►
From YouTube: KubeCon EU Office Hour: OKD Working Group
Description
OKD is the open source upstream Kubernetes distribution that powers OpenShift, and we hosted an open office hours with a brief introduction to the OKD project and followed by OKD Working Group Members from UMich, Rohde & Schwarz, MarketAmerica, BCIT along with Red Hatters discussed how how they're using the distribution in their own organizations and take questions from the live audience.
For more information: visit http://okd.io and join the conversations in the OKD Google Group https://groups.google.com/forum/#!forum/okd-wg
A
Now,
good
morning,
good
afternoon,
good
evening,
wherever
you're
healing
from
welcome
to
another
special
kubecon
eu
2021
edition
office
hours
today,
I
am
joined
by
members
of
the
wonderful
okd
community.
I
would
like
to
thank
all
of
you
for
joining
today
and
I'm
actually
going
to
have
to
probably
turn
my
video
off
at
some
point
so
that
all
of
you
can
fit
in
the
screen
if
you
end
up
sharing
anything
on
screen.
A
C
D
F
Hello,
I'm
joseph
meyer,
I'm
a
cloud
architect
at
the
company
rhodey
and
schwartz.
It's
a
german,
unique
located
company.
We
are
in
a
telcom
and
I'm
using
okd
since
more
than
three
years
now
we
used
it
for
starting
our
digital
business
journey
and
we
used
a
lots
of
distributions
before
okd
and
yes,
okd
was
was
our
favorite
ones.
In
our
evolution
phase,
we
are
using
it
for
our
development
engineers
and
also
we
use
it
in
production.
Currently,
a
few
hundred
developers
are
on
it
and
it
has
great
potential
for
us.
It
yeah.
We.
F
G
G
G
H
B
D
D
D
That's?
What
we're
going
to
talk
about
here,
we're
going
to
give
a
quick
overview.
Give
you
a
quick
update
on
current
state,
where
we
are
talk
a
bit,
but
not
a
lot,
because
we'll
save
it
for
some
of
your
q,
a
about
what
what
differentiates
okd
and
openshift
from
some
other
kubernetes
distributions
and
finally
give
you
some
info
on
how
you
can
connect
with
the
working
groups,
both
the
okd
working
group
and
our
fedora
core
os
working
group
so
okd.
D
It
is
a
community
distribution
of
kubernetes,
end
of
story
that
it
is
a
distribution
of
kubernetes.
So
it
is
100
kubernetes
right.
This
is
not
a
fork
of
kubernetes
or
something
slightly
different.
This
is
all
kubernetes
that
that's
a
question
that
we
get
from
time
to
time.
It
is
also
open
shift.
So
so
any
of
you
that
might
be
red
hat
subscribers
to
ocp
the
openshift
container
platform.
D
D
D
D
They
are
going
to
be
able
to
bootstrap
themselves.
I
don't
know
if
any
of
us
have
experimented
with
that.
Yet
I
know
that
I
have
not,
I
hope
too
soon,
but
that's
something
I'm
really
excited
about
we're,
also
getting
a
lot
more
collaboration
now,
really
starting
in
our
4.6
release,
but
it's
accelerating
in
our
4.7
release,
with
a
lot
of
the
providers
of
the
operators
that
you
find
in
operator
hub.
D
I'm
going
to
talk
for
just
a
minute
about
the
operator
pattern,
because
this
is
one
of
the
things
that
that
really
differentiates
okd
from
other
kubernetes
distributions
is
that
from
the
bottom
up,
okd
is
based
on
operators.
The
very
first
thing
that
starts
running
on
that
bootstrap
node
to
get
a
cluster
up
and
running
is
an
operator
operators
manage
the
entire
ecosystem
operators
control
the
updates.
They
control
the
versioning
they
control
and
maintain
the
health
of
your
cluster.
D
A
D
Maintenance
of
one
of
these
clusters
much
easier
than
what
you
may
have
experienced
in
a
more
traditional
kubernetes
environment
and
again,
like
I
said
operator
hub,
is
the
place
to
get
just
about
anything.
You
want
to
run
in
your
cluster
and
as
we
continue
to
work
with
our
companion
projects
that
provide
the
operators
for
the
ocp
platform,
you'll
see
more
and
more
red
hat
official
operators
showing
up
in
the
operator
hub,
as
well
as
a
host
of
community
provided
operators.
D
D
This
is
our
our
slack
channels,
our
google
group.
We
have
bi-weekly
meetings
that
you
can
find
at
that
fedora
project
calendar
site
and
we
maintain
a
couple
of
repositories
that
we
track
our
issues
and
features
and
and
things
of
that
nature
in
fedora
core
os.
By
the
same
token
and
I'll
leave
this
slide
up
for
a
few
seconds
again
has
a
very
active
community,
and
timothy
is
here
with
us
today.
D
So
if
there
are
any
fedora
core
os
specific
questions,
he
is
the
man
to
answer
those
for
you
here
are
a
few
more
resources
that
you
can
grab
I'll
leave
this
up
for
a
few
seconds
for
any
screenshots,
and
because
this
is
an
office
hour
and
not
a
slight
presentation.
I
talked
very
fast
and
blew
through
these
slides.
So
let's
get
to
the
office
hour
and
answer
your
questions
thanks.
Everybody.
B
Okay,
so
we
have
questions
all
over
the
map
from
both
existing
okay
to
users
and
from
folks
who
are
looking
at
it
because
of
this.
One
of
the
first
ones
that
came
in
here
was
from
sanjay,
who
is
interested
in
using
dns
mask
I
as
a
service
or
something
else
for
caching
dns
inside
okd,
and
wants
to
know
if
there's
anybody's
experience
with
that,
if
somebody
has
put
up
a
how-to
somewhere.
D
That
one
may
be
dumper
I
I
have
not.
I
have
not
attempted
to
do
anything
with
that.
Yet
I'd
be
interested
if,
if
he
could
type
it
in
asynchronously
in
what
the
specific
use
case
is
what
the
need
is
for
a
dns
cache,
because
I
have
it
personally
in
my
lab
or
or
even
in
the
production
ocp
clusters
that
I
that
I've
worked
with
come
across
a
scenario
where
the
native
dns
provider
of
the
cluster
itself
wasn't
sufficient.
B
Okay,
I've
asked
him
to
expand
on
his
use
case,
so
we've
got
some
other
more
sort
of
basic
questions
about
installation.
One
was
somebody
who
wants
to
try
out
okd
and
they're
having
some
issues
with
the
installation
being
complicated
earlier.
You
talked
about
the
bootstrapless
installation.
That's
coming
is
that
is
that
sort
of
your
me,
our
main
installation,
simplification
effort.
E
If
you're
running
into
issues
check
it
out,
because
there's
a
lot
of
the
sort
of
the
nuanced
detail
that
you
might
not
find
in
in
the
the
standard,
documentation,
standard,
okd
documentation,
and
it
provides
some
background
on
prerequisites.
For
example,
that's
really
helpful.
I
think
in
the
case
of
the
vsphere
upi,
I've
actually
got
a
repository
with
a
script
that
does
the
whole
process
for
you
pretty
much
after
you
set
a
couple
variables.
C
And
I
I
would
just
put
a
pitch
in
our
next
okd
working
group
is
next
tuesday
at
1700
utc
and
it's
open
to
everybody.
And
if
you
go
to
okd.io,
you
can
find
the
guides
in
the
nav
bar,
and
you
can
also
find
the
information
to
navigate
to
all
the
repos
latest
release
and
everything
else
and
hopefully
find
your
way
to
the
okd
working
group
forum
on
google
as
well.
I
put
that
in
the
chat
earlier.
C
C
Technical
support,
supported
thing
so
and
the
the
good
news
is
it's
a
really
active
community
and,
as
you
can
see
by
all
the
folks
that
are
here
today
as
well
as
we've
got
great,
a
great
collaboration
going
on
with
the
fedora
coreos
folks,
because
I
really
do
want
to
emphasize
okd
runs
with
fedora
core
os
as
opposed
to
rail
core
os.
So
it
is
a
very
different
we'd
like
to
say
sibling
upstream
or
sibling
stream
to
okay
ocp.
G
That's
right,
hey
one
thing
I
can
say
with
installation
is
that
it
you
can
make
it,
as
you
know,
easy.
As
you
know,
a
20
or
30
line
configuration
file
to
as
complex
as
you
want
it
to
be.
I
run
ipi
so
pretty
much.
I
have
a
set
of
one
configuration
file
that
I
can
repeat
over
and
over
and
over
again,
and
if
I
have
to
change
something,
it's
just
a
small
tweak.
B
D
B
Take
a
look
at
that
and
if
you
have
follow-up
questions
you
can
ask
them
later
on.
The
are
our
dns
sanjay.
Just
got
back
to
me.
He's
basically
he's
got
services
that
make
heavy
use
of
dns
and
he's
looking
to
not
overwhelm
core
dns
in
his
infrastructure
for
application,
dns
requests
and
also
maybe
get
better
response
times
by
guess,
having
pods
that
are
local
to
each
node
doing
dns,
and
I
was
wondering
if
anybody
has
any
advice
there
related
to
this.
Since
there's
a
community
on
chat
user.
D
D
B
B
B
F
D
If
if
they
mean
that
their
underlying
hypervisor
provider
is
centos
stream,
I
just
updated
my
lab
environment
to
centos
stream.
The
operating
system
of
okd,
though,
is
fedora
core
os,
and,
and
we
don't,
we
don't
support
any
other
operating
systems
currently
other
than
fedora
core
os
running
on
the
master
nodes
and
the
compute
nodes.
E
Yeah,
so,
basically,
just
to
give
a
little
bit
of
background
on
it.
There's
a
machine
configuration
operator
and
different
parts
of
okd
that
talk
directly
to
the
os
and
and
manage
the
updates
and
they're
pretty
tied
to
each
other,
and
so
there's
more
documentation
on
that
on
our
website.
But
if
you
want
more
nuanced
detail
on
sort
of
how
they're
all
connected
feel
free
to
pop
into
the
channels
or
look
over
the
document.
H
H
B
So
a
couple
other
things:
one
thing
is:
I
love
these
community
things
I
just
I
want
to
call
out.
We
had
a
question
in
here.
Another
community
answer
answered
it
right
with
somebody
asking.
Is
there
a
way
to
install
okd
on
poxmox
as
a
virtualization
platform
yeah,
someone
else
actually
jumped
into
the
chat
and
said
they've
already
done
that
so
yes,
what's.
B
Cool
okay:
here's
a
more
technical
question
which
I'm
going
to
read
out
directly
because
I
don't
get
it
wrong,
which
is
niels
wants
to
know,
is
the
okd
ipi
process
the
same
as
the
ocp
ipa
process,
they're
using
okd
ipi
as
a
test
platform
to
test
a
migration
from
upi
to
ipi
installations
on
vsphere,
and
they
want
to
know
when
they
do
this
in
production.
Can
they
use
the
same
process
for
ocp.
G
Yeah
I've
done
both,
so
I
took
my
configuration
file
for
okd
and
with
a
couple
of
very
small
tweaks
to
make
sure
I
was
pointing
at
the
ocp
stuff
work
like
a
charm,
okay,
I
think
you
can
even
use
the
same.
Well,
I
don't
know
I
used,
I
use
the
the
ocp
installer
and
the
okd
installer,
I'm
not
sure
if
you
can
switch
those
out
and
have
them
installed,
but
the
process
itself
is
the
same.
B
F
F
G
The
amount
of
resources
needed
to
run
okd.
I
think
that
once
we
get
to
that
single
node,
okay
d
for
people
use,
you
know
you'll
find
testing
wise
you'll
be
able
to
do
a
lot
more
for
a
straight
install
of
okd.
You
know
even
for
testing
it's
still
fairly
beefy.
You
have
to
have
fairly
beefy
nodes.
So
that's
one
of
the
things
you
have
to
keep
in
mind
with
whatever
distribution
you
decide
to
use
is:
do
you
have
the
utilization
or
the
resources
you
know
to
be
able
to
run
it.
D
Yeah
the
the
trade-off
there
is
that
you
know
that
all
of
the
automation
and
self-healing
that
comes
with
okd
right
that
does
require
compute
and,
and
so
so
okd
is
a
bit
fatter
from
its
resource
requirements
than
you
know.
A
plain,
vanilla,
kubernetes
or
or
another
container
supporting
platform,
but
the
the
the
trade-off
is
a
self-managing
environment
that
that
really
gives
you
a
cloud-native
experience.
In
my
opinion,.
G
Absolutely,
which
is
one
of
the
reasons
why
we
chose
it.
I
don't
have
time
to
manage.
You
know
every
five
minutes.
You
know
the
the
cluster.
I
can
turn
over
a
lot
of
stuff
to
my
developers.
They
get
a
lot
of
authority
to
do
things
within
their
projects
and
I
don't
have
to
manage
it.
They
can
manage
it
themselves.
E
Yeah,
I
would
echo
that
I
think
that
you
know
from
the
devops
perspective
that
sort
of
okd
and
openshift
provide
that
sort
of
middle
layer.
That's
really
good!
If
you
look
at
some
of
the
other
container
orchestration
options
out
there,
they
involve
a
lot
of
glue,
scripts
and
stuff
that
you
have
to
bring
to
the
table,
and
I
think
okd
provides
a
lot
of
that
and
makes
it
very
easy
for
developers
to
sort
of
slide
in
to
the
role
of
a
container
developer
and
devops.
So
it's
really
good.
D
Yeah
and
with
the
maturity
now
of
projects
like
rook,
ceph
and
cube
vert,
you
know
which,
which
also
have
productized
versions,
that
you
know
that
you
can
get
subscription
support.
For
I
mean
my
goodness,
you,
you
have
a
full
hyper
converged
cloud
native
environment
that
that
can
replace
most
of
what
you
find
in
a
data
center.
H
Is
that
strong?
A
strong
point
we
have
with
will
kitty.
Is
that
so
that's
my
fedora
cross
hat
on
that
we
test
the
full
os,
the
full
everything
that
comes
with
the
os
with
everything
in
the
okd
platform.
So
what
you
get
is
when
you
update
from
one
version
to
another
of
okay,
you
get
a
platform
which
has
been
fully
tested
and
to
end
the
os
and
everything
on
top
of
it,
and
you
practically
show
that
everything
works
together,
which
is
like
something
that
only
happens
in
okd
and
ocp
platforms.
B
H
Sure
so
rpm
is
3,
it's
it's
the
company,
the
main
component
of
federal
course
node,
and
that
helps
us
manage
versions
of
the
node.
So
when
you
update
your
kde
cluster,
you
update
all
the
components
of
the
kubernetes
components
and
you
also
update
the
nodes
at
the
same
time
and
to
do
that
you
use
rpms3
and
what
it
does
is
essentially.
Rpms3
is
a
git
way
to
manage
the
operating
system.
H
So
you
move
from
one
version
of
the
other
of
of
your
fedora
course
node
and
you
you
just
pull
a
new
image
and-
and
you
essentially
reboot
to
it
to
update,
and
so
that
has
a
lot
of
attention
advantages
so
that
the
first
one
is
that
we
actually
test
all
those
matches.
We
have
several
streams
in
federal
careers
and
we
make
sure
that
they
work
and
then
the
second
one
is
that
every
single
update
you
do
on
your
system
is
atomic
and
it
just
either
happens
or
does
not
happen
at
all.
H
B
C
Ocm
is
the
upstream
project
of
acm
for
those
of
you
who
don't
know,
and
it's
we're
open
sourcing,
all
of
that
and
you
can
find
it
what
I
think
this
is
a
great
topic
for
in
maybe
a
future
office
hour
to
demo
this
bits,
because
the
folks
from
the
teams
have
been
reaching
out
to
the
okd
teams
and
would
like
some
airtime
too.
So
I
think,
there's
going
to
be
a
nice
bit
of
collaboration
there
yeah.
C
The
benefit
of
the
okd
working
group
is,
we
are
on
the
bleeding
edge.
We
test
all
of
this
stuff
for
people.
So
if
you
have
a
new
project
or
new
operator-
and
you
want
it
tested,
come
to
the
okd
working
group
because
we
are
we
love
being
guinea
pigs,
maybe
not
for
some
of
your
production
environments
guys.
But
I
know
that's
one
of
the
things
that's
really
useful.
The
upstream
open
source
side
of
openshift
is
to
get
stuff
tested
over
by
the
okd
working
group.
C
I'm
going
to
be
quiet,
but
I
can't
there
is
a
whole
sig
around
edge
and
working
groups
reach
out
on
the
okd
forum
in
the
google
group
and
post
that
question
and
we
will
find
the
person
who
has
the
answer
to
that.
I
don't
think
anybody
here
has
really
done
that,
but
we
have
talked
about
edgy,
stuff,
edge,
related
stuff
and
arm
as
well
as
arm
stuff
in
the
okd
working
group.
So
there
are
people
who
are
covering
those
topics.
B
Yeah,
which
is
still
a
relatively
new
project,
and
one
of
those
things
we
should
be
doing
more
sessions
about
so
people
know
how
to
use.
I
don't
know
how
to
use
it.
I'd
like
to
do
a
session.
For
that
reason,
let's
see
a
few
things:
somebody's
actually
looking
for
best
practices,
information,
particularly
when
installing
on
bare
metal
in
terms
of
what
they
need
and
how
to
configure
it.
E
Yeah,
so
there's
so
best
practices,
I
think,
can
probably
could
be
construed
from
the
guides
that
we
have
and
the
documentation
we've
actually
tried
to
do.
Although
there's
a
lot
of
variety
in
upi
and
in
bare
metal
in
general,
the
guides
actually
have
a
some
good
practices
outlined
that,
basically,
you
could
determine
best
practices
from
those.
E
G
G
A
A
A
B
B
Yep,
yeah
yeah,
hey,
listen,
one
of
the
areas
I
used
to
consult
was
actually
you
know:
department
of
energy
power
generation,
etc
and
the
onboard
computers
they
have
on
power
equipment.
All
of
these
are
actually
real
hardware
constraints,
unfortunately
the,
but
we
have
questions
from
the
audience
actually
on
specific
features.
They're
features
that,
like
they
know
from
ocp
and
the
supported
versions,
they
want
to
know
what
the
equivalents
are
on
okd.
B
A
D
F
D
B
G
D
B
Yeah-
and
so
so
people
know
some
terminology,
we
know
a
lot
of
names
here.
So
one
of
the
things
that
we
do
at
red
hat
is
we
try
to
give
the
open
source
community
supported
version,
a
different
name
from
the
red
hat
subscription
version
so
that
people
don't
get
confused
and
so
for
openshift
service
mesh?
That
name
is
maestra,
which
is
istio,
plus
the
management
tools
and
ui
and
stuff
that
we've
developed
to
make
istio
easier
to
use.
C
I'm
not
sure
that
that
keeps
it
less
confusing
or
not,
but
keep
them
separate
in
terms
of
support
availability.
Well,
I
think
that's
really
the
key
here
yeah
I
I
tend
to
think
it
just
makes
everything
so
confusing
and
we
do
have
a
catalog
for
operators
for
for
okd.
So
if
you
go
into
the
repo
and
follow
the
links
there,
you
can
find
that
and
we're
working
on
getting
making
sure
we
have
equivalency
with
ocp
and
that
they're
all
tested
by
the
community.
C
So
if
you're
looking
at
our
guides
or
at
docs.okd.io-
and
you
find
something
that's
out
of
whack
or
not
quite
clear
enough
login
issue-
we
definitely
will
take
any
even
grammar
checking,
helps
us
and-
and
we
really
have
had
a
great
swell
of
testers
from
on
all
kinds
of
implementations
and
deployments
of
of
okd,
and
it's
really
helped
a
lot
driving
innovation
and
getting
ocp
and
the
ups,
the
the
product
itself
very
stable.
So
it's
really.
C
The
engineers
at
red
hat
have
been
incredibly
supportive
of
the
okd
community
and
you
know
shout
out
to
vadim
and
christian,
and
you
know
timothy
and
other
folks
who
have
really
you
know
a
lot
of
their
spare
time:
weekends
and
others
getting
releases
out
and
helping
the
community
move
things
forward.
So
we're
incredibly
grateful
for
that
support
too.
C
But
this
is
really
a
very
community
driven
effort
and
it's
you
know
any
and
all
feedback,
and
if
you
have
some
strange
interesting
contortion,
a
stack
that
you
want
to
write
a
guide
about
or
if
you
have
an
alternative
to
one
of
the
routers
we've
been
using
or
a
different
way
of
doing.
Dns
just
hit
us
up
in
the
the
working
group
forum
on
google
or
in
the
kubernetes
slack
on
openshift
dev,
we're
all
hanging
out
there
most
of
the
time
and
we're
pretty
global
in
our
coverage.
I
B
Speaking
of
community
support,
so
we
have
a
highly
technical
question
in
the
streaming
chat.
I
don't
know
if
any
of
you
guys
are
actually
logged
into
the
streaming
chat.
If
not,
we
might
want
to
recommend
that
he
actually
take
that
somewhere
else,
because
this
is
this
is
not
something
I
even
read
aloud.
It
has
network
errors
and
stuff
in
it
yeah
I
and
and
definitely
looks
like
an
occasion
for
community
troubleshooting.
E
I
would
say
if
the
person
could
take
that
to
the
channel
or
the
email
list,
that
we
have
the
the
kubernetes
based
slack
channel,
that
we
have
openshift
dev
on
the
kubernetes
slack
server
or
our
email
list
either
one
of
those.
Then
we
can
take
a
look
at
it
for
sure
yeah
yeah.
We
try
to
avoid
doing
troubleshooting
at
that
level
during
office
hours,
because
sometimes
that
means
bouncing
around
documentation
and
and
things
like
that.
But
we're
happy
to
help
for
sure.
E
I
C
D
D
E
We
get
questions
like
this,
where
maybe
none
of
us
currently
in
the
group
have
experience
with
it,
and
this
isn't
everyone
in
the
working
group.
But
if
we
don't
have
experience,
if
there
is
a
someone
out
there
in
the
world
who's
asking
us,
if
you
could
try
it
and
and
let
us
know
how
it
goes,
that
would
be
awesome
because
then
we
can
write
up
some
documentation.
F
F
It's
much
more
suited.
I
think
we
also
get
these
questions
about
windows
containers
all
the
time
and
I
always
recommend
cupid,
because
keyboard,
I
think,
is
yeah.
It's
a
vm
running
in
your
kubernetes
cluster
and
you
can
run
any
windows
version
you
you
like,
and
I
think
it's
more
capable
for
migration
scenarios,
at
least
than
windows
containers,
because
you
client
can't
pack
anything
in
windows,
containers.
A
A
C
If
not
one,
one
of
the
things
I
want
to
say
is
as
as
someone
who's
been
playing
with
since
the
early
days
of
origin
in
okd.
I
am
really
grateful
to
all
the
folks
who
are
on
this
call
and
all
the
other
folks
in
the
okd
working
group
that
are
community
members
who
have
come
and
given
their
feedback
over
the
years
and
really
encourage
everybody
sends
its
open
source.
We
don't
know
that
you're
out
there,
it's
not
gated.
C
We
have
no
clue
who
you
are
so
if
you
raise
up
your
hand
and
and
join
the
okd
working
group,
we'd
love
to
have
you
there
and
ask
your
questions
again
in
the
kubernetes
slack
in
openshift
dev,
and
this
has
been
one
of
the
most
wonderful
and
rewarding
groups
of
people
to
work
with
over
the
past
few
years.
It's
really
grown
exponentially,
and
you
know
the
feedback
we've
gotten
and
the
testing
and
the
deployment
marathon
we
did
a
year
ago
and
then
another
okd
testing
and
deployment
workshop,
we're
all
community
driven.
F
C
Openshift
there's
a
whole
playlist
of
videos
of
people
walking
through
deployments
for
okd
and
they're
really
great
and
the
guides
are
a
great
supplement
to
those
and
we
really
are
looking
for
you
know
I
keep
saying
it
other
permutations
of
how
you
want
to
deploy
okd
and
that
and
just
stretching
the
edges
of
where
openshift
lives
and
breathes
out
there.
So
please
do
join
us.
We're
really
one
of
the
most
receptive
and
included
groups
of
folks.
There's,
no
silly
questions
only
silly
walks,
you
never
walk
alone,
walk
with
the
ocado.