry,
I
just
yeah
because
there
wasn't
a
vacation.
I
didn't
get
to
review
the
pr
I'm
back
today.
G
F
It
is,
it
is,
I
think
I
spoke
with
radic
about
the
idea.
I
I
guess
the
design
is.
Is
I
didn't
don't
think
we
we
had
like
a
specific
design.
There
was.
There
is
a
there
is
an
introduction
in
the
in
the
pr
about
the
structure.
So
if
you,
if
anyone
has
a
different
idea,
then
you
are
welcome
to
tell
me
I'll
adjust.
A
So
the
next
section
is
the
review
of
the
mailing
list.
There
were
a
couple
of
messages.
Let
me
charge
my
screen
real,
quick.
A
Okay,
so
I
do
I'll
just
skip
those
that
have
plenty
of
attention,
which
is
this
one
another
one
is
for
igor
and
he
says
the
issue
was
resolved.
So
let's
skip
this
one
as
well
here
release
of
cdi.
Congratulations.
A
I
have
some
from
vladimir
that's
great.
Let
me
just
share
in
the
chat
for
anybody
who
would
like
to
take
a
look
and
on
the
agenda
as
well.
A
Oh,
no,
my
browser
got
stuck
okay,
the
the
next
one,
our
meeting
notes
and
finally,
the
last
one
from
a
week
ago,
ipv6
is
resisting
bridge
mode
to
how
to
make
the
click
on
figure
ip.
This
is
a
no
limitation,
known
limitation,
and
we
discussed
this
one
on
the
slack
channel
for
anybody
who'd
like
to
follow
the
responses.
A
So
that
will
be
for
the
list
and
let's
just
spend
the
next
15
minutes
or
so
with
the
issues.
If
you
don't
mind
to
use
a
bug,
scrubbing.
A
Anybody
understands
or
is
familiar
with
this
enhancement
proposal,
and
is
this
something
we
would
like
to
have.
C
C
Maybe
I
don't
know,
there's
a
precedent:
does
anyone
know
there's
a
precedent
for
this
already
with
maybe
other
operators
or
components.
A
Okay,
thanks
david.
E
I
wanted
to
reply
to
david
that
we
discussed
it
for
starting
with
ssp
operator
that
we
look
at
the
virtual
vertical
and
horizontal
portal
scales,
so
we
can
analyze
this
and
set
limits
and
that's
the
only
reason
we
should
try
to
keep
those
in
balance.
But
I
think
there
is
no
actual
reason
why
we
need
to
change
those
right
now,
except
scheduling
and,
I
think,
being
on
a
high
priority
class.
We
should
be
fine,
so
we
won't
get
killed.
E
Change
the
requests,
that's.
C
Actually,
so,
if
we're
gonna
create
an
alert,
the
one
of
the
criteria
for
that
alert
is
needs
to
be
actionable.
If
it's
not
actionable
yeah.
E
G
A
Okay
thanks.
Second
one,
some
methods
are
missing:
namespace
parametering,
open
api,
specs,
regular
json
file.
A
A
And
it
is
an
active
discussion,
so
I
will
and
it
seems
to
be
resolved,
so
I
will
just
close
it.
A
H
Yeah
I
mean
this
is
because
vert
launcher
didn't
manage
to
do
to
reply
in
a
meaningful
way,
but
there
there
should
be
something
in
invert
launcher.
I
guess.
C
Bird
launcher
isn't
responding
because
something
within
the
sync
vmi
call
it's
stuck,
and
my
guess
is
that
we're
trying
to
access
the
storage
and
it's
just
whatever
is
it's
just
blocked,
so
the
sync
bmi
call
can't
progress,
which
means
it
eventually
times
out.
We
never
get
an
error,
because
it's
not
an
ac.
A
Nice,
so
would
it
help
if
we
get
the
vert
larger
log
and
see
just
what's
the
last
thing
we
heard
about
and
then
find
the
place
in
the
code?
That's
the
next
line,
or
whatever.
A
C
A
A
Has
been
reviewed,
okay,
okay,
sickness
work,
please
remove
all
before
all
usages
in
the
test
yeah.
We.
I
brought
it
up
with
my
team,
and
we
will
look
into
this
when
we
find
time,
but
it's
certainly
something
we
would
like
to
do
so.
I
marked
this
accepted.
E
E
E
E
A
A
Enhancement,
fabian
opened
a
couple
of
them
seven
days
ago.
Templating
mechanism
for
vms.
A
C
E
Wasn't
this
one
of
the
milestone
issues
where
we
said
we're
gonna
have
a
big
discussion.
A
E
I
I
disgusted
a
lot
with
him
because
we
have
the
openshift
templates
in
in
with
us,
but
they
were
just
like
we
discussed
if
we
can
do
something
that
is
like
more
upstream,
like
there's
helm
but
helm,
never
felt
like
the
right
thing
for
it
for
this
and
what
we
can
do
to
provide
some
some
templating
mechanism
also
because
hell
needs
its
own
registry
and
all
that
kind
of
stuff.
E
It
didn't
feel
like
it's,
that's
native
templating,
for
what
we
could
use
in
a
cr,
but
I
think
it's
a
bigger
topic
and
we
need
to.
We
need
this
on
an
agenda
item
and
have
more
people
here,
including
fabian,
to
brainstorm
about
this.
C
E
C
E
E
C
Yeah
I'd
like
to
explore
that
more.
I
think
what
we'll
probably
do
for
this
topic
is
target
the
workflow
that
we
want
so
figure.
E
C
E
I
think
all
those
milestones
when
we
don't,
we
don't
need
to
go
over
because
we
we
actually
like
last
meeting.
We
said
we're
gonna
dedicate
a
whole
meeting
or
the
start
of
a
whole
community
call
to
one
every
week.
So
we
should
put
one
on
the
next
agenda
and
see
how
people
respond
to
it
or
send
a
instant
mail
to
the
mailing
list
and
say:
hey
next
week,
we're
gonna,
discuss
the
user
guide
and
groom
those.
A
C
For
a
topic,
we
could
probably
rule
out
the
persistent
container
disk
volumes.
We'd
want
somebody
from
kind
of
the
storage
realm
to
have
the
discussion.
Let's
see
who's
on
the
call
today,.
A
Okay
in
no
man,
no
reason
to
you
know
or
organize
it
here,
I'll
try
to
do
it
offline
and
see
when
he's
coming
back.
A
Okay,
so
let's
do
one,
that's
not
by
fabian.
E
E
A
wall-
this
is
six
seven
days
old.
We
should
have,
I
think,
I've
yeah
we've
talked
about
that
one.
E
E
A
Okay,
any
closing
words
from
anybody.