►
From YouTube: OKD Working Group Meeting 04-26-2022
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
Welcome
to
the
okd
working
group
meeting
for
april
26
of
2022,
and
we
have
a
pretty
full
agenda
here,
so
we're
gonna
jump
right
into
it.
We'll
start
out
with
a
gender
review,
take
about
30
seconds
to
take
a
look
at
the
agenda.
A
Please
put
your
name
in
the
attendees
section
so
that
we
know
that
you
were
here
and
we
can
sort
of
keep
an
eye
on
if
we
need
to
reach
out
to
someone
who
wasn't
able
to
make
it
that
maybe
has
tests
or
maybe
will
be
affected
by
some
of
the
things
that
we
talk
about
so
we'll
get
started
with
some
release
updates
so
christian,
why
don't
you
go
ahead
and
take
it
away.
B
B
Regarding
the
question
that
came
up
that
just
came
up
earlier
before
the
recording
started
with
regrets
to
the
lcd
corruption
issue,
I
am
not
actually
sure
this
solves
it,
so
the
question
somebody
brought
to
me
was
whether
we
could
do
another
4.9
release
to
fix
the
issue
that
is
in
the
latest
on
the
last
4.9
release
that
we
have.
B
We
can't
actually
do
that,
since
we
have
this
rolling
release
model,
we
are
now
at
4.10,
so
we
can't
go
back
and
release
another
4.9
version
if
the
version
of
ncd
that
it
that
we
that
we
released
this
weekend,
though,
is
bigger
than
I-
and
I
forgot
the
specific
version
numbers
here.
But
if
it's,
if
it's
bigger
than
one
of
the
affected
versions,
it
should
be
safe
to
upgrade,
but
I
am
not
sure
whether
that's
actually
the
case.
So
if
there's
anybody
with
more
knowledge,
I'd
be
very
thankful
for
them
to
chime.
B
A
So
this
actually
came
up
as
a
question
in
the
channel,
and
I
couldn't
answer
it
directly
off
the
top
of
my
head.
How
can
we
check
the
xcd
version
either
from
the
command
line
of
the?
I
guess
we
could
log
into
a
node
and
do
lcd
control
right
to
get
the
version
that's
running?
Is
there
a
way
to
do
it?
Just
by
looking
at
the
manifests
or
the
containers
or
any
aspect
outside
of
an
existing
install?
Does
anyone
know.
D
A
That's,
I
think
something
we
should
figure
out
that
way.
We
can.
I
think
we
should
post
something
about
this.
The
documents
documentation
subgroup
can
take
this
up.
A
We
can
post
something
about
it
and
actually
say
hey
if
you
want
to
know
where
you're
at
these
are
the
versions-
and
you
can
always
test
yourself
by
looking
here,
see
which
version
you're
running,
because
that
email
today
I
mean
this
has
been
out
actually
as
an
issue
for
what,
like
three
weeks
now,
I
think,
or
almost
a
month
and
like
that
email
today,
just
everyone
just
went.
It
was
like
the
don't
panic
scene
from
airplane.
A
I
the
yes
it
is,
but
if
you
do
28,
I
think
it
is
or
whatever
just
came
out
it.
It
actually
is
the
edge.
So
take
a
look.
Actually
I
don't
have
it
in
front
of
me
but
yeah.
So
actually
it
is.
The
releases
do
open
it
up.
For
that.
F
I've
just
posted
the
I've,
just
posted
the
knowledge
page
article
in
the
chat.
We
can
also
put
that
in
the
the
notes,
but
it's
fcd
350-352.
A
Yeah
and
we're
talking
about
data,
corruption
and
essentially
not
all
of
the
members,
are
getting
the
updates
under
heavy
load.
A
So
yeah,
I
think
the
docs
group
should
should
take
this
up
as
something
to
let
folks
know
about
just
because
we're
gonna,
I'm
sure
we're
gonna,
get
questions
about
it.
Moving
forward
christian
anything
else
you
want
to,
let
us
know
about
in
terms
of
yeah.
B
One
one
noteworthy
thing
is
that
we
that
the
latest
release
now
fixes
or
moves
away
from
the
cryo
version
that
is
affected
by
the
tve.
I
don't
have
the
number
right
right
now,
but
there
was
a
cryo
cve
and
unfortunately
we
were
stuck
on
that
version
for
a
couple
of
weeks,
as
you
know,
and
that
is
fixed
in
the
new
release.
B
A
B
So
yes,
essentially
there's
a
few
things.
I
wasn't
able
to
mirror
out
the
images,
because
only
vadim
has
a
mirror
of
the
images
to
quay,
because
only
vadim
has
access
to
that.
That
is
fixable
from
our
sides,
though,
so
I
can.
I
should
probably
just
request
that
access.
It
wasn't
given
to
me
back
in
the
day
and
then
there's
the
the
other
issue
that
the
whole
build
system
is
essentially
owned
by
vadim,
so
it
runs
kind
of
in
his
name
space
using
his
his
secrets.
B
I
can
only
say
right
now:
there
are
going
to
be
changes
in
in
the
build
system
and
we
are
going
to
pull
that
into
prowl
back
when
we
set
up
the
system,
it
wasn't
possible
to
do
those
rpm
os,
3,
composers
within
pro,
that
is
entirely
possible
now,
and
we
also
have
something
we're
investigating,
which
is
the
core
os
layering
effort,
where
we
would
reuse
an
actual
f
cos,
release
and
just
layer,
our
things
on
top
and
thereby
preserve
the
actual
f
cos
versions,
because
we've
always
had
have
this
package
version
mismatch
mismatches
for
all
the
packages,
because
we
we
only
pulled
in
the
the
fedora
corvette's
manifest
and
then
did
a
new
compose.
B
You
already
knew
rpm
os
3
compose
for
for
okd,
and
with
this
new
concept
we
will
be
reusing
an
existing
compose
and
we
we
can
now
do
both
things
within
rci
within
pro
completely
new
composers,
and
also
these
layer,
layered,
composers
or
just
layering
of
an
existing
compose
and
we're
going
to
be
changing
towards
that
in
the
near
term.
B
A
B
A
A
E
There
were
a
couple
of
small
but
important
fixes
in
this
release.
One
there
was
a
kernel
issue,
so
we
had
to
actually
pin
the
near
kernel
or
pin
a
previous
kernel.
The
new
kernel
was
was
breaking
on
vmware
and
we've
also,
hopefully
permanently
until
we
have
a
real
solution,
fixed
caillou's
issue,
with
the
search
dot
issue.
E
We
put
that
into
a
spot
where
that
will
always
get
resolved
properly.
It
was
a
timing
issue
before
we
had
a
five
second
timer
in
there
and
with
410
it
looked
like
it
was
acting
between
four
and
six
seconds,
so
sometimes
it
would
work.
Sometimes
it
won't
well
we're
able
to
take
that
sleep
out
and
put
it
into
a
proper
spot.
So
now
it
happens
after
networking
is
up
and
it
just
does.
Its
thing
have
not
been
able
to
make
it
break
yet.
A
John,
can
you
drop
that
issue
in
the
meeting
notes
for
folks
that
aren't
sure
what
we're
talking
about?
Basically
there's
a
a
dns
search
issue
because
of
the
three
different
competing
systems
for
name
lookups,
on
fcos
or
on
fedora,
and
so
there's
things
that
get
written
to
resolve
and
that
needed
to
happen
after
other
things,
and
so
information
was
not
found
in
terms
of
external
domains.
E
Ahead,
jeff,
there
is
one
outstanding
thing
that
will
be
fixed,
probably
with
the
next
release.
We
we
found
that
there
is
an
issue
with
the
installer
and
that
will
be
the
installer
with
creating
a
bootstrap
node
and
not
being
able
to
resolve
api
int.
E
B
B
B
Those
are
exactly
the
two
machine
configs
we
are
adding
in
our
custom
okd
build
process,
though
it
yeah
and
I
don't
know
why
they
sometimes
get
picked
up,
and
sometimes
they
don't,
but
yet
also
seems
like
a
timing
issue.
They
seem
to
be
created
to
at
too
late
a
stage
where
the
mco
or
the
mcc
has
already
begun,
which
is
the
machine.
Copy
controller
has
already
begun,
rendering
that
that
configuration
and
therefore
doesn't
include
those
two
yeah,
not
sure
how
are
we
yeah?
B
I
I
don't
really
know
the
way
forward
or
what
the
fix
would
be.
Maybe
it
just
goes
away
with
our
new
build
system,
I'm
kind
of
hoping
hoping
that
but
yeah,
I
don't
really
know
what
causes
it
it's
it's
it's
definitely
specific
to
our
builds,
because
it's
it
only
really
yeah.
It
only
happens
with
our
two
right.
B
I
did
look
at
that
issue
again
and
I
wasn't
sure
there's
no
bugzilla
linked
to
it.
I
think
it
would
be
a
good
idea
to
open
a
bugzilla
and
put
that
on
on
the
machine
config
operator
team
to
I
know
they
they've
already
been
tagged
on
the
github
issue,
but
if
they
have
a
bugzilla,
they
really
have
something
to
to
look
at,
and
they
have
to
look
at
that.
E
B
E
Yeah,
if
he
can't
do
it
I'll,
do
it
and
I
I
haven't,
looked
at
it
yet
there
is
a
an
issue:
that's
going
to
hit
us
with
411
unless
the
fix
comes
again
with
the
installer
there's
something
they've
gone
to
a
new,
a
non-forked
version
of
some
of
the
installer
code,
and
that
is
causing
issues
with
bootstraps
being
deleted
or
some
issues
in
there.
So
that's
out
there
also.
B
And
about
just
about
that
issue,
a
yeah,
a
colleague
of
mine
on
my
team,
is
looking
just
looking
at
that
as
we
speak,
and
I
I
think
will
probably
be
fixed
before
4.11
is
released,
because
that
will
also
affect
effect
openshift
compute
platform,
the
product
that
will
probably
go
away.
E
Yeah
he
reached
out
to
me
earlier
today,
so
I
gave
him
some
information,
so
hopefully
that'll
help.
That's
just
and
the
other
thing-
and
I
know
bruce,
is
interested
in
this.
They
may
have
a
fix
for
the
stuff,
that's
ffs
stuff,
I
guess
they're
testing
it
in
develop
right
now.
Hopefully,
they'll
push
it
to
stable,
but
what
I
don't
know
is
how
long
it
takes
for
that
to
get
to.
A
Dusty
can
probably
speak
to
that.
Let's
let
I
know
dusty
came
in
sort
of
dusty's
on
a
limited
time
schedule,
so,
let's
actually
transition
quickly
to
f,
cos,
updates
and
then
bounce
back
to
anything
else,
just
so
that
we
can
let
dusty
take
care
of
other
things
dusty.
Any
f-cos
updates.
D
Hey,
I
don't
have
anything
in
particular
we're
getting
pretty
close
to
the
fedora
36
release,
so
at
that
point
our
testing
stream
will
get
moved
over
to
fedora
36
and
then
shortly
after
that
staple
so
christian,
I
don't
know
if
we
have
any
like
advanced
testing
of
like
next,
but
that
would
be
useful
if
we
find
issues
there
to
report
those
and
see
if
we
can
figure
them
out
early
on,
rather
than
you
guys
having
a
freeze.
B
Yeah,
we
do
have
that
one
branch
that
builds
off
of
the
the
devel
fcos
develop
so
we'll
we'll
probably
get
it
built
from
that
yeah.
D
Cool
yeah
and
just
to
mention
one
of
one
of
the
changes
there
was
iptables
moving
to
nf
tables
by
default,
which
shouldn't
be
an
issue
because
red
hat
core
os
or
rel8
made
that
change
some
time
ago,
and
then
there
was
what
else
was
there?
We
have
podman
4.0
in
fedora
36,
which
had
some
breaking
changes
to
the
podman
api.
D
I
don't
remember
if
okd
ocp
still
uses
podman
for
anything
at
all
these
days.
I
know
early
on.
There
was
some
you
know
pulling
from
a
registry
type
stuff
that
was
happening,
but
I
don't
know
if
we
do
that
anymore,
but
that's
just
something
else
to
consider.
B
I
think
it's
still
potman
is
still
used
as
a
fallback
in
case
oc
isn't
there,
but
I'm
not
sure
in
which
cases
that
would
actually
happen,
and
I
I
hope
that
api
hasn't
changed
too
much.
Full
images,
yeah.
D
If
it's,
if
it's
just
a
scripted
change
like
if
it's
executing
podman
in
a
script,
it
should
be
fine.
The
only
thing
I
think
that
really
changed
in
a
breaking
way
is
if
you're,
using
the
podman
api,
so
not
the
docker
api,
which
podman
also
supports,
but
the
podman
specific
api.
There
were
some
breaking
changes
there.
A
Excellent
any
questions
or
comments
for
dusty
in
regards
to
fedora
west
stuff.
E
I
guess
this
is
the
christian
and
dusty.
I
mean.
I
know
that
we
have
the
kernel
pinned
right
now.
I
know
we
don't
want
to
keep
that
pin.
How
do
we
go
forward
in
in
testing
new
kernels
to
see
if
we
can
unpin
this
issue.
B
There
is
a
bugzilla
for
this
and
we
should
definitely
follow
what
what
happens
there-
maybe
they'll
patch
the
kernel
in
fedora,
maybe-
and
I
think
there
was
already
an
upstream
fix
proposed-
so
maybe
that
gets
back
parted
by
the
kernel,
maintainers
or
even
upstream,
so
we'll
we'll
just
have
to
wait
for
now.
B
Oh,
oh
sorry,
yeah
that
that
might
be
a
few
months
out
still
so
yeah,
I'm
not
sure,
maybe
that
we
can,
if,
if
there's
no
cbes
or
other
big
things
popping
up
in
the
meantime,
maybe
we
can
do
that.
Yeah.
D
Yeah
once
there's
a
fix,
we'll
fast
track
it
in
fedora
core
os,
but
like
it's,
not
it's
not
currently
pinned
in
fedora
core
os,
just
because
we
didn't
catch
it
before
it
hit
stable
and
because
people
are
working
on
fixing
it
upstream
we're
not
necessarily
chasing
the
fedora
kernel
maintainer
to
revert
that
commit
so
we're
hoping
that,
basically,
a
fix
comes
in
and
we
can
just
get
it
back,
ported
yeah
I
can
as
soon
as
it's
ready
I
can
test
it.
I.
A
See
me
drop
off
too
okay,
someone
to
bounce
back
to
speaking
of
kai.
A
He
brought
up
an
issue
with
the
registry
409
pointing
to
or
the
catalog
in
409
samples
catalog,
pointing
to
centos
images
that
were
that
they're,
it's
pointing
to
registry
redhat
registry.redhat.io
instead
of
quay
or
key
as
brian
would
say.
So
we
know
there
aren't
any
four
nine
updates
coming
out.
Does
anyone
know
if
410?
If
that
properly
points
in
the
correct
place,
I
haven't
checked
yet,
but
kai
pointed
out
that
the
centos
based
images
all
of
the
centos-based
images
point
to
non-existent
places
in
registry.red
hat.
I
o.
A
Okay,
all
right
man,
we
need
to.
C
A
A
Okay,
all
right,
let's
move
on
now
to
actually
brian
with
the
documents
updates.
F
Okay,
so
we
had
our
meeting
last
week
a
couple
of
things
so
we're
gonna
start
planning
to
move
to
the
new
git
organization,
the
okd
project,
git
organization,
and
what
we
want
to
do
is
you
want
to
plan
it
and
do
a
single
move,
so
we're
not
left
with
people
spanning
multiple
repos
and
we
get
all
confused.
So
we're
going
to
start
that
work.
We've
shut
down
the
community
repo.
Already
we've
moved
everything
out
of
there,
so
that,
hopefully,
is
gone
now
we
requested
it
be
deleted.
G
I
I
believe
it
was
christian,
I
think,
put
in
the
the
request
for
me.
B
F
In
the
works,
okay,
so
that's
going
to
mean
leave
the
main
okay
d,
repo
and
then
obviously
the
okay,
the
I
o
repo
that
will
move
across
to
the
new
home
at
some
point
and
we'll
keep
you
updated
as
we
sort
of
make
that
plan
and
we're
going
to
do
that
move.
We
are
working
on
the
technical
documentation,
so
you
want
to
add
a
new
section
to
ok
to
io,
which
is
covering
technical
documentation
in
the
tasks,
and
I
did
put
a
link
to
the
discussion
forum.
So
we
have
this
conversation
last
meeting.
F
If
you
have
any
best
practices
tips,
if
you
know
of
any
sort
of
obscure
places
within
the
muscada
logs,
where
it's
useful
to
go
check
and
any
little
sort
of
magic
scripts
that
you've
got
to
help
work
out
when
things
go
wrong,
if
you
can
just
dump
them
in
that
discussion
thread,
I
again
the
links
in
the
in
the
in
the
tasks
at
the
bottom
of
the
of
this
meeting,
and
then
the
docs
groups,
probably
me,
will
render
that
into
something
worthwhile
on
the
okd
site
and
also
I'm
still
working
through
a
a
work
sample
of
how
to
compile
modules
and
john
gave
me
some
very
useful
information.
F
So
I'm
working
through
that
we'll
write
that
up
and
we'll
also
get
that
pushed
there
at
some
points.
Christian,
as
we
sort
of
move
the
build,
we
do
want
to
sort
of
write
up
some
things
like
the
machine
config
operator
and
some
of
the
other
sort
of
bits
of
magic
that
form
up
a
release
just
to
help
the
community
understand
how
it
all
works
and
and
when
they're
looking
at
the
repos
and
help
them
do
that.
F
So
that
is
sort
of
a
long-term
project
to
actually
get
some
technical
documentation
there
and
hopefully
keep
it
updated.
As
as
the
project
moves
on
regarding
okay
leader
io,
we're
going
to
reorganize
the
working
group
site,
we're
going
to
make
the
this
main
group
and
more
prominent
and
we're
going
to
rename
all
the
subgroups
as
subgroups
and
they're
going
to
be
in
a
separate
section
as
subgroups.
Just
so.
It's
very
clear
that
this
is
the
primary
group
and
then
there
are
sub
groups
underneath.
So
we'll
just
just
be
tidying
that
out.
F
And
I'm
going
to
pass
it
on
to
jamie
to
talk
about
the
sub
chairs,
because
we
need
to
do
one
thing
to
get
our
become
in
line
with
what
we
put
in
our
sort
of
code
of
conduct.
A
Right
so
in
terms
of
the
subgroup
chairs,
I
wanted
to
be
able
to
make
it
a
public
like
hey
the
okay
word
group
is
doing
this
and
I
wanted
to
be
able
to
have
post
something
on
twitter.
That
says
the
okd
working
group
is
doing
this.
It
turns
out.
I
don't
have
everything
I
need
to
get
into
the
twitter,
so
I'm
going
to
work
with
diane
on
that
it's
tomorrow
it
turns
out.
A
I
need
an
email
address
for
verification,
so,
but
once
we're
able
to
communicate
out
that
we
are
actually
having
a
process
of
defining
shares,
then
we'll
move
forward
with
that
it
hasn't
been
a
big
rush.
Until
now,
I
didn't
want
to
rush
it,
because
what
I
want
to
do
is
make
sure
that
we
let
the
community
know
that
we
are
a
as
much
as
possible,
open
and
and
democratic-ish
community,
and
let
people
know
that
that
their
participation
is
appreciated
and
that
there's
a
path
for
them
to
get
involved.
A
And
this
factors
into
the
larger
discussion
of
like
outlining
a
path
for
involvement
which
we've
talked
about
before
so
once
the
twitter
stuff
is
cleared
away
in
it,
and
we
can
actively
like
promote
that
we're
following
this
process
then
and
that'll
be
in
the
next
couple
days.
Then
we
will
actually
do
that,
and
the
other
thing
is
I'm
still
behind
in
posting
videos
it.
A
F
A
Yes-
and
I
think
it
might
be
something
that
we
look
at
to
see
if
we
can
get
a
a
copy
yeah,
oh
yeah,
go
ahead.
G
I'm
trying
to
find
a
raise
your
hand
button,
but
I
can't
find
it
on
my
screen.
I'm
I'm
trying
there's
a
group
inside
of
red
hat.
I
think
I
mentioned
this
on.
The
docs
called
called
operate
first
and
they
have
a
cloud
called.
I
think
it's
based
on
mass
open
cloud,
it's
in
massachusetts
somewhere
and
they
have
some
cloud
computing
resource
and
some
bandwidths
and
they
like
building
the
icd
stuff.
So
I've
asked
them
and
shared
with
them.
G
Charo's
build
upstream
without
a
paddle
blog
post
and
asked
them
to
take
a
look
at
trying
to
recreate
that
on
operate
first,
and
so
I'm
going
to
try
and
get
them
to
come
to
the
working
group
and
talk
about
their
adventures
there.
G
But
what
my
goal
is
is
kind
of
using
this
as
a
poc
with
this
group
to
see
if
we
can
do
if
we
can
get
a
community
build
process
for
crc
on
this
cloud,
that
we
can
help
manage
and
and
automate,
and
that
I
I
was
hesitating
at
first,
because
it's
at
boston
university
thanks
like
that
that
I
was
hesitating
at
first
because
I
didn't
think
there
was
that
much
uptake
on
crc.
G
I
didn't
want
to
spend
bandwidth
on
that,
but
I've
been
told
that
there
are
people
actually
using
it
and
who
want
it.
So
we'll
see
we'll
use
that
and
then
maybe,
if
it
all,
goes
well,
we
might
even
be
able
to
build
a
community
build
process
for
okd
on
that
cloud
too,
of
some
ilk.
G
So
I
I'm
just
just
so
I'm
being
transparent
about
what
I'm
doing
in
the
background
here,
trying
to
find
some
additional
resources
and
the
compute
resources
to
do
the
crc
and
to
do
a
community
build
of
okd
on
f
class
sometime
in
the
not
too
distant
future.
If
that
works
out,
I'm
not
positive
that
the
okay,
the
operate
first
cloud
is
optimal.
For
building
it,
but
I
thought
the
crc
would
be
a
first
good
test.
G
So
there's
your
quick
update
and
I
will
now
go
back
and
hunt
them
down
and
see
if
they
got
any
any
further
than
I
in
the
past
week,
since
I
last
talked
to
them.
A
Fantastic,
thank
you
diane.
What
does
this
group
think
about
tasking
the
documentation
subgroup
with
maybe
moving
charo's
document
into
the
okd
space
so
that
it's
actually
an
okd
document
and
we
can
build
and
update
it
without
rely
having
to
rely
on
charo
charo's
incredibly
busy
these
days,
and
I
think
the
more
that
we
can
make
it
institutional
knowledge
versus
an
individual's
knowledge
it
would
benefit.
Does
anyone
see
any
downsides
to
that.
A
Yeah
one
of
the
things
that
I
talked
about
as
we
were
going
through
the
guides
and
those
folks
aren't
here
today,
but
the
folks
that
are
going
through
and
helping
get
the
the
the
guides
updated.
A
A
You
know
it's
its
freshness
and
verify
that
it's
not
stale
and
that
we
do
that
on
a
regular
process.
So
a
lot
of
groups
I'm
involved
with,
will
do
that
with
their
policies,
their
guidelines
or
whatever.
I
think
we
could
do
that
with
our
documentation
is
just
create
a
calendar
and
get
volunteers
right
and
if
we
yeah
brian.
F
Yeah
something
another
part
of
that
is
a
few
meetings
ago.
We
talked
about
automation,.
G
F
We
can
automate
a
lot
of
the
processes
that
are
within
our
documentation.
Then
we
get
kicked
by
a
failure
notice
if
our
documentation
is
out
of
date.
I
think
that's
also
a
very
useful
way
of
of
keeping
documentation
up
to
date
and
just
make
sure
that
we
link
the.
If
you
fix
the
build,
you
got
to
fix
the
documentation.
A
F
So
the
two
stay
in
sync,
I
think,
that's
also
sort
of
reviving
your
project.
I
know
that
you
gave
us
the
link
to
some
scripts,
that
you'd
written
jeremy
just
reviving
that
project
and
actually
trying
to
automate
as
much
of
the
the
instructions
that
we
put
on
okdio.
So
we
know
when
it's
our
date.
A
All
right,
I
think,
that's
it
for
documentation.
Let's
keep
moving
on
so
discussion
items
are
there?
Is
there
anything
in
the
discussion
of
the
repo
discussion
section
of
the
repo
that
folks
want
to
pull
out
and
talk
about
or
think
need
special
attention?
E
Those
were
specifically
added
in
410
future
api
stuff
is
coming
out
and
they
got
to
get
off
of
it
before
they
upgrade
at
a
certain
point.
C
Right-
and
I
think
specifically
for
it
was
like
internal
resources
that
were
that
were
being
flagged,
and
he
was
wondering
if
there's
anything
you
should
do
to
fix
those.
I
don't
think
it
was
any
anything
that
any
like
user
workloads
which
yeah
I
think
we
just
have
to
hope
that
upstream
gets
to
them,
which
they
will
I'm
sure.
A
Yeah,
I'm
sure
they
will.
I
think
we
like
one
of
the
things
I've
listed
in
this
is
like.
Does
it
make
sense
in
the
meeting
notes?
One
of
the
things
I
put
in
the
meeting
notes
is:
does
it
make
sense
to
kick
this
to
the
documentation
working
group
for
us
to
come
up
with
our
own
little
explanation
of
this
and
maybe
help
folks
understand
what
is
an
internal
workload
and
what
is
an
external
workload,
because
the
red
hat
documentation
itself
doesn't
do
a
very
good
job?
A
I
notice
folks
are
running
the
command
and
not
really
clear.
It's
not
clear
what
they're
looking
at
right
and
so
I've
seen
several
comments
that
are
like
yeah.
I
got
the
warning
messages.
I
ran
the.
What
is
it
mcp
or
whatever
command,
and
I
don't
know
what
I'm
looking
at?
Does
it
make
sense
for
okd
to
write
something
up
that
sort
of
helps
our
users
sort
of
understand?
What
they're
looking
at.
A
A
Got
your
hand
raised.
B
Yeah,
I
found
the
button
somewhere
in
the
menu
it
for
we
do
actually
because
it's
a
separate
build
for
the
okd
docs.
We
can
actually
put
some
of
those
comments.
We
don't
need
to
put
them
on
our
okd
io,
specific
docs.
We
can
actually
put
them
in
the
main
docks
and
just
put
them
in
like
this.
If
statement
that
says
this
only
display
this
for
for
okd
there's.
B
Actually
we
have
this
one
docs
person
from
red
hat,
michael
burke,
who
has
been
helping
a
lot
with
with
our
issues
and
he's
now
I
think,
planning
another
pr
to
add
a
list
for
the
openstack
variant
that
just
yeah,
essentially
the
compatibility
matrix,
because
in
current
4.10
we
only,
I
believe,
support
open
is
that
red
hat
openstack
16,
which
is
openstack
train
upstream
and
he'll,
he'll-
put
that
that
list
for
for
the
upstream
version
names
in
there
too.
B
I
think
we
can
actually
use
that
more
heavily
and
just
put
okay,
these
specific
comments
in
the
actual
docs
and
then
and.
A
Yes,
we
can
definitely
can
kick
that
there
if,
if
folks
feel
that
that's
the
appropriate
place
versus
something
on
the
site,
would
it
be
sorry.
A
Well,
you
know
the
question
might
be:
is:
are
there
any
downsides
to
putting
it
into
the
product
document
work
so
christian
in
the
in
the
documentation
working
group
we
separated
into
product
documentation
versus
sort
of
community
documentation?
Are
there
any
downsides
that
anyone
can
think
of
of
putting
it
in
the
product
documentation
versus
the
community
documented?
I
don't
want
to
belabor
this,
but
I'm
just
curious
if
anyone's
using
it.
B
I
I
think,
if
you
have
like
specific
work,
arounds
or
commands
for
for
things
you
might
have
to
do
in
that
in
in
that
instance,
we
have
these
things
as
kind
of
warnings
or
notes
on
the
on
the
product
docs,
and
if
there
is
something
you
need
to
do
specifically
for
for
fedora.
B
In
that
case,
I
I
think
it
would
be
perfectly
fine
to
put
them
in
the
docks
as
well
and
then
later
on,
with
that
version
of
of
whatever
software
you
have
to
use
comes
into
the
product.
People
can
just
move
that
that
note,
so
I
think
it's
yeah
and
obviously
it'll
get
reviewed
by
the
product.
B
Folks,
all
the
you
know
the
the
docs
folks
that
maintain
that
repository
so
trying
to
move
as
much
valuable
information
into
the
docs,
I
think
is,
is
a
good
thing
and
since
we
can't
just
can
exclude
them
from
the
product
docs
by
by
putting
them
into
the
the.
If
statement,
I
think
yeah,
we
should
try
to
to
use
that
as
heavily
as
possible
for
for.
J
B
Obvious
obviously,
concise
notes
and
tips
and
warnings
and
whatnot
if
yeah,
if
you
think
it's
a
good
place,
I
think
you
should
try
to
put
them
in.
A
A
Oh,
what
is
it
what's
the
I've
seen
a
couple
of
these
float
by
about
f
costs
and
our
costs
getting
mixed
up
in
the
installer
for
bare
metal?
Does
anyone
know
about
that
and
what
what's
happening
there?
We
did
receive
an
issue
about
that,
and
someone
mentioned
in
the
channel
about
this,
but
I
don't
know
anything
about
it.
Has
anyone
seen
this
where
a
bare
metal,
okd
install,
pulls
up
archos
notes.
E
A
Exactly
I
I'm
at
a
loss,
but
they
were
asking
in
the
channel
and-
and
I
think
it's
the
same
person,
but
I
can't
be
sure
because
they
have,
I
think,
slightly
different.
That's
me
actually.
J
Okay,
yeah
yeah.
I
don't
I
don't
really
know
what's
happening
other
than
I
extract
the
bare
metal
ipi
installer
from
the
okd
release
that
the
previous
one,
not
the
really
recent
one
and
yeah.
K
K
E
So
you
know
that
other
pr
that
I
have
out
there
about
making
the
the
installer
artifacts.
E
B
I
think
I
think
it's
I
I
think
it
might
make
sense,
though,
because
I
think
we
do
use
different
containers
for
the
different
release,
artifacts
and
you're,
trying
to
to
pull
the
openshift
bare
metal,
install
binary
specifically,
and
it's
it's
a
command
that,
like
this,
I've
actually
never
used
it
myself.
B
I've
only
just
you
know,
pulled
the
openshift
installer
from
it,
but
not
bare
metal
install,
and
I
think
you
might
be
right
that
we
aren't
actually
propagating
the
okd
environment
tag
environment
variable
into
the
other
artifacts
build
containers
and
just
the
main
one-
and
I
think
that
is
the
issue
you
yeah-
changing
it
in
the
docker
file
might
actually
be
might
actually
be
sufficient
here,
which
is
the
pr
you
you
opened
yeah
right
right.
A
Yeah,
if
you
could
put
that
into
the
meeting
notes,
that
would
be
great,
and
is
it
eric?
Thank
you
for
bringing
that
to
our
attention.
This
is
very
helpful
to
know.
J
Yeah,
because
the
bootstrap
node
that
gets
set
up
as
a
virtual
machine
on
the
provisioner
thing
that
installs
with
fedora
chorus,
but
the
fedora
coros
whenever
my
machine's
pixie
boot
they're
getting
served,
read
that
chorus.
J
Also,
since
I
have
you
here,
do
you
know
how
I
can
modify
the
like
send
something
in
that
would
allow
me
to
get
serial
output
while
I'm
installing
the
bare
metal
nodes.
F
B
That
would
be
a
good
question
for
dusty
and
I
think
you
might
want
to
pass
a
kernel
argument.
J
A
So
that
happens,
yeah
that
would
be,
you
could
do
it
with
a
with
a
k
arc,
probably
passed
in
for
upi
for
ipi.
I
don't
know,
tell
you
what
we
will
look
it
up
and
find
out
and
we'll
put
it
did
you
open
a
discussion
item
on
this
one
on
the
site,
or
did
you
just
mention
it?
A
I
saw
your
question
in
the
channel,
but
can
you
open
up
a
discussion
item
on
the
okd
repo
and
if
you
put
that
discussion
item
there
we'll
respond
to
you
with
a
response
we'll
reach
out
to
dusty
and
also
find
out
if
this
is
something
that
that's
doable
for
ipi
or
probably
just
upi,
but
we'll
check
and
see.
I
B
B
All
right,
and
if
it's
not
in
the
manifest
already,
you
can
do
it
via
ignition
and
it
should
at
least
follow
the
second
boot
or
for
the
boot
into
the
actual
real
root.
Then
have
that
kernel
arc
applied
right,
which,
which
still
is
a
bit
natural
in
in
okd,
because
we
first
boot
a
standard,
fedora
core
os,
and
then
we
pivot
into
our
machine
os.
So
yeah
your
experience
may
vary
here.
I
think,
but
we'll
definitely
find
that
out.
I
think
that's
a
interesting.
B
Yeah
you
can,
you
can
specify
it
through
machine
config,
but
that
actually
isn't
that's
not
a
pure
ignition.
The
mco
currently
consumes
it's
an
mco
api,
so
the
mc
yeah,
the
machine
config
operator,
will
consume
that
and
then
the
mcd
will
apply
it
and
reboot
the
machine.
So
you
don't
get
it
from
the
from
the
get
go
like
you
would
with
natively
right
right,
but
there
is.
There
is
a
juror
card
for
actually
making
use
of
the
native
ignition
api
to
have
that
come
up
right
away
with
the
kernel
heart
applied.
K
G
A
All
right
I'll,
just
pop
that
there
all
right
well,
thank
you
so
much
for
bringing
this
to
our
attention
and
we
will
do
our
best
to
get
these
addressed
and
get
the
information
to
you
posthaste
and
thanks
for
your
patience,
while
we
figure
this
out
got
two
very
good
things
to
our
attention.
So
much
appreciated,
and
thanks
for
coming
to
the
meeting
too,
it's
nice
to
have
some
folks
showing
up
here
that
isn't
the
usual
crowd
all
right.
A
Moving
on
to
the
survey,
so
I
did
have
a
chat
discussion
with
3t
night
before
last.
A
Things
got
sidetracked
and
so
we're
going
to
get
a
copy
of
that
survey,
revamp
it
and
then
send
it
out
to
our
users
to
get
a
sense
of
what
they're
doing
and
how
they're
using
okd
and
how
they'd
like
us
to
what
they
might
look
for
in
okd
in
the
future
and
how
the
community
can
grow
and
change
to
meet
their
needs
as
a
community
open
source
community.
A
Any
questions
or
comments
on
that.
Any
feedback
on
that.
A
All
right
we
did
the
project
updates
we're
pretty
close
to
the
end.
If
folks
could
add
task
items
in
there
so
that
we
remember
all
of
our
tasks.
That
would
be
helpful,
I'll,
try
to
put
in
a
couple
too,
but
if
folks,
who
put
in
task
items
that
will
help
us
remember
all
the
things
that
we
said
that
we
were
going
to
do
here
and
I.
G
G
And
and
brian
ennis
I
will-
I
do
still
owe
you,
the
the
dublin
and
the
london
dates
for
those
and
if
anybody
else
is
in
the
uk
area,
we
have
two.
I
think
july,
6th
and
june
20
something
two
gatherings
there
that
I'm
looking
for
okd
collaborators
with
brian
to
give
the
community
update.
Otherwise,
it's
going
to
be
me
doing
the
song
and
dance
with
brian.
E
So,
by
the
way,
I
I'm
looking
at
that
installer
quick
segway
since
we're
running
out
of
time.
That
is
a
that
is
an
ftos
installer.
I
don't
see
anything
in
it
that
references
route.
E
A
E
A
This
a
shot-
if
someone
else
can
can
contest
this,
and
maybe
we
can
get
a
couple
folks,
testing
it
and
then
narrow
this
down.
That
would
be
awesome,
so
we
need
some
bare
metal
testing.
I
might
be
able
to
throw
it
on
a
machine
and
test,
but
if
anyone
else
can
that'd
be
awesome.
I
On
the
other
side
of
this
I'll,
try
to
give
a
review
on
that
installer
pr,
john.
Just
for
what
it's
worth
our
the
installer
team
has
been
under
like
a
lot
of
pressure.
We
know
that
they
have.
They
have
not
necessarily
had
a
lot
of
spare
cycles,
just
to
look
at
pull
requests,
but
I'll.
Try.
E
To
yeah
it
was
a
minor
one.
For
me
I
was
like
yeah.
You
know.
I
just
happened
to
notice
that,
because
going
through
bernie
had
to
do
the
build-
and
I
happened
to
see
that
it
wasn't
passing
the
tax
through
because
I
didn't
know
how
to
build
an
okd
installer.
You
know
until
vadim
put
that
information
out.
So
just
happened
to
notice
that
that
wasn't
being
passed
down
so
not
a
huge
deal.
You
got
the
important
ones
out.
I
A
Excellent
all
right
we've
got
a
few
minutes
left.
Is
there
anything
else
that
the
folks
would
like
us
to
talk
about?
Take
a
look
at
or
anything
else
before
we
sign
off
it's
an
excellent
meeting.
E
A
Yeah
a
lot
of-
and
I
think,
we're
a
large
in
part,
I
think,
due
to
brian's
documentation,
work,
john,
your
work
and
christian's
work
and
everyone
shipping
and
we're
starting
to
understand
the
nature
of
the
beast
better.
I
think
we're
actually
like
understanding,
okd
and
what's
underneath
this
in
terms
of
the
build
process
in
terms
of
the
installer
et
cetera,
and
I
think
that
that's
fundamental.
E
Well,
testing,
too,
I
mean
looking
at
looking
at
the
pro
results
and
digging
into
when
they
do
the
tests
and,
seeing
you
know
what
in
the
world
you
know
is
proud
doing,
and
the
results
that
they're
getting
digging
into
that
is
almost
as
bad
as
digging
into
a
log
into
a
lock
bundle.
E
B
We
should
do
a
a
deep
dive
into
the
prowl
artifact
sometimes,
and
I
I
do
hope
that,
with
the
with
the
change
in
build
system
that
will
will
soon
have
with
pulling
it
into
pro
natively,
not
shelling
it
out
to
cirrus
ci.
A
Christian,
who
do
you
think,
is
the
best
person
if
we
were
to
have
like
a
like
a
session,
a
meeting
dedicated
just
to
exploring
prowl,
who
would
be
the
best
person
to
lead
that.
B
I
I
think
we
have
any
any
of
you
red
hat
employees
here
have
have
knowledge
of
most
of
them.
I
would
say
that
system
and
they
would
bring
a
piece
to
that
meeting
and
the
experience
you
look
at
different
things
for
different
components.
Obviously-
and
I
I
don't
want
to
like
pull
in
a
member
from
from
the
test
platform
team
right
now-
who
isn't
here,
who
could
probably
explain
it
best
but
yeah
because
they
are,
they
also
have
a
lot
of
things
to
do.
B
Yeah
so
we'll
put
it
on
your
shoulders.
There
is
a
there
is
a
they
actually
have
great
documentation
about
the
system,
and
I
think
it's
docs.ci.openshift.org.
B
And
I
haven't
checked
it
right
now,
but
yeah
that
should
be
it,
and
that
is
actually
a
canonical
documentation
for
that
system
and
it's
very
helpful.
I
I
think,
instead
of
you
know,
instead
of
trying
to
like
look
at
the
pro
system
and
understand
it
like
in
abstract,
through
like
one
presentation
or
a
series
of
presentations,
more
likely,
I
would
say
it's
probably
if
we
could
come
up
with
with
tasks
that
we
could
do
like
a
deep
dive
around.
That
might
be
the
most
beneficial
like
so
take
something
john
just
said,
you
know
if
we
took
the
notion
of
like
I'm
a
community
member
and
I'm
putting
a
pr
up,
and
this
is
what
I
see
happen
right.
I
How
do
I
then
investigate
what
are
these
tests
that
failed?
Which
ones
do
I
need
to
pay
attention
to?
How
do
I
find
out
where
it
failed?
How
do
I
debug
it?
You
know,
maybe
if
we
focused
on
it
like
that,
we
could
probably
easily
record
like
an
hour
session
where
someone
kind
of
leads
it
from
the
perspective
of
like
I've,
just
put
up
a
pr
and
here's
what
happens
next.
G
Yeah
and
their
time,
let's
christian-
and
I
can
talk
to
some
of
the
powers
that
be
too
and
see
if
we
can
find
someone
between
christian
and
I,
as
opposed
to
just
putting
on
christian
shoulders.
So
let
me
see
what
we
can
do
with
maybe
talk
to
michelle
creaky
and
others
and
find
some
folks
to
help,
because
I
think
training
up
the
community
on
this
is
a
huge
benefit.
So
let
me
see
what
I've
got.
A
Excellent,
thank
you
diane
all
right
folks,
I'm
going
to
hit
stop
on
the
recording
and
thank
you
very
much
for
attending
the
meeting.
Thank
you
eric
and
all
the
other
guests
that
contributed
as
well
and
so
next
week.
At
the
same
time
is
the
documentation
subgroup
meeting
and
then
this
meeting
happens
again
in
two
weeks.