►
From YouTube: OKD Working Group 2020 08 18 Full Meeting Recording
Description
OKD Working Group 2020 08 18 Full Meeting Recording
co-chairs: Diane Mueller and Christian Glombek
B
A
C
A
Nice
yeah,
so
red
hat
had
a
sponsors
keynote
and
we
decided
to
use
it
for
saying
happy
birthday
to
oci,
irvashi,
mohani
and
sally
o'malley
did
the
presentation
and
in
it
they
mentioned
okd
a
couple
of
times.
So
I
I
tweeted
that,
and
that
was
that
was
really
very
nice
to
see
because
we
don't,
we
don't
have
a
lot
of
brand
recognition
out
there
shall
we
say
we're
kind
of
overshadowed
by
our
big
brother
openshift.
A
D
A
Well,
the
thing
I
think
the
cheaters
way
just
the
back
door
and
I'm
recording
this
is,
if
you
are
a
member
of
the
cloud
native
foundation
slack
if
you're
there.
If
you
go
in
to
the
slack
channels,
you'll
see
number
one
number
two
number
three
and
all
of
these
other
slack
channels
that
are
set
up
specifically
for
kubecon
and,
to
be
quite
honest,
I
think
the
more
interesting
thing
today,
there's
chat
through
the
entrado
platform
hosted
thing
which.
D
A
Really
hard
to
figure
out
and
disappears
after
the
event,
and
then
the
cloud
native
foundation
has
set
up
for
all
the
kukan
different
tracks,
and
so
I'm
hanging
out
in
slack.
I
have
one
window
open
for
the
booth
because
I'm
on
booth
duty,
but
I
would
say
that
the
slack
is
really
where
the
networking
is
going
on.
D
A
E
A
D
A
Wrong
by
the
end
of
the
day,
I
I
actually
literally
had
almost
lost
my
voice
because
I
was
maybe
not
talking
as
much
as
some
of
the
presenters,
but
it
was.
I
first
used
my
voice
after
vacation,
so
I
wanted
to
say
again
on
record.
Thank
you
to
christian
and
charo
and
every
one
of
the
working
group
members
who
presented
it
was,
I
know,
a
big
effort
to
get
the
resources
to
do
that
charo.
A
A
You
really
can't-
and
I
I
need-
and
I
know
the
open
stackers
aren't
probably
here
today,
but
it
was
wonderful
to
close
on
the
openstack
one
and
have
that
private
cloud
full
open
source
stack
as
the
closer,
so
I'm
gonna
start
making
a
bigger
deal
and
try
and
get
some
of
the
open
stackers
to
come
more
often
to
these
things.
But
that
was
for
me
that
was
really
good,
because
the
yeah
bringing
that
community
in
too
was
very,
very
very
nice.
A
What
did
I
say,
462
viewers,
so
that's
people
coming
in
at
different
times
whether
it
was
from
the
tweets
looking
for
a
few
minutes-
and
I
don't
have
the
youtube
live
stuff
because
youtube
unbelievably,
because
with
google
compute
you
think
they'd
be
faster,
but
their
analytics
page
only
shows
the
number
of
people
who
engaged-
and
I
think
146
people
engaged,
but
it's
kind
of
wonky
and
the
viewer
number
is
usually
bigger
on
youtube
than
either
the
other
two
platforms.
But
we
won't
know
for
a
few
days.
A
So
it's
gonna
take
me
a
little
bit
of
time
to
edit
all
those
videos.
I
thought
I
could
turn
them
around
last
night,
but
it
didn't
happen
so
and
kubecon
is
happening
today
and
tomorrow.
So
I'm
going
to
try
and
get
all
of
the
videos
done
and
out
the
door,
at
least
on
the
playlist
by
the
end
of
day
tomorrow,
that'll
be
pushing
it.
But
that's
that's
my
goal.
I
also
I'm
going
to
get
dental
work
done
tomorrow
afternoon.
So
you
know
you'll
see
how
that
all
works
out.
So
I
am.
A
I
can't
say
enough
that
those
kinds
of
events
are
huge
for
us
building
the
brand
of
okd.
I
can't
whether
I
said
this
before
I
hit
the
record
button
on
before.
We
also
got
a
shout
out
this
morning
on
the
keynote
stage
at
kubecon,
the
red
hat,
sponsored
speech,
which
we
never
really
do
to
talk
about
ourselves
or
red
hat
products.
A
So
that's
the
work
that
we
normally
do
and
also
starting
to
engage
with
the
public
in
terms
of
people
who
are
using
okd,
and
so
one
of
the
folks
who
was
on
the
call
most
of
the
day
yesterday
was
a
gentleman
from
redfield,
and
I
think
neil-
and
I
were
talking
about
this.
A
little
bit
in
the
chat
red
pill
is,
is
migrating
from
311
okd
to
4..
So
pepin
is
a
great
use
case
for
okay
and
production.
A
E
A
A
So
it's
engaging
and
doing
things
like
yesterday
help
us
get
these
people
out
of
the
closet,
get
them
to
give
us
feedback
and
enable
them
to
do
things.
You
know
enable
us
to
know
what
they're
doing
with
okd
and
get
that
feedback
into
openshift
into
all
the
other
things
as
well.
So
that's
you
know,
that's
sort
of
my
quick
and
dirty
update
on
what's
going
on
or
what
went
on
yesterday.
A
B
E
The
container
stick,
I
believe
that
he
wants
to
set
up
some
kind
of
dedicated
workshop
to
like
kick-start
the
whole
thing,
and
so
we
should
figure
that
out
and
get
a
date
set
up
and
maybe
take
advantage
of
the
fact
that
fedor
now
has
this
virtual
platform
thing.
So
we
can
all
like
just
jump
in
and
start
figuring
that
stuff
out.
D
A
A
B
A
What
I
would
suggest
is
that
maybe
we
do
something
like
a
joint
meeting,
ukd
working
group
in
the
container
working
group
and
maybe
use
the
next
two
weeks
from
now
time
to
do
something
and
then
just
give
him
a
half
an
hour
to
talk
about
what
they,
what
their,
how
they
want
to
kick
start
it
or
kick
start.
It
then,
and
do
something.
B
So
I
think
he's
really
trying
to
set
up
a
two
hour
workshop.
That
kind
of
does
work
walk
through
and
then
also
like
a
hacking
session
to
get
things
started,
because
I
don't
think
there's
too
many
active
members
of
the
container
stick
right
now,
that's
clemont,
who
set
that
up.
I
joined
in
from
the
beginning,
and
I
I'm
not
sure
neil
you
were
probably
there
as
well.
A
E
Yeah,
the
the
main
reason
that
the
containers
take
died
off
was
the
the
openshift
origin
community
that
essentially
supported
it
in
the
beginning,
stopped
caring
and
openshift
stopped
being
buildable
and
runnable
on
fedora
and
a
lot
of
the
infrastructure
that
was
being
built
around.
It
was
actually
developed
by
the
openshift
team
and
those
all
rotted
away,
and
so
it
got
to
a
point
where
nothing
worked
and
there
was
nobody
there
to
help
because
nobody
cared,
and
so
naturally
the
sick
just
fell
apart.
A
B
There's
actually
the
issue,
I
just
pasted
it
in
the
chat
about
the
next
meeting
and
somebody
posted
a
when
is
good,
which
I
think
is
probably
like
a
doodle
to
find
the
to
find
a
good
time
for
everybody.
So
you
could
probably
just
enter
yeah
whenever
you
can
in
there
and
then
we'll,
hopefully
find
a
good
time
for.
A
A
A
F
F
A
If
we
can
use
the
next
time
to
go
through
those
and
see
where
we
need
to
reach
out
to
people,
because
the
one
thing
that
I'm
a
little
concerned
about
is
because
their
operators
come
and
go
and
we
are
going
to
get
more
and
more
of
them
is
that,
if
we're
holding
the
bag,
excuse
me
for
maintaining
the
community
versions
for
okd,
and
we
don't
have
a
connection
to
that.
We're
going
to
be
in
trouble.
B
So
it's
essentially
what
at
least
for
the
enhancement
proposal,
what
I
want
to
make
sure
that
all
the
at
least
the
red
hat
made
operators
that
all
of
them
get
released
for
okd
by
their
teams.
So
we,
as
a
working
group,
don't
have
that
additional
work,
because
we
we
won't
be
able
to
do
it.
We
won't
be
able
to
go
into
every
operator
and
fix
it
up
and
then
also
maintain
that
for
all
releases,
that's
the
thing
the
teams
have
to
do
themselves.
They
it's
essentially
like
packaging.
E
I
read
makes
me
more
and
more
confused
on
how
to
do
it
and
it's
it's
just
I
I
feel
lost
in
the
sea
of
of
buzz,
speak
and
confusion
and
technical
confusion,
and
that's
not
helpful,
and
I,
in
order
to
help
better
support
having
a
larger
community
of
operators.
We
should
make
sure
that
that
is
something
that
people
can
feel
self-sufficient
in
learning.
How
to
do
this.
A
D
A
B
Those
would
be
the
ones
you
you
may
be
wanting
to
approach
for
getting
operators
on
to
the
platform,
while
okay
working
we're
really
just
working
on
the
core
operators,
which
are
part
of
okay
of
open
shift.
So
anything
else,
it's
like
a
feature.
We
will
not
support
from
our
side,
because
it's
just
not
our
realm
here.
Obviously
we
want
people
to
use
that,
but
yeah
we're
really
focused
on
the
core.
E
We
don't
have
a
specific
group
for
doing
that
and
the
operator
framework
if
it's
going
to
build
a
community
if
you
want
to
like
feed
a
community
development
and
interest
in
this
particular
pattern
of
running
services
and
infrastructure
on
on
kubernetes
platforms,
that's
probably
a
place
to
start,
and
so,
if
there
is
going
to
be
some
kind
of
community
there,
maybe
a
sig
that's
connected
to
fedora.
To
do
that
would
help
make
it
so
that
it's
a
lot
more
accessible.
F
F
A
I
don't
know,
I'm
not
positive,
you
need
a
fedora
operator,
sig
the
main
cause.
Let
me
just
finish:
okay,
red
hat
has
donated
the
operator
framework.
The
sdk
and
the
olm
side
of
things
is
all
part
of
the
cncf
and
is
an
incubated
project,
so
there
will
be
sigs
over
there
for
fedora
people
to
join
and
they
can
create
a
subset
there
if
they
need
to
rather
than
a
fedora
operator
thing-
and
I
I
know
that's
not
normal
mode,
but
that
would
be
to
me
the
natural
place.
A
To
make
requests
for
more.
You
know
more
documentation
and
to
donate
that
and
to
build
awareness
for
fedora
in
the
kubernetes
world
would
be
a
better
way
for
doing
community
development
and
doing
some
collaboration
in
that
space,
rather
than
doing
something
separately
that
might
die
on
its
own
if
it
doesn't
have
care
and
feeding
from
the
operator
framework
community.
E
E
The
thing
is
they
want
those
to
move
to
being
things
that
run
on
top
of
openshift
for
good
or
bad.
There
are
no
patterns,
there's
no
procedures.
There's
no
practices,
there's
nothing
for
how
to
do
that.
If
you
want
people
to
be
able
to
do
that,
you
need
it's
not
just
that
there
has
to
be
a
community
upstream,
there's,
always
a
community
upstream
for
all
the
crap.
I
honestly
doesn't
matter.
The
problem
is
that
you
also
need
you
need
communities
where
people
are
actually
doing
work
who
are
actually
able
to
take
advantage
of
those
technologies.
E
That's
the
reason
why
linux
distributions,
while
they
like
open
susan
fedora,
they
both
use
rpm.
They
both
have
their
own
packaging
guidelines.
They
all
do
their
own
packaging
stuff,
but
there's
also
the
rpm
upstream
community,
where
things
that
do
make
sense
to
put
together
there
happen
there.
From
my
perspective,
I
see
operators
in
the
same
manner.
E
They
it's
essentially
the
foundational
aspect
for
delivering
applications
and
services
on
a
kubernetes
for
lack
of
a
better
phrase,
operating
system
platform
right
and
so
er
different
communities
are
going
to
have
different
patterns,
different
styles,
different
guidelines,
different
tools
and
like
when
I
talk
about
with
the
container
sig
and
having
people
be
able
to
make
operators
using
those.
The
idea
is
that
you
take
fedora
content
and
you
build
upon
it
upon
it
and
be
able
to
make
operators
and
stuff
and
they
would
be
pushed
into
the
federal
registry.
E
They
could
even
be
pushed
out
to
operator
hub.
That,
probably
is
a
good
thing
to
do
and
there
should
be
a
liaison
between
the
fedora
side
and
the
and
the
upstream
part,
I'm
not
disagreeing
with
that
part.
But
the
problem
is
that
without
without
a
way
to
distribute
that
community
load,
you
have
two
potential
problems.
E
So
you
wind
up
having
you
know,
you
have
an
operator,
you
have
a
you,
have
a
release,
sig,
which
then
has
a
subset,
for
you
know
xyz
thing
and
those
people,
just
it
kind
of
fans
out
over
time
and
where
that
fans
out
doesn't
matter
if
it's
on
cncf
side
or
the
distro
side
or
somewhere
else
altogether.
What
I'm
saying
is
that,
because
we
have
all
this,
we
have
this
nucleus
of
people
that
exist
in
the
fedora
community.
To
do
these
sorts
of
things
we
should
have
them.
B
E
B
And
I
have
it's
kind
of
like
software
packaging,
for
linux,
distros,
right
and
we'll
we'll
have
to
make
sure
that
fedora
packages
its
own
operators,
because
we
just
can't
consume
or
we
don't
want
to
consume
operators
that
aren't
built
with
our
guidelines
elsewhere.
They
may
work,
but
they're,
not
you
know
not
specifically
made
for
the
environment,
which
is
the
reason
why
I
want
to
get
the
red
hat
teams
to
specifically
also
release
for
fedora
core
os
based
operators
and
then
all
the
other
operators
teams
that
we
don't
really
have.
B
I
think
it's
easiest
if
it
comes
from
the
upstream
directly,
if
the
operator
team
that
creates
an
operator
that
they
have
a
few
release
artifacts,
they
want
to
push
out
every
time
and
fedora
core
os
specifically
for
okd
should
be
one
of
those.
So
that
is
really
the
goal.
So
I
mean
if,
if
that
really
works
out,
we
wouldn't
need
the
the
operator's
sig
in
fedora,
because
there
would
be
nothing
to
do
all
the
upstreams
would
release
for
for
us.
But
I
don't
yeah.
B
A
A
Of
doing
the
hackathon
workshop-
and
you
know
it
with
craig
put
in
the
link
to
the
book
that
josh
and
jay
wrote,
I
think
we
can
get
get
them
to
come
and
you
know
help
teach
the
class
and
give
away
the
book
and
talk
about
you
know.
You
know
how
to
get
involved
in
all
of
these
other
places,
but
also
create
some
educational
materials
that
are
fedora
specific,
that
address.
You
know
things
that
are
fedora
coreos
things
that
are
okd
and
as
as
a
starting
point
at
the
very
least.
E
B
And
you
release
operators
as
container
images
right,
so
we
want
them
to
be
fedora
based
container
images
and
those
will
be
running
super
fine
on
fedora
core
os.
The
problem
with
having
containers
like
fedora
containers
is,
they
won't
necessarily
run
on
older
kernels
because
they
assume
the
fedora
kernel
to
be
present.
So
only
from
that
version
up,
it's
really
supported,
and
we,
you
know
we
wouldn't
want
to
run
container
images
on
rail
7.
B
E
B
E
Yeah
and
really
the
core
of
is
that,
because
we
we're
starting
to
have
an
open
shift
and
there's
this
push
from
red
hat
and
from
other
people
that
we
should
be
running
our
infrastructure
applications
in
open
shift,
whether
it's
a
good
idea
or
not,
is
beside
the
point.
That's
sort
of
the
thing
that's
starting
to
happen.
E
I
don't
want
to
think
about
how
much
worse
it
would
be
if
they
also
had
to
try
to
figure
out
how
to
bring
all
the
applications
on
top
of
it,
and
they
all
worked
in
inconsistent,
weird
ways
that
are
not
necessarily
good
patterns
for
handling.
You
know
those
those
concerns
because,
like
this
is
all
new
stuff.
E
This
is
all
new
to
everybody,
and
so
there
are
very
few
people
who
understand
what
the
good
practices
are
and
how
this
should
work
and
how
this
is
optimized
for
this
case,
in
that
case
and
whatnot
and
every
application
is
different,
and
so
that's
why
it
helps
to
to
be
able
to
have
that
sort
of
thing,
both
upstream
and
the
cncf
side,
as
well
as
handle
fedora
specific
things.
You
know
yeah
in
the
fedora
level,
for
fedora
applications.
A
A
That's
embedded
in
openshift
and
and
we
had
to
do
a
whole
lot
of
handholding
and
that's
been
going
on
for
the
past
two
years,
so
these
guys
now
have
it
down
to
a
science,
how
to
teach
it
and
under
you
know,
and
they
can
flip
over
to
whatever
the
specific
platform
is
or
operating
system
or
ecosystem
that
they're
working
in
on
a
dime
that
you
know
so
it's
it's
not
like.
We
don't
have
people
to
do
this.
We
just
need
the
connections
and
a
date
and
a
place
to
do
it.
A
B
B
A
So
that
for
that,
that
side
of
it,
so
when
you're
talking
with
clement
mention
matt
and
michael
and
jay
and
and
josh,
and
let's
connect
all
those
dots
there
and
I
can
help
with
intros
or
whatever
there.
The
other
piece
of
it
is
supporting
your
request
to
the
red
hat
engineers
team
to
make
sure
that
they
do
that,
and
so
we'll
have
to
there'll,
be
some
internal
stuff
that
I'll
have
to
help
negotiate.
A
I
hope-
and
I
don't
feel
like
it's
going
to
be,
that
hard
of
an
ask
because
there's
a
lot
of
automation
involved,
so
it
might
be
just
that
so
there's
there's
sort
of
two
sides
to
this
coin
and
and
what
you're
going
through
with
fedora.
We
have
gone
through
with
everyone
who's
in
the
certified
operator
program
at
red
hat,
everyone
who's
in
the
community
side
of
things
that
are
doing
upstream
projects.
A
B
Yeah
and
for
for
okd,
specifically,
we
don't
always
need
fedora
operators.
So,
as
I,
as
I
mentioned
before,
the
ubi
ones
are
expected
to
work
so
for
most
red
hat
operators.
They
won't
even
need
to
change
big
things
in
their
operators.
They
just
need
to
make
sure
they
they
respect
the
limits
of
the
boundaries
that
fedora
coreos
sets
them,
and
I
think
the
biggest
one
is
that
we
don't
have
python
on
the
host
in
fedora
coreos.
We
do
have
in
rel
core
os,
so
many
operators
still
assume
that
and
won't
therefore
work
on
fedora.
B
So
if
we
make
them,
I
think
we
just
had
the
the
bird
lip
which
one
was
it
a
cubert
operator
fixed
for
that
and
that's
kind
of
the
thing
I
want
to
see
here.
I
don't
want
to
add
extra
work
for
them
to
release
really
a
second
operator
built
on
fedora.
They
just
have
to
make
sure
that
the
operator,
the
one
they
release
will
also
work
on
okd
and
will
also
be
added
to
the
catalog
on
you
know
we
that
we
ship
with
okd
yeah.
A
E
E
F
F
E
Some
kind
of
taboo
on
it,
but
like
the
best
thing
that
I
can,
the
best
thing
that
we
can
do
is
ask
diane
to
like
poke
somebody
with
a
stick
to
make
this
like
go
faster,
but
like
there
the
amount
of
potential
success
there
is
so
low,
but
that's
the
best
we
can
actually
do,
and
I
I'm
not
willing
to
talk
about
it
in
detail.
While
this
is
being
recorded.
E
F
Completely
different,
I'm
I'm
desperately
searching
for
a
place
where
we
can
collect
training
material
for
for
okd,
because
I
think
there
is
a
lot
of
people
who
made
videos
block
entries,
but
everything
is
spread
spread
around
and
I
yeah
I
would
love-
would
love
to
have
a
place
where
we
can
collect
that.
I.
E
F
E
F
A
E
F
A
Taro,
I
see
your
name
is
on
this
too,
I'm
just
going
to
put
it
towards
the
top,
because
I
remember
to
do
something
about
it
between
now
and
next
week.
So
yeah.
This
is.
This
is
something
that
it's
near
and
dear
to
my
heart
too,
and
creating
it
in
a
way
that
it's
updatable,
maybe
a
yaml
file,
that's
easy
for
people
to
add
things
automagically!
A
You
just
put
the
link
in
and
a
description
that
something
along
that
line,
so
just
give
us
a
little
bit
of
time
and
we'll
figure
that
out
and
start
thinking
about
you
know
what
it
is
that
you'd
like
to
see
there.
We
may
have
to
create
a
couple
templates
so
that
there's
some
consistency
in
what
we're
delivering,
but
yeah
and
and
part
of
it
is.
It
is
also
everything
that
we
just
did
yesterday.
A
A
Yes,
for
sure
move
that
move
that
forward
all
right,
so
we
got
12
minutes
left
of
what
was
supposed
to
be
for
me,
a
very
short
meeting,
and
it
obviously
means
that
I'm
not
downloading
any
of
yesterday's
videos.
So
I'm
not
actually
moving
anything
forward
at
the
moment.
Is
there
anything
else
that
we
should
talk
about,
and
I
would
just
ask
once
you
have
the
update
from
clement
christian?
A
B
B
C
E
E
E
C
E
C
E
E
E
Oh,
you
know
you
know
what
I've.
If
you
look
at
the
the
ipixie
branch,
I
actually
I
need
to
push
an
update
and
because
I'm
using
a
newer
version
of
rook
seth
now,
so
it
may
be
that
if
you're
still
using
the
older
version
of
rook
seth,
that's
included
in
my
github
page.
That
might
be
the
issues
there
may
have
been
something
that
that
broke
with
a
newer
4.5.
E
E
A
D
A
A
But
it
flew
by
so
I
think
I
was
on
premise
and
in
rally
for
two
days
I
was
then
I
flew
back
to
canada
anyways.
I
do
have
to
jump
today
at
here.
I
sent
you
the
email
christian
and
I
cc'd
cheryl
and
vadim
as
well.
So
to
start
the
thread
internally
with
the
operator
framework,
training,
training,
people
and
they
may
all
barf
over
it.
Now
that
I've
volunteered
them
all
but
they've
been
told.
So
we
should
be
able
to
get
something
going
pretty
quickly
with
clement.