►
From YouTube: Orchestrator with refactored Provisioner 2020-08-10
A
Right
disclaimer,
I'm
not
on
my
best
internet,
so
in
case
I
drop
out
in
the
middle
for
no
reason.
You
know
why.
A
Okay,
so
today
I
think
I'll
actually
go
through
one
of
the
orchestrator
marks,
which
is
not
a
new
feature,
but
essentially
refracting
the
existing
functioning
module
to
use
better
module
so
to
remove
all
the
v1
codes.
So
in
case
you're
wondering
this
is
multi
request
77
in
the
orchestrator
pro.
A
A
I
have
the
rc
file
corresponding
to
the
necessary
configuration
for
orchestrator
up
some
configuration
settings
for
orchestrators
to
use
okay,
so
those
are
all
actually
demo
things.
I
I
don't
have
any
actual
sensitive
information
in
there.
A
A
A
A
C
B
B
C
Yes,
this,
this
does
away
with
the
split
development.
C
A
A
B
B
C
C
Inputs
and
just
for
the
purpose
of
history
of
the
call
we
now
support
using
a
project
as
a
back
end,
so
you
can
use
you
want
to
step
back
and
type
gitlab,
and
then
you
need,
like
your
personal
token,
for
that
project
for
the
api,
and
you
can
set
that
up
and
you
can
actually
use
git
lab
as
your
backend
instead
of
google.
Thank
you
mitchell.
Nelson.
For
that
merge.
B
A
B
A
B
A
A
B
B
C
A
C
Although
it
looks
like
you're
running
non-geo,
so
I
think
the
answer
will
be
13
minutes.
It's
23
geo
adds
10
minutes,
which
is
something
we're
gonna
need
to
probably
need
to
look
into,
because
adding
two
more
nodes
should
not,
like
you
know,
double
ish
the
time.
So,
okay,
that's
why
we
did
the
type
again.
Thank
you
mitchell
for
adding
the
timer
flag.
A
So,
but
for
geo,
rail
stuff
also
comes
into
play
right.
You
need
to
set
up
replication
and
all
those
things
too,
not
just
spinning
up
the
machine.
C
Yeah,
there's
replication
and
yeah
there's
a
bunch
of
other
stuff
it
may
be.
It
is
what
it
is
because
of
all
the
database
stuff,
but
the
big
thing
was
when
we
started
this,
I
was
I
really
wanted
instrumentation
and
I
didn't
have
time
for
it
and
mitchell.
Surprised
me
with
an
mri
is
like
hey:
here's
timings.
I
was
like
yay
so.
A
A
B
A
C
That's
the
thing
we
actually
have
kind
of
an
open,
open
question
about
that,
one
of
our
other
issues
as
well,
because
mitchell-
and
I
like
we-
keep
all
of
our
notes.
Just
like
we're
doing
certain
work.
We
just
keep
all
of
our
shared
findings
in
one
issue,
but
it
confuses
the
heck
out
of
everyone
else
because
we're
treating
a
google
doc
where
we
just
put
stuff
in
but
there's
no
deliverable
around
it,
and
so
we
found
out-
and
I
found
out
last
meeting-
it's
confusing
everybody
because
they're
like
is
this
deliverable.
C
C
D
C
E
C
E
E
I
left
a
comment,
though
thinking
you
know
if,
if
this
research
issue
is
really
just
us
taking
notes,
if
we're
eventually
planning
to
have
like
an
epic
or
an
overarching
issue
for
the
goling
conversion,
we
could
put
those
kinds
of
notes
there.
It
might
make
more
sense,
since
that
those
would
be
notes
on
an
overarching
project
instead
of
just
a
standalone
issue
for
some
notes.
If
that
makes
sense,.
F
I
do
think
you
know
there
may
be
some
confusion
around
the
way
it's
being
used
now,
but
you
know
if
the
folks
that
are
doing
the
work
are
able
to
make
progress
quickly
on
it,
then
that's
kind
of
like
I'd,
say
the
prime
directive,
and
then
we
can
explain
how
we're
using
it
to
others
but
yeah.
I
like
the
idea
of
dog
food
and
other
other
features.
E
F
The
model
we've
been
using
is
like
here's,
the
big
epic,
here's.
The
big
thing
that
we
want
to
do
is
the
epic
and
then
the
first
issue
is
here's
the
spike
to
go
figure
out
all
of
the
things
that
we
have
to
do
to
complete
this
epic
and
then
and
then
more
issues
fall
out
of
that
spike
to
populate
the
overall
epic.
E
F
F
That
model
works
really
well
when
working
with
like
product,
because
then
they
see
like
hey,
here's
the
big
thing
and
then
they
start
to
see
the
progression
over
time
of
you
know.
As
the
spike
is
completed,
the
plan
starts
to
formulate
itself.
E
F
After
this
I'll
go,
take
a
look
and
see
robert
and
me
are
gonna
chat
later,
so
we
can,
we
can
look
at
it.
We've
been
doing
a
lot
of
interesting
things
on
the
orchestrator
project,
in
terms
of
you
know,
using
weights
and
other
things
that
we
haven't
done
with
some
of
the
other
projects,
just
because
this
one's
new,
it's
a
nice
opportunity
to
spread
our
wings
a
little
bit.
A
B
C
C
B
A
B
C
So
I
don't
see
it
there
looking
through
and
is
the
one
who
like
helped
write
most
the
documentation,
I'm
going
to
say:
let's
call
that
out
it's
the
thing
that
it
should
be
written
down
somewhere
what
it
is
it's
likely
to
change
to,
like
maybe
the
2k
reference
architecture,
but
for
right
now
it's
the
on
the
us
pipeline
because
we
didn't
want
to
spin
up
a
whole
humongous
thing
by
default.
Yeah.
A
So,
by
default,
this
now
does
what
provisional
does
right
since
pg
scale
configuration
and
okay.
C
And
I'm
gonna
throw
out
another
question
here
this
this
is.
This
is
for
you
is
the
person
using
this.
Do
you
feel
like
having
the
documentation
in
the
script
directory
like
you
go
into
deploy
and,
as
you
read
me
there
with
deploy,
does
that
make
sense
to
you,
or
would
you
rather
see
that
in
the
doc
directory
and
had
everything
kind
of
living
around
in
the
doc
directory?
Is
this.
A
C
A
I
don't
mind
it
living
in
another
directory,
but
dogs.
It
should
be
linked
somewhere
from
dog,
because
I
will
be
looking
first
at
the
doctor
and
I
expect
it
to
have
links
to
wherever
else
I
need
to
go.
So
if
something
is
written
in
scripts
doc,
something
it
should
be
linked
from
the
parent
doc
directory.
That's
that's
how
I
I
personally
expect
stuff
to
be
because
I
might
not
even
bother
to
check
the
project
directory
structure
and
see
where
all
will
be
documentation.
C
They
are
cross-linked,
but
I
don't
think
it's
necessarily,
I
think
discoverability.
This
is
one
of
those
things
in
the
project.
Everything
almost
everything
is
written
down.
It's
like
95
coverage,
but
it's
about
as
clear
as
diving,
head
first
into
a
mud,
puddle
sure
it's
all
dirt,
but
which
one
do
you
want
like
so.
A
B
A
With
that
in
this
is
kickstart.
B
A
B
C
A
F
A
B
B
B
C
C
E
A
A
C
Yes,
because
prime
oil
remember
what
that
tag
actually
means.
I
think
yes,
because
everything
gets
its
primary,
it's
a
tag
to
differentiate
later,
and
so
it
doesn't
get
turned
off.
When
you
do
it,
it
doesn't
get
turned
off,
because
if
you
add
the
tag
later,
you'll
destroy
your
primaries.
If
you
add
geo
after
the
fact,
you'll
have
to
destroy
the
primary
cluster
to
redeploy
the
secondary.
A
B
C
There
you
go
gx,
I'm
going
to
ask
questions.
C
C
C
What
is
the,
what
is
the
tip
of
the
branch
for
this?
What's.