►
From YouTube: Kubernetes kops office hours 20201023
Description
Recording of the kops office hours meeting held on 20201023
A
Hello,
everybody
today
is
friday
october
23rd
2020..
This
is
the
bi-weekly
chaops
office
hours
meeting.
I
am
a
moderator,
facilitator,
justin,
santa
barbara.
I
work
at
google
a
reminder.
This
meeting
is
being
recorded
and
will
be
put
on
the
internet
and
to
please
be
mindful
of
our
code
of
conduct,
which
generally
boils
down
to
being
a
good
person
and,
using
the
raised
hand,
feature
if
we
get
too
many
people
talking
at
once.
A
We
have
an
agenda.
I
put
a
link
to
that
in
the
chat.
Please
do
feel
free
to
add
your
items
there
in
presumably
in
the
open
discussion
section,
but
anywhere
that
where
it
fits,
please
also
feel
free
to
add
your
name
into
the
attendees
list.
It
can
be
helpful
for
people
watching
the
video
on
youtube
to
know
who
you
are
otherwise
we
will
just
write
jump
right
into
it.
We're
a
couple
of
minutes
late,
so
we
will
get
going
first
up.
A
Our
first
sort
of
section
is
to
review
action
items
from
last
time
and
it
sounds
like
peter.
You
created
the
release,
119
branch.
A
Wonderful,
thank
you
for
doing
that
and
on
a
similar
note,
119
0
alpha
5
cyprian.
Do
you
want
to.
C
Oh,
that
went
out
pretty
smooth
this
time,
didn't
need
any
help
to
get
it
released,
at
least
on
github,
so
the
other
parts
are
still
with
justin,
but
at
least
for
alphas
and
betas,
where
we
don't
have
that
much
traffic,
I
think
it's
reasonable
to
do
it
first,
to
do
it
any
one
of
us
and
maybe
for
official
releases
until
we
get
the
artifacts
automatic
publishing,
you
will
still
have
to
to
do
it.
A
Right
and-
and
so
the
big
win
I
think
right
now
is-
I
think
you
changed
the
order,
so
we
don't
need,
we
don't
need
any
of
the
privileged
buckets
yet
or
the
more
privileged
buckets.
Yet
we
don't
need
them.
They
are
helpful
for,
as
you
say,
for
the
more
popular
releases,
but
we
can
a
github
really
suffices.
Is
my
understanding?
Is
that
right,
yep.
A
Perfect,
the
automation
is
continuing,
and
next
up
is
the
motivation
for
the
automation,
which
is
118.2,
which
I
was
supposed
to
do
and
is
not
automated,
and
I
did
not
do
it
and
I
apologize
and
I
would
have
written
it
down
on
my
list
and
I
will
do
it.
I
think
this
afternoon.
I
think
there
was
one
blocker
which
was
resolved
previously,
so
there
are
no
active
blockers
assuming
that
stays
the
same
I'll.
Do
it
this
afternoon.
A
D
Or
anyone
else
well,
last
I
talked
with
him
he's.
I
think
peter
had
asked
him
to
split
out
the
was
the
length
of
the
target
group
and
we're
trying
to
figure
out
how
that
fits
with.
You
know
whether
what
should
be
a
separate
task
and
such
it
looks
like
inside
the
nlb
are
structs
for
the
target
groups
which
aren't
passed
and
then
the
tasks
seem
to
be
associated
with
the
target
groups.
D
A
Peter
you're
here
right:
do
you
wanna?
Is
there
anything
any?
Is
that
anything
particular
there
you
wanna
bring.
B
A
Well,
we
also
previously
talked
about
splitting
up
that
pr
into
introducing
nlb
support
and
then
a
smooth
transition
to
in
to
add
nlb
support
to
existing
clusters
is
is:
are
we
doing
that
or
not.
A
It
nice,
okay,
great
the
next
action
item,
review
action
items
for
last
time
is
deciding
on
the
next
meeting
day
and
time
for
both
our
regular.
A
I
want
to
say
bi-weekly,
but
our
regular,
this
this
currently
bi-weekly
friday
meeting
and
the
october
next
oktoberfest.
We
have
been
doing
these
oktoberfest
meetings,
impromptu
gatherings,
discussion,
things
which
I
think
have
been
valuable
and,
I
think,
are
different
from
what
we
are
doing
here.
A
I
don't
know
if
we
get
have
enough
data
to
pick
a
different
day
for
this
meeting.
I
don't
know
if
anyone
feels
that.
That's
that
there's
a
strong
reason
to
pick
another
day
right
now,
when
are
the
hacked
over
fast
sessions
happening?
Are
there
they're?
Just
it's
not
at
a
fixed
time,
so
we
will.
We
can
now
decide
the
next
one,
which
will
be
next
week,
ideally
the
regular
time,
okay,
yeah
we're
trying
to
like
find
a
good
time.
A
A
Should
we
try
another?
Should
we
try
a
later
one
on?
I
think
it
also
we.
We
do
actually
lose
some
people
in
europe
if
we
go
later
as
well
as
we
expected,
but
and
we
I
think
we
lose
in
west
coasters
as
well,
because
unless
we
go
really
late
just
because
of
the
way
the
day
falls.
So
I
don't
know
if
anyone
has
a
particular
time
they
would
like
to
try
for.
Ideally,
we
would
do
one
next
week
either
on
tuesday,
wednesday
or
thursday
would
be.
A
A
Should
we
try
a
10
o'clock
on
tuesday
eastern
time,
7
o'clock
am
pacific
time.
How
does
that
feel
for
people?
C
So
I'm
okay
any
hour
as
long
as
it's
not
monday
or
friday.
Okay,.
A
So,
okay,
so
then
it
sounds
like
10
a.m.
On
on
tuesday,
which
it
looks
like
it's,
the
27th
would
be
our
next,
so
impromptu.
E
A
A
E
E
D
Okay,
yeah:
it's
it's
not
good
for
me,
but
you
know:
go
ahead,
yeah
the
the
two
hours
before
this
are
difficult
for
me
because
their
morning
routine,
so
I
can
do
before
then,
but.
A
Okay,
I
mean,
I
think,
that's
okay,
yeah,
we
will
miss
you,
but
I
think
that's
sort
of
part
of
the
idea
already
so
try
to
understand
fair
enough
cool,
all
right
and
then
the
I
think,
that's
the
that's
the
last
item
on
our
review
action
items
from
last
time.
Unless
there's
anything
else,
anyone
recalls,
if
not,
we
can
jump
into
the
open
discussion.
A
The
first
item
on
our
open
discussion
is
a
fun
one
cyprian
and
I
were
chatting
about
the
the
brand
cops
and
chaos
and
the
potential
to
go
like
do
a
rebrand
and
to
emphasize
the
ops
bit
and
perhaps
de-emphasize
the
cops
which
seems
reasonable
at
this
particular
moment
in
history.
Just
because
it's
it's
yes
and
then
so
I
don't
know
how
people
feel
about
that.
A
C
C
A
That
right,
we
would
then
be
have
a
very
clear
like
it.
You
wouldn't
yeah,
because
I've
seen
I've
seen
old,
lower
case.
I
know
uppercase
and
I
to
be
clear.
The
the
cli
tool
would
remain
the
same.
You
wouldn't
have
to
type
a
capital
o
unless
you
wanted
to
aliase
it
yourself,
the
yeah.
So
I
I
don't
know
what
people.
E
E
It's
been
fun
because
we've
described
how
to
say
cops
or
chaos
to
people
and
had
for
so
many
years
to
actually
have
an
opinion
on
it
and
a
stated
opinion
is
quite
it's
it's
exciting
in
a
way.
It's
also.
You
know
it
feels
a
little
awkward
to
me
but
like,
but
but
also,
but
that's
probably
just
because
I've
been
saying
cops
for
so
long.
E
A
You
can
try
it
and
go
back,
it's
like.
I
don't
think
I
don't
think
we're
gonna
start
dictating
to
people
how
they
will
pronounce
the
words.
E
C
A
I
didn't
know
you
were
an
artist
yeah
right
copy
paste.
Well,
that's
really
good!
I
also
really
really
like
the
the
o
with
the
the
stylized
k
in
it,
like
I
actually
really
like
that,
like
I
could
see
us
making
hoodies
or
something
with
that
on
it.
That's
that's
pretty
cool
all
right.
A
This
is
in
this
pr
for
people
that
are
not
it's
in
one:
zero,
zero,
seven,
seven.
E
Yeah
one
thing
I
don't
like
about
our
current
logo
is
that
the
the
fav
icon
on
the
top
of
our
web
page
like
looks
like
a
little
sheriff's
badge,
which
is
doesn't
help
don't
fit.
The
point.
C
From
yeah,
I
would
cut
this
o
and
make
it
the
fav
icon.
E
I
I'm
generally
in
favor,
I
got.
I
have
no
complaints,
but.
A
It's
one
of
those
things
which
is
like
you
know.
I
think
it's
a
good
idea.
People
are
more
inclined
to
like
you
know,
never
cheat
whenever
gmail
does
a
redesign
everyone's
like.
Oh
no,
you
broke
my
gmail
and
so
like.
If
we're
not,
if
we're
not
screaming
about
it,
then
it's
probably
a
good
sign
that,
like
it's
actually
a
good,
a
good
design,
because
it
is
always
there's
a
big
like
resistance
to
change
natural
natural
resistance
to
change.
So
if
we're
not,
if
we
don't
hate
it,
that's
a
very
concerning.
E
Is
it
would
it
be
reasonable
or
useful
to
put
some
sort
of
a
statement
alongside
like?
Is
it
or
is
that
taking
too
political?
Like
a
viewpoint,
I
I
think
we
already
have
the
the
we.
A
I
don't
think
it's
I
don't
I
I
I
don't
think
I
I
don't
think
we're
making
that
much
of
a
statement,
or
I
don't
intend
for
us
to
make
that
statement.
It's
more
just
like
we're
trying
to
avoid
something
that
is
avoid
making
the
wrong
statement.
Yeah.
G
A
A
Yeah,
the
you
can
also
just
put
it
in
if
it's
an
svg
that
people
can
edit
right,
you
can
also
put
it
in
and
like
that's
the
sort
of
great
like
help
wanted
type
issue
for
a
sort
of
a
new
type
person
or
a
different.
I
mean
you
obviously
are
capable,
but
you're.
You
know
there
are
people
that
that
are
dedicated
designers
or
full-time
designer
type
people.
A
I
mean,
or
just
yes
and
put
it
in
yeah,
do
that
put
it
put
in
your
version,
as
is
don't
worry
too
much
about
it
and
say
like
look
if
someone
has
an
eye
for
this
it'd
be
great
to
like
tweak
it
and
move
it
yeah.
E
A
Yeah
he
shows
up
to
the
the
hecktoberfest
meetings.
A
I
think
he
said
he
might
not
be
able
to
show
up
if
it
was
at
10
eastern,
but
we
can
see
if
he
shows
up
and
or
just
ping
him
again
and
say
like
we
generally
thought
this
was
a
no
one
had
particularly
strong
objections
to
this.
So
are
you,
okay
with
us
trying
it
cool.
A
Okay,
that
that
was
great,
like
I,
was
expecting
a
lot
more
abstract
discussion.
So
let's
thank
you
for
that
and
let's
go
on
to
a
technical
topic,
one
zero,
zero.
Four
eight
is
around
as
another
one
from
cyprian
hackman
container
runtime
packages
as
assets.
C
Well,
there
are
already
been
two
reviews
on
it.
One
is
justin's,
one
is
jones,
I
think
I
addressed
john's
concerns,
or
at
least
most
of
them.
A
Yeah,
I
I
can
clarify
my
concern,
which
I
wanted
to
just
bring
to
the
group.
I
think
it's
a
good
change.
A
The
concern
was
specifically
around
these
old
versions
and
they
are
old
versions
that
we
thought
we
think
it's
prior
to
kubernetes,
116,
that
we
are
changing,
potentially
changing
the
behavior
in
that
we
are
definitely
changing
how
docker
gets
installed
where
previously
it
was
installed,
using
the
distro
specific
packages,
and
now
it
is
installed
from
the
tar
files
and
the
question
is
like
do
we?
Do
we
adopt
the
position?
A
We
think
it's
the
same.
It's
an
older
version
of
kubernetes
really
pretty
far
back
like
if
it
breaks
we're.
Okay
with
that
and
we'll
sort
of
fix
it,
because
we
think
it's
pretty
unlikely
or
do
we
want
to
adopt
the
contrary
position
where
we
say
it's
an
old
version
of
kubernetes.
We
don't
want
to
break
it
because
the
last
thing
we
want
to
do
is
like
go
back
and
have
to
like
deal
with
those
old
versions.
So
let's
just
keep
the
code,
as
is
for
those
old
versions.
A
So
that's
that
is
the
trade-off
that
I
was
wrestling
with,
and
I
think
cypriot
has
pointed
out
correctly
that
we
do
have
test
coverage
or
more
some
test
coverage
of
these
older
versions.
So
we're
not
flying
blind,
but
we
are,
you
know
we
certainly
are
not
going
to
be.
I
think
ourselves
running
these
running
testing
them
as
much
in
our
development.
C
Well,
we
install
it
and
run
conformance
testing,
but
that
should
be
quite
enough
for
the
container
runtime.
A
Yeah,
I
imagine
we
can
get
adequate
coverage
on
that.
I'm
more
concerned.
It
is,
but
the
the
the
square
is
the
distro
and
the
kubernetes
version
that
I'm
worried
about
and
we
do
have
a
grid
that
can
give
us
some
coverage.
I
just
don't.
I
was
trying
to
look
up
whether
we
have
you
know
like
rel
with
116,
and
it
looks
like
it
looks
like
we
don't
go
back
that
far.
A
We
do
like
specifically.
We
we
do
test
with
116,
but
we
don't
necessarily
test
the
rel
with
116
or
rel
with,
I
guess,
technique,
115,
and
so
we
don't.
A
D
G
A
Realistically
that
that
is
true,
we
we
have
tried
to
say
that,
like
newer
versions
of
cops,
do
work
with
older
versions
of
kubernetes,
specifically
so
that
we
don't
have
to
maintain
those
older
versions
of
cops
for
as
long
okay
apps
for
as
long
the
I
do
agree
with
you
that
likely
a
user
will
stick
with
their
like
older
version
of
cops.
A
C
Older
versions
are
picky
with
tests,
also
from
what
I
remember
when
I
did
the
backwards
right
matrix,
so
I
don't
know
I
could
run
some
tests
and
fix
rel
if
that's
the
case,
but
keep
in
mind
that
until
so,
it's
only
a
rail
seven,
because
until
recently
amazon
linux
and
rail
8
didn't
work
at
all.
So
I
think
it's
a
very
small
percentage
of
users.
A
C
So,
okay
same
no
it
so
I
took
samples
from
all
the
packages
look
inside
the
rpms
and
devs.
So
what
files
they
have
and
didn't
notice
anything.
C
A
A
I'll
I'll
look
at
those
and
then
if,
if
I
don't
find
anything
obvious,
then
I
think
we're
which
I
suspect
we
won't
then
I'm
okay
to
to
merge
this,
as
is
because
it
is
older
versions
and
as
richard
points
out,
it's
unlikely
that
we
that
those
users
will
be
bang
up
to
date
on
on
cops
as
well,
and
this
would
be
in
120
and
not
in
119.
Is
that
true.
A
C
G
G
A
Okay,
let's
move
on
to
the
next
thing
that
does
anything
else
on
that
epic,
because
this
next
one
is
an
amazing
topic.
Yay
peter,
do
you
wanna?
You
wrote
it
down.
I
think
there
are
some
representatives
of
of
I
guess
cloudnatics,
but
do
you
want
to
introduce
it
peter.
B
Sure
so
we've
been
azure.
Support
has
been
our
most
thumbs
up,
github
issue
for
a
long
time
and
recently
kenji,
who
I
believe
is
on
the
call.
A
That
sounds
like
peter
might
have
had
some
background
things
happening,
but
thank
you
yeah.
This
is
around
adding
azure
support
and
we
see
that
there
is
a,
I
guess,
a
fork
or
an
open
source
fork,
which
is
really
great
to
see
and
great
to
have
you
here
welcome,
kenji
and
nick
and
yeah.
A
G
That's
a
great
so
from
us.
I
don't
just
want
to
know
what's
interesting
what
kind
of
procedure
we
should
take
to
create
appear
because
I
can
send
that
one
giant
field
or
that
probably
govern
different
review,
but
also
I
don't
want
to
know
the
procedure
also
also
want
to
know.
I
don't
really
follow
what
community
is
doing,
and
there
was
some
previous
discussion
on
the
cluster
api
and
the
crops
and
want
to
know
if
we
still
want
to
add
this
kind
of
new
csv
support
good
course.
A
Absolutely
yes,
so
I
think
I
think
the
the
great
first
step
would
be
just
just
to
on
your
first
topic
to
just
create
a
an
overarching
pr,
just
with
all
the
changes
in
it,
and
we
can
read
through
that
and
maybe
suggest
some
areas
that
that
could
be
broken
down
into
smaller
pieces.
If
need
be.
A
That
bad,
then
it's
then
we
can
do
that.
What
we'll
probably
do
is
put
the
azure
support
behind
a
feature
flag
so
that
people
won't
accidentally
get
it
and
what
that
also
means
for
you
is
you
aren't
necessarily
you
can
make
breaking
changes?
If
you
want
to
you,
can
you
know
we
can
merge
it
without
as
strict
code
review
policies,
for
example,
okay,
and
so
our
review
of
the
code
will
really
focus
on
the
the
changes
to
the
shared
code.
A
So
anything
that's
azure
specific,
we
can
sort
of
say
well,
that's
behind
a
feature
flag
and
that's
okay,
but
anything
that
changes
the
shared
code
will
be
that
we're
the
where
we
sort
of
really
need
to
pay
more
attention.
A
A
So
that
would
be
wonderful
to
see
that
would
be
we.
You
know,
I'm
you'd
have
a
lot
of
people
very
happy
when
you,
when
you,
when
that
one
goes
in,
that
will
be
wonderful.
So
thank
you
and
then
on
the
cluster
api.
We
do
it.
I
I
do
personally
intend
to
get
cluster
api
support
in
into
into
cups.
I
think
that'll
be
it'll,
be
a
great
thing.
It's
still
a
long
way
away,
and
we
we
also.
A
We
also
want
to
support
the.
We
always
want
to
support
the
not
using
cluster
api,
and
even
when
we
get
cluster
api
in
there,
it's
likely
that
we
will
use
it
primarily
for
the
worker
nodes
rather
than
the
control
plane
nodes.
A
So
there's
still
a
great
need
there
and
there's
also
a
question
around
you
know:
how
do
you
create
your
first
cluster
yeah?
So
I
think
I
think
the
yeah
there's
definitely
still
appetite
for
non-cluster
api
implementations.
I
think
one
of
the
things
that
will
be
very
interesting
is
as
we
move
forwards.
A
To
what
extent
can
we
start
to
share
code
with
cluster
api,
but
that's
two
or
three
big
steps
down
the
road
in
terms
of
first
of
all,
we
have
to
get
you
know
the
support
for
cluster
api
in
and
then
we
can
start
to
think
about.
You
know
how
to
how
to
work
better
together,
but
so
it's
years
away,
four
years
away
from
that,
but
in
the
meantime
yes
it'd
be
wonderful
to
get
this
as
your
support
landed,
and
I
don't.
I
don't
think
it
precludes
cluster
api
support
in
any
way.
Okay,.
A
No,
it's
not
it's.
It's
john
has
been
doing
some
wonderful
work
to
clean
up
the
cops
controller,
and
I
think
I
finally
oh
dm
that
that
pr,
john
sorry
about
the
ridiculous
delays-
hey,
I
see
the
approved
sign
and
so
that
the
the
missing
pieces
to
to
add
node
support
worker
node
support
are
not
that
big
anymore.
A
We
we
can
probably
get
that
in
fairly
easily.
I
will
put
it
on
my
list
to
rebase
and
see
where
we
are.
It's
probably
we're.
Probably
whenever
we
want
to,
we
could
probably
get
it
done
within
a
cycle
if
we
needed
to.
If
we,
if
we
really
wanted
to,
we
could
focus
on
it
and
get
it
done.
It's
not
even
that
much
it's
not
that
much
work
to
be
honest
so,
but,
and
that's
specifically
for
worker
nodes.
Only.
E
A
Not
overly,
it
would
be
a
bit
like
like
the
the
gke
model
or
yeah
okay,
where
you
don't
you
sort
of
configure
a
control
plane
externally,
externally
being
either
the
gke
apis
or
in
our
case
the
cops
cli
and
then
you
can
bring
up
additional
nodes
in
your
cluster.
But
you
you
do
that
in
a
different
place.
D
Okay,
definitely
some
you
know
little
tricks.
We
can
do
like
get
vfs
support
in
that's
separable.
D
G
A
A
A
A
Yes,
but
please
do
send
the
like
to
send
the
big
pr
first
and
we
can
talk
that
through
yeah
and,
of
course
you
can
also
during
the
process.
We
can
use
temporarily
your
like
a
different
ltd
manager
image
right.
So
I'm
sure
you
have
a
fork
of
the
hcd
manager
image
and
we
can
use
that
that
fork
temporarily.
Okay
on
your
branch.
A
Sorry
I
clicked
mute
when
I
meant
to
click
the
next
tab
all
right.
Our
next
item
is
from
cyprian
the
flannel
cni
version
0.13.0.
C
So
I
added
well,
I
made
it
multi-arch
and
well.
It
was
time
to
upgrade,
but
in
the
newer
manifest
they
are,
removing
the
chaos
k8
s
app
label.
C
A
It
looks
like
they're
just
removing
it
without
any
or
they're,
just
changing
it
to
app.
A
I
know
that
sig
apps
has
some
recommended
labels
and
it
is
not
just
app.
It
is
like
a
prefixed
thing
with
like
a
name,
I
got
kate's,
I
o
apps.kate
studios
name
or
something
I'm
wary
of
changing
it.
If
it's
just
to
get
rid
of
the
kate's
thing.
That
seems
a
little
odd.
I
don't
know
if
we
can
figure
out
why
they
did
that.
A
C
A
I
think
it
was
a
de
facto
standard
and
then
the
sig
apps
decided
to
come
with
a
different
standard
and
of
course,
because
there
were
two
standards
that
neither
one
of
them
took
off.
D
C
A
D
You're
asking
that
in
the
your
final
review
comment,
there
is
a
of
what
appears
to
be
a
bug
where
we
weren't
putting
all
the
labels
in
so.
For
that
reason,
I
think
we
should
probably
backport
it
at
119.
A
A
A
D
D
G
D
A
A
Doctor
removed
some
older
versions,
and
we
said
well
we're
not
going
to
like
worry
about
it,
because
this
is
like
kubernetes,
1.0,
1.4
or
something
yeah
so
like
we
effectively
have
this
policy
already,
and
I
think
it
is
the
right
policy,
and
particularly
with
cyprian's
pr
when
it
lands.
A
It's
not
going
to
be
a
it's
not
going
to
be
a
hardcoded
list
anymore,
so
people
will
be
able
to
add
their
own
tar
files
if
they
do,
if
they
do
want
to
run
the
very
first
version
of
docker
and
see
what
happens,
they
can
do
that
in
theory.
I
imagine
it
won't
work,
but
that's
that's
their
choice.
So
yes,
I,
if
you're,
proposing
that
we
should
introduce
deprecation
policy.
I
would
be
very
happy
to
see
such
a
deprecation
policy
that
I
think
tying
to
kubernetes
version
sounds
fine.
A
We
should
not
tie
to
the
upstream
deprecation
annoyingly.
We
should
not
tie
to
the
upstream
deprecation
policy,
because
the
upstream
application
policy
is
very
aggressive
and
doesn't
meet
the
kubernetes
lts
or
even
pre-lts
lifespans.
C
They
changed
it
a
bit
okay,
so
the
policy
with
docker
is
not
every
quarter.
They
chose
to
do
it
that
way,
but
when
they
release
a
new
major
version
or
whatever
minor
version,
then
they
don't
release
any
patch
versions
for
anything
previous.
Unless
your
enterprise
customer.
A
C
A
D
Yeah
yeah,
I
think
I
think
it's
you
know
a
long
line.
So
if
you're
writing
this
kubernetes,
you
can't
run
container
runtimes
older
than
that
that's
sort
of
the
shape.
I
was
thinking.
C
A
D
That's
a
reasonable
policy
yeah
because
you
know
one
is
you
know:
when
do
we
get
rid
of
the
tv
workaround
for
1892.,
for
that
we
have
to
get
rid
of
18.92.
D
There's
a
the
marking
run
c
as
immutable
right.
Yes,.
C
A
Well,
let's,
let's,
let's
let
john
investigate
this
policy
we're
we
are
running
short
on
time,
so
I
want
to
finish
our
we're
at
the
end
of
our
open
discussion
topics
unless
there's
anything
critical,
I
suggest
the
next
and
if
there's
anything
else,
we
we
do
about
the
remainder
of
our
time
to
the
release
plans
and
the
blockers.
A
Yet
I
certainly
have
some
which
are
my
responsibility,
and
I
will
tackle
some
of
those,
the
the
one
which
I
think
is
more
pressing,
so
we
will
likely
not
do
a
119
0
beta
1,
given
that
list
of
blockers,
I
suspect
the
unless
there
was
that,
seems
to
be
what's
what's
going
on
and
then
the
118
two
it
looks
like
we
had
one
blocker,
which
has
been
closed
as
resolved.
So
I
will
do
118
two
this
afternoon.
A
C
A
The
certificate
with
acm
do
we
feel
that's
a
blocker.
A
That's
not
really
that
yeah,
the
like
you
know
like
it
can
be
a
blocker.
Even
if
it's
unfortunate
it's
a
blocker.
A
Unlike
the
always
export,
I
think,
as
long
as
we
treat
it
as
a
blocker
for
release,
I
think
that's
acceptable,
like
well
big
old
release,
notes,
saying:
okay,
this
particular
scenario
is.
F
Okay,
good
question:
around
nine
seven,
nine
four:
I
finally
have
some
time
to
work
on
it.
Do
we
wanna
hold
off
on
that
until
an
lb
support
lands,
or
can
we
land
that
beforehand
they
may
conflict,
which
is
why
I'm
asking.
A
9794
is
prevent
unintended
resource
updates
to
load
balancer
attachments,
yeah.
C
I
so
don't
get
me
wrong.
I
think
it's
important,
but
the
person
that
does
the
nlb
part
should
really
concentrate
on
getting
it
in
and
not
rebasing
and
fixing
the
issues.
I
think
you
are
much
more.
C
A
I
wanted
to
call
it
two
things
that
peter
mentioned
in
chat:
I'm
guessing.
He
can't
speak
on
the
microphone
right
now,
which
is
that
he
suggests
having
the
runtime
deprecations
announced
in
the
119
release,
notes
if
we're
going
to
deprecated
119
and
that
we
should
try
to
have
a
pre-release
with
nlb
and
acli.
We
shouldn't
just
do
that
at
the
last
last
pr
that
goes
in
before
we
release
119.0,
which
I
agree
with.
I
think
those
are
those
are
good
points
and.
D
C
Right
so
I
would
like
to
do
the
beta
to
get
more
coverage
on
119.
That's
the
reason.
Even
with
the
alphas,
we
got
some
very
good
feedback
on
bugs
and
issues.
So
I
think
if
we
don't
release
a
beta
soon,
we
will
lag
quite
a
lot.
We
can
release
one
beta
every
two
weeks,
so
it's
very
easy
now,
but
waiting
like
we
did
in
the
past.
A
True,
we
did,
we
did
fight
hard
to
get
to
parity
and
we
are
risking
slipping
again
behind.
So
from
that
point
of
view,
yes,
can
I
suggest
why
don't
we
and
receiver
over
time?
Why
don't
we
I'll
work
on
some
of
these
blockers
and
we
can
figure
it
out
on
tuesday
make
sure
that
we
are
happy
with
betas.
A
Okay,
let's
see
the
reason
position,
anyone
disagree
with
thumbs
up.
I
like
that
thumbs
up,
like
that's
awesome
all
right.
Well,
then,
we
are
overtime.
I
rush
us
through
the
last
couple
of
minutes
of
argent.
I
apologize,
but
if
there's
nothing
else,
we
can.
Some
of
us
can
regroup
on
tuesday
at
10
a.m,
eastern
for
a
sort
of
cops
oktoberfest,
and
otherwise
I
will
see
everyone
in
two
weeks.