►
From YouTube: KubeVirt Community meeting 2021-01-27
Description
Recording of the KubeVirt Community meeting on 2021-01-27
A
Start:
okay
welcome
everyone
to
the
kibber
community
meeting
on
27th
of
january
2021.
So
today
in
the
agenda,
the
first
topic:
we
have
a
topic
from
roman
on
multi-factor
authentication
for
the
repos.
B
So
I
just
wanted
to
head
up
that
if
you're
on
the
cube
list,
you
have
probably
seen
my
emails
regarding
google
to
verification
appeared
sent
to
think
street.
So
at
the
end
of
the
week
it
will
finally
enforce
them
station
left.
I
have
sent
sorry.
C
B
Okay,
wait
much
better!
Now
I
don't.
A
A
A
I
think
it
was
earlier
today
to
the
keeper
that
mailing
list
announcing
that
well
for
a
while
we
have
been
you
have
roman,
has
been
asking
people
to
enable
to
factor
authentication
to
github.
A
This
would
become
a
requirement
to
remain
a
member
of
the
github
work
for
the
giveaway
github
work,
and
the
announcement
here
is
and
in
the
email
and
to
emphasize
here
is
that
the
plan
is
to
start
enforcing
this
by
the
end
of
the
week,
meaning
that
after
after
that,
so
starting
from
next
week,
if
your
github
account
does
not
have
two-factor
authentication
enabled
you
will
not
be
a
member
of
the
github
org
and
the
tuber
github
work
anymore.
A
D
A
There
is
a
way,
but
everybody.
E
G
Just
a
quick
heads
up
that
we
were
talking
recently
about
automating,
the
cuber.org
membership
and
the
pr
on
this
has
initially
landed.
Now.
I
think
no
one
has
yet
used
it
exactly
but
yeah
I
at
least
at
least
it's
it's
working
now
and
then
you
can
propose
your
membership.
If
you
aren't
already
with
the
pr
I
linked
the
original
pr
so
from
there,
you
should
just
see
where
the
configuration
for
the
membership
can
be
added.
G
I
think
we
already
have
some
documents
in
the
works,
which
will
be
added
soon
to
add
also
the
instructions
in
the
community
documents.
I
guess
I
guess
pep
and-
and
chris
are
already
working
on
that,
like,
as
I
understand.
H
Yes,
we
actually
have
a
mention
of
that
later
in
the
on
the
open
floor,
but
since
we
brought
it
up
now
might
make
sense,
I
don't
have
a
link
to
the
pr,
but
we,
I
believe
the
plan
was
to
move
from
the
google
doc
that
we've
been
using
for
updating
the
or
to
set
the
rules
for
membership
application.
I
think
that
now
exists
as
a
github
pr.
I
know
it
was
being
in
the
process
of
being
marked
down,
so
I
think
a
pr
exists
for
that,
but
chris
can
speak
to
that.
I
Yeah,
I
created
the
pr
for
that
late
last
night.
I
posted
a
chat
and
then
the
meeting
notes.
I
The
comments
are
still
in
the
google
doc,
so
there's
still
a
number
of
comments
that
have
to
be
resolved.
H
Okay,
thanks
yeah,
so
if
we
could
reasonably
resolve
the
comment
by
basically
incorporating
the
suggestion,
we
went
ahead
and
did
that
and
I
guess
the
reason
to
bring
this
up
and
stress
it-
is
that
we
it's
the
conversation
is
going
to
become
less
productive
and
more
fragmented.
H
If
we're
trying
to
maintain
both
the
google
doc
and
github,
I
would
encourage
moving
to
github
comments
for
being
able
to
to
you
know
to
move
forward,
because
we
can,
you
know
once
again
still
there
get
the
same
effect
where
you
comment
on
the
line
in
github.
Instead
of
commenting
over
on
the
google
doc
site.
I
It's
so
the
google
docs
is
almost
getting
too
hard
to
follow
it's
and
then
we're
losing
we're.
Losing
initiative
on
on
getting
things
resolved
days
are
gone
by
now
without
comment
and
or
a
resolution
to
to
a
comment.
D
I
G
And
also
I'd
say
that
github
pr
is
much
more
open
to
the
qubit
community,
also
for
externals
and
for
everyone
to
see
than
a
google
doc
right
so
and
yeah.
Everyone
is
con,
confident
that
that
he
can
work
with
with
guitar
prs,
and
I
think
let's
just
stick
to
that,
because
just
we
will
have
to
incorporate
it
into
documents
anyway.
So.
A
Cool
thanks,
yeah,
it's
just
just
for
reference
purposes.
I
added
a
link
to
the
the
original
working
document
on
the
meeting
meeting
notes
but
yeah
as
just
to
emphasize
that
conversation
should
be
moving
to
the
pr.
I
We
do
want
this
published
before
summit.
I
A
That
would
be
the
goal.
Yes,
I
mean,
but
it's
more
important,
that
we
reach
a
proper
agreement
and
then
that
we
meet
a
specific
date.
But
let's
change
the
question
slightly
to
let's
say:
should
we
give
a
kind
of
a
lay?
I
mean,
let's
take
a
look,
let's
give
some
time
to
for
people
to
comment.
Maybe
it's
worth
following
up
on
the
kubertev
mailing
list.
I
Yeah,
that
sounds
good.
I
did
send
a
brief
email
to
cooper
dev
last
night,
just
saying
that
the
poll
has
been
created.
H
And
you
know
one
final
thought
on
the
matter:
once
we
publish
it
or
merge
it
in
github,
that's
not
to
say
that
it's
final
forever,
obviously
pr
can
be
raised
against
it.
We
can
amend
it
at
some
point
if
needed
or.
A
True
yeah,
of
course,
and
then
just
I
mean
the
merging
of
this
argument
into
the
repo
implies
a
few
follow-ups.
I
mean
there
are
other
documents
in
actually
different
repos
that
talk
about
the
consumer
process
and
those
who
need
to
be
updated
as
well
as
well
as
the
user
guide.
So
there
will
be
there's
more
to
that
than
just
this
pr.
A
C
Okay,
I'll
just
talk
about
real
quick.
I
wasn't
prepared
for
this,
but
so
there's
a
pr
that
that
vladik
made,
which
is
really
neat
that
allows
us
to
dynamically
increase
log
verbosity
on
the
keyboard
components,
it's
something
we
can
do
in
the
keyboard
cr
and
it
happens
dynamically.
C
So
if
you're
debugging
and
you
want
to
see
increased
verbosity
of
log
messages
check
out
that
pr,
it
gives
a
an
overview
of
how
you
can
change
keyboard
cr
to
enable
different
verbosity,
based
on
what
components
you're
interested
in
targeting
or
potentially
what
node
you're
interested
in
targeting.
So
if
you
have
a
specific
node
that
you
want
to
just
increase
the
verbosity
on
for
the
vert
handler,
then
you
can
target
just
that
node.
C
If
you
want
to
do
it
overarching
across
the
entire
cluster,
you
can
do
it
using
just
targeting
the
specific
component
and
the
api
and
that
pr's
description
describes
how
to
do
that.
So
anyway,
it's
a
neat
feature.
I
think
it's
going
to
help
out
with
development
and
kind
of
just
tracking
down
some
of
these
more
complex
issues,
and
it
actually
gives
our
law
verbosity
a
meaning.
Now
we're
in
the
past,
you
can
set
verbosity
high
and
then
essentially
you'd
never
see
those
messages
unless
you
went
into
some
custom
things
to
enable
it.
J
J
Yeah,
I
just
really
wanted
to
encourage
people
if,
wherever
it
makes
sense
to
add
more
logging,
it
will
definitely
help
in
debugging.
A
Okay,
moving
on
the
next
to
the
open
floor,
the
next
topic
was
a
status
update
on
the
keyboard
summit.
Two
things
on
that
we
received
and
by
the
way
david
I
saw
you.
You
were
also
adding
an
item
on
the
agenda
about
this.
I
don't
know
if
you
want
to
talk
about
the
proposals.
Let
me
give
a
quick
update
on
the
general
status
of
things
so
we're
working
on
on
preparing
all
the
logistics.
There
has
been
a
change
this
week
on
the
logistics
side.
A
That
kind
of
is
delaying
a
few
days,
but
we
aim
to
have
details
by
the
end
of
the
week
that
we
can
share
for
everyone
to
attend
on
the
proposals
or
the
session
sites.
So
one
of
the
things
that
we
need
to
publish
is
the
schedule
and
what's
the
content
and
that
relates
to
the
sessions,
I
don't
know
david.
Do
you
want
to
talk
about
this,
or
should
I.
C
I
got
slightly
distracted,
but
I
will
talk
briefly
that
we
got
maybe
hope
you
didn't
just
say
this
because
I'll
be
saying
the
exact
same
thing.
We
got
a
tremendous
amount
of
turnout
way
more
than
we
expected,
and
that
meant
we
had
to
kind
of
shore
up
a
review
team
quickly,
and
we
did
that
across
multiple
companies
and
we
looked
at
the
presentations
we
all
kind
of
independently
raided
them
and
it
ended
up
that
we've
had
to
cut
a
few
which
I'll
I'll
be
responding
on
the
pr's
and
things
like
that.
C
What's
been
accepted,
what's
been
cut
and
we'll
potentially
have
a
waiting
list
as
well
for
the
ones
that
are
in
the
rating
kind
of
close
to
that
line
and
well
I
thank
everyone
for
their
submissions.
This
was
again
way
more.
It
turned
out
than
we
expected
the
event.
Isn't
that
long,
and
we
look
forward
to
kind
of
next
year
being
able
to
do
this
again
and
having
potentially
a
longer
a
longer
event
with
where
we
can
accept
more
people.
C
A
Have
yeah
a
good
problem
to
have
indeed
but
yeah?
I
want
to
emphasize
thank
you
to
everyone
who
submitted
because,
beyond
the
quantity
I
I
would
say
that
the
quality
of
the
submissions
or
the
the
interest
of
the
topics
was,
is
also
significant,
so
yeah
thanks
again
for
every
for
the
interest
and
all
the.
A
Proposals-
okay,
anything
anything
else
on
this
topic
or
any
other
topic
that
was
the
last
one
on
the
agenda.
If
not,
we
will
move
to
the
box
scrap.
E
Hello,
so
if
you
don't
mind
spending
20
minutes
on
the
back
scrubbing
please
stay
around.
I
will
share
my
screen.
E
Yes,
all
right,
so
the
first
one
deploying
william
vm
with
uf5
files
was
it
addressed
to
us.
L
Hi,
let
me
catch
up
here.
L
L
So
if
you
enable
dfi,
you
also
need
to
enable
smm,
unless
you
disable,
which
is
a
bit
convoluted
for
users
to
comprehend.
L
So
I
was
wondering
what
we
could
do
to
improve,
that.
The
reason
why
secure
boot
is
enabled
by
default
is
that
we
want
to
default
to
secure
configuration.
The
reason
why
smm
is
not
enabled
by
default
is
because
it
could
be
a
security
concern
from
users.
C
C
It's
making
people
explicitly
understand
what
these
features
are
and
turn
them
on
that
they
need
them.
But
what
would
you,
what
would
be
your
intuition
on
how
to
make
this
more
user-friendly.
L
So
what
I
did
already
is
in
the
api
documentation
it.
It
says
exclusively
that
security
will
be
enabled
and
will
require
this
command
yeah.
I
I
I
think
the
the
the
way
it
is
right
now
is
probably
the
best
for
security.
L
M
L
Yeah,
it
is
documented
in
the
in
the
api
we
could.
I
could.
C
One,
so
I'm
not
familiar
with
ssm.
What
exactly
is
that
feature.
L
It's
also
called
ring
minus
two,
so
it's
like
a
super
super
privileged
cpu
mode,
where
you
can
do
anything
you
want,
including
on
the
firmware.
C
Yeah,
I
don't
think
we'd
want
to
enable
that
silently,
I
think,
we're
in
the
best
situation
we
can
be
in
at
the
moment,
but
I
agree
with
that.
N
N
E
All
right
done
here
we
go
our
back
roll
binding,
convert
io
and
it
failed
to
list
converts
and
good
vertio.
E
I
see
that
should
be
looking
into
it.
Okay,
so
I
we
can
skip
it
if
it's
being
addressed,
I
love
one
unless
somebody
wants
to
erase
anything
for
this.
E
Okay,
this
one
package
hooks
uses
the
same
for
the
file
name,
so
for
our
hooks,
we
have
defined
several
protofiles
for
the
product
buffers
and
they
all
share
the
same
name,
which
seems
to
be
an
issue
so
and
the
what
it
calls
this
is
this
tiny
warning.
E
I
guess
there
is
no
reason
not
to
address
this
right.
I
would
even
consider
it
a
good
first
issue,
since
it's
just
renaming
supposedly.
C
Yeah,
that's
the
data
volume.
The
original
proposal
for
data
volume
was
to
call
it
a
vm
image,
but
we
wanted
to
at
the
time
cdi
was
considered
to
be
something
that
might
be
more
generic
and
something
that
might
actually
have
a
use
case
beyond
cube
vert.
So
we
picked
a
more
generic
name.
O
That's
what
cdi
is
for
okay,
maybe
we
can
find
them
cti.
N
C
N
N
Okay,
thanks
david.
C
P
I
can
comment
a
bit
about
this
issue
yeah.
I
am
vasila,
I'm
actually
working
on
it
and,
as
romans
suggested,
I
pick
this
so
basically,
the
issue
is
that
on
the
nodes
we
have
c
groups
version
two
which
are
incompatible
with
the
version
one
and
produces
an
error
during
slice
detection,
which
is
actually
written
in
issue
description.
P
C
H
I
would
expect
so
in
general,
if
this
is
going
to
play
in
where
it's,
because
we're
using
system
d
inside
of
our
container
and
running
multiple
services
are
we
doing.
C
P
Found
three
three
places
where
the
issue
may
arise
so
that
they
are
described
in
the
pull
request
which
I
created
so.
P
Slice
detection
yeah
that
uses
groups
to
detect
basically
the
slides.
The
second
one
is
with
hot
plug
discs.
It
uses
a
device
controller
and
the
third
one
is
the
reading.
The
cpu
set
file
yeah,
so
the
written
here
so
in
general,
the
issues
can
be
fixed.
There
is
another
problem
with
testing
this
thing,
because
the
sci
now
is
configured
to
use
group
v1
and
in
order
to
enable
the
support
for
v2,
there
should
be
some
kind
of
a
new
vm
with
more
recent
kernel
and
container
runtime.
C
Okay,
so
for
my
own
understanding,
can
we
even
start
a
virtual
machine
today
on
a
host
with
secret
version,
two.
P
Are
you
asking
if
it's
possible?
Well,
yes,.
C
Well,
I
mean
it's
possible
it
today,
given
the
latest
release
a
convert,
will
it
work
or
will
it
fail.
P
It
will
fail:
okay,
yeah,
I
don't
know.
Maybe
in
some
use
cases
it
will
actually
run
the
vm,
but
in
the
I
think,
if
you
run
an
example
with
container
disk,
it
will
certainly
fail.
Q
See
groups
v2
and
leave
the
version
one
behind.
P
P
P
Oh
yeah,
but
as
far
as
I
know,
some
distances
are
already
just
switching
to
the
sigrid,
so
so
that
there
are,
let's
say
distros
that
enable
both
the
kernel
allows
such
configuration
but
container
runtimes.
They
do
not
usually
support
that
so
either
v1
or
v2.
C
I
think,
even
when
we
look
at
like
how
key
verse
productized
there's
a
certain
level
of
support
for
v1
right
now,
that
still
needs
to
exist.
E
Like
the
most
expensive
option
would
be
to
run
to
like
two
set
of
lines,
one
with
only
one
one
with
v2,
but
that
clearly
cannot.
It
would
be
huge
waste
problem,
but
maybe
having
a
nightlight
job
to
verify
that
v2
functions
would
be
enough
for
this.
C
C
Time
I
think
that
might
be
my
preference
and
we'd
still
keep
c-group
version
one
unit
tests,
so
I
think
the
things
that
actually
parse
the
parsing
logic
and
all
that
we
can
probably
kind
of
solidify
that,
with
unit
tests
to
ensure
a
certain
level
of
compatibility,
even
in
the
future.
If
we
we
don't.
C
For
that
specific
functionality
that
you're
talking
about,
I
would
recommend
syncing
out
with
alexander
wells,
because
he's
the
one
that
wrote
that
recently-
and
he
even
mentioned
it's
known
to
him-
that
secret
version
two
would
not
work
with
that
and
I'm
sure
he
wants
to
get
that
addressed
as
well.
That
might
be
a
little.
P
A
I
have
one
sad
question:
do
you
guys
think
this
topic
is
worth
some
documentation
beyond
aligning
the
release
notes
or
something
I
mean
it
sounds
confusing
to
me
enough.
Maybe
you're
just
wondering.
C
Hopefully
it
all
just
ends
up
working.
I
consider
it
a
bug
right
now
we
could
kind
of
describe
we
could
describe
what's
going
on,
if
I'm
not
sure.
C
E
Sense
all
right,
I
guess
vessel,
if
you
thanks
for
the
contribution
and
once
it's
ready
for
review,
please
move
it
to
assigned.
P
E
J
E
Bird
all
right
admission,
blogs,
integers
and
cpm
memory
resources
we
talked
about
this
last
week.
Haven't
we.
C
C
Looks
ready
to
go?
What
was
the
last
comments.
E
All
right
remove
the
ducks
directory
from
pr
testing.
E
Do
we
have
danielle
here?
Would
you
danielle,
would
you
like
to
give
us
a
highlight
what's
going
on
here.
E
G
Right:
okay:
the
problem
is
that
I
digged
a
little
bit
into
how
regex
is
working
in
in
golang,
and
the
problem
is
that
we
don't
have
really
an
easy.
J
G
To
to
use
this,
one
have
changed
so,
in
my
opinion
at
least,
there
is
no
easy
solution.
I
just
didn't
want
to
close
it
right
ahead,
because
I
think
it's
a
valid
point
and
I
think
we
should
do
something
about
that,
but
I'm
not
sure
what
the
I
I
think.
The
only
thing
we
could
do
would
be
to
move
the
documents
folder
somewhere.
G
E
J
E
G
To
be
honest,
I
think
there
are
not
that
often
pr's
against
the
against
the
markdown
documents.
Actually,
at
least
as
I
see
these,
I
think
there
has
been
one
or
two
in
the
recent
weeks,
and
I
don't
think
this
is
really
pressure
pressing.
E
A
A
quick
point
of
order
that
you
we've
been
more
than
20
minutes
on
this.
I
believe
we
can
continue,
but
oh.
E
I
think
that
this
is
kind
of
a
success
that
we
are
14
days
ago
on
the
box
curbing
so
eventually
we
will
go
get
through
all
of
them
thanks
for
the
timekeeping,
but
let
me
stop
sharing
them.
A
Okay,
actually
about
about
this,
I
was
just
wondering
we.
We
always
start
the
box
chrome.
I
find
it
great,
but
we
always
start
from
the
top,
and
you
know
we
we
review
issues
that
have
been
updated
a
few
hours
ago.
Would
it
be
worth
maybe
next
week
taking
a
look
at
the
older
ones
or
start
from.
E
Yeah,
I
guess
we
can,
but
the
issue
is
the
oldest
ones,
are
like
a
year
or
two
years
old.
Then
it
would
be
more
difficult
to
get.