►
From YouTube: OKD Working Group 2020 02 18 Full Meeting Recording
Description
OKD Working Group 2020 02 18 Full Meeting Recording
co-chairs: Diane Mueller, Christian Glombek
https://github.com/openshift/community/projects/1
A
Rinku
all
right,
everybody
welcome,
we'll
give
everybody
a
few
minutes
to
join.
This
is
the
okk
ad
working
group.
If
you
are
here
for
any
other
working
group
meeting
I
apologize,
please
check
your
links
a
bit
of
a
collision
with
the
operator
framework
in
the
past,
but
I
think
we've
cleared
that
up
so
again
in
the
chat.
There
is
a
link,
please
add
your
name
into
the
the
attendee
list
that
waiting
the
many
meetings
going
on
all
right.
So,
let's
start
driving
today.
I
know
we
have
Christian
here
today,
but
Vadim
is
out
sick.
A
So
just
going
to
catch
up
with
a
few
people
and
see
where
we're
at
in
terms
of
events,
we
have
still
coming
up
again.
The
coop
con
EU
event
on
March,
30th
and
Christian
will
have
I
think
15
to
20
minutes
to
talk
about
ok,
D
and
Fedora
core
OS
there,
depending
on
how
tightly
packed
I
put
that
schedule
and
the
same
will
happen
again,
but
a
little
bit
extended
I
think
you
get
25
minutes
at
Red,
Hat
summit.
The
other
thing
is:
if
any
of
you
are
coming
to
Red
Hat
summit.
A
B
B
B
There
is
the
branching,
the
4.5
branching
happening
hopefully
later
today
and
after
that
we
will
probably
have
most
most
of
these
CIA
platforms
or
most
of
the
platforms
working
again
on
CI.
That
there's
been
some
hiccups.
We've
added
an
an
Etsy
D
operator
and
the
switch
over
to
that
has
caused
some
some
failures
in
the
in
the
past
few
weeks.
But
we
I
think
we
sorted
that
out
and
the
switchover
is
supposed
to
happen
tonight.
B
We've
been
working
on
yeah
get
getting
all
the
the
blockers
done.
There's
been
some
progress
on
on
the
as
your
fixed,
disc
and
upload
a
problem
so
hopefully
very
soon,
we'll
have
we'll
have
discs
or
we'll
have
images
on
on
Azure
that'll
that'll
be
nice
for
the
automated
installer
driven,
install
yeah.
There's.
Nothing
really
big
right
now
to
show
off
I
really
do
hope
by
the
next
meeting.
We'll
have
much
more
to
show
after
the
4.5
branching
happened
and
we'll
we'll
still
be
building
off
master,
but
it'll
be
the
4.5
master.
At
that
time.
B
B
D
B
D
B
That
was
probably
Vadim
and
I
still
don't
have
an
answer
to
that,
but
we'll
definitely
because
bare
metal
isn't
really
our
focus
right
now,
but
it's
done
of
the
next
big
thing
we
want
to
tackle.
After
all,
the
you
know
all
the
cloud
platforms
we
we
sort
of
tried
to
support
in
the
product
work
in
okd
as
well.
We
want
to
tackle
that
yeah
and
because
we
have
some
some
downstream
images
for
the
cluster
I
have
funny
bare-metal
things.
I
will
need
to
make
sure
that
they
that
we
get
them
from
upstream
I.
B
Think
it's
we're.
Taking
some
images
from
our
do.
They're
from
the
OpenStack
I'm,
not
100%,
sure
another
question
I
will
follow
up
on
bare
metal
is
definitely
one
thing
we
want
to
tackle
very
soon
now
the
difficulty
sort
of
for
me
with
that
is
it's
it's
running
on
a
different
CI
system.
So
I
don't
really
see
that
a
lot
but
yeah
I
will
follow
up
on
that.
Thanks
for
taking
notes,
I
am.
A
B
D
C
B
A
So
what
the
other
thing
that
I
would
offer
cuz
dado
is
also
a
member
of
open
shift.
Commons
the
openshift
Commons
member
booth
is
going
to
be
in
community
central
as
well.
So
if
you
wanted
to
have
like
a
half
an
hour
of
time
to
give
a
presentation
in
the
booths
there
we'd
love
to
have
you
do
something
kneel
there
and
it's
really
random.
A
C
A
C
E
A
Right
I'm,
just
putting
your
name
down
here
and
just
for
you
cool
all
right.
All
the
lurkers
are
gonna
get
outed.
Today
there
you
go.
Let
me
just
go
back
to
this.
So
are
there
other
things
that
people
have
questions
about?
I
know,
there's
a
lot
of
you
on
there
and
just
noticed
the
UMich
guy
are
at
the
University
of
Michigan.
Are
you
guys
testing?
Okay
D
now
are
on
VMware
yeah.
F
B
So
the
way
we
test
UPI
in
in
our
CI
infrastructure,
you
can
actually
find
you
can
find
docker
images
in
in
those
platform,
specific
folders,
I,
think
docker
file,
GPI
or
something
they're
called,
though
that's
mostly
how
we
test
that.
So
we
download
you
know
matchbox
and
all
the
the
UPI
specific
things
in
in
those
darker
yeah
in
those
docker
images
and
then
run
those
in
an
automated
manner.
B
B
F
Provide
the
ability
of
other
folks
to
have
an
easy
way
to
get
UPI
provisioned
anyone
else
who
wanted
to
test
right
so
providing
scripts
that
would
allow
folks
to
get
there.
Okay,
D
cluster
up
quickly
without
having
to
manually
type
in
each
step
as
it
were,
you
know,
or
run
each
individual,
gutsy,
command,
etc,
so
being
able
to
provide
that
to
the
greater
community.
F
F
All
of
the
slides
for
the
roadmap
and
it
actually,
these
fears
listed
in
the
UVI
like
for
versions
up
to
like
four
or
five
and
later
at
least
the
roadmap,
as
it
was
published
like
three
months
ago
or
so
that
I
saw
a
slide
for
it.
So
that's
what
that's
what's
fueling!
This
is
the
notion
that
my
understanding
is
that
it's
going
to
be
UPI
for
the
foreseeable
future,
so.
C
The
development
of
IP
i4
v
fear
is
that
the
interface
library
is
4c
and
go
are
not
free.
Software
they're,
not
open-source,
bundling
them
and
including
them
in
the
Installer,
is
a
no-go
go
VMI
and
a
couple
of
the
other
libraries
are
proprietary
and
they
don't
have
source
code
available.
So
building
and
shipping
them
is
going
to
be
problematic.
C
Pyv
ami
actually
partially
relies
on
the
proprietary
c
libraries,
which
is
a
problem
in
itself.
There
is
a
version
of
it
that
doesn't
use
the
c
library
and
that's
for
python
and
that
kind
of
works,
but
it's
far
more
limited
for
doing
all
the
really
crazy
stuff
with
vSphere
you.
You
really
need
the
full
sdk.
C
A
A
Ok,
so,
but
I
think
I
want
to
go
back
to
Jamie's
original
suggestion
is
that
is
there
any
of
any
reason
why
we
shouldn't
create
a
folder
in
the
repo
to
hold
some
of
these,
and
you
know
community
installation
scripts,
for
people
to
use
during
the
testing
period,
with
some
sort
of
general
surgeons
warning
on
them,
or
is
that
just
fixed
for
work
and
I?
Guess
the
questions
for
Christian
is
there?
It
is
oh.
B
A
B
I
I
wouldn't
say
not
origin,
maybe
in
the
okp
repository
we
already
have
what
we
we
have
the
documentation
in
the
okd
repository
as
well.
So
if
that's
kind
of
ripped
yeah,
maybe
in
the
okd
repository
or
maybe
we'll
have
to
like
I'm,
not
sure
if,
if
Clayton
wants
those
scripts
in
there,
but
if
he
doesn't,
we
might
just
create
another
repository
for
that.
So
I'll
follow
up
on
the
best
I,
don't
think
the
community
report
well,
it
could
could
be
in
the
community
and
repository
as
well.
I'm,
not
yeah
I.
F
A
A
G
B
H
B
A
J
So
I
built
the
center
s
core
OS,
just
on
my
own
I,
basically
rebuilt
the
OpenShift
rpms
and
then
on
top
of
that
I
also
rebuilt.
The
latest
ignition
I
haven't
actually
attempted
to
install
okd
on
top
of
it,
but
my
thought
is:
if
I
can
make
that
work,
then
that
would
give
the
archives
folks
a
nice
pass
to
switch
to
ignition
v3
I,
don't
know
what
the
hang-up
is
there,
but
at
least
that
might
prove
that
it
could
work.
I
Blatant
here
so
at
this
point
there
is
no
clear
picture
of
whether
we
will
have
st.
OS
chorus,
but
regardless
your
work
is
remarkable
and
my
I
might
have
a
place
to
to
put
it
while
st.
OS
figures
out
what's
supposed
to
happen
with
Colo
it.
So
look
at
me
on
the
kubernetes
is
like
I,
usually
hang
out
on
the
opposite:
dev
Channel!
Don't
look
for
me
over
there
and
let's,
let's
sync,
our
clocks
and
work
out
what
we
need
to
do
to
get
almost
official:
okay,
D
own
st.
OS,
chorus,
okay,.
C
I
mean
I'm
a
little
waffle
e
about
the
thought
of
having
having
a
sin
to
ascore.
Os
I
mean
aside
from
the
fact
that
pretty
much
every
time,
CentOS
does
something
like
this.
All
the
oxygen
gets
sucked
out
of
the
fedora
based
effort,
which
does
not
make
me
a
very
happy
camper
in
the
first
place,
but
also
because
CentOS
is
a
bad
place
to
try
to
develop
new
technology,
build
on
top
of
new
technologies,
which
is
part
of
what
I
hope.
C
Okay
Dee's
mission
is,
rather
than
just
playing
catch-up
to
OCD,
because
that
eventually
being
the
driver
of
new
container,
related
technology
usage
and
development
and
Santos
is
the
wrong
place
for
that,
because
in
most
cases,
rel
doesn't
get
anything
until
Fedora
proves
it
out.
First
and
CentOS
is
not
going
to
be
a
place
where
that
kind
of
stuff
will
happen.
C
Unless
it's
already
happened
in
Fedora
and
then
the
catch-22
is
well,
then
does
that
mean
we
have
to
wait
three
plus
years
before
the
next
Fedora
branching
before
we
start
even
getting
to
play
with
these
things
in
in
CentOS?
Well,
because
nobody's
testing
them
I
mean
a
reasonable
fashion
or
using
them
in
any
production
like
quality
with
Fedora
as
the
base
there.
Like.
A
So
well,
this
is
what
I
think
about
that.
I
think
it's
amazing
that
James
built
this
thing
I,
just
when
I
look
at
the
scent
o
s,
resources
that
are
assigned
to
working
on
anything
scent,
OS
related
they're,
pretty
much
non-existent
people
who
other
than
James
doing
this
out
of
the
amazement
of,
in
goodness
of
your
heart
in
your
spare
time
and
and
I
look
at
the
Fedora
one
and
that's
where
we're
putting
all
where
all
the
resources
are
that
are
working
on
this,
and
it
is
the
driver
of
the
new
technology
development.
A
It's
where
we're
doing
ignition
3
work,
it's
almost
like
Center
West
chorus
and
is
downstream
from
Fedora
core
OS
in
a
weird
way:
I,
don't
I,
understand.
Historically,
the
threat
that
sent
OS
might
have
been
to
fedora,
but
I
think
we've
we've
come
we're
coming
down
a
different
path
with
Fedora,
core
OS
and
ok
D
this
time
around,
but
and
and
that's
in
my
commitment
is
to
can
is
almost
fedora
first
and
everything
else
is
fine.
A
I
A
I
I
OS
coral
s,
we
will
know
whether
ok
D
works
or
not,
and
what
are
the
Missy
points
so
until
you
say,
let's
figure
out
a
way
to
test
and
relieve
ok
D
using
Korres
these
walk
around
the
chains
create,
it
might
be.
You
know
something
does
to
let
us
know
if
pursuing
for
an
official
site,
OS
Coretta,
stop
being
bored
or
not.
C
Again,
I'm
fine,
with
the
existence
of
a
Santos
chorus,
I'm,
just
worried
that
you
know,
president
has
been
each
time.
Something
like
this
comes
up
on
the
oxygen
gets
sucked
out
of
the
equivalent
Fedora
effort,
like
that's
happened
with
OpenStack.
That
happened
with
OpenShift
three
two
and
3x
that
happening
with
over
it
right
now
and
like
I,
really
don't
want
history
to
repeat
itself
again,
because
every
time
that
happens
like
it
means
that
fewer
people
are
working
on
the
on
the
proper
upstream
features
and
helping
integrate
them
in
fedora.
C
They
only
come
in
just
at
the
last
second
to
integrate
these
things
so
that
they're
in
they're
branching
for
rel,
not
like
anyone
else.
The
community
knows
when
the
branching
happens
but
like
when
that
happens,
and
whole
surge
of
random
starts
happening,
and
that's
that's
the
kind
of
stuff
that
I
just
want
to
avoid.
So
that's
why
I'm
stating
that,
because
it's
something
that
is,
it
is
a
fear
for
me
when
it
comes
to
this.
D
Given
that
our
starting
point
here
is
Fedora
core
OS,
the
especially
given
the
tight
coupling
now
between
the
four
dot
X
family
of
openshift
and
the
operating
system,
whereas
with
three
dot
X,
the
operating
system
was
just
a
place
to
run
it
right.
And
so,
if
you
were
running
CentOS
or
you
were
running,
Fedora
was
less
impactful
to
how
you
ran
it.
Whereas
now
they're
much
more
tightly
coupled.
C
A
I'm
glad
you're
worried
about
it
because
it
means
that
we'll
think
about
it
and
be
conscious
of
it.
I.
Don't
think
personally
that
it's
that
that
it's
anything
that's
going
to
stop.
Okay,
Dee
and
Fedora,
core
OS
collaboration
and
and
our
release
cycles
are
going
to
be
in
in
this-
is
Diane's
opinion.
I.
A
The
commitment
that
I
have
and
will
continue
is
that
it's,
the
collaboration
is
between
okd
and
Fedora
core
OS
and
that
all
the
other
OS
is
have
to
bring
resources
to
the
table
like
James
has
done
and
and
Gelson
is
hoping
to
find
an
incubation
space
there.
But
I.
Really.
My
commitment
to
you
all
and
I
think
is-
is
to
really
continue
the
collaboration
tightly
with
the
fedora
fedora
community
and
make
sure
Fedora
core
OS
is
a
success.
B
Just
sort
of
as
a
side
note
that
center
stream
will
be
upstream
from
rel
while
pentose
these
standard
Santos
is
downstream,
which
is
why
it
does
not
make
a
lot
of
sense
for
us
to
put
an
effort
there,
because
changes
will
never
really
trickle
down,
naturally
into
into
anything
else.
It's
of
the
last
yeah
yeah.
The
last
thing,
the
most
downstream
thing
here,
yeah.
B
A
A
A
What
is
what
is
the
thought
on
this
with
this
few
group
of
people
hear
about
this
and
I'm
gonna?
Leave
it
at
that,
because
I'm
trying
to
get
my
messaging
for
when
people
ask
me
the
question
correct
and
I
would
just
like
to
hear
and
I
know
that
a
Christian.
Maybe
you
could
go
first
and
explain
what
you
explained
to
me
earlier
today.
B
Sure
there
was
the
question
on
the
mailing
list
again
last
week
with
regard
to
to
the
one
note
set
of
call:
ok,
d4
and
I
linked
a
an
enhancement
proposal
that
got
merged
that
was
written
by
Clayton
and
I.
Think
that
sort
of
that's,
not
the
stance
we
take
or
the
architects
take
on
it
right
now.
So
it's
a
long-term
goal
to
sort
of
get
there,
but
the
OpenShift
codebase
isn't
really
ready
for
it
at
the
moment,
especially
a
few
of
the
core
operators.
B
Don't
support
just
having
one
one
master
node
I
mean
there's
EDD,
there's
the
Machine
config
Damon.
How
are
they
the
Machine
config
operator
and
all
of
those
don't
support
that?
Yet
so
it's
that
there's
quite
a
few
big
changes
needed
to
support
that
in
the
in
the
code
ready
Pena's
use
case,
it's
sort
of
done
by
disabling
a
few
of
the
operators.
I
think
the
machine
config
operator
goes
degraded
in
in
the
CRC
Nicole
ready
container.
B
We
could
aim
at
creating
an
okd
version
of
the
CRC
I'm,
not
sure
what
what
the,
what
more
value
we
would
get
CRC
already
is,
is
free
and
for
everybody
to
use,
but
yeah
I,
don't
really
have
any
strong
opinions
on
that.
I
know
it's
sort
of
a
gap
we
have,
we
don't
have
anything
like
mini
cube
or
mini
shift
in
the
past
or
okay,
deep
or
and
I.
Don't
really
have
a
good
answer
to
that
couple.
C
K
C
C
But
right
now
there
is
no
seriously
it
doesn't
have
a
no
Katie
variant
right
now.
That's
the
problem
so
right
now,
CRC
requires
no
Red
Hat
account
and
because
it
is
only
going
to
deploy,
OCP
and
actually
I
really
do
think
it'd
be
super
valuable
to
be
able
to
have
in
a
CRC
that
does
earth
cause
and
Oh
Katie,
because
okay
knee
is
moving
further
along.
Then
then
then
OCP
is
and
it'd
be
usable,
useful
to
be
able
to
just
try
new
things
super
easily.
B
B
B
Let's
say
before
may
even
like
starting
developing
it
at
least
not
me,
because
we'll
need
to
get
up
the
the
okay
d4g
a
version
before,
and
then
we
can
sort
of
look
at
creating
in
yeah
a
CRC
version
of
it,
but
that
that's
not
one
of
the
release
criteria
we
have
right
now
and
various
of
those
are
more
pressing
at
the
moment.
But
if
you,
if
you
guys
say
we
we
want
that
will
definitely
tackle
it,
but
it
won't
happen
before
me.
That's
kind
of
the
catch
yeah
I.
C
Think,
that's
fine!
It's
just
it's!
Just
it's
I
think
going
to
be
a
prerequisite
to
make
it
easier
for
broader
consideration
of
the
okd
branch
of
kubernetes
for
more
people,
because
it
is
the
main
like
down
down
point
when
people
talk
about
it's
like
okay.
So
how
do
I?
How
do
I,
just
you
know,
give
it
a
shot?
It's
like!
Well,
do
you
have
a
monster
machine
and
some
magical
magical
things
that
you
can
tell
the
open
ships
installer
to
run
limit
on.
C
A
platinum
credit
credit
card
so
not
having
to
require
all
those
things
to
just
make
it
so
that
people
can
give
it
a
shot
and
evaluate
it
on
on
the
mere
fact
of
its
other
features
that
exist
at
present.
That
is,
that
is
a
super
valuable
thing.
I
mean
yes,
it's
going
to
be
a
memory
hug,
it's
going
to
be
kind
of
a
resource
sink
and
it's
going
to
kind
of
suck
compared
to
just
deploying
it
properly
in
production,
but
I
think
that's
sort
of
the
point.
C
It
helps
people
like
see
what
it's
gonna
be
like
not
without
requiring
it
to
be
a
like,
so
full-fledged
whatever's
I
mean
I
would
love
to
have
single
note,
whatever's
that
makes
CRC
actually
simpler
and
easier,
but
but
the
existing
CRC
I
don't
know
how
difficult
it
would
be
to
adapt
to
okay,
D
but
I
feel
like
it
wouldn't
be
too
terribly
awful,
but
doing
the
reason
the
resources
for
it's
happening
after
GA.
That's
fine,
like
I,
mean
I,
can
kind
of
wait
at
some
point.
C
I
will
get
an
openstack
resource
and
I
will
rebuild
my
over
cluster
and
then
deploy
it
everywhere
and
we'll
see
how
it
goes,
but
like
for
all
the
all
the
other
people
who'd
like
to
try
out
Katie
I,
think
that's
that's
a
really
valuable
thing
to
have,
because
many
Cuban
mini
shift
keep
getting
brought
up.
There
are
people
still
deploy
mini
shift
and
trying
out
open
shift
that
way
and
they're
getting
a
poor
experience,
because
that's
not
what
open
shift
looks
like
at
all
anymore.
A
A
Pause
for
a
minute,
and
just
for
Diane's
education
in
the
very
old
days
open
shift
two
days
in
open
ship.
Well,
we
had
a
VM
that
we
just
sat
up
there
on
the
origin
page.
Is
there
any
possibility
of
just
creating
a
virtual
machine
that
had
it
running
or
is
that
just
gonna
be
too
huge
into
clunky?
It's.
C
Technically
possible
I
mean
the
the
real
the
real
kick
is
making
it
so
that
the
F
cos
machine
on
first
boot
when
it
starts
up,
is
automatically
configured
to
pull
and
set
up.
Okay,
D,
that's
that's
pretty
much
it!
Everything
else
is
mostly
Sympatico.
It's
it's
once
you
get
past
that
point,
everything
gets
pretty
easy.
It's.
D
C
D
Because
it's
the
Installer
that
requires
that,
so,
if
you
built
the
Installer
from
the
okd
source,
you'd
have
to
you'd
have
to
set
it.
You
know
the
tags
for
libvirt
and
then
set
the
release
image
before
you
build
the
Installer
to
point
to
one
of
the
okd
releases.
I
think
that
would
work
I
might
I
might
give
it
a
try
this
weekend,
if
I'm
not
still
working
on
some
Korkis
conversions
from
spring
food,
I.
C
Mean
that's!
That's
our
I'd
love
to
try
that
out
too,
because
I
think
it
actually
should
like
the
current
state
of
affairs
for
okd
right
now.
Is
that
it
once
it's
up
and
running
it
works
pretty
okay,
so
so,
and
the
Installer,
the
okay,
the
Installer
is
mostly
okay
when
it's
setting
things
up,
so
it
might
just
kind
of
work
with
some
small
adaptations.
A
A
If
we
can
supply
that,
that
would
be
great
and
not
just
that's
that's
why
I
figured
I
knew
we
didn't
have
a
full
meeting
today
and
I
figured
I'd,
bring
it
up
again,
just
because
of
that
so
yeah.
If
you
can
figure
out
a
way
to
do
that,
that
would
be
great,
and
but
the
problem
with
that
is
that
every
time
we
do
a
new
release,
we
need
to
be
able
to
rebuild
that
image.
So
some
and
I
know
this
was
the
problem
in
the
old
days
of
open
ship
for
origin.
A
Is
that
the
Evangelist
team
dating
back
to
Christian,
Rama
and
Steve
poo
Steve
would
build
the
VM
for
us
and
we'd
have
to
upload
it
with
each
release
so
that
the
thing
what
I
don't
want
to
do
is
make
make
more
work
for
people,
but
it
also
has
to
be
something:
that's
repeatable,
so
Charro.
If
what
you're
doing
is
a
complete
hack
and
can't
be,
you
know
repeated
for
each
release,
then
maybe
it's
it's
a
nice-to-have
for
Red
Hat
summit.
A
A
D
Actually
I
am
I
will
be
preparing
some
materials
to
teach
it
internally
to
our
Java
development
group.
It
probably
won't
be
until
May
or
June
before
I
do
that,
but
I've
got
what
I
did
is
I
wrote
an
app
in
spring
booth,
that's
representative
of
the
types
of
apps
that
these
folks
were
right
and
then
fork
that
off
in
a
in
a
spring
boot
branch
and
then
converted
it
to
Korkis
so
that
they
can
see
the
same
code
bring
boot
ranch
to
Korkis
branch.
What
the
difference
is
in
the
code
are
yeah.
A
C
C
A
Alright
well
I'll,
keep
you
in
mind
and
I'll
hit
you
up
and
maybe
we'll
get
Marcus
so
join
as
well,
but
I
think
it's
it's
a
I'm,
I'm,
really
thrilled
with
Quercus
and
and
I
think.
If
we
can
do
more
I
think
you
know
it
also
probably
has
a
booth
somewhere
in
probably
on
the
red,
Red
Hat
summit
flora
at
Expo,
that
you
could
do
some
demos
on
this
as
well.
So
this
sort
of
chocolate
mine
is
there
anything
else
Christian
we
should
be
carried
covering
off
today.
A
B
B
B
A
A
B
Yeah,
maybe
one
one
more
thing:
I've
set
up
a
calendar
on
the
Fedora
calendar
thing,
fado
Cal
I'll,
just
quickly
paste
the
link
in
the
chat
in
a
minute
and
I've
set
up
or
just
a
regular
meeting
every
two
weeks.
So
it
should
be
there
forever
on
on
the
calendar.
So
you
can
just
import
that
calendar
via
iCal
and
yeah
and
I've
also
set
up
remind
us
on
slack
on
a
few
in
a
few
channels
to
sort
of
ping
everyone.
A
few
minutes
before
the
meeting
starts.
B
I
think
I
send
it
out
on
various
channels.
I
also
had
a
config
to
send
emails
automatically
reminders
to
the
working
group.
Google
group
and
also
I
think
the
users
and
death
lists,
but
apparently
the
fatal
Cal
thing
doesn't
support
sending
to
multiple
recipients.
So
for
now
it's
just
sending
it
out
to
the
okee
working
group,
Google
Group.