►
From YouTube: 20190220 kubeadm office hours
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
A
First
up
was
this:
PR
fix
config
over
rights
for
our
PMS.
This
is
existed
for
a
long
time
seems
pretty
trivial.
I
could
double-check
this
in
the
sinus
to
myself,
your
hand-signed
I
could
take
a
look
at
that
one
later
convenient
config
precedents.
I
know
that
we
had
a
bunch
of
ongoing
conversations
about
this.
One.
B
C
C
C
There
is
the
risk
to
leftovers.
One
is
to
a
low
user
to
provide
a
certificate
manually
and
dizzy.
It
was
decided
with
JSON
for
the
following
automation
in
okay
and
the
second
is
is
to
add
more
tests
for
the
init
part,
I
agreed
with
Yaga
to
move
the
testing
and
separated
PR.
So
we
can
unblock
the
join
part
for
the
join
part
down
lower
said
we
are
already
in
good
shape.
There
is
a
PR
from
Rafael
and
is,
if
you,
if
or
your
JSON
with
take,
can
take
a
take
a
look.
C
C
We
didn't
implemented
the
PR
yet
I
have
pink
at
today
and
we
will
decide
if
someone
who
should
take
over
soon
in
the
last
bit
that
the
weather
working
is
to
make
some
improvement
to
rossette
related
to
hae.
One
is
to
clean
up
the
cluster
status
when
we
later
master
node
and
there
is
a
PR
which
is
there
from
some
time
now,
but
it
is
if
I
remember
what
the
outer
was
his
mark,
but
it
is
not.
C
He
is
not
working
on
this
PR,
so
I
pink,
also
mark
asking
if
we
plan
to
tell
to
start
this
effort
or
someone
as
jumping
and
the
last
one
is
the
removing
the
TC
team
member
from
the
TCD
cluster
when
we
Rosetta
a
master
node
and
this
PR
was
his
merger
is
going
to
merge
sooner
if
something
that
we
have
to
test
the
hell
out
of
it.
In
my
opinion,
but
let's
get
this
map
and
that's
also
my
purse.
C
A
We
are
currently
off
track
with
regards
to
artifact
publishing,
but
part
of
this
is
not
a
problem.
We
have
to
interlock
with
the
release
team
so
that
they
can
actually
do
artifact
generation.
That's
comes
from
kak,
so
and
because
there's
no
one
on
the
other
side
to
whom
this
work
we
need
to
get
people
spun
up,
leaves
gonna
get
proposal
merged
that
we
all
agree
to
for
115
and
they
get
the
other
pieces
in
place
by
the
beginning
of
115.
So
that
way,
the
release
cycle
can
actually
go
smoothly.
A
That
work
will
mercs
current
work,
good
land
or
cannot
later
it's
not.
It's
made
sure
it
was
not
on
a
blocking
path
for
us.
So
if
it
doesn't
merge
this
fine,
the
proposal
is
pretty
clear
and
what
we
want
to
accomplish.
So
we
could
do
that
in
the
beginning
of
115
and
it
will
be
okay,
but
the
actual
release
work
and
getting
someone
on
the
release
site
own
that
that's
that's,
gonna,
be
a
challenge
for
me.
In
probably
the
115
cycle.
C
A
A
We
still
need
to
do
that
in
150
plus
time
for
you
to
actually
enable
that
separate
test
suite
we
created
so
we're
we're
verifying.
You
know
the
actual
component
bit
that
we
update
for
the
hosts
is
updated
on
the
config
map.
You
know
as
well
as
all
the
all
the
other
information
that
we
expect
is
part
of
it.
C
C
D
C
C
There
is
no
are
clear
at
a
clear,
defining
way
in
how
to
add
the
second
kubernetes
version
to
the
cluster,
because
and
now
kind
can
do,
create
the
cluster
and
create
a
cluster
ready
for
upgrade,
but
for
upgrading
I
need
to
bring,
in
my
my
notes,
all
the
depth
and
the
images
for
the
target.
A
version.
C
C
A
C
C
So
commit
program
was
an
experiment,
a
set
of
script
that
we
developed
and
we
put
in
the
konami,
mean
report.
It
was
designed
to
support
the
developer
experience
so
to
help
people
to
get
Cooperman
tested
and
to
test
company
in
PR
and
whatever
we
decided
to
duplicate
this
set
of
script,
it
was
based
on
a
vagrant
Python,
whatever
so
in
favor
of
kind,
but
along
the
way
we
understood
that
these
specific
skills
so
supporting
kuben
mean
or
seed
cluster
lifecycle
developer
in
their
user
experience.
C
In
the
main
kind
project,
so
what
we
discussed
with
Ben
was
to
introduce
her
and
separated
command,
that
use
kind
as
a
library
and
and
put
on
top
of
kind
a
set
of
additional
command
that
helps
the
developer
to
to
make
their
daily
job
of
to
test
PR
to
get
quickly
up
a
multi
master
cluster.
That
is
something
that
in
I
have
many
many
new
contribute
or
are
paying
me
how
I
can
do
these
locally,
because
I
don't
apply
a
cloud
provider
and
so
on,
and
so
on.
C
C
A
A
Think
that
seems
totally
fine
if
we
wanted
to
put
a
PR
against,
because
we're
going
to
clean
up
the
Canadian
repo
if
we
put
a
PR
up
against
Covidien
repo,
but
make
sure
we
highlight
that
you
know
experimental
something
with
that.
So
that
way,
you
know
that
that
people
don't
rely
too
heavily
on
it
and
we
can
decide
at
a
later
date.
People
can
kick
the
tires
on
it.
I
think
that
would
be
a
highly
useful
thing
in
a
highly
leveraged
thing.
My
long-term
goal
has
always
been
to
kill
a
local
up
cluster.
A
C
A
C
Okay,
so
we
basically
kinda
is
a
library
in
business.
Don't
find
what
gives
you
gifts
all
the
kind
come
and
also
build
the
create
delayed
and
so
on,
plus
some
specific
comment
which
are
copy
and
and
do
I
will
show
the
Manaus
so
I
put
I
I
repeated,
oh
cool
I
get
I
created
an
entry
for
all
the
kind
convert,
because
I
find
unknowing
to
switch
from
one
condom
comment
to
another.
C
Basically,
what
gives
you
you
can
you
can
create
your
cluster
and
it
gives
you
one
flag,
which
is
important.
That
is,
allows
you
to
prevent
prevented
the
car
library
from
studying,
kubernetes
and
some
other
flag,
which
are
then
a
forest
sudden.
That
that'll
was
you
to
change
the
topology
of
your
cluster
quickly,
like
the
number
of
worker
node,
and
this
is
the
first,
the
first
div,
the
the
only
difference
in
the
in
the
create
part
of
the.
C
C
Concrete
file
already
set
up,
but
cool
meanwhile
was
not
executed,
and
this
is
your
package
just
yes,
but
it
is
also
for
testing
the
Ku
band,
meaning
were
flow
because
kinda
executed
the
common
meaning
e4
flow
overflow.
For
me,
I
don't
have
a
chance
to
change
to
change
it.
I
don't
take
this
chance
to
set
the
they
were
both
flag,
I
I,
don't
I
cannot
see
the
logs,
so
kind
execute
the
you
need
workflow,
but
it
I'd
the
output
from
me.
It's
almost
more
phases.
C
Yes,
but
kinda,
but
the
problem
that
maybe
one
day
we
will
get
feces
there.
But
it's
not
it's
not
when
the
question
is
when
so,
basically
now
I
have
a
container
that
are
like
too
small
to
move
too
much
to
empty
machines,
and
now
I
I
can
manage.
I
can
work
on
this
machine
using
or
or
using
docker
exact
because
they
are
docker
container
like
SSH,
SSH
or
I
can
use
the
kinder
offered
me
office
me.
Some
proper
topology
aware
wrapper
on
top
arrived
deck
or
copy.
C
For
instance,
I
can
do
kind
exec
on
all
the
nodes
and
then
I
give
a
give.
A
comment
could
mean
version
and
checking
which
version
is
stolid
on
on
all
the
the
nodes.
Another
things
that
I
can
do,
I
can
do.
I
I
mean
I
I
was
testing
a
PR,
so
a
local
admin
binary
built
on
on
my
machine
and
I
want
to
copy
it
inside
the
container,
because
I
want
to
test
that
these
are
this
release
so
I'm
doing
the
copy.
A
C
Yes,
this
can
be
I
can
move
these
also
in
in
the
in
the
unit
part,
but
out
works.
A
kind
now
is
that
you
build
everything
into
the
images
and
I.
Don't
want
to
do
that.
Maybe
I
don't
want
to
die,
build
the
image
every
time,
because
I'm
testing
I
mean
in
a
faster
test
and
build
the
tester
change
a
build
test,
and
so
on
so
it
it
can
be
usable
useful
to
do
coffee
instead
of
rebuilding
the
image
other
or
it
gives
me
some
the
possibility
to
execute
some
developer
friendly
war
flow.
C
And
I
can
I
can
set
use
phases
because
I
want
to
press
the
cobble
mean
in
it,
not
cooling
in
it,
but
colleague,
phase
1
phase,
2
phase
3
and
in
order
and
because
I
want
to
test
a
new
release
of
could
mean
I
want
original
regression,
and
this
is
ok
and
now.
I
is
downloading
an
image
because
something's
changed
but,
as
you
see,
is
running
phase
pre-flight
is
giving
you
the
full
output
of
cubed
mean.
D
A
I,
like
it
a
lot
I,
do
think
that
there's
no
reason
in
the
long
haul
why
we
couldn't
build
some
of
this
functionality
into
kind
itself.
I
do
I
do
think
that
some
of
these
workflows-
they
could
be
you
know,
listed
as
experimental
or
mischievous
developer,
listed
as
something
else
for
power,
admins
or
whatever,
but
nothing
you've.
Demoed
here
highlights.
A
A
B
A
C
Agree
with
you,
my
first
solution
was
to
was
to
get
this
in
kind,
but
I
understand
that
when
people
managing
kind
that
they
they
are
a
little
bit
scared
by
by
conflating
too
many
use
case
in
this
initial
phase
of
the
project,
I
accept
these.
The
the
what
I'm
proposing
here
is,
in
my
opinion,
something
which
is
really
experimental
and
the
the
long-term
of
goal
was.
These
two
may
move
almost
everything
to
the
other
side.
C
We
have
to
design
together
with
Ben,
maybe
I,
don't
know
some
plugin
and
mechanism,
something
that
makes
clear
different
from
water.
What
is
the
main
use
case?
What
what
is
para
mental
that
gives
people
the
the
good
the
the
necessary
speed
that
we
need
for
testing?
For
instance,
I
developed
the
yesterday
a
new
use
case
for
testing
joint
phases.
I
cannot
wait
three
weeks
well
because
otherwise
I
will
miss
the
deadline
for
the
cycle.
I.
A
Make
it
so
you
don't
need
to
justify
and
I
totally
get
it
I
think
I
think
when
we
go
through
a
planning
session
for
next
cycle.
I
will
open
an
issue
now
talking
about
based
workflow,
because
that
absolutely
makes
a
ton
of
sense
even
for
kind
itself,
and
I
can
open
an
issue
now
and
I
think
we
can
probably
PR
some
of
the
details
of
you
know
as
a
as
an
experimental
thing
to
Cuban
iam
repository
itself
right.
A
So
if
we
want
to
get
committee
and
rolling
and
get
people
unblocked
and
testing
out
these
ideas
as
long
as
it's
denoted
in
Cuba
diem
that
it's
experimental
and
that
some
of
these
features
are
prone
to
change,
I
think
you'll
find
the
I
do
long
term
what
to
get
to
the
stage
of
exactly
what
you're
doing
is
to
have
a
turnkey
multi
node
each
a
solution
for
a
local
developer
in
a
very
seamless,
clean
fashion.
So
that
way,
rubidium
developers
can
actually
develop.
C
Multi
node
is
supported
already
supported
because
I'm
using
in
kind
of
for
these,
and
what
is
not
supported,
is
upgrade
I.
Am
prototype
is
a
fee
for
a
grade
even
I,
don't
know
what
we
the
target
in
the
main
kinda
repo
so
for
unit
phases
for
joy,
enjoying
control,
plane,
join
phases.
I
am
I,
am
sure
that
I
can
copy
paste
this
piece
of
code
in
in
the
other,
in
the
main
kind
Reaper.
As
soon
as
there
is
agreement
on
this
yeah,
that's.
A
Fine
I
think
I'm
totally
supportive
of
this,
because
folks
gotta
get
rolling.
Let's
just
make
sure
that
we
mark
in
document
it
is
experimental,
maybe
even
have
a
folder
called
experimental
right
underneath
when
you
do
up
the
pr2
cube
ATM
to
get
it
in
there
just
make
the
folder
even
say
def,
experimental
or
development.
Yes,.
C
C
A
D
D
I
updated
the
kept
with
your
updates,
so
I
took
all
your
comment
and
moved
it
around
some,
so
I'm
up
for
a
meeting
about
that
whatever
and
I
so
I
actually
initially
made
the
cap
as
implementable.
So
I'm
like
halfway
done
with
that
as
well,
so
because
I'd
like
to
get
it
moved
to
implementable
as
soon
as
possible,
so
that
you
know
we
can
start
making
pr's
and
tickets
for
it.
You.
A
Can
make
you
can
make
PRS
right
now
it
doesn't
matter,
there's
no
prerequisite.
This
isn't
asked
by
Aaron
for
us
to
get
a
cap
in
place,
but
the
way
we
staged
the
idea
of
how
to
do
this,
it
doesn't
change
anything
we're
not.
You
know
we
can
build
this
totally
as
a
parallel
bridge
and
we
don't
actually
route
traffic
over
it
until
we
have
all
everything
else
done.
So,
even
if
we
miss
the
114
cycle,
that's
completely
fine
I
do
I.
Do
ideally
would
like
to
have
the
busy
in
place.
A
I,
don't
necessarily
want
to
route
traffic
across
the
bridge.
Yet,
but
as
long
as
we're
exercising
the
parallel
builds
and
folks
can
take
a
look
at
it
and
I
think
it's
important
and
then
in
115.
If
we
need
you
because
we're
we're
getting
kind
of
late
in
the
cycle
now
we
have
only
a
couple
two
days.
You
said
for
ratio
of
development
left.
B
D
C
D
A
Got
10
more
days
left,
so
the
the
idea
of
getting
everything
switched
over
is
slim
because
it's
not
gonna
have
enough
test
cycles
right,
but
that
doesn't
mean
we
can't
PR
and
get
some
of
the
pieces
in
the
build
apparatus
in
place
and
if
we
hold
it,
we
hold.
If
you're,
not,
we
wait.
Let's
just
get
the
PR
is
up
there
and
we'll
just
go
right.
Okay!
Well,
we'll
wait
to
tear
down
the
other
bridge
once
we
hit
like
15
at
this
point.