►
From YouTube: Kubernetes SIG Node 20210310
Description
Meeting Agenda:
https://docs.google.com/document/d/1j3vrG6BgE0hUDs2e-1ZUegKN4W4Adb1B6oJ6j-4kyPU
A
Welcome
everyone
to
today's
sig
node,
triage
and
ci
meeting
I
am
currently
hosting
today,
because
sergey
is
unavailable.
So
let
us
go
through.
The
board
looks
like
we've
got
eight
things
in
triage
20
issues
to
do
some
issues
in
progress.
Oh
and
let
me
also
pull
up
the
agenda.
A
Oh
and
I
forgot
the
date,
it
is
wednesday
march
10,
2021.
A
So
it
looks
like
we
have
just
one
agenda
item,
which
is
the
d
flick
of
the
sig
note
alpha
tab,
which
I
think
we
have
successfully
done.
I
mean
do
you
want
to
talk
more
about
this
yeah.
C
Looks
like
italy,
there
was
missing
the
runtime
config.
This
is
some
parameter
used
by
api
server.
So
cube
test
was
solving
this
parameter,
not
passing
to
the
signal
test
at
all,
and
now
we
have
this
on
the
runtime.
A
That's
great
okay!
So
that's
our
one
agenda
item
and
I
guess
let's
look
at
it
and
be
odd.
Ooh!
Look
at
all
those,
not
failures!
I
don't
see
red
every
day.
That's
amazing!
There
are
actually
tests
running.
That's
amazing!
Go
us!
That's
great
work!
Thank
you
so
much
for
looking
into
that,
because
I
know
that
that's
been
a
slog
and
this
is
before
when
you
know
it
failed
every
single
time.
So
look
at
that
stark
difference
as
of
whatever
this
time
and
date
is
looks
like
march
9th.
So
yesterday.
A
Brilliant
okay,
so
do
we
have
any
other
items
for
the
agenda.
A
Sounds
like
no,
let's
jump
into
triage,
so
I
have
not
actually
looked
into
this
sergey
started
and
I
had
also
started,
but
I
got
distracted
by
caps.
So
it
looks
like
we've
got
a
bunch
of
stuff
to
triage
for
pr's
and
issues.
Let's
maybe
start
with
the
issues.
A
And
this
one
is
probably
a
good
place
to
start,
because
I
have
already
kind
of
looked
into
this.
I
tried
to
report
this,
but
it
was
the
day
of
the
great
everybody
gets
logged
out
from
github
and
then
github
like
banned
me
and
ate
my
flake
report
for
like
eight
hours.
So
I
think
that
matthias
is
looking
into
this.
There
is,
as
far
as
I'm
aware,
a
working
progress
pr.
I
think
it's
this
one
99998,
so
I
will
attach
that
and
maybe
assign
it
to
him.
A
Race
condition
from
liggett
and
it
looks
like
janju
kim-
has
taken
a
look,
that's
great
and
there's
a
pr
up,
which
I
think
is
in
our
column
here.
So
let
me
just
double
check,
so
that's
taken
care
of
which
is
good
like
having
assignees.
A
Oh
you
put
a
life
cycle
frozen
on
there.
Don't
do
that?
Let's
see
so
now.
I
guess
we've
just
got
some
prs.
Would
you
like
me
to
go
over
the
triage
prs
first
or
the
to
do
issues?
First,
we've
got
a
lot
on
the
board.
It
looks
like,
or
do
we
want
to
go
over
updates
on
in-progress
issues?
A
Okay,
we
have
a
very
a
very
amenable
audience.
I
see
that's
totally
fine,
let's
I'm
gonna,
I
think,
probably,
since
we're
focusing
on
issues.
Let's
continue
focusing
on
issues.
A
I
think
that
there's
a
little
bit
of
mixing
of
things
right
now,
because
I
see
there's
a
pr
in
this
column
and
I
think,
there's
also
a
couple
of
pr's
in
this
column,
and
so
let
me
just
maybe
fix
that
as
pr
I
will,
I
think,
like
these
ones
are
they
got
sent
back
to
like
the
owners
to
work
on,
because
we
don't
really
have
a
pr's
in
progress
column,
but
this
one
is
a
to-do
for
the
review,
so
I'll
put
it
into
this
column
and
we
can
make
sure
that
somebody
who's
not
just
derrick,
is
assigned
to
review
it.
A
This
one
has
an
lgtm
on
it,
so
we
can
come
back
to
that.
One.
E
A
Okay,
I
will
let's
look
at
this
one
and
then
I'll
drag
this
into
the
approver
column,
and
then
I
don't
know
what
the
status
of
is
this
one.
Is
it
still
a
work
in
progress
or
just?
Is
this
ready
for
review
the
adding
restart
cubelet
flag
on
end
to
end
test.
A
Oh,
that
is
directed
at
I
mean
who
is
the
author.
B
A
A
A
A
So
issues
in
progress.
I'm
gonna
go
through.
Let's
do
those
first,
because
otherwise
we
won't
get
there's
a
lot
of
things
to
do
so.
Do
we
have
matt
on
the
call.
A
Looks
like
no,
so
we're
probably
not
going
to
be
able
to
request
updates
from
him.
However,
I
don't
know
if
these
are
still
failing
tests,
so.
A
Cubelet
master
and
node
conformance
seem
like
duplicates.
This
is
quite
an
old
issue.
Does
anybody
know
what
the
status
of
this
one
is.
D
A
A
Yeah
it
looks
like
this.
One
has
not
moved
since
december,
so
node
conformance
has
continually
failing
tests.
That's
bad!
Is
that
still
true?
Yes,
what
is
the
test?
That's.
A
A
A
A
A
Issue
I
know
that
we
recently
removed
the
container
d
1.2
tests,
because
I
lgtm
to
that.
So
it
looks
like
we
have
an
assignee
here
any
status
updates
on
this
one.
This
looks
like
it's
well
taken
care
of
just
in.
C
A
I
know
this
one
is
in
progress.
The
assigned
enforced
ownership
of
all
end
to
ends.
Is
there
anyone
on
the
call
who
has
an
update
on
this
one
like
there
have
been
a
lot
of
prs
for
this,
so
I
don't
know
if
there's
anything
that
desperately
needs
a
review.
It
looks
like
there's
a
new
pr,
but
I
don't
know
if
this
is
for
node.
D
A
Okay,
well,
this
was
only
assigned
last
week.
So
let
me
just
triage
accept
this
and
we
can
move
on
and
it
may
be
fixed,
there's
a
attached
pr,
oh,
which
is
even
the
one
that
I
have
open.
So
it
looks
like
there's
been
progress.
Huzzah
node,
end
end,
lock,
contention
job.
A
Is
this
still
irrelevant?
I
guess
it's
good.
I
must
be
going
in
the
opposite
order
of
sergey
because
I
guess
we're
looking
at
all
this
old
stuff,
okay,
okey-doke
and
we
all
have
two
more
in
progress,
and
that
was
one
that
I
just
assigned
and
then
this
one
was
already
assigned,
so
that's
great
cool
so
to
do's
and
some
have
helped
wanted.
A
A
A
It's
because
it
was
frozen.
I
see
I
see
so
I
just
removed
that
it's
why
I
generally
think
people
should
not
put
frozen
on
things
because
it
is
applied
irresponsibly,
sometimes
like.
Surely
that
no
longer
applies.
It
was
a
test
failure
thing
and
it
was
like
a
test
failure
thing
from
you
know
the
stone
age.
So,
let's
see
failing
test
probing
containers
should
be
restarted
with
interesting.
A
One
not
to
pick
on
people
but
swati,
you've
been
very
quiet.
Are
you
interested
in
taking
a
look
at
this.
A
A
A
Progress
because
that
one
is
quite
important,
that's
a
release.
Blocker
so,
of
course
feel
free
to
ask
help
with
that
one.
There
was
a
thread
in
the
slack
about
some
issues
with
liveness
probe
tests
not
behaving
properly,
so
you
might
want
to
read
through
that
and
I'm
very
happy
to
help
if
you
have
any
questions
about
like
the
code
paths
or
things
like
that,.
A
B
A
This
is
green.
So
what
about
these
two.
A
A
A
A
Okay,
that
was
f
romani.
A
Okay,
let's.
A
A
Often
enough,
okay
sounds
like
we
should
investigate
this
one.
We
need
an
assignee
who
wants
to
look
at
this.
D
A
I
don't
know
exactly
how
many
overlaps
we're
gonna
find
here.
I'm
sure
they're
gonna
be
some.
So
this
one
looks
new.
This
one
is
a
duplicate
of
what
was
it
just
the
the
one
that
the
bot
marked
as
stale,
that.
A
A
okay,
but
this
one
is
new,
should
not
start
app
containers
if
in
it,
okay,
just
timed
out,
anybody
want
a.
A
E
E
A
A
Oh,
that's
an
old
one.
Some
of
these
are
very
old.
We
have
12
more
to
do's,
but
we're
making
good
time.
So
I'm
glad
that
we're
doing
this,
I
know
this
is,
I
think
I
think
pod
deletes
when
sources
are
ready
for
qos.
A
A
A
A
A
A
A
Does
anybody
know
if
we
still
have
this
test?
I've
never
seen
this
dashboard
before.
A
So
I'm
gonna
leave
that
one
unclosed.
A
V2,
if
it
was
just
another
timeout,
oh
yeah,
this
is
different,
so
who
did
we
assign
to
the
not
v2
one?
Would
that
person
also
like
to
take
a
look
at
this?
A
D
A
D
A
A
A
A
I
was
gonna,
say
I've
reviewed
some
mirror
code
or
mirror
pod
code.
I
should
say
and
yeah
I'm
pretty
sure
we
have
coverage
for
this.
It's
just
somebody
sticks
a
life
cycle
frozen
on
it
and
then
it
never
gets
closed.
A
A
A
And
if
you
want
to
add
yourselves
to
the
agenda
for
this
week
or
sorry
to
the
attendees,
I
will
add
myself,
you
can
each
add
yourselves.
I
think
I
put
the
link
in
the
chat,
but
if
not,
we
can
do
that.
A
A
A
A
Oh,
maybe
it's
just
because
it's
a
deletion,
so
the
bot
doesn't
know
what
to
do
with
it,
but
then
it
should
at
least
have
like
cl
or
I
guess
it
does
have
the
cla
check.
A
A
A
B
A
And
apparently
my
internet's
lagging
there.
We
go
and
let
me
throw
this
one
to
in.
A
Progress:
okay,
I
think
we
got
through
all
of
the
to
do's.
Go
us
very
exciting
and
I
guess
we've
got
this
one
more
to
do
here
and
then
a
few
pr's.
So
how
are
people
feeling
we
are
43
minutes
into
this?
So.
A
A
Yeah,
I
don't,
I
feel
you
yesterday
code
freeze,
I've
been
working
pretty
long
days,
it
looks
like
adisk
is
looking
at
this
one.
I
will
maybe
assign
her
and
see
if
she'll
look
at
this.
A
Okay,
great,
so
we
got
five
more
prs
and
then
I
think
the
board
will
be
in
goods
good,
no,
four,
more
pr's,
because
the
card
counts
as
one
thing,
and
then
we
can
just
throw
these
all
to
the
review
in
progress
column.
I
think-
and
we
can
try
to
find
some
reviewers
for.
E
A
A
A
A
A
Now
I'm
gonna
encourage
everybody
just
sort
of
offline
to
we've
got
like
10
prs
here
that
I
think
need
review,
I'm
pretty
sure
that
most
of
them,
you
know
anybody,
can
just
go
ahead
and
pick
them
up.
I'm
not
gonna
run
through
these
and
assign
them,
because
I
think
that
we
might
have
some
trio.
I
should
do
on
the
other
board
for
the
second
half
of
this
meeting,
but
is,
are
there
any
other
node
ci
topics
for
today?
I
think
we
did
really
great
in
terms
of
board
cleanup.
A
Sounds
like
no
okay,
I'm
gonna
stop
recording
and
then
we
can
jump
into
the
second
half,
which
is
just
all
of
the
other
things
triage.