►
From YouTube: Kubernetes WG K8s Infra 20180831
Description
k8s-infra-team bootstrap meeting - Aug 31, 2018
A
Anyway,
Brendon
yes
noted,
we
should
see
if
we
can
do
that.
What
I
heard
around
the
GCR
stuff
was
that
they
can
give
us
a
new
name
that
points
to
an
arbitrary
other
name,
but
that
it
was
more
difficult
to
take
a
name
move
it,
maybe
maybe
as
a
one-off.
They
can
do
it.
The
other
advantage
of
actually
I
mean.
B
A
Right
I
think
that's
what
Ben
was
mostly
saying:
yeah.
We
can
do
something
like
that
with
GCS,
it's
just
racy
yeah.
Well,
it's
not
racy.
It's
it's
a
definite,
visible
outage
like
we
can
do
it.
We
can
schedule
it
for
midnight
or
something,
but
it's
there's
gonna
be
a
few
minutes
or
or
where
it
doesn't
exist.
Right.
C
A
We'll
have
to
shut
the
door
on
that
or
something
so.
This
is
why
it's
not
at
the
top
of
the
list.
The
other
thing,
that's
nice
about
doing
the
sort
of
facade
pattern
that
they
gave
us
with
GCR
is
that
it
auto
region
eliza's.
So
if
we
have
these
scripts
that
distribute
things
to
various
regions,
if
we,
if
we
care
about
it
for
the
for
this
GCS
stuff,
we
might
care
about
it
for
releases.
A
D
A
C
D
Currently
so
we
currently
have
two
individuals,
Rachael
chemistry,
so
it's
team
myself
and
Taylor
from
C&C
of
Taylor's
just
managing
like
the
budget
question,
so
she
is
not
related
to
the
technical
discussions.
We
have
team
in
myself
and
we
can
add,
add
the
entire
team
of
individuals
that
we
have
here
at
disco.
All
the
people
who
are
going
to
be
involved
into
the
process
seems
reasonable
from
there.
B
A
I
think,
just
as
policy,
we
should
endeavor
to
have
a
list
of
people
who
have
effectively
like
project
level
ownership
over
all
projects,
but
that
keep
that
list
small
and
then
everybody
else
be
added
to
projects
with
the
smallest
access
control
that
we
can
figure
out,
acknowledging
that
not
every
product
has
a
fine-grained
ackles.
But
many
do
now.
B
A
A
I
know
we
have
some
capabilities
around
that
I've
never
used
it,
but
I
can
take
a
look
at
that
too.
So
as
a
group
which,
should
we
establish
well,
let's
talk
about
admin,
access
right
now,
there's
a
small
list
of
people
who
have
org
admin,
which
means
you
can
create
new
projects.
That
list
probably
can't
be
very
large.
How
are
we,
how
do
we
want
to
establish?
Who
can
be
in
that
list
and
who's
willing
to
be
in
that
list,
because
it
will
mean
being
able
to
do
things
for
other
people
if
I
create
projects.
D
A
D
B
E
F
Can
throw
my
name
in
there
as
well
if
he
needs
another
non
Googler
like
I
and
I?
Think
as
what
we're
eventually
going
to
need
some
governance
around
this
like
who
gets
involved?
Who
gets
those
permissions?
What
hoops
do
you
need
to
jump
through?
Eventually,
we're
gonna
need
to
establish
that,
whether
it's
a
safe
working
group
or
something
else,
but
you
know
we
can
start-
we
can
start
with
a
small
team
now
and
then
establish
the
governments
later
when
we
know
what
we're
covering.
A
E
That
that's
fine,
but
another
thing
that
I
was
talking
just
before
you
guys
joined
was.
We
are
missing
a
lot
of
people.
We
are
missing.
You
know
people
from
hippy
Oh
who
volunteered
on
Twitter,
for
example,
people
from
Red
Hat
who
traditionally
helped
out
with
a
lot
of
the
infra
work.
So
we
are
missing
people
too.
So
if
we
pick
our
names
right
now,
I'm
a
little
bit
worried
that
we,
you
know,
we
don't
have.
A
D
We
shouldn't
agree
with
that
equality.
Yeah,
we
shouldn't
have
the
unlimited
list
of
people
who
have
access
to
everything.
My
assumption
that
we
currently
have
did
a
good
group
of
people
who
are
great
on
bootstrapping
and
all
these
people
on
the
current
call.
So
we
definitely
may
have
announced
individuals
from
other
teams
and
other
companies
who
will
help
us
with
the
efforts,
because
we
can
skip
awesome
volunteers
and
definitely
their
help
will
be
absolutely
welcome.
But
I'm
not
sure
that
we
should
have
a
really
big
list
of
people
who
can
have
an
access
to
everything.
E
I
agree
with
you
on
that:
it's
just
that.
We
need
to
make
sure
that
we
have
enough
people
who
will
respond
to
the
things
that
need
to
be
in
a
fixed
or
approved
or
set
up
or
configured.
That
needs
to
be
done
by
hand
right,
but
then
the
other
thing
here
is
we
need
some
Googlers
who
know
work
that
we
can't
be
just
people
who
are
from
the
outside
as
well
so
I
see
that
be
a
missing
Robert
Bailey
from
from
the
call
yeah.
E
Good,
okay,
yeah,
so
I'm
happy
to
you,
know,
I
rolling,
but
who
we
have
here
on
the
call
and
then
figuring
out.
You
know
lit
steering
committees,
figure
out
some
of
the
stuff
and
then
go
on.
Let's
pick
a
name,
and
then
is
the
name:
okay
that
suggested
there
there's
one
from
Aaron
one
from
Chris:
okay,
Karl.
A
Coincidentally,
there
are
ten
people
listed,
at
least
in
the
attendees.
If
there's
anybody
on
the
call
who's,
not
in
the
attendee
list,
please
make
sure
you're
added
at
the
bottom
of
the
linked
document,
but
there's
ten
of
us.
We
consider
that
maybe
the
inaugural
leadership
group
of
this
effort
and
as
more
people
get
added,
we
will
figure
out
sort
of
which
efforts
they're,
actually
interested
in
helping
administer
non-exclusive
yeah.
C
So
I
just
wanna,
throw
out
there
on
that.
We've
been
planning
on
this
for
a
while,
hoping
that
the
credits
would
be
finished
and
out
Amin
and
I.
At
least
this
quarter
for
sure
have
like
a
have
time
allocated
specifically
to
moving
all
of
this
from
the
Google
side
and
I
think
we
will
continuing
as
forward
as
long
as
we
need
to
do
this
as
well.
Awesome.
A
A
A
D
A
C
B
H
A
A
That's
a
question
that
steering
has
been
wrestling
with:
we
didn't
want
to
make
the
kubernetes
at
I/o.
G
sweet,
be
a
really
dynamic
thing.
We
didn't
want
it
to
become
a
status
thing
or,
like.
Oh
I,
have
access
to
the
G
sweet
groups,
but
it
actually
sort
of
makes
sense
and
there's
a
lot
of
things.
You
can
do
there
that
you
can't
do
on
public
Google
Groups
Brendan.
What
do
you
think
since
you're
on
steering
and
I'm,
not
anywhere
Aaron.
B
Well,
I
kind
of
feel
like
given
that
we're
talking
about
potentially
even
hiring
employees
like
it
kind
of
starts
feeling
like.
Maybe
it
should
be
part
of
the
juice
sweet
I
mean
this
feels
like
plumbing
I,
guess,
I
feel
like
there's
no
harm
and
and
I
don't
think
anybody's
gonna
have
any
status
associated
with
it,
but
I
Erin
might
have
no.
H
A
Sweet
is
funded
separately,
but
groups
don't
cost
money
accounts
do
but
groups
don't
so
we
create
as
many
groups
as
we
want
and
add
people
to
those
groups,
and
it
shouldn't
cost
anything
and
I
think
that
I
mean
it
seems
like
a
reasonable
way
to
administer
it.
The
the
Gotcha
is,
there's
a
very
small
number
of
people
who
have
access
to
that
gee,
sweet
and
now
are
putting
another
group
of
people
in
the
way
of
being
able
to
do
progress.
Yeah.
B
A
H
Argue
probably
stick
to
public
Google
Groups
for
now
yeah
in
the
interest
of
deep
bottlenecking,
and
if
it
turns
out
as
we
look
to
more
formalize
this
or
realize
we
do
need
some
tighter
access
controls
that
are
not
available
as
a
result
of
the
public
groups,
then
we
can
consider
doing
it
for
the
communities
that
I
have
google
groups.
But
I
agree
strongly
with
you
about
the
whole
status
thing.
H
B
A
A
E
A
Yeah
for
the
last
item
there,
let's
just
grab
anything
that
starts
with
Kate's
infra.
That
seems
reasonably
useful
and
we'll
go
from
there
and
I
guess:
I
agree
with
her
and
if
we
start
with
public
we
can
always
move
at
private
later.
Although
the
transition
is
hard
because
you
can't
add
a
group
to
a
group
through
the
public
groups,
but
you
can
through
private
groups
whatever
we'll
deal
with
it.
When
you
get
there
in
terms
of
official
nests,
we
will
need
to
have
the
conversation
with
steering.
H
There
has
not
been
we've
had
some
chatter
on
the
private
of
the
private
mailing
list,
since
we
could
discuss
this
while
elizy
embargoed
publicly,
we
had
a
little
bit
of
a
chat
before
you
showed
up
and
I.
Think
next
steering
committee
meeting
I'll
make
sure
its
front
and
center
as
a
thing
to
digest.
Address
and
I
have
a
next
item
to
make
sure
we're
clarifying
the
whole
working
group
thing.
A
Think
we
should
let
Aaron
be
the
the
bridge
for
now
and
get
a
recommendation
from
steering
committee
as
to
what
structure
they
would
like
to
use
for
this.
Since
this
is
a
little
bit
different
than
anything
in
the
past,
and
once
they
have
that
structure,
then
we
can
draft
the
Charter
or
the
mission
or
whatever
is
appropriate
for
the
structure
that
use
I'm.
H
H
A
B
I
think
so,
though
it
doesn't
seem,
I
mean.
B
H
A
Close
yeah,
okay,
so
looking
back
at
the
dock
name
and
effort
list
of
priorities
will
need
to
come
up
with
the
naming
and
structure
convention
for
GCP
projects.
I'll.
Take
the
action
to
go
in
a
little
bit
and
just
to
see
sort
of
with
the
first
few
items
on
our
list.
What
the
right
level
of
access
control
would
be
with
respect
to
Google
Groups
I
haven't
actually
played
with
groups,
and
I
am
very
much
so
I
can
play
with
that.
A
little
bit
interesting
challenge,
I
just
found.
A
A
A
Can't
you
like
create
it
if
you're,
if
you're,
writing
that
program
Brendan.
If
you
have
the
other
direction,
then
I
it's
done.
Okay,
so
I
can
point
it
out
existing
DNS
have
it
dumped
the
ammo
and
then
we
can
sort
of
look
at
it
scrub
it
decide
if
it's
what
we
want
and
and
then
like
use
this
as
a
proof
of
concept,
I
like
we
can
get
either
you
or
the
group
added
to
the
the
first
project
and
we
can
import
it
and
see
if
it
all
came
out
the
way
we
want
it
to
okay.
B
C
A
C
J
All
right,
one
other
ten
dental
question
I
had
just
wanna
make
sure
we
do
have
to
fact
required
on
these
projects
or
this
or
it
would
be
a
really
good
idea.
I
think
it's
something
you
can
require
with
GCP
I
was
looking
up
so
I
think
as
long
as
we
like
and
enable
that
on
whatever
or
this
against
EF
or
whatever
I.
A
Think
it's
already
enabled
on
these
organ
I
remember
the
hassle.
Brian
went
through
to
turn
on
a
year
and
a
half
ago.
Okay,
so
we're
gonna
use
github,
kubernetes
Kate's
at
I/o
or
yeah.
That's
that's!
The
right
wing,
github,
kubernetes
Kate,
said
I/o
will
make
some
directories
under
there.
So
Brendan.
If
you
want
a
home
for
your
yeah
mole
importer,
that
could
be
a
home
for
it
unless
you
want
to
stick
it
in
a
separate
repo
and
just
use
it
from
there
like.
C
C
A
B
B
A
B
A
A
A
Be
interested
absolutely
and
anybody
who
comes
up
with
more
ideas
of
things
that
are
need
to
be
cover
here
or
wants
to
flesh
out
more
details
of
how
to
make
these
transitions
like
Ren
and
Jeff.
You
probably
have
thoughts
around
how
to
make
the
flop
of
the
testing
infrastructure
this
either
link
to
a
dock
or
start
throwing
notes
in
here
somewhere
yeah.
A
And
as
I'm
thinking
about
the
ordering
that
we
talked
about
here,
we
can
do
DNS
manually,
but
eventually
we'll
want
to
have
this
tool
and
the
tool
probably
wants
to
run
to
keep
cluster
so
we'll
pretty
quickly
end
up
in
the
second
category,
which
is
fine,
actually
I.
Think
that's
the
right
way
to
go,
because
that
that
unblocks,
you
know
15
different
efforts
that
want
to
run
in
the
Q
cluster
right.
K
A
A
Priority
right,
quick,
just
to
put
it
very
concretely
like
I,
knew
the
GCR
stuff
would
be
a
big
deal
to
get
there
in
director
up
and
to
get.
You
know,
hundreds
of
pull
requests
against
all
the
different
repos
changed.
So
I
started
that
months
ago,
and
it's
still
not
done
it's
still,
there's
still
some
that
haven't,
merge
those
PRS
or
have
regressed
in
some
way
or
another,
so
we'll
have
to
do
another
pass.
I
E
B
H
D
A
D
H
Sure
some
of
you
have
gotten
this
on
Twitter
or
whatever
the
other
awesome
social
media
live
streaming.
Things
are,
but
I
just
wanted
to
really.
Thank
you
so
much
for
all
of
the
pushing
and
effort
that
has
gone
forward
to
get
us
this
far
I
know
there
are
some
outside
who
think.
Oh
great,
the
switch
has
been
flipped
and
it's
all
done-
and
this
is
the
group
of
people
who
recognize
that
yeah
we're
talking
on
the
order
of
months
worth
of
work.