►
From YouTube: kubeadm backlog grooming 2019-07-25
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
A
Me
the
master,
so,
let's
see
what's
new
first
and
then
we
can
go
through
where
we
were
last
time.
So
this
is
a
ticket
a
lot.
Yesterday,
I
started
seeing
a
pattern.
People
are
having
problems
with
load,
both
balancers
and
I,
know
the
immediate
reaction,
like
a
load,
balancer,
honest
and
thorough
scope
for
cab,
a
diem
that
is
true,
but
we
don't
have.
We
don't
define
the
requirements
of
our
greatest
concern
in
terms
of
a
load
balancer,
for
instance,
I
added
some
items
here,
like
we
don't
define.
A
A
Basically,
you
have
to
provide
feedback
on
what
they
are
experiencing
in
different
compromises
like
Asia
is
quite
broken.
I
need
promised
Savannah
boosters
to
comment
on
this
I
like
Jason,
provided
a
very
good
outline
of
what
load
balances
aid
operations,
but
overall
I
don't
think
we
should
write.
Oh
sorry,
quote
provider
load
balancer
details
the
docs.
A
Instead,
we
should
probably
like
have
a
list
of
requirements
like
what
do
you
need
from
a
load
balancer
in
a
call
provider,
so
I
started
started,
seeing
problems
in
the
cube
ATM
channel
on
the
VMware
internal
slack
and
also
Stack
Overflow,
and
basically,
people
are
trying
different
stuff
and
some
of
the
load
balancers
simply
don't
work.
For
instance,
if
you
have
I,
don't
have
helping
routing
it's
a
problem,
so
this
is
a
new
item.
Hopefully,
Richard
can
actually
help
with
documenting
parts
of
this.
Anybody
has
any
comments.
D
It
will
prevent
some
headaches
and
also
we,
we
bacon
throughout
a
bunch
of
our
things,
there's
some
ideas
about
going
from
1
to
N
for
AJ
and
it
might
be
beneficial
to
like
you
know
before.
Maybe
it's
something
like
planning
your
deployment.
We
actually
have
this
and
other
Docs
that
I've
seen
or
at
least
used
to
here,
like
it
could
be
listed
out
as
a
set
of
like
preconditions
before
you
actually
get
to
kuba
DM
deployment
that
you
should
think
about
these
things
as
constraints
for
planning
your
deployment.
A
C
This
was
a
user
that
was
asking
the
two
that
the
control
plane
phase
basically
does
not
create
a
manifest.
Is
that
my?
If
the
manifest
are
ready
exist
on
the
disk
and
because,
basically
the
user
R
is
provided?
He
wants
to
provide
in
its
own
manifest
because
they
are
doing
Cedeno.
If
I
remember
word,
but.
A
So
this
is
a
part
of
the
problem.
I
tried
to
explain
yesterday,
but
I
think
I
failed
when
you
are
presenting
the
customized
cap.
A
A
A
C
C
C
A
C
C
It
would
be
beneficial
for
the
PR
for
the
issue.
At
least
you
mean
if
we
can
comment
on
on
this
issue.
At
the
end,
this
is
going
to
be
fixed
at
the
anisa
T
V
and
the
link
to
the
ad
city
issue
that
I
saw
I,
know
I,
don't
remember
if
I
saw
in
Guam
was
locked
or
in
public
dock
gosh
I
don't
have
the
link.
Let
me
check
this.
C
D
D
Yeah,
there's
well
there's
a
way
to
work
around
this
anyways
I,
don't
know
if
it'll
work,
you
have
to
change
the
list,
order
applied
to
each
of
the
Supreme
NCD
notes
and
that
that
prevents
the
province's
problem.
There's
a
worker,
no
I
do
believe
that
there
was
planning
to
get
this
done
for
116,
I'm,
gonna
repaint
him
again.
I
do
know
that
DIMMs
reached
out
there's
patches
available
for
SD,
but
we
need
to
merge
that
through
KK
and
we'd
also
need
to
vendor
an
update
at
CD.
A
C
B
A
D
B
B
A
A
A
C
A
B
D
B
A
A
A
D
D
A
test
it
with
a
different
location
to
verify
to
see
if
the
settings
work
properly.
Even
if
it's
not,
you
can
check
to
make
sure
that
the
overrides
through
the
configure,
all
working
as
they
should
be,
and
that's
like
a
first
step
before
closings
I-
think
that's
that
good
diligence
is
probably
reasonable.
D
C
C
A
D
There
is
a
DNS
discovery
capability
for
Etsy
team
members.
I'd
be
surprised
if
we
could
support
that,
given
our
configuration
unless
it
was
like
totally
external
at
CD
and
somehow
plumb
that
through
that
would
have
to
be
how
how
you
configure
it.
But
that's
like
a
weird:
it's
like
a
bootstrap
problem
right
because
you
need
it
need
a
special
ordering
to
make
this
even
possible,
so
like
you'd,
have
to
plumb
through
I
know
how
it
works.
You'd
have
to
plumb
through
the
members
from
DNS
discovery
into
the
API
server
arms,
and
it's
just
I.
C
I
think
they
did
pretty
well
to
the
this
comment,
divide
the
program
to
pass
a
create
and
estimate
the
city.
Cluster
user
Cooper
mean
to
create
a
class
that
they
use
and
estimate
the
city.
Let's
seal
the
way
you
can
do
it
I
think
it
should
close,
it
say
sure,
I'm
foreclosing,
there
are
not
no
feedback.
C
C
D
We
could
document
this
as
a
proposed
configure
like
part
of
the
setting
up
your
deployment,
and
if
we
have
part
of
the
external
entity
configuration
you
could
possibly
add
a
blurb.
That
basically
says
you
know,
so
this
is
not
a
fully
dynamic
system.
If
you
want
to
sort
of
externality
DNS
rally
point
stuff,
then
you
need
to
set
us
up
music
tournament.
Cd.
C
D
A
A
But
we
never
I
know
these
documents
that
we
never
talk
about
discovery
service
and
that
it's
not
supported
for
stacked
I
think
we
should
add
a
sentence
somewhere,
possibly
in
the
high
availability.
Doc.
Let's
say
hey.
If
you
want
to
use,
you
know,
name
name
server
for
for
the
endpoints
of
a
CD.
You
should
use
external
HDD
or
apply
is
possibly
not
mentioned
this
hack
that
users
drink
nobody.
It.
C
A
D
D
C
A
C
B
A
C
A
Marker
this
one
initiation
was
a
so.
This
is
Lucas
also
planned
to
duplicate
to
this,
but
I
closed
it
originally
Rafael
reported.
You
know
this
is
an
artifact
of
the
face
level
where
we
have
to
specify
the
face
up
on
certs,
but
we
also
need
experimental
upon
certain,
which
is
now
a
broad
search.
So
we
have
not
much
limits
a
broad
search.
The
confusion
here
is
that
we
need
an
extra
part
pass
to
the
command
that
has
a
verb
in
the
name
as
well.
A
D
D
A
Can
give
some
feedback?
There
is
a
guy
who
works
for
a
pivotal
called
Ben
moss
and
he
is
a
colleague
of
people
like
Leah
who
attempted
to
contribute
to
comedian
and
they
walked
our
backlog.
We
saw
some
like
lost
cue
tickets.
They
started
sending
Pierce
immediate
without
discussion.
We
closed
some
peers.
A
There
helped
with
writing
the
notes
of
you
know
the
notes
during
the
meeting
taking
notes,
but
then
explained
to
me
that
this
was
also
not
a
guy
who
attempted
to
join
question
about
developments
from
Utah,
but
they
are
really
discouraged
and
I.
Don't
think
he's
read
the
people
I
think
it's
a
complexity
of
the
projects
and
maybe
the
whole
development
model
is
really
not.
D
What
the
it's
more,
the
development
model
and
pivotal
in
particular,
so
I,
wouldn't
necessarily
take
I,
wouldn't
extract
pivotal
as
a
general
user
story
because
of
how
pivot
operates,
they
rotate
people
and
they
rotated
them
off.
So
unless
you
actually
get
engaged
with
the
SIG's
and
understand
the
flow
and
the
people
yeah,
and
especially
if
they
rotate
people
in
and
out,
because
really
difficult,
because
by
the
time
they're
done
or
the
rotation
and
by
the
way,
Ben
moss
and
léa
léa
are
both
rotated
out.
D
A
Ben
is
I,
think
I'm
working
on
sig
windows
in
his
free
time,
I'm,
not
sure
anyway,
I.
Don't
it's
really
complicated,
then
to
the
question
of
how
is
a
complete,
we
should
my
opinion
is
that
we
should
have
got
more
and
more
videos
explaining
different
parts
of
debate.
Am
the
testing
process.
We
can
explain
kinder.
We
can
explain
how
we
work
on
features
during
the
cycle.
Things
like
that.
If
we
have
a
bunch
of
videos,
we
can
just
do
this.
Another.
D
Thing
that
might
help
to
is
like
we
have
an
incredibly
user
base
like
it's
just
just
by
posting
in
the
Kebede
ian.
I
don't
really
go
to
the
convenience
lecture
anymore,
because
it's
like
a
fire
hose,
but
just
by
posting,
a
cuvette
DM
slack
channel,
saying,
like
hey,
we're
looking
for
new
contributors,
if
you're
interested
here's
a
video
for
onboarding
that
might
be,
or
if
you're
interested
in
becoming
a
new
contributor
to
sit
closer
lifecycle,
here's
non-boring
video
that
alone
might
be
a
first
pass
stuff
and
I
can
take
as
an
actually.
C
The
other
problem
that
I
see
is
that
in
the
backlog
we
are,
we
have
a
lot
of
nasty
issues
which
are
not
very
well
documented
or
not
easy
to
address.
They
are
there
for
forever,
and
so
when
people
arrive,
Cooperman
is
easy
a
so
there
is
no
area
and
you
see
flow
in
inflow
of
new
feature
or
new
new
blogs
that
we
ever
so.
That
is
a
kind
of
stuck
in
it
or
not
to
call
it
it.
We
are.
We
are
getting
stable,
so
I.
A
So
we
have
to
remember
another
big
problem
is
that
if
you
are
not
on
a
salary
to
work
on
something
that
is
ready
to
to
cube
ATM
or
coaster
API,
it's
really
quite
special
case
for
people
to
join
these
projects
for
bid
show
you're
one
example.
There
are
not
many
people
like
you,
Lucas
is
I,
guess
also
an
example,
but
he's
kind
of
more
involved.
But
it's
really.
We
don't
have
contributors
that
do
this
in
the
three
times,
so
it's
tricky
well
I
think
there.
D
Has
to
be
a
subtle
motivation
like
internal
motivation,
if
they're
using
PDM
to
solve
their
problems,
if
it
just
works
from
90%
of
their
use
case,
they're
very
inclined
not
to
care
right,
but
I
think
finding
those
people,
maybe
in
the
selection,
who
are
having
problems
for
our
developers.
You
know
we
can
help
guide
them
along
the
way,
but
typically
I
don't
have
the
resources
or
bandwidth
to
identify
those
people.
Yeah.
C
D
A
lot
of
the
issues
we
have
right
now
are
very
thorny
right.
It
required
deep
knowledge,
but
if
we
were
to
do
a
Code
walkthrough,
that
might
be
a
good
way
because
there's
a
lot
of
low-hanging
cleanup
work
that
could
be
done,
but
we
haven't
really
enumerated.
That
stuff,
like
Richie,
was
talking
about.
Yes,.
C
D
A
A
A
Let
me
let
me
get
back
to
this
offline
after
the
meeting.
Okay.