►
From YouTube: Kubernetes SIG CLI 20230719 Bug Scrub
Description
Kubernetes SIG CLI bug scrub and triage on July 19th, 2023.
B
Everybody
to
the
July
19
2023
bug
scrub.
Please
remember
that
all
kubernetes
meetings
are
governed
by
the
code
of
conduct,
so
please
be
nice
to
have
everybody.
C
B
B
A
B
Okay,
I
can
follow
up
with
somebody,
or
we
can
try
and
track
down
that
being
the
precedent
for
this
yeah.
Okay,
Cube
cuddle
watch
fails
after
first
event.
B
E
I
think
I
looked
into
it
and
I
think
they
figured
it
out.
I
told
told
them
to
use
verbose
logging
at
the
very
bottom
and
I
think
they
figured
out
their
own
problem.
Are
we
doing
something
else?
They
were
doing
something
they
didn't
say,
I!
Think
in
the
in
the
original
description,
okay,.
B
A
B
No
okay,
yeah
probably
cannot
handle
Json
path.
Filter
expressions
like
at
dot
x,
equals
y.
E
So
yeah
it
was
wrong.
Yeah
you're
right
actually
I
forgot
about
that.
It
might
still
be
open.
I
can
check
yeah.
B
E
A
E
So
I
think
this
person
is
confused
because
somewhere
I
think
in
the
readme's.
Sometimes
it
says
dot
IO,
slash,
Coupe
pedal,
but
it's
really
referring
to
the
GitHub
repo.
Not
it's
not
intended
to
be
an
actual
clickable
URL.
E
B
B
Okay-
and
this
is
at
the
they.
F
E
A
B
Right
yeah,
so
they're,
not
even
clickable,
unless
this
is
the
one,
but
no
this
one
goes
to
the
repo
properly
and
I.
Don't
see
it
yeah,
it
was
not
not
an
issue.
B
C
D
B
B
B
E
B
E
And
there
might
be
two
parts
to
it
as
well,
so
maybe
maybe
the
first
step
is
like
hey.
Let's,
let's
update
client
go
so
that
it
can,
you
can
programmatically
get
the
ports
and
then
once
we've
got
that
in
place,
then
we
could
look
at
adapting
the
coupe
cuddle
port
forward
command
to
then
provide
different
output
types
or
something
I'm.
Just
trying
to
think
I
haven't
looked
at
the
fork,
so
I
don't
know
how
big
of
a
thing.
F
E
D
B
Is
this
the
same
one
from
before?
It
is
not
the
same
one
from
before.
Okay,
actually,
it
looks
like
the
same
one
from
o
I
think
that
means
I
had
a
comment
written
for
this
and
it
never
got
posted
right
because
after
this.
B
B
Rewind
to
verify
that
this
is
what
was
expected,
basically,
which,
if
I
remember
right,
it
was
not
what
was
expected,
and
so
it
is
definitely
a
bug,
and
the
main
reason
is
that.
C
F
A
A
C
A
B
E
I
mean
it's
not
a
bad
issue.
I
am
if,
if
somebody
wants
to
do
a
pull
request,
so
I
think
this
came
about
because
they
had
a
pull
request
about
some
cleaning
up
the
help
and
then
we're
kind
of
getting
going
down
the
line
of
finding
more
and
more
inconsistencies
and
I
was
just
like
hey,
let's
just
let's
just
fix
the
ones
that
you
started
to
set
out
to
fix.
Instead
of
because
I
mean
you
could
probably
go
through
all
the
commands
and
find
a
lot
of
improvements
to
to
the
help
text.
E
I
guess
you
know,
my
caution
would
be
against
just
like
word
smithing
and
trying
to
make
it
perfect
versus
good
enough,
and
you
know
we
don't
want
to
just
constantly
be
tweaking
the
health
text.
But
if
there's
places
that
are
where
things
are
clearly
missing,
then
it's
good
to
fix
that
yeah.
That's.
B
Where
I
come
down
on
it,
yeah
I
for
sure
it
looks
like
all
the
PRS
that
were
open
for
this
have
been
merged.
B
C
B
E
I
think
there
are
I
think
the
pr
that
was
referenced
in
this
one
that
was
linked
from
this
issue
was
oh,
okay,
the
one
that
yeah
that
that
PR
was
the
one
that
kind
of
kicked
it
off
kicked
off.
This
issue
was
the
May
24th
one,
okay,
so
I
think
there
are
still
some
things.
D
B
The
1430
config
that's
right.
B
No
like
yeah,
this
is
a
workaround
for.
B
B
B
They
could
update
their
hosts
file,
they
could
they
could
replace
the
hostname
with
the
IP
address.
They
could
run
a
DNS
server
that
doesn't
return
values
that
its
Network
can't
handle.
E
B
E
I
don't
know
I
mean
there
are
several
commands
that
are
popular
commands
that
do
support
those
flags,
but
it
doesn't
necessarily
mean
we
should
I
guess,
but
it
it
does
kind
of
add
more
weight
to
the
case
that
hey
you
know.
These
are
kind
of
kind
of
standard
flags
that
are
used
in
certain
situations.
E
I
guess
I,
don't
know
I'm
mixed
on
it,
but
I
guess
I'm,
just
saying
if,
if
I
mean
we
could
accept
it,
but
if,
if
somebody
submitted
a
pull
request
to
add
a
couple
of
global
Flags,
I
think
there
would
be
a
lot
of
questions
about
whether
we
really
need
those
flags
or
not.
C
D
A
B
B
C
Oops
view.
A
D
B
E
Yeah
so
there
so
there's
an
issue
that
this
PR
is
linked
to,
that
is
unexpected,
eof
and
I.
It
involves
changing
some
of
the
Speedy
stuff,
which.
D
E
Know
is
already
going
to
be
changed
anyway
to
websockets
in
the
future,
so
this
PR
didn't
I
mean
it
does
fix
some
of
the
it
does
fix
the
unexpected
eof
problem,
but
I
don't
know
if
there's
really
any
interest
in
in
merging
this
PR
or
not
and
I,
don't
know
if
it's
it
might
be.
This
problem,
it's
hard
to
say:
okay,.
E
So
yeah
there's
a
workaround
with
the
retries:
it's
not
a
good
workaround,
but
yeah.
It
will
at
least
let
people
have
a
chance
of
succeeding
with
us
if
they
need
to.
B
Yeah
I'm
wondering
if.
E
B
Well,
it
could
be
the
sure.
A
B
A
B
E
E
B
A
E
B
B
E
F
B
Yeah,
okay:
we
don't
really
need
to
keep
on
with
that
one
all
right.
So
with
the
last
15
minutes,
we
will
talk
about
this
one.
Sorry,
the
six
CLI
stuff
in
the
kubernetes
repo.
A
E
E
And
these
a
bunch
of
these
are
untreated.
Our
needs
triage,
but
they're
really
old
in
the
cube
cuddle
repo,
and
this
was
one
of
them
and
it
it
came
up
because
it
it
went
rotten
or
something
and
then
I
looked
at
it
and
I
was
like.
Oh,
this
really
is
a
problem
and
in
fact
it
happens
all
the
time
in,
in
the
end,
tests
in
a
flake
situation
and
I
don't
know
so
I
guess
we
I
feel
like
we
should
at
least
triage
it.
But
I
don't
know
if
it's
a
simple
solution
or
not.
E
So
so
the
rule
of
the
matter
is
that
so
we
have
a
timeout
for
the
overall
HTTP
requests,
that's
made,
but
there's
a
TLS
handshake,
timeout,
that's
less!
So
it's
like
a
10
second
timeout
or
something
so,
regardless
of
what
the
overall
timeout
is.
You
can
have
a
TLS
handshake
timeout
that
happens
in
like
10
seconds
and
then
it'll
just
fail
and
it
only
happens
under
heavy
load
right.
B
B
A
E
B
So
you
you
decided
to
Frozen
I
think
this
should
definitely
be
accepted.
A
B
B
E
I
guess
that
was
part
of
my
question
today.
I
mean
we
I
I,
don't
know
we
could
I
mean
I.
Could
post
something
in
I
could
maybe
go
on
slack
and
ask
in
the
API
machinery
Channel
what
they
think,
whether
that's
an
API
machine
or
anything,
the
the
triage
thing,
the
screenshot
one
of
the
screenshots
I
posted,
said
API
Machinery
on
it.
So
I
don't
know.
E
That's
probably
why
I
thought
that
one
like,
maybe
maybe
it
is
API
machinery
and
not
us
yeah,
or
at
least
they
should
be
involved,
probably
so
I'll
so
I
guess
we
could
accept
it.
I
don't
know
if
we
want
to
accept
it
or
not
before
we
figure
out
who
it
belongs
to
or
I'll
ask
in
the
slack
Channel
who
it
belongs
to.
B
Yeah
is
it
actually,
it
seems
like
well,
okay,
so
it's
being
reported
as
a
cute
head
also,
however,
it
may
end
up
being
something
that's
actually
inside
of
API
Machinery,
instead
of
especially
because
it's
a
client
go
thing,
probably
or
sorry.
What
what
Eddie
links
to
is
a
client
goes
thing.
I
would
not
say
that's
a
guarantee,
that's
where
the
source
of
the
issue
relies,
so
I
didn't
accept
it
as
six
a
lot
CLI
for
the
moment,
if
you
find,
if
it
is
Brian
when
you
talk.
C
A
B
My
favorite
thing
about
doing
this
is
I
just
get
like
a
torrent
of
emails
constantly
from
all
of
these
issues
and
stuff
that
I
comment
on
okay
under
restricted,
PSA
ephemeral
containers
can't
be
created
with
restricted
profile.
B
D
B
A
F
B
Idea,
EOL
apply
in
favor
of
more
descriptive
verbs
all
right.
What
do
you
saying
here.
B
This
is
probably
going
to
be
the
last
one
illustrate
some
surprising
Behavior,
with
apply
I
found
repeatedly
that
humans
are
bad
at
predicting.
What
apply
will
do
so
from
talking
to
a
lot
of
people?
The
apply
action
feels
like
it
means
create
or
replace,
but
it's
really
not
it's
much
more
sophisticated
resource
co-ownership
system.
B
B
Just
let
them
any
bugs
over
the
years
in
hindsight,
I
wonder
if
apply
is
the
right.
Oh
dear
God,
no
I,
don't
think
that
we're
gonna
I,
don't
think
we're
gonna
rename
this.
A
B
I
do
see
that
they're
tagged
in
it
all
right,
yeah
sure.
B
Oh
I'm
not
going
to
put
my
personal
thoughts
into
this
from
my
into
the
comment
here,
but
my
personal
thought
is
user
education
rather
than
try
them.