►
From YouTube: OKD Working Group Meeting 09-14-2021
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
B
C
C
C
All
right,
let's
get
started
with
the
meetings,
so
vadim
go
ahead
and
give
us
the
latest
on
release
updates.
E
So
there
hasn't
been
much
happening
in
our
releases,
mostly
because
we're
blocked
on
well
multiple
ways.
The
most
pressing
issue
was
that
fedora
crosses
regressed
and
stable
and
testing
develop.
There
was
a
file.
There
was
a
bug
filed
for
the
request,
streams
related
to
network
manager
and
some
weird
systemd
setting,
which
prevents
it
from
booting
when
we
pass
the
master
ignition.
E
The
picks
for
this
is
to
revert
back
to
the
older
durocross
table.
That's
this
one
from
may
or
from
june.
E
I
have
updated
all
the
nightlys
for
for
eight
for
nine
and
four
seven
and
on
ci
for
seven
still
fails
on
vsphere,
so
we
would
need
manual
confirmation
before
we
cut
out
a
new,
stable
release
that
will
be
working
or
we
need
to
look
into
something
more.
E
This
is
one
of
the
reasons
why
we
didn't
cut
the
release
two
weeks
ago,
so
that
we
would
make
sure
another
news
is
that
kubernetes
would
be
rebased
and
for
aid
and
for
seven
branches,
we're
waiting
for
the
final
tax
to
be
added
to
the
prs,
and
hopefully
it
would
make
it
into
the
nitrous
and
we'll
cut
any
force
and
release
based
on
120
120
10.
I
think
the
largest
problem
is
probably
that
we
still
don't
have
a
safe
upgrade
path
from
47
to
48..
E
First
of
all,
there
are
multiple
issues
and
regressions
in
cryo
which
supposed
to
be
fixed
and
we'll
get
them
in
the
new
release
which
are
affecting
ocp
as
well,
and
in
our
case
we
have
a
bug
with
ovn
team,
where
the
applications
are
still
getting
disrupted
more
than
we
expect
them
to.
This
is
why
the
ci
cannot
create
an
upgrade
pass
and
we
have
no
annual
way
to
override
this.
E
E
We
created
an
okd
specific
version
for
ironic
ipa
downloader.
I
think-
and
we
have
great
chat
with
the
bare
metal
team,
and
I
think
we
had
a
progress
on
the
actual
ironic
image,
so
we
would
need
landed.
E
We
would
need
to
land
a
pull
request
and
a
promotion
vr
in
release
repo,
and
we
should
have
all
the
bits
and
pieces
necessary
for
the
night
list
once
we
confirm
that
all
of
these
are
ready.
We
can
port
them
back
to
four
eight
and
four
seven.
Four
nine
and
another
interesting
issue
with
fedora
was
that
they
have
changed
the
sea
linux
profile
rules
which
led
to
sugar
release
tests.
E
Failing
we
have
a
code
request
to
actually
fix
that,
but
until
then
we
cannot
upgrade
cylindex
pro
dynamics
targeted
packages
in
our
fedora
cross,
which
leads
to
inability
to
update
chilepsy,
which
leads
to
an
ability
to
update
a
bunch
of
packages.
One
of
them
is
network
manager
132.
E
So
at
this
point
we
have
to
revert
back
to
network
manager
130,
which
has
an
issue
with
reverse
pointer
resolution.
So
it's
an
unfortunate
casualty
while
we
figure
out
things
with
upstream
and
how
to
land
the
fixes
and
should
take
them
back
to
four
seven.
So
that's
something
I
will
write
a
known
issue
about
when
we
cut
out
a
new,
stable
release.
E
C
Excellent
any
questions
on
what
vadim
covered
there.
Anyone
have
any
questions
or
comments.
C
Looks
like
we're
good
for
that.
Thank
you.
Vadim.
Moving
on
to
timothy
with
fedora
quarter,
west
updates.
F
Just
that,
on
the
on
the
various
releases
under
the
upgrades
from
section
there
are
sometimes
some
individual
letters,
you
know
f
or
s,
which
presumably
means
something
failed
or
succeeded,
and
occasionally
there
are
four
of
them.
Sometimes
there
are
three
of
them
and
I
haven't
been
able
to
find
anywhere
what
those
actually
mean.
E
The
exact
job,
so
I
have
to
start
them
manually.
I
think
I
covered
this
previously,
but
I
can
get
more
into
detail.
So
the
way
we
create
an
upgrade
path
is
actually
make
ci
run
and
upgrade
from
one
release
to
the
other.
E
There
is
some
automation
which
runs
upgrades
from
previous
nightly
to
the
latest
nightly,
but
there
is
no
automation
which
runs
this
same
sort
of
jobs
on
stable
branches.
So
I
have
to
pick
these
myself,
usually
I'm
guided
by
latest
two
releases
and
the
most
popular
for
six
from
telemetry.
E
Yeah,
if
you
need
a
specific
stuff,
just
ping
me,
I
can
run
the
job,
but.
F
F
F
Jobs
that
I
can
probably
go
like
in
that
case,
the
the
the
change
log
makes
a
bit
more
sense.
So
thank
you.
E
Yeah,
I'm
trying
to
split
them
between
aws
and
gcp,
just
power
all
the
platforms.
We
probably
could
try
a
vsphere
upgrade,
but
the
pass
rate
would
be
very
low
again.
All
that
is
done,
health
manually
at
least
I
invoke
those
jobs.
You
cannot
affect
them
once
they
get
started.
C
H
Hey
hi,
I'm
tibo
te
raviere
from
the
chorus
team
at
redhat
working
on
federal
chorus,
mostly
so.
I've
linked
several
issues,
so
the
one
of
the
big
thing
we
fixed
recently
is
around
the
uefi
booting,
especially
the
live
iso.
If
you
had
troubles
in
the
past
booting
the
live
iso,
it
should
now
be
fixed
in
federal
course.
H
H
It
will
have
an
impact
for
the
first
boot
and
we
still
have
one
in
progress
regarding
five
members
configuration
two,
which
is
if
we,
if
you
had
issues
like
when
we're
installing
nodes
on
with
federal
crores
and
in
the
past,
if
you
have
some
kind
of
issues-
and
probably
this
is
something
you've
hit
before
so
take
a
look
and
chime
in
if
you
have
any
any
anything
to
hide
on
the
discussion.
H
Principles
live
iso
type
of
trees,
so
it's
essentially
the
live.
Isos
are
live
versions
of
the
recorders,
so
it's
essentially
a
live
system.
Fully
live
system,
fedora
core!
It's
running
from
an
iso,
but
it
just
boot
it
from
an
iso
either,
be
it
like
a
real
disc
or
you
if
you
write
your
iso
onto
a
usb
stick
or
whatever
it
boots
from
that
yeah.
So
that's
one
of
the
issues
with
fixed
presently
and
one
progress.
Another
thing
in
progress
that
is
coming
up
is
the
platform
console
defaults
changes.
H
So
the
the
goal
is
we're
working
on
making
it
possible
to
have
per
platform
different
console
default,
console
configuration
per
platform,
so
that
would
make
sure
that
on
aws
you
get
the
the
version
of
the
console
that
you
need
to
get
on
aws
and
on
there
metal
you
get
the
version
of
the
console
that
works
best,
environmental
and
things
like
that
and
for
each
different
platform.
We
have
different
defaults,
and
this
is
this
work
in
progress.
So
it's
not
there
yet!
H
But
if
you
have
like
anything
interesting,
if
you,
you
think
that
we
should
use
a
specific
version
of
of
the
specific
configuration
for
the
console
for
bare
metal
or
whatever
feel
free
to
chime
in
into
the
issue.
H
Okay,
so,
like
that's
like
two
biggest
items,
then
we
have
a
couple
of
deadlocks
and
a
couple
of
pictures
with
the
kernel
and
fedora
that
have
been
reported
to
us.
So
that
might
we
have
interest
too,
if
you
encounter
those.
Obviously
you
won't
yet
because
you
don't
probably
don't
have
the
same
kernel
yet
on
on
okd.
H
So
I'm
just
thinking
there,
then
there
in
case
you
have
something
that
looks
like
this
and
you
can
chime
in
too
and
yeah,
and
I
forgot
we
are
really
really
close
to
have
a
h,
a
arch,
64
artifacts.
Now,
so
we
essentially
have
them
available.
It's.
H
H
Maybe
we
already
have
one
yeah.
We
just
have
one
right
now,
so
we've
just
had
one
stable
release
for
hr
64,
so
you
should
be
able
to
try
them.
We
have
adb
smi's,
we
have
perimeter,
we
have
a
lot
of
stuff,
openstack
images
all
but
yeah
ear64
is
definitely
coming
to
federace.
H
It's
there
yeah
and
the
remaining
things
are
mostly
cosmetic
like
make
sure
that
it's
listed
on
the
website
and
anything
like
that.
So
that's
just
coming
up
right
away
so
yeah.
Let's
just
link
this
one
there
into
I've
already
linked
it.
No
bills.
Are
there
I'm
updating
the
notes
at
the
same
time,
to
make
sure
I
don't
forget
oops
there.
It
is.
H
Okay-
and
one
last
item
that
I
want
to
point
right
point
out
is
if
you
have
only
fiber
channel
or
fiber
channel
over
ethernet
connections
on
baritone
nodes,
for
example,
you
might
currently
have
no
way
to
get
a
network.
H
Any
network
working
we've
we're
looking
into
that
to
see
if
we
need
to
add
by
default
well,
including
the
image
any
other
utilities
to
make
fiber
channel
fully
working
on
fedora
caress
and
so
yeah,
if
you're
interested
in
that,
if
that's
something,
that's
just
really
interesting
for
you
in
your
use
case
on
their
metal
mostly
again
feel
free
to
chime
in
and
and
we'll
on
the
discussion,
and
that
should
be
it
for
me.
I
think
that's
like
the
top
items
from
federal
careers
from
the
last
two
weeks.
C
We've
got
a
couple
questions
in
the
chat
channel.
H
A
Okay,
sure
you
mentioned
the
f
cause
supports
live
iso
stuff.
I
didn't
know
that
that
was
a
thing.
Does
it
support
like
hybrid
persistence,
type
stuff
as
well
like,
if
you
put
it
onto
a
usb,
stick
that
var
home
would
automatically
be
provisioned
in
the
remaining
space
and
you
could
use
it.
H
So
it's
not
automatically
done,
but
you
can
do
it.
So
if
either
you
use
the
space
you
have
on
the
usb
stick,
but
it's
probably
not
what
you
want,
but
most
really
you
want
to
do.
What
you
want
to
do
is
use
the
disk
on
the
system
to
like
persist
data
so
basically
essentially-
and
you
can
do
that
so-
the
live
iso
is
a
fully
fully
working
philadelphis
environment
and
his
running
ignition.
H
So
if
on
first
boots,
so
if
you
want
to
configure
the
live,
iso
vr
ignition,
it's
perfectly
possible,
it's
supported
and
it
will
do.
And
so,
if
you
want
like
say,
run
dot
your
life
iso,
you
can
even
like
embed
inition,
any
ignition
configs
into
the
iso,
and
it
will
be
picked
up
by
ignition
that
first
boot
and
then
applied.
And
so
let's
say
you
want
a
partition,
slash
bar
and
put
slash
bar
on
the
business
and
disk
and
keep
running
the
live
iso.
As
at
the
british
fresh
and
like
ephemeral
environment.
H
You
can
do
that
and
let's
say
when
you
reboot,
then
you
will
get
the
live
iso
again,
so
you
have
to
figure
out
like
an
another
way
to
do
updates
in
this
case,
if
you
use
it
like
that,
the
the
main
use
case
for
that
kind
of
stuff
that
kind
of
way
of
working
is
pxc
booting.
Well,
where
you
just
like.
You
update
the
version
of
federacores
directly
on
the
server
and
you
reboot
your
nodes
and
they
just
simply
reboot
into
the
new
version.
H
Yeah
you,
you
can
either
use
it
to
provision
bermuda
to
install
federal
chorus
onto
nodes,
because
you
have
a
fully
working
environment
and
you
can
run
chorus
and
store
manually
there
or
automatically
the
ignition
or
you
or
kernel
arguments
directly
provided
to
to
to
the
kernel
to
the
on
the
on
the
common
line
during
the
live
iso
boot
up
or
you,
you
can
directly
use
that
as
as
an
environment.
It's
it's
it's
possible,
but
yeah.
It's
probably
better
use!
C
Okay,
let's
move
on
now
to
docs
updates
brian
go
ahead
and
take
it
away.
I
Okay,
so
we've
been
working
on
updating,
okdio
and,
if
anyone's
actually
tried
to
update
the
main
site.
You'll
understand
that
it's
not
trivial
and
the
learning
curve
to
understand
the
various
packages
that
make
up
that,
and
you
really
do
need
to
be
a
full
front-end
developer
to
actually
achieve
that.
So
what
we've
done
is
we've.
Let
me
actually
share.
I
We've
actually
set
up
a
beta
branch.
Can
you
see
that
yep
yep,
okay.
I
A
branch
within
the
main
repo-
and
this
has
now
been
switched
over
to
use
mk,
docs
and
you'll,
see
there's
a
link
on
the
side
which
actually
takes
you
to
the
beaver
site.
So
this
is
being
rendered
using
github
pages
and
on
a
github
push
you
get
all
of
the
automation
to
build
and
update
this
site.
So
when
we
accept
a
pull
request
or
a
commit
onto
the
branch,
it
will
actually
automatically
update
and
the
documentation,
there's
no
infrastructure
needed
within
red
hat
or
anywhere
else.
I
The
site
is
fully
marked
down,
so
even
the
the
home
page
now
is
being
fully
marked
down.
It
means
that
we
don't
have
all
the
fancy
animations
that's
on
the
current
site,
but
in
terms
of
updating
it
it
is
much
simpler.
We
also
have
a
much
better
navigation
where
we've
got
the
high
level
site
navigation
here
on
the
left
and
an
in-page
navigation
on
the
right
and
just
click
and
navigate
around
based
on
the
headings.
One
two
and
three.
This
is
all
sort
of
automated.
I
If
I
look
at
the
site,
it's
fully
fully
and
responsive.
So
if
I
come
onto
a
phone,
you'll
see
that
the
menus
vanish
and
they
sort
of
switch
up
to
here
and
get
to
the
in-page
navigation.
I
So
it's
a
fully
responsive
design
and
it
also
will
flick
between
light
and
dark
mode
depending
on
your
browser,
and
if
your
system
settings
allow
that
whenever
you
open
it,
it
should
adopt
to
whatever
your
your
system.
Settings
are
so
the
basic
site
is
there
and
I'm
not
precious
about
this.
If
anybody
thinks
there's
a
better
color
scheme
or
anything
like
that,
it's.
I
H
F
A
The
title
where
you
see
there
it
says
community
distribution
of
powers,
red
hat
openshift,
like
it's
actually
a
little
hard
for
me
to
read
it
because
of
the
the
background,
reducing
the
contrast.
I
I
Yeah,
which
is
which
is
just
from
the
outside
I'm
happy
to
take
that
out,
yeah
just
other
than
that.
I
I
just
wanted
to
add
a
little
bit
more
and
interest
rather
than
having
solid
you'll,
see
that
as
we
get
into
the
if
we
get
into
the
main
site
we
lose
all
of
that,
so
it
just
becomes.
K
F
I
You've
got
navigation,
the
search
again,
I
can
type
vsphere
and
it
it
just
pops
up,
and
then
these
are
all
the
places
vsphere
and
appears.
So
we
have
a
sort
of
a
an
automatic
search
index
on
the
side
as
well.
So
if
you
want
to
find
stuff
and
what
I
have
done
in
the
community
section,
if
you're
interested,
I
started
putting
documentation
about
how
the
markdown
we
use
standard
markdown,
but
we
also
use
an
extension
library.
So
we
can
do
things
like
have
these
sort
of
information
boxes
we
can
put
tabs
in.
I
I
I
I
have
gone
and
adopted
us
english
as
the
standard
default,
so
everybody
needs
to
sort
of
use,
u.s,
english,
if
you
want
to
on
a
page,
you
can
actually
put
a
comment
in
which
you
can
add
words.
So
if
you've
got
technical
words,
you
can
add
them
in
to
pass
a
spell
checker,
it
will
not
publish
if
there's
spell
checking
errors
or
if
there
is
link
checking
errors.
The
other
thing
we
do
is
a
full
link,
checker
on
publish
so
again
we
should
never
have
broken
links
or
bad
spelling
within
the
content.
I
So
it's
there
go
play
with
it.
As
I
say
it's
in
the
beta
branch
on
on
on
the
main
github
repo,
I'm
now
looking
at
the
content
and
trying
to
sort
of
beef
up
the
content
and
get
there,
one
thing
that
we've
talked
about
is
on
the
install
page,
maybe
moving
some
of
the
content
from
the
okd
readme.
I
C
Yeah
I
wanted
to
actually
loop
vedim
in
on
this,
so
vadim,
the
readme
from
the
repo
has
sort
of
a
mix
of
a
lot
of
different
aspects,
right
so
installation
and
the
releases
nightlys
and
stuff
like
that.
So
the
thought
was
that
we
would
break
this
up
a
little
bit
and
make
the
read
the
readme
more
of
just
a
true
readme
of
this
is
the
project
and
then
have
links
to
the
various
okd
dot
io
web
pages.
For
those
particular
topics
is:
do
you
have
any
concerns
about
that?
E
That
sounds
that
sounds
very
good.
I'm
just
not
sure
which
side
the
people
first
encounter.
Do
they
look
for
okie
dio,
or
do
they
go
to
github
and
figure
out
that
it's
okay?
So
I
think
there
should
be
some
text
very
brief
description
on
both
sides
so
that
people
would
not
get
confused,
but
other
than
that
we
definitely
should
upload
a
bunch
of
texts
from
readme
way
overload.
E
G
I
I
I
We
can
actually
move
that
move
into
here
as
well.
I
mean
I
actually
do
have
clc
section
in
there,
but
I
I
I
think
that
we
can
actually
add
and
make
this
a
much
more
dynamic
site.
Hopefully
it's
easy
enough
to
use
now.
So
if
we
have
like
a
a
headline
on
build,
x's
broken,
please
use
a
previous
version.
I
There's
no
sort
of
steep
learning
curve
and
I
think
mike
appreciated
just
how
horrible
it
is
having
content
with
his
guides,
and
so
hopefully
it
all
just
makes
sense
and
sees
you
now
but
yeah,
I'm
hoping
to
have
the
core
of
the
reorganization
and
this
in
the
pages
in
place
within
the
next
two
weeks.
So
then
we
can
maybe
this
call
in
two
weeks
time.
People
can
have
a
look
through
and
actually
make
make
take.
The
decision
needs
more
work
or,
let's
switch.
It
live
now.
C
I
think
a
primary
thing
will
be
getting
that
readme
parsed
out
and
that
text
that
vadim
mentioned
that's
going
to
be
a
necessity,
and
once
we
have
that,
I
think
it's
safe
to
to
do
the
other
stuff
sort
of
as
needed.
F
Yeah,
I
know
it
looks,
looks
fantastic
and
I'm
really
happy
that
you
know
you
made
this
switch.
F
We
should
make
sure
that
we
get
the
color
scheme
reasonable,
because
when
I
look
at
it
it
your
sort
of
lighter
background
in
in
the
the
middle
looks
like
it
has
sort
of
a
bluish
tinge
to
it
and
then
the
highlights
are
red
and
the
sort
of
red
blue
is
the
worst
on
the
eyes.
You
know
the
human
eye
can't
focus
simultaneously
on
both
red
and
blue,
because
they're
opposite
ends
of
the
spectrum.
I
F
Oh
sure,
well
I
guess
from
the
red
hat
side
you
know
we
like
red,
but
that
doesn't
mean
that
we
should
destroy
everybody's
eyes.
You
know
in
in
the
process
yeah
so
well.
C
We
can
use
the
colors
incorporated
into
the
rocket
man
right
and
just
sort
of.
F
But,
and
I'm
guessing
the
like
is
it
is
it
themed
is,
is
there
like,
I
haven't,
looked
at
how
hard
it's
changed,
the
the
template
for
colors.
I
I've
broken
them
all
out,
so
it's
really
easy.
If
you
go
and
look
at
the
there's,
a
css
folder
in
the
docs
folder
it's
in
there,
so
every
color
has
got
a
css
variable
name
against
it,
so
we
can
just
go
and
change
them.
C
And
there's
a
switch
for
changing
right
so
that
people
we
we
need
to.
Let
people
know
that
that's
there
all
right.
I
do
want
to
move
on
because
we
do
have
a
full
agenda
and
some
we
have
some
guests
here.
So
I
need
to
get
through
some
of
the
other
things
that
we've
got.
I'm
real
briefly,
the
I'm
noticing
in
the
channels
that
there's
a
lot
of
folks
asking
questions
without
providing
a
lot
of
context.
C
Has
anyone
asked
us
in
the
docs
group
and
folks
really
sure?
Is
there
like
a
kubernetes
based
how
to
ask
questions
document
that
you're
aware
of,
like
I
know,
there's
that
one
site
that's
been
around
for
years?
That's
like
how
to
ask
questions,
but
does
anyone
know
of
a
good
text
that
we
can
point
to
or
build
upon
about?
C
How
folks
should
ask
questions
when
they
are,
you
know
seeking
help
in
the
channels
or
in
a
discussion
item
or
in
it,
vadim
started
to
put
some
stuff
together,
which
was
really
helpful
and
if
maybe
we'll
just
put
that
into
a
document.
If
we
don't
know
of
anything
else,.
C
C
People
excellent,
very
cool,
it
doesn't
sound
like
there's
a
kubernetes
specific
page,
so
I
think
we'll
just
build
on
what
vadim
has
and
the
docs
working
group
can
tackle
that
next
week.
We'll
just
add
those
items,
team
came
up
with
a
list
a
couple
weeks
ago,
we'll
just
put
those
items
together
and
then
put
that
in
a
format
that
can
be
put
into
the
description
of
the
slack
channel
and
then
at
the
top
of
the
google
well
in
the
slack
channel.
C
Then
we're
going
to
change
the
google
group
to
actually
point
to
the
discussion
section
and
then
in
the
discussion
section
we
can
put
that
there
and
and
have
folks
provide
that
information
in
the
video
that
can
be
done
and
there's
templates
in
the
discussion
section
as
well.
Right.
E
There
are,
you
can
have
templated
answers,
but
I
don't
think
these
are
shared
across
all
the
work.
At
least
I
have
a
list
of
mine,
okay
and
I'm
not
sure
how
to
share
them
like
these
are
saved.
In
my
account,
okay
get
up
might
be
weird:
okay,.
C
C
And
training
materials
for
docs
support,
resource.
C
That
was
basically
docs
about
docs,
and
so
we
answered
that
a
little
bit
at
the
last
docs
meeting
and
we'll
be
talking
more
about
that
at
this
next
meeting.
Repo
issues
is
there
anything
that
people
want
to
pull
out
of
the
repo
to
discuss
real
quick
from
the
issue
section
of
the
repo
any
actual
issues
vadim.
C
I
know
you've
been
pushing
a
lot
of
stuff
to
discussion
because
a
lot
of
it
was
discussion
oriented
and
you
covered
a
lot
of
issues
at
the
beginning
here,
but
anything
you
want
to
bring
to
our
attention.
E
No,
I
think
I
haven't
seen
anything
super
interesting
or
at
least
which
something
which
cannot
be
covered
with
a
log
bundle.
I
wanted
to
discuss
the
whole
procedure
when
I
learned
that
github
has
discussions.
I
was
thinking
that
would
be
a
great
place
to
figure
out
what
this
issue
is
about,
and
then
you
can
move
it
back
to
issues
now
turns
out.
E
Github
has
a
label
called
bug
and
if
you
mark
it
as
a
bug
that
will
give
you
a
brand
new
tab
outside
right
nearby
between
issues
and
discussions,
and
it's
going
to
be
a
dedicated
place
for
bugs.
So
you
can
have
three
places
where
you
can
figure
out.
If
it's
a
bug
or
not,
I
don't
know
if
I
should
go
on
with
moving
issues
to
discussions
or
should
we
keep
them
as
issued
just
mark
them
as
bugs
my
head
is
spending
on
this
honestly.
J
It's
probably
because
a
lot
of
people
use
the
issues
with
the
bug
label
or
something
like
that:
anyways
they're
thinking,
oh,
let's
make
it
easier
for
them,
but
we
kind
of
came
up
with
our
own
thing.
In
the
meantime,
I
don't
know
honestly,
like
I've,
been
keeping
an
eye
on
the
discussion
sections
and
watching
what
vadim
moves
in
there.
A
lot
of
it
isn't
really
like
a
bug
bug.
J
You
know
it's
not
like
it's
just
a
bunch
of
people
over
and
over
having
trouble
deploying
their
clusters,
which
is
understandable
because
it's
really
hard,
but
it
kind
of
illustrates
like
how
far
we
have
to
go
here.
E
F
Right
now,
like,
conversely,
a
curated
bug
like
not
everybody
should
be
able
to
say
that
it's
a
bug
right
like
like
to
me,
like
the
normal
usage
that
I
used
to
see,
is
that
anybody
can
raise
an
issue
and
then,
after
discussion,
you
decide
whether
or
not
it's
a
bug,
and
then
I
guess
a
discussion
would
then
be
something
that
people
don't
even
think
is
an
issue.
It's
just.
They
want
to
talk
about
things,
maybe
propose
new
things
or
whatever.
J
It
is
also
hard
to
figure
out
because
oftentimes
I
do
notice
that
a
lot
of
bugs
that
I
run
into
personally
already
have
are
tracking
and
being
fixed,
but
it's
very
hard
to
figure
that
out
from
the
outside
looking
in,
and
that
I
don't
think,
that's
necessarily
related
to
here,
but
I
think
other
people
are
having
that
problem
as
well.
We
could
probably
go
a
lot
further
to
closing
a
lot
of
these
things
out
quickly
if
we
had
some
way
us
as
external
people.
J
Not
just
you
know
those
of
us
inside
of
the
red
hat
bubble,
to
be
able
to
triage
that
kind
of
stuff.
G
G
You
know
communication
with,
like
the
red
hat
people
who
are
working
on
ocp.
You
know.
I
know
that
like
if
these,
if
there
were
bugs
getting
into
bugzilla
and
that's
a
you
know,
I
know
we're
migrating
jira
in
the
future
and
everything.
But
if
we
had
bugs
going
into
that
pipeline,
then
you
would
have
people
responding
more
quickly,
at
least
from
people
who
are
working
on
the
code.
I
think
now
it's
just
there's
this
gap
that
we
need
to
kind
of,
like
you
know,
fill
up.
C
Okay,
I
want
to
table
this
discussion
because
we've
got
20
minutes
left
and
we
have
a
guest
and
we
have
a
couple
of
other
important
items
that
I
want
to
do.
But
let's
put
this
on
the
agenda
for
the
next
meeting
and
try
to
come
up
with
an
actual
plan
for
how
to
move
folks
around
vadim.
E
Much
better,
I
don't
think
we
have
a
guide
where
to
look
first
in
the
most
gathered
ra,
however,
a
bunch
of
tools,
centered
around
action
was
getters,
which
you
just
needed,
and
it
gives
you
a
very
short
guess
of
what's
happening,
but
then
it
depends
on
on
basically
specific
components
like
api
server
issues
are
very
broke,
it
can
be
audited,
it
can
be
just
simply
bottom
starting.
So
for
a
start,
you
can
have.
E
C
All
right,
I
didn't
want
to
spend
too
much
time
on
that,
but
I
think
I
noticed
there's
a
lot
of
stuff
that
if,
if
we
can
get
through
the
must-gathers,
if
community
members
can
get
through
the
must-gathers,
we
can
save
vadim
and
a
lot
of
other
folks
a
lot
of
time
and
and
we're
concisely
answer
people.
I
think
okay,
I'm
gonna
breeze
through
these
really
quick.
C
G
Like
I
don't
I
don't
know
if
this
is
ever
something
the
community
would
accept,
but
like
is
it
like?
Would
it
ever
be
possible
for
us
to
create
the
kind
of
service
where
someone
could
like
make
a
must
gather
and
be
able
to
like
drop
it
in
a
place
where
we
could
review
later?
We
can't
do
that
very
much.
C
All
right:
well,
let's,
let's
move
on
because
I
do
want
to
get
to
this
other
stuff,
the
user
questionnaire.
The
the
group
decided
to
the
docs
group
decided
to
use
your
questionnaire
would
be
great,
and
at
this
meeting
we
talked
about
it
last
time
and
so
user
questionnaire
is
up
and
posted
adriti
posted
a
response
with
some
great
suggestions.
C
If
other
folks
can
chime
in
the
goal
is
to
get
the
survey
out
to
okd
users,
what's
missing,
what
is
the
you
know
their
biggest
roadblocks?
What
do
they
like
the
most
etc
so
check
that
out
code
ready
containers,
sub
group
we've
got
neil
daniel
driti,
myself,
charo
brian
and
anyone
else
want
to
be
on
that.
Anyone
else
on
the
call
who
wants
to
be
part
of
that
sub
group.
That's
helping
maintain
good
rating
containers,
put
your
name
in
the
chat
or
speak
up.
C
Okay
bruce
is
interested
as
well.
Okay,
we'll
add
your
name,
okay,
very
cool
and
we're
in
the
process
of
navigating
that
out.
Basically,
I
think,
neil
and
and
daniel
are
going
to
be
sort
of
the
leads
on
this
and
yeah.
It
looks
like
we
got
a
fair
amount
of
people
that
are
interested,
so
that's
awesome
and
that's
that
was
diane's
measure
of
success,
for
the
subgroup
was
that
there
were
people
interested
to
obtain
it.
Vadim
answered
the
operator
questions.
Please
check
those
out.
C
We
won't
go
over
them
here,
but
there's
some
clarity
about
operators
and
we'll
talk
about
that
at
the
next
meeting.
New
business
was
a
migration
path
outline.
Would
it
be
possible
vadim
for
us
to
create
a
simple
table
that
we
could
direct
people
to,
because
it
seems
like
a
lot
of
questions
are.
Can
I
up
upgrade
from
this
version
to
this
version,
etc?
Is
it
too
complex
for
that
or
do
you
think
we'd
be
able
to
pull
it
off.
E
For
a
cp,
we
have
a
full-blown
service
which
gives
you
an
upgrade
path
for
okay,
it's
a
bit
untrivial
I
mean
if
you
can
parse
json
manually,
we
have
an
endpoint
but
visualizing.
This
is
a
bit
tricky.
E
Although
I
think
we
have
like
a
static
type
diagram
which
can
give
you
the
edges,
because
we
don't
use
any
channels,
that's
much
easier
than
ocp
again,
visualization
of
this
agreement
is
always
a
tricky
question.
E
C
You
so
now
I
want
to
move
on
to
our
guest
sandro,
who
is
here
to
talk
a
little
bit
and
did
anyone
else
show
up
or
is
it
just
you.
C
So
was
it
you
who's
going
to
talk
who's,
going
to
talk
about
cube,
vert
who's
here
for
talking
about
cuber?
Did
they
show
up.
K
C
Okay,
all
right!
Well,
our
discussion
item
is
because
of
a
desire
for
okd
to
be
more
stable
or
the
switch
to
kubert
for
virtualization,
the
upstream
of
red
hat's
virtualization
offering,
and
we
received
an
email
about
about
that
and
invited
folks
to
come
here
and
talk
about
it.
K
I
see
that
I
see
that
fabian
drop
it
okay,
so
you
probably
know
that
within
openshift
we
are
shipping
openshift
virtualization
and
we
have
not
really
a
fully
integrated
solution
corresponding
to
that
based
on
okd.
A
I'd
like
to
see
that
for
sure
like
it,
it's
a
very
interesting
thing
to
me
and
I
think
one
of
the
bigger
weaknesses
of
kubernetes
as
a
whole
is
you
know
how
poor
it
is
at
doing.
You
know
what
people
tend
to
expect
for
virtualization.
C
K
Yeah,
I
I
think
so
the
idea
is
to
get
the
community
involved
in
this
process
and
getting
more
closer
to
providing
something
that
can
be
tested
by
the,
for
example,
for
the
over
community.
K
To
see
what's
the
future
and
getting
more
in
touch
on
what
will
happen,
for
example,
with
red
and
openshift
virtualization,
it
will
help
getting
more
understanding
about
the
differences
and
the
possibilities
that
okd
can
provide
in
that
area.
I
Let's
work
on
multiple
sort
of
different
backgrounds,
because
obviously
okie
d
works
on
cloud
I
mean,
I
think,
we've
got
a
fairly
large
v
sphere
population-
I
I
run
overt,
but
I
think
majority
tend
to
run
vsphere
and
then
obviously
we
got
some
bare
metal.
What
is
the
dependency
on
the
underlying
platform
on
the
on
the
virtualization
layer?.
K
A
The
other
thing
is
that
I
remember
reading
this
in
the
open
shift
documentation.
I
don't
know
if
this
is
still
true,
but
if
you
pair
openshift
container
platform
running
the
openshift
virtualization
module
on
top
of
red
virtualization,
you
can
still
orchestrate
vms
from
the
openshift
interface.
They
just
get
provisioned
through
rev,
and
so
that
means
that
you
get
a
second
opera.
A
You
get
the
ability
to
do
direct
virtualization
management
through
the
openshift
api
for
for
rev,
so
you
get
an
open
stack,
cloudy
super
scalar
type,
provisioning
api,
combined
with
the
more
traditional
virtualization
resource
management,
interface
and
api
of
overt
with
one
solution.
So.
K
Yeah
we
got
a
novel
conference
today
and
gal
zeidman
from
my
team
presented
the
integration
between
okd
and
oviatt.
In
these
terms,
so
I
added
a
link
to
the
agenda
to
that
presentation
and
I
would
suggest
to
have
a
look.
The
integration
works
pretty
well.
K
G
A
Oh,
my
god,
I
don't
want
to
think
about
that.
The
the
the
only
real
problem-
I
you
know,
actually
it's
not
the
only
real
problem,
but
the
biggest
problem
with
openshift
virtualization,
at
least
from
my
perspective,
like
I've
played
with
it
a
little
bit
at
some
demos
and
whatnot,
is
that
it's
too
hard
to
work
with
like
it
is
extremely
alien
to
someone
who's
more
used
to
like
how
the
overt
ui
works
and
how
virtual
machines
tend
to
be
managed
that
way
how
provisioning
works?
A
With
that,
like
I,
I
would
class
openshift
virtualization
closer
to
openstack,
rather
than
closer
to
overt,
whereas
like
say
something
along
the
lines
of.
What's
it
called
harvester
harvester
is
closer
to
overt
than
it
is
to
openstack
in
terms
of
how
virtual
machine
management
works
and
they're
both
using
cubeverd.
A
So
like
at
least
having
used
overt
and
okd
like
I.
If
I
wanted
to
move
from
over
to
okd,
I
think
a
prerequisite
would
be
some
kind
of
ux
migration
that
makes
it
so.
People
who
are
working
that
way
can
still
continue
to
function
in
an
openshift
world,
and
I
don't
think
we're
there.
Yet.
Maybe
we
will
be
someday,
I
don't
know,
but
that
that's
really.
The
big
gap
that
I
see
is
that
people
who
work
the
vmware
over
rev
style
of
virtual
machine
stuff
basically
have
no
home
in
openshift
virtualization.
C
Okay,
let's
see
if
we
have
any
questions
that
were
there
no,
mostly
discussion
in
the
discussion.
I
just
just
yeah.
G
Go
ahead.
Sorry
jimmy!
I
just
want
to
add
a
side
note
about
some
of
the
tech
behind
this
too.
If
people
are
curious,
oh
so,
like
you
know,
in
openshift
now
we
have
a
kubrt
cloud
provider.
G
You
know
like
machine
api
provider
that
we've
been
working
on
for
a
while,
I'm
not
sure
how
up
to
date
that
is,
but
recently
apple
has
come
to
the
cluster
api
community
and
they've
put
a
lot
of
people
behind
trying
to
create
a
cluster,
a
full
cluster
api
provider
for
coobert,
and
I
know
that
our
red
hat
team
that's
been
working
on
kubevert
is
now
like
kind
of
linking
up
with
some
of
these
apple
people
to
help
maybe
see
if
we
can
bridge
the
gap
between
the
machine
api
provider
that
we
have
now
for
kubevert
and
what
apple
wants
to
do
in
the
yeah.
F
G
Neil
apple
wants
to
do
with
kubevert
in
the
upstream
I
like
so
in
the
cluster
api
project.
It
seems
that
apple
is
very
curious
about
building
and
automating
clusters
of
kubernetes
on
top
of
kubert.
You
know
kubernetes
clusters,
basically,
and
so
they
want
to.
They
want
to
put
a
lot
of
into
that,
and
I
know
that
they
had
been
looking
at
the
openshift
machine
api
provider,
because
that
was
the
only
thing
out
there
for
a
while,
but
it's
very
open
shift
specific.
G
I
mean
that
that's
a
great
question.
You
know
like
I,
I
talked
with
some
of
the
apple
folks
and
I
talked
with
some
of
the
red
hat
folks
and
basically
made
some
introductions
and
tried
to
hook
them
up
because
I
don't
work
on
kubert.
Like
specifically,
my
impression
is
that
apple
is
really
curious
to
get
this
cluster
api
provider
running
and
they
would
love
to
have
red
hat's
help
working
on
it.
So
I
know
those
two
groups
are,
you
know,
there's
some
communication
happening
there.
G
I
I
don't
have
my
nose
in
the
middle
of
it
to
know.
Really,
if
like
could
they
be,
could
apple
be
convinced
to
do
this
all
on
okd
or
something
like
that
would
be
really
cool?
Do
you
still
have
those
contexts
where
you
could
yeah
yeah?
I
mean
apple's
still
coming
to
the
cluster
api
like.
I
could
continue
this
conversation
with
them
and
see
if
we
can
maybe
convince
them
to
come
here
or
something.
C
G
C
About
this
yeah,
all
right,
we've
got
three
minutes
left.
So
let's
do
a
sub
working
group
on
this.
I
think
it'd
be
great
to
have
a
sub
working
group
on
this,
we'll
get
the
red
hat
folks
sandro.
Anyone
else,
that's
interested,
and
then
folks,
here,
okay
d,
folks
I'll,
send
something
out
over
the
working
group,
google
group
email
list
and
then
we'll
get
volunteers
for
that
and
actually
solidify
this
as
a
project.
C
So
it's
a
great
offering
and
then
mike
will
reach
out
to
apple
folks,
and
let
them
know
that
we're
starting
this
as
a
sub
project
and
invite
them
maybe
they'll,
be
interested.
We
have
three
minutes
left
someone
wanted.
Someone
asked
did
I
get
to
this
because
I
know
it's.
We
want
to
be
fair
to
everyone
in
terms
of
access.
Someone
asked
about
changing
the
meeting
time
to
support
basically
folks
that
are
in
europe
and
middle
east
and
africa.
C
This
is
a
little
bit
late.
This
is
so
it
starts
at
1
pm.
For
me,
we've
got
some
other
folks
where
this
is
starting
at
9.
00
a.m,
like
I
think
it's
9
a.m
for
diane.
I
don't
know
that
we
can
go.
What
are
people's
thoughts?
I
don't
want
to
direct
this
conversation.
What
are
people's
thoughts
about
going
earlier?.
A
I
don't
think
I
can
go
any
earlier,
like
my
days
are
already
kind
of
sandwiched
hard
as
it
is
like.
I
I
front
haul
most
of
my
community
meetings
through
the
first
half
of
the
week
and
then
the
latter
half
of
the
week
is
when
I
actually
get
to
do
other
things
and
like
some
of
y'all
have
seen
my
calendars,
like
my
mondays
and
tuesdays
are
basically
chock
full
and
I
basically
don't
have
any
more
free
time
left.
A
So
I
don't
think
I
can
really
move
it
anymore,
even
though,
and
like
the
west
coast
folks,
I
don't
think
diane
could
get
up
earlier.
I
don't
think
she
wants
to
get
up
earlier.
F
I
mean
same
time
zone
as
diane,
we're
actually
not
that
far
from
each
other,
and
it
is
actually
currently
at
10
because
of
daylight
savings
time.
But
then
I
guess
october,
whenever
we
flip
back
to
standard
time,
it'll
go
to
nine
right
and
I
I
wouldn't
be
able
to
make
it
any
earlier.
Diane
said
quite
a
while
ago
that
it
was
impossible
negotiating
a
time
in
the
red
hat
calendar
to
get
it
somewhere,
which
makes
it
really
hard
to
move.
C
E
That's
like
eight
that's
acceptable.
Moving
it
later
would
be
very
bad.
Moving
earlier
might
might
make
a
conflict
with
my
play:
joke
music.
So
I'm
fine
with
the
time,
and
I
can
understand
that
some
folks
might
not
be
able
to
make
it.
I'm
pretty
sure
someone
from
you
can
make
him
to
one
of
these
meetings,
and
that
means
they
could
hit
me
or
christian
beforehand.
So
we
can
pass
the
information
and
don't
watch
the
recording.
That's
probably
the
only
solution
to
the
stupid
time
zones.
We
have.
C
D
C
Yeah
all
right,
let's,
I
think
we're
going
to
stick
with
where
we
are
for
now.
This
is,
I
don't
mean.
I
understand
that
this
for
some
folks
it
might
be
difficult,
like
driti
just
pointed
out
that
it's
11
30
there
we'll
keep
it
here
for
now
and
we'll
come
up
with
a
communication
mechanism
to
make
sure
that
people
get
the
recordings
after
the
fact
very
quickly
and
that
people
who
have
input
into
meetings
can
get
input
to
point
people
and
their
ideas
across
via
point
people
if
they
can't
make
it.
C
Okay,
awesome:
okay,
folks,
we
are
over
right
now,
so
I
I
guess
it's
a
good
time
to
end
and
thank
you
so
much.
This
is
a
hugely
productive
meeting
and
I'll
get
the
video
up
as
soon
as
possible
and
I'll
talk
to
you
all
soon.