►
From YouTube: Kubernetes SIG CLI 20230104 - Bug Scrub
Description
No description was provided for this meeting.
If this is YOUR meeting, an easy way to fix this is to add a description to your video, wherever mtngs.io found it (probably YouTube).
A
A
Hello
and
thank
you
for
joining
us
for
the
first
six
CLI
bug
scrub
of
2023.
My
name
is
Eddie
zaneski
and
I
will
be
your
host
today.
Just
a
quick
reminder
that
this
meeting
in
all
kubernetes
meetings,
abide
by
the
cmcf
code
of
conduct.
So
please
be
excellent
to
each
other.
A
B
Yeah
sure
the
way
I
got
started
with
kubernetes
is
through
contributing
to
the
Hindi
localization
project.
Over
there
we
were,
you
know,
referencing
about
120
or
128
comments
lately,
and
that
led
me
to
discover
more
what
there
is
to
kubernetes
and
just
jumping
on
to
the
various
ways
to
get
involved.
B
A
Awesome
welcome
anyone
else.
C
Hi
I'm
Caitlin
I've
been
around
a
few
times,
not
very
consistently
I
work
with
six
security
and
have
just
recently
started
the
kubernetes
foundations
program
that
contrida
is
running
at
Shopify,
so
I'll
be
working
directly
with
her
for
the
next
six
months.
So
you'll
see
a
whole
bunch
of
me
working
on
as
much
sexy
live
stuff,
as
I
can
well.
I
have
her
knowledge
to
to
benefit
from.
A
And
managing
that
cool
I
think
that's
everyone
on
the
list.
So
with
that
anyone
have
any
issues
or
PR's
they'd
like
us
to
take
a
look
at
first.
D
Yeah
I
think
about
one
I
can't.
D
The
threat
was
Kim
as
usual.
Kim
is
doing
great
job
for
us
and
it's
something
that
I.
C
D
We
could
improve
significantly
the
user
experience
around
when
there
is
no
keep
contact
specified.
I
know
that
I
think
we
still
do
provide
some
weird
messages
about.
Oh
we're
trying
to
default
you
to
localhost,
which
is
pointless
at
this
point
in
time,
and
the
proposal
from
Kim
was
to
put
a
fat
warning
about.
Oh,
we
just
didn't
found
any
reasonable,
Cube
Conflict
for
you
to
work
with
so
or
something
like
that.
D
Pretty
reasonable
yeah
and
the
fix
could
be
pretty
simple,
so
yeah.
If,
if
someone
is
interested
to
picking
it
up,
I
think
this
could
even
be
somewhat
reasonable.
First
thing:
issue:
I
have
also
something
slightly
related
on
my
list:
I
just
need
to
find
about
a
default
into
localhost,
which
we
still
have
somewhere
in
I,
think
triangles,
but
I.
Oh,
maybe
I
found
it.
D
Oh,
it
was
closed
already
I
tasted,
another
PR
about
stopping
defaulting
key
conflict
to
localhost
8080
I.
Remember
when
I
tried
this
a
while
back,
a
ton
of
a
ton
of
tests
started
failing
because
we
have
a
bunch
of
places
that
will
lie
on
on
that
localhost,
and
so
the
change
itself
was
reasonable.
D
It's
just
that
it
required
a
ton
of
other
places
that
would
need
to
be
chased
down.
I
can
probably
find
even
mrcr
that
I
had
a
while
back,
which
was
trying
to
fix
it.
E
D
Oh
I
remember:
we
had
to
revert
it
so
I
remote,
the
defaulting
a
while
back
and
then,
unfortunately
the
it
changed
the
behavior
with
local,
which
I
remember
correctly,
so
a
couple
of
interconnected
stuff
that
probably
could
be
sold
at
once.
If
someone
is
interested
in
the
ink
through
several
layers
or
PRS
and
history.
D
Quite
a
bit
of
stuff
to
go
through,
but
that
would
be
actually
something
really
worth
it
fixing
because
we've.
D
This
issue
for
for
quite
a
while
and
yeah.
That's
definitely
that
particular
issue
about
having
local
requiring
a
valid
conflict
is
completely
one
of
the
worst
things
that
we
could
have,
because
local
should
be
local
without
any
concert,
but
yeah.
B
A
A
All
right,
this
is
1341
jsonpath
support,
Marley
I'm,
going
to
point
to
you.
I
just
forgot
to
do
the
triage
accepted.
D
I,
remember
that
we,
oh
there's
a
set
of
networking
related
issues.
I
think
that's
even
in
the
client
so
where
we
do
retry
the
operation,
but
other
than
that
I'd
be
very
careful.
Extending
to
the
list.
I
would
have
to
look
it
up
for
where
that
was,
and
adding
some
kind
of
a
retry
would
be
will
be
seen
as
something
of
a
higher
level.
D
I
guess,
I'll,
probably
just
cuddle
invocation
and
some
kind
of
retries,
rather
than
adding
this
kind
of
thing
in
in
Cube
cuddle,
especially
that
he
did
not
mention
any
particular
command.
G
H
I
think
the
CI
tools
probably
have
retry
mechanisms
for
steps
and
that's
what
I
would
suggest
if
it's
a
problem
in
CI
or
building.
If
it's
CI
then
like
it
shouldn't
be
too
bad
to
like
build
it
into
your
script,
to
do
retries,
it's
not
like
you
have
to
type
it
all
the
time
or
anything.
I
As
as
much
I
was
saying,
the
only
place
that
we
do
do
the
retry
in
Discovery,
and
we
we
did
that
because
there
was
you'd,
ask
for
a
group
and
then
you'd
ask
for
the
resources
in
that
group.
It
was
it
used
to
be
the
case
that
the
discovery
was
in.
It
was
a
multiple
ask.
You
had
to
ask
one
part
and
then
from
that
you
ask
for
another,
but
I:
don't
I,
don't
think
we
do
it
elsewhere.
E
I
H
H
Correct,
but
what
Katrina
was
saying
I
think
is
valid
like
if,
if
you
use
like
Dash
o
name
and
we
Emit
and
I
don't
know,
does
it
emit
that
in
that
format,
but
if
they're
just
piping
something
into
it
and
it's
got
the
resource
in
front
of
it
that
can
be
kind
of
annoying
I
don't
know,
would
they
really
be
doing?
That,
though,
is
that
a
normal
use
case?
A
A
Continue
to
make
a
good
point
any
objections
to
that.
If
someone
does
open
a
PR
to
to
add
this
in.
I
The
there
still
might
be
a
possibility
that
we
remove
copy
altogether
from
Coop
control.
Right
I
mean
that
was
a
suggestion,
because
there
were
so
many
cbes
and
you
know
pull
it
out
and
put
it
in
a
plug-in
like
we
did.
Convert.
D
A
Yeah
we
talked
to
Micah
housler
and
wherever
the
last
kubecon
was
Detroit
Mike
a
house
slurred
on
the
security
committee,
definitely
was
in
favor
of
pulling
it
out.
I
think
I
brought
up
last
time.
Every
time
I've
talked
to
Phil
Estes
and
the
oci
folks
and
the
not
oci
CRI
folks.
Everyone
thinks
that
a
an
API
built
into
CRI
would
make
sense.
It's
just
no
one
wants
to
work
on
it.
D
E
A
A
A
G
I
guess
by
that
label,
do
we
also
mean
like
we
need
evidence
from
the
plug-in
prototype
that
this
is
popular
enough
to
except
it's
a
feature
not
just
that
the
Prototype
needs
to
exist.
G
Yeah
that
may
be
worth
developer's
standard
message
for,
since
it's
something
you
run
into
all
the
time
it's
like
and.
F
I
What
Sig
was
in
charge
of
the
Pod
disruption,
budget,
resource,
yeah,
I'm,
sorry
take
off,
say,
gaps.
Yes,.
D
Well,
it's
a
shared
responsibility
between
sick
apps
and
cigarette
piano
missionary
because
part
of
the
disruption
or
the
eviction
API
is
living
inside
of
the
cube
API
server,
but
the
PDP
itself
as
a
controller
is
part
of
the
say,
gaps
yeah.
A
G
F
G
Yeah,
basically,
that
the
current
behavior
makes
sense-
and
it
kind
of
boils
down
to
like
somebody
trying
to
mix
config
strategies
problem
and
then
their
response
is
like.
Oh,
can
you
make
a
cap
to
ultimately
make
the
config
strategies
work
together?
Well,
that
was
their
last
comment
of
the
looks
of
it
and
I
I.
Don't
think
that's
in
the
cards.
G
Yeah
I
think
I
think
so.
I,
don't
think
the
feature
request
makes
sense
and
I
don't
think
it's
really
feasible
for
us
to
promise
to
make
the
config
management
strategies
compatible
right.
The
they
want
to
be
able
to
use
apply
with
all
the
other
operations.
I.
Don't
think!
That's
something
we've
ever
considered
trying
to
make
work.
B
H
I
guess
I'm
thinking
that,
like
this
code's
in
client
go,
do
we
do
we
have
are
we
are
we,
the
owners
of
client
go
or
oh
right,
because
I
feel,
like
other
times,
I've
tried
to
make
changes
in
client
go.
It
ends
up
being
API
Machinery
that
has
to
approve
it.
I
G
And
the
same
thought
that
it
does
sound
like
if
they're
right
about
what's
happening,
which
we
can't
verify
it
sounds
like
a
client
goes
probably
the
one
that's
doing
the
behavior.
A
E
A
G
F
G
Don't
know
that
it
does
sound
like
a
very
specific
Network
situation
that
is
causing
this,
and
that
could
be
a
reason
for
lower
reports,
whether
it's
coming
from
clyco
or
us
like
it
could
be
a
thing
that
does
happen
exactly
as
described,
and
people
just
don't
run
into
it
because
they
don't
have
they
don't
have
this
very
Niche
problem
coming
up
in
their
environments,
it
doesn't
mean
that
it's
like
worth
fixing
or
I,
don't
know
the
difficulty
of
us
reproducing
it
could
be
related
to
the
difficulty
or
to
the
low
number
of
reports.
H
A
F
E
H
Yeah
I
mean
I'm
somewhat
empathetic
on
on
this,
but
I
do
think.
I,
don't
know
that
we
can
build
all.
We
can't
like
keep
building
a
bunch
of
exclusions
into
into
our
dip,
so
I
think
there's
a
workaround,
so
I
I
think
that
would
be
okay
with
closing
it.
A
H
D
A
Well,
it's
using
the
one
in
my
Go
pass.
Okay,
we
can
take
a
look
at
this
offline.
H
So
there
was,
if
it,
if
it
had
an
equal
sign
in
it,
was
where
the
code
was
basically
just
assuming
that
it
should
split
on
equal
sign,
yeah,
I,
guess
I,
just
pinged
you
on
that.
As.
H
I
H
I
H
Yeah,
oh
well,
I
guess
Echo
template
or
something
like
that
or
yeah
or
or
just
parse
Json
themselves,
but
I
mean
it
doesn't.
It
sounds
I
mean
it
doesn't
sound
crazy
to
me,
but
I,
don't
know
what
other
people
think.
H
H
I
guess
it
kind
of
comes
down
to
like
what
is
the
purpose
of
annotations
versus
I
mean?
Should
they
be?
Are
they
using
annotations
when
they
should
be
using
labels
and
annotations
are
as
if
the
purpose
of
an
annotation
is
to
like
provide
some
additional
control
for
for
a
resource
versus
being
able
to
identify
it?
So
a
label
would
be
more
for
identification.
H
G
The
additional
information
section
links
very
close
duplicate.
D
Yeah,
it
was
expanding
on
the
the
extra
columns
and
I
I.
Think
I
was
okay.
With
that
I
had
to
some
comments.
I
just
never
got
a
chance
to
return
to.
We
reviewed
this.
A
E
D
Yeah
I
mean
that's
weird,
because
what
he
would
have
to
do
is
request
a
ble
to
Output
with
including
the
entire
resource,
because
you
can
do
that
and
then
you
would
have
to
extend
the
current
printing
mechanism
by
adding
those
additional
columns
that
you
care
about.
So
it
shouldn't
be
that
hard.
D
It's
just
that
yeah
somebody
would
have
to
do
it,
because,
basically,
what
currently,
how
this
currently
works
is
the
server
tells
us
what
kind
of
columns
and
values
we
have
to
print
and
by
default
we
don't
request
the
full
resource,
which
is
just
another
column
in
the
table
that
you're
going
to
get
from
the
class
from
the
server.
D
But
you
can
do
that
with
additional
flag
that
in
not
in
Cub
cuddle,
but
in
the
in
the
request,
because
we
do
that
for
sorting
and
then
you
would
have
to
build
on
top
of
the
current
printing
mechanism
such
that
it
is
printing,
as
it
is
printing
parenthesi
and
then
at
the
additional
columns
coming
from
the
extra
column.
Slack.
D
So
theoretically
should
be
simple.
I
haven't
looked
that
deeply
into
the
code,
although
I
still
have
that
particular
PR
on
my
list.
So.
D
Yeah
I
would
prefer
to
have
more
generic
mechanisms,
like
the
one
NWR
that
Eddie
has
opened,
rather
than
adding
specific
columns,
because
if
we
will
open
the
doors
for
annotations
I'm,
pretty
sure
that
someone
else
will
will
show
up
soon
with
yet
another
field.
Next
time,
I
agree,
so
that
other
option
is
just
a
a
better
option
for
the
future.
It's
just
that
we
would
have
someone.
D
You
need
to
pick
it
up
again
and
see
where,
where
the
crew
HDR
ended
and
I
could
probably
walk
someone
through
how
this
could
be
done
or
where
to
look
for
places
to
extend.
A
A
F
E
A
A
D
Yes,
because
we
are
the
ones
invoking,
the
eviction
on
the
pots
can
I.
Remember
correctly,.
A
G
G
Comes
right
to
me
for
me
is
like
do
you
need
to
think
through
what
else
people
might
want
to
order
their
training
based
on?
Could
there
is
this?
The
only
thing
is
this:
the
most
important
thing
like,
especially
if
we're
doing
the
flag
is,
you
know
we're
introducing
a
flag
here
for
changing
the
ordering
Behavior.
Are
we
then
going
to
get
well?
We
actually
want
this
different
ordering
behavior,
in
which
case
there's
like
a
bigger
feature
about
ordering
that
has
different
modes
or
something
like
that.
D
Yeah
more
like
a
generic
approach.
If
we
would
want
to
do
and
maybe
rather
than
having
a
specific
approach
having
more
generic
approach
will
be
better,
he
did
mention
in
the
one
of
his
last
comments
on
the
pr
he
will
be.
D
He
got
to
talk
about
it
on
contributor
meeting
and
finalize
so
I'm,
not
sure
if
she's
talking
about
sexy
live
or
domain
computer
all,
but
maybe
it
would
be
nice
to
get
him
jump
on
one
of
the
60i
max
calls
and
present
the
topic
and
maybe
ask
him
for
what
are
the
other
use
cases
that
people
might
have
before?
We
will
decide
on
any
particular
direction
together.
H
A
D
So
from
this
quick
skin
that
he
has
on
the
pr
we
just
list
pots,
so
I
don't
expect
in
a
particular
order
and
I
was
considering
the
By
Priority
being
the
default,
which
would
be
okay-ish.
D
D
If
we
will
hear
more
options
about
oh
yeah,
we
want
to
wanna
at
additional
options
in
the
future.
Maybe
then
doing
something
generic
would
be
better
from
start.
If
the
prioritization
is
the
only
one,
then
I
could
probably
start
with
just
the
prioritization,
but
then
again,
I
could
probably
easily
be
persuaded
that
the
priorities.
So
that
means
that
the
critical
should
be
more
important
than
less
critical
or
the
other
way
around,
because
you
want
to
have
those
less
critical
or
the
more
credible
to
be
moved
as
the
last
one.
A
With
that,
we
are
out
of
time
super
great
job.
Today
we
got
through
a
lot
and
we
don't
have
much
on
triage
right
now,
which
is
awesome,
so
people
weren't
breaking
things
over
quality
break.
It
seems
so
I'll
get
those
issues
cleaned
up
and
we'll
see
you
all
next
week
at
the
Sig
meeting
thanks
everyone.