►
From YouTube: Kubernetes SIG Azure - 2019-06-12
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,
hello,
everyone.
This
is
June
12th,
it's
the
signature,
meeting
bi-weekly
meeting.
This
is
a
meeting
that
is
recorded
and
on
the
internet.
So
please
be
mindful
of
what
you
say.
Please
adhere
to
the
kubernetes
code
of
conduct,
which
essentially
amounts
to
don't
be
a
jerk
all
right.
So
we
don't
have
so
I
want
to
give
an
opportunity
to
people
who
haven't
been
on
the
call
to
introduce
themselves
first
and
foremost,
and
then
we
can
walk
through
our
one
item
agenda.
B
A
A
So
last
week,
I
was
it
last
week.
I
think
it
was.
Last
week
last
week,
I
had
a
visit
at
Microsoft,
with
Craig
and
and
Rita
to
talk
about
some
of
the
things
that
we
were
trying
to
achieve
for
for
cigar
in
the
second
half
of
the
year.
So
one
if
you
have
not
seen
the
agenda
for
this
meeting
I'm
going
to
link
it
and
you
can
take
a
moment
and
just
drop
a
drop,
your
name
on
there
and
and
then
secondarily,
we've
got
the
planning
doc.
A
There
and
that's
some
of
the
meeting
notes
that
we
took
it
still
has
a
little
bit
to
go
to
get
cleaned
up.
I
want
to
walk
through
the
cleanup
of
the
cig
cloud
provider
consolidation,
so
one
of
the
biggest
things
that
we
will
be
working
on
in
the
next
few
cycles
is
to
consolidate
a
cigar
under
cig
cloud
provider
right
so
as
the
sub-project
interested
cloud
provider.
So,
what's
going
to
be
involved,
there
is
a
lot
of
one
figuring
out
like
the
sessions
that
we're
going
to
do
for
the
next
coupon.
A
So
we're
not
we're
not
announcing
that
this
is
done
until
Cuba
on
San
Diego
in
November,
but
there
are
lots
of
little
administrivia
things
to
do
like
figuring
out,
meeting
Cadence's
and
doing
zoom
links
and
changing
slack
channels.
A
lot
of
that.
A
lot
of
that
process
can
start
now
so
I'm
thinking
what
I'm
gonna
do
is
put
together
a
Gerber
Nettie's
project
board
and
start
just
chunking
chunking.
A
So
in
so
something
that's
a
little
bit
better
fleshed
out
is
a
you
know.
One
of
the
things
that
we're
working
on
is
moving
the
entry
cloud
provider
for
Azure
out
of
tree
right
so
from
kubernetes
kubernetes,
slash,
staging
legacy
cloud
providers
ladder
into
kubernetes
right
now.
It's
kubernetes
cloud
provider
Azure,
but
it'll
become
kubernetes
sakes
cloud
provider,
Azure
all
right.
So
the
work
that's
involved
and
forklifting
that
code
right
so,
first
and
foremost
we're
going
to
update
the
caps
that
we
have.
A
So
we
have
a
few
open
right
now,
I
think
across
region
nodes,
also
availability
zones
and
the
the
most
important
one,
which
is
the
actual
out
of
tree
Cloud
Controller
manager,
right
updating
that
for
116.
So
after
you
know,
after
I
clean
up
this
planning
doc
and
start
chunking
out
some
of
the
tasks
on
sub
project
board
I'm
going
to
start
on
updating
the
caps
for
116.
So
anyone
who
is
not
familiar
with
the
cap
is
cap
is
a
kubernetes
enhancement
proposal
and
basically
those
are
the
the
documents
that
we
provide
to
describe
a
change.
A
That's
happening
to
the
ecosystem
right,
so
that
you
know
that
dock
includes,
like
the
motivation
for
the
change
like
what
we're
trying
to
achieve,
what
we're
not
trying
to
achieve
how
we're
implemented
some
of
the
design
details
of
that
change,
as
well
as
how
we're
going
to
test
it
and
validate
that
it's
something
that
can
be
graduated
through
alpha
beta
and
GA
stages
right.
So,
if
you're
curious
about
the
current
state
of
that
kept,
let
me
pull
that
up.
A
A
Rita
has
been
working
diligently
on
actually
getting
us
like
reasonable
tests
for
for
the
cloud
provider,
both
on
the
entry
side
and
the
out
of
tree
side.
Actually
that
question
there
Rita
it's
something
that
I
intend
to
do
or
I
think
that
we
should
do
being
able
to
test.
Do
the
intend
tests
in
both
aks
engine
as
well
as
cluster
API
badger?
C
A
D
A
A
Yeah
yeah
I
think
you
know
forcing
us
to
kind
of
adhere
to
whatever
our
definition
of
conformance
is
for
an
azure
cluster
and
kubernetes
would
be
our
kubernetes
cluster
around
a
sure
would
be
super
useful,
so
I
think
you
know
first
step
these.
You
know
these
end-to-end
tests
kind
of
like
see
if
we
can
create
a
cluster
see
if
we
can
create
a
machine
still
some
work
to
be
done
there.
Also.
A
If
anyone
is
interested
in
contributing
to
cluster
API
Azure
happy
to
have
more
contributors,
I
think,
that's
you
know
our
we're
doing
our
recruitment
run
before
this
kind
of
stuff.
Right
now,
as
well
as
you
know,
I
think
D.
So
you
know
for
me
kind
of
the
requirement
when
someone
submits
a
PR
is
like
we
have
some.
We
have
some
general
tests
that
do
linting
does
some
unit
testing,
but
we
don't
have
an
end-to-end
test
and
as
a
result
of
that,
if
I
see
a
big
change,
lend
or
a
change,
that's
a
little
sketchy.
A
A
Get
back
to
the
agenda
right
so
yeah,
so
in
in
doing
some
of
the
CCM
work,
we
need
to
establish
timelines
right,
so
we're
basically
going
to
be
working
through
the
1/16
to
about
the
1:18
time
cycle
kind
of
like
end
of
year,
end
of
your
in
a
little
beyond
to
get
some
of
the
stuff
done.
So
what
we
really
need
to
do
is
figure
out
what
happened
needs
to
happen
in
each
cycle.
A
A
So
the
from
the
outside
looking
in
it,
looks
like
the
code
is
ready
to
forklift
all
right,
so
I've
moved
it
from
staging
cloud
provider
or
whatever
it's
called
into
legacy
cloud
providers.
So
it's
it's
being
staged
there,
but
there
are
some
kind
of
like
intangibles
that
I
don't
fully
understand
at
this
point,
I.
C
Think,
as
you
have
outlined
in
the
planning,
doc
right,
I
think
there's
some
missing
features
like
the
CSI
stuff
right
and
then
credential
provider.
So
let's,
let's
talk
to
pink,
Bay
and
Andy
to
understand
to
give
us
a
rough
estimate
as
to
one
that
will
be
done
and
I
think
that
seems
to
be
the
last
items
out
there
until
we
can
move.
Is
that
your
understanding
as
well?
Yes,.
A
Yes,
so
I
kind
of
like
issued
a
PR
to
like
actually
move
the
stuff
and
pink
fan
and
you're
like
wait,
hold
on
here's
some
more
stuff
to
consider
so
yeah
I
kind
of
did
that
as
a
signal
signaling
to
say
like
okay.
Let's
do
this
but
yeah
so
I'll
I'll
gather
my
thoughts
and
and
reach
out
to
them,
because
they'll
they'll
have
the
best
idea
of
what's
going
on
there.
Yeah.
A
A
C
A
You
know
we're
definitely
a
sub-project
that
are
currently
a
cig,
but
soon
to
be
a
sub
project
that
needs
more
contributors
right.
So
I
think
that
you
know
a
well-documented
manifests
of
how
to
get.
You
know
how
to
get
a
Azure
cluster
running
with
with
a
cloud
for
what
the
cloud
provider
would
be
good
right.
So
there
is
an
active
kept
for
that.
So
when
it's
it's
part
of
the
out
of
tree
provider
cap
and
then
there's
also,
let
me
link
another
one.
A
All
right
and
that
one
kind
of
describes
the
expectation
for
documentation
for
each
of
the
cloud
providers
so
having
a
an
API
server,
manifest
cube
controller
manager,
manifest
a
cubelet
manifest
and
same
like
for
the
entry
and
out
of
tree
providers
and
as
well
as
documenting
like
how
we
do
resource
management.
So
they
all
the
resources
associated
in
the
cluster,
node
objects,
so
service
objects
and
the
labels
and
annotation
to
attach
to
those
right.
A
So
that
is
rolled
into
part
of
our
auditory
provider.
Cap
and
I.
Think
that
is
probably
like
to
me
at
least
that's
one
of
the
more
urgent
things.
I
think
the
sooner
people
have
better
understanding
of
how
to
do
this
and
how
to
do
it
successfully
that
we
can
get
more
people
on
so
I.
Have
you
know
why
I
have
a
few
co-workers?
Actually
that
have
been
like
their
interest
is
piqued
in
cluster
API
Azure
and
it
would.
It
would
be
nice
to
point
them
in
both
directions.
D
Take
my
mics
working
now,
yes,
Oh
beautiful
I
haven't
I'll,
just
quick,
introduce
myself
too
I'm
I've
been
working
with
kubernetes
last
like
six
months
and
trying
to
get
more
involved
with
the
community
on
it.
So
I'm
kind
of
geared
some
more
of
a
fly
on
the
wall,
but
I
hope
to
contribute
soon
enough,
yeah
for
sure
for
sure.
B
A
A
So
you
know
it's
kind
of
solving
the
meta
problem
like
if
I
could
take
the
you
know
the
same
way:
I
define
different
objects
and
applications
within
kubernetes.
What,
if
I,
could
do
that
with
a
group
like
what,
if
I
could
do
that
for
a
kubernetes
cluster
right
so
being
able
to
have
a
like
a
manifest
and
a
set?
A
and
a
controller
manager
around
the
cluster
object
and
the
machine
objects
right?
So
cluster
is
one
type
of
object
and
then
the
machine
is
the
machine.
A
It
lives
within
the
cluster
object
right,
so
it's
intrinsically
linked
to
the
two
they're
intrinsically
linked
together.
So
you
know
what
we've
done
is
each
provider
are
a
large
portion
of
the
providers
that
are
in
kubernetes
have
started
working
on
individual
provider,
implementations
of
the
cluster
API
API.
A
So
right
now
we're
at
v1
alpha
1.
So
what
that
means
is
we
can
expect
a
lot
of
things
to
change,
but
right
now
we're
pretty
stable
like
we're.
You
know
we're
we're
like
we've
stabilized
on
0
1
0
or
for
you
1,
alpha
1
as
cluster
API,
and
the
next
step
is
like
figuring
out
what
V
1
alpha
2
looks
like
right.
A
So
on
the
azure
side,
what
we
want
to
do
next
is
make
sure
that
we're
building
things
that
people
actually
care
about
right-
and
you
know
people
to
us-
are
you
know
on
the
Microsoft
side,
it's
its
customers
right
people
who
are
consuming
IKS
engine
and
aks
overall
and-
and
you
know,
I
so
I
am
I
I'm
at
VMware.
You
know
and
customers
matter
to
me
too
right
so
I
want
to
make
sure
that
we're
we're
kind
of
pulling
in
the
feedback
that
that
we
need
to
to
make
it
a
successful
provider.
A
C
Yeah
we're
we're
working
on
it
well.
Well,
obviously,
like
we
all
dying
to
know
like
when
that
will
be
done,
but
I
socio
is
the
person
driving
that
right
now.
So,
okay.
C
A
Yeah
it's
like
26
pages
or
something
like
that
and
I
was
like
mm.
This
is
a
later.
This
is
a
later
thing.
It's
read,
so
you
know
so
so,
for
the
second
part
of
this
is
like
the
beach
fat
provider
right.
So
there
is
a
LinkedIn
that
second
half
planning
doc
for
people
on
the
call.
There
is
a
proposal
for
the
idea
of
machine
states
and
pre-boot
bootstrapping,
the
idea
being.
A
We
can
start
to
have
a
common
way
of
describing
like
the
dependencies
that
are
required
to
bootstrap
the
machine
right
so
like
right
now,
the
way
we
bootstrap
machines
and
cluster
API
yeah
sure
the
initial
implementation
of
it
was
via
a
deployment
template
within
go
code
right,
so
template
it
out,
and
then
we
kind
of
write
the
template
to
to
arm
and
we'll
build
a
machine
that
way
right.
That
has
been
rewritten
to
use.
A
Use
all
of
the
azure
API
as
as
opposed
to
just
the
deployment
just
deployment
API
right
so
in
in
that
in
rewriting
that
there's
also
like
a
bash
script.
That's
template
it
out
in
in
in
the
code
right
which
is
kind
of
silly
right,
which
means
it's.
You
know
it's
extremely
brittle.
Actually,
as
I
was
doing
a
test
a
few
days
back,
you
know
one
of
the
one
of
the
dependencies
failed
to
pull
and
it
killed
the
entire
installation
right.
A
So
that's
you
know,
that's
how
brittle
so,
bringing
you
know
bringing
the
pushing
the
dependencies
out
of
the
provider-specific
code
and
into
an
API
that
can
describe
what
they
need
to
be
and
how
they
change
over
time,
also
being
able
to
version
them
right.
So,
if
I'm,
you
know,
if
I'm
going
to
install
a
Kerber,
Nettie's
1:14
cluster,
these
are
the
things.
I
need
right,
I'm
being
able
to
describe
that
as
an
API
as
opposed
to
okay.
A
Like
one
of
the
next
steps
for
everyone
right
and
then
finally,
you
know
there's
interest
in
talking
about
what
it
looks
like
to
provision
aks,
the
cluster
API
adder,
all
right,
so
that's
active
work,
that's
happening
in
Microsoft
and
more
details
will
be
forthcoming
as
I
have
them,
but
the
first
you
know
the
next
big
thing
that
we're
we're
trying
to
suss
out
is
vmss,
because
that
has
come
up
so
many
times
and
I.
Think
that
you
know
it's
an
obvious.
It's
an
obvious
win
for
all
of
us
to
figure
out
so.
B
A
A
Technical
leads
to
get
this
done
all
right,
figuring
out
the
sessions
for
cube
con,
making
sure
that
we
have
like
a
reasonable
meeting
cadence
so
that
we
can
incorporate
the
technical
leads
that
are
in
a
pack,
as
well
as
the
users
in
a
pack
and
potentially
amia
and
and
cover
the
u.s.,
the
u.s.
Eastern
and
Pacific
time
zones
right.
A
So
that's
something
that
we
have
to
figure
out,
we'll
probably
end
up
doing
two
meetings,
two
or
one
meeting
that
is
skewed
time
to
better
accommodate
the
APEC
region,
changing
the
zoom
links
and
the
slack
channels,
and
all
that
good
stuff
and
figuring
out
how
to
deliver
quarterly
updates
and
then
the
really
really
administrivia
like
changing
github
labels
and
and
migrating
projects
from
migrating
projects
to
the
kubernetes
sakes.
Org
right
also
figuring
out
like
how
we
test,
you
know
so
we're
doing
a
good
job
on
the
testing
side
and
then
how
we
release
right.
A
So
we
are
able
to
release
the
project
right
now,
but
I
think
that
we've
only
cut
one
release
for
the
out
of
tree.
One
I'm
not
super
concerned
there,
because
the
out
of
tree
is
essentially
a
vendor
of
the
entry
provider
right.
So
that's
not
a
super
big
deal
now.
I
had
posted
on
Twitter,
because
I
was
like
really
curious
about
like
what
happens
when
you
change
a
project
name
or
what
happens
when
you
move
a
go
project
into
a
different
org
and.
C
A
A
A
So
because
we
have
the
vanity
URLs
as
it
moves
from
the
cloud
provider
Azure
as
it
moves
from
Kate's
as
it
moves
from
the
kubernetes
org
into
the
kubernetes
SIG's
org.
We
should
see
no
outage,
hopefully
right.
The
first
reason
is
that
the
vanity
URLs
do
redirects
to
the
github
URLs
great
and
the
github
URL
itself.
Once
we
change
it
as
long
as
we
only
change
it
once
github
automatically
does
a
redirect
of
that
stuff
right,
so
I
think
we're
covered
across
both
the
import
path,
as
well
as
just
generally
browsing.
A
Those
links
right
so
I
want
to
I,
actually
want
to
do
that
as
soon
as
possible.
So
I'm
going
to
start
looking
into
what
it
takes
to
to
do
that.
The
reason
I
say
as
soon
as
possible
is
because,
right
now
we
have
the
opportunity,
while
people
are
not
consuming
this
as
frequently
like
once
we
move
out
of
tree.
That
will
be
the
one
thing
right
and
anyone
who's
doing.
Imports
will
do
the
imports
there
right.
A
A
But
it's
a
little
better
to
have
a
pretty
front-end
for
it
right
so
hound
to
CS
kto,
and
what
that
allows
you
to
do.
Is
you
can
search
any
search
any
regex
across
all
of
the
kubernetes
repos
that
are
visible
and
get
an
answer
right,
so
I
can
search
for
all
the
things
that
are
imports
of
cloud
provider
Azure
or
listed
in
documentation
as
cloud
provider
Azure
and
change
that
to
kubernetes
SIG's.
A
Slash
glad
for
my
treasure
right
so
I
think
that
we're
we're
mostly
safe
and
I
think
that
you
know
right
now
is
a
good
time
to
try
to
do
that
stuff.
So
as
soon
as
I
gather
the
rest
of
this
proposal
together
and
put
together
the
project
board,
we
can
we
can
get
rolling
on
that
any
questions
we
have
a
minute
left,
I.
C
A
C
A
Okay,
so
once
so,
because
he's
going
to
be
doing
some
presentations
for
Azure
over
there,
I
want
to
make
sure
that
we
don't
confuse
anyone
by
changing
meeting
times.
Just
yet
make
sure
they
join
the
channels.
And
then
we
can.
We
can
Annette.
We
can
decide
toss
out
a
doodle
or
toss
out
a
few
doodles
and
figure
out
if
we're
doing
two
meeting
times,
if
we're
doing
one
and
and
and
then
update
the
calendar
accordingly.
Okay,.