►
From YouTube: Kubernetes SIG CLI 20220202 - 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
Hello
and
thank
you
for
joining
us
for
the
february
2nd
six
cli
bug
scrub.
My
name
is
eddie
zaneski
and
I'll
be
your
host.
Today
we
have
a
code
of
conduct
that
everyone
is
expected
to
abide
by,
but
from
the
cncf.
So
please
be
excellent
to
each
other,
and
with
that
do
we
have
any
new
folks
on
the
call
that
would
like
to
introduce
themselves.
B
Hey
hi
hi,
guys,
hello,
hi,
hi,
I'm
william,
and
I
I
work
for
chuck
to
talk.
No,
no,
I
don't
I
work
with
him
yeah.
I
would
like
to
join
and
see.
What's
this
all
about
and
hope
to
contribute,
that's
awesome.
Thank
you
for
joining
us.
Hey.
C
C
Yeah,
I
just
stopped
I'm
nash,
I've
attended
as
six
eli
meetings
earlier,
but
this
is
my
first
time
on
the
box
club.
So
I'm
looking
forward
to
thank
you.
A
A
This
is
a
feature
request,
add
an
option
for
cube
control
diff
to
exit
zero.
When
differences
are
found,
this
is
I'm
just
going
to
link
them
to
the
cap.
A
I
don't
think
this
is
actually
in
the
spec
at
all,
because
I
actually
looked
through
the
ingress
definition
and
caps,
and
so
this
was
implemented
with
the
original
v1
ingress
and
the
user
has
an
issue
because
their
users
don't
have
our
back
permissions
to
read,
cube
names,
cube
system
which
totally
makes
sense
in
a
real
organization,
and
so
their
users,
when
describing
an
ingress
without
a
default.
A
A
I'm
hesitant
to
remove
what's
in
there
as
a
kind
of
convention
or
default,
I
suggested
to
use
that
we
could
wrap
the
error
a
bit
better
to
make
it
clear.
A
F
Is
there
really
a
default
backend
and
if
so,
why?
Why
is
the
server
not
not
telling
us?
I
guess
I'm
wondering
why
describe
would
ever
make
any
assumptions
about
about
something
being
there
when
it's
when
it's
reading
it
out
like
it
seems
like
a
default.
I
don't
know,
I
guess
I'm
just
asking.
Should
the
server
tell
us
that
this
is
the
back
end,
even
though
it's
not,
maybe
it
is
the
default.
A
Yeah,
I
agree
with
you
brian
the
I
mean
like
I
said
it's
not
in
the
spec,
but
this
comment
kind
of
calls
out
that
it
is
so
yeah.
I
don't.
I
don't
have
a
good
way
to
approach
that
without
knowing
so
maybe
I'll.
Just
ask
whoever
implemented
the
original
pr.
A
A
D
G
A
There's
a
lot
of
like
more
rich
terminal
stuff.
We
could
do
in
the
future
when
bam
with
permits.
This
is
definitely
one
of
them.
Another
thing
I'd
love
to
have
is
just
like
you
know
we
have
the
config
set
and
marcus
has
been
working
on
the
pr
to
get
that
to
work
with
all
fields.
A
I'd
love
to
just
have
like
a
built-in
terminal,
ui
editor
type
thing
to
make
it
easier
for
users
to
like
switch
and
edit
their
contests
on
the
fly,
and
but
that's
like
you
know
that
those
are
all
nice
to
have
things
and,
and
I'm
sure,
there's
cis
calls
that
clear,
the
terminal
and
refresh
it-
and
I
don't
know
I
know
n
curses-
is
used
for
that.
So
yeah,
I
don't
know
what
the
go
equivalent
of
all
that
is.
A
A
A
F
Yeah,
I
think
it
was
basically
just
preventing
it
from
getting
into
this
state
where
it's
still
running,
but
you
can
never
connect.
A
Because
chance
like
it
could
possibly
be,
the
user
was
just
that
was
happening
like
their
pod
was
dying
and
they
were
just
okay
with
that
and
didn't
realize
it
or
I'll
I'll.
Ask
some
more
questions
on
that
thread.
Unless
you
want
to
chime
in
there
brian.
F
A
A
A
I
talked
to
katrina
about
this
for
a
bit.
It's
a
tricky
one,
because
it's
it's
a
real,
simple
ad,
but
we're
trying
to
avoid
adding
more
flags
to
the
name
space,
and
this
is
like
a
one-way
door
type
decision
where
you
know
once
we
add
it,
we
have
to
keep
it
so
I
I
didn't
think
that
the
benefit
was
there
to
add
a
alias
for
another
flag.
Take
up
another
flag
there.
A
A
A
A
Me
yeah,
okay,
I
did
the
exact
same
thing
basically,
but
I
guess
since
we're
I've
made
this
because
we
were
gonna,
I
was
gonna.
Do
two
pr's
and
some
things
happen.
I
guess
so
now,
I'm
only
doing
one
pr
for
all
of
this
edit
well,
edit
all
keys
you,
okay,
if
I
just
edit
the
title
like
that
yeah
I
mean
you
can
also
just
close
your
duplicates
or
because
of
the
person
who
the
original
off
info
issue
that
I
was
responding
to.
Oh
okay,.
A
Can
you
would
you
take
care
of
that?
Please
yeah
I'll,
take
a
look
at
it.
I
don't
know
if,
if
I
need
permissions
I'll,
just
ping,
you,
oh
you
should
be
able
to
close
your
own
issue.
Okay,.
A
A
A
A
A
A
A
E
F
Trying
to
understand
so
is
the
case
use
case
here
that
you
don't
know
what
image
name
is
running
on
that
project,
or
on
that
in
that
that
you
don't
know
what
image
name
is
running
in
that
deployment.
You
just
want
it
to
go
to
a
certain
version,
whatever
image
it
happens
to
be.
Is
that
what
it's
saying.
E
F
E
A
Yeah,
I
agree
with
you
y'all
that
that's
our
would
you
agree.
This
is
our
sentiment.
This
trade
of
all
trade
off
of
more
complexity
for
convenience
isn't
worth
it
here.
A
E
The
convenience
and
and
risk
associated
increase
in
risk.
F
Yeah,
it
also
kind
of
feels
like
something
that
you
could
script,
something
for
like.
If
you
find
yourself
constantly
having
to
rev
the
version
like
you
can
just
create
a
small
dash
script
or
something
that
that
retrieves,
the
version
increments
it
or
sets
it
to
the
new
version
and
then
runs
this
command
yeah.
That's
fair!.
A
The
logs
are
just
full
on
hitting
the
timeout,
so
there's
this
the
timeout
seconds
of
566,
which
is
what
like
nine
minutes
or
something,
and
that
is
what
the
the
timeout
is
hitting.
So
it's
basically,
it's
printing
out
success
and
then
it's
sitting
on
success
hanging
with
the
connection
open
for
nine
minutes,
waiting
for
the
round
tripper
to
timeout
and
close.
A
F
A
E
So
this
is
where
brian
picked
me
up
there
and
figured
that
out
thanks
brian.
F
A
I
I
You
basically
mean-
and
I
honestly
don't
see
a
reason
where
we
would
expose
all
the
customized
facts
through
this.
I
wonder
if,
if
we
could
somehow.
F
I
E
I
E
A
I
A
A
Okay,
cool
this
one
is
an
older
1154.
We've
talked
about
this
a
few
times.
This
was
the
context
and
namespace
environment
variables.
A
I
had
said
basically
everything
we
talked
about
that
this
changes
couldn't
live
in
cube,
control,
it'd
need
to
live
in
client
go,
so
it
would
probably
need
a
cap
to
introduce
and
change
this
behavior,
and
we
come
back
to
the
problem
we
always
come
back
to,
which
is
whenever
we
make
a
change
to
how
configurations
are
loaded.
There
is
a
huge
period
of
inconsistency
with
tools
and
upgrading,
and
we
just
haven't
been
able
to
do
that.
So
I
think
I'll
just
close
this
for
now.
I
Which
reminds
me
did
you
manage
to
put
together
that
proposal
for
for
the
rc
files?
No,
is
it
too
late?
You
basically
have
today
and
tomorrow
to
get
it
through
reviews
all
right.
I
are
you
available
to
review
it,
because
I
I
can
ship
it.
I
If
you
put,
if
you
pull
me,
if
you
send
me
the
link
to
the
to
the
enhancements,
yes,
the
the
blocker
might
be
the
prr.
You
need
to
get
someone
from
the
prr
team,
which
includes
john
ilana,
david
or
wojtek,
so
at
least
one
of
them
is
from
europe.
Does.
I
Still,
everything
requires
pr
review,
even
if
it's
alpha,
okay,
you
can
check
the
ricardo
had
to
add
the
pr
just
recently,
which
is
which,
by
the
way,
we
got
it
merged
yes,
today
or
earlier
today,
so
yay
yeah,
so
that
at
least
one
thing
is
slowly
moving
forward.
All
right.
A
The
thing
I
struggle
with
is,
I
still
don't
know
how
we
want
to
do
it,
especially
from
all
the
conversations
and
the
one
we
had
with
phil
where
phil
was
like.
I
want
to
be
able
to
do
this
per
context
and
have
different
configs
per
context,
and
so
I've
just
been
sitting
at
a
dock
and
rewriting
it
over
and
over
again,
but
we
can
talk
about
that
later.
A
I
My
dad
it's
not
it's
not
that
you
have
to
finish
it
even
if
it
slips
124,
there's
nothing
stopping
us
from
playing
with
ideas.
We
don't
have
to
merge
anything,
but
we
can
have.
I
don't
know
branches
where
we
will
be
working
on
it
and
playing
with
it,
and
you
know
that's
fair.
Maybe
we
could
maybe
in
124
we
could
work
on
having.
I
I
don't
know
two
three
proposals
up
into
proof
of
concepts
working
and
see
which
one
we
will
like
more
both
from
the
implementation
standpoint
as
well
as
from
from
the
usability
we
can.
I
don't
know,
devote
entire
6
cli
call
one
of
the
future
calls
to
discussing
this,
but
it
would
be
good
to
have
some
demo
initially
and
then
we
can,
based
on
that,
make
the
final
call
and
from
that
create
the
the
cap.
Instead
of
rushing
it
through
okay.
A
I
A
A
All
right,
we
got
15
minutes
left
11
20.
waiting
for
pods
hangs
indefinitely.
If
a
pod
is
concurrently
scheduled
for
termination
and
removed
by
a
deployment.
I
B
D
All
right,
I
I
was
saying
I
didn't
know
this
issue
exists
before
basically
like
way
just
getting
the
thought
like
every
every
few
seconds
and
then
refresh
it
and
see
if
that
match
what
we
want.
I
So
I
I
don't
remember
what
the
implementation
details
look
like
for
weight
currently,
but
there
were
problems
with
wait
until
and
all
those
functions,
at
least
in
controllers,
so
we
switch
to
informer
based
weight,
loops,
not
sure
if
or
how
we're
approaching
that
in
the
weight
command.
Maybe
that's
something
that
we
could
easily
solve
with
with
this
switch.
D
D
I
was
just
looking
at
how
how
the
the
original
way
function
was
implemented
and
just
how
to
do
the
same
thing.
I
A
I
Hold
on
give
me
a
sec.
Yes,
it's
still
on
my
list.
I
haven't
looked
at
it
and
I
think
I
find
the
least
watch.
I
Yeah
but
of
course
we
don't
have
it
in
the
trying
to
find
where
we
are
using
this
and.
D
I
List
watch
and
that's
a
that's,
a
wafer
part
implemented
using
the
the
watch
tools
with
informers
right
in
the
run
command.
I
I
remember
we
were
working
on
a
lot
of
this
those
bits
and
we
were
shifting
all
the
controllers,
all
the
commands
tests
and
everything
towards
using
this
approach,
just
using
the
informers
behind
the
scenes
and
it's
much
more
reliable,
especially
when
something
happens
in
between
in
the
communication,
so
that
it
actually
retriggers
reconnects
the
sync
again
and
you
can
actually
work
and
have
this
do
some
initial
conditions.
A
I
It's
it's
something
that
if
weight
isn't
using,
it
should
not
sure
how
far
we
can
go,
because
that
might
be
only
for
the
built-in
types
which
might
be
tricky
for
for
the
weight
command
itself.
A
E
Yeah
exactly
it's
the
go,
laying
version,
so
you
can't
build
it
before
version
121.,
so
kubernetes
is
being
built
with.
I
think
it's
go
117
now
I
think
maybe
it'd
have
been
116
back
to
121.
E
There
is
no,
you
know,
the
previous
versions
didn't
have
the
ability
of
go
lag,
didn't
have
the
ability
to
build
darwin
md64
before
116
golan,
116
and
before
121
kubernetes
was
built
with
115.
A
A
I
A
I
That
looks
like
a
bug
on
on
the
client
go.
Something
tells
me
because
we
are
constructing
the
to
get
the
wrong
way.
Honestly,
I
have
no
idea
what
it
should
look
like
correctly
for
ipv6.
I
I
A
A
I
assume
it's
just
a
convention.
I
don't
have
the
context
as
to
why
we
did
it.
This
way.
Do
you
happen
to
know
and
the
the
root
cause?
Is
the
users,
the
users?
Users,
don't
have
our
back
permissions
to
get
cube
system
which
makes
sense
in
a
real
world
and
so
they're
getting
confusing
errors
when
they
describe
this.
G
H
A
H
It's
a
it's
a
whole
big
project,
because
we
have
dozens
of
issues
that
have
been
opened
against
customized
for
the
helm,
integration,
it's
not
working
out
for
us
and
it
we
have
problems
like
this,
where
people
have
expectations
around
it
and
working
and
keep
controlled
over
it.
Never
will
so
yeah
we're
well
on
our
way
towards
having
an
external
plug-in
for
it
that
folks
can
use
so
that
it's
it
is
making
progress
and
I'll
take
up
the
issue.