►
From YouTube: KubeVirt Meeting Meeting 2021-03-17
Description
KubeVirt Meeting Meeting 2021-03-17
A
Okay,
hello,
everybody
I'm
chris
caligari
and
I'm
the
community
manager
for
coobert-
and
this
is
our
weekly
meeting
to
discuss
important
topics
and
community
affairs,
and
I
shall
share
my
screen.
B
A
Hi
fabian
sorry
I
had
my
I
had
my
share
screen
window
open.
I
I
could
hear
a
voice,
but
I
couldn't
see
you,
but
I
know
your
voice.
Well,
no,
you
don't
know.
Do
you
don't
need
to
see
me?
A
We've
all
got
that
lockdown
look
yeah.
My
my
wife
made
me
shave
my
head
a
few
weeks
ago,
so
I'm
a
little
thin
up
here,
the
contrary
up
here,
you're
asking
for
it.
A
Screen,
I
think
I
have
it
okay
here
we
go
okay
for
everybody,
who's
new.
We
do
keep
a
log
on
who
attends
the
meeting.
So,
if
you
could,
please
add
your
name
here,.
A
And
then
I'll,
I
will
add
the
recording
afterwards.
A
For
anybody
who's
new,
if,
if
you'd
like
to
take
a
a
few
seconds
to
introduce
yourself,
that
would
be
really
cool.
C
Hi,
sorry,
I
guess
I'll
start.
First,
I'm
the
only
new
one,
I'm
sammy
walker,
I'm
a
infrastructure
consultant
in
red
hat,
I'm
coming
on
board
actually
from
a
person
with
slack
conversations
regarding
some
it
was
at
se
linux
and
we
do
testing
among
different
things.
So
I'm
trying
to
get
more
involved
with
keyboard
as
a
whole
too,
since
I've
been
using
it
late
before
openshift
adopted
it.
So
definitely
looking
forward
to
meeting
up
one
and
getting
to
this
community
here.
A
I
always
like
to
hear
about
how
people
are
using.
C
Cooper,
I
can't
seem
to
edit
the
document
requested
edit
access
to
add
myself,
but.
A
Just
for
for,
for
sake
of
speed,
I'll
go
ahead
and
add
you,
you
have
to
join
the
the
mail
list
and
there's
instructions
up
at
the
top
on
how
to
do
that
can
do
it
can
do.
Thank
you
for
that.
A
E
A
So
I
can
talk
to
about
the
the
conference's
topic.
I've
been
getting
swamped
with
with
requests
to
be
present
at
various
conferences.
We
have
kubecon
eu
coming
up.
We
have
red
hat
summit
coming
up.
A
We
flat
out
missed
suzatan,
that
was
in
november.
I
believe
it
was
wasn't
even
on
our
radar,
and
that
was
I
think
that
was
a
a
big
mistake.
On
our
end
personally,
I
was
eyeballs
deep
into
getting
a
user
guide
migrated
into
from
docsify
over
to
make
docs
and
drop
the
ball
on
getting
out
to
events
in
the
fall.
A
So
in
the
future
I
would
love
to
be
present
at
suse,
suze,
khan
and
and
rancher.
A
Of
course,
I
was
a
one-person
team
and
now
chandler
is
with
me,
so
we
have
a
lot
more
bandwidth
to
to
set
up
demonstrations
and
whatnot
for
for
conferences.
So
we
we're
going
to
work
much
harder
this
year
in
getting
out
there.
A
I
have
a
colleague
at
nasa
who's
very
interested
in
using
kubernetes
and
kubert
to
manage
his
his
legacy
of
virtual
machine
workload.
Alongside
of
his
future
workloads
that
will
be
containerized.
A
So
in
the
long
term,
that's
going
to
be
a
great
opportunity
for.
A
A
Anybody
have
any
other
questions,
concerns
about.
A
A
And,
of
course,
volunteers
are
always
welcome
if
you
like
to
talk
with
people
in
chat
and
and
looks
like
the
platforms
are
going
to
have
a
some
kind
of
video
conferencing
feature.
So
if
you
like
to
talk
to
to
users
or
potential
users,
this
is
an
excellent
opportunity.
Please
reach
out
to
me
and
we
can
get
you.
C
A
Yeah
excellent,
thank
you.
Let's,
let's
talk
and
slap
and
and
I
can
help
you
get
help
you
on
board
with
what's
happening
for
events
perfect.
A
A
And
and
you'll
find
all
the
places
that
we
hang
will
out.
A
A
Knew
it
was
you
because
I
just
approved
your
pull
request.
F
Yeah,
just
just
a
quick
heads
up
to
the
community
that
we
have
now
something
in
place
to
make
transparent
and
automate
memberships
of
of
keyboard
contributors
and
also
for
management
of
repos
and
teams,
and
so
on
and
yeah.
I
put
together
yesterday's
small
document
on
just
where
to
file
prs
against
I.
This
is
not
exactly
completely
detailed
for
exactly
details.
F
A
Very
good
thanks
for
that.
If
you
don't
mind,
I'm
going
to
tack
a
link
into
our
appendix
on
how
that
works,
to
get
right
to
that
page.
F
G
H
Man,
so
we
have
this
very
long,
standing
issue
that,
while
you
can
use
when
you
do
make
and
make
test
or
whatever
you
know,
keep
it-
you
can
use
podman
or
docker
for
quite
some
time
now,
but
if
you
do
make
last
app
and
make
funk
testers
on,
you
still
need
docker,
and
I
spent
some
time
on
this
topic
lately,
because
it
could
really
help
us
to
make
final
progress
there
and
podman.
H
H
It's
not
yet
merged,
but
we
are
basically
able
to
just
talk
to
podman
through
the
docker
api
and
then
start
our
clusters
through
that,
apart
from
taking
away
the
need
that
people
install
docker
locally
if
they
have
partner
already,
this
will
also
help
a
lot
with
ci,
because
podman
allow
is
demon
less
and
therefore
it
has
features
for
sharing
image
repositories
which
the
docker
demon
does
not.
This
will
allow
us
to
have
very
nice.
H
Caching,
between
different
bot
lanes
on
end-to-end
testing
will
help
us
with
enter
and
test
sea
load.
For
in
order
to
that,
that
was
the
point
during
the
unfoldment
in
google
ti.
Are
there
any
questions
or
comments?
Did
you.
A
Say
that
sorry
go
ahead,
go
on,
did
you
say
podman
3.
A
I
H
I
am
not
sure
if,
to
be
honest,
I
have
no
clue.
I
have
to
check
on
github
if
portman
3
is
released
at
all.
It's.
H
Three,
zero
one
is
released
since
first
and
honestly,
I
didn't
check
if
it
was
released
already.
I
just
went
with
the
assumption
that
latest
master
has
the
best
stock
api
support,
which
turned
out
to
be
true,
and
we
did
some
issue
reports
there.
Department
people
were
great
fixing,
fixing
them,
and
then
I
was
able
to
do
that.
A
H
H
At
least
portman
three,
so
I
just
started
portman
three
with
the
with
the
socket
support,
and
it
even
worked
great
for
me
there
to
run
directly
with
the
docker
binary
against
the
spot.
Man.
I
H
I
mean
it
means
that
for
storage
you
will
have
to
use.
How
is
it
called
fuse
overlay,
a
storage
back
end,
which
is,
of
course
not
as
performant
as
overlay
to
withdraw,
but
it
should
at
least
be
much
better
than
any
other
of
the
emulation.
D
Okay,
applause
on
that
I
I
spent
quite
some
time
with
that
when
I
joined
and
with
pop
into-
and
there
were
still
too
many
problems
or
I
didn't
get
it
to
work-
I
I
got
cluster
up
but
plus
the
sync
never
worked.
So
that's
awesome
to
see.
D
H
So
there
are
still
only
two
issues
remaining
one.
Is
that
filtering
applying
a
filter
when
you
get
the
list
of
containers,
the
filter
was
ignored
and
the
other
part
was
that
the
pods
section
when
you
read
when
you
fetch
the
parts
the
report
section
was
not
filled
that
were
the
only
yeah.
H
Yeah,
so
for
for
a
few
days
now,
our
end-to-end
tests
don't
need
to
build
the
image
anymore,
because,
with
the
introduction
of
bazel
dnf,
we
can
now
build
all
containers
and
everything
and
use
the
shed
libraries
from
rpms
without
having
to
install
them.
So
for
just
make
cluster
up
cluster
sync
and
make
funk
tests.
There
is
no
need
for
the
builder
image
anymore
for
as
long
as
you
have
a
new
enough
glc
version
which
helps
us
to
just
not
have
to
load
or
pull
the
build
image
at
all
for
end-to-end
tests.
H
That
will
help
us
in
the
long
run.
Also,
this
report
on
getting
read
from
read
from
the
builder
image
also
means
that
steps
like
mac
generate
and
make
tests
and
make
funk
test
can
then
be
executed
without
the
image
which
makes
everything
more
sniffing,
more
light,
which
there
is,
by
the
way,
no
intent
to
remove
the
image
in
any
way.
This
is
really
just
to
have
better
hermetic
and
predictable,
builds
and
save
ci
load
and
make
maintenance
tasks
easier
like
not
having
to
keep.
H
Code,
generators
for
kubernetes
in
sync,
with
with
the
used
golang
version
or
stuff
like
this.
This
is
then
all
controlled
by
the
go
mod
file,
and
this
will
help
us
also.
A
A
It's
been
there
for
quite
a
while.
Now
I
have
a
meeting
with
the
red
hat
office
of
open
source
program
and
as
long
as
he
doesn't
have
any
more
concerns,
I'm
going
to
go
ahead
and
merge
that
please
review
it
add
comments
where
you
think
they're
needed
and
anything
in
the
future
will
just
create
a
an
issue
and
adjust.
A
A
Okay,
sam
first
meeting
and
your
first
topic.
C
So
yeah
this
kind
of
stems
around
some
conversations
I
had
in
slack
with
some
people
within
keyboard
community
and
the
keyboard
dev
channel.
I
you
know
I
ran
into
this
new
with
keywords
where
it
just
wouldn't
run
on
anything
cannot
coil
or
I
think
they'd
be
in
as
well,
and
this
was
because
of
a
missing
se
linux
resource-
and
you
know
everyone
knows
they
use
app
armor
versus
elegance
on
that
side
of
things.
C
So
you
know,
obviously
it
was
resolved
that
heroes
merged
and
it
was
fixed
quickly,
but
you
know
worst
case
scenario.
If
I
had
you
know,
say
there
was
some
enterprise
customers
and
they
were
running
production
and
they
were
running
conical
and
they
updated
okay,
no,
their
their
vm,
suddenly
stopped
working
after
this
family
figure
out.
So
I'm
I'm
wanting
to
you
know,
contribute
to
some
ci
testing
or
something
similar
to
where
we
can
kind
of
broaden
the
scope
of
this.
But
I'm
also
trying
to
learn
what
you
already
have
in
place.
C
So
I
can
know
how
to
contribute
properly,
because
I'm
you
know,
like
I
said
new
to
this
community.
While
I've
been
using
the
tooling
for
a
long
time-
and
I
love
it,
I'm
trying
to
figure
out
the
best
way
to
start
a
give
back
and
I'm
really
good
with
ci
cd.
So
I
figured
this
is
the
best
place
to
start.
H
So
I
think
yeah
we
were
talking
yeah
federico
is
also
on
the
call
here
I
think
federico
is
so
now
for
a
few
weeks
or
ci
leads.
So
I
think
federico
is
also
good
contact
to
talk
to-
and
I
think
regarding
to
the
issue
which
you
had
testing
keyboard
completely
fast
on
maybe
a
little
bit
farfetched,
but
I
could
think
that
we
could
start
we
could.
H
If
you've
spent
half
the
time
we
could
come
up
with
something
like
doing
limited
testing
on
different
operating
systems
with
periodic
job
or
something
so
that
we
have
basic
knowledge
that
we
don't
pay
break
core
assumptions.
We
have,
we
even
have
a
conformance
test,
suite
a
very
small
one
which
can
be
run.
So
maybe
you
can
also
add
your
voice
here.
We
could
add
something
like
starting
a
very
simplistic
kubernetes
cluster
on
on
ubuntu
and
just
run
the
10
performance
tests,
which
ref
right
now
just
to
be
sure
on
a
regular
basis.
K
C
C
Yeah
for
sure
I
want
to,
I
mean,
if
anything,
I'm
not
sure
how
you
work
with
testing.
I
do
a
lot
with
ltsp,
so
it
gives
me
a
quick
experimental
cluster
to
quickly
split
up
and
speed
down
phd
engagement,
proxy
and
other
things.
So
I
can
contribute
resources
locally.
C
I
can
provide
templates
and
we
can
definitely
just
start
with.
You
know
a
quick
convincing
page
and
work
across
generic
os's
I'd
be
more
than
happy
to
start
contributing.
There.
H
And
I
would
see
various
ways
here
like
like:
we
can
probably
deploy
on
various
external
systems
or
do
it
similar,
like
we
do
with
keyboard
ci.
I
guess
it
also
depends
a
lot
on
on
what
you
can
contribute
and
and
how
permanent
your
engagement
would
be,
but
to
have
to
to
have
at
least
a
pc
up.
Pretty
quickly
is
probably
not
much
of
an
issue
if
it's
just
periodic
that
doesn't
block
anything.
Maybe
a
couple
already.
C
A
And
that
takes
us
to
the
end
of
the
open
floor
about
one
last
call
for
open
floor
items.
A
A
Okay,
we'll
move
on
from
there
then
would
who
would
like
to
run
bug
scrub
this
week.
J
J
Okay-
and
I
will
mark
it
as
accepted
meaning
we
want
if
this
gets
resolved,
we
would
be
happy,
but
it's
it's
not
the
highest
priority,
and
do
you
know
what
your
igor
is
he's
not
looking
to
it,
because
it's
not
the
high
spirit,
sorry
for
playing
yeah,
it's
not.
H
It's
it's:
it's
not
a
problem
for
keyword.
If
you
just
build
it
and
deploy
it,
it's
more
a
problem.
When
you
want
to
build
your
10
language
launcher
directly,
then
you
have
to
be
aware
that
you
need
to
put
in
the
vertender,
also
liberate
lip
rpm
dependencies
and
so
on,
just
to
have
the
key
more
user
present
there
or
somehow
differently
at
the
cumulus,
and
it's
not
really
necessary.
I
mean
we.
We
actually
want
to
look
invert
launcher
for
the
qemu
user
in
principle.
In
practice
it
doesn't
make
any
difference.
J
A
Doesn't
that
have
to
do
with
expanding
the
file
system?.
J
J
C
When
I
use
vert
control
to
upload
the
image
to
a
pvc,
that's
larger
than
the
cloud
image,
it
usually
takes
care
of
care
of
expanding
the
file
system.
So
I'm
not
sure
how
they've
done
this
here
or
if
they're
reading
from
the
file
system
or
at
cfs
10..
Sorry,
not
the
devices.
The
disks.
J
Okay,
so
I
will
ask
her
and
what
your
the
discussion
is
copied
here.
I'm
blind.
J
K
J
J
J
Do
we
have
virtual
machine
snapshots
documented
in
the
user.
A
A
I
am
checking
on
that
as
we
speak,
it
sure
doesn't
look
like
it
isn't
that
driven
by
the
virtual
disk
type.
A
Like
whether
or
not
raw
versus
q,
cow
2
gets
used.
H
So
the
the
snapshot,
restore
and
and
create
he's
talking
about
is
a
feature
which
is
done
on
the
cluster
level.
It's
not
related
to
cue
card.
If
that
was
the
question,
it's
more
like,
I
think,
he's
looking
for
how
to
how
to
enable
the
feature
at
all
on
the
cluster
right.
G
J
H
A
J
J
J
J
All
right,
the
next
one,
vmi's
failure
against
light
migration
when
performing
workload
updates.
J
Thanks,
okay,
so
let's
skip
it
word
launchers
start
filling
in
the
version
of
earth.
039.
J
J
I'm
just
saying
if
it's
being
handled
and
it
seems
like
a
little
taking
care
of
it.
J
Oh
awesome
thanks,
so
I
will
give
it
a
neat
info
since
there's
still
some
exchange
happening.
H
H
J
All
right,
the
boy
will
take
care
of
it
right.
L
J
Okay,
okay,
so
just
disable
a
c
linux
or-
and
I
plot
one
of
the
other
unit
tests
results-
are
nci.
What
kind.
G
H
Well,
it's
about
disabling
it
because
he
thinks
it's
not
that
much
of
load
and
may
probably
help
showing
flakes
or
something.
J
H
H
We
can
stop
doing
anything
at
all
in
this
case,
because
we
can
configure
bases
slightly
different
and
it
can
then
not
can
then
just
basically
directly
go
to
go
to
the
end
and
say
everything
is
good
or
just
download
for
the
two
tests
where
something
changed,
something
the
rest
just
isn't
the
cache,
so
it
should
be
planning
to
be
more
aggressive.
So
I
guess
I
have
to
talk
to
vasily
and
see
where
we
can
proceed
here,
because
that's
kind
of
that
the
two
requests
are
contradicting
put
it
this
way.
H
J
For
taking
this
vermont,
all
right
still
have
eight
minutes
so
end
of
discovery
failed.
H
J
It
should
be,
would
you
mind
linking
it
on
the
issue
cool?
I
can
link
it
in
the
chat
to
make.
A
There's
also
an
activity
on
a
user
guide
regarding
gpus
and
mdev.
A
Just
this
past
week,
I've
been
working
with
a
contributor
to
get
that
in
place.
J
Awesome:
okay,
there
was
this
one.
It
was
also
erased
on
the
on
the
mailing
list
by
me.
J
So
we
had
we
had
this
issue
where
we
wanted
to
merge
bug
fixes
first
for
the
flaky
ci,
but
instead
the
merge
queue
was
taking
other
pr
that
were
not
so
related
to
the
to
the
issue.
So
the
the
purpose
enhancement
on
this
issue
is
to
have
a
mechanism
that
allows
us
to
say
set
the
priority
of
a
pr
and
say
this
one
should
get
first,
that's
like
the
top
priority
pr
for
emerging,
like
makes
sense.
H
So
I
know
that
this
on
the
last
incident
in
the
last
two
weeks
we
had
over
three
weeks.
We
had
huge
issues
in
the
eye
and
in
the
past
mostly,
I
resolved
that
by
just
manually
merging
important
prs
when
they
were
ready
and
just
this
time
I
didn't
do
it
in
general.
I
was
fine
with
that.
If,
if
maintainers
just
do
manual
merges
I'm
not
sure
if
we
need
the
process,
I
mean,
maybe
it
would
have
helped
this
one
incident.
H
But
before
that
we
had
a
lot
of
other
incidents
and
we
resolved
it
without
it.
So
I'm
not
yet
convinced
that
this
process
is
needed.
I
mean
to
somehow
cut
out
other
pr's.
I
mean
I
mean
writing
something
to
people
deaf
and
insisting
that
maintainers
look,
which
did
not
happen
that
this
time
so
fast
of
course,
makes
sense.
J
Okay,
so
I
mean
doing
this
makes
sense,
but
we
don't
it's
not
the
first
thing
that
we
should
automate.
Did
I
rephrase
it
correctly
yeah.
H
J
J
J
Right,
do
we
have
eddie
on
the
call,
and
do
you
want
to
yeah?
Do
you
want
to
comment
on
this
or
you
do
agree
with
what
has
been
said.
G
H
I
would
prefer
that
we,
as
I
said,
and
we
didn't
do
it
this
time,
which
was
from
my
perspective,
a
mistake,
but
we
did
it
a
lot
of
times
before
and
it
worked
good
enough.
That's
why
I
would
not
go
for
a
more
strict
process
right
now,
but
if
we
will
do
it
again,
I'm
happy
to
discuss
this
again
from
my
perspective,
but
if
others
want
to
have
it
now
too,
you
can
also
discuss
it
now
or
in
a
better
form
than
this
crease
call.
J
Okay,
so,
but
then,
like
I'm
optimistic
now,
I'm
feeling
that
we
won't
need
this
too
often
to
so
it's
immersed
in
automation.
So
does
anybody
opposed
to
close
this
and
eddie.
J
J
Okay,
I
will
do
it
offline,
so
we
can
use
the
remaining
two
minutes.
J
And
chris
here
he
made
me
think
that
we
may
be
should
look
at
the
user
guide
as
well
in
this
form
to
see
if
there's
an
issue,
that's
immersed
conversation.
Here
I
mean
if
there
are
any
just
seven
of
them.
Would
you
like
to
give
shout
outs
to
any
of
these
to
get
some
input.
A
How
does
that
happen
in
today's
day
and
age
apologize?
I
had
to
mute
because
my
family
was
going
crazy
right
outside
my
doors,
yeah,
I'm
the
team
of
one.
So
we
have
some
some
issues
with
user
guide
for
all
of
our
non-us
contributors.
A
Utf-8
characters
have
been
getting
into
into
user
guide
and
that
causes
our
html
proofer
app
to
crash.
So
I've
had
to
go
around
and
and
find
solutions
for
those
characters,
and
so
I'd
like
to
get
this
automated,
but
I
don't
want
to
create
too
many
surprises
for
people
on
on
their
keyboard
character
sets.
C
Enhancement,
it's
something
that
would
help
me
get
more
familiar
with
the
automation
and
testing.
I
would
be
more
than
happy
to
kind
of
look
into
that.
I
mean
it
sounds
simple
enough
on
paper
to
where
I
can
just
check
on
a
lot
of
markdown
files
and
say
what
kind
of
format
I'm
using.
I
could
try
for
it.
I
would
be
more
comfortable
if
I
just
make
sure
I
had
something
to
pair
with,
but
you
know
something
I
can
definitely
work
on.
A
Yeah
sounds
good.
Sam
I've
been
using
this
great
big
long,
gretz
command
to
find
utf-8
characters.
I'm
sure
there's
better
ways
to
to
go
about
that.
A
Let's
have
a
chat
and
slack
and
we'll
probably
get
together
on
on
on
the
zoom
call
and
I'll
give
you
a
big
overview
on
how
the
tooling
works
perfect
understood.
J
All
right,
we
are
out
of
time,
but
we
we
got
some
work
done
today.
So
thanks
all.
A
Yeah
great
lots
of
participation
this
week.
That
makes
me
a
happy
cooper
project
organizer,
and
I
will
let
you
guys
go
have
a
good
week
and
we
will
see
you
around
the
project.