►
From YouTube: Kubernetes SIG Node 20210105
Description
Meeting Agenda:
https://docs.google.com/document/d/1j3vrG6BgE0hUDs2e-1ZUegKN4W4Adb1B6oJ6j-4kyPU
A
Okay
signals
meeting,
as
I
said
it's
january
5th,
2021,
so
yeah
we
didn't
do
so
bad
during
the
holiday
season.
We
skipped
few
meetings
but
still
like
we
have
a
negative
like
we
have
less
prs
than
we
had
before
holiday,
so
people
were
approving
and
closing.
Some
many
pr's
were
closed
by
like
by
people
not
by
automation.
That
is
good.
A
A
There
is
a
link
with
a
dashboard
on
open,
prs
and
signaled,
and
you
can
notice
that
last
year
we
clearly
turned
the
wave
around
so
before
that
we
only
grew
on
pr
account,
and
now
we
are
slowly
decreasing
pr
count.
Every
every
release.
We
have
a
smaller
and
more
baseline
for
open
prs.
A
A
Okay,
so
next
one
is,
we
did
some
triaging
end
of
last
year
for
cisc
group
ci
subgroup,
and
there
is
a
few
pull
requests
that
are
lgtm
by
the
group
but
not
approved.
So
we
ask
this
group
to
like
anybody
who
has
approval
rights.
Please
take
a
look
and
try
to
help
us
to
move
this
forward.
A
Thank
you
and
lastly,
I
started
this
migration
to
docker
shim,
my
migration
out
of
docker
documentation.
There
is
a
place
in
kubernetes
io
and
I
propose
to
start
putting
documentation
around
and
some
that
will
help
guide
customers
how
to
do
migration.
A
We
identified
few
areas
where
customers
may
need
help,
some
of
them
like
monitoring
and
security
agents
that
may
not
work
when
docker
film
will
be
duplicated
with
that
switch
from
the
kershim
other
areas
like
how
to
detect
that
they've
been
using
docker
stream
at
all.
So
I
started
this
documentation.
Please
review
and
contribute
help
really
appreciate
it
here.
A
C
Hi,
sorry,
I
didn't
think
my
number
would
come
so
soon
swede
happy
new
year,
everyone
for
those
of
you
who
don't
know
me,
I'm
karen
from
google,
I
work
with
sergey.
We
are
starting
to
look
into
swap
support
and,
as
part
of
that
last
quarter,
I
spent
some
time
just
digging
into
what
some
of
the
blockers
for
it
will
be,
both
technical
and
in
terms
of
design
itself,
and
I
put
some
thoughts
in
a
dock.
That's
linked
to
in
the
agenda.
C
It's
like
three
pages.
It's
not
too
large.
C
Oh,
my
god,
one
second.
C
Okay,
I
know
why
sorry
about
that
I
will
share.
I
think.
C
Okay,
you
should
have
access
now.
I
think,
based
on
what
I've
seen
supporting
swap
is
too
broad
of
an
ask
for
us
to
make
significant
progress
on
or
impactful
progress
on,
anyways,
and
so
I
think,
even
in
a
couple
of
bugs
there's
a
lot
of
use
cases
that
are
being
discussed,
but
I
think
it
would
be
prudent
for
us
to
narrow
it
down
to
a
few
that
we
think
we
can.
C
We
can
make
progress
on
quickly,
so
I
have
my
own
opinions
as
you
can
expect,
but
I
think
if
there's
a
if
there's
a
few
people
who
would
like
to
collaborate
on
that
and
maybe
for
121
scope
or
in
121
timeline,
if
we
can
have
a
cap
with
high
level
design,
you
know
use
cases
we
want
to
tackle
and
what
kind
of
exact
problems
we
want
to
solve.
C
That
would
be
great,
I
mean,
or
if
that's,
if
that's
not
something,
that's
not
a
good
way
to
approach
it,
I'm
still
open
to
options.
There.
D
So
I
started
looking
at
this
right
before
sort
of
heading
into
the
holiday
break
and
I
guess
it's
kind
of
precipitous
because
I
spent
most
of
my
time
working
on
the
use
cases.
So
I
have
a
dock
put
together
of
a
bunch
of
use.
Cases
documented
on,
like
the
public
issue
sort
of
classified
by
type,
as
well
as
from
some
other
conversations
that
I've
had,
which
should
help
guide
some
of
the
implementation.
D
It
looks
like
you've
looked
at
most
of
the
challenges
in
terms
of
the
implementation,
so
essentially
very
complimentary.
I
have
not
yet
had
a
chance
to
like
throw
that
into
a
public
doc.
Mostly,
I
was
seeking
feedback
on
that
and
also
everybody's
on
vacation.
So
I
think
it
would
make
sense
for
us
to
put
together
some
sort
of
like
sig
public
dot,
get
feedback
on
it
as
far
as
timelines
go
like.
D
I
think
that,
like
121
alpha
at
this
point,
probably
like
pretty
aggressive
but
maybe
like
122
alpha
so
hopefully
getting
like
the
initial
cap
done,
this
release.
C
E
F
Yeah,
a
big
plus
one
on
the
two
of
you
getting
together.
B
F
On
the
red
head
side,
we've
been
having
a
lot
of
use
cases
where
swap
has
been
coming
up,
as
potentially
useful,
like
if
you're
running
smaller
kubernetes
deployments
outside
of
a
cloud
context,
if
you're
using
things
like
cuber,
there's
a
lot
of
community
ecosystem
projects
around
cube,
where
maybe
some
of
the
more
ideological
positions
we
had
taken
in
kubernetes
when
we
assumed
the
presence
of
permanent
elastic
infrastructure
was
always
present
are
not
always
holding
now
for
where
cube
is
being
placed,
and
so
big
plus
one
on
that.
F
I
think
the
timeline
that
elana
raised
is
good.
What
I'm
wondering
is
anyone
on
the
call
from
any
particular
user
community
right
now.
That
would
benefit
from
swap
that
wanted
to
help
as
well.
I
know
we
had
cube
vert
members
joined
previously,
but
maybe
if
we
could
pair
up
with
a
set
of
instead
of
platform
providers
but
more
like
platform
consumers
to
help
enrich
that
use
case.
That
would
be
good.
D
Yeah,
I
do
have
a
list
of
some
folks
who've
indicated
interest
both
like
for
contributing
like
on
the
issue
as
well
as
people
who,
I
think,
indicated
interest
in
one
of
the
previous
meetings,
so
yeah
I'll.
I
don't
want
to
like
you,
know,
dox
people
and
put
their
contact
info
in
like
a
public
thing
without
asking
them
first
but
I'll
start
reaching
out
to
people
as
we
work
on
this
and
hopefully
get
together
a
group
and
get
get
people's
feedback
and
make
sure
that
everybody's
kind
of
on
board.
B
D
Like
a
you,
gotta
find
the
balance
between
like
yak
shaving
and
like
five
big,
like
sort
of
appropriate
scope
for
like
an
mvp.
That's
actually
going
to
be
helpful,
so
I
think,
having
as
many
people
involved
in
those
conversations
as
possible,
sort
of
coming
to
some
sort
of
consensus
on
that
and
then
going
ahead
and
figuring
out
what
that
sort
of
initial
implementation
looks
like
that.
That's
that
would
be
my
plan.
C
All
right,
then,
at
least
on
the
gke
side,
we
have
a
couple
of
big
customers
who
are
willing
to
be
design
partners
in
it.
So
I'm
sure
they'll
be
happy
to
chime
in
and
provide
some
of
the
use
cases
as
well,
but
elena,
I'm
assuming
you're
from
red
hat
as
well.
C
Okay,
okay,
I
think
yeah.
I
can
also
post
on
slack
and
see
if
there's
more
interest
in
joining
this,
but
whenever
your
doc
is
ready
to
be
shared,
I
can
I
can
upstream
some
of
stuff
from
my
doc
in
it.
B
D
C
I
think
the
big
one
is,
you
can
set
memory
low
and
memory
high,
so
it
just
has
finer
crane
support
for
setting
memory
swaps.
I
think
in
v1.
Again
I
don't
know
the
details
exactly.
This
is
just
based
on
cursory
reading.
In
v1,
you
only
set
a
upper
bound.
F
Yeah,
so
I
know
in
v1
you
can
only
set
the
upper
bound,
but
there
was
also
the
ability
to
set
memory
and
then
mem
plus
swap
as
a
separate
value.
I
didn't
know
if
renault
or
maybe
giuseppe
is
on
if
they
had
to
memory.
What
was
in
b2
or
what
might
exist
in
oci
today,
with
respect
to
swap.
B
B
F
I
just
want
to
make
sure
that
we
identify
any
v1
versus
v2.
F
Yeah,
so
we
already
had
work
started
to
have
cuba
tolerate
v2
now
that
had
merged,
and
so
I,
my
bias
would
be
to
you
know,
plan
for
the
future,
where
we
accept
that
v2
is
going
to
be
this.
The
the
way
things
go
forward
and
not
like
do
something
that
stops
v2
from
being
possible.
C
D
Great,
I
will,
as
soon
as
I
have
a
doc
up
I'll
link
it
in
the
meeting
notes
and
I'll
also
send
an
email.
E
Okay,
it's
me
hello,
so
happy
new
year.
I
will
be
very,
very
quick.
So
basically
we
have
this
pod
resource
api
extension
we
are
proposing.
We
have
a
cap
already
approved
for
the
for
the
topology,
we're
scheduling
to
enable
that.
Basically,
this
is
the
overarching
goal,
so
the
pod
resources
changes
are
pretty
much
simple.
In
itself
I
mean
the
cap
was
mostly
uneventful.
E
F
E
I
didn't
I
got
reviews
from.
I
know,
I'm
sorry
I
misurd
probably
reynold.
Yes,
I
got
in
touch
with
him,
but
last
year,
not
in
2021,
so
I'll
ping.
Him
here.
E
D
Thanks
cool,
I
put
something
on
the
agenda
for
the
status
of
the
sigs
caps.
I
don't
know
if
we
have
any
other
topics.
So
if
we
want
to
jump
into.
A
That
I
think
it's
a
good
transition,
because
pod
resources
is
a
cap
and
it's
likely
on
this
list
and
we
need
to
understand
whether
it's
scheduled
for
121
or
not.
D
Yeah,
one
of
the
things
that
I
try
to
do
with
sig
instrumentation
since
I've
become
a
chair
of
the
sig,
was
make
sure
kind
of
at
the
beginning
of
every
release
cycle
to
go
through
all
of
like
what
are
our
caps.
What
are
we
planning
on
doing
this
release?
What
are
we
bumping
to
next
release
and
that
kind
of
thing
and
the
sig
node
list
is
quite
long
there's
I
think
54
open
issues
in
the
enhancements,
repo
and
of
those
the
vast
majority
are
not
tracked.
D
37
are
not
tracked,
so
I
was
wondering.
Is
there
like
a
plan
to
to
tackle
this?
Is
anybody
currently
assigned
like
basically
like?
Why
are
things
the
way
they
are
and
is
there
like
a
task
force
for
like
okay?
These
are
the
121
things.
These
are
maybe
the
122
things,
or
is
this
just
kind
of
organically
gotten
in
this
state.
F
Obviously,
there's
a
number
of
features
that
are
in
flight
or
habit
and
flight.
I'm
trying
to
look
through
this
list
to
see
if
there
are
explicit
ones
that
we
hadn't
talked
through.
D
One
of
the
things
that's
like
that's
a
bit
special
about
signature.
I
mean
like
one
of
the
things
that's
open
here,
for
example,
is
like
issue
number
five
pod
security
policy
which
is
currently
untracked
and
so
like
yeah.
F
So
if
that
was,
if
I'm
not
mistaken,
so
that
that's
a
good
example
of
like
things
that
we
should
probably
split
between
exclusively
signaled
versus
having
another
sig
label,
so
sigoth
actually
did
the
deprecation
of
pod
security
policy
right
before
the
holiday
period,
so
at
least
that
one
its
fate
is
at
least
clear.
Now
in
the
cube
community,
I'm
not
sure
if
you
looked
at
others
that
might
be
spanning
cigs
but
yeah.
It's
probably.
F
Through
the
the
less
remaining
particularly
older
ones,
to
see
if
they've
been
cleared
out
or
known.
D
From
like
2016
and
2017
and
stuff,
that's
just
not
kind
of
sat
since
then,
and
a
number
of
them
have
like
life
cycle
scales
set
on
them.
But
I
don't
know
if
that
necessarily
means
that
they're
actually
stale.
I
think
it's
just
one
of
those.
The
scope
of
node
is
so
large
and
I
don't
know
that
people
have
necessarily
had
the
bandwidth
to
even
look
into
this
stuff.
So
no.
F
So
one
of
the
things
I
had
started
to
do
a
little
bit
120.
I
didn't
look
at
the
enhancements
issues
list
because
that
was
actually
not
even
as
accurate
as
we
would
have
wanted
relative
to
the
feature
gates,
and
so
I
just
pasted
a
link
there,
which
was
my
first
attempt
to
try
to
track
some
of
this
to
push
for
forward
progress.
F
We
did
make
a
lot
of
progress
on
a
number
of
these
items
and
maybe
what
we
can
do
as
a
next
step
is
like
update
this
to
say:
is
there
actually
a
tracking
enhancement
issue
or
not,
and
where
we
had
some
consensus
on
this,
maybe
close
out
the
issue,
but
I
think,
even
in
this
stock
there
were
like
30
odd
feature
gates
that
were
in
various
states
of
work,
and
we
did
make
a
fair
bit
of
progress
last
release
pushing
a
number
of
them
forward
so
but
but
in
general
I
think
alana
as
a
sig
we're
all
open
to
like
learning
on
maybe
things
that
work
well
in
instrumentation
that
you
want
to
help.
D
Yeah,
it's
definitely
it's
hard
to.
You
know,
for
example,
compare
like
instrumentation
with
node,
because
instrumentation
has
what,
like,
I
think,
eight
kept
or
something-
and
this
is
like
you
know-
order
of
orders
of
magnitude
larger
so
as
and
in
terms
of
just
like
the
the
scope
in
terms
of
time
and
the
complexity
of
the
features,
it's
much
larger,
I'm
not
sure.
Actually,
if
there
are
like
other
cigs
working
on
things
that
are
like
similar
in
scope,
like
maybe
api
machinery.
So
I
can.
D
I
can
go
look
into
that
kind
of
thing
and
I
can
also
sync
with
the
enhancements
folks
and
see
what
they've
been
doing
in
terms
of
tracking
for
the
sig
node
stuff,
because
I'm
sure
that
they've
spent
a
bunch
of
time
on
that
as
well.
F
I
think
it's
just
a
matter
of
like
us,
committing
as
a
group
to
maybe
pick
the
next
three
to
five
items
that
we
all
feel
comfortable
moving
forward,
whereas
some
gotten
kind
of
a
stasis
once
they
reached
a
beta
state
and
I'm
not
even
sure
if
this
list
was
accurate
with
everything
that
we
did
push
forward
in
120.
But
I
think
anything
that
was
labeled
green.
I
think
was
this.
You
sergey
who
had
gone
and
updated.
A
Yeah
I've
been
updating
it
like
on
the
top
of
document.
I
put
some
example
how
to
ga
feature,
and
then
I
try
to
map
it
to
enhancement
when,
whenever
I
looked
at
it
so
most
of
greens
are
done,
but
then
I
realized
that
the
release
team
is
doing
the
same,
so
release
team
was
also
taking
all
the
documentation
issues
or
the
like.
They
have
a
spreadsheet
for
that.
A
So
maybe
what
is
this
document
helped
a
lot
with
is
to
identify
which
areas
we
want
to
work
on
and
get
the
owner
for
this
area,
so
somebody
driving
it
forward
and
once
this
like
this
identifies,
is
great
and
then
maybe
we
also
need
to
look
at
what
is
critically
dangerous,
like
maybe
something
that
needs
urgent
access
and
nobody
looking
into
that.
So
maybe
just
reviewing
all
the
caps
may
help
to
identify
those.
G
Thanks
for
brought
this
up-
and
I
just
quickly
look
through-
and
I
think
at
least
that
we
can
have
the
smuggle
and
review
the
review.
Those
like
a
lot
of
tracked
stuff
make
sure
and
what's
the
plan
move
to
the
next
stage.
At
least
we
understand
what
it
is
not
the
tracked
right
now
so
so
I
think
last
release
and
the
direct
drive
initiate
effort
and
try
to
move
a
lot
of
like
the
from
one
stage
out
of
pre-ga
stage
to
the
next
stage.
G
So
there's
a
lot
of
more
academic
progress,
but
it's
not
as
progress
what
we
are
originally
expected,
and
so
maybe
this
time
this
quarter
and
we
can
start
look
at
at
least
we
see
the
pro
more
actively
to
close
some
stealth
issues.
Still
the
cap.
G
Specific
available
for
the
certain
use
cases
and
is
there's
no
good
way
to
come
up
or
agree
or
reach
our
consensus
on
the
generic
solution
so
far.
Yet
so
that's
why
we
are
after
a
while,
we
kind
of
get
the
organic
native
is
organically
grown
and
have
some
like
solution,
and
then
we
hopefully
like
the
copper
solution,
come
up
for
the
concrete
use
cases.
Then
we
can
evolve
into
the
generic
solution.
G
A
D
Okay,
I
can
take
an
action
to
at
least
what
I'm
gonna
plan
on
doing
is
following
up
with
the
enhancements
team,
seeing
like
the
work
that
they
did
in
the
last
release
and
then
maybe
trying
to
put
together
something
for
like
this
is
the
status
of
all
these
54
things
and,
let's
see
you
know
who
owns
them
and
who
to
follow
up
on
them
and
next
sort
of
steps.
So
we
can
at
least
try
to
get
this
disentangled
cool.