►
From YouTube: Platform Sync: 2021-01-13
Description
Meeting notes: https://bit.ly/38pal2Z
A
B
I
can
throw
in
at
least
the
status
of
the
windows
ci
stuff.
The
one
of
the
runners
was
failing
for
the
last
couple
days
on
pack
or
maybe
less
than
that,
but
david
had
reached
out,
and
I
think
we
got
it
sorted
out,
there's
a
pr
that
it
still
is
going
to
require
a
little
bit
more
looking
into
because
there's
a
test
that
seems
to
be
consistently
failing
now
around
network,
so
I
think
david's
going
to
go
in
and
recreate
the
runner.
B
A
Cool,
I
think
I
could
kind
of
lead
into
my
status
update,
so
I
have
two
going
into
the
pack
release
so
paco160
that
was
released
maybe
30
minutes
ago,
or
something
like
that.
So
that
is
out.
It
fixes
a
lot
of
issues
related
to
build
pack
uris
so
which
funny
enough
cora
had
just
gotten
into
our
slack.
You
know
reporting
an
issue
and
we're
like.
A
Oh,
we
just
released
it
here
and
you
know
it's
fixed
right,
so
that
worked
out
fairly
well
as
far
as
timing
goes,
so
that
is
out
there
outside
of
that
I've
been
focusing
on
techton.
I've
got
the
techdon
integrations
test
working.
So
this
right
now
enables
us
to
test
the
tecton
tasks
that
we
have
in
the
catalog
with
kind.
A
What
is
it
gke
and
openshift
openshift?
Definitely
being
the
more
you
know
the
the
most
different
platform,
I
guess
I
should
say,
and
so
that
has
a
couple
little
additional
changes
necessary
there
and
probably
would
wanna
be
more
focused
on
the
testing
around
openshift,
since
that
seems
to
have
a
pretty
good
number
of
user
base.
A
A
There'll
be
a
couple
hurdles
there,
so
I'll
probably
spend
a
week
or
so
and
then
the
other
thing
is
taking
the
tacton
issues
and
kind
of
migrating
them
over
into
our
repository
and
kind
of
considering
our
repository
like
the
development
repository
for
the
tasks
as
opposed
to
the
catalog
itself.
So
I
think
that
will
help
us
organize
a
lot
of
the
efforts
there.
C
Yeah
sure
so
I
guess
this
might
be
something
that
we
want
to
jump
into
a
little
bit
later
on
in
this
meeting,
but
I've
been
working
off
and
on
for
a
little
while
on
this
dive
tool
and
trying
to
get
it
to
a
place
where
we
can
start
pushing
features
to
it,
so
that
it'll
help
us
be
able
to
like
visualize
or
explore
builders
and
app
images,
etc.
That
we
produce
and
kind
of,
I
guess
this
is
this-
is
like
ended
up
being
a
pretty
large
amount
of
work.
C
A
Sounds
good
moving
on
so
we
talked
about
the
release,
planning
paco160
is
out,
and
so
we'll
start
our
next
iteration
for
o17.
Oh
there's
a
milestone
been
created
there
already,
and
so
there's
a
couple
issues
already
tagged
to
that.
So
we'll
keep
an
eye
on
that
one
all
right!
Next,
we
have
a
standing
item
for
needs.
Discussions
of
pac,
I
think,
for
the
purpose
of
this
meeting.
A
A
C
C
C
C
C
A
So
if
I
were
to
think
about
this
in
phases
right
we're
saying
that
this
pr
is
necessary
in
order
for
pac
to
reuse,
a
lot
of
the
logic
that
dive
has
in
order
to
provide
that
same
functionality
that
dive
does,
but
with
a
build
pack
specific,
you
know
notion
around
it.
Is
that
right,
that's
like
phase
one
yeah,
and
then
I
think
you
know
if
we
talked
about
it
earlier.
A
The
other
phase,
or
maybe
the
you
know
like
a
side
goal
right,
is
to
enable
dive
to
be
able
to
be
like
extensible
in
a
way
where
it
could
have
information
from
third
parties,
such
as
build
packs
and
be
able
to
display
that
information.
Is
that
oh
correct?
Okay,
so
that
would
be
like
a
phase
two
or
again,
maybe
a
side
quest,
cool.
B
A
C
Yeah,
so
I
guess
I
kind
of
wanted
to
bring
this
up
to
get
a
little
bit
of
feedback
on
where
the
like
direction
that
we
want
to
get
to
like
this
fork
in
the
road
is
like
the
first
thing.
We
want
to
do
like
pull
this
stuff
into
pack
and
use
it
or
because
I
think
that
probably
it'd
be
pretty
easy
to
get
changes
accepted.
That
would
let
us
visualize
our
images
more
easily
just
in
this
tool,
so.
A
C
Yeah,
so
I
I
mean
I've
talked
with
this
guy
andrew
goodman
a
bit
and
he
seemed
like
seems
very
receptive
to
this
as
like
something
that
he
needs
to
do
a
to
like
change
his
ui
libraries,
because
they're
no
longer
supported.
So
that's
good.
The
level
of
effort
left
is
I'd,
say
I
feel
like
I'm
at
the
90
point
right
where
the
last
couple
things
are
around
coloring
and
some
initial
configuration
that
you
can
set
up
in
these
dive.
A
A
C
A
A
Yeah
right,
how
does
that
work?
So
I,
like
I
said
openshift,
is
probably
the
most
difficult
one
kind.
It
works
on
most
ci's
with
no
no
problem,
gke
very
similar
with
tech.
Sorry
terraform,
you
could
stand
up
and
tear
down
gke
instances
pretty
easily
the
open
shift
that
one
right
now
at
least
the
way
I
have
it
set
up
is
using
something
called
code,
ready,
containers
right,
and
so
it's
essentially
like
a
the
full-fledged
vm.
A
That
then
runs
open
shift
and
it's
like
essentially
like
easy
to
set
up
right
and
it
works
really
well
for
local
development.
I
don't
know
how
feasible
that
is
to
get
it
in
a
ci
environment,
the
installation,
time
startup
time
and
all
that
is
like
pretty
it
crazy.
It's
like
30
minutes,
or
something
like
that.
So
I
I
there
is
something
called
I
forgot
like
mini.
Something
mini
shift,
I
believe,
is
what
it
is
and
I
haven't
looked
too
deep
into
it.
A
But
my
hope
is:
it's
a
lot
like
a
mini
cube
where
it's
a
very
bear
down
version
of
openshift
and
it
still
provides
a
lot
of
the
same
functionality
in
which
it
can
be
tested.
So
that's
probably
going
to
be
my
effort
in
that
github
action.
It's
like
exploring
a
more
lightweight
version
of
that.
If
I'm
not
able
to
do
that,
then
obviously
I'd
have
to
look
for.
B
Yeah,
I
think
so.
I
was
because
david-
and
I
were
talking
this
morning
about
what
to
do
for
windows
if
it's
worthwhile
to
yeah
just
having,
I
guess
more
data
points
to
think
of
how
all
the
runners
will
be
laid
out,
helps
us
think
about
what
we're
going
to
do
for
windows,
we're
still
kind
of
stuck
in
in
the
in
the
lcao
tests
of
having
a
bespoke
runner.
B
C
B
Yeah
and
we've
been
using
the
windows
runners
a
lot
on
other
projects
and
they
they
seem
really
nice
and
they're,
not
as
fast
as
the
like
packet
bare
metal
ones,
but
the
state
management's
really
nice
with
it
the
debugging,
it's
not
great,
to
not
be
able
to
like
hop
on
and
debug
it,
but
you
can
make
a
vm
that's
very
similar
to
it
pretty
easily
and
we
put
some
of
the
docs
and
it's
actually
already
in
that
pack
wiki
on
how
to
make
a
vm
that's
similar
to
the
windows.
Printer.
A
B
These
ones
are
bare
metal
or
not
bare
metal,
they're,
just
plain
old
templated,
gcp,
vms
or
azure
vms.
We
did
both
flavors
and
then
some,
like
probably
30
lines
of
powershell
to
run
on
each
one
to
make
it
ready
to
run
pack
so
nothing
too
sophisticated,
and
I
think
all
that
could
be
automated
if
we
wanted
to
go
to
a
packer
image
or
something,
but
so
far
it's
pretty
pretty
boring.
B
We
had
also-
and
we
mentioned
about
like
kubernetes
testing-
is
interesting
too.
I
did
figure
out
recently
that
you
can
run
a
container
on
a
kubernetes
that
exposes
the
docker
socket
over
tcp
and
use
that
as
a
docker
demon.
If
you
wanted,
you
have
to
do
some
things
to
make
it
secure,
but
so
there's
a
few
different
like
creative
ways
that
you
can
avoid
having
to
make
a
custom
vm
there's
some
like
pretty
close
to
already
templated
vms
that
are
out
there.
B
A
I
I
totally
I
can
see
that
and
yeah.
I
think
if
our
stance,
or
or
at
least
guiding
principle,
is
to
minimize
runners,
I
think
that
would
be
beneficial
because
again,
if
we,
if
we
start
creating
runners,
I
think
we
could
end
up
in
a
really
bad
state
where
it
takes
a
lot
more
effort
to
manage
than
it's
worth.
Yeah.
B
B
One
of
those
runners,
as
far
as
I
could
tell
the
the
bin
format,
like
the
the
the
fun
trick
where
you
can
build
arm
images
on
a
air
x86
machine.
I
think
that
only
works
for
docker
desktop,
but
potentially
you
could
just
have
one
custom
runner
that
runs
docker,
desktop
and
use
that
for
stuff,
but
I
think,
obviously
there's
a
bigger
pieces
to
the
arm
puzzle
to
solve.
Besides
that,
so
I
don't
want
to
throw
that
off.
A
B
A
Yeah
yeah-
let's,
let's
hold
off
on
that
grenade
here
for
a
little
bit
and
wait
till
we
get
a
little
bit
more
focus
on
that
arm,
even
though
I
believe
this
is
totally
like
on
the
side,
but
the
leadership
team
core
team,
I
guess,
is
looking
to
create
a
roadmap
for
this
year.
I
do
have
a
feeling
that
arm
is
gonna,
be
on
there.
Just
given
how
prevalent
it's
been
in
the
ecosystem,.