►
From YouTube: KubeVirt Community Meeting 2019-05-19
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
Welcome
everybody:
this
is
the
cube
root
community
meeting
may
15th
2019,
probably
a
bit
of
a
light
agenda
today,
as
red
hat
summit
was
last
week
and
I,
don't
think
we
had
a
lot
of
features
but
then
mainly
focused
on
testing,
but
with
that
I
will
turn
it
over
to
Roman
to
discuss
red
hat
summit.
All.
B
There
were
already
a
lot
of
people
coming
over
which
are
following
the
project,
but
on
it,
yet
using
it
and
I
think
what
I
it's
a
steroid
now
a
little
bit
is
that
we
are
sticking
with
alpha
on
the
API
salon
right
now.
It's
mostly
about
not
breaking
the
API
for
all
the
existing
projects
and
less
about
the
stability
of
running
basic
virtual
machines.
Right
I
already
have
some
users
or
for
some
time
already,
which
use
it
very
successfully
in
a
stable
way.
B
Apart
from
that,
yeah
showed
some
demos
on
how
easy
it
is
to
install
Cubert.
That
was
pretty
well
received
it.
You
really
only
need
KVM
on
the
node.
Nothing
else
and
and
migrations
were
also
an
very
interesting
topic
for
many
people.
They
were
happy
to
see
it
further
regarding
two
features
which
we
don't
yet
have
but
which
they
would
want
to
see
is
cloning,
snapshotting
and
hot
block.
A
C
A
B
Also
now
got
the
credentials
and
the
budget
for
GCS
to
store,
build
artifacts
and
the
logs
and
the
test
results
so
all
set
there
still
I
will
be
out
next
week
on
Q
one
Q
Khan
I
would
prefer
if
we
didn't
switch
over
after
impact
from
there.
The
switch
would
include
not
not
fully
enabling
Pro
just
the
first
set,
and
the
first
set
is
that
we
just
run
all
tests
like
we
do
now
with
standard
CI
own
Pro,
which
gives
us
the
slash,
retest
and
and
all
the
other
features
from
pro,
but
not
yet.
B
B
And
the
immediate
immediate
benefits
you
should
have
is
less
reruns,
so
Pro
just
reruns
tests
based
on
changes
if
they're
Kurt,
Simon
master
and
if
they're
failing
so
you
know,
we
run
our
usual
tests
like
they
don't
really
have,
and
very
often
it
fails
because
of
one
test
which
succeeds
every
second
run
or
something
you
just
it
retest
and
it
just
reruns
the
ones
which
were
which
are
failing
or
if
master
changed.
That's
pretty
nice
already.
B
B
We
can
now
easily
push
build
artifacts
like
manifests
on
every
poor
request,
which
is
open
so
that
people
can
just
install
this
for
requests.
We
can
push
it
to
dr.
Harper,
gray
and
release
move
manifests
there,
which
then
can
be
installed
and
tested
out.
That
might
be
something
interesting
if
other
people
on
other
projects
want
to
try
out
a
pull
request
already,
which
can
sometimes
hang
for
a
few
weeks
until
it's
in
a
final
shape.
I
always
like
didn't
wanted
to
bring
it
up
as
a
possibility.
D
D
E
And
Wednesday
there's
a
link
there
in
the
notes,
too
well
more
details
and
register.
So
far
over
20
people
registered
to
we'll
have
a
nice
crowd
of
people
talking
about
keep
dirt
and,
as
I
said
common
to
probably
do
should
we
I
mean
so
many
people
will
be
there
I'm
wondering
if
we
should
cancel
the
meeting
this
meeting
next
week.
F
C
Maybe
just
multi
mark
regarding
go
models
because
I
created
the
pair
to
move
forward
to
visitor
to
work
with
go
models,
so
you
drop
support
project
might
and
it
also,
if
limitation,
on
the
raw
version.
So
role
models
exist
only
from
version
1.2
Levin,
but
again,
probably
the
most
part
of
the
people
just
build
it
under
the
container.
So
it
should
not
be
a
problem
and
at
least
from
what
I
saw
already.
It
works
much
faster,
then
glide
and
it
really
dropped
a
lot
of
unused
code
from
our
vendor
directories
low.
It's.
C
So
and
again,
I'm
currently
blocked
by
CGI
unit
testing,
because
they
currently
don't
use
updated
client
set
so
I
created.
They
are
also
topped
a
CGI
repository
to
use
good
night,
so
special,
what
dot,
13.4
and
generated
new
client
set
from
it.
So
when
it
will
be
most
I
will
update
my
PR
and
you
need
to
secure
new
path.
Yes,
Rachel
would
I
have
to
say
and
and.
C
C
Exactly
it's
like
you,
have
possibility
by
default,
go
models
just
trying
to
fetch
all
dependency
from
from
the
internet
and
put
it
under
some
cache
directory
and
there
you
go
go
on
page
ago
H,
but
you
have
possibilities
to
specify
environment
variables.
So,
first
of
all,
you
can
render
all
the
best
from
the
cache
locally
like
like
we
have
today
and
what
you
will
run
go
built.
For
example,
you
have
possibility
to
specify
to
environment
waterfalls
the
one
you'll
say
just
do
fetch
dependency
from
the
Internet
and.