►
Description
The OKD Working Group's purpose is to discuss, give guidance to, and enable collaboration on current development efforts for OKD, Kubernetes, and related CNCF projects. The OKD Working Group includes the discussion of shared community goals for OKD 4 and beyond. Additionally, the Working Group produces supporting materials and best practices for end-users and provides guidance and coordination for CNCF projects working within the SIG's scope.
https://okd.io
A
To
the
okd
working
group,
documentation,
subgroup,
meeting
for
may
17th
of
2022
and
let's
jump
right
into
the
latest,
don't
forget
to
put
your
name
in
the
meeting
notes
if
you're
here
that
way,
we
know
that
you
are
here
post
a
link
to
those
notes.
There
we
go
to
the
hack,
md
and
we'll
start
out
brian
updates
to
the
technical
documentation.
B
B
Why
not
just
do
it
on
the
actual
site,
so
it
is
now
you'll
see.
We've
now
got
a
development
section
on
the
site.
I
have
moved
things
forward
a
little
bit,
but.
B
I
can't
get
it
to
build
of
the
six
or
seven
docker
files
in
the
project.
I
can't
actually
build
the
console,
so
I'm
currently
working
through
trying
to
work
out.
What's
going
on
with
that,
but
the
good
news
is,
I
actually
have
a
cluster
that
when
I
do
get
it
built,
I
cannot
test
it
against
it.
B
I
I've
rebuilt
my
home
cluster,
so
yeah,
it's
it's
there
and
I
incorporated
john's
feedback
and
the
idea
is
we
just
want
to
build
it
out
from
here
and
just
get
people
to
do.
Pull
requests.
B
A
C
A
A
Yeah
because
they
were,
it
was
a
combination
before
of
of
prow,
and
I
can't
think
of
the
name,
cumulus,
maybe
serious
serious.
That's
it
seriously.
A
A
A
B
I
noticed
that
it
rendered
quickly,
but
it
didn't
finish
downloading
for
a
little
while
and
the
other
thing
I've
noticed
and
some
sites
have
started,
putting
robot
blockers
on
which
I
think
breaks
as
well.
So
right
so
the
question
comes:
is
it
more
valuable
to
have
the
link
checker
on
automated,
or
should
we
have
it
as
a
manual
task
that
we
occasionally
run
to
verify?
B
Broken
links
are
horrible
on
a
site,
but
this
happens
more
often
than
not
that
our
build
breaks
because,
as
you
see,
we
have
a
couple
of
hundred
links
on
our
site
and
the
chances
of
them
all
being
there
and
all
responding
in
a
timely
manner,
and
we
seem
to
be
hitting
this
more
more
frequently
now.
So
is
it
worth
just
taking
the
link
checker
off.
A
I'll
have
a
look.
That
would
be
one
thing
yeah,
let's
add
that
to
as
a
to
do
just
because
if
it
could
give
warnings
instead
of
that,
that
might
be
interesting.
So,
let's
see.
A
A
A
Okay,
updates
I've
got
a
commit
to
make
to
the
faq
about
the
release,
differences,
I'm
still
collecting
some
information
about
four
10
okds
specific
stuff,
because
there's
a
few
specific
things
in
terms
of
stuff,
that's
slightly
different
from
ocp
theft,
testers
and
crc.
I
haven't
written
anything
up,
no
one's.
I
haven't
seen
anyone
actually
talking
about
seth,
the
pat,
except
for
the
past
couple
past
two
weeks
or
crc.
The
new
focus
seems
to
be
bare
metal.
Everyone
keeps
asking
questions
about
bare
metal
now.
A
D
I
think
that
the
interest,
I
think
the
interest
is
still
there,
it's
just
that
everybody's
waiting
right
and
so
there's
not
much
point
in
flooding
the
channels.
While
we
wait
and
let's
see
if
I
can
find
so,
I.
B
D
Yeah,
unless
last
I
looked
it
hadn't
gotten
out
of
the
kernel,
there
was
still
a
pull
request
outstanding.
C
A
D
So,
probably
I'm
not
going
to
wait,
I'm
going
to
bite
the
bullet
and
back
up
all
my
pvs
and
recreate
them.
D
B
C
D
Then
you
lose
ceph
and
you
can
get
it
back,
but
you
lose
all
of
the
stories
that
you've
had
and
hopefully,
as
I
can
tell
you'll
still
have
the
the
pvcs.
So
everything
will
request
stories.
But
then,
when
you
get
stuff
up
and
running
again,
which
provisions
the
storage,
then
they'll
all
be
blank
and
they'll
have
to.
C
D
Okay,
I
I
think
you,
but
I
mean,
if
you
were
aware
of
all
this,
it
wouldn't
be
a
big
deal,
because
you
could
do
your
new
install
and
we
don't
have
an
operator.
So
you
have
to
sort
of
install
ceph
by
applying
the
crds
right
and
then
waiting
for
a
while.
D
It's
actually
not
that
big
a
deal
and
then
go
in
and
make
the
fix
before
you
actually
start
using
it
and
what's
the
fix,
the
the
fix
is
basically
changing
the
I
I
guess
changing
the
the
storage
yaml
file
so
that
it
doesn't
use
the
the
default.
D
Way
of
starting
up
which
they've
changed
from
synchronous
to
asynchronous,
so
you
have
to
change
it
back
to
synchronous,
okay
and
then,
and
then
it
will
work
with
the
existing
kernels
and
everything.
Now
I
haven't
tried
it.
D
Because
I've
been
chasing
down
the
rabbit
hole
of
the
silly
other
thing
that
I
uncovered
so
yeah
to
see
whether
or
not
I
get
anybody
to
agree
that
there's
actually
some
underlying
bug
or
if
it's
just
life
right
all
right,
but
but
by
the
way
I
noticed
did
anybody
else
notice
that
all
of
a
sudden
about
100
operators
appeared.
It
looks
like
the
on
the
red
hat
side.
Something
happened.
D
And
my
that
sort
of
uncovers
it
one
of
the
things
that
we're
not
very
good
at
is
history.
A
D
Like
when
there's
a
change,
the
only
record
of
the
previous
was
in
your
memory
and
like
I
think
we
we
used
to
be
at
about
40
operators,
or
so
until
I
don't
know
when
sometime
the
last
week
or
two
or
so.
B
But
I
mean
the
challenges:
do
they
all
work?
Because
even
when
you
have
40,
there
were
still
some
community
operators
where
red
hat
had
their
version
of
an
upstream?
They
changed
the
community
operator
to
use
the
red
hat
version,
which
means
you
needed
a
subscription
to
get
to
the
registry.
The
red
hat
registry.
A
B
A
Water,
yet
so
we
can't
yeah,
maybe
put
it
on
the.
I
really
wish
we
had.
The
twitter
thing
worked
out:
let's
do
it
on
the
blog
and
just
do
a
post
like
it's.
I
don't
know
I
feel
like
we
should
just
do
something
like
okay
status
of
of
okd
or
updates
or
changes,
or
something
something
that
we
can
like
update
the
blog
post
with
latest
news,
or
something
like
that,
like
maybe
just
the
latest
news
blog
post.
Does
that
make
sense?
Well,
I.
C
A
Next
week
yeah,
we
don't
want
to
do
that.
My
thought
was
that
I'll
reach
out
to
christian
and
see
if
he
knows
anything
about
it.
So
because
I
don't
what's
the
current
time
in
spain.
B
A
Yeah,
so
folks
are
probably
all
out
drinking
right
now
right
because
that's
what
kukan
eu
so
I'll
shoot
a
message
to
him
and
see
if
I
can
get
some
clarity
on
that
and
then
I'll.
Let
folks
know
because
it
seems
to
me
like
we
should
say
something,
because
it's
like
wow
our
operators
just
quadrupled
right.
So.
B
D
No,
that's
all
right,
that's
fair!
I
started
trying
some
of
them
on
on
my
test
cloud,
in
particular
the
certificate
manager
one
but
and
let's
see
what
else
is
there
well,
there's
a
whole
pile
of
aws
ones
which
aren't
really.
I.
D
And
the
the
ones
that
we're
most
interested
in,
which
would
be
like
you
know,
tecton
and
serverless
and
maybe
istio
or
you
know
those
things
I
don't
think.
Are
there
yet.
D
The,
although
the
the
arco,
cd
well.
B
B
D
B
A
Right
exactly
okay,
oh
yeah
go
ahead.
A
So
crc
we're
waiting
for
diane
to
get
back
to
us
about
that
external
group
see
if
they
want
to
do
it.
She
said
she
reached
out
to
them,
we'll
see
so
the
scoop
on
twitter
is
that
I
still
don't
have
access
and
diane
it's
unclear
if
diane
has
access
to
it
as
well,
because
it's
based
on
that
employee's
email
address
and
I
haven't-
got
a
response
to
diane
yet
about.
A
A
So
nothing
there
survey
folks
haven't
really
said
anything
about.
It,
haven't
really
commented
or
updated
anything
to
it.
My
sense
is
that
it
that's
okay,
because
we
can
wait
until
kubecon
to
you
know,
wait
like
until
next
week
or
the
week
after
to
send
this
out
to
folks
yeah.
B
I
think
the
last
two
weeks
have
been.
We
had
red
hat
summit
and
now
we've
got
kubecon,
which
is
it's
the
open
it's
the
okd
session
tomorrow.
Isn't
it.
C
A
Yeah
so,
but
I
do
want
to
keep
this
on
our
palette,
because
I
think
that
doing
this
would
be
fantastic.
Why.
B
C
A
Fantastic
bare
metal
request
for
testers
bare
metal.
We
don't
have
twitter
to
be
able
to
request.
So
I
don't
know
if
we
want
to
do
something
other
than
twitter
and
maybe
just
send
something
through
the
working
group
email.
I
was
going
to
do
that,
but.
B
I
because
just
seeing
a
bare
metal
tester
is
it
an
ipi?
Is
it
a
upi?
Is
there
anything
specific
about
it
with
storage
or
anything?
Did
we
create.
A
A
We're
just
getting
a
lot
of
bare
metal
stuff
that,
like
no
one
knows
how
to
answer,
because
I
don't
have
bare
metal
and
a
lot
of
other
folks
and
the
red
hat
folks,
don't
have
a
lot
of
bare
metal
experience,
and
so
the
idea
is
that
we
at
least
firm
up.
It's
we
get
some
testing
so
that
we
have
some
data
about
how
bare
metal
is
working,
because
there
is
now
the
assisted
installer
like
running
it
locally
right
and
then
the
other
aspect
is
we
want
to
shore
up
some.
A
B
How
do
I
get
a
serial
console
or
things
that
are
very,
very
bare
metal,
specific
in
terms
of
or
a
particular
fiber
driver
for
a
bit
of
hardware
or
a
bit
of
of
a
predominantly
bare
metal,
specific
stuff,
I.e?
The
virtual
machine
gets
rid
of
all
that
complexity
in
terms
of
device
drivers
and
things
like
that,
or
are
they
generic
bare
metal
stuff
where
a
vm
would
actually
work?
C
B
D
D
And
but
whenever
I
look
at
the
bare
metal,
you
know
documentation,
it
seems
like
they're
all
assuming
that
they
have
linux
on
it,
which
to
me
is
not
bare
metal.
I
mean
that's
like
a
liver
or
something
like
that
which
is
sort
of
a
different
branch,
but
somehow
those
are
conflated
in
my
mind,
so
I
don't
know
that
and
it
could
be.
Maybe
it's
just
me
that's
confused,
but
do
we
know
what
exactly
people
mean
when
they
say
bare
metal.
A
And
I
think
that
that
so,
if
I
look
at
the
documentation-
which
you
know
you
know
michael,
you
can
speak
to
this-
probably
better
than
any
of
us
is
it's.
It's
bare
metal
is
just
in
one
of
the
install
options
in
the
documentation
and
it
doesn't
actually
show
like
oh
michael
dropped.
It
doesn't
actually
show
any
type
of
explanation
of
what
they
mean
by
bare
metal.
It
just
says
you
can
install
okd
on
bare
metal.
B
And
then
I
think
it
does
use
live
verts,
it
does
builds
the
builds
the
the
containerization
on
top
of
that
yeah.
B
A
D
Different
like,
as
brian
was
saying
you
know,
if
it's
you
know,
when
you're
saying
bare
metal,
what
specific
things
are
you
worried
about.
A
All
right,
let's
put
that
as
a
question
to
the
larger
group,
I'll
put
it
in
the
to
do.
D
A
All
right
style
updates,
I
think
that's
all
done
so
we
can
pull.
B
Yeah,
I
I
think
we
need
to
ask
diane
if
the
friend's
available
to
do
any
more
work.
There's
just
I
mean
just
thinking
about
what
we're
supporting
today
is
having
an
area
on
the
home
page
for
latest
news,
where
we
can
put
something
on
that'll,
be
there
for
two
weeks
and
then
automatically
go
away.
So
we
don't
have
to
worry
about
having
latest
news
and
it's
certainly
nine
months
old
and
so
a
feature
like
that
added
onto
the
home
page,
I
think,
would
be
useful.
B
B
As
far
as
as
far
as
I'm
aware,
the
styling
update
is
finished,
he's
delivered
it,
it's
live
and
that
piece
is
done,
but
I
think
there
are
updates
and
what
I'm
talking
about
updates.
I
will
just
last
weekend
I
did
update
all
the
tooling
in
the
automation,
so
I
went
through
and
made
sure
that
we
had
the
latest
versions
of
mk
docks
of
material
on
toolings.
We
were
still
using
an
old
centos
image
to
actually
base
the
build
on.
B
C
C
B
A
A
Yeah,
oh
okay,
so
now
we're
up
to
new
business,
I'll,
just
pull,
publishing,
updates
or
style
update
anything
new
about
the
publishing.
I
think
I
are
we
agreeing,
then
that
we
should
disable
at
least
temporarily
until
you
find
out.
If
those
can
be
warnings
instead
of
errors,
can
we
just
disable
it
for
now
or
do.
B
I
mean
I,
I
get
notifications,
so
whenever
anything
fails,
I
go,
I
go
poke
it
and
prod
it
and
okay
and
if
it
is
blocked,
I
bypass
it.
B
Yeah,
I
I
do
go
review
the
files
and
check
it's
okay
and
look
at
what
the
failure
was
and
making
sure
that
it's
not
the
changes.
That's
introducing
the
breaking
feature,
it's
something
old
and
then
I
I
get
it
working,
but
I'll
probably
do
that
this
evening
or
within
by
the
weekend
I'll
have
that
either
sorted.
C
A
Zoom,
okay,
so
this
came
up
before
currently
we're
using
blue
jeans.
It
doesn't
have
as
much
adoption.
I
don't
have
direct
access
to
the
recordings,
because
this
is
actually
diane's
account
and
so
we're
behind
again,
because
I
have
to
reach
out
to
diane.
Diane
has
to
make
time
in
her
busy
schedule
to
grab
the
recordings
and
and
share
them
out
with
me,
etc,
etc,
and
also
zoom
has
is
more
feature
rich
out
the
gate
it's
it
seems
I
could
be
wrong.
A
A
Well,
so
here's
here's
the
thing
it
would
have
to
be
community
funded,
meaning
one
of
us
would
find
more
than
happy
to
contribute,
but
we
would
want
an
email
address
that
multiple
of
us
have
access
to
right
and
you
know
one
way
or
another.
We
need
to
find
find
a
resolution
to
the
email
thing.
We
need
a
dedicated
email
for
all
of
these
accounts
that
we
want
to
set
up.
You
know
in
different
services
and
stuff
like
that,
so
assuming
we
figure
out
the
email
thing,
what
do
folks
think
about
moving
to
zoom.
B
A
It's
like
less
than
an
hour.
Isn't
it
it's
40
minutes,
40
minutes
right!
Yes,
I'd
be
willing
to
to
to
contribute
towards
the.
I
think
that
the
least
expensive
one
is
like
60
a
year
or
something
like
that.
I
don't
know.
Let's
take
a
look.
A
Plans
and
pricing-
oh,
I
guess
it's
a
little
bit
more.
It's
like
150
a
year
for
the
pro
that's
up
to
100
participants,
large
meetings,
one
gigabyte
of
recording,
whiteboard.
A
B
A
A
B
I
I
actually
think
that
that
is
part
of
the
problem.
Is
that
we're
not
truly
an
open
source
community,
because
at
the
minute
everything
is
built
around
the
red
hat
product,
the
build
the
process
and-
and
yes,
we
can
do
things
in
a
community
way,
but
we
can't
really
decide
the
direction
of
okd
right.
That's
that's
a
red
hat
product,
manager's
activity.
B
So
getting
an
external
company
to
support,
I
think,
is
going
to
be
a
little
bit
more
challenging,
maybe
because,
at
the
end
of
the
day
it's
a
red-hot
commercial
product,
that's
driving
a
direction,
not
a
community.
C
D
D
A
But
if,
if
no
one
has
had
success,
building
it
from
source
yet
and
multiple
people
have
tried,
then
we're
not
really.
C
D
Right,
well
I
mean
when,
when
christian
is
talking
about
prow
he's
talking
about
the
red
hat
internal
running,
exactly
right,.
C
D
D
B
A
A
So
brian,
did
you
catch
that
conversation
between
john
and
christian,
where
christian
said,
oh
there's,
there's
parallel
for
all
of
these
is
is
there
and
I
asked
the
question
like?
Is
it
every
single
one
at
that
particular
registry?
Well,.
B
B
Product
around
it
right-
and
it
may
be
that
everything
here
is
an
open-
is
in
open
source.
It's
it
may
or
somewhere
be
in
the
in
the
several
hundred
reposes
in
the
open
source
github.
But
the
challenge
is:
do
you
have
time
to
actually
go
and
put
all
the
pieces
together.
A
D
B
Yes,
exactly
exactly
so,
and
this
is
what
I'm
going
through
with
the
console
project
at
the
minute.
It's
just
like
out
of
the
box,
I
did
a
podman
build
failed.
I
did
a
docker
build
just
in
case
failed
and
now
I'm
well.
Why
did
it
fail?
You
know
I'm
not
going
down
the
rabbit
hole
of
let's
try
and
figure
this
out.
As
I
say,
there's
a
dockerfile.product
which
suggests
it's
the
product
build
rather
than
the
dev
build,
and
that
then
uses
these
internal
images
which
I
don't
have
access
to
so
yeah.
B
I
I
think
this
is
now
that
we
started
this
technical
track.
These
are
the
sort
of
things
and
we
want
to
start
off
when
we
actually
figure
out
what
the
problem
is.
We
want
to
start
raising
issues
and
make
sure
that
every
project
in
the
release
info,
where
we
point
the
the
url
every
project
there
is
buildable
from
the
readme
file
within
the
project.
B
There's
no
ambiguity,
there's
no
guessing,
and
so
I
think
that's
where
we
want
to
get
to-
and
I
think
might
make
the
the
the
the
comment
that
within
prow
they
don't
actually
take
note
of.
What's
here
because
they
have
a
global,
replace
where
they'll
actually
look
at
the
thing
and
proud
will
actually
replace
the
base
image
with
the
current
one.
B
A
Yeah
all
right!
Well,
that's
that's
a
lot
for
us
to
digest,
but
we'll
approach,
diane
and
we'll
find
out-
and
this
just
sprang
from
the
discussion
of
the
zoom.
But
it's
a
reoccurring
topic
like
what
what
actually
and
if
we
had
something
definitive
from
red
hat.
I
would
feel
better
if
it
was
actually
like,
even
if
they're
saying
we're
going
to
continue
to
control
this
at
x
level,
just
to
hear
them
say
that
or
write
it
down
somewhere
so
that
we
know
so
that
we're
not
always.
A
A
All
right:
well,
we
have
our
to
do's
ahead
of
us,
and
the
to
do's
are
check
on
link
checking
give
to
give
warnings
instead
of
errors.
That's
for
you,
brian!
Let's
see
what
else
announcement
on
operator
change
I'll
check
with
christian
first
I'll,
bring
bare
metal
discussion
to
the
greater
group
and
brian
get
the
status
of
of
brandon
for
more
style
work
find
out.
I
don't
know
if
maybe
you
can
reach
out
directly
to
him.
B
B
A
B
B
A
You
know
and
to
be
fair.
Some
of
it
is
also
my
hesitancy,
because
of
not
knowing
what
this
change
is
and
what
it's
going
to
mean.
B
Yeah,
I
I
was
going
to
say,
because
I
mean
in
reality
we're
really
talking
about
the
open
shift
and
the
okay
data.
Io
repos
they're,
the
only
two
that
we
really
control.
Everything
else
is
going
to
have
to
stay
what
happens
with
the
operators,
where
the
community
operators
end
up
living
if
they're
being
made,
maintained
and
managed
and
built
by
proud
they're,
probably
going
to
be
in
the
open
ship
dog.
A
Yeah
and
I
don't
know
well
I
mean
they
could
now,
let's
from
a
technical
perspective,
they
could
pull
in
from
an
external
repo
and
use
it
in
their
internal
prowl.
So
if
that
that
would
be
possible,
but
they
would
have
to
set
that
up
to
happen
well,
but
you
know
what
I
mean
like:
allow
us
to
submit
code
and
then
have
it
pulled
in.
B
A
My
sense
is
that
it,
the
discussion,
will
probably
be
about
a
lot
of
this
stuff,
because
if
we
can
get
christian
there
or
and
or
vedine,
I
think
we
need
to
sort
of
hammer
away
at
this
like
okay.
What
can
we
really
do?
What's
going
to
get
impacted
by
this
change?
That's
happening
on
the
red
hat
side
and
how
much
effort
should
we
be
putting
into
this?
You
know
at
the
last
main
meeting
christian
said
yeah.
A
B
I
mean
I,
I
I
think,
if
they're
open
to
us,
doing,
pull
requests
against
the
sources
right,
but
for
me
that
that
means
that
I
have
to
be
confident
that
my
build
and
my
local
build
and
test
isn't
going
to
fall
down
on
the
first
hurdle
when
it
gets
to
the
red
hat
system.
So
for
me
to
feel
confident
about
doing
a
pull
request,
I
have
to
feel
confident
that
I
can
do
a
build.
A
No,
but
that's
the
in
the
same
vein
as
what
I
was
saying
a
few
minutes
ago.
They
could
create
a
a
page
that
gets
generated
with
results
from
full
request
that
isn't
directly
within
rao
right,
like
they
could
mirror
some
of
the
results
from
prow
out
to
the
outside
world.
To
let
people
know
the
success
you
know
of
their
tests
and
stuff
like
that
and
details
and
whatnot.
B
D
Yeah,
basically,
there
should
be
something
that
you
can
fork
and
you
know
that's
a
reliable
copy,
yeah
and
you
you
build
your.
You
know,
change
against
your
fork
and
you
know
before
you
do
your
pull
request
you.
You
know
you
update
the
fork
to
make
sure
it's
current
and
then
you
try
it
again
and
then
you
put
it
in
and
the
the
way
it
is.
That's
not
possible.
D
Yeah
I
mean,
like
you
know,
you
can
fork
tiny
bits
and
pieces
of
things,
but
you're
still
missing
some
critical
parts
that
you
can't
build
and
it
doesn't
make.
Well,
I
guess
the
like
the
okd
part
is.
D
And
you
know
somehow
there's
these,
but
you
know
like
a
lot
of
a
lot
of
the
ocp
stuff,
is
presumably
open
source,
but
somehow
there's
these
little
sticklers
that
I
don't
know
if
it's
so
much
red
hat
philosophy
as
just
that
everybody's
busy,
and
they
don't
have
time
to
think
about
it.
D
And
I
I
would
hope
that
it's
the
they
don't
have
time
to
to
worry
about
it,
because
if
that's
the
case,
then
you
have
some
chance
of
fixing
it.
B
So
they
the
way
that
they
work
in
the
projects.
They're
not
deliberately
trying
to
to
stop
us
doing
things,
but
they're
really
doing
it
for
the
pro
bill.
So
they
know
what
prowl
needs
and
what
the
internal
systems
need
and
that's
where
their
effort
is
going.
They're,
not
really
considering
a
community
having
to
sort
this
stuff
out
that
doesn't
have
access
to
internal
red
hat
systems,
and
I
think
that's
part
of
the
challenges
that
that
you're
then
trying
to
pick
up
that
that
content
from
outside
in
and
it
you
don't
always
have
a
great
time.
D
Yeah
but
but
by
the
way
on
the
cef
side,
it
looks
like
the
patch
is
made
into
the
testing
branch.
Okay,
okay,.
A
Right
that
too,
all
right!
Well,
let's
go
ahead
and
end
for
today,
brian,
I
will
see
you
in
two
weeks
proof
so
I'll
probably
see
you
next
week,
we'll
put
this
meeting
to
a
close.