►
From YouTube: OKD Working Group Meeting 07-19-2022
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
All
right,
let's
go
ahead
and
get
started.
This
is
the
okd
working
group
meeting
for
july
19
2022.
Please
put
your
name
in
the
attendees
section.
If
you
could
that
way,
we
know
if
you
were
here
or
not
if
we
need
to
get
important
information
to
you
and
take
about
30
seconds
to
look
at
the
agenda.
If
you
haven't
already-
and
let
us
know
if
there's
anything
you'd
like
to
add
change,
modify
move
whatever
so
take
about
30
seconds
to
do
that,
please.
B
A
I
don't
think
so.
I
think
that
was
a
cut
and
paste
error
by
someone
who
added
something
the
other
day.
So
I.
B
B
A
D
D
That
should
give
us
the
door
36
that
should
give
us
a
stable
federal
courts
as
a
basis
and
then
so
once
of
course,
I
don't
think
any
significant
features
are
landing
this
release,
but
we
have
fixed
for
battle
metal
installation
that
should
make
it
work
on
okay,
natively
and
a
few
more
bugs
were
found
related
to
q
beard.
Hopefully,
l36
would
fix
it.
The
installing
keyboard
on
latest
opd
seems
to
break
cylindex
policies.
D
I'll
drop,
a
link
to
the
bug
in
the
chat
later
on,
so
hopefully
moving
to
six
should
should
fix
it.
D
D
Using
that
to
simplify
this
work,
we
prepared
a
bunch
of
build
configs
for
most
of
the
components
which
are
rebuildable
for
okay,
so
that
it
would
be
easy
for
you
to
find
the
reference
how
to
do
that.
D
Most
of
them
are
fairly
straightforward.
Just
take
the
dockerfile
and
use
the
centos
9
provided
builder
as
a
base
for
some.
You
need
to
replace
the
docker
file
because
those
were
written
for
relate
some
contain
bugs
which
are
fairly
trivial.
So
we
would
need
to
work
with
upstream
to
do
that,
and
I
tried
to
rebuild
the
whole
okay
payload
using
this,
but
I
don't
think
I
succeeded.
D
It
failed
on
some
modest
problems,
so
we'll
keep
on
digging,
but
in
general,
that
would
be
useful
for
both
okd
and
ocp,
because
eventually,
we'll
move
to
rail
9
base
for
all
the
packages
and
having
this
tested
beforehand
would
be
incredibly
valuable
and
show
both
to
the
community
and
to
write
engineers.
Why
bookity
is
such
a
cool
thing.
D
And
I
think
that's
pretty
much
it
since
we
have
this
as
a
bunch
of
build,
complex
and
tecton
pipeline.
This
is
where
we
can
work
with
operate
first
folks
and
provide
some
kind
of
interface
or
a
community
cluster
for
folks
on
the
community
to
login,
build
their
own
versions
for
okay
and
proceed
further.
D
So
once
we
have
this
stabilized,
we
can
start
working
with
them
on
how
to
approach
this.
I
think
resources
and
the
authentication
would
be
the
largest
problems,
but
I'm
pretty
confident
we
can
tackle
this
and
I
think
that's
pretty
much
it
from
from
the
release
site.
B
A
E
So
I'll
take
this
one:
okay,
the
way
fibs
work
as
I
do,
and
as
far
as
I
understand,
the
complexity
of
the
whole
thing
is
that
to
have
something
fully
certified,
you
submit
a
lot
of
paperwork
to
a
certain
government
entity
in
the
u.s
about
how
your
cryptography
stack
is
implemented
and
what
it
does,
and
things
like
that.
So
right
now,
some
like.
If
we
want
to
have
anything
based
on
centos
stream,
cryptography
stack,
which
is
essentially
open,
ssl
be
fips35.
E
We
need
someone
to
submit
that,
as
with
all
the
paperwork
do,
and
this
I
think
I
remember
quickly-
I
don't
think
that
it's
going
to
happen
in
as
part
of
the
work
where
that
radar
does,
because,
essentially,
when
you
ship
something
like
that-
and
you
want
to
have
certified-
you
don't
want
to
have
the
whole
stack
change
frequently.
So
it
works
for
well
because,
because
well,
updates
are,
should
be
only
critical
fixes
or
purchases
like
that
for
a
specific
release.
E
A
given
specific
version
of
the
crypto
package
that
would
be
certified
would
stay
the
same
for
given
rail
release,
so
the
certification
itself
would
be
preserved
in
code
for
central
stream
change.
App
changes
happen
as
they
go
as
they
come,
so
when
does
the
certification
lose
its
meaning?
I
don't
know
it's
probably
up
to
this
to
answer
that,
but.
A
C
B
C
A
The
paperwork
thing
is,
you
know
one
aspect,
but
from
an
engineering
perspective
I
think
is.
I
was
asking
from
that
because
of
it's.
Basically,
it's
making
sure
that
you
have
right
those
modules
that
are
certified
and
I
don't.
I
don't
know
that
that
would
change
a
lot
I
mean.
Does
it
do
do
those?
Does
that
change
a
lot
in
fedora
in
general?
I
don't
know
I
don't
I
don't
pay
close
enough
attention
so.
E
The
the
the
general
idea
is
that
essentially
center,
a
stream
is
the
next
role,
so
everything
that
goes
into
central
stream
will
be
into
rail
at
some
point,
so
the
crypto
stack
that
is
in
central
stream
is
really
close
to
what
will
be
in
real
life.
E
C
B
Yeah,
I
think
that
was
that's
the
shortest.
I
think
it's
something
we
could
consider
doing
in
the
future
when
we
have
it.
You
know
when
it's
real
and
when
the
process
is
there
then
yeah,
but
it
would
take
like
another
sub
group
and
someone
dedicated
to
you
know
resubmitting
on
a
regular
basis,
these
fixed
things,
and
I
don't
think
we're
there
yet.
But
a
third
party
might
you
know?
Maybe
you
know
something
like
data
or
somebody
wanted
to
be
in
the
business
of
doing
that
and
having
a
release
that
was
fips
compliant.
A
That
is
true:
okay,
moving
on
to
the
fedora
core
os
updates
and.
E
E
Is
is
the
idea
to
use
central
stream
as
the
base
for
cluster
operations,
or
is
it
to
use
central
stream
as
a
host
node
os?
So
there's
two
things:
there's
one
thing
that
has
that
valiant
talked
about
before,
which
is
using
center
stream
images,
container
images
to
build
okg,
which
is
something
that
is
only
done
on
the
build
in
frasca
to
generate
the
okd
pedals.
The
thing
that
you
install
in
the
end
containers,
so
that's
he
is
doing
that
right
now
then,
there's
the
centos
stream
chorus
work,
which
is
didn't
progress.
C
D
This
option
is
on
the
table,
but
that
would
bring
us
back
to
the
okd
three
days
where
we
had
to
rebuild
every
single
container
which
gets
us
from.
We
do
not
need
to
do
this
on
daily
basis.
We
would
rather
just
take
them
probably
from
musabi,
but
for
the
community
we
tell
them
hey
we're
building
them
based
on
ubi
and
some
are
built
on
top
of
rail,
because
we
need
packages
like
open
speech
and
and
sdn,
and
things
like
novi
and
things
like
that.
D
But
you
are
able
to
build
them
using
your
uvi
most
of
the
cases
and
the
rest.
You
should
be
able
to
build
them.
That's
what
we're
trying
to
prove
here,
but
we
probably.
D
Them
because
it's
it
just
makes
life
harder
for
us.
As
for
the
host
node
os,
this
option
is
on
the
table,
but
I'm
not
sure
which
benefit.
Does
it
bring
to
us
like
when
it's
fairly
clear
what
fedora
gives
us
when
compared
to
rail
it
gives
us
a
stream
kernel.
It
gives
us
fresh
packages,
it's
not
clear
what
centos
stream
gives
us.
D
It
doesn't
give
us
bit
a
bit
compatibility
to
rile
and
it
doesn't
give
us
fresh
enough
packages
it
can
it
doesn't.
We
don't
see
the
benefit
of
it
at
this
point,
but
it's
still
of
course,
discussable.
Why
we
want
this
experiment
so
that
we
could
prove
that
otd
is
experimentation
platform.
You
can
turn
it
into
whatever
you
want
to,
and
it's
helpful
for
ocp.
That
means
we
get
more
trust
from
from
the
engine,
but
it
probably
would
be
just
a
ci
level,
experimentation
thing.
D
A
All
right,
I
think
this
is
stuff
that
would
go
into
a
nice
little
faq
item
I'll
bring
it
to
the
docs
meetings.
I
think
at
this
point
we
should
start
the
documentation,
working
group
or
community
outreach
working
group
or
whatever
we're
going
to
call
it
maybe
should
put
something
together
that
talks
about
this
project
a
little
bit
and
maybe
shed
some
light
on
it,
so
that
we
can
prepare
folks,
because
it
sounds
like
we're
getting
close
to
something
being
available.
A
We
did
f,
updates
documentation,
subgroup
updates,
brian.
E
We
do
yeah,
no
problem,
I
don't
I
don't
have
much
so
you
won't
take
long.
The
first
one
is
a
link
to
this
week
this
month
in
federal
rest,
which
made
through
addition
that
we've
had
it
s3
e90x
images
in
federal
rs,
but
I
don't
think
we
have
s319x.
E
Okg
builds
right
now
so
well.
The
base
is
here.
The
next
might
come
later,
and
the
second
item
is
that
yeah
we've
talked
about
century
stream
chorus
in
previous
meetings.
It's
still
working
progress
right
now,
it's
block
on
engineering
on
working
on
it,
and
hopefully
we
can
give
an
update
at
some
point
later.
B
And
just
there
was,
and
I
apologize
I've
been
offline
for
a
couple
of
days.
I
think
it
was
peter
hunt
who
is
on
the
call.
Now
was
I'm
working
on
that
blocker
for
you
so
peter?
I
don't
know
if
you're
on
the
call
and
can
china.
H
Yeah,
so
I
don't
have
a
huge
status
update.
We,
the
cryo
community,
has
had
a
couple
of
requests
for
el9
builds
and
we
haven't
gotten
there.
Yet
I
want
we
had
talked
on
some
form.
I
don't
remember,
which
form
about
having
an
okd
working
group
within
centos,
which
would
give
us
a
place
to
build
the
packages
for
cryo
and
cry
tools
and
the
cubelet,
but
so
probably
need
to
chat
a
little
bit
about
that
but
yeah.
I.
A
Don't
have
any
specific
updates,
because
we
have
to
kind.
H
Of
determine
where
we're
gonna
be
putting
the
packages
and
exactly
how
we're
gonna
build
those
up.
E
E
Okay
right
now,
right
now,
yes
right
now,
gublet
will
be
built
on
rpm
cryo
needs
to
be
built
in
rpm
and
include
it
into
escos.
E
Maybe
in
the
future
we
could
use
chorus
layering
to
avoid
that
and
directly,
but
we'll
still
need
to
have
rpms
but
or
maybe
not,
maybe
not,
but
well,
right
now,
yeah.
We
need
it.
H
That's
that's
another
thing
that
I
was
kind
of
thinking.
We
could
avoid
having
any
any
official
place
to
put
the
pack
like.
We
don't
need
the
centos
working
group,
necessarily
if
we
create
you
know
with
those
proper,
a
copper
repo
and
have
the
a
layer
pull
from
the
copper
repo
and
build
on
top
of
s-cos.
And
I
I
mean
I
think
that
that
could
end
up
being
an
idiomatic
way
of
handling
the
packages.
H
And
I
wanted
to
talk
about
that
because
that
could
be
easier
than
you
know,
having
a
full
working
group
and
you
need
to
do
all
the
paperwork
for
all
of
that.
B
Is
it?
Is
it
something
that
you
could
have
a
sub
group
underneath
the
okd
working
group,
or
is
it
does
it
necessarily
have
to
be
in
in
the
center?
I
would
like
to
start
creating
a
collaboration
bridge
with
the
centos
community.
So
I'm
you
know
I'm
open
to
I'm
personally,
I'm
not
speaking
for
the
working
group
interested
in
how
we
do
that
and
reach
out
to
the
centos
folks
to
make
them
aware
that
this
is
happening.
H
Yeah,
I
don't
know
so
so.
The
thing
that
we
have
to
determine
is
how
officially
we
want
to
ship
cryo
in
centos,
because,
like
so,
for
instance,
in
fedora,
now
we're
shipping
it
as
a
module
that
has
been
cumbersome
for
me
as
the
packager
to
put
together
the
cryo
packages,
and
it's
not
exactly
matching
what
we
do
with
openshift
proper,
because
we're
the
the
cadence
of
our
release
in
a
release.
Branch
is
too
frequent
for
an
official
package
system.
H
Like
you
know
the
fedora
packaging
system,
whereas
if
we
had
a
copper
repo,
we
could
just
you
know,
put
whatever
we
want
in
there
and
that
would
just
be
pulled
straight
to
the
you
know
built
on
top
of
the
release
image.
So
it
kind
of
like
I'm
open
to
any
level
of
collaboration,
but
I
think
generally,
what
we
need
to
determine
is
like
if
we
want
these
cryo,
the
cryo
rpm
to
be
generally
available
in
centos
or
if
it.
E
Yeah
it's
a
little
bit
up
to
you
like.
If,
obviously
you
want
to
support
folks,
but
I
would
mean
also
to
the
later
like,
if
you
want
to
start
and
focus
only
okd
or
kt.
First
then
yeah.
We
could
probably
make
a
couple
somewhere
in
different
right
in
front
and
if
you
push
changes
there
and
we
use
that
build
skus
for
now
and
we'll
see
here,
if
other
folks
want
to
take
a
maintenance
or
make
up
a
century
working
group-
and
they
are
free
to
do
so
too
so.
H
Yeah,
yeah,
and,
and
and
mike
to
that
point
we
have
some
people
who
are
using
it
on
centos
right
now
we
build
through
open
souza's,
build
system
obs
and
people
install
ontentas
with
that
it
doesn't
yet
support
centos
9,
so
we
wouldn't
be
able
to
use
it
for
that.
Quite
yet.
H
Basically,
what
the
cryo
community
has
is
a
lack
of
time
to
put
into
packaging
and
there's
all
these
different
packaging
solutions
for
all
these
different
distributions
that
our
users
need
and
the
more
one-offs
that
we
had
but
have
the
higher
the
burden
of
packaging
becomes,
which
is
why
you
know
it's
my
preference
to
have
it.
I
you
know
my
ideal
would
be
if
the
it
would
be
pulled
directly
from
obs
or
something
or
like
some
auto
build
service
that
we
create
to
reduce
our
overhead,
because
we're
already
spread
really
thin.
H
B
The
only
thing
and
and
a
lot
of
what
you
said
to
me
with
just
then,
was
a
lot
of
acronyms
and
a
lot
of
places.
So
I'm
just
say
that
out
loud.
B
The
thing
that
that
I
worry
about-
and
we've
done
in
the
past
is
had
things
that,
were
you
know,
specific
build
for
okd,
and
then
we,
you
know,
vadim
has
done
tremendous
amount
of
work,
keeping
us
in
a
release
path
that
that's,
and
I
just
the
more
that
this
could
be,
and
I
know
it's,
the
burden
of
the
resources
of
creating
this
package
can
be
just
part
of
the
automated
build
of
cryo,
and
I
understand
resources
are
thin
the
better.
B
This
will
be
not
that
you
shouldn't
do
what's
gonna
work
for
this
first
mvp,
but
I
long
term
it
never
works
out
for
us
to
have
something.
That's
custom
for
okd.
It
always
ends
up
being
on
some
poor
sod
like
vadim's
lap
or
you
know
mike
or
somebody
doing
something
special
like
daryl
and
his
wonderful
snowballs
of
things
yeah,
it's
it's
just
the
more.
B
It
can
just
be
part
of
the
cryo
release
process
and
packaged
up
the
better,
and
if
I
need
to
go
in
and
sit
and
arm
wrestle
with
dan
walsh
or
somebody,
let
me
let
me
know
he'll
win,
but
he'll
do
it.
If
I
have
to
you
know,
jump
up
and
down,
but
do
what
you
need
to
do
to
get
to
the
mvp,
which
is
you
know
this
first
release
of
it
and
then,
let's
not,
let's
try
not
to
make
it
little
custom
things
for
okd,
if
possible,.
H
A
Excellent,
that's
fantastic!
Oh!
By
the
way
we
have
a
policy
if
there's
a
guest
on
video
that
you
have
to
introduce
them
who's
your
guest,
peter.
H
This
is
kitkat
and
she
every
time
I
talk
she
decides.
She
remembers
that
I
exist
and
wants
attention.
C
A
All
right
and
sorry,
timothy
that
I
sort
of
skipped
over
when
we
got
diverted.
I
apologize,
I'm
good.
Okay,
let's
move
now
to
the
documentation,
subgroup
updates
with
brian.
G
Okay,
so
we've
got
quite
an
interesting
discussion
last
weekend.
First,
one
is
the
actual
repo
move,
so
we
actually
have
some
use
of
the
okd
project-
github
retail-
and
it's
now
trying
to
line
up
the
red
hat
resources
to
move
the
open
shift
and
the
open
shift
mcs
repose,
which
is
the
okay
I
o
and
the
main
okd
repo,
and
we
are
going
to
do
a
a
move
rather
than
try
and
sort
of
copy,
because
that
means
everything
goes.
The
discussion
groups
the
the
issues
everything
just
moves
across
to
one.
G
So
we
need
the
red
hatters
that
control
the
dns
and
to
actually
switch
the
okd
to
io
across,
and
then
we
also
need
a
red
hatter
that
can
actually
do
the
move.
That
has
the
authority
to
actually
pass
over
the
repos
from
the
open
shift
and
the
open
shift
cs
organizations
to
the
okd
project.
G
So
we
need
to
sort
of
do
the
cap
herding
to
get
these
people
together,
set
a
time.
Do
the
move,
so
I'm
sort
of
looking
at
diane's
direction.
B
Like
that
poor
sod
who's
in
in
charge
of
that
will
gordon
and
we've
been
going
back
and
forth.
I
think
he
was
on
vacation
a
little
bit
last
week.
Will
gordon
and
jerry
folla
are
the
two
that
I
rely
on
for
that,
and
I'd
also
had
asked
the
will
to
track
down
the
mx
stuff
too
for
the
email
address.
B
So
I
think
you
guys
have
his
email
will
gordon
so
maybe
reach
out
to
him
with
a
little
email
cc
me
I'll
make
him
answer,
and
let's
find
a
time
that
we
can
just
do
it
and
get
this
off
of
our
plates.
He
would
be,
he
would
be
the
person
the
mx
one.
He
was
asking
legal
for
one
more
sign
off
of
something
and
I'll
check
in
with
him
on
that
too,
but
yeah
it's
will
garden
and
the
other
person
is
jerry.
B
And
but
his
time
zone's
not
quite
great.
B
Yeah
and
basically
owns
openshift.cs
so
which
is
different.
That's
for
customer
success
because
we
use
customers
everywhere,
charo.
G
Okay,
so
hopefully
in
the
next
week,
or
so
we
can
get
that
done,
which
means
that
we'll
then
be
within
the
okd
dash
project,
github
organization,
so
everything
so.
The
next
thing
is
we
had
a
discussion
around
what
are
currently
the
guides,
so
we've
got
a
number
that
are
outdated,
there's
a
couple
that
we're
going
to
change
to
blogs,
and
but
I
think
we
need
to
have
the
discussion
around
and
I
noticed
jamie
you.
G
You
raised
a
discussion
item
as
well
in
terms
of
how
do
we
help
people
get
started,
because
we
had
a
discussion
that
the
docs
are?
Okay,
I
o
are
great
once
you've
got
over
the
hurdle
of
doing
your
first
install
and
you
understand
all
the
terminology,
but
when
you
come
into
it
fresh,
the
first
time
you
just
hit
with
a
wall
of
acronyms
and
and
choices
that
you
may
not
be
able
to
actually
make
a
an
informed
decision
about.
So
is
that
the
role
of
the
guide?
G
So
so
we
had
a
discussion
about
what
should
the
guides
be
and
a
lot
of
them
are
out
of
date.
So
I
mean
there
is
the
one
about
the
the
unofficial
snow
face
back
in
4.7
time
frame.
We've
now
got
the
official
snow.
So
again
things
need
to
be
updating,
so
there's
some
work
to
be
done
there
and
we
need
to
get
that
done.
G
We
then
had
quite
a
discussion
with
the
with
the
work
going
on
with
things
like
the
discussion
from
brian
cook
last
week
on
the
what
is
it
called
the
red
hat,
hybrid
application
cloud
being
an
early
adopter
and
then
there's
also
the
possible
operate
first
cloud
and
the
community
having
a
build
environment
outside
the
red
hat
firewall,
and
we
talked
about-
which
is
the
next
item
so
I'll
cover
this
now
as
well.
The
operators
as
a
community.
G
We
are
looking
to
build
an
okd
catalogue
and
red
hat
engineers
haven't
found
the
time
to
actually
go
and
create
that
for
us,
so
as
a
community
we're
looking
to
enable
a
community
to
build
that
catalog
of
red
hat
operators
as
they
appear
in
the
ocp
product,
that'll,
be
a
community
built
catalog
and
so
we're
looking
at
what
is
the
overall
governance?
We've
now
got
our
okidi
project
repo
and
I
think
christian's
gone
and
created
a
key
to
io
and
project
that
we
can
use
to
actually
host
images.
G
So
it's
all
around
the
project
management,
the
governance
and
we've
heard-
and
I
certainly
experienced
it
when
new
people
want
to
join
the
community,
it's
difficult
for
them
to
find
out
how
they
can
contribute,
what
needs
doing.
We
don't
have
an
activity
or
a
project
board,
so
it's
quite
difficult
to
actually
join
the
community
and
feel
it
can
be
useful
and
where
you
can
be
useful.
G
I
was
lucky
that
I
could
offer
the
nk
docs
conversion
and
that's
how
I
sort
of
got
into
the
community
and
then
once
you
get
to
know
people
it's
easier
to
to
be
feel
that
you're
part
of
the
community,
but
there's
no
way
of
actually
doing
that.
A
lot
of
projects
have
the
good
first
task
or
good,
first
activity
flag.
So
we
need
to
look
at
things
like
that.
G
So
I
have
put
a
discussion
up
today,
I'm
going
to
put
a
link
in
the
chat,
and
we
can
also
put
that
in
the
the
minutes,
and
I
really
just
want
to
have
a
discussion
and
get
people's
ideas
just
to
work
out.
How
do
you
want
to
do
this?
G
We're
also
aware
that,
although
we
do
have
quite
a
big
community,
we
are
a
very
small
group
of
active
non-red
hatters
within
the
community
that
actually
do
things,
and
so
we
also
need
to
be
careful
that
we
don't
take
on
too
much
or
we
actually
try
and
build
that
community,
and
so
a
lot
of
discussions
to
be
had
around
the
community
governance
and
then
also
with
the
operators
and
last
week
was
red
hat
project.
I
forget
what
the
official
term
is
and
christian
found
a
team
that
actually
went
and
created
a
text
on
pipeline.
G
I
Okay,
so
brian
I'm
curious
a
little
bit
like
I'm,
I'm
intensely
curious
about
helping
to
bridge
the
gap
for
developers
who
want
to
contribute
to
okd,
and
so
I'm
curious,
like
what
could
we
do
to
help?
I
realize
getting
started
can
mean
different
things
for
different
people,
but
like
what
could
I
do
or
what
could
we
do
as
red
hatters
to
help
in
creating
materials?
For
like
that,
you
know,
I'm
a
developer.
I
I've
come
to
opec
okd
like
what's
my
next
step,
because,
like
it's
going
to
be
a
really
long
time
before,
we
have
a
board
that
says:
oh
here's,
a
good
issue
you
could
look
at
like
you
know
how
it
is
all
the
projects
are
separated.
We
don't
do
our
planning
in
github,
so
the
question
is
like:
is
there
material
or
could
we
create
like
a
video
series,
or
could
we
could
we
do
something
where
we're
helping
to
show
users
like
this
is
how
you
set
up
okd.
G
G
I
did
my
first
install
by
looking
at
a
video
guide
and
it
was
the
easiest
way
for
me
to
get
going.
So
I
I
think
things
like
that
are
a
great
help.
I
think
the
operators,
because
getting
started
now
once
you've
got
the
core
platform
you
end
up
with
this
you're,
then
stuck
because
if
you
want
to
do
techtone
or
githubs
or
run
the
code
ready
workspaces
or
whatever
they
call
debt
workspaces
now
or
you
want
to
do
some
storage
or
you
want
to
have
an
s3
bucket.
G
These
are
all
operators
which
are
there
for
for
ocp,
but
at
the
minute
we
don't
have.
So
I
think
that's
another
sticking
point
that
yeah
I've
got
my
first,
my
first
cluster
up
and
running,
oh,
but
everything
I
want
to
do
on
it.
There
isn't
the
operator
that
I
need
to
actually
make
it
that
easy.
So
I
think
that's
where
the
community
operator
comes
in,
I
have
created
a
technical
documentation
section,
so
that's
where
I'm
trying
to
pull
the
things
together
in
terms
of
if
you're
a
developer,
and
you
want
to
help.
G
How
do
you
actually
build
an
operator?
How
do
you
build
a
bit
of
how
do
I
do
a
bit
of
debugging?
How
do
I
do
this?
So
if
you
can
help
with
that
technical
section
now,
I've
been
bugging
christian
for
actually
putting
the
the
docker
files
for
build
images,
so
someone
can
actually
see
what's
in
it
and
how
it
works.
So
we've
now
put
it
that
in
the
opd
project,
repo
so
trying
to
expose
resources
that
are
hidden
or
obscure
in
terms
of
finding
what
they
are
within
prow
and
the
the
internal
ci.
G
I
mean
all
the
transformations
that
go
on
there.
It's
really
difficult
to
actually
understand
what
a
build,
how
a
build
works.
So
anything
you
can
do
to
help
there
contribute
documentation
is
probably
for
a
technical
point
of
view
and
yet,
within
the
documentation
group,
we
do
need
to
look
at
that
getting
started
experience.
C
B
I'm
not
I'm
not
muted,
so
austin
mcdonald,
who
is
leads
the
operator
sdk
work
and
is
basically
the
de
facto
community
manager
for
the
operator
framework
from
the
red
hat
side
of
things,
and
I
have
been
talking
about
not
just
doing
an
okd
s
cause
hackathon,
but
also
doing
an
operator
a
hackathon
on
this,
and
I
think
that's
what
this
tekton
pipeline,
that
they
did
during
what's
called
shift
week
last
week
is
a
starting
point
for
so
one
like
we
did
years.
B
It
feels
like
years
ago
for
the
deployments
and
the
testing
and
getting
the
videos
out
of
that
trying
to
put
together
this.
I
call
them
hackathons
and
probably
misnomers,
but
like
the
series
of
like
this
is
how
the
tech
tongue
pipeline
works.
B
So
that
they
could
explain
what
it
was,
do
the
okd
folks
people
could
answer,
you
know,
ask
questions
of
them
and
maybe
connect
the
dots
and
maybe
hopefully
get
those
maintainers
of
those
operators
actually
own
some
other
process,
or
at
least
the
feedback
on
that,
and
that
was
the
thinking
and
then
I
can
edit
the
chisel
out
of
all
those
videos
and
get
them
up
and
you
can
use
them
on
the
in
the
documentation
and
blog
posts
that
that
was
my
that's
the
best
game
plan.
A
A
This
is
where
you
know,
people
contribute
or
whatever,
and
actually
just
have
that
as
a
as
a
video
that
we
feature,
maybe
on
the
website
or
something
so
that
there's
a
walkthrough
of
what
the
community
even
is
without
having
to
read
through
a
paragraph
and
then
try
to
follow
all
the
resources
and
whatever
so
I'd
be
willing
to
do
something
like
that
to
get
us
started
in
the
video
realm.
I
think
video
is
the
place
to
be.
B
A
B
If
you
want
to
think
about
what
the
outline
of
that
would
be-
and
maybe
I
mean
I'm
happy
to
edit
the
chisel-
if,
if
you
want
to
do
it
at
night,
you
know
we
can
interview
it,
we
can
do
an
interview,
style
or
just
talking
style
and
make
it
I've
done
a
couple
of
what
is
okd's
in
the
past
or
different
events,
and
you
know
down
to
four
or
five
minutes,
but
I
think
a
walk-through
of
even
you
know
flip
sharing
your
screen
showing
okay.
This
is
where
the
documentation
is.
B
I
know
this
is
confusing.
These
are
guides.
This
is
the
blog.
This
is
the
dot
io
docs
that
sort
of
thing,
and
you
know
we
can
do
a
couple
takes
of
it,
and
I
can
edit
it
happy
to
do
that
I'll.
B
Love
it.
If
it's
not
me,
that's
the
talking
head
that
it's
an
external
participant,
so
that
would
be
great
for
me.
A
But
one
thing
I
would
say
is
that
it
might
be
good
to
get
someone
who
isn't
me
and
who
isn't
maybe
the
people
on
this
call,
because
I
don't
know
if
you
noticed,
but
if
you
look
at
there's
a
there's,
a
common
denominator
in
in
participation
of
of
this
working
group,
meaning
there's
a
lot
of
people
who
are
all
sort
of
the
same
from
the
same
event
right,
and
I
think
it
would
be
nice,
particularly
if
we
could
get
you
know
some
people
from
from
different
perspectives
and
also
maybe
get
more
of
a
of
a
female
presence
in
the
community.
B
Yeah,
no,
I
think
we
could
find
somebody
and
if
we
build
a
script,
but
if
we
to
do
a
first
take
and
walk
through
jamie
as
you
as
the
the
chair
of
the
working
group
right
now
might
be
nice
to
have
the
first
one,
and
then
we
can
do
little
side
vents
here
too,
but
I
do.
I
would
like
to
promote
awareness
of
you
being
our
point
person
now
too.
A
B
A
So
we've
got
some
great
ideas:
let's
make
sure
we
document
them,
and
I
want
to
acknowledge
that
brian
has
been
just
just
kicking
ass
and
doing
some
great
work.
Brian.
Thank
you.
So
much
for
you've
lifted
up
the
group
so
much
in
in
just
the
short
time
that
you've
been
involved
with
all
this
stuff
and
really
been
valuable.
A
All
right,
so
next
up
is
customizing
okd
and
if
vadim
is
still
on
vadim.
If
you
want
to
just
highlight
your
comment
that
you
put
in
the
notes
and
explain
this
a
little
bit
more
feel
free,
but
basically
it
was
sort
of
riffing
on.
So
how
do
you
know
what
commit
the
repo
that
the
repo
was
built
from
points
to
the
commit
that
this
was
built
from
when
you're
looking
at
source
repositories
for
images?
So
lithium
go
ahead
and
maybe
walk
us
through
that?
A
little
bit.
D
D
G
Just
so
you
know
that
is
that
that's
now
actually
any
docks.
Your
kitty
docks
I'll,
put
a
link
in
the
technical
section.
A
All
right,
well,
that's
good,
and
so
brian
has
something
up,
but
if
edem,
if
you
want
to
flesh
it
out,
that
would
be
fantastic.
If
you
have
the
time,
we
would
very
much
appreciate
that
so
basically
yeah.
There
is
a
way
to
find
out
more
details
about
where
it
came
from,
and
I
think
that
that
is
really
important.
G
I
think
so
just
the
other
point
that
jack
raised
last
time
was:
is
there
a
way
to
actually
link
a
feature
and
understand
what
I
see
on
screen?
Where
does
that
live?
What
what
operator
is
looking
after
that
bit,
because
I
think
there
was
also
within
cern.
They
had
some
issues
where
they
knew
what
they
wanted
to
change,
but
they
weren't
they
couldn't
figure
out
which
operator
or
which
bit
of
the
source
code
manage
that
bit
of
functionality.
G
So
I
think
that
was
one
of
his
other
which,
which
I
think
is
the
the
last
bullet
point
in
the
agenda,
how
to
determine
if
an
ocp
feature
is
in
okd.
I
think
that
was
linked
to
linked
with
that
in
terms
of
okay,
so
I
actually
understand
the
feature,
but
how
do
I
link
that
to
a
bit
of
source
or
a
module?
G
A
Yeah,
this
is
something
I
brought
up
at
the
meeting,
maybe
before
you
even
on
on
the
group
ride,
but
the
idea
that,
like
we
don't
have
anything
that
says,
oh
when
you
need
to
look
at
logging,
where,
where
are
the
logging
pods,
when
you
need
to
look
at
you
know
alert
manager?
Where
are
the
actual
alert
manager
pods?
So
you
can
look
at
logs
and
stuff
like
that
mike,
do
you
know
of
anything
or
vadim?
Do
you
know
of
anything
that
of
it?
Oh
video,
I'm
still
there.
A
I
We
I
do
not
think
there
is.
We've
talked
about
something
like
that
internally
and
you
know
for
those
of
you
of
a
certain
vintage.
You
might
remember:
microsoft
used
to
distribute
this
class
hierarchy
as
a
poster
when
you
bought
like
the
old
microsoft
foundation
classes,
we
talked
about
trying
to
make
something
like
that
with
okd
or
with
a
openshift,
where
you
could
see
what
all
the
operators
do
and
kind
of
like
which
parts
they
go
into,
but
I
don't
think
we
have
anything.
Currently,
that's
just
like
a
quick
reference.
I
Is
there
an
operator
that
owns
this
object
and
if
there's
an
operator
that
owns
this
object,
that's
probably
the
one
that's
doing
most
of
the
reconciliation
work
on
it
that
you
know
so,
like
it's
a
little
bit
right.
Now,
it's
a
little
bit
of
a
sherlock
holmes
kind
of
thing.
You
know
where
it's
like.
I
see
something
happening,
and
I
saw
some
object
that
I
didn't
recognize.
Okay,
where
did
that
object
come
from?
I
And
then,
if
you
look
in
the
fcd
records,
you
can
oftentimes
see
if
there's
an
owner
reference,
you
know
and
if
there's
an
owner
reference
that
gives
you
another
breadcrumb
to
start
tracing
back
but,
like,
I
think,
that's
the
manual
process
of
doing
it.
I
I
haven't
seen
anything
that
like
makes
it
explicitly
clear
where
all
these
things,
maybe
the
demands,
though.
D
No,
it's
it's
more
complicated
because,
for
instance,
scaling
up
new
nodes
is
shared
between
outer
scaler
machine
api
and
machine
config,
and
they
are
supposed
to
work
together
to
give
you
a
seamless
experience.
But
when
something
breaks,
machine
config
says
I
did
my
best.
I
generated
the
ignition
with
machine,
api's
fault
or
maybe
machine
as
a
scalar
didn't
work.
So
it
definitely
would
be
a
useful.
A
very
rough
map
of
what
alert
manager
is
in
monitoring
operator,
but
it
probably
won't
be
a
very
detailed
one
and
we
keep
on
adding
new
things.
D
So
it's
really
complicated
job.
We
definitely
need
what
would
be
useful.
Something
for
the
beginners
like
here
is
what
you
get
monitoring
solution,
including
alerting
observation
on
dashboards
from
easiest
and
high
level.
Overview
of
the
components
would
suddenly
be
useful
for
other
people
to
try,
but
mike
is
correct.
Once
you
get
on
a
more
complicated
level,
it's
operators
all
the
way
down.
Once
you
find
out
which
cr
is
responsible
for
this,
you
would
know
which
operator
is
reconciling.
B
D
C
I
Yeah,
just
to
add
a
little
to
what
vadim
is
saying
there,
because
vadim
mentioned
something:
that's
like
really
important
right.
He
mentioned
like
okay,
you
might
have
a
situation
where
the
cluster
is
growing,
because
the
auto
scaler
is
doing
something
with
the
machine
api,
which
is
doing
something
with
the
machine
config
right,
but
those
pieces
are
not
tightly
coupled.
So
the
key-
and
this
is
the
this-
is
something
we
probably
need
to
make
more
explicit.
The
key
here
is
kubernetes
right.
Kubernetes
is
the
data
store
you
know
at
cd?
A
An
idea
for
a
blog
post,
literally
just
exporting
oc,
get
co
to
a
table
and
then
just
have
a
description
column.
That's
just
like
hey.
What
does
this
operator
do
and
literally
just
give
someone
the
1200
foot
overview
of
like
what
the
operators
are,
that
get
installed
by
default
on
a
fresh
install
and
what
they
actually
do
all
right.
So
we
have
about
four
minutes
and
I
have
a
hard
stop
at
two.
Are
there
any
last-minute
thoughts
that
people
want
to
cover
before
we
sign
off.
G
Brian,
I
got
one
it's
mainly
because
charles
turned
up:
what
are
we
doing
about
crc
or
whatever
the
new
name
of
it
is
now
because
we
still
don't
have
a
solution
for
it.
F
I
haven't
touched
it
in
fun,
so
I
the
last
I
heard
from
the
the
actual
the
engineering
team
it
sounded
like
they
were
actually
getting
close
to
having
the
ability
to
build
an
okd
release
with
the
tooling
that
they
have
natively.
F
A
A
long
time-
and
I
wanna-
we
don't
expect
you
to,
and
we
appreciate
the
things
that
you
did
in
the
past
and
I
think
I
think
really
at
this
point.
We
need
to
go
back
to
that
thread.
The
engineers
in
question
are
actually
in
that
discussion
thread
in
the
repo.
Let's
talk
to
them
directly,
I
I
don't
want
to
burden
charo
with
with
this
any
longer.
You
sorrow
contributed
when
he
could
and
I
and
I
think,
let's
reach
out
to
those
engineers
who
are
in
that
threat.
G
A
All
right,
well,
documentation
group
will
make
a
final
decision
on
moving
forward.
I
think
so
all
right.
That
brings
us
to
the
end
of
the
meeting.
Thank
you
very
much
for
joining
us
and
be
sure
to
get
all
of
your
to-do
tasks
done
and
we'll
talk
to
you
online
soon.