►
From YouTube: Kubernetes SIG Release for 20221129
Description
Kubernetes SIG Release for 20221129
A
All
right,
everyone
Welcome
to
our
weekly
sick
release,
meeting
this
meeting
at
Houston
GF
code
of
conduct,
which
basically
boils
down
to
be
mindful
what
you
said
and
be
excellent
to
each
other.
I
already
mentioned
it.
So
please
add
yourself
to
the
list
of
attendees
and
for
today
we
have
Ray
as
Note
Taker.
So
thanks
for
that
and
now
we
can
jump
directly
in
our
recurring
topics.
Do
we
have
anyone
on
the
call
who
would
like
to
say,
hi
and
who
is
new
to
sick
release,
foreign.
A
B
Okay,
I
can
give
it
a
try,
so
a
lot
has
been
happening
in
the
past
week.
First
of
all,
really
stuff,
I
think
that
we
had
rc0
last
week
and
in
general
it
went
pretty
well
to
my
knowledge.
We
had
some
minor
issues:
history,
jobs
generation
which
are
now
fixed
with
that
fix,
we
interested
the
new
version.
Marker
called
stable.
Four,
it
is
now
official.
It
has
been
documented.
It
has
been
properly
integrated
in
the
release
tooling.
So
this
is
one
thing
to
mention:
there
is
rc1
scheduled
for
today.
B
Maybe
Jeremy
can
tell
a
little
bit
more
about
it
later
for
release
team
update,
but
this
is
it
mostly
from
126
for
patch
releases.
We
have
petrolysis
upcoming
next
week
called
Destiny.
I
would
like
to
point
it
out
that
the
cherry
pick
deadline.
Is
this
Friday
December
2nd?
They
will
send
an
announcement
for
this,
but
keep
in
mind
if
we
have
anything
to
share
epic
to
the
store
before
the
deadline.
B
One
thing
that
I
would
like
to
point
out
and
shout
out
to
Carlos
and
Sasha,
and
everyone
involved,
is
that
now
we
are
having
a
signing
for
binary
artifacts.
This
has
been
merged
today.
I
think
it
will
be
already
kicked
off
today
for
RC
Von,
so
we
will
have
a
chance
to
see
how
it
works.
It
is
a
new
subcommand
in
Corel
and
it
is
integrated,
as
I
said,
in
the
release
step.
B
B
It
has
been
going
pretty
good
and
we
managed
to
successfully
complete
it,
like
all
packages
are
getting
successfully
create
and
installed
published
and
all
this
stuff,
and
additionally,
this
weekend,
I
got
queue
package,
somehow
integrated
with
all
that,
so
that
is
properly
generating
these
spec
files
that
is
expected
by
OBS,
that
is
generating
the
turbo
with
all
needed,
the
files
and
binaries
and
other
stuff
that's
consumed
by
packages.
Build
process
to
yeah
then
build
the
packages,
and
this
is
something
that
I'm
going
to
discuss
a
little
bit
later
today.
B
Also
I
think
for
registered
Cherry
picks
that
we
are
doing
pretty
well,
so
most
of
them
should
be
merged
by
now.
As
far
as
I
know,
if
I'm
not
mistaken
and
I
think
that
would
be
most
of
it.
I
might
forgot
the
stuffing.
There
has
been
a
lot
of
moments,
but
yeah,
please
feel
free
to
add.
If
there
is
any
figures.
A
C
First,
one
is
I,
started
work
on
the
provenance
generation,
so
if
anybody
wants
to
join
me
on
that
ping,
me
and
I
also
started
with
Carlos
working
on
a
new
repository
of
actions,
so
that
everybody
can
use
the
tools
and
the
pr
should
be
coming
this
week.
If
you
want
to,
if
you
want
to
reuse
our
Tools
in
your
project,
be
it
kubernetes
or
not,
you
can
you'll
be
able
to
use
it
much
more
easily.
C
Once
we
have
those
but
expect
some
news
over
the
week,
oh
and
the
other
one,
an
important
one,
was
we
released
a
blog
post
about
the
backboards
together
with
Kate,
stick
infra
about
the
registry
change,
so
I
will
add
it
to
the
notes
to
the
to
the
meeting
notes
so
that
anybody
can
see
it
in
case
you
didn't
capture
the
announcement
and
awesome
work
on
the
markers
marker
and
without
also
very
happy
to
see
that
the
binary
sending
happened
during
the
night.
It's
exciting
years
to
wake
up
to.
A
D
E
We
were
going
to
discuss
that
now.
Xander
is
going
to
run
with
that
today,
so
we
can.
We
can
schedule
around
what
what
we
need
to
do.
I
think
the
promotion
times
are
so
much
better
now
that
it
we
can
start
it
later
today,
and
it
won't
be
that
big
of
a
deal
rc1,
we
won't.
We
don't
have
all
of
the
extra
work
of
the
branch,
cuts
and
job
creations
and
stuff,
so
I
think
this
will
be
fairly
straightforward
and
we
can.
G
G
We
are
chasing
the
last
couple
of
of
open
PRS
to
match
the
docs
in
and
for
the
blog
post
for
comms.
There
are
a
few
authors
which
who
are
now
out
of
office,
so
we
are
basically
moving
basically
to
others
to
check
if
they
can
review
what
has
already
been
done
and
maybe
then
the
ones
who
are
maybe
in
office
can
take
over.
So
we
are
basically
just
facilitating
a
little
bit,
but
nothing
concerning
nothing
major.
G
G
E
E
If
we
don't
open
it
until
January
I
think
that
limits
the
it
compresses
the
time
we
either
would
want
to
start
a
little
bit
later,
to
give
folks
time
to
pick
shadows
and
onboard
them
before
the
release
really
gets
going,
especially
for
enhancements,
or
we
would
just
need
to
shorten
the
time
for
the
survey
down
quite
a
bit.
Yeah.
G
D
Yeah
we
we
would
roll
it
out
before
Christmas
before
everyone
goes
out
on
vacation
and
ideally
one
blocker,
for
that
is
finding
an
EA
which
I
am
doing
right
now,
but
hopefully
like.
We
would
have
our
one
to
two
seven
EA
decided
by
end
of
this
week
and
then
I
can
start
handing
over
the
1.27
tasks
to
the
incoming
El
and
I.
Also
plan
to
like
roll
out
the
shadow
surveys
for
the
end
of
cycle
service.
D
This
week
plus
assembling
the
release
team
issue
as
well.
I
believe
all
the
leads
are
right
now
in
the
process
of
selecting
the
next
leads
and
I
hope
that
they
would
have
an
answer
by
made
of
next
week
for
most
of
them,
but
in
my
mind,
finding
an
EAS
of
priority
right
now
to
get
things
for
1.27
Rolling
as
soon
as
possible.
A
C
H
H
We
just
basically
say
all
the
new
time
will
still
can
still
be
pulled
from
that
country
registry,
but
in
the
future,
if
you
want
to
use
basically
anything
after
January
2023,
you
have
to
use
the
new
version
stream
and
even
the
old,
basically
Old
Tag,
like
pause
1.1.
You
still
can
pull
from
from
the
new
registry
because
we
did
a
full
copy
of
that.
So
because
we
already
change
the
change
log
in
125,
we
can
basically
say
do
a
cherry
pick
and
say
in
the
change
of
oh
all,
basically,
all
the
older
version
of
KK.
H
C
E
Yeah
I
was
just
gonna
point
out
that
there
was
a
threat
already
in
case
again
for
where
somebody
was
asking
very
specifically,
when
can
I
when
will
I
not
be
able
to
pull
from
kate's.ccr.io
and
I.
Think
that
that's
Clarity,
that
people
are
gonna
either
look
for
or
they're
going
to
continue
to
use
it
until
they
can't
use
it
like
we've.
F
E
E
The
longer
that
we
keep
this
going
where
like
126,
is
going
to
land
in
two
weeks
and
it's
going
to
land
in
case
you
see
our
IO,
so
people
will
still
be
able
to
pull
from
there,
which
is
just
going
to
prolong
the
activity
that
people
are
going
to
keep
using
that
registry.
So
I'm
in
favor
of
trying
to
accelerate
this
as
much
as
possible.
H
Yeah,
if
we
can
stop
I,
think
the
we
we
have
like
one
I
would
say
three
weeks
left
before
Christmas,
so
I
would
like
to
basically
go
a
little
fast
on
this
and
issue
some
communication
either
first
week
of
January
of
end
of
December,
like
depending
of
timeline
like
I,
think
the
first
thing
like
make
sure
generally
next
year,
we
we
are
sure
we
can
stop
pushing
to
the
old
registry.
So
I
think
that's
the
basically
the
one
thing
I
want
to
investigate
during
December
like
trying
to
figure
out.
C
But
one
thing
I
think
we
can
be
smart
about
is
communicating
early
and
a
lot
more,
so
I
I
think
it
doesn't
hurt
if
we
start
putting
out
the
word
and
we
can
iron
out
the
implementation
details,
I
mean
we've
had
like
sensitive
discussions
already
about
it
of
how
to
implement
it.
But
we
can
draft
a
plan
in
the
next
Gates
and
infra
meeting
and
then
we
can
bring
back
the
details
of
what's
to
be
done
on
the
security
side
and
but
I'm
sure
it's
going
to
be
a
lot
of
communication.
C
H
A
I
H
So
I
I
I,
oh
I'm,
gonna,
say
that
I
think
I
think
the
the
main
idea
is
to
basically
the
the
real
impact
is
mostly
the
patch
releases.
I
think
this
is
where
we're
gonna
have
that's
a
real
impact
because
for
I
would
say
for
any
new
release.
If
we
basically
decide
today,
we
stop
publishing
that
it's
fine,
but
the
real
impact
is
patch.
H
Release
is
because
we
are
basically
I
think
people
will
have
the
Assumption
to
still
pull
from
the
duplicate
one,
so
they
are
like
basically
shoot
things
we
can
do,
but
my
real,
my
real
concerns
is,
is
basically
how
we're
gonna
handle
patch
releases
in
the
future.
I
think
that's
like
my
real
concern
right,
but
we
can.
We
can
have
this
conversation
as
you
think
about
this
yeah.
C
Okay,
because
some
of
the
discussions
were
that
batch
releases
for
the
stallion
support
branches
like
23
and
24.,
we're
still
going
to
be
pushed
there
and
then
once
they
those
end
of
life,
then
we
stop
all
all
releases
but
yeah.
We
can
definitely
discuss
the
implementation,
details
and
okay
different.
H
Right
this,
my
question
is
about.
Basically,
this
is
about
getting
approval
from
Sig
release.
To
not
push
to
do
all
which
is
right,
so
I
think
we're
still
out.
I
need
to
establish
a
plan,
and
from
that
plan
we
make
that
decision,
because
I
think
yeah.
We
just
kind
of
make
the
decision
right
now
stop
publishing
stuff.
We
need
to
look
at
what's
the
meaning
of
this,
like
configuration
of
proud
how
we
handle.
I
H
Okay,
so
I
think
right
now
out
of
time
is
not
accurate
here,
because
it's
up
to
us
to
define
the
timeline.
We
can
also
decide
to
make
that
in
12028
like
this,
like
nothing
blocking
us
to
do
that.
So
that's
why
I'm,
basically
I
prefer
to
first
look
at
the
technical
detail
before
we
make
the
communication.
H
C
One
one
thing
that
could
be
helpful
is:
if
anybody
on
the
security
side
has
any
objections
about
not
trying
to
Sunset
the
registry
for
126
I
mean
we
still
need
to
discuss
it
a
little
bit
further
and
deeper.
But
if,
on
the
next
Gate
Cinema
meeting,
we
should
try
to
start
working
on
the
implementation.
C
So
if
anybody
has
any
concerns
or
that
well,
please
let
yourselves
be
heard
now
and
I
can
bring
it
to
the
casing
for
a
meeting
or
just
raised
them
up
in
the
chat
as
soon
as
you
can,
and
so
that
we
can
I
mean
the
project
needs
everybody's
help.
So
if
we
can
go
in
a
Cooperative
attitude
and
bring
propositive
ideas
and
some
settings
earlier
the
better,
but
if
anybody
has
any
concerns,
I
probably
announce
that
time.
C
C
About
this,
like
it's,
not
it's
not
a
like
you
have
to
commit
now
you
can
think
about
it
and
and
but
I
think.
Overall,
there
is
Goodwill
from
RC
with
gets
infra
to
work
this
and
make
this
happen
as
soon
as
possible,
but
there's
always
a
chance
for
people
to
think
about
it
and
raise
any
objections.
Comments
concerns
whatever.
H
H
H
B
Okay,
so
I
guess
we
can
take
off
the
cube
Pikachu
topic,
okay,
so
here's
the
thing
I
have
been
working
on
experimenting,
so
first,
what
I'm
trying
to
accomplish,
and
we
need
a
way
to
build
these
spec
files
to
generate
them
and
to
publish
them
on
the
open,
build
service
so
that
those
can
be
picked
up
and
that
we
can
based
on
that,
generate
depths
and
RPMs.
And
the
thing
is
that
we
need
a
tooling
for
that,
because
some
something
has
to
generate
those
spec
files
and
also
besides
spec
files.
B
Before
when
we
had
a
Rapture
and
Google
build
folks
taking
care
of
packages,
it
worked
in
a
way
that
we
were
pulling.
For
example,
we
are
real
the
cube
CTR
package.
We
pull
Cube
CTL
binary
from
the
Google
Cloud
bucket
and
include
that
in
the
package
field,
when
doing
that
for
OBS,
that's
not
going
to
work
because
OBS
builds
are
air,
app,
there's
no
internet
access
and
you
have
to
provide
your
binaries
before
the
build
starts
and
what
I
manager
to
do
it.
How
I
accomplish
it
to
get
both
networking
and
multi-arged
builds?
B
Is
that
I
create
a
turbo
that
contains,
for
example,
Cube
CTL
binaries
for
all
architectures
put
that
with,
for
example,
maximum
compassion,
so
that
we
save
on
storage
and
stuff
like
that
and
push
that
to
the
OBS
service?
What
I
expect,
for
example,
this
case
Q
packaging
to
do
is
to
generate
spec
files
and
to
generate
that
torque
turbo
that
bot
can
be
pushed
to
OBS
service.
My
question
is
how
cubic
cubicg
is
standing
right
now?
It
is
a
bit
of
experimental
tool
and
not
something
that
really
took
off.
B
For
example,
we
never
implemented
it
in
Creve
or
something
like
that,
at
least
not
to
my
knowledge,
Google
build
adminster
using
different
spec
files
than
those
that
are
generated
by
crowd
that
qpg
so
I'm
wondering
how
much
do
we
want
to
keep
of
the
breakfast
compatibility
like
do
we
want
to
start
over
to
completely
to
completely
get
rid
of
spec
files
that
we
have
right
now
to
create
new
spec
files
that
are
working
with
OBS
and
to
make
behavior
that
it
just
works
with?
What's
in
it
with
OS,
and
that
will
stop
maintaining?
B
H
B
Well,
for
each
release,
when
we,
for
example,
want
to
release
124.5,
we
need
like
to
take
this
back
file
and
to
put
stuff
like
versions,
dependencies
and
other
stuff,
so
for
each
release
we
need
to
generate
a
dedicated
spec
file
because
it
might
be
different
from
release
to
release.
Same
goes
for
between
minor
release
and
even
twin
patch
releases.
It
can
be
different,
so
Q
package
is
only
here
to
mitigate
that
and
to
make
it
easy
to
generate
stuff.
D
H
Because
I
think,
like
my
question,
is
really
to
a
comments
different
my
made
by
Stephen.
Maybe
a
few
minutes
before,
like
about
the
fact
that
those
spec
are
consumed
by
Downstream.
So
if
we
want
to
leverage
Cube
PKG,
we
is
is
going
to
be
delicate
to
change
the
template
of
those
spec
because
there's
like
underlying
impact
for
people
using
that.
B
Yeah,
so
that's
one
of
the
Perpetual
series
for
like
do
we
really
need
to
maintain
stuff?
We
have
right
now
for
downstairs
customers
like
do
we
have
any
idea
if
anyone
is
actually
using
that
story
at
all,
or
can
we
say
I
think
there
is
some
know
that
it
is
experimental.
So
read
me
or
something
like
that.
B
Can
we
just
say:
okay,
we
will
just
make
it
work
with
OBS,
because
this
is
what
we
use
and
we
will
not
only
maintain
the
other
way
that
you
had
before
so
I
mean
yeah
for
what
was
in
Q
package.
We
will
still
maintain
what
is
used
by
Google.
Will
that
means,
but
that
is
a
different
thing
that
is
being
generated
by
qpj
right
now,.
B
A
Yeah
so
Marco
generally
I
think
in
the
future
we
would
like
to
have
spec
files
which
follow
the
best
practices
for
those
operating
systems.
So
I
would
say
that
we
probably
will
not
keep
them
as
they
are,
but
for
an
intermediate
solution.
We
would
just
like
to
to
keep
them
right.
So
what
I
would
expect
from
Krell
afterwards
that
it
will
still
modify
the
specifiers
and
OBS
will
take
care
of
the
hosting
and
building
from
them?.
B
A
B
Then
I
will.
This
is
what
we
have
right
now.
I
will
include
this
in
cap
I
plan
to
update
the
campus
again
right
now,
however,
I
would
first
like
to
meet
with
OBS
folks
to
discuss
if
the
solution
that
we
have
right
now
is
collateral.
I
have
sent
Doodle
for
that,
but
it
seems
like
the
best
time
slot
is
tomorrow
or
Monday
next
week.
B
B
B
If
someone
wants
to
take
a
look,
maybe
I
can
create
a
br
with
the
new
spec
files
that
we
are
using,
so
that
you,
for
example,
can
take
a
look
and
see
what
will
be
the
difference
and
then
basically
decide
what
do
we
want
to
do
with
all
files
like
do
we
want
to
keep
them
drop
them
or
something
else
like
yeah?
Do
we
want
to
maintain
real
behavior
of
this
tool.
A
Yeah
sounds
good
and
for
the
for
the
build
process
itself
that
are
some
details.
We
can
probably
then
catch
up
in
the
in
our
meeting
with
the
OBS
folks,
because
there
are
those
service
definitions
which
allow
internet
access
to
download
source
files
and
things
like
that.
But
we
have
to
create
a
workflow
for
us
which
yeah
actually
works
so.
B
One
thing
that
might
be
nice
to
answer:
it
comes
to
qpkg
right
now
we
have
some
Logic
for
building
packages
like
to
run
the
appropriate
dpkg
and
RPM
tools.
Any
idea,
if
you
still
need
that
or
do
we
want
to
actually
to
drop
that
completely
and
completely
depend
on
OBS,
that's
not
used
by
like
Google
build.
That
means
so
yeah.
This
is
something
that
I
am
thinking.
If
you
have
like
any
objections
to
that
chopping
goal.
B
If
you
have
any
objection
to
change
the
Volcan
packaging
tool,
please
reach
out
to
me
so
that
I
can
plan
for
it
and
that
I
can
put
that
in
caps
when
creating
it.
So
yeah
I
think
that's
the
best
way
out
so
yeah
I
will
create
that
PR.
We
can
discuss
it
there
and
it.
Meanwhile,
if
you
have
objections,
please
let
me
know,
and
then
after
we
have
the
meeting
with
OBS
folks,
I
will
start
working
on
cap
and
follow
up
with
that.
A
I
Yep
yeah,
okay,
so
that's
a
very
familiar
thing
for
everybody
that
uses
Max.
So
this
is
a
problem
that
I've
solved
for
another
project
that
I
work
on
right.
So
I
would
like
to
do
the
same
for
kubernetes,
but
I'm
not
quite
familiar
with
the
build
process
so
somewhere
very
early
in
the
build
process.
We
need
to
run
some
code
that
basically
sticks
a
certificate
to
the
file
and
submits
it
to
Apple
monitorization,
and
that's
all
I
need
to
do
so.
D
I
So
somewhere,
so
I
need
to
find
out
how
the
build
process
works
and
somewhere
very
early
in
the
build
process.
After
the
binary
is
built.
We
didn't
go,
build
I
need
to
insert
the
shell
script
that
basically
uploads
the
file
to
another
to
an
Apple
service
and
it
notarizes
it
so
Gatekeepers,
of
course,
problems
when
people
download
the
binaries.
C
I
B
E
I
I
That's
really
so
there
is
this
program
that
gets
called.
You
basically
have
to
zip
the
Darwin
binaries,
and
then
it
passes
this
program.
It
gets
uploaded.
You
need
to
supply
like
a
some
API
Keys,
a
certificates
and
apple
issues
to
you
and
then
once
you
do
that
you
wait
like
about
less
than
a
minute.
Usually,
and
then
you
know
you
get
response
back
from
those
and
yeah
we're
happy
with
this
binary,
and
then
it
gets
Enterprise,
don't
process.
B
I
H
Sorry,
just
one
comment
about
this,
so
I
talked
to
this
about
about
getting
developed,
a
developer
program
account
with
steering,
and
there
were
like
some
legal
issue
related
to
that
I
think
the
one
thing
is
like
we
should
not
I
mean.
Basically
what
steering
says
like
we
should
not
basically
have
an
individual
owning
the
icon
to
basically
access
to
API
keys.
So
the
idea
is
to
try
basically
try
to
talk
to
cncf
about
this,
so
we
basically
cncf
get
the
account
pay
for
it
and
give
us
the
access
the
access
key.
H
I
I
Have
to
work
on
for
Canada
yeah,
they
were
like
yeah,
you
should
keep
now
and
then
we'll
change
it
when
they
get
to
it,
which
I
mean
works
for
me,
because
I
get
a
feature
out
the
door.
But
if
you
look
at
the
the
chart,
the
script
is
there.
So
if
you
put
on
the
screen
real
quick,
not
there,
I
didn't
put
an
issue
so.