►
From YouTube: Kubernetes SIG Azure - 2019-03-06
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
A
Well,
first
off
I
thought
the
link
in
the
chat
for
the
agenda
itself.
So
if
you're
at
the
meeting
please
pop
into
the
agenda,
just
write
down
your
name
that
you're
attending
and
if
you've
got
any
topics
that
you
want
to
discuss,
feel
free
to
pop
them
on
the
agenda
too,
these
meetings
are
usually
short
and
sweet,
so
the
first
one
I
have
seen
is
that
we're
looking
at
the
project
board
for
backlog,
grooming,
Dave
I
think
that
you
might
have
put
that
up.
So
if
you
want
to
take
that
away.
C
A
A
Yes
No,
maybe
so
looks
good
cool
awesome,
so
looks
like
a
lot
of
a
lot
of
movements
happening
in
cluster
autoscaler
or
a
few
issues
have
been
put
up
for
a
cluster
artists
Kaler
and
support
around
newer
versions
of
a
gas
with
autoscaler
labeling
vmss
supports
availability
zones
and
then
a
lot
of
these
seem
to
be
issues
around
moving
cloud
provider
forward
out
of
tree
cloud
provider
forward,
so
I
know
Rita
may
be.
Do
you
want
to
speak
a
little
bit
about
the
intent
test?
Work
that
you've
been
doing.
D
Yeah
sure
so
for
kubernetes
kubernetes.
As
you
know,
we
have
like
a
like
a
pre
submit
Lennox
job
crowd
job
already
out
there.
We
are
using
it
for
PRS
and
there's.
Obviously
the
wouldn't
knows
one
as
well
and
therefore,
as
far
as
like
cloud
provider,
I
sure
I'm
updating
that
to
use
cube
test,
it's
currently
using
its
own
like
custom
script.
That's
a
work
in
progress,
okay,
I.
D
A
A
Sounds
good
Thanks
all
right,
so
update
on
my
side,
we've
been
working
on
a
migration
of
after
disc
CSI
and
as
your
file
CSI
drivers
from
Andy's
repos
into
into
kubernetes,
cigs
or
kubernetes
CSI,
so
that
their
official
drivers
there
was
a
little
bit
of
slow
down
there.
Just
kubernetes
owners
churning
through
their
their
own
backlog,
so
that
is
seemingly
cleared
up
now
looks
like
Nikita,
has
access
to
admin
access
to
Andy's
repos.
So
she
is
going
through
the
verification
process
of
some
of
the
collaborators
who
have
contributed.
A
Make
sure
that
the
clas
are
in
check
that
the
the
repo
is
actually
adhering
to
kubernetes
standard.
So
it
has
a
contributing
documents.
Readme
is
its
properly
labeled.
It
has
a
appropriate
access
for
kubernetes
contributors
as
well
before
that
migrates
over,
as
well
as
doing
some
cursory
checks
on
the
license
and
dependency
management
tool
Shane
before
getting
that
over.
So
we
run
in
the
background
something
called
fossa,
which
is
a
open
source
license
and
dependency
management
scanning
tool,
so
we're
able
to
I
won't
say
fairly
easily,
but
that's
kind
of
what
we're
moving
towards.
A
So
there's
a
licensing
sub-project
in
cig
release,
myself
Nikita
dims
and
Steve
Winslow,
who
is
one
of
the
legal
dudes
from
Linux
Foundation,
are
part
of
our
the
sub
project
owners
for
that.
So
what
we're
looking
at
doing
eventually
is
being
able
to
turn
on
fossa
across
all
of
kubernetes
to
enforce
licensing
checks
for
repos,
alright.
So
right
now
it
runs
passively
and
occasionally
and
doesn't
enforce
any
checks,
but
we're
able
to
use
that
data
at
least
two
to
kind
of
message.
A
A
That's
all
I
have
on
the
board
personally,
and
it
looks
like
most
of
things
are
end
to
end
most
of
things
active
or
intend
or
are
out
of
tree
related,
so
I
think
I
will
be
in
I.
Don't
know
how
many
of
you
are
based
in
San
Francisco,
but
I'll
be
in
San.
Francisco
next
week,
I'll
be
in
California
period
next
week.
The
first
part
of
the
week
for
the
open
source,
Leadership
Summit,
we're
doing
a
panel
for
that
and
then
Friday
I'll
be
at
Microsoft
with
lucky.
A
A
There
we
go,
it
was
cap,
see
progress
right.
So,
in
the
background
we
have
been
working
on
myself,
Tariq
and
Nick.
Lane
has
been
working
on
one
getting
cap
C
to
a
point
where
we
can
open
it
up
to
contributors
and
then
also
getting
it
to
some
baseline
expectation
of
like
this
is
a
cluster
now
right
so
immediately
before
this
meeting,
I
was
doing
a
demo
with
with
Nick
and
and
Tariq
to
kind
of
show
off
that
we
actually
can
build
it
for
master.
A
Now,
right
and-
and
there
are
some
baseline
things
that
happens,
it
will
create
a
load
bout,
a
load,
balancer
standard
public
IP,
a
standard
NIC.
The
control
plane
connects
the
backend
pools
and
all
that
good
stuff
and
stand
up
a
one.
Thirteen
point:
four
one,
thirteen
point:
four
cube
ADM
cluster
right,
so
that
is
pretty
exciting
news.
Okay,
so
that's
pretty
exciting
news.
I
want
to
show
off
the.
E
A
I'm
gonna
show
us
the
backlog
a
little
bit,
sir.
Prior
to
that
meeting
last
night
and
earlier
today,
I
cleaned
up
some
of
the
things
that
we
need
to
be
working
on,
filed
issues
for
some
of
the
stuff,
and
so,
if
anyone
is
actually
interested
in
working
on
a
cloud
provider,
cluster
API
provider,
a
sure
we're
getting
to
the
point
where
we
can
actually
open
it
to
contributors.
A
A
So
some
of
the
things
that
I
have
on
my
plate,
too,
is
make
sure
that
we
there
are
a
few
bugs
in
in
the
logic
for
reconciling
the
the
load,
the
backend
pools
for
the
load
balancers
and
the
public
IPS
against
Nicks,
and
those
are
the
things
that,
once
those
are
done,
I'll
be
able
to
say
that
like
master
actually
builds
and
builds
consistently,
I
also
want
to
document.
So
I
started
documenting
the
developer
workflow
so
how
to
actually
get
involved.
A
A
So
these
are
all
available
as
issues
if
anyone
is
interested
in
starting
to
work
on
any
of
the
cap.
C
stuff
have
a
little
extra
bandwidth
in
your
schedule
and
you're
interested.
Let
me
know
some
of
the
things
that
we're
looking
at
investigating
longer
term
support
for
availability
zones
across
the
board.
Right
now
there
are.
There
are
some
parts
of
the
logic
that
hard
codes
against
zone
3,
and
it
also
makes
the
assumption
that
you
will
be
using
an
azure
region
that
supports
availability
zones
so
moving
forward.
A
We
need
to
write
logic
that
can
intelligently
figure
out
what
availability
zones
are
available
in
that
region,
whether
or
not
you're
in
a
region
that
supports
availability
zones,
what
availability
zones
are
available?
So
you
know
design-wise
I
went
with
load
balancer
standard
because
of
the
support
to
attach
virtual
machines
that
are
in
that
our
availability
zone,
virtual
machines
into
the
load,
balancer
I,
think
that
you
can
only
do
that
with
load
balancer
standard
if
I'm
correct,
but
I've
got
a
bunch
of
Microsoft
people
on
sale.
If
I'm
wrong.
Let
me
know.
A
Sure
so,
right
now
in
trying
to
get
us
to
the
point
where
it's
actually
working
and
it
like
master
one
passes
all
of
the
tests
and
then
also
stands
up.
Cluster
I
made
the
decision
to
one
hard
code
availability
zones,
so
they're
all
pin
to
zone
3,
but
also
for
to
be
able
to
use
availability
zone.
Virtual
machines
I
believe
that
you
need
to
use
a
load,
balancer
standard
and
pip
standard
right.
A
I'd
also
like
to
look
at
moving
to
virtual
machine
scale,
sets
I
know
that
there
are
so
when
I
was
working
on
tectonic
Fraser,
so
I
wrote
a
decent
chunk
of
the
tectonic
crash
or
logic.
I
know
there
were.
There
is
a
difference
in
the
virtual
machine.
Api
and
the
virtual
machine
scale
sets
API
where
it
makes
it
tricky
to
reach
into
a
scale
set,
and
we
have
the
identity
of
a
machine
essentially
so
based
on
what
has
been
built
in
cluster
API.
A
So
far,
a
lot
of
a
lot
of
it
is
targeting
using
the
the
machine
API
as
opposed
to
machine
lists
and
machine
set
machine
deployment.
Right
that
said,
I
think
machine
sets
would
probably
be
analogous
to
what
a
virtual
machine
scale
set
would
be,
but
it
would
still
need
logic
to
reach
into
the
scale
set
and
understand
the
machines
in
an
action
on
one
machine
as
opposed
to
the
scale
set.
A
If
anyone
is
interested
in
jumping
and
picking
something
up-
or
you
know,
people
well
I
would
love
to
see
or
people
who
are
interested
in
writing
tests.
So
if
there's
anyone,
if
there's
anyone
that
you
can
think
of,
that,
might
be
interested
in
writing
tests
for
this
repo,
let
me
know-
and
then
okay,
so
cig
a
cloud
provider
consolidation.
A
A
Actually,
I'm,
not
I'm,
not
sure
I
can't
share
just
yet,
but
we
have
been
in
discussions
about
like
what
it's
going
to
take
to
to
consolidate
under
it
so
cm
channels.
You
know
it's
assumed,
accounts
slack
channels.
What
the
what
the
cloud
provider
repos
would
be
named
going
forward,
what
the
sub
projects
would
be
named,
who
owns
them?
How
to
redefine
some
of
that?
What
we
have
to
do
on
the
label
side,
what
we
have
to
do
on
the
communication
side?
A
If
there's
anything,
we
have
to
do
on
the
on
the
github
automation
side
right
so
trying
to
drive
down
a
lot
of
those
those
little
nitti
things,
so
that
proposal
is
almost
complete
and
I
think
we're
planning
on
presenting
it
to
steering
committee
at
the
end
of
the
week,
we're
just
giving
a
few
more
people
time
to
review
it.
So
that's
coming
soon.
What
that
means
for
these
meetings
in
the
future,
I
don't
know
yet
I
would
like
to
I
would
imagine
that
each
provider
would
like
to
continue
meeting
I.
A
If
we
were
to
enumerate
all
the
potential
providers
I,
don't
think
that
it
would
necessarily
be
a
valuable
use
of
time
for
the
project
right.
So
so
that's
coming
soon.
I
think
that
you
know
there
were
some.
There
were
some
thoughts
about
making
it
happen,
a
towards
coupon,
Shanghai
and
I
kind
of
think
that
we
should
push
that
out
a
little
bit.
So
you
know
some
of
the
discussion
was
also
around.
Do
we?
A
How
do
we
present
provider
changes
right
so
who?
How
often
do
people
present
at
the
community
meetings?
Who
gets
cube
con?
Who
gets
cube
con
sessions?
Things
like
that
right
so
because,
because
Barcelona
and
Shanghai
are
coming
up
soon,
I
kind
of
said,
like
hey,
we
should
not
make
a
decision
or
like
whatever
decision
we've
made
about
provider
sessions.
Right
now
needs
to
stand
right
because,
like
people
have
already
submitted
proposals
around
these
tracks,
so
the
soonest
we
can
make
a
change.
A
There
is
for
San
Diego,
keep
con,
saying,
San,
Diego
right
but
yeah,
so
that
that's
in
flight
right
now,
as
I
have
more
details,
I'll
I'll,
let
you
all
know
I
will
stop
talking.
You
know.
We've
got
seven
minutes
left.
If
people
have
questions
comments,
concerns
things
that
they
want
me
to
track
down
for
them.
Let
me
know.
A
So
anybody
else
I
think
I
think
everyone
knows
everyone
here.
If
anyone
wants
to
introduce
themselves
is
new
to
the
call
I
see
actually
I
see
two
names
I
have
not
seen
before.
So,
if
you
want
to
just
shout
out,
say
hi
what
you're
here
for
what
you're
looking
to
do,
if
we
can
help
you
in
any
way
go
for
it,
hey.
F
It's
speaker
I'm
from
PTC.
Actually
we
are
thinking
we,
we
you're
using
Cardinal
project
for
architects
for
provisioning,
artists
cluster
another,
and
now
we
are
thinking
about
moving
into
aks
and
we
are
kind
of
interested
in
how
we
can
start
contributing
to
the
either
upstream
issues
about
other
or
just
a
case
engine
or
anything
about
a
case,
because
we
will
be
asked
in
quite
many
clusters
and
I
guess
more
likely,
so
I
just
wanted
to
get
warm
up
for
the
team
here.
So
that's
why
I'm
joining
picture
welcome.
A
G
A
Welcome
welcome
so
next
time
around
I'm
gonna.
Do
the
I'll
do
the
interest
at
the
beginning,
instead
of
leaving
them
for
the
end,
but
yeah
so
I
I
think
that
we,
you
know
there
have
been
discussions
about
like
the
what's
the
appropriate
aks
aks
communication
platform.
I
know
I've
noticed
that,
like
a
lot
of
the
a
lot
of
the
Microsoft,
people
have
come
to
hang
out
in,
say
gosh
or
the
cig,
a
short
rack
slack
channel
within
the
kubernetes
workspace.
A
So
that's
been
awesome
in
terms
of
like
just
general
support
and
having
having
a
lifeline
into
Microsoft
there.
I
would
like
to
see
I
would
love
to
see
like
more
interaction.
I
would
love
to
see
eventually
something
of
a
a
roadmap.
I,
don't
know
what
people
are
willing
to
share
on
the
Microsoft
side,
but
it
would
be
cool
to
have
a
public
road
map
that
we
can
kind
of
tied
together
with
a
kubernetes
project.
So
I
know
that
there
sorry
different.
A
A
C
Yeah
so
I
know
we
had
this
discussion
offline,
a
little
were
kind
of
the
kubernetes
sig
edger
slack
channel
of
what
that
will
become
talking
what
Paris
they
would
almost
like.
Seeing
that
not
be
a
support,
channel
and
I
think
that
goes
across
all
cloud
providers
that
it's
more
just
for
sig
as
a
related
stuff
and
not
a
support
form
for
cloud
providers.
I,
don't
know
if
you
heard
any
more
on
that
or
yeah.
A
So
part
of
that
proposal
is
actually
discussion
around
the
delineation
between
like
so
the
first
part
of
the
consolidation
was,
do
we
consolidate
into
a
working
group
or
do
be
consolidated
into
sub-project
right
I
think
that
sub
projects
are
more
appropriate
because
they
sub
projects
by
definition
own
code,
and
we
have
code
that
we
need
to
own
right
within
the
kubernetes.
It
works
right.
So
that
is
that's
part
of
it
right.
The
second
part
of
it
is
they
they've.
A
Recently
they're
working
out
the
details
for
a
new
organizational
group,
called
user
groups
right,
so
user
groups
would
eventually
be
the
place
where,
if
people
wanted
to
organize
and
talk
about
a
thing
whatever
that
thing
may
be,
if
it
is,
if
it
has
the
requisite
amount
of
interest,
I
think
they
would
be
willing
to
spin
up
a
user
group.
I
know
there
was
so
just
in
Santa.
Barbara
was
saying
that
you
know
on
the
AWS
side
there
has
been.
You
know
they
kind
of
have
two
modes
of
communication
right.
A
One
is
sick,
Channel
and
then
there's
also
I.
Think
there's
like
a
user
group,
ish
channel
right.
So
his
you
know
his
want
is
to
be
able
to
organize
people
in
the
right
places.
Alright,
my
my
thought
was:
if
we
have
people
going
to
two
places
right,
the
conversation
kind
of
gets
disconnected
I.
Think
that
you
know
the
the
argument
can
be
made
that
users
there's
no
one
definition
of
a
user
right.
I
am
a
you.
You
know,
I
am
an
end-user
of
azure,
but
I'm
also
chair
for
a
sure,
I'm
also
developer
I'm.
A
Also
a
you
know,
product
manager,
right
and
like
and
people
will
fall
into
a
bunch
of
like
different
buckets
like
that.
I'm,
not
sure
that,
like
having
separate
channels
for
that
is
the
best
use
for
moderator
time,
the
the
channel
is
I.
Think
that's
probably
the
funny
enough
naming
is
is
one
of
the
final
points
of
contention
that
we're
working
on
for
like
across
the
board,
because
everyone
loves
a
good
bike
shed
and
and
then
also
what
we
do
with
the
channels
right.
So
that
is
TBD
right.
A
If
it's
going
to
be
like
a
provider,
Azure
Channel,
right
or
a
platform,
a
church
annal
right
where
it
can
be
both
the
user
discussion
and
and
kubernetes
specific
interests
right
for
development
chatter.
So
the
short
version
is
we
don't
know
yet
and
I
think
we'll
start
to
see
it
a
little
bit
better.
Once
we
button
up
the
proposal
and
prevent
presented
to
searing
so
stay
tuned
for
that,
that
was
a
long-winded
way
of
me.
Saying
I,
don't
know
yet.