►
From YouTube: Kubernetes SIG Node 20210223
Description
Meeting Agenda:
https://docs.google.com/document/d/1j3vrG6BgE0hUDs2e-1ZUegKN4W4Adb1B6oJ6j-4kyPU
A
Okay,
now
recording
okay,
let's
restart,
say
good
morning
and
everyone,
so
a
quick
follow-up
before
we
start
everything
like
direct
cannot
join
today's
meeting
and
but
he
just
informed
me
earlier
this
morning
and
he
upload
all
the
past
of
the
video
or
record
video
and
the
current.
I
think
you
already
done
all
the
last
year,
video
and
the
current
instill
in
the
process
to
upload
this
year's
process.
A
I
think
by
before
today's
pacific
lunchtime
and
all
the
video
will
be
uploaded
successfully.
Okay,
so
you
need
to
start
our
regular
pro
agenda
circley
and
elena.
Maybe
you
want
the
update
of
the
pr
and
the
test
status
and
the
also
challenges
theaters.
B
I
can
start
with
a
number
of
pr's.
We
we're
still
growing
a
number
of
pr's.
I'm
not
sure
why
it's
happening
like
we
probably
don't
have
like.
I
mean
some
pr's
just
stuck
in
requiring
approvals,
and
maybe
we
don't
provide
enough
information
for
approvers
to
approve
it
fast.
B
Some
pr
stuck
with
reviewers
needed
so
yeah
just
need
to.
C
C
A
Yeah,
anyway,
you
want
to
write
a
screen
and
let
me
know-
and
I
will
go
ahead.
C
Okay,
so
here
is
the
the
node
triage
board
and
I
just
went
through
it
this
morning
in
advance
of
the
meeting
to
try
to
make
sure
that
I
went
through
everything
that
was
urgent
and
so
right
now,
I
think,
probably
one
of
our
bigger
there.
We
have
two
bottlenecks.
Basically,
we
have
the
bottleneck
of.
We
don't
have
enough
approvers
and
we
have
the
bottleneck
of.
C
We
don't
have
enough
reviewers
and
I
think
one
of
the
things
so
I
have
had
a
little
bit
less
time
to
review
things
the
last
two
weeks
and
I
think
that
that
has
been
felt
in
terms
of
impact
on
pr
growth.
So
we
currently
have
you
know
about
32
things
in
the
needs.
Reviewer
column.
We
have
40
things
in
the
needs,
approver
column,
that's
a
lot
of
stuff
and
of
course
the
they
don't
stop
coming.
So
if
you
are
interested
in
helping
out
with
this,
you
feel
free
to
reach
out
to
me.
C
If
you
have
any
questions
or
anything
like
that,
but
there
is
a
link
on
the
board
for
how
to
triage
this
stuff
on
the
community
wiki,
and
so
you
can
read
through.
Basically
how
do
I
know
when
to
put
something
in
the
triage
column?
How
do
I
know
when
to
put
something
in
the
waiting
on
author
column?
What
labels
should
I
put
on
prs
that
kind
of
thing
so
that
documentation
is
all
available?
C
And
so
long
as
you
are
a
kubernetes
org
member,
you
will
be
able
to
move
things
between
columns
and
apply
almost
all
the
labels.
You
need
to
be
an
approver
to
apply
an
approved
label
for
those
who
are
not
yet
org
members,
but
still
want
to
help
out.
You
can
still
go
through
stuff
in
the
triage
column
and
apply
priority
labels
and
kind,
if
they're
missing
for
some
reason
and
provide
feedback
initially.
So
if
that's
a
way
that
you
want
to
start
getting
involved,
that
is
an
option
but
yeah
really.
C
We
just
need
more
reviewers
and
more
approvers,
and
so
long
as
there
are
things
that
are
you
know,
I
just
saw
something:
got
moved
to
the
done
column.
So
thank
you,
someone
for
proving
something.
D
No,
I
just
did
that,
I
think
like
does
it
make
sense
to
maybe
have
a
periodic
like
a
review
meeting
or
something
if
people
want
to
just
get
together
and
make
a
pass
whenever
we
feel
like
we
are
getting
underwater.
C
Yeah
we've
been
doing
that
at
the
sort
of
in
the
second
half
of
the
the
ci
triage
meeting.
I
wasn't
able
to
make
it
last
week
during
a
conflict,
but
typically
once
sergei
is
done
going
over
the
ci
board,
we
can
take
a
pass
during
that
meeting.
So
that's
wednesday
from
10
to
11..
I
don't
know
if
you
can
make
that
one.
I.
C
Great
yeah,
so
we
can
go
over
that
then.
The
only
other
thing
I
wanted
to
mention
is:
there
was
a
little
bit
of
confusion
recently
over
cherry
picks.
So
I
wanted
to
briefly
talk
about
cherry
picks,
so
this
one,
for
example,
now
is
in
good
shape,
but
like
it
had
it
and
some
of
the
other
pr's
associated
with
it
have
been
sitting
for
about
a
month
because
they
hadn't
been
labeled
correctly.
C
So
if
there
is
a
node
cherry
pick,
you
need
to
ensure
that
it
has
the
node
label
on
it.
Hopefully,
the
priority
set
properly
and
all
of
the
branches
that
it
affects
must
be
in
sync.
So
if
you
are
concerned,
you
want
to
get
a
backboard
to
120,
but
it
also
affects
119
and
118.
You
also
need
to
back
port
to
those
versions.
You
can't
just
backport
to
120.
C
You
can
not
back
four
to
119
and
118
if
it
doesn't
affect
those
branches,
but
if
they
are
affected,
then
they're
all
supported
versions,
so
they
all
should
receive
the
cherry
picks.
So
there's
been
a
lot
of
confusion
around
that
sort
of
thing,
because
the
cherry
picks
have
not
necessarily
been
like
staying
in
sync
and
then
they
sit
for
a
while,
because
people
aren't
sure
what
to
do
with
them
and
they
don't
get
attention.
C
I
have
been
trying
to
make
sure
that
all
of
the
cherry
picks
are
aligned
and
in
good
shape
and
getting
in
before
the
cherry
pick
deadlines.
So
you
can
reach
out
to
me
if
you
have
any
questions
about
that
or
if
you
want
to
do
cherry
picks,
I'm
also
happy
to
go
and
do
for
high
priority
things
to
go
and
submit
the
prs,
but
there's
a
whole
bunch
of
documentation
on
how
that
process
works
as
well.
A
B
Just
looking
at
numbers
of
closed
and
nourished
prs,
we
are
not
doing
like
more
than
average
like
we're
doing
as
good
as
we
do.
In
average,
we
don't
do
as
good
as
we
did
so
when
we
actively
purging
the
prq,
but
I
mean
it's
a
regular
number
for
us
typical
over
weeks.
B
If
you
look
at
history,
I
think
the
problem
right
now
exaggerated
by
the
fact
that
there
are
many
prs
at
promoting
something
into
ga,
so
there
are
like
pr's
needed
for
feature
work
and
that
makes
incoming
queue
a
little
bit
higher.
So
you
need
to
double
down
on
that.
E
Hi,
so
this
is
a
quick
one.
I
think
it's
we
discussed
this
last
week.
The
there
are
two
pr's
for
in
the
in
the
enhancements,
repo
that
need
to
be
merged,
and
one
of
them
is
tim.
That's
a
major
one.
Tim
hopkins
changed
that
poll
number
1883
and
I.
E
To
look
at
it
review
his
suggestions,
and
if
there
are
any
questions
about
it,
I
can
I'm
happy
to
answer.
We
both
worked
on.
It
went
back
and
forth
quite
a
bit
and
agreed
upon
what
the
what
we
should
have
in
the
api
change
from
what
what's
currently
in
the
repo
and
the
second
one
is
is,
is
mainly
a
housekeeping
change.
E
Prr
section
was
needed
for
for
the
according
to
the
new
process,
and
I
added
that
now
it
should
be
quick
to
look
at
that
and
review
that
and
merge
it.
This
will
set
it
up
for
the
next
one,
two
two
release
when
we
start
getting
it
in,
did
you
or
derek
have
a
chance
to
drag
his?
Not
here.
Do
you
have
a
chance
to
take
a
look
at
the
changes
that
tim
suggested
or
mother
demon
hi.
A
I
did
look
at
that
one
did
you
ask
me
or
derek,
I
didn't
look
at
it
and
okay.
Yeah
derek,
also
take
a
look
because
the
team
also
talked
to
me
on
those
things.
So
I
don't
have
any
content
and
I
hope
that
I
want
to
take
a
lighter
look.
Yes
right,
so
that
cover
because
we
do
have
like
the
diverging
in
the
past
right.
So
so
I
already
look
at
this
one.
So
I'm,
okay
with
this
yeah.
E
Yeah,
I
think
we
we
already
agreed
upon
checkpointing
in
the
in
the
node
and
moving
the
resources
allocated
to
status,
and
the
main
thing
to
really
look
at
is
whether
that
resize
new
field
that
we
added
in
status,
makes
sense
and,
of
course,
the
future
future
edition,
potentially
in
beta
for
sub
resource,
to
call
in
that's
about
it.
E
A
I
will
also
follow
up.
We
start
after
this
meeting
because
I
think
the
in
the
past.
We
have
some
diversity
on
that
very
topic
that
we
are
in
the
meeting
a
good
converging
in
the
signal
meeting,
but
I
think
that
maybe
still
need
more
time
for
the
digest.
After
read
your
proposal
here,
I
mean
team's
problems
most
okay,
thanks!
A
Thank
you
next
one.
I
just
give
the
quick
update.
It's
just
that's
the
reliability,
effort
from
the
kubernetes
overall
and
one
effort
what
they
are
doing
is
try
to
which
is
in
the
past.
We
already
did
a
while
back,
but
so
every
single
test,
e3
test
they
want
to
have
like
the
sig
owner.
So
we
did
that
long
time
back
and
but
a
lot
of
the
falling
apart,
because
a
new
set
of
the
conform,
types
and
all
new
side
of
the
rules,
all
those
kind
of
things
and
people
also
keep
adding
tests.
A
So
there's
the
one
effort
and
the
people
start
to
looking
into
the
signaled
e2e
test.
Second
of
the
etv
test
used
to
be
completed
owned
by
signal,
because
we
created
of
the
signal
test
and
also
many
signal
tests.
Actually,
it
is
later
contributed
back
to
the
conform
test
or
cluster
level
of
the
e2e
test,
so
but
they
still
have
the
common
set.
So
there's
the
people
start
looking
into
that
one
and
then
realize
a
lot
of
you're
confusing.
A
So
I
just
want
to
see
that
the
second
node
e2
it
has
the
most
node,
the
it
has
the
even
like
the
storage
and
the
next,
this
quarter,
how
to
using
this
water
and
how
to
use
support
of
the
nuclear
storage
and
a
lot
of
things
actually
initiate
bad
signals,
and
also
mainly
the
test
that
I
did
on
the
know.
The
level
single
node
of
the
e3
test.
A
Actually,
that
feature
is
driving
com
mostly
is
complete,
driving
by
the
signal
or
maybe
sometimes
with
partner,
is
like
local
storage
as
the
feature
we
partner
with
the
sig
storage
and
add
that
feature,
but
back
to
neck
auto
for
resource
handling
and
monitoring
these
disk
usage,
all
those
kind
of
things
actually
complete
driving
bad
signal.
So
all
those
kind
of
tests
should
be
owned
by
signaled.
A
I
just
want
to
mention
these
kind
of
things
and
make
sure
those
not
fall
into
crack
and
because
those
is
quite
important
feature
to
make
sure
on
load,
reliability
and
happiness
and
last
time
maybe
it's
direct
or
maybe
a
lila
mentioned
that
in
the
in
the
open
shift
production.
A
There's
the
auto
disk
is
one
awful
concern,
so
those
tests
actually
try
to
help
us
to
prevent
such
things
happen
just
to
share
here
and
that's
just
for
your
information,
and
I
already
helped
to
clean
those
kind
of
things
and
see
what
kind
of
things
we
should
do.
F
That's
me
francesco
here,
so
a
couple
of
updates.
Basically,
first
of
all,
I
was
checking
the
board
de
lana
mentioned
and
I
I'm
not
sure
it
is
on
the
board.
I'm
not
sure
why
and
well
just
raised
it
does
the
net.
They
need
the
triage
label
label
mpr.
So
really
and
the
other
thing
is
we
got
the
cap
approved
with
a
few
comments
which
are
all
addressed,
but
unfortunately
the
review
is
not
really
progressing,
so
this
work
missed
the
1.20
deadline.
F
So
I'm
just
trying
to
do
all
my
well
all
my
homework
to
make
sure
it
it
may
end
up
in
121.
So
this
is
just
raising
attention.
It
still
needs
reviews
and
everything
else.
I've
been
trying
to
reach
out
to
renault
and
few
others,
but
with
little
success
so
far,
so
again
just
raising
awareness.
So
thank
you.
Everyone
suggestions.
Welcome,
thank.
F
A
A
Okay,
I
will
follow
up
with
the
dark
to
see
what
what's
in
his
plan
and
who
is
this
one
yeah.
F
C
F
A
G
Yeah
hi
everyone,
so
it's
basically
a
review
request,
so
I
prepared
the
bunch
of
pr's,
so
I
linked
one
and
the
rest
is
inside
that
pr,
so
like
documentation,
changes
and
cap
update
stuff
like
that.
So
please
review
because
we
still
have
chance
to
get
it
into
the
upcoming
release.
G
A
I
I
don't
know
who
has
the
review
benefits,
but
I
want
to
see
that
actually,
this
time
we
spend
a
lot
of
time
go
over,
do
the
planning
and
go
over
1.21
planning.
So
I
can.
Maybe
you
miss
those
meetings,
so
we
basically
carefully
to
put
everything
is
in
the
root
map
and
if
it's
a
new
map,
you
can
see
that,
for
example,
for
the
power
resource
api
we
basically
identify.
Who
is
the
reviewer?
Who
is
the
approver
and
and
also
what's
the
stage
status
for
this
for
this
release.
A
So
of
course
we
are
not.
I
I'm
not.
I
didn't
propose
to
hold
or
pass
the
progress
that
is
using
what
we
can
move
progress,
but
I
just
don't
think
about
the
at
this
moment
we
can
guarantee
and
because
the
real
benefits,
just
you
heard
like
a
leila
and
the
circuit
earlier
and
also
other
people
propose
something
already
in
the
roadmap,
and
we
we
just
don't
have
that
enough
review
of
bandwidth
and
approval
benefits.
I
just
want
to
explain
to
you
and
yeah
yeah,
definitely
yeah.
A
I
share
with
that
dog.
Let
me
just
share
this
talk
with
you,
and
so
you
can
take
a
look
so
the
menu
and
actually
help
to
create
this
dog
make
us
to
management,
can
manage
our
full
planning
and
and
manage
our
progress
much
better.
So
I
share
with
you.
So
this
is
why
huge
page
is
not
in
that
roadmap.
G
C
C
Like
that's,
not
just
us,
we
would
only
have
a
week
to
actually
get
the
code
merged,
so
this
is
a
concern
of
mine
in
general,
like
the
fact
that
we
only
really
have
a
month
between
enhancements,
freeze
and
code
freeze
like
it's
a
really
short
period
in
the
release
cycle.
I've
raised
this
and
I'm
hoping
that
maybe
we
can
increase
the
amount
of
time
that
we
have
for
these
things,
but
as
of
now,
I
think
it's
it
would
be
very,
very
tight,
so
I
wouldn't
have
bandwidth
to
review
it.
H
H
G
The
pr
I
I
linked,
it's
actually
about
code
change,
so
have
I
understood
correctly
that
I
should
read
this
document
and
and
proceed
like
it's
like
for
including
these
changes
into
the
next
release.
So
there
is
like
no
chance
to
to
get
it
into
the
121.
A
A
I
just
think
about
it's
not
worth
your
effort
to
do
that,
one,
because
you
need
confidence
in
this
team
for
that
exception
and
at
the
same
time-
and
I
just
feel
we
don't
have
that
much
of
the
review
bandwidth
in
signal
also
even
your
currency.
A
That's
the
problem
right
because,
because
we
didn't
identify,
everyone
have
to
allocate
the
actual
review
benefits
and
to
to
do,
and
so
that's
so
that's
like
I
don't,
but
at
the
same
time
we
we
can
make
a
progress,
move
forward
same
thing
like
that,
we
pa
we
still
move
forward.
But
it's
not
committed
into
this
just
a
seminar.
We
will
be
still
doing
those
things.
Yeah.
G
A
That's
okay,
it
is
new.
I
know
it's
new
because
we're
doing
this
kind
of
planning
in
the
past,
but
last
year
falling
apart,
we
didn't
really
follow
that
one
so
manu
this
time
take
extra
effort
to
pick
up
this
one,
and
so
we
can
have
the
better
planning
yeah.
If
you
notice
that
after
1.14,
we
haven't
done
this
for
a
while.
So
I
think
this
is
really
good
things
for
us
to
especially
giving
our
review
bandwidth
is
the
not
enough.
A
This
is
make
us
up
upfront
communicate
with
the
every
developer
and
the
community
members
all
the
things
yeah
thanks
for.
A
I
So
I
there
was
a
I
started:
a
thread
on
the
sig
release,
slack
channel
and
the
121
release
leads
have
decided
that
this
can
go
into
the
121
release.
So
that
is
good.
We
had
the
other
kep
the
windows,
node
log
viewer
or
the
the
node
log
viewer
cap
that
one
likely
will
not
make
it
into
this
cap
all
right
into
this
release.
I
So
we're
planning
on
moving
forward
with
the
windows,
privilege,
container,
stuff
and
I'll
have
some
of
the
implementation
peers
up.
Hopefully,
in
the
next
couple
of
days,.
A
Okay,
that's
great.
I
look
at
the
our
planning
dock.
I
think
most
other
things
actually
in
the
in
the
right
track.
Of
course,
the
review
benefits
is
still
for
the
real
implementation.
That's
the
big
issue
for
us
any
other
things.
People
want
to
broaden
up.
G
Just
one
quick
question
about
the
the
planning:
so
what
what
I
see
in
the
document
is
121
planning
and
when
the
next
release
planning
will
start
and
whom
should
I
contact
about
it?.
A
We
should
have
like
the
1.20
string,
but
I
think
I
will
suggest
the
reason
I
share
that
doc
like
this
is
not
meet
the
120
one.
Then
I
better
use
you
can
add
item
there,
but
then
we
decided
like
that.
You
can
see
the
color
code
and
the
suggested
by
the
eyeliner
I
like
that
one,
but
there's
a
lot
of
is
without
the
color.
Basically,
it's
just
pumped
right
to
the
next
or
maybe
even
further,
so
I
suggest
right
there.
So
at
least
we
have
that
who
it
is.
We
start
can
talk
about.
A
Who
is
the
reviewer
an
approver?
And
then
you
start
those
kind
of
process.
So
the
next
time
when
we
kick
out
the
start,
the
new
new
document,
so
this
one
will
be
automatically
moved
to
the
new
dock,
that'd
be
easier
for
us
to
track.
C
A
C
A
A
A
Yes,
yes,
so
so
just,
but
we
can
talk
about
the
negative
planning.
So
yes,
so
so,
okay,
so
we
can.
I,
I
think
this
is
good.
Things
just
add
those
things
into
the
kind
of
planning
doc,
but
we
know
that's
not
commitment
and
then
makes
people
can
pick
up
to
do
the
planning
they
will
know
where
the
base
nine,
where
to
start
that'll,
be
easier.