►
From YouTube: SIG Azure Valentines Day Edition 2018
Description
A
Hey
everybody:
if
you're
joining
us
you're
watching
the
singer
community
meeting
for
Wednesday
February
14
the
filled
with
love,
Valentine's
we'd
love
you
from
Microsoft
as
you're
sort
of
signaling,
not
pleasant
music,
this
time
necessarily,
but
you
will
have
the
soothing
mellifluous
voice
of
mr.
Collett
back
telling
you
all
the
things
that
are
going
right
and
wrong
a
measure
and
we'll
go
from
there.
So
all
right,
Kel!
Kick
it
off!
Tell
us
what's
going
on
and
just
let's
run
through
the
the
agenda.
That's
a
bit!
Dalia!
/!
Stick!
Yes!
So.
B
And
she's
working
diligently
on
it
we
expected
available
for
one
time
the
vmss
support
is
done.
Actually
it's
out
there
it's
merged
and
it's
ready
to
go.
So
you
have
time
in
your
hand,
please
go
ahead
and
test
it.
The
autoscaler
was
still
questionable.
I.
Think.
The
reason
behind
that
is
was
the
vmss
availability
sets
everybody's
trying
to
figure
out.
Where
should
we
go?
B
I
promise
you
next
week,
I'll
have
a
clearer
status.
Honor
at
Gateway
is
not
gonna,
make
it
to
1:10.
The
reason
behind
that
is
Ivan
was
working
on
a
cache
of
today's
focus.
A
bit
on
to
the
visual
studio
extension
for
Cooper
natives
produced
a
new
code,
all
the
funky
stuff
that
you
see
on
visual
studio
code
right
now
is
actually
his
work
and
I.
Think
it's
some
small
of
a
priority
for
for
developers,
then
just
be
a
bit
way
for
now
out
of
tree
provider.
You
all
know
it's:
it's
there.
B
The
agile
disk
encryption
and
the
explicit
MSI
does
not
look
like
it
will
make
it
Chinese.
New
Year's
is
next
week
and
it's
eating
up
a
good
week
or
10
days
of
our
time
to
actually
do
something
about
these
things.
Specifically,
the
encryption
I'll
skip
the
windows
and
oh
nice
integration
for
now,
because
I
don't
have
any
information
on
them.
The
field
driver
is
on
track.
I
have
a
promise
that
it
will
actually
make
it
to
110.
Please
to
understand
that
this
is
not
an
OP
stream
per
se.
It
doesn't
really
stop
stream.
B
Schedule
modification
of
I
have
I
have
started,
and
it
looks
like
we
may
need
to
change
the
something
in
the
disk
to
mark
that
just
plug
in
to
mark
disks
was
either
premium
or
numb
bring
me
in
order
for
the
schedule
to
understand
that
this
disk
shouldn't
go
into
that
mode
and
so
on.
I'm
still
wasted
working
through
this
I'll
have
more
information
next
week
on
it.
B
Natural
as
refinery
sighs,
I'm
gonna
drop
like
I'm
on
I
skipped
and
I'm.
Now,
at
the
bauxite
refinery,
sighs
I
almost
hit
on
sake.
We
need
to
do
this.
For
now,
the
mount
option
is
on
track.
The
Flex
volley
I
mean
I'm,
just
gonna
file,
everything
Windows
in
one
pocket
until
I
get
more
information
on
it.
Next
week
on
different
news,
we
did
merge
two
cars,
actually
one
PR
and
the
other
one
is
come
to
merge
really
soon.
The
first
PR
is
we
the
services,
the
controller
control
Nigel.
B
Had
this
black
code
services
sync
period,
which
was
advertised
as
this
will
control
how
much
how
the
interval
between
synchronizing
the
services
with
external
load
balancer,
which
I
got
excited
about,
because
we
have
a
problem
with
large
clusters
and
arm
of
the
parent
about
everything.
And
then,
after
careful
consideration
this
this
flag,
we
discover
that
it's
not
used
it.
Just
there
is
dummy
to
confuse
people,
so
we
should
we
moved.
We
worked
on
removing
it
and
we
managed
the
community
to
convince
the
yeah.
B
A
Out
something
real
quick
in
regard
to
that
that
this
is
this
is
I,
mean
sure.
This
is
great
that
we
don't
have
to
deal
with
this,
but
this
is.
This
is
actually
a
service
to
the
community
at
large.
This
sort
of
cleanup
effort,
so
I
just
want
to
call
out
Callie
you
stepped
up
above
and
beyond,
and
you
did
a
lot
of
hoop
jumping
to
do
the
right
thing
and
I
sincerely
appreciate
it.
I
mean
that's
the
level
of
commitment
to
community.
That's
really
making
a
difference
here.
So
thank
you
and.
B
B
B
When
you
go
and
configure
a
cluster
on
Android,
you
get
what
we
call
a
rate
limit
right.
There
is
limited
that
controls
the
number
of
calls
that
cloud
provider
execute
against
arm.
Now,
when
you
do
that,
the
problem
is
arm,
separates
between
get
calls
and
write
a
specific
goal
such
as
put
and
post
all
right
and
delete.
So
with
one
in
the
big
limiter.
It
gets
really
confusing
because
it's
two
counters
but
we're
losing
one
limiter
for
them.
So
we
started
separating
the
because
the
rate
limiter.
B
I
really
would
love
a
call
or
other
people
who
take
a
peek
at
that,
because
we're
really
changing
the
behavior
of
how,
as
a
cloud
provider
Kosar
this
entire
effort
of
caching-
and
this
is
the
other
one,
I'm
cashing
without
saying
what
it
is.
Let
me
just
I
am
running
with
two
cups
of
coffee:
I
usually
operate
after
6:00.
So
how
long.
B
This
other
link,
I'm
posting
on
the
chat
window,
has
a
cache
with
bet
for
the
ends
for
Anna's
reroute
table,
along
with
the
work
we've
done
on
the
rate.
Limiter
is
all
driven
by
working
diligently
to
ensure
that
we
have.
We
can
support
large
clusters
by
default
without
really
any
hiccups
or
any
specific
configuration
for
a
default,
the
code
to
run
yes,
people
are
expected
thrown
at
City
on
separate
nodes,
highly
available
masters
and
the
other
other,
which
is
normal,
but
I
don't
want
them
to
do
anything
in
order
to
create
large
cluster
than
that.
B
C
B
Last
piece
merge,
I
think
yesterday,
which
was
disc
support,
so
that
was
the
last
piece.
I
know
everybody
is
happy
and
everybody
wants
to
like
use
it.
Please
do,
but
my
advice
to
you
is
because
it's
new
and
we've
tested
it
and
everything
I
would
take
it
slow
like
use
it
and
Destin
and
f-for
now,
right
until
we
were
absolutely
sure
that
it's
ironclad
like
this
is
what
you
should
do
so.
C
I
had
a
meeting
with
one
of
our
support.
Guys
got
me
a
meeting
with
ping.
Fei
is
it
who
did
this?
Yes,
that
you
tested
with
like
ten
fifteen,
but
he
couldn't
get
a
successful
one
hundred
nodes.
So
I'm
just
wondering
if
you
have
more
updates
on
this,
or
at
least
yeah
I'm
gonna
be
trying
a
hundred
nodes
like
I'm
trying
to
do
this
by
tomorrow,
yeah
back
so
so.
C
B
When
you
do
your
test,
can
you
please
just
do
your
feedback
as
either
contribute
an
MD
file
with
what
you
did
or
if
you
found
then
I
should
just
create
an
issue
for
us
to
track
it
all
right.
Okay,
the
reason
behind
that
is
I
want,
so
you
gonna
do
this
I
want
the
other
people
who
will
do
this
after
you?
Do
it
to
start
from
where
your?
When
will
from
that
point,
you're
done
basic
I,
don't
want
to
repeat
the
work.
I
want
people
to
like
hey.
We
already
know
how
to
do
these
things.
C
C
It's
like
it's
been
going
on
for
a
week
now
and
if
I
go
into
portal
and
I
remove
the
disk
from
the
note
that
I
currently
attached
mm-hmm
everything
kicks
in
properly,
like
the
controller
picks
up
the
disk,
attach
it
to
the
new
nodes,
announce
it
and
then
the
pod
gets
it
and
everything,
but
for
some
reason,
I
can't
find
the
reason
why
it
won't
unattached
it
or
unmount
slash,
remove
it
from
a
node.
Is
this
managed
manage
diskant
blob?
This
manage
this
asier
disk.
B
It
is
what
I
wanted
to
do
with
you.
Can
please
I
want
you
to
file
an
issue
and
in
order
to
short-circuit
the
process,
I
want
you
to
share
the
control
of
this
case
scrub
any
PII,
or
anything
that
you
don't
want
to
put
on
the
public
and
just
change
the
control
log
for
us
and
be
generous,
like
don't
don't
don't
Cheryl
part
of
the
chair
as
much
as
you
can
of
the
control
evie
a
tweet
nv8.
B
I'll
be
honest
with
you:
there
is
that
there
has.
There
is
an
issue
raised
yesterday
about
blob
disks
and
when
I
looked
at
it,
there
is
something
extremely
EFI
about
it.
So
I
want
to
make
sure
if
he-
and
it's
not
kubernetes
side,
it's
probably
the
PM's.
Something
is
off
so
I
want
to
compare
contrast.
The
two
issues
together
and
if
so
I
will
follow
it
without
dependents.
Important
and
just
you
will
be
used
as
tracking
point.
Basically
for
this.
C
C
I'll
do
both
so
those
are
gonna,
be
my
next
two
things
right
now:
it's
testing
the
VMS
s
and
really
getting
to
the
bottom
of
why
I
like
PVCs
and
TVs
with
Azure
discs
are,
are
not
flawless.
I'm
running
one,
seven
twelve,
which
I
saw
was
supposed
to
be
to
receive
some
some
some
back
ports
of
issues
related
to
disk
but
I'm
guessing.
B
The
issue
that
was
reported
yesterday
was
on
1/9,
so
one
mine
at
least
had
the
same
things
that
you
will
get
in
1,
7
12
in
terms
of
disks,
so
either
we
have,
we
really
really
messed
up
and
one
of
our
merges
or
we
have
something
that
we
don't
understand
and
underlying
p.m.
and
so
on.
In
all
cases,
it's
it's
a
priority
for
me
to
figure
this
out,
because
it's
looks
to
me
like
started
to
be
a
widespread
issue,
so
we
need
to
work
it
out.
Ok,.
C
And
just
to
complete
I,
it
was
more
stable
before,
but
I
love
strike
of
what
changes
I
introduced
that
caused
it,
but
I
know
for
sure
before
was
it
wasn't
it
was
stable,
but
now
it
I
don't
know
it.
Was
it
like
when
I
moved
to
that
7.11
that
7.12
but
I
know
one
of
those
just
made
it.
It
works
all
right.
I'll
show
some
luck.
Yeah.
B
B
A
B
A
All
right,
I,
assuming
we
don't
have
anything
else
to
talk
about
I'm,
going
to
adjourn
the
official
part
of
the
meeting,
stop
the
recording,
then
I'm,
relaxing
piano
music
and
let
the
let
anyone
who
wants
to
have
some
sense
in
moments
stay
for
a
bit
and
then
go
about
your
day.
So
thank
you.
Everybody
truly
appreciate
it
as
always,
but
thank
you
thank.