►
From YouTube: OKD Working Group 2020 03 17 Full Meeting Recording
Description
OKD Working Group 2020 03 17 Full Meeting Recording
A
Hello,
everybody
and
please,
if
you
could
add
your
names
to
the
attendee
list,
the
link
is
in
the
chat
and
we'll
get
started
with
this.
Ok
key
working
group
meeting
today,
Christian
Blum
back,
are
one
of
our
co-chairs.
Has
that
flu
I'm
not
sure
which
fluid
is
but
we're
wishing
him
well
and
he
is
staying
home
and
not
available
for
today's
call.
A
All
right,
so
the
team
you're
here,
which
is
great
because
Christian
tapped
you
on
the
shoulder
to
give
the
ok
d4
update
as
well
as
Vadim,
has
an
okay,
Dion
GCP
tour,
though
that
will
use
that
today
to
drive
some
conversations
and
just
a
quick
update
on
events.
Upcoming
events.
There
still
is
not
a
date
posted
for
pukani
you.
If
it
does
happen,
we
will
have
face
to
face
ok
d
working
group
meeting
there.
A
We
were
going
to
hold
a
virtual
open
shift,
Commons
gathering
to
replace
the
Amsterdam
one,
but
we
have
now
cancelled
that,
but
we
are
having
a
red
hat
summit,
OpenShift
Commons
gathering,
and
it
will
be
virtual
on
April,
27th
and
I'm,
going
to
ask
Vadim
and
Christian
to
record
pre-record
and
probably
get
dusty.
The
pre
record,
ok,
d4
Fedora
core
OS
talk.
It
won't
be
part
of
the
agenda
for
the
day.
The
agenda
is
very
curated
down
to
mostly
customer
talks
in
one
state
of
openshift.
A
410
kubernetes
released
talk
just
because
we
don't
think
people
are
going
to
watch
an
entire
day
long
event,
but
we'll
see
I
think
we'll
watch
stuff,
and
so
we
gonna
add
all
the
talks
that
aren't
getting
broadcast
as
part
of
that
day.
We're
gonna
pre-record
all
of
those
that
have
them
available
so
we'll
be
doing
that
and
as
far
as
I
know,
almost
all
of
the
other
things
over
the
next
few
months
have
been
cancelled
or
postponed.
A
But
my
offer
to
anyone
here
if
you've
got
a
topic,
whether
it's
ok,
D,
related
or
Fedora,
core
place
that
you
want
to
record
a
briefing
on
we're
just
basically
taking
all
the
content
from
the
coop
cons
and
summit
that
aren't
getting
broadcast
and
re-recording
them
and
making
them
available
on
our
YouTube
channels,
which
is
our
H
OpenShift.
So
that's
sort
of
the
state
of
state
of
events
things
here.
Any
questions,
comments
about
that.
A
B
So
here
are
the
notes:
there's
been
quite
active
two
weeks
for
us.
The
biggest
news
is
that
we
finally
have
a
proper
documentation
on
thoughts.
Okay,
D
bowtie.
Oh
it's
pretty
much
a
carbon
copy
from
OCP
for
the
three
documentation,
so
it
has
a
lot
of
inaccuracies,
a
lot
of
mentions
of
velcro
s.
So
what
we
would
like
you
to
do
is
to
skim
through
those
find
the
parts
which
are
irrelevant
to
fedora
crow
s
like
I,
DMZ
and
file
box.
To
be
okay,
dear
echo,
we
will
arrange
them.
B
Together
and
ask
openshift
dog
steam
to
get
a
priority
on
those
later
on,
we
would
be
able
to
I'm
not
too
familiar
with
documentation
and
the
way
they
template
things.
But
after
a
couple
of
tries,
we
would
be
able
to
submit
quick
fixes
ourselves
manually,
I'm
status,
so
the
platform
support
so
AWS
vSphere
UPI
a
bare-metal
unknown
to
work,
we've
cycled
quite
a
few
bugs
during
that
cycle,
but
there's
a
pretty
stable.
B
We
finish
the
support
for
gzip
EIP.
I
in
my
case
bootstrap
now
did
not
get
destroyed,
but
we
need
more
confirmation
and,
if
that
happens,
for
everybody,
that's
odd.
Google
thinks
now
that
there
is
no
official
image
uploaded
to
the
door
OS
controlled
bucket.
Yet
so,
for
now
you
would
have
to
upload
it
yourself
manually
and
override
the
path
to
its
using
a
show
and
bar
and
I
also
be
showing
the
install
process
that
later
today,
next
o
beards,
Rev
and
all
things
like
should
also
work.
There
isn't
one
park.
B
A
B
I
didn't
try
the
UPI,
but
since
UPI
is
really
much
worth
IPA
does
with
modifications
that
should
also
work.
Of
course,
that's
more
complex.
The
biggest
pain
point
right
now
is
a
sure,
because
we're
still
blocked
on
tickets
from
Fedora.
Hopefully
those
would
get
results
soon
on
the
better
status.
We
have
a
better
milestone.
B
B
Next,
on
the
official
release,
where
the
main
criteria
is
that
we
don't
use
poor,
installer
and
m-city
and
we
test
upgrades,
the
upgrades
are
fairly
easy
to
enable.
We
don't
do
that
so
that,
because
we
want
to
push
took
away
first
and
once
it's
done,
we
will
enable
upgrades
and
have
them
properly
tested.
The
forks
is
a
bit
more
complex.
We
need
to
work
with
installer
team,
only
approach
that,
but
we
started
working
on
installer
nm
co-integration
now
so,
hopefully,
if
okay.
C
B
A
C
Bare
metal
I'm
getting
ready
to
publish
fairly
comprehensive
Doc's
on
how
to
do
it
in
my
github
page
that
it's
targeted
to
my
internal
team
here,
so
they
can
so
it'll
it'll
have
a
lot
more
instruction
than
you
guys
would
need,
because
it'll
tell
you
how
to
install
DNS
and
set
up
your
vine
server
and
everything.
That's.
C
For
just
everyone
yeah,
so
probably
this
weekend,
I'll
have
it
done.
If
you
troll
around
in
the
project
on
my
github
site,
the
docs
are
all
in
there,
but
the
readme
said
the
readme
doesn't
link
to
them
yet
because
there's
still
kind
of
work
in
progress
and
I'm
working
now
on
making
it
work
with
fixed
IP
addresses,
but
we
may
have
run
into
an
F
cause
bug
or
something,
because
it's
not
keeping
the
fixed
IP
addresses,
but
it
works
great
with
DHCP
reservations.
B
D
D
But
they
don't
have
to
be
bashful.
That's
why
they're
declarative
it
could
just
be.
You
know,
read
config
things
they're,
not
like
they're,
not
like
Etsy
network
interfaces,
which
is
actually
a
shell
script
that
has
to
be
executed
to
run,
which,
by
the
way,
if
you
never
knew
that
today,
you
learned.
B
A
B
B
A
So
what
would
I
would
I
would
like
to
see
is
if
you
can
flag
the
whole
group
and
especially
me
when
the
overt
thing
gets
done
and
and
resolved
properly,
that
we
we
do
some
sort
of
beta
announcement
with
or
without
a
sure,
because
if
we
seem
to
have
some
consensus
here,
that
we've
got
got
a
good
cross-platform
support
and
if
we
could,
we
could
aim
for
next
week.
That
would
be
great
to
do
that.
A
A
C
A
Know
because
it
would
be
nice
to
have
the
announcement
then
coincide
it
with
an
external
person.
Writing
a
blog
post
about
it
and
I'm,
not
sure
where
we
would
publish
it,
but
yeah
and
Chris
is
offering
some
editorial
help
as
well.
So
I
think
I
know
part
of
my
problem
with
getting
the
Fedora
magazine.
Article
out
is
just
time
and
having
at
least
a
minimal
amount
of
content,
I'm
good
at
editing
things
in
shape-shifting
things
and
getting
the
messages
right.
A
D
A
A
A
C
A
Would
be
great,
then,
even
that
that
is
a
baseline
and
then
between
if
Chris
is
willing
and
myself,
and
hopefully,
Christian
gets
a
little
bit
better
if
we
can
just
get
everybody
else
to
add
in
their
two
cents
about
it
too,
and
and
have
some
content
to
get
out
there,
so
that
would
be
real.
That
would
be
a
really
nice
thing,
because
we've
had
so
many
crazy
things
happening
to
be
able
to.
You
know,
give
somebody
all
the
folks
out
there
in
virtual
self-isolation
something
to
play
with
over
the
next
few
weeks.
B
That's
the
final
result,
but
the
first
step
is
here
so
I
didn't
bother
with
a
proper
video.
I've
got
a
bunch
of
screenshots
and
a
lot
of
talking,
but
here's
what
we
have.
We
will
show
how
to
start
looking
for
GC
be
using
ivi
flo-jo.
If
you
know
an
issues
and
take
a
quick
peek
of
what
installer
is
actually
doing,
you
guys
should
be
pretty
different
because
you
have
to
set
up
things
yourself
and
our
case
you
want
to
take
care.
We
want
installer,
take
care
of
that.
So
here's
in
cell
conflict
template
I'm
using.
B
There
is
not
much
different
from
other
providers.
Actually,
the
difference
is
of
course,
platform.
Gcp,
your
project,
ID
and
the
region
you're
using
the
rest
is
pretty
much
the
same
I'm
using
it's
in
fact,
ginger
template
because
I'm
templating,
those
using
ansible
and
create
clusters
on
demand,
I
wrote
my
own
wrapper
around
the
Installer,
where
I
can
do
make
GCP
or
make
AWS
and
I
have
multiple
clusters
running
outside
nodes
that
it's
not
designed
to
keep
a
long-term
cluster
living.
B
B
B
B
Later
on
this
huge
terrible
command,
which
starts
the
Installer
from
the
image
we
have
just
pulled
so
we're
passing
it
like
I,
have
a
bit
of
line.
First
of
all,
we
mount
the
folder
with
our
cluster
case.
You
can
have
different
ones
running
in
a
while
and
we're
making
the
Installer
output
to
this
particular
directory.
B
B
So
the
first
thing
would
cube
service
would
do
is
to
upgrade
us
to
the
latest
image
from
our
release.
Payload,
it
would
extract
the
Machine
us
content,
pull
it
and
apply
it
as
a
Norway,
Street
content
and
then
finally,
reboot
to
after
the
reboot.
We
would
see
that
the
request
version
has
changed
the
latest
one
and
six
of
March
when
I
run
this
and
boot
cube
service
continues.
That's
basically
a
huge
difference
from
our
costs,
where
the
initial
image
does
not
peel
it
into
itself.
B
So,
during
the
process
you
can
on
the
bootstrap
node,
you
can
already
export
the
cube
config
from
this
directory
and
slash
of
OpenShift
start
watching
what's
happening
in
the
cluster.
Wants
it
assembles
the
API
service.
Oh,
this
is
pretty
small
sweet.
So
first
thing
we
see
is
that
the
version
number
there's
like
this.
That's
much
better
than
you
can
you'll
see
that
yeah.
B
So
what
we
see
here
is
that
the
first
operator
called
version,
it's
a
cluster
version
of
greater
started
progressing
and
it
has
started
the
network
operator
and
it's
also
progressing.
This
is
why
three
of
our
masters
are
not
yet
ready.
The
network
configuration
is
being
installing
them
and
we
have
tons
of
pods
hanging
and
pending,
because
nodes
are
not
yet
ready.
B
B
That's
expected
because
we
use
machine
API
to
create
them
dynamically.
We
define
three
machine
sets
each
of
them
in
different
availability
zone,
and
we
want
one
machine
in
each
and
after
some
time,
when
she
an
API
browser,
starts
progressing,
it
would
create
machines.
They
would
get
status
provisioning
that
one
is
not
yet
probably
processed,
but
later
we
would
see
that
new
workers
have
started
appearing
in
our
GCD
console
and
eventually
that
work
configuration
would
be
installed
on
them
as
Siri
files
simply
copied,
they
would
become
ready
and
more
operators
would
start
to
your
progress.
B
B
B
We
can
run
OC
status
and
yep
pretty
much.
That's
pretty
much
it
two
known
issues
again.
You
would
have
to
use
the
openshift
install
OS,
which
overrides
gives
us
not
yet
uploaded.
Once
it's
done,
we
will
update
the
Installer
to
point
to
correct
location
and,
in
my
keys,
oops
trapping
never
actually
completed,
rather,
if
they're
complete
but
open
shift
installer
on
my
site,
never
noticed
that
and
never
asked
to
destroy
the
bootstrap
node,
pretty
sure
the
problem
is
on
my
ant,
because
all
things
look
like
issue
work.
C
So
those
out
of
memory
killed,
pods
I,
don't
know
if
I
see
it
on
Etsy
d,
but
I
see
the
same
thing.
When
I'm
doing
the
UPI
install
in
my
lab.
It
doesn't
seem
to
adversely
affect
anything
because
the
correct
pods
do
appear
to
be
running,
but
if
that's
a
bug,
I
might
be
able
to
provide
some
additional
information
because
it's
happening
on
the
UPI
side
as
well.
Yeah.
B
There
is
a
Bugzilla
for
that
and
those
parts
they
update
the
cube
api
EEMA
and
once
they
are
done,
they
are
using
way
too
much
memory
because
we're
limiting
them
and
they
get
all
kills
because
they
have
used
all
of
their
living
okay.
So
it
shouldn't
affect
anything
you
can
affect,
though,
so
it's
a
bug
which
needs
to
be
fixed.
Of
course.
Yes,
but
since
so
kitty
is
just
a
Forex,
installer
and
MCO.
B
The
rest
is
the
very
same.
What
we
have
a
no
CP,
so
every
single
bug,
which
is
outside
of
the
Installer
failed,
builds
MCO
failed
to
put
the
proper
file.
All
of
that
goes
prey
to
OCP,
of
course,
but
we
would
like
you
to
notify
us
that
you
found
something
and
file
a
bug
and
opt
rifle
as
well,
so
that
we
would
hope
how
bad
things
are.
E
E
F
B
E
B
B
D
D
E
C
Yeah
Neil
in
the
in
the
lab
work
that
I've
been
doing,
even
though
it's
UPI
with
libvirt,
if
you
provision
just
with
the
vert
install
if
you
provision
another
machine,
pointed
to
booting
off
of
the
worker
ignition
config.
The
only
thing
you
have
to
do
is
approve
the
CSR
and
up
with
all
yeah
and
so
before.
I
left
the
for
the
weekend.
I've
got
a
cluster
running
at
home
with
the
usual
three
masters
and
six
workers,
and
you
can
just
keep
adding
workers
to
it.
Till
you
run
out
of
CPUs
and
RAM.
D
B
B
A
That
was
really
great,
I.
Think
that's
that's
awesome.
Work
I
just
wanted
I
think
you
said
that
earlier
in
the
meeting
that
the
okd
content
on
documentation
was
under
okd
preview,
I
I
believe
it's
actually
listed
as
okay.
The
latest
just
wanted
to
double
double
check,
as
opposed
to
the
word
preview.
A
B
D
G
D
That
like
having
the
late
of
having
a
section
latest
section
of
docks
that
actually
is
ok.
T4
is
really
really
good,
but
it
is
very,
very
weird
that
I'm
still
seeing
stuff
from
openshift,
3
and
ok
t3
in
the
latest
tab,
like
mini
shift,
is
not
a
thing
on
ok
t4
at
all.
We
don't
have
an
equivalent
to
mini
shift
right
now.
Yes,.
A
So
that
those
are
the
issues
that
I
would
ask
you
to
actually
log.
Ok,
if
you
find
them-
and
we
can
just
then
we'll-
have
a
list
of
things
that
are
odd
and
need
to
be
removed.
So
I
think
this
is
the
one
that
that
we
should
be
looking
at
the
ok,
the
latest
one,
that's
up
there
and
I
think
the
right
way
to
do
this
is
to
and
I
keep
trying
to
bump
to.
A
This
is
to
put
an
issue
in
the
open
ship,
ok,
d-list
and
then
you
know
it
says,
drop
section
on
adding
rail,
computes
or
whatever.
You
know
those
kinds
of
things
and
then
the
documentation,
team
and
the
other
folks
can
review
this
and
make
either
make
the
corrections
or
take
your
cut
and
paste.
If
you
have
a
sentence,
if
we
did
bad
grammar
cut
and
paste
that
in
as
well
with
the
correction,
that
would
be
helpful.
Ok,.
C
H
D
D
A
D
Don't
have
any
options
at
this
point.
That's
that's
the
problem
right
now,
we're
being
told
if
we
file
issues
on
the
dock
on
the
open
ship,
docks
repo
they'll
get
ignored
because
they're
on
okay
di,
oh,
so
we're
filing
them
in
okay,
Dion,
OpenShift,
okay
d,
but
that
doesn't
mean
a
whole
lot,
because
I'm
pretty
sure
the
docks
people
aren't
really
looking
at
those.
Somebody
has
to
go
and
prune
them
and
then
go
forward
them
back
to
the
docks
people,
and
this
feels
like
a
whole
lot
of
busy
work
for
no
particularly
good
gain.
So.
A
G
E
E
If
folks
will
will
look
an
issue
and
it's
not
that
can
both
log
that
as
okay
D
versus
OCP
and
then
obviously
we're
gonna
have
labels
mentioning
this
is
okay
T
and
this
is
OCP,
but
then
we
can
differentiate,
and
maybe
you
guys
can
you
know,
pay
attention
to
more
OCP
and
less
sources.
Okay,
D
and
vice-versa.
Obviously
based
on
your
workload
and
suchlike,.
E
D
B
Is
why
I
wanted
to
start
with
opt
repo?
First,
because
we
can.
We
can
aggregate
a
bunch
of
issues
and
on
a
huge
PR
to
open
ship
docks
and
get
more
attention
to
that.
Instead
of
multiple
small
PRS
from
contributors
which
we
don't,
we
don't
have
options
to
track
because
they
can
affect
both
OCP.
You
know
Keely,
so
was
the
process
which
we
came
up
with.
It
might
not
be
effective,
but
let's
discuss
that.
Well,.
B
Think
we
should
follow
the
latest
map.
This
is
why
the
latest
here
and
I'm
hoping
you're
posting
the
latest
codes
you
have
to
the
master,
so
one
so
Kitty
derives
from
that.
They
automatically
get
all
the
changes
and
the
contributors
would
update
master
and
eventually
I
would
make
two
open
share
for
the
four,
but
this
is
different.
D
D
E
This
I
would
say
for
folks
who
wants
to
see
the
master,
they
can
build
it
locally
and
they
can
locally,
they
can
see
and
render
and
real-time
they
made
changes
and
that's
how
do
you
normally
contribute
to
the
dock?
To
the
docks?
That's
how
I've
been
doing
3x
in
whether
the
the
the
current
master
is
pointing
to
two
dogs.
E
Latest
that's
a
different
topic:
cuz
I,
don't
ideally
personally.
One
of
the
frustration
I
had
in
3x
is
when
you
say
latest
latest
doesn't
mean
anything
because
if
you
switch
to
OCP,
they
don't
have
latest,
though,
ideally
it
would
be
nice
to
kind
of
follow
exactly
what
OCP
is
when
when
we
say
you
know,
4.3
it's
for
the
free,
and
that
is
latest
for
the
4
is
whatever,
rather
than
saying,
latest,
and
have
these
magic
stuff
being
done,
which
points
to
either
for
free
for
two
for
one
etc.
D
I
mean
one
of
the
key
differences
between
okay,
Dee
and
Ossie
P.
That
okt
is
essentially
rolling
forward
on
the
latest
code
anyway.
So,
ideally,
the
documentation
on
master
could
correspond
to
the
active
rolling
development
of
okd
as
it
stands,
because
then
at
some
point,
they'll
branch
it
and
then
that'll
become
OCP
and
ifs
and
a
quote-unquote,
stable,
ok,
d-sub,
release
or
whatever.
B
A
B
Right,
so
any
changes
would
be
rendered
on
Doc's
like
ad
I/o,
and
we
can
go
just
great
flying
bugs
just
a
matter
of
how
to
track
them
properly.
Should
we
go
straight
to
open
ship
docks
and
ask
and
label
them
with
Oh
kitty
or
add
some
kind
of
a
tag
in
the
title
which
which
process
what
you
prefer
you.
A
Could
just
put
okd
in
the
label
and
then
the
subject
line
of
the
propeller
girl
yeah
it
started.
Thank
you
in
the
title
of
the
thing,
if
you
did
ok
d
:
and
then
whatever
the
issue
was
you
found
going
back
to
where
he
was
showing
them
earlier?
Where
you
log
an
issue
I
think
that's
that's
the
proper
way.
G
A
A
A
A
So
I
think
that
the
basic
thing
here
is,
if
you
find
a
mistake
or
a
misstatement
or
grammar
or
whatever,
in
the
okay
D
documentation,
go-to
OpenShift,
doc,
docs
in
the
open
ship,
repo
log,
an
issue
with
okay
D
in
brackets,
that'll
flag
it
and
we
will
and
give
as
much
detail
as
you
can
about
where
you
found
the
mistake
and
how
you
found
the
mistake
or
the
correction
that
you
want
to
be,
have
made
and
submit
it
and
we'll
you
know
Michael.
We
need
to
come
back
and
have
another
call.
A
A
Anyone
else
make
any
new
discoveries
new
platform
things.
Okay,
then
I
am
going
to
call
this
meeting
and
thank
you
vetting
for
taking
on
the
GCP
demo.
That
was
great
I'll,
probably
try
and
sniff
that
out
as
a
short
video
as
well
and
share
that
and
Craig
for
your
medium
blog
post.
That's
wonderful
and
everybody
else
for
all
your
efforts
to
make
this
happen
and
I'm
really
looking
forward
to
getting
beta
out
the
door
so
enjoy
your
week
and
read
some
Doc's
sounds.