►
From YouTube: KubeVirt Community Meeting 2019-06-26
Description
No description was provided for this meeting.
If this is YOUR meeting, an easy way to fix this is to add a description to your video, wherever mtngs.io found it (probably YouTube).
A
B
B
Okay,
all
right
I'll
start
over,
so
we
had
a
good
discussion
on
the
mailing
list.
It
may
still
be
going
on
I'm,
not
sure
what
we're
looking
for
in
the
discussion
is
blockers.
Things
that
prevent
us
from
being
able
to
maintain
our
API
going
forward
is
G
a
thing
that
would
cause
a
backwards,
incompatible
change
for
us
to
be
able
to
support
it
moving
forward.
So
the
kinds
of.
A
B
We
can
do
moving
forward
is
we
can
add
things
to
our
API
and
that's
that's
really
about
it.
So
we
can
add
things
and
say
we
deprecated
previous
workflows
to
still
maintain
them.
So
there's
anything
that
you
guys
perceive
as
being
something
that
just
is
it
gonna
work
moving
forward
that
will
require
us
to
either
remove
a
field
or
modify
the
name
of
the
field.
We
need
to
be
aware
of
that
and
reconsider
whether
or
not
with
GA
that
at
the
moment
it
looks
like
we're
we're
moving
forward
with
it.
C
A
D
Yeah,
so
what
this
change
is
going
to
do
it,
it
will
prevent
virtual
machines.
From
from
my
say,
this
change
will
prevent
like
live
migration
for
virtual
machines
with
a
bridge
interface
that
simply
because,
after
live
migration,
virtual
machines
will
lose
connectivity
until
they
will
refresh
and
get
a
new
IP,
and
even
then
there
will
be
issues
in
order
to
prevent
losing
connectivity.
D
We
just
want
to
switch
by
default
for
using
the
masquerade
interface
and
this
this
basically
will
be
able
to
be
configurable,
which
interface
is
the
default
in
the
cube
config,
and
that's
pretty
much
it.
This
also
will
enable
allow
of
allow
users
to
use
the
slurping
slurp
interface,
and
they
can
also
disable
it.
E
It
has
some
assumptions
made
about
what's
available
to
it
and
since
we're
running
in
isolated
container
with
limited
Texas,
let's
say
PS,
nothing,
it's
not
working!
So
an
alternative
solution
that
we're
currently
approach
is
to
just
do
all
the
configuration
and
see
and
I
plugged
in,
and
there
is
PR
for
that.
It's
involving
but
reviews
welcome
if
you're
interested
in
this
kind
of
thing-
and
hopefully
it
will
not
require
anything
changes
on
convert
side
since
we
have
generic
my
commutation
support
yeah.
A
D
E
It
took
me
a
while
to
realize,
what's
going
on
until,
but
I
was
helping
me
with
the
buggy
in
it
and
he
connected
to
V
and
C
for
the
same
machine
and
then
console
unblocked.
So
it
turns
out
that
when
he
moved
miles
they
are
he
generated
some
entropy
and
they're.
Apparently
the
boot
process
was
blocked
because
no
RNG
available
and
it
was
waiting
for
some
source
of
entropy
I,
had
a
key
presence
or
a
mouse
or
whatever
and
was
never
going
anywhere.
E
So
once
he
moved
mouse
that,
like
the
system
continued
to
boot,
it
was
very
weird
but-
and
it
seems
like
it's
happening
it
on
fresh
kernels.
I
have
like
a
bunch
of
I,
have
more
details.
If
you're
interested
you
can
find
them
in
the
issue.
But
I
found
a
bunch
of
good
bad
reports
and
specialist
mailing
lists
around
that.
There
was
a
change
latest
in
latest
in
old
versions
that
made
it
worse,
so
I've
posted
a
bunch
of
PRS,
just
you
know
to
add
the
orangey
device
for
for
Fedora
examples.
E
I
have
a
documentation,
update
or
user
guide
mentioned
in
that,
but
also
documents
in
our
energy
in
general
it
was
not
documented,
but
there's
a
legit.
There
are
some
legit
comments,
questions
in
in
the
PRS
like
when.
Why,
wouldn't
we
just
add
orangey
by
default
to
all
vmi
and
I-
know
the
answer
to
that,
but
maybe
we
should
so
I
can
follow
up
and
just
push
a
patch.
That
would
just
do
that
always
add
RNG,
but
they
specified
it
not.
E
C
A
F
Yeah
I
also
think
I
think
that
still
the
concern
I
would
have
today.
The
question
to
me
is
what
devices
do
we
want
to
add
by
default?
If
we
know
ahead
of
time
that
there
are
not
recognized
for
all
guests?
That's
true,
for
example,
for
the
ballooning
device
for
the
virtual
orangy
device
and,
for
example,
for
of
the
trial
serial
console.
We
could
add
them
and
the
two
things
there
are
to
it
or
that
they
would
be
existing
every
guest.
F
D
F
Be
honest,
I,
don't
know,
I,
don't
really
don't
know.
I
actually
I
suggested
to
yard
that
we
rediscovered
I
wonder
actually
what
over
it
and
what
OpenStack
does,
and
maybe
we
can
take
to
the
meanness
because
they
have
experienced
with
it
and
they
have
judged
it
for
themselves
already.
You
know
do
we
do
we
advertise
reverse
this
by
default?
F
E
Maybe
I
wouldn't
I
suspect
it's
a
match
of
what
the
kernel
you're
running
on
the
host,
what
Cindy
and
they
image
and
maybe
something
else
about
the
hardware
it
solved
it
I'm,
not
exactly
where
just
was
very
confusing
when
and
you
hit
it
I
think
that
maybe
we
may
want
to
take
an
extra
step
and
prevent
this
happening
for
anyone,
even
if
maybe
sometimes
we
exposed
too
much
but
so
okay.
So
it's
it's
only
to
fall
up
and
check
what
what
others
are
doing
and
see
which
reasons.
A
E
The
X
in
the
first
reports
somewhere
around
the
middle
of
the
triggers-
yes,
it's
not
like
it
completely
fresh,
but
fresh
enough
and
I
also
suspect
that
you
know,
even
if,
like
one
of
the
one
of
the
weird
things
about
the
issue
that
I
feel
he
did,
it
was
it
Fedora
was
affected,
but
not
serious
or
Alpine,
because
because
they
they're
they
were
using,
they
were
starting
different
services
that
didn't
require
access
to
two
random
data
and
Fedora
is
doing
a
lot
more
stuff,
like
I,
think
it's
generating
this
HP.
So
maybe
that's
that's.
E
What
was
working
it
yeah,
but
I
think
but
I
think
like
it's
incorrect
to
suggest
that
Oh,
Cirrus
and
Alpine
are
not
affected
because
they
they
are
still
affected.
It's
just
that
their
boot
process
is
not
locked
by
the
issue,
but
if
you
run
something
inside
the
day
at
the
games
that
will
need
access
to
entropy,
then
it
will
probably
fail
in
similar
way
like
being
stuck
or
something.
F
A
A
G
But,
to
be
honest,
countless
a
bust
a
little
bit
blocked
because
of
the
because
he
cannot
bring
the
new
package
that
will
make
possible
to
support
who've
served
on
top
of
okey
D
4.1
the
to
cluster,
because
just
radical
sr8
system
does
not
have
iptables
model
under
it
in
the
only
option
is
to
use
enough
tables.
So
it's
it's
pretty
blocker
country.
We
can
work
around
it
like
in
different
hats,
like
we
use
the
little
lived
at
those
Liberty
much
and
create
the
new
image
that
will
take
from
this
image
and
Justin
thousand
you
package.
A
A
H
Yeah
and
basically
it
says
it
all,
so
you
come
in
a
closing
to
go
for
papers
in
a
couple
of
weeks
from
now.
So
this
is
a
cool
forum,
one
of
the
theory
they
have.
When
accepting
proposals
is
you
know
it's
easier
if
they
are
co-presented
by
teams
across
different
organizations,
so
I
would
encourage
people
to
to
say
you
know
first
obviously
find
an
interesting
topic
and
obviously
kuipers
is
a
very
interesting
topic,
but
there
are
many
ideas
here,
but
the
other
thing
would
be.
If
we
can,
another
goal
would
be.