►
From YouTube: Kubernetes SIG Node CI 20220601
Description
SIG Node CI weekly meeting. Agenda and notes: https://docs.google.com/document/d/1fb-ugvgdSVIkkuJ388_nhp2pBTy_4HEVg5848Xy7n5U/edit#heading=h.2v8vzknys4nk
GMT20220601-170415_Recording_1392x1032
A
Okay,
hello,
it's
the
first
day
of
June
2022..
It's
a
signaled,
CIA
meeting,
welcome
everybody.
We
go
to
agenda
first
and
the
Francesca
is
here
right.
Yes,.
B
I
am
so
okay,
so
it's
really
about
the
status
update
for
this
PR,
which
wants
to
re-enable
the
device
plugin
test
which
we're
skipping
because
they
got
broken
sometimes
ago.
So
this
is
the
context
now
quick,
good
news.
The
test
should
be
fixed,
the
I
it
took
a
while
with
some
giant
things,
but
they
should
be
fixed
I'm,
pretty
confident
it
should
be
in
good
shape.
Why
mostly
good
news?
Because,
yes,
a
lane
is
failing
and
you
can
see
in
the
Pro,
but
that
line
fails
doing
the
setup
well
before
any
end-to-end
test
run.
B
So
first
of
all,
I
want
just
to
make
sure.
We
are
aware
that
this
Lane
is
the
serial
exactly
they're.
Not
so
popular
sailing
is
failing
and
this
feeling
was
set
up
and
second
and
the
question
is
how
we
move
forward.
Do
we
want
to
track
this
failure
and
ignore
for
this
case,
or
do
we
want
to
solve
it
for
move
forward
I
just
asking,
because
I
don't
have
any
strong
feeling
in
this
case.
A
B
A
I'm
curious
Mike,
if
you
don't
have
any
contacts
with
this
test,
it's
a
swap,
there's
a
you
may
have
worked
on.
C
A
Okay,
let
me
put
it
here:
yeah
I
think
we
can
ignore
it,
since
it
has
never
worked
before,
and
does
anybody
wants
to
get
assigned
to
this
PR
to
review
it.
B
B
A
A
And
I
think
we
go
okay,
let's
go
through
this
board.
I
filled
up
this
week
and
yeah,
it's
mostly
same
as
the
last
week
Ubuntu.
A
So
what
I
found
is
there
is
a
summary
tests.
Still
flaky
on
some
environment
like
on
some
Lane
I,
didn't
look
into
details
like.
Is
it
the
same
flake
as
before,
as
we
fixed
before
the
previous
release
or
something
new
I
hope
it's
something
new,
and
then
there
is
another
test
at
failing.
I
already
started
character
account
for
something-
and
this
is
the
same
Lane
as
last
week
but
last
week
as
a
bug
report
here
is
different.
A
It's
about
device
plugins
that
was
fixed
by
Francesca,
I,
think
so
yeah,
it's
a
separate
failure,
different
failure,
so
I
will
create
a
separate
block
and
maybe
close
previous
bug.
A
A
Maybe
we
need
to
close
this
one
because
it's
separate
and
then
then
good,
to
see
you
and
then
open
separate
one
for
of
a
new
failure.
Frequently.
A
Yeah
I
will
create
about
what's
this
meeting
and
performance
are
still
failing,
I
had
a
fix
that
fixed
the
image.
The
image
is
not
failing
any
longer
now
it's
failing
on
some
timeout,
so
maybe
it
takes
too
much
time
to
to
perform
this
like
tender
flow.
Computation
I
think
we
have
a
bug
for
that.
So
we'll
take
a
look.
A
Everything
else
is
pretty
much
the
same
as
last
week.
Not
much
not
much
movement
so
should
be
fine,
and
this
one
is
which
I
discussed
in
Mike
I
will
create
a
bug
and
ask
you
to
take
an
initial
look
at
list.
A
A
Any
questions
for
for
the
health
of
our
tests
I
know
the
install
also
created
a
summary
back
recently,
at
least
in
all
the
failure
failed
tests.
I
think
it
serves
the
same
purpose
as
this
one.
Just
you
really
need
to
get
get
it
down
to
healthy,
but
I
mean
now
we
have
new
flakes,
we
just
fixed
device
plugin
and
have
a
new
place
now,
anyway,.
B
Hey
trash.
A
E
A
Okay,
let's
take
a
look
so
yeah
unready,
endpoints
test,
400
endpoints.
A
Yeah
I'm
not
sure.
Yes,
somebody
needs
to
take
a
look
but
I'm,
not
sure
which
PR
will
merge
it.
Maybe
it
will
go
with
a
fixed
PR.
Does
anybody
want
to
be
assigned.
F
Yeah
we
integrated
this
test
into
my
PR
to
fix
certain
points.
Okay,
so
we
should
be
able
to
close
it
I
think
shortly.
D
A
A
It
has
enough
for
the
universe,
let
me
assigned
specifically
to
David,
as
he
is
working
very
closely.
Okay,.
A
A
A
So
yeah
Daniel
you
you'll,
be
even
assigned
here.
Did
you
have
a
chance
to
take
a
look.
E
Apparently
I
reviewed
it
I
think
this
looked
good.
I.
Think
the
testing
for
changes
also
looked
good
I,
don't
think
either
of
us
has
enough
approval
power
to.
A
A
A
Context,
right
now
same
I,
think
so
safe
to
accept
it
and.
A
E
Take
a
look
at
the
alternative,
PR
I
think
it's
the
next
one
opening
your
list
of
tabs
too.
Okay,.
E
A
E
A
Yeah,
it's
safe
to
triage
to
do
yeah
I
think
we
track
all
of
them.
You
know
tested
it,
but
it's
another
way
to
track
it.
Yes,.
E
E
They're
attracting
comments
mostly
but
yeah
eviction
is
the
unfun
one.
A
A
E
A
Okay,
she
repeats
clearly
not
CI
specific.
A
D
A
Almost
done
yes,
this
is
just
a
to-do.
You
know
configs.
We
use
this
legitimeter
setting
because
it
is
duplicated.
So
we
need
to
switch
to
new
way
to
configure
things.
A
A
In
mainly
increase
configuration,
it's
pretty
straightforward.
A
Is
anybody
interested
in
getting
issues
from
to
do,
but
it
has
three
Cycles
if
not
I
will
catch
up
on
approver
needed
and
some
reviews
I'm
still
not
up
to
date
with
everything
and
yeah.
If
you
have
something
assigned,
please
try
to
clean
it
up,
so
we'll
have
more
a
better
idea.
What
what's
left
to
be
done.
A
A
A
Okay
yeah:
this
is
this
is
what
we
looked
at.
We
need
to.
A
D
A
You're
assigned
to
accepted.
E
A
Was
sweetie
here
before?
Did
you
know
that
she's
looking
at.
B
D
D
A
A
A
yeah,
it's
understandable-
maybe
we
can
add
six
storage
here
as
well,
because
clearly,
six
storage
look
into
that
and
since
there
is
a
PR.
E
Yeah
the
tricky
one
of
this
is
figuring
out
how
this
is
gonna
break
things.
E
And
this
one's
a
bug
with
unclear
like
Behavior
as
to
how
we
should
actually
fix
it.
E
A
E
No
so
when
it
creates
the
lease
things
get
weird
like
so
the
least
tries
to
be
updated,
but
at
least
got
deleted
by
the
cascading
delete.
And
so
then
we
created
a
new
lease
with
no
owner
reference
attached,
because
this
code
recreates
stuff,
if
things
fail.
E
A
Did
you
have
a
chance
to
look
at?
Why
are
we
ignoring
issues?
It
says
that
it
ignore
some
errors
and
try
to
recreate
it
anyway.
A
A
A
E
So
the
current
state
of
this
is,
it
definitely
happened
on
Docker
I
couldn't
reproduce
it
on
continuity,
I
need
to
try
again
and
cryo,
and
then
I
can
close
it
out.
A
What
was
the
name
of
a
person
who
looked
at
it?
Yeah
I
think
we
we
looked
at
it
and
we
had
this
checked
I.
Just
don't
remember
the
person
from
Microsoft
who
looked
at
that.
Let
me
set
a
reminder
for
myself.
E
A
E
D
A
A
Right
right,
it's
another
buckets,
122.
yeah.
It
was
because
David
switching
okay.
Can
you
take
a
look
yeah.
E
A
D
A
E
Just
oh
so
the
action
item
for
this
was
figuring
out.
If
we
had
any
precedence
of
fixing
stuff
and
release
branches,
that's
removed
from
current
versions.
E
I
mean
so
like
leaking
sandbox
containers
could
eventually
host
the
node.
So
like
it's
not
exactly,
you
know,
leave
it
alone
thing.
A
A
A
Not
working
for
everybody
but
yeah
at
least
there
is
a
workaround.
A
I'm
not
I
mean
many
vendors,
didn't
support
continuity
for
or
like
craft
for
a
long
time
and
surprise
events
also
same
boat.
A
E
E
This
one
is
weird
and
that
it
I
I
think
that
using
a
fork
of
kubernetes
and
that
I
can't
find
what
this
would
happen
and
the
fixed
they
proposed
wouldn't
fix
it
for
anywhere
that
they
seem
to
be
referencing
like
it's
really
weird.
E
A
A
D
A
E
Like
how
that
probe
worked
mostly
for
me,
is
it
trying
to
make
requests
and
they
can't
get
through
like
what
you
expect
to
happen?
Or
is
there
something
weirdo
with
like
an
exact
probe
or
something.
E
Like
I
want
I
want
to
like
pay
someone
to
write
a
browser
extension
that
gives
me
auto
complete
for
proud
commands.
B
A
Yeah,
this
is
the
same
bug
right,
David
I
think
it's
related
to
what
you're
doing
all
right
right
here.
F
This
is
it
so
yeah
I
looked
at
this,
so
this
is
not
the
same
bug
as
the
one
that
I'm
fixing.
F
I
need
to
update
that
comment,
which
I
will
do
can.
F
A
Feel
free
to
unassign
yourself
if
it's
completely
undulated,
okay
and
since
you
trusted
in
I,
will
put
it
in
triage.
But
again,
if
it's
not
Embark
just
comments
here,.
A
D
E
A
A
A
A
D
B
Feel
free
1.18.18,
it's
very
old,
but
still
you.
You
cannot
send
it
to
me
if
you
want,
but.