►
From YouTube: Kubernetes SIG Node 20210623
Description
Meeting Agenda:
https://docs.google.com/document/d/1j3vrG6BgE0hUDs2e-1ZUegKN4W4Adb1B6oJ6j-4kyPU
A
Hello,
everyone
and
welcome
to
today's
edition
of
the
sig
node
ci
subgroup
today
is
wednesday
june
23rd
2021
and
we
don't
have
anything
on
our
agenda
today,
but
we
have
a
bunch
of
stuff
from
last
time.
So
I
may
as
well
share
my
screen
with
the
agenda
and
I
was
thinking
we
should
at
least
review
some
of
the
topics
from
last
time.
So
there's.
A
Nothing
on
the
agenda,
but
we
discussed
a
few
things,
including
removing
the
node
feature
flags
and
having
some
better
documentation
for
node
conformance
and
how
that
is
different
than
conformance
and
then
also
the
status
of
serial
tests.
So
is
there
any
particular
order?
We
would
like
to
cover
these
things
in.
A
A
A
Yeah,
I
I
would
rather,
we
not
have
a
single
point
of
failure
on
sergey,
so
does.
Is
anybody
else
interested
in
this?
I
think
that
this
is
like
aaron's
written
a
pretty
straightforward
description
here.
For
some
strange
reason
excuse
me:
we
have
a
lot
of
tests
that
use
like
node
versions
of
selectors
instead
of
just
regular
selectors
and
right
now.
I
guess
it
doesn't
necessarily
make
a
lot
of
sense.
A
So
that
way
we
can
pick
them
up
easily
in
the
tests,
but
I
think
probably
what
would
need
to
be
done
for
this
looking
in
test
infra,
if
I
can
I'm
always
bad
at
this,
but
let
me
see
if
I
can
find
our
tests.
C
A
Yeah,
so
we've
got
a
number
of
these
things
like
node
feature
and
node
alpha
feature.
So
I
think
my
suggestion
would
be
probably
the
order
that
we
would
need
to
do
this
in
is
we
would
have
to
update
all
the
selectors
in
the
test,
infra
jobs
anywhere
that
node
feature
and
node
alpha
feature
is
used
to
like
pick
up
something
else,
for
example
like
feature
like
colon
node,
blah
blah
blah,
and
then
that
would
be
followed
up
with
some
prs
in
kubernetes
kubernetes
to
update
the
names.
A
C
Yeah
something
like
that,
I
think
one
of
the
reasons
we
had
not
feature
is
not
selector
like
before.
We
clearly
separated
sigs,
we
may
have
had
test
run
from
like
common
folder,
and
then
we
needed.
D
C
Prefix
but
now
we
clearly
separate
all
the
tests
and
so
like
you
can
always
do
extra
filtering
by
sig
note
prefix,
because
it's
currently.
A
Oh
yeah,
that's
right!
That's
right!
We
have
the
what
you
would
call
it
the
sig
selectors
now
and
all
the
tests.
C
Yeah,
so
it's
all
fixed
now
like
we
just
need
to
clearly
articulate
the
plan
and
that's
what
I
wanted
to
do.
Sorry.
I.
A
C
B
A
Cool,
let
me
just
add
a
couple
of
notes
for
that.
A
A
History
well,
there's
that
which
is
different,
node
conformance.
C
Maybe
it's
not
confirmed
versus
couplet
versus
regular
tests.
We
have
the
well.
A
A
I'm
not
sure
that
we
have
an
issue
to
document
the
criteria
for
node
conformance
tests.
A
Yeah,
it
will,
but
I
mean
at
least
for
this
one
I
think
the
main
ask
we
really
need
to
get
the
docs
updated.
So
people
know
what
node
conformance
means,
and
so
I
think
that
would
live
here
and
then
we've
got
like
at
least
three
or
four
other
issues
talking
about
node
conformance
jobs
in
both
test
infra
and
kk
and
kind
of
all
over
the
place.
So
I
this
is
probably
the
right
place
to
put
the
documentation
issue.
C
A
B
A
Steering
wouldn't
look
at
something
like
this
because
steering
mostly
deals
with
governance.
Sig
testing
might
care
about
this
or
the
conformance
subproject
in
sig
architecture.
So
I
could
add
sig
testing
to
this.
A
B
And
in
the
openshift
issues
you
don't
have
something
related
since
you
were
looking
at
it
from
the
cryo
point
of
view.
Now.
A
Oh
yeah,
so
cryo
wants
to
add
some
like
tests,
but
those
are
mostly
to
cover
like
so
that
we
don't
just
have
one
container
runtime
in
the
right
like
we.
If
they're
cri
conformance
tests,
we
want
to
ensure
that
we're
testing
more
than
one
container
runtime
and
the
other
thing
about
the
cryo
test
is
they
have
coverage
for
the
system
d
driver
and
I
don't
believe
that
the
container
d-tests
do
so.
A
A
A
I
don't
know
if
it
would
either
be
in
signode
or
sig
testing,
because
a
bunch
of
the
test
documentation
already
exists
here
in
sig
testing,
including
conformance
tests.
So
maybe
we
would
have
a
link
in
the
conformance
test
docs
into
the
sig
node
repo,
which
talks
about
node,
conformance
or
cri
conformance.
Or
what
have
you.
A
A
Yes,
I
can
hear
something
now
francesco.
Do
you
have
any
updates
on
serial
tests.
A
Okay,
sorry,
I
was
not
looking
at
the
chat,
not
much
updated,
still
working
on
them,
so
we
split
out
the
eviction
tests.
I
don't
know
if
we've
done
that
for
the
test
repo
yet
for
cryo.
I
think
we
just
skipped
them.
A
A
D
D
A
I
understand
okay,
and
then
I
guess
the
only
other
thing
we
had
on
the
agenda
is
we're
doing
a
bug
scrub.
It
starts
in
my
evening
today
and
will
be
the
next
two
days
so
thursday,
friday,
depending
on
where
you
are
in
the
world,
and
we
don't
have
a
lot
of
people
signed
up
right
now
to
help
with
the
americas.
So
if
you
want
to
help
out
in
nasa
region,
we
only
have
one
mentor
signed
up
outside
of
me
and
sergey.
A
So
I
would
appreciate
if
we
had
some
folks
on
the
call
who
would
be
interested
in
doing
that,
it
would
be
preferable
that
either
you
are
well
that
you
would
be
both
an
org
member
and
like
have
been
to
a
bug
scrub
before
and
kind
of
know
what
mentors
at
bug
scrubs
do,
but
you
probably
one
or
the
other
is
totally
sufficient.
A
A
So
there
is
a
sign
up
form
which
I
will
link
from
the
slack.
A
And
I
think
we've
got
lots
of
volunteers
in
aipac
and
ameya,
so
that
should
be
good.
B
So,
just
to
understand
the
you
created
a
zoom
meeting
for
the
scrub
and.
D
A
Yeah,
that
was
that
was
my
sort
of
idea.
I
mean
probably
people
will
drop
in
and
out,
but
we
should
make
sure
that
there's
at
least
somebody
who
has
host
in
there
at
all
times,
just
to
make
sure
you
know
if
some
rude
person
drops
in
that
they
can
be
kicked
out
and
because
the
the
links
are
public.
So
there's
always
a
risk
of
that.
I
think
it's
mostly
going
to
be
that
kind
of
hallway
track
for
people
to
chat.
A
I
don't
know
if,
like
you
know,
people
might
want
breakout
rooms
or
something
like
that.
I
think
that
functionality
is
available
so,
but
not
entirely
sure
I
haven't
messed
around
a
lot
with
zoom,
so
yeah.
C
A
B
Me
also,
I
will
probably
be
online
much
earlier
than
when
emilia
is
supposed
to
start
so
yeah.
I
would
say.
A
Yeah
feel
free,
I'm
probably
going
to
be
around
for
like
a
bunch
of
the
at
least
early
in
the
apac
shifts
as
well.
So
it
just
depends
on
like
most
of
our
apac
leads
are
like.
We've
got
a
bunch
of
folks
in
india
and
I
don't
think
we
have
any
volunteers
from
australia,
for
example,
but
like
india,
people
are
like
it
starts
so
early,
I'm
like
well,
it's
not
early
for
new
zealand,
so
you
know
somebody's
gotta
be
there
to
receive
them.
So
it's
a
global
event.
A
We've
got
people
everywhere,
not
just
one
place
or
one
time
zone,
so
yeah,
I'm
looking
forward
to
it.
You
can,
of
course
let
me
know
if
you
have
any
questions
or
need
anything,
and
I
hope
that
all
of
the
folks
who
are
on
this
meeting,
I
will
see
you
the
next
couple
of
days
scrub
and
bugs.
B
Yeah,
maybe
just
one
one
more
question:
do
we
assign,
and
I
mean
how
do
we
keep
track?
Who
is
looking
at
which
bug.
A
So
that's
up
to
the
game,
mentor
or
region
captain
to
sort
of
coordinate.
That's
why
we're
splitting
it
because
otherwise
it's
just
way
too
much
to
track
for
one
person,
so
the
idea
would
be
you
know
you
have
a
bunch
of
newcomers
and
you've
got
a
list
of
bugs
as
like
a
mentor
or
a
region
captain
and
so
somebody's
like.
I
need
a
bug,
and
so
you'll
say:
okay.
Why
don't
you
look
at
this
bug?
A
Why
don't
you
look
at
this
bug
and
so
on
and
we've
got
you
know
I
think
each
player
has
about
somewhere
between,
depending
on
the
player
20
to
40
bugs.
So
that
should
be
enough
to
last
for
two
days.
If
you
need.
A
Out
you
can
just
like
leave
a
note
in
the
channel
because,
right
now
we
don't
have
enough
mentors
in
nasa,
so
we
might
have
extra
bugs.
B
Yeah,
but
I
I
mean
more
like
we,
we
make
like
a
google
sheet
or
something
to
keep
track
to
remember,
to
poke
someone
if,
if,
if
he
has.
A
I
think
that
would
be
much
too
complex
and
liable
to
get
out
of
date.
I
would
try
to
assign
people
in
the
github
issues
themselves.
A
But
in
a
lot
of
cases
like
some
of
these
can
be
closed
pretty
quickly.
So
there's
no
point
in
assigning
somebody
to
something
that
you're
just
going
to
close
out.
If
you
want
someone
to
investigate
something,
then
certainly
assign
them
to
it.
So
they
can
then
put
a
report
on
it
and
know
that,
like
other
people
are
not
looking
at
the
same
thing
but
yeah,
that's
yeah,
yeah.
B
Yeah,
that's
yeah,
yeah
yeah,
it's
four!
Yes,
yeah
yeah
yeah!
It's
like
putting
a
mutex
on
it
to
make
sure
we
don't,
especially
if
we
are
like
two
captains
and
and
make
many
mentors
and
okay.
A
Yeah,
the
only
thing
that
I
did
create
sort
of
an
external
thing
for
that's
like
not
in
github
is.
I
know
that
there
are
a
lot
of
like
features
and
generally,
we
probably
shouldn't
be
tracking
features
in
github
issues
in
kk,
because
it's
just
kind
of
it's
hard
to
go
through
them
all
it's
hard
to
prioritize
and,
like
anybody
can
file
a
feature
request
there.
A
So,
if
you
see,
for
example,
like
you
know,
there's
a
bunch
of
different
issues
all
related
to
one
feature
or
there
are
like
a
lot
of
different
requests
for
a
particular
thing.
You
might
want
to
note
them
in.
I
created
a
dock
for
like
feature
stuff
so
that
way
sometime
in
the
future,
maybe
at
a
sig
node
meeting
we
can
maybe
go
through
and
like
prioritize,
hey,
we've
gotten.
All
these
feature
requests
is
that
on
our
roadmap
kind
of
thing,.
D
Hey
alana:
does
this
involve
to
finish
the
work
and
like
do
a
pr
or
only
like,
find
out
what
is
the
bug
about
and
dig
down
on
these
things?
What's
the.
A
But
if
you
find
a
bug
and
you're
just
like
I'm
really
passionate
about
this
bug,
and
I
want
to
fix
it,
and
you
know,
I
think
it
has
high
impact
and
it
will
be
helpful
to
fix.
Please
fix
it
like
that's
what
this
sort
of
event
is
for,
because
otherwise
we
wouldn't
know
that,
like
oh
yeah,
there's
this
bug
and
it's
like
bad,
that's
often
one
of
the
funnest
parts
of
doing
a
bug
scrub
is
going
and
submitting
prs
to
actually
get
the
bugs
fixed.
Instead
of
just
you
know,
updated
and
closed.
C
Even
if
you
put
enough
details
how
to
fix
this
bug
just
apply
such
help,
and
it
will
help
like
it
will
help
new
contributors
to
find
issues
to
contribute.