►
From YouTube: SIG Azure Community Meeting 20180207
Description
The most relaxing SIG meeting. http://bit.ly/sig-azure
A
A
A
C
B
C
A
A
C
B
The
whole
idea
is:
if
we
manage
to
have
a
node
cluster
in
one
subscription,
our
users,
whoever
they
are
either
using
any
service.
It's
a
respective
of
what
it
is.
It
will
be
able
to
run
either
smaller
clusters
or
many
many
smaller
clusters
in
the
single
sketchy,
and
that's
the
primary
drive
for
doing
this
as
basically
we
suffer
so
other
people
boat,
so
I'm,
sacrificing
myself.
B
There
are
two
to
two
primary
limits
that
were
facing
right
now.
Two
primary
issues
that
were
facing
will
be
scaled
beyond
certain
number
of
nodes
or
number
of
clusters
per
subscription.
Those
are
related
to
arm
itself
our
main
point,
so
we
the
way
the
controller
works.
The
way
the
club
provider
works
right
now
when
it
supports
all
the
controllers
that
talks
to
it.
It
just
creates
a
lot
to
get
calls
against
our
main
point,
so
we
quickly
consume
arm
coated
pair
subscription,
so
we're
working
against
that
okay
working
to
fix
that
with
reserved
a
PR.
B
The
second
topic,
which
is
really
around
that
how
the
availability
sets
the
network
and
all
of
that
things
work
right
now
with
a
Dracaena.
We
should
be
in
a
better,
better,
better
place
to
actually
experiment
beyond
2500
nodes,
and
we
should
be
able
to
just
go
and
create
additional
nodes.
So
those
are
the
two
limits
that
we're
trying
to
work
out
with
this
project.
B
B
And
then
yeah,
so
this
is
this
is
progressing
slowly
and
it
with
someone
many
months
project
is
not
a
small
like
100
approach.
It's
one
activity
at
every
week
on
it
and
then
we'll
see
how
it
progress
right
now
there
are
some
stuff
we
know,
such
as
the
armed
limits
and
some
stuff.
We
don't
know
that
would
make
once
we
go
and
add,
let's
say
3000
and
so
on
sure
he
has.
These
nodes
are
running
with
God
provider
configured,
which
is
kind
of
interesting
challenge.
To
do
what
you
see.
A
B
A
B
You're
keeping
an
eye
call
and
the
rest
of
the
people
watching
this.
We
recently
discovered
that
some
settings
are
actually
not
working
so,
for
example,
the
service
sync
variant
on
the
controller
manager
is
not
used,
angular,
which
is
kind
of
confusing
for
users,
because
this
was
saying
that
I
can
configure
the
service
sync
period
tense.
I
am
slower
not
not
as
good
consuming
resources
as
as
I
should,
but
this
is
actually
not
not
correct
and
we
created
up
in
order
to
remove
that,
but
that
setting
is
not
used
and
we
created
appear
to
me.
C
C
A
B
Hope
a
mess
we're
on
progress.
We
have
great
time,
I,
don't
know
if
you
guys,
Marita
is
working
on
it
and
she's
progressing.
I
talk
every
now
and
then
about
this
and
it's
progressing
it's
on
track,
I'll
leave
by
this
end
of
this
week
or
next
week
we
shall
have
something
to
share
VMs
as
support.
We
have
everything
except
disks
and
that's
building
on
the
other.
The.
B
Arm
so
the
API,
the
SDK,
doesn't
have
the
API
to
attach
disks
individually
to
nodes
and
that's
not
bending
on
the
endpoint
itself.
It's
pending
on
the
SDK
imperative.
So
we're
waiting
for
this
right
very
strong
point
here
is
once
this
is
done.
We
we
really
need
to
eat
to
like
we
need
to
be
careful
people
using
it
immediately
in
production,
because
we
need
to
keep
it
and
test
for
a
while
and
then
move
it
into
proper
clothes,
because
it's
a
massive
change
on
the
way.
A
B
B
So
Ivan
Ivan
is
moving
ahead
with
this,
and
E
is
actually
started,
doing
doing
some
like
Christian
examples
and
so
on.
So
this
is
progressing
out
of
three
provider.
We
still
at
the
stims
situation,
where
we
have
the
code,
but
we're
waiting
for
clearance
to
get
it
to
just
open
up
the
repo
okay,
but
that
meaning
is
actually
today
I'll
check
in
on
that,
and
everybody
now
can
I
ask
you
to
update
from
this
one
next
time,
because
I
really
don't
know
how
to
get
updates
on
this
particular
action.
I'm.
B
B
B
D
B
B
So
that's
that's
one
of
the
use
cases
that
this
is
going
to
be
interesting
for
sure.
Okay,
it
comes
with
fine
print
fuse
block
drivers,
or
this
is
user
space
drivers
they
come
at.
They
are
a
little
bit
slower,
but
the
agile
block
block
muck
will
op
storage
is
in
terms
of
data,
can
support
massive
amount
of
data,
so
we
will
be
getting
it
from
Doc's
around
this.
Certainly
enough.
The
storage
folks
as
well
stores
full
of
scale
in
the
driver
and
just
don't
have
enough
Nueske's
dogs
around
it
so
we'll
have
to
own
this.
B
B
A
B
Thing
I
want
to
talk
about
is
there
is
a
bunch
of
very
old
issues
and
either
we
are
referenced
directly
as
agile,
bulk
or
issue,
or
we
just
get
referenced
in
some
discussion
on
either
the
cloud
provider,
networking
issues
and
so
on
right.
Some
of
these
districts
are
extremely
old,
just
to
make
a
point,
the
search,
a
sure
on
issues
all
right
on
coconuts
issues,
and
you
will
see
at
least
10
pages
of
issues
open
all
right.
This
is
any
issue
that
has
the
word
hazard
in
it.
B
So
what
I
started
to
do
is
every
day
I'd
spend
15-20
minutes
just
going
back
and
I
start
like
talking
either
we
need
this.
We
don't
need
this.
Can
we
close
it
just
for
us
to
until
focus
on
a
smaller
set
of
issues
and
I
will
be
going
back
to
you,
Jase
that
close
or
move
or
move
these
things?
Okay,
yeah,
just
at
me
and
I'm.
Also
writing.
I'm
gonna
share
it
probably
next
week,
I've
just
a
simple
Python
script.
That
goes
against
that.
B
This
is
a
side
project.
I
have
no
comment
on
it
right,
just
a
simple
tool
that
grabs
all
the
issues,
all
the
PRS
and
just
was
date,
issue
ID,
and
we
start
having
some
consistent
way
of
just
because
the
worst
that
could
have
as
you
you
find
yourself
mentioned,
or
an
issue
that
has
been
in
discussion
three
months
for
three
months
now
and
then
hey
agile.
This
also
affects
agile
yeah.
Well,
thank
you
very
much,
but
with
you
guys,
three
months
into
the
discussion,
there
is
nothing
nothing
we
can
do
now
other
than
all
right.
B
C
B
Appreciate
that
a
lot
of
people
and
the
community
realized
this
like
realize
some
people,
not
even
in
natural
communities,
you
realize
the
impact
and
dimension
of
either
you
Brandon
myself
or
just
take
agile
directly
and
that's
highly
appreciated.
But
it's
not
exactly
something
you
should
be
defending.
Oh.