►
From YouTube: Kubernetes SIG Node 20220209
Description
Meeting Agenda:
https://docs.google.com/document/d/1j3vrG6BgE0hUDs2e-1ZUegKN4W4Adb1B6oJ6j-4kyPU
A
B
You
know
what
it
is:
yeah
you
have
this
blog
post
with
alana
to
celebrate
our
achievements
over
two
years.
It's
not
yet
two
years,
but
I
think
it's
a
good
beginning
of
the
year
and
annual
reports
by
six
will
be
starting
soon,
so
yeah,
it's
a
perfect
time
for
this
blog
post.
B
A
Yeah,
I
recommend
that
folks
take
a
look
at
this
so
sergey,
and
I
sort
of
put
this
together
and
we
gave
shout
outs
to
a
lot
of
folks
for
like
people
that
have
sort
of
gone
up
the
contributor
ladder
through
the
sig
and
we
sort
of
wanted
to
treat
the
sub
project
just
like
a
little
case,
study
to
show
other
like
cigs
and
sub
projects,
how
this
sort
of
thing
might
work
and
how
people
can
kind
of
like
get
more
involved
within
the
project.
A
So,
please,
like
read
over
for
accuracy,
add
anything
that
we
might
have
missed
we're
very
worried
about
missing
things.
We
don't
want
to
miss
things
so
like
if
you're
worried
about
commenting
publicly
feel
free
to
message
either
of
us
privately
but
yeah.
We
wanted
to
sort
of
celebrate
the
work
that
everybody's
been
doing
here
and
make
sure
that
you
get
a
little
bit
of
visibility
for
like
some
of
this
sort
of
behind-the-scenes
effort
that
doesn't
necessarily
get
a
lot
of
applause.
A
We
mentioned
it
in
passing,
but
make
sure
you
just
go
ahead
and
like
take
a
look
over
it
and
let
us
know
if
there's
anything
major,
that
we
should
have
commented
on.
A
Okay,
what
what's
the
time
limit
the
time
limit?
Well,
it
looks
like
they
have
the
blog
scheduled.
A
A
Okay,
anything
else
with
us.
A
No
moving
on
danielle
is
adding
tests
to
various
parts
of
the
cubelet.
Thank
you.
I've
done
this
before
it's
not
very
fun.
I
appreciate
this.
D
Yeah,
no,
it's
horrible
going
from
nothing
to
something
is
like
the
worst
part,
because
you
have
to
write
so
many
garbage
tests
yeah
just
to
like
build
a
mental
model
and
also
to
like
make
sure
things
are
testable
by
other
things.
Yes,
but
yeah.
My
plan,
when
I
have
you
know
the
occasional
five
minutes
of
not
continually
unbreaking
e
to
e
tests
is
going
through
package
by
package,
the
kubelet
and
adding
tests.
Whether
or
not.
A
No
they're
really
not
having
done
a
lot
of
unit
test
refactoring
reviews
you
have
to
go
through
every
single
test
case
line
by
line.
We
thank
you
danielle.
D
A
Yeah,
I
know
that
this
has
been
something
that,
like
I
have
talked
about
at
least
internally
at
red
hat,
about
like
questions
of
like.
Do
we
want
to
invest
in
adding
a
bunch
more
test
coverage?
Do
we
want
to
rewrite
portions?
Because
it's
going
to
be
too
difficult
to
add
test
coverage
and
like
we
may
as
well
just
like
start
from
scratch,
because
a
lot
of
better
understanding
of
it
are
there
things
that
we
want
to
spin
out
of
tree
and
so
on?
A
So
you
know
just
kind
of
like
big
futuristic
things
for
cubelets,
so,
yes,
some
mix
of
the
three
I'll
let
I'll
let
folks
know,
because
certainly
adding
tests
are
going
to
be
a
lot
of
a
lot
of
investment.
Oh
sure
I
can
review
and
approve
this
or
no
approve.
A
There
you
go
done
next
agenda
item.
B
So
I
would
suggest
we
go
to
status
of
ci
just
to
update
everybody.
What's
going
on,
I
put
a
new
column
today
and
yeah
on
signal
critical.
We
start
seeing
more
and
more
random
failures.
Last
week.
It
was
a
couple
this
week
is.
A
A
Yeah,
okay,
I
added
I'll,
maybe
as
an
action.
I
might
file
an
issue
about
this
afterwards
because
yeah,
I
just
don't
think
that's.
B
A
So,
oh,
your
credentials
are
invalid.
I
couldn't
say
that.
B
B
B
Yeah,
besides
that,
there's
a
small
progression
group,
we
do
not
major
breakthrough
and
the
idt
keeps
testing
this
good
subscripts
for
continuity
yeah.
This
is
still
red.
B
A
few
david
cathey
with
a
few
so
something
happening.
D
Yeah,
that's
next
on
my
priority
list,
because
we
would
like
c
groups
v2
to
be
like
our
default
testing
target
soon.
A
D
Has
almost
had
been
an
increasing
sea
of
red
lately,
it's
possible.
A
This
looks
like
an
ant
issue:
yeah.
These
are
otherwise
all
green.
So
maybe
just
I
don't
know,
there's
some
issue
with
container
dnc
group
v2,
because
when
this
is
not
failing
all
the
time
it
looks
okay,
we
don't
have
any
actual
test
failures.
That's
so
weird!
Okay!
I
do
this
pretty
critical.
D
Yeah,
it's
next
on
my
list
after
seeing,
if
like
after
trying
to
figure
out
if
something
is
a
regression
or
expected
behavior
that
was
previously
broken
yeah.
Is
this
the
same
issue.
B
B
So
few
yeah
those
contaminants
issues
are
boost
up
issues
that
ajit
is
looking
at
all
right.
There
is
already
pr
for
that.
So
I
mean
there
was
pr
before
we
merged
it
and
turned
out
that
we
need
more
changes
anyway.
B
Else
is
pretty
much
unchanged
from
last
week
that
okay.
A
B
A
A
That's
fair,
okay.
I
also
want
to
make
sure
we
get
the
flakes
and
failing
tests.
A
Okay,
that
looks
like
it's
assigned
and
accepted,
so
it
looks
like
it's
in
progress.
Probably
this
looks
like
it's.
B
A
I
this
is
interesting.
We
typically
don't
backport
end-to-end
tests,
but
there's
I
guess
this
isn't
just
an
end-to-end
test.
This
is
a
whole
patch.
I
had
a
discussion
with
my
team.
I
think
internally
about
this
one
before
I
disappeared
for
a
couple
of
days,
so
sure
I
will
throw
that
in
needs
approver
and
we
can
discuss
more
later,
but
I
think
these
cherry
picks
honestly.
I'm
not
sure
that
they're,
the
ci
board
isn't
the
right
word
for
them,
because
there's
a
pretty
large
change
in
the
cubelet
as
well.
A
So
I
guess,
keep
test
up
of
that
branch
is
failing.
A
B
A
Okay,
triaged,
I
will
assign
it
or
I
will
put
it
into
removed
references
to
docker
in
the
api.
A
A
This
is
a
pr,
oh,
I
don't
think
this
is
the
ci
group
we
want
to
go
through,
so
I
know
that
this
yeah
it
looks
like
nobody's
working
on
this.
C
C
And
they're
working
directly
on
this
one,
I'm
working
with
the
version,
but
I
think
the
failure
is
basically
the
same.
I
did
another
pr
recently
yeah
see
there's
a
couple
yeah.
Unfortunately,
this
one
didn't
fix
the
issue
itself,
I'm
still
trying
to
find
out.
What's
up
with
this.
A
Well,
let's
see
exclude
non-failed
tests
well
that
test
isn't
on
here
at
all,
so
that's
another
one,
then
yeah
yeah.
This
is
a
different
test.
The
the
summary
one
is
not
showing
up
as
failing
on
here.
So
that's
a
good
sign.
Perhaps
your
change
fixed.
It.
A
For
this
is
ubuntu
not
fedora
right,
so,
okay
yeah,
it's
definitely
still
failing
on
up
on
two,
but
not
fedora.
So
let
me
at
least
note
that.
A
Cool
this
is
sort
of
a
feature
I
don't
know
what
the
status
of
that
is,
but
I
need
to
check
in
on
that
pods
extended.
Is
this
still
flaking.
A
Okay,
so
we
triage
accepted
this
one
a
while
back
and
then
we
have
not
assigned
anybody
to
it.
I
guess,
let's
see
if
sandbox
yeah,
that
one
is
still
flaking
for
sure.
Does
anybody
want
to
work
on
this?
One.
A
A
Feel
free
to
assign
yourself,
okay,
okay,
I
think
that's,
probably
everything
that
we
need
to
look
at
on
this
board.
Let
me
just
make
sure
there's
no
okay,
it's
probably
done,
but.
A
Then,
let's
move
on
to
yeah
you
merged.
Let's
move
on
to
bugs.
A
Okay,
pod
with
sri
ov
device,
failed
on
ungraceful
mode
reboot
a
bunch
of
logs,
but
not
very
many,
not
enough
to
debug
this
okay
they're
on
122
cloud
provider,
openstack
no
other
in
details.
Okay,
they
at
least
said
they're
on
susie.
A
A
A
I
mean
to
me
this
doesn't
sound
like
a
cubelet
thing.
It
sounds
like
a
storage
thing,
because
if
restarting
the
cubelet
fixes
it
then.
A
Okay,
maybe
let's
check
for
some
things
that.
A
Since
I
commented
on
basically
all
of
these,
we
can
just
look
at
the
ones
that
have
other
updates.
Look
at
these
two.
A
A
I'll
just
leave
this
one
open.
Since
I
need
to
comment
on
it.
There's
no
point
in
me
commenting
to
say
I
need
to
comment
okay,
what
happened
on
this
one?
Let's
see.
A
A
A
I
honestly
just
don't
think
that
we
should
be
spending
a
lot
of
time
trying
to
help
an
individual
like
come
up
with
their
cri
implementation,
and
that's
probably
not
a
great
use
of
our
time
if
they
are
breaking.
If
we
have
broken
something
if
we
violated
guarantees
and
by
all
means,
but
they
actually
have
to
show
that
and
like
one
line
is
not
really
enough
to
show
that
so,
okay,
I
think
that's
it
for
bugs.
We
are
at
half
past.
If
you
have
anything
else
you
want
to
cover
today.
B
Can
refuse
to
run
time
no
matter
here,
one.
B
A
And
apparently
pods
are
using
less
memory
and
the
same
amount
of
cpu
cool.
Okay,
thank
you.
I
think
we're
done
for
today.
Oh
there's,
two
things
in
chat.
I
missed
oh,
but
one
is
nice.
It's
probably
when
I
merged
or
noticed
thing.
Okay,
I
think
that's
all
for
today,
thanks
everybody
for
joining
us,
hi
cube.