►
From YouTube: Ask an OpenShift Admin (E95) | Agent Based Installer
Description
OpenShift's installation options continue to grow, with more and more flexibility based on the type of platform you're installing to and the environment your cluster will be running in.
The agent based installer is one example of this, providing an experience to facilitate rapidly getting the first cluster up and running in your environment so that you can establish a foothold and begin expanding.
Want to learn more?
Try Red Hat OpenShift: https://red.ht/3uVxkya
Follow us:
Andrew Sullivan https://twitter.com/practicalandrew
Jonathan Rickard https://twitter.com/jrocktx1
OpenShift https://twitter.com/openshift
Subscribe to Red Hat's YouTube channel: https://www.youtube.com/redhat/?sub_c...
About Ask an OpenShift Admin:
Do you have questions? We want to answer them! Every Wednesday at 11am ET, hosts Andrew Sullivan and Jonathan Rickard sit down to answer your questions about Red Hat OpenShift.
About OpenShift:
Red Hat OpenShift is an open source container application platform based on the Kubernetes container orchestrator for enterprise application development and deployment.
OpenShift on redhat.com https://red.ht/3rYz5ZS
OpenShift on GitHub https://github.com/openshift/
#RedHat #OpenShift #Kubernetes #ContainerPlatform
A
A
A
Here
we
go
too
many
tabs
open,
as
is
sometimes
the
case,
so
I
I
thought
I
had
muted,
all
of
the
other
YouTubes
and
twitches
and
everything
else,
but
yeah
turns
out.
I
had
one
of
them,
so
I
was
hearing
myself,
which
is
incredibly
distracting
for
anyone.
Who's
ever
had
to
go
through
that
so
yeah,
hello,
everyone
and
welcome
to
this
week's
stream.
So
this
week
I
am
by
myself.
A
A
So
today's
episode,
or
today's
stream,
rather
is
going
to
be
focused
on
the
agent
based
installer,
so
agent
based
installer,
is
something
that
has
been
in
progress
for
a
while
and
was
recently
G8
alongside
openshift
4.10.
So
this
one
is
interesting.
It's
kind
of
it's
a
little
bit
different
experience.
You
know.
We've
talked
a
lot
about
the
install
experience
here
on
the
stream
over
the
last
two
and
a
half
years,
gosh
more
than
that.
It's
almost
been
three
years.
I.
A
Think
since
we
started
streaming
march
today
is
March
1st
I.
Think
the
first
of
the
live
streams
here
on
red
hat
live
streaming
were
in
early
April
I
want
to
say
like
or
no
maybe
maybe
it
was
May
I
want
to
say,
May
4th,
because
I
remember
somebody
making
a
joke
about.
You
know
May
the
4th
May,
the
4th
so
yeah
it's
been,
it's
been
quite
the
exciting
three
years.
A
But
I
was
really
pleasantly
surprised,
as
I
was
going
through
and
testing
things
and
experimenting
with
it
and
learning
all
this.
So
if,
if
I'm
being
honest,
it's
kind
of
a
boring
demo
and
that's
a
good
thing,
the
install
experience
should
be
easy.
It
should
be
low
effort.
It
should
be
not
exciting
if
you
will
I
I
always
find
it
interesting.
A
Strange
hilarious
when
it's
like
just
getting
something
installed
is
exciting,
like
no
that's
the
least
exciting
part
of
the
whole
experience
like
everything,
after
that,
all
of
the
openshift
goodness
and
everything.
That's
that's
what
we
really
want
to
talk
about
and
focus
on,
so
yeah
I'll
definitely
show
a
demo
like
I
said
it's
there's
a
couple
of
nuances,
I'll
say,
but
it's
really
straightforward.
A
So
before
we
get
touched,
I
will
do
a
quick
couple
of
top
of
Mind
things,
but
I
also
want
to
remind
everyone
who
is
with
us
here
that
this
is
one
of
our
office
hours.
Series
of
live
streams
here
on
red
hat
live
streaming,
so
it
means
that
you
can
ask
a
question
about
anything
that
comes
to
mind.
Anything.
That's
bothering
you
that
you're
having
problems
with
that
you're
just
curious
about
feel
free
to
ask
those
questions
and
chat
and
we'll
go,
do
our
best
to
get
to
those
here
on
the
stream.
A
A
So
you
can
find
me
at
andrew.sullivan,
redhat.com
or
at
practical
Andrew
on
the
social
media
platforms.
I
will
say
that
these
days,
email
is
almost
easier
for
me.
I
just
realized
this
morning,
because
it's
the
first
time
that
I've
done
any
Tweeting
in
quite
a
while
Johnny's
been
handling
a
lot
of
that
stuff.
That
I
had
a
whole
bunch
of
like
unread
messages.
So
my
apologies
to
anybody
who
has
sent
me
anything
I
think
there
was
some
from
as
far
back
as
like
last
December,
I
I,
hope
I
have
responded
to
everyone.
A
A
We'd
love
to
hear
from
you
all
our
hope,
not
I,
don't
know
if
you're
on
I
got
your
email
this
morning,
I've
already
sent
it
off
and
it's
already
over
with
engineering
where
they're
validating
the
the
issue
that
you
encountered,
which
was
around
a
incorrect
URL,
being
specified
for
the
insights
operator
and
whitelisting.
So
they're
validating
everything
and
they'll
they'll
get
all
that
updated.
So
they
move
very
quick
on
on
that
one.
It
was
literally
within
like
two
minutes
of
me
sending
the
email,
like.
A
Oh
yeah
we'll
check
on
this
and
then
five
minutes
later
yeah.
This
looks
like
it.
It
changed,
but
we're
not
sure
why
you're
seeing
it
now
and
the
conversation
continued
from
there.
So
all
right
top
of
Mind
topics
got
way
too
many
tabs
open
as
usual.
So
first,
let
me
share
my
screen
and
the
first
one
I
want
to
talk
about.
Is
this
openshift
virtualization
tuning
and
scaling
guide?
A
So
if
we
scroll
down
here
like
tested,
maximums,
here's
their
tested
maximums
for
virtual
machines,
you
know
10
000,
concurrently,
running
virtual
machines
inside
of
a
single
cluster,
but
I
thought
this
was
interesting
down
here.
So,
like
scheduling,
I
did
not
know.
For
example,
that
you
can
change
this
node
status,
Max
image
count
in
order
to
affect
scheduling,
so
they
walk
through
the
explanation
here,
but
essentially
one
of
the
things
that
affects
the
scheduler
score
for
any
particular
pod
is
whether
or
not
the
image
that
that
pod
uses
already
exists
on
the
Node.
A
So
if
I
have
you
know,
10
nodes
in
the
cluster
and
each
one
of
those
nodes
is
running,
you
know
100
different
pods.
That
could
be
a
hundred
different
images
and
if
I
schedule
a
new
pod,
it
will
factor
in
whether
or
not
you
know
maybe
two
of
those
10
nodes
already
has
the
image
there.
What
would
more
heavily
favor
those
two
nodes
but
and
as
we
see
here,
the
default
number
of
container
images
that
it
uses
to
to
adjust
that
or
factor
into
that
scheduling
is
only
50..
A
A
Basically,
unlimited
so
that
redrings
back
all
of
the
images
and
it
plays
into
that
score,
I
will
say
like
everything
or
anytime,
you
do
any
kind
of
Performance
Tuning
or
anything
like
that,
always
be
cognizant
of
behavior
before
and
after
so
what
do
I
mean
by
that?
A
long
time
ago,
I
was
a
DBA,
and
you
know:
I've
been
a
storage
admin,
a
virtualization
admin.
A
You
know
every
vendor
ever
has
these
types
of
performance
and
tuning
guides
and
it's
important
to
understand
what
your
performance
Baseline
is
now
and
then
make
a
change,
preferably
only
one
at
a
time
and
then
establish
what
that
change
or
how
that
change
affected
your
environment.
So
before
I
was
seeing
X
number
of
iops
I
made
a
change.
Was
that
change
positive?
Did
it
impact
you
know
positively
or
negatively
against
what
was
happening
in
your
cluster?
Did
it
have
the
desired
effect?
A
A
This
is
what
this
position
is,
or
this
is
what
this
type
of
thing
means
and
stuff
like
that.
But
internships
are
ultimately,
you
know
who
gets
selected
is
ultimately
based
on
the
manager.
That's
hiring
that
intern
and
I
am
a
manager,
but
I
don't
have
any
internships
on
my
team,
unfortunately
so
anyways.
So
as
you
flip
through
this,
as
you
look
through
this,
let
me
share
the
URL
for
this,
so
I
should
have
done
that
earlier,
shouldn't
I.
A
So
as
you
look
through
this
particular
article,
definitely
take
an
eye
to
you
know
just
all
the
different
things
that
can
be
in
here.
So
it
is
specific
openshift
virtualization,
but
there
were
some
things
in
here
that
I
haven't
seen
before
you
know
here:
Cube
API
burst
queries
per
second,
here's,
some
VM
tuning
right
so
on
and
so
forth,
so
it
even
if
you're
not
using
openshift
virtualization.
It
may
have
some
interesting
stuff
in
here
for
you
just
about
other
aspects
of
openshift
of
kubernetes,
and
you
know
how
to
manage
those.
A
A
So
the
other
thing
that
I
wanted
to
talk
about
here
in
the
top
of
Mind
topics
is
around
vsphere
and
in
particular
we
get
asked
fairly
regularly.
You
know
by
customers
who
are
deploying
openshift
onto
vsphere.
They
ask
their
account
team
who
comes
back
and
and
asks
us
on
the
openshift
team.
You
know
hey,
can
VMware
or
can
openshift
control
and
manage
things
like
Affinity
or
anti-affinity
rules
for
openshift
nodes
and
the
answer
to
that
is
at
least
for
now.
A
No,
so
the
rationale
being
a
couple
of
different
things,
so
one
we
don't
necessarily
understand
what
the
goals
or
the
configuration
the
the
desired
state
of
that
vsphere
cluster
is
according
to
what
the
the
administrator
wants.
So
if
we
were
to
just
go
in
and
say,
oh
yeah
I
want
to
create,
you
know
a
new
anti-affinity
group
or
a
new
affinity
group,
and
you
know
set
it
to
anti-affinity
against
these
hosts
and
all
that
that
could
be
completely
incorrect
based
off
of.
What's
you
know
the
the
administrator
wants.
A
You
know
you
know
running
or
pens
to
nodes,
one
two
and
three
of
any
of
fancy
anti-afinity
set
on
them
at
all,
so
we
don't
understand
necessarily
what
those
mean
and
we
need
to
introduce
a
lot
of
logic
in
order
to
help
Define
those
rules
and
stuff
like
that.
In
addition
to
that,
usually
the
first
complaint
or
the
first
pushback
we
get
back
on
those
is
simply
permissions.
A
You
know
vcenter
admins,
don't
necessarily
want
to
give
cluster
level
permissions
because
you
need
permission
at
this
level
in
order
to
set
or
create
those
rules
and
manage
those
rules.
So
a
lot
of
times,
that's
where
we
get
a
lot
of
pushback
as
well.
For
example,
I
had
a
conversation
with
some
folks
earlier
this
week
because
their
customer
didn't
want
to
have
tagging
based
permissions.
So
if
we
look
in
the
docs
and
go
down
to
12.
A
A
If
I
come
over
here
at
this
level
inside
of
my
vcenter,
it
needs
to
have
tagging
permissions,
so
these
tags
are
created
by
the
CSI
driver
by
the
machine
API
interface
in
order
to
basically
identify
what
resources
belong
to
a
specific
cluster
and
they
for
whatever
reason
in
the
VMware
worlds.
Those
exist
at
the
top
level.
A
So
this
customer,
you
know
very
understandably,
for
admins-
were
hesitant
around
giving.
You
know
that
type
of
permission
at
such
a
large
level
or
such
a
high
level.
Again
we
hear
the
same
thing
about
you
know:
DRS
rules
about
Affinity
rules
when
you
look
at
these,
so
if
I
come
in
here,
look
at
configure
and
I
go
to,
DRS
is
turned
on
and
now
I
don't
remember
how
to
get
to
Affinity
rules.
A
Here
we
go
host
groups,
so
in
here
I
can
create
host
groups,
so
maybe
I
have
you
know
one
or
all
the
way
up
to
five
of
these
hosts
that
I
want
to
put
into
a
specific
group
for
whatever
reason
and
then
I
create
a
VM
to
host
rule
where
I
can
then
either
associate
those
VMS
with
those
nodes
or
anti-affinity
prevent
them
from
running
on
those
prevent
them
from
running
together.
All
of
that
type
of
stuff.
So
again,
that's
that's.
A
A
Generally
speaking,
in
my
opinion,
most
everyone
will
want
to
have
anti-affinity
rules
for
the
control
plane,
notes,
meaning
we
don't
want
to
have
all
three
control,
plane,
nodes
running
on,
say
a
single
physical
server.
We
want
to
have
them.
You
know
ideally
distributed
across
one
on
each
each
one
on
a
different
node.
So
that
way,
if
a
single
physical
server
fails,
we
don't
lose
our
control
plane
quorum.
A
That
being
said
would
suggest
that
they
are
soft,
Affinity
rules
or
anti-affinity
rules,
not
hard
anti-finity
rules,
so
the
difference
is
a
hard
rule
means
that
it
will
not
turn
on
that
virtual
machine.
If
it
violates
the
rule,
a
soft
rule
means
allow
the
VM
to
run,
but
when
it
can
meet
that
that
requirements
or
that
take
action,
so
basically
DRS
in
that
case,
within
the
motion,
the
virtual
machine
to
different
node
in
order
to
regain
the
the
Affinity
status.
A
The
reason
why
this
is
important
is
you
know:
let's
say
that
I
have
you
know:
I
lose
several
of
these
hosts
and
now
I
don't
have
enough
hosts
to
actually
restart.
You
know:
I
lose
three
of
my
five.
Assuming
I
have
enough
capacity
to
tolerate
that.
Without
violating
the
hard
rule
it
wouldn't
be
able
to
restart
the
the
control
plane
node
with
a
soft
rule.
It
will
restart
the
VMS
that
way.
I
can
re-protect.
A
You
know
at
CD,
it'll
have
three
nodes
for
Quorum
and
then
as
I
reintroduce
Hardware,
it
would
continue
to
protect
things
so
infrastructure
nodes,
similar
type
of
decision
I,
think
it'll
depend
on
what's
running
on
those
infra
nodes.
So,
for
example,
if
you
have
infra
nodes
that
are
doing
Ingress,
then
you
know,
maybe
you
want
to
protect
those
in
a
similar
way
right
of
anti-affinity
rules
to
keep
them
on
separate
hosts.
So
that
way,
if
one
host
goes
down,
you
don't
suddenly
lose
all.
You
know
Ingress
capabilities
into
your
cluster,
so
having
I.
A
Think
two
is
the
minimum
recommended
number
of
Ingress
controllers
with
openshift
anyways,
but
you
can
scale
that
up
as
well.
If
you
so
choose,
other
info
services
are
going
to
depend
if
you're
running
odf,
you
know
odf,
you
would
probably
not
want
to
have
those
running
on
the
same
node.
For
the
same
reasons,
you
know
losing
too
many
odf
nodes
means
that
your
storage
would
go
down,
which
means
that
your
apps
that
are
using
that
persistent
storage
would
go
down.
A
A
A
A
So
as
a
cluster
administrator
I
have
access
to
look
at
this
secret
right,
I
can
get
the
username
and
password
these
are
just
base64
encoded
and
as
I
scroll
down.
I
can
also
do
things
like
look
at
the
vsphere.conf
file
and
all
this,
so
you
could
again
as
that
VMware
or
as
that
cluster
openshift
cluster
administrator.
A
You
know
you
can
extract
the
credentials
and
then
down
here
it'll
have
you
know
here's
my
vcenter
server,
so
you
could
Point
your
browser
at
that.
You
could
use
those
credentials
and
you
would
have
whatever
permissions
have
been
granted
to
the
openshift
cluster
at
that
point.
So
take
that
into
account.
You
know,
definitely
work
with
your
VMware
administrators.
So
the
way
they
understand
any
impact
in
any
threat,
that's
associated
with
that
some
customers-
just
anecdotally.
You
know
from
from
my
experience
talking
with
a
lot
of
different
folks.
A
A
A
So
in
the
what's
new
in
unshift
4.12,
we
announced
the
is
that
it?
No
the
ability
to
do
Zone
based
provisioning,
eventually
I'll
find
this
as
I'm
trying
to
click
and
talk
at
the
same
time,
which
is
never
a
good
idea
for
me
here
we
go
Zone
awareness,
so
this
is
a
tech
preview
feature
where
you
can
go
in
and
effectively
can
I
make
this
bigger.
A
So
this
enables
Behavior
that's
very
similar
to
how
AWS
works
so
essentially,
I
go
into
vcenter
and
I
create
labels
for
different
hosts,
different
groups
of
hosts
inside
of
my
inside
of
my
environment.
So
you
see
the
example
they
use
here
is
I
have
a
data
center
that
has
been
assigned
to
region,
so
Us
East,
one
mimicking,
AWS
and
then
I
have
different
clusters
that
have
been
assigned
different
zones,
so
U.S
east
one
a
b
and
c,
and
then
over
here
I
have
a
separate
data
center.
So
keep
in
mind.
A
A
Maybe
one
for
you
know
SFO
or
something
like
that.
So
these
data
centers
would
have
labels
and
then
each
one
of
the
Clusters
would
have
labels
if
I
had
more
than
one
and
then
basically
what
happens
is,
and
it
doesn't
show
it
in
here.
What
happens
is
when
I
go
and
create
my
cluster
when
I
go
into
deploy.
My
openshift
cluster
I
specify
different
nodes
to
be
created
in
each
one
of
these
zones,
so
for
comparison's
sake.
What
is
this
here
we
go.
So
here
is
a
AWS
cluster.
A
A
Apparently,
I
got
logged
out,
that's
fun
here
we
go
so
machine
sets.
So
you
see
here,
I
have
a
machine
set
for
us,
East,
2A
2b2c,
so
that
2A
I,
look
at
the
yaml
here
represents
the
availability
Zone
in
the
region
that
it's
using.
So
we
effectively
implement
the
same
type
of
mechanism
in
vsphere.
With
that
just
keep
in
mind
it's
Tech
preview,
which
means
that
it's
not
supported,
but
if
you're
interested
you
can
go
in
and
you
can
check
that
out
today
with
openshift
4.12..
A
So,
as
you
saw
just
a
moment
ago,
when
I
was
showing
my
vcenter
I,
don't
have
enough
resources
to
actually
test
that
or
demo
that
rather
so,
maybe
we
can
figure
out
something
and
I
can
show
that
at
a
future
date,
I'm
going
to
turn
off
my
name
down
in
the
corner
there
just
to
keep
it
from
interfering,
okay,
so
yeah,
please!
Let
me
know
if
there's
any
questions
about
that
I
know,
VMware
being
by
far
the
most
popular
data
center
hypervisor.
You
know
most
folks
have
some
cluster
that's
running
on
vsphere.
A
A
So
Dwayne
VLAN,
not
really
because
you
know
vlans-
are
meant
to
be
Network
level
segmentation,
so
if
you
think
of
it
as
a
host
and
I'm
kind
of
making
this
up
as
I
go
here.
So
please
forgive
me
so
if
you
think
of
it
in
terms
of
a
VLAN,
vlans
provides
segmentation,
but
not
necessarily
resiliency.
The
goal
here
or
the
goal
with
the
VMware
zones
is
to
provide
resiliency
so
that
if
my
DRS
cluster,
for
whatever
reason
fails,
you
know
maybe
I
have
one
cluster
in
each
rack.
A
Row
in
my
data
center,
maybe
I
have
multiple
server
closets
and
each
VMware
DRS
cluster
represents
one
closet
at
that
particular
site
effectively.
I
can
the
goal
is
to
spread
my
openshift
nodes
across
all
of
those
to
provide
resiliency
against
a
larger
scale,
failure
so
larger
scale
here
being
larger
than
a
single
or
a
couple
of
nodes
right
all
the
way
up
to
a
DRS
cluster
level,
but
not
a
full
site-wide
failure,
because
it
would
be
I
mean
I,
guess
it
depends
on
how
you
do
architect.
A
A
A
So
guidance
for
deployment
spanning
multiple
sites,
so
I'll
share
this
in
chat
and
effectively
this
walks
through.
Like
yes,
it's
supported
to
do
this,
but
you
have
to
meet
certain.
You
know
certain
things
in
order
to
be
successful,
so,
for
example,
Network
requirements
store
garments,
workload,
placement
considerations,
so
this
is
really
great
KCs.
If
you're
interested
in
you
know
spanning
a
cluster,
usually
the
biggest
thing
is
quite
simply
latency.
A
You
know
you
want
to
order
that
SCD,
the
control
plane
nodes
have
low
enough,
both
Network
average
latency
to
be
able
to
meet
the
requirements
of
of
your
cluster.
So
recurrence
of
your
cluster
here
being
like
how
many
API
calls
are
happening,
how
many
object
reads
and
writes
and
updates
are
happening
in
at
CD
all
of
that
type
of
stuff.
Remember:
adding
latency
to
etcd
limits
scalability
for
your
cluster
because
it
has
to
be
able
to
commit
those
objects
has
to
be
able
to
write
those
objects.
A
So,
if
you're
doing
a
bunch
of
operations-
and
it
can't
commit
them
fast
enough-
you
basically
have
to
reduce
the
number
of
operations
mirror.
No,
we
haven't
gotten
to
the
agent
based
installer
yet
because
I've
been
rambling
about
vsphere
and
yeah
or
DRS,
and
Affinity
rules,
agent-based
and
seller
compatible
with
UPI
or
only
IPI.
Actually,
I
have
not
gotten
it
to
work
with
IPI
I've,
only
gotten
it
to
work
with
UPI
I've
done
both
a
platform.
A
None
install
and
I'll
show
that
here
in
just
a
moment,
and
then
I've
also
done,
although
I
will
admit,
not
fully
successfully
a
platform
vsphere
deployment,
but
I've
only
created
the
VMS
with
that
I
have
not
done
an
IPI
with
that
foreign
considerations.
I'm,
not
sure
what
you
mean
there,
Tim
role
considerations
across
zones
in
the
cloud
I,
don't
know
if
you're
referring
to
like
permissions,
you
know
IAM
roles
or
something
like
that.
A
A
So
this
is
particularly
true
with
say,
disconnected
installs.
You
know
where
it's
hard
to
get
access
to
a
Bastion
host
or
something
like
that
inside
of
that
environment,
where
I
went
to
from
a
connected
node
to
be
able
to
go
in,
you
know,
create
this
image
and
then
as
well
as
the
registry
that
it
needs
to
deploy
and
then
move
those
over
to
that
environment.
Basically
attach
my
image,
my
my
one
ISO
to
my
host
boot
it
up
and
now
I
have
a
cluster.
A
That's
up
and
running
that
I
can
then
use
as
that
foothold
to
deploy
further.
You
know,
add
ACM
or
something
like
that
and
deploy
further
clusters
in
that
environment.
I
think
it's
also
useful
in
a
couple
of
other
scenarios.
So,
for
example,
maybe
I
am
you
know:
I
have
a
bunch
of
smaller
sites,
so
I
I'm
sitting
at
the
regional
or
the
central
site
for
my
company
and
one
of
the
smaller
sites
needs
an
openshift
cluster
instantiated.
Well,
I
can
create
the
image.
I
can
hand
it
off
to
somebody
who's
there.
A
A
So
this
blog
post,
which
was
created
by
the
product
management
and
product
marketing
teams,
blog
posts
through
a
number
of
different
aspects,
including
down
here
as
we
get
through
and
I
think
this
was
created
before
ga
yeah,
so
it
was
created
before
GA,
so
they
mentioned
some
quirks
that
are
no
longer
applicable,
but
then
is
that
we
walk
through.
We
create
our
traditional
install
config.
A
That
tells
us
about
our
cluster.
You
know
all
things
are
standard
for
a
bare
metal
cluster.
Here
you
can
see
created
this.
He
is
specifying
a
disconnected
mirror,
so,
just
like,
like
you,
would
with
any
other
disconnected
cluster,
you
use
OC
mirror
to
go
in,
create
the
install
registry
for
that,
and
then
we
also
specify
an
additional
trust
bundle,
so,
basically
the
certificate
that
it
will
use
for
that
particular.
A
A
The
other
aspect
that
we
see
here
is
configuring
the
network
associated
with
each
one
of
these
here
on
my
first
host
here,
I'm
telling
it
what
interfaces
I
want
to
use
and
then
I'm
telling
it
how
to
configure
those.
So
here
it's
a
very
simple.
You
know
one
network
adapter
that
I'm
going
to
assign
an
IP
address
on,
tell
it
where
the
DNS
server
is
and
tell
it
where
its
default
route
is.
But
you
can
also
do
things
like
create
a
bond.
You
can
add
VLAN
interfaces
all
that
other
stuff
and
telling
it
so.
A
So
one
thing
that's
important
here:
you'll
notice
in
this
example
that
Ben
has
created
there's
three
different
hosts
when
we
do
this
and
I'm
going
to
show
an
example
of
this.
In
just
a
moment
it
generates
a
single
ISO,
and
that
means
that
it
provides
the
same
network
config
to
each
of
the
hosts.
In
this
case
three
each
three
hosts
that
get
booted
and
that's
why
our
Mac
address
information
is
really
important
here,
because
essentially,
if
that
Mac
address,
if
that
host
doesn't
have
this
Mac
address,
then
it's
just
skip
over
this
config.
A
Here
we've
got
fdf3
and
b918.
So
when
this
ISO
is
booted
on,
you
know
this
VM,
it's
going
to
say:
well,
I,
don't
I,
don't
have
this
network
adapter,
so
I'm
not
going
to
apply
that
config
I!
Don't
have
this
network
adapter
I'm
not
going
to
apply
this
config
I
only
have
this
network
adapter,
so
this
is
the
config
I
will
apply.
So
that
is
something
Mac
addresses
are
required
when
we
go
through
and
do
this
and
for
Mira
who's
just
asking
about
IPI.
A
A
A
Yeah
I
should
have
read
your
very
next
message,
so
no
today
it
only
works
with
the
either
a
non-integrated
so
platform,
none
cluster,
a
platform
bare
metal
cluster
or
a
vsphere
integrated
cluster
I
believe
that
they
have
plans
to
add
others
in
the
future.
But
I
don't
know
what
those
are
off.
The
top
of
my
head.
A
Our
hope
not
in
Reading
feature
for
console
banners
only
available
during
upgrades.
Hopefully
we
see
another
customization
episode
soon.
Oh,
yes,
yeah!
That's
a
good
point.
Our
hope9
one
of
the
things
that
Johnny
and
I
have
talked
about
is
the
dynamic
plugins
for
the
openshift
console.
There's
some
really
cool
stuff.
There
I
think
we
talked
about
during
that
during
the
what's
new
stream,
and
it's
something
that
we
hope
to
show
off.
A
A
A
A
So
I
have
already
created
my
virtual
machine
inside
of
here,
so
you
can
see.
This
is
just
a
relate
virtual
machine,
just
like
you
would
always
use
with
VMware
and
openshift
and
importantly
over
here
under
networking,
not
under
that
networking.
Under
this
networking
it
tells
me
the
MAC
address
that
VMware
has
assigned
to
the
virtual
machine,
so
that
Mac
address
is
what
I
want
to
use
for.
In
my
install
conference,
I'll
config,
my
agent
configure
so
zero
six,
eighty,
which
I
think
I've
already
got
in
here
some
more
agent
config
in
my
agent
config.
A
You
can
see
I'm,
giving
it
the
Rendezvous
IP,
which
is
just
going
to
be
this
single
node
and
then
of
my
hostname
hostname
is
for
this
one
should
be
an
important
but
we'll
see
and
then
yep
zero.
Six
Alpha
3
is
the
MAC
address
that
we
just
saw
I
will
say
that
sometimes,
when
creating
a
VM,
it
won't
give
you
the
MAC,
address
right
away:
it'll,
wait
to
generate
it
and
assign
it
until
after
the
VM
has
been
powered
on
the
first
time.
A
A
Each
network
adapter
will
have
a
sticker
that
has
the
MAC
address
on
it.
You
can
also
something
like
access,
your
ilum
or
your
lights
out
management
interface.
Typically,
all
of
the
major
server
vendors
in
their
interface
there
they
will
have
the
Mac
addresses
and
other
information
about
the
hardware.
So
you
can
pull
it
off
that
way
and
then,
of
course,
you
can
do
something
like
boot
into
a
live
ISO.
A
So
you
know
boot
into
a
Rel
or
a
fedora
ISO
or
something
like
that,
and
then
just
do
a
an
IP,
a
or
an
IP
int
I.
Don't
remember
the
command
off
the
top
of
my
head.
Are
there?
Is
there
any
dependency
for
nfs3
an
open
shift
regarding
compatibility,
no
no
NFS,
so
it
depends
the
only
time
there
would
be
any
a
dependency
on.
Any
storage
is
if
you're
using
PVC
PVCs
inside
of
openshift,
but
the
cluster
itself
doesn't
need
NFS
to
deploy
or
anything
like
that
foreign.
A
So
I'm,
not
sure,
I
fully
understand
the
purpose
of
agent-based
installer
yeah,
so
agent-based
installer
is
targeted
at
a
well.
One
of
the
main
use
cases
is
disconnected
environments,
so
I
can
go
in
and
basically
I
generate
my
ISO
and
then
I
hand
that
over
and
I
boot,
the
node
in
a
disconnected
environment,
and
it
goes
and
it
installs
the
cluster.
So
you
do
need
just
like
with
any
openshift
install.
A
A
Just
by
setting
you
know
removing
Enzo
and
setting
DHCP
to
true
telling
it
where
DNS
is
at
and
I'm
telling
it
it's
default
route,
and
then,
aside
from
that,
so
if
I
do
a
head
on
dash
20
on
my
install
config,
my
install
config
is
just
a
standard,
install
config
I'm,
only
showing
the
first
20
or
so
lines
here,
so
that
I
don't
show
off
my
policy
good,
so
install
or
single
node
openshift
deployment
here.
So
the
workers
I
have
zero
replicas.
My
control
plane
I
have
one
replica
beyond
that.
A
My
cluster
name
telling
it
to
machine
networks
like
I,
always
do
cluster
Network,
you
know
so
on
and
so
forth.
Service
network
pull
secret,
SSH
key,
just
like
with
any
other
install
so
I.
It's
no
cluster
folder
is
empty
here.
I'm
going
to
copy
ammo
into
ISO
cluster
directory
got
my
two
files
in
there
just
like
with
any
other
openshift
config.
It's
going
to
consume
those
files.
A
A
A
It
sets
the
Rendezvous
IP
racting
the
ISO,
from
the
release
payload.
So
the
iso
is
in
this.
You
can
see
the
command
that
it
uses.
The
iso
is
found
within
that
container
image
that
it
pulls
down
blah
blah
blah
assuming
agent
config
consuming
install
config.
It
builds
out
everything
that
I
need
and
if
I
go
into
my
directory
there,
let's
clear
this,
so
it's
easier
to
read.
A
Oh
there
it
is
oh,
oh
I
didn't
set
the
directory.
That's
what
happened
so
when
I
ran
my
config
I
I
set
the
log
level
to
debut,
but
I
did
not
add
the
dash
error
equals
no
cluster
like
I
should
have
so
it
used
ones
that
were
in
my
directory,
anyways
attention
to
detail.
Folks,
it's
important
so
anyways
same
results
and
I
still
have
my
backup
or
my
two
files
here
that
I
can
use
to
regenerate
this
if
I
need
to
so
the
end
result
is
I.
A
So
what
you'll
notice
isn't
here
is
all
of
the
ignition
files
that
I
would
normally
use
for
something
like
a
UPI
deployment.
So
let's
take
this
agent
and
I
want
to
copy
that
over
to
my
over
to
my
esxi
host,
all
I'm
doing
is
an
SCP
taking
that
agent
ISO
and
then
I'm
putting
it
over
on
my
vsphere
host
here.
In
that
one
data
store.
A
A
A
A
So
coming
back
to
our
virtual
Sheen
here,
so
one
thing
I
am
going
to
do
because
I
use
this
virtual
machine
as
a
test
yesterday
I'm
going
to
remove
the
disk,
that's
here
and
add
a
new
one.
So
the
reason
I'm
doing
that
is
because
by
default
the
boot
order
will
it'll
boot
to
the
disk
first
and
the
CD
second,
since
it
already
has
an
install
on
there
on
that
disk
I.
A
A
You
can
see
it
already
booted
to
the
core.
Os
live
ISO,
so
it'll
go
through
and
it'll
continue
along
its
merry
way,
booting
into
core
OS
so
from
over
here.
If
this
is
you
know,
if
whatever
host
you're
using
for
this
is
you
can
connect
to
your
agent
based
install
node?
If
that
makes
sense?
Basically,
my
admin
host
here,
which
is
Anchor
I,
have
network
connectivity
over
to
where
this
is
being
deployed
at
which
I
can
do
things
like
use.
A
A
So
you
can
see
down
here
get
out
of
here
and
here
at
the
info
level
waiting
for
the
cluster
install
to
initialize.
So
basically,
it's
waiting
for
something
to
happen
over
here.
We
can
see
some
activity
happening
as
it's
going
through
various
aspects
of
deploying
the
cluster,
so
one
host
expected
zero
host,
not
validated
it's
going
through
and
doing
all
these
things.
It'll
do
things
like
if
you've
ever
done,
an
agent
or
a
assisted
installer
based
installed
before
it
does
a
bunch
of
validation.
A
A
A
If
I
look
at
my
persistent
volumes,
you
can
see
that
there
are
none
like
I
deployed
the
entire
cluster
and
it
has
no
persistent
volumes
that
I've
created
some
services
that
you
deploy
can
optionally.
Take
advantage
of.
Rwx
read:
write
many
storage,
for
example
NFS
storage,
but
you
don't
have
to
you,
don't
need
that
so,
for
example,
if
you
deploy
the
registry,
the
registry,
if
you
want
to
have
a
scaled
registry,
so
two
three
four
instances
of
the
registry
pod
running
with
the
same
storage
backend,
then
you
would
want
rwx
storage.
A
However,
I
would
also
recommend
object.
Storage
in
that
case,
if
that's
an
option
but
yeah
there's
no
dependency
for
NFS,
but
your
applications
can
absolutely
take
advantage
of
it.
If,
if
you
want
to
or
need
to
switch
back
over
here,
so
you
can
see
it's
continuing
to
give
me
information
about
what's
happening
inside
of
there.
Cluster
is
ready
for
install
ing
cluster
for
installation
host
the
first
one,
so
Master
zero
that
I
had
booted
is
preparing
for
installation
coming
over
here
you
can
see
I
get
similar
information
about
it,
cluster
installation
in
progress.
A
So
from
here
we
just
wait
so
yesterday
it
was
taking
I,
don't
know
about
15
minutes
for
the
cluster
to
deploy-
or
at
least
for
this
process
to
finish,
and
then
it
takes
another
little
while
for
it
to
completely
finish
deploying
the
cluster.
It
does
exactly
what
it
says
it
does:
it's
really
straightforward,
which
is
really
nice,
so
you
can
download
it
400
megabytes,
I.
A
A
A
So
this
is
not
an
IPI
install
right,
it's
not
going
to
boot
up
and
it's
not
going
to
connect
and
it's
not
going
to
provision
a
bunch
of
VMS,
rather
I'm
going
to
hand
it
the
VMS
that
it's
going
to
use,
and
it's
going
to
use
this
for
my
quote:
unquote:
load
balancer,
just
like
IPI,
would
and
then
I
provide
all
the
other
information.
That's
here
so
well,
I
hope
it's
not
having
dropouts.
Please
let
me
know
if
that
continues
Peter,
so
yeah.
That
would
be
please.
A
A
A
Okay,
hopefully,
that
will
improve
things
a
little
bit
if
it
continues.
Just.
Let
me
know
please
in
the
case
of
a
non-single,
node
cluster.
How
do
we
differentiate
the
provisioning
of
control,
plane,
slash
worker
nodes
if
we're
using
the
same
ISO
for
both,
so
that
would
come
in
the
agent
config
file
or
the
Edgington
yeah
agent
config
file.
A
A
A
A
So,
apologies
for
the
confusion
there,
all
right
so
moving
back
over
to
our
this
is
my
attempt
or
I'm
going
to
attempt
to
deploy
a
three
node
compact
vsphere
cluster
with
this,
so
here's
my
platform
stanza
with
the
VIPs
that
I'll
be
using
and
then
everything
else
is
standard
for
an
openshift
install
and
I've.
Also
have
my
agent
config
yaml.
A
You
can
see
here
I'm
doing
the
same
thing,
so
Rendezvous
IP
is
going
to
be,
in
my
instance,
I'm
just
using
the
first
node
as
my
bootstrap
and
then
I'm
configuring
everything
else
inside
of
here,
according
to
how
I
need
it
to
be,
but
that
this
root
device
hunt
isn't
really
necessary.
If
you
have
a
one
device
on
here,
I'm
just
leaving
it
in
there
for
completeness
sake
and
see.
What's
in
here,
okay,
so
I
do
have
an
empty
directory.
A
Equals
blue,
so
what
you're
going
to
see
here
in
just
a
moment
is
it's
actually
going
to
say
that
it's
ignoring
a
bunch
of
the
config,
including
all
of
my
VMware
configuration
inside
of
here
irritatingly?
It
prints
those
values.
So
what
we
should
see
come
on,
you
can
get
there.
We'll
basically
see
is
to
say
you
know,
ignoring
vcenter
value,
ignoring
your
name
value,
ignoring
password
value
and
all
that
so
the
password
that
I've
put
in
here
is
not
the
real
password,
so
just
FYI,
if
it
ever
actually
gets
there,
I
don't
know.
A
This
cluster
is
rather
busy,
so
sometimes
these
things
lag,
I,
don't
I,
don't
know
that
I
will
expect
my
cluster
deployment
here
to
actually
succeed.
Just
due
to
how
busy
the
underlying
VMware
environment
is,
it
gets
laggy.
Sometimes.
So,
if
I
look
at
my
monitor
and
go
to
Advanced
performance
yeah,
you
can
see
it's
there.
We
go.
A
A
A
A
Anyways
that'll
eventually
get
there,
and
what
we'll
get
is
the
exact
same
thing
we'll
get
an
ISO
and
my
goal
if
I
don't
run
out
of
time
here,
my
goal
is
to
do
effectively
the
same
thing
so
we'll
come
over
here
to
our
nodes,
I'm
going
to
go
back
to
our
mediums
here,
because
it's
easier
to
read
and
for
each
one
of
these
nodes
I
will
attach
that
ISO.
So
I
will
edit
our
VM.
A
It'll
spit
out
an
ISO
and
then
I
will
copy
it
over
to
the
data
store.
Just
like
I
did
with
my
other
environment,
attach
it
to
our
nodes
and
then
let
them
boot
can
we
use
Pixi
to
deploy
automatically
yeah
I
would
imagine
so,
but
I
haven't
actually
tried
that
so
in
this
instance,
yeah
there's
clicking
way
more
clicking
than
you
would
probably
need.
A
As
a
result
of
me,
you
know
doing
things
like
adjusting
disks
Etc,
but
you
could
absolutely
automate
all
of
this,
so
maybe
use
ansible
to
create
a
you
know,
VMware
or
you
know,
Rel
or
whatever
virtual
machine.
If
you're
installing
on
bare
metal
on
a
physical
server.
Many
of
those
have
the
ability
to
automate
against
their
ipmi.
You
know
ilomb,
whatever
interface,
you
could
do
it
that
way.
I
can't
I,
don't
see
any
reason
why
you
couldn't
pixie
boot
in
order
to
to
accomplish
that.
A
So
same
sort
of
thing,
click
back
over
here
yep,
so
it's
going
through
and
doing
its
thing,
I
assume
that'll,
eventually
reconnect
this
guy
over
here
is
doing
his
thing.
So
I
don't
know.
What's
going
on
there,
these
poor
hosts.
Are
you
memory
and
CPU
super
high
storage
is
very
highly
utilized,
so
I
don't
know.
What's
going
on.
A
A
A
A
You
can
see
that
Ben
when
he
wrote
this,
he
used
KVM
running
on
a
Rel
host.
You
don't
need
to
do
this
anymore
when
you
get
down
to
that
part
that
was
fixed
before
it
was
G8
just
like
I
showed
before.
If
one
of
these
would
ever
complete
here,
we'll
get
out
of
that.
So
just
like
I
showed
before
in
the
auth.
So
here
we
have
our
Cube
admin
password
as
well
as
the
coupe
config
that
I
can
use
to
connect
to
that.
A
Responding
nope,
oh
because
I'm
connected
to
I'm
connected
to
the
corporate
VPN,
so
the
the
corporate
DNS
Server
doesn't
resolve
locally
from
from
my
network,
so
that
won't
work
until
I
disconnect
from
the
VPN,
which
is
probably
why
this
is
not
responding
actually
now
that
I
think
about
it,
because
it's
trying
to
look
it
up
by
is
trying
to
do
connect
to
this,
and,
as
you
can
see,
it
fails
on
that.
One
thing
that
I
can
do
is
add
that
into
my
Etsy
hosts
I
think.
A
So
it's
being
slow
to
respawn,
but
hopefully
maybe
it'll
work.
This
guy's
still
doing
his
thing.
All
right,
I'm,
just
gonna,
give
up
channel
in
my
inner
Johnny
here,
Demo's
going
wrong.
All
kinds
of
things
happen
in
the
way.
They're
not
supposed
to
it
goes.
Connection
was
refused,
did
I,
specify
the
right
host
or
report,
maybe
not.
A
A
It'll
give
me
the
same
output
yeah.
You
can
see
here,
42
complete,
so
our
single
nodes,
non-integrated
install,
is
going
according
plan.
It
takes
it
a
few
minutes,
just
like
always
does,
and
eventually
this
won't
succeed
and
then
I'll
have
a
fully
running
single
node
openshift
cluster.
You
can
see.
The
point
here
is
that
I,
provided
it
all
of
the
information
that
it
needed,
including
the
static
IP
configuration
data.
A
It
generated
an
ISO
I,
simply
boot
that
to
an
ISO
I
could
also
see
where
this
could
be
interesting
for
some
scenarios,
for,
if
maybe,
if
your
vsphere
admins
don't
want
to
hand
over
the
permissions
or
something
like
that
to
do
an
IPI,
install
I
know
when
I
was
a
a
VMware
admin.
You
know
going
back
a
few
years
I.
It
always
took
a
lot
of
convincing
for
our
team.
A
When
somebody,
you
know
an
application
team
came
to
us
and
said:
hey,
you
know,
can
you
know
I
need
you
to
create
these
VMS,
but
don't
turn
them
on.
Don't
boot
them
until
I
tell
you
to
because
I
have
to
Stage
all
of
these
other
things
in
order
to
have
it
correctly
installed,
which
is
what
an
openshift
install
has
to
do
right.
You
go
to
the
VMware
team
and
you
say:
hey
deploy
this
OVA
creates
you
know
six
or
eight
or
ten
VMS
three
control
plane
to
and
fro.
A
However
many
workers
so
create
them
from
the
OVA,
but
don't
turn
them
on
yet,
and
then
you
know,
okay,
openshift
install,
you
know,
create
ignition
config,
put
everything
up
where
it's
supposed
to
be
tell
them
to
set
the
VM
config
options.
You
don't
have
to
do
any
of
that
here
right.
So
here
it's
just
here's
an
ISO,
please
boot
the
VM
to
this
ISO
and
it
goes
through
and
it
sets
everything
for
us.
We
can
get
that
first
cluster
up
and
running
from
there.
A
We
can,
you
know,
work
out
how
to
get
the
rest
of
the
stuff
set
up
and
configured
with
the
VMware
install
what
I
hope
to
show
over
here.
That
is
just
not
not
happening.
I,
don't
know
here.
It's
finally
to
the
point
of
consuming
those
I
was
hoping
to
show
over
here
is
that
the
end
result
would
be
an
integrated
cluster.
No,
it
wouldn't
work
for
things
like
the
CSI,
provisioner
or
machine
set
integration.
Because
again,
these
credentials
are
not
correct,
but
it
would
in
theory
it
would
still
work
right.
A
I
would
still
have
you
don't
need
these
to
work
for
a
a
vsphere
integrated
cluster
to
deploy,
UPI
IPI
does
need.
It
does
need
those
to
deploy.
So
that's
not
going
to
work
here.
I
don't
know
what's
going
on,
but
my
other
cluster
is
continuing
along
it's
doing
its
thing.
Slowly,
but
surely
so
I
will
check
in
with
the
product
management
team,
as
well
as
the
tech
marketing
team
for
the
agent
based
installer
and
see
if
they
have
any
video
demos
that
show
this.
A
If
you
want
to
see
the
whole
thing
all
the
way
through
again,
this
is
not
meant
to
be
exciting.
It's
meant
to
be
pretty
boring
of
I
create
two
files:
I
generate
an
ISO
I
boot,
my
nodes
to
that
ISO
and
then
wait
a
few
minutes
and
I
have
a
cluster
up
and
running
so
live
without
a
net.
That
is,
that
is
every
day
of
my
life
Peter
outside
scope,
question
ovn
versus
sdn,
which
one
is
preferred
and
why
so
we
have
not
had
an
episode
about
that
in
a
long
time.
A
Mira
and
my
apologies-
if
I
am
butchering
names,
I'm
terrible
with
names,
as
my
children
can
attest
to
so
a
long
time
ago,
I
want
to
say
four
seven
or
four
eight.
When
ovn
kubernetes
was
first
introduced
into
openshift,
we
we
chatted
with
Mark
Curry
about
it,
so
the
change
with
4.12
is
now
OV
and
kubernetes
is
the
default
sdn
for
all
deployments,
since
4
10,
4,
9
or
410.
It's
been
the
default
to
sdn
in
some
scenarios
already
so,
for
example,
sale,
node
openshift.
A
So
the
major
change
is
that
effectively
all
of
the
testing
validation
Etc
was
done
and
improvements
as
needed,
so
that
OV
and
kubernetes
has
both
feature
and
performance
parity
with
openshift
sdn
I
say
that
openshifts
sdn,
no
ovn
kubernetes
also
has
many
more
features
than
openshift
sdn.
So
if
I
come
here
to
the
docs.
A
Here
we
go
so
I'll
post
this
into
chat
here,
real
quick,
so
you
can
see
here
a
list
of
different
features
that
are
available
so,
for
example,
if
you
want
to
do
IPv6,
you
need
to
use
ovn
kubernetes.
If
you
want
to
do
ipsec
encryption
at
the
sdn
layer
between
nodes,
you
need
to
use
ovn
kubernetes,
my
stance
up
until
even
now
my
stance
has
always
been
go
with
the
default
unless
you
have
reason
not
to
use
the
default.
So
for
a
long
time,
openshift
sdn
was
the
default.
A
There
was
no
reason
to
change
it
unless
you
specifically
needed
one
of
these
features
and
I
would
say
that
the
same
continues
to
be
true.
Now
continue
with
the
default
OV
and
kubernetes.
Unless
you
have
a
specific
reason
not
to
use
it,
I
will
say
that
openshift
sdn
there
is
a
process,
and
it's
documented
in
here
somewhere.
A
Basically,
you
can
convert
from
openshift
sdn
to
OV
and
kubernetes.
It's
a
it,
but
it
is
a
one-way
conversion.
So,
if
you
deploy
OV
and
kubernetes,
you
can't
go
back
to
openshift
sdn,
but
yeah.
There's,
there's
tons
of
additional
features.
All
the
network
observability
stuff,
that's
coming
out
is
all
going
to
be
OV
and
kubernetes.
A
There's
a
bunch
of
other
features
around
various
things.
Just
a
ton
of
stuff,
that's
happening
where
OV
and
kubernetes
is
going
to
be
the
future.
That
being
said
so
far
as
I
know,
openshift
sdn
has
not
been
deprecated
yet
I,
don't
know
if
it
does
get
deprecated,
it
will
it'll
be
a
long
time
before
it
ever
actually
gets
removed
or
support
is
removed.
A
Having
chatted
with
that
team,
one
of
the
major
drivers
for
the
move
from
openshift
sdn
to
OV
and
kubernetes
is
really
Community.
Support
openshift
sdn,
as
the
name
implies,
is
used
by
well
openshift
and
there
are
no
other
kubernetes
distributions
that
are
using
it.
There's
no
other.
You
know
third
parties
that
are
using
it.
You
know
nothing
like
that,
so
the
community,
for
it
is
one
it's
a
community
of
red
hat,
OV
and
kubernetes
is
used
by
a
bunch
of
different
kubernetes
is
kubernetes
kubernetes
distributions.
A
A
A
So
if
you
have
something
that
is
sensitive
to
that,
you
know
having
lots
of
or
potentially
having
lots
of
packets
just
take
that
into
account.
What
I
was
looking
for
here
is
the
deprecated
features
here
so
deprecated
and
remove
features
and
there's
nothing
in
here
operator.
Images,
monitoring,
installation.
A
Sorry
for
all
the
scrolling,
Network
yeah,
the
only
thing
that's
deprecated
with
4.12
is
Courier,
so
openshift
sdn
continues
to
be
fully
supported
fully
generally
available
again,
if
they
ever
deprecate
it
and
I
I
do
not
know.
I
have
not
heard
if
they
have
any
plans
to
deprecate
it.
I
expect
it
to
be
one
that'll,
be
deprecated
for
a
a
large
number
of
releases,
so
that
way,
folks
have
plenty
of
time
in
order
to
figure
out
how
to
migrate
off
and
whether
or
not
it'll
affect
anything
or
anything
like
that.
A
A
Is
there
a
switch
possible
on
ovn
to
Calico,
so
you
can
use
Calico
but
I
as
far
as
I
know,
you
cannot
convert
that
being
said,
that
conversion
would
be
managed
by
the
tigera
automation
tiger
operator
or
whatever,
so
you
would
have
to
check
with
their
documentation
and
see
what
they
have
possible.
My
experience
I've
deployed
Calico
a
handful
of
times.
It's
always
been
at
cluster
install
time,
so
I
I
can't
really
answer
that
it
does
not
use
Calico
by
default.
So
by
by
default.
A
A
A
We
can
see
here
our
cluster,
our
single
node
openshift
cluster
is
progressing,
albeit
slowly
because
of
my
ancient
Hardware,
but
it's
at
84
plus
percent,
complete,
stop
watching
that
and
do
an
OC
get
Co
and
you
can
see
things
are
things
are
Marching
along,
however
slowly
they
happen
to
be
going
but
yeah.
So
that's
agent
based
installer,
again
straightforward,
well,
I,
hope,
it's
straightforward!
I
hope
you
all
understood
the
process.
If
not,
you
can
always
reach
out
to
me.
A
You
can
reach
out
directly
andrew.sullivan
redhat.com
I'm
happy
to
answer
any
of
those
questions
as
well.
As
put
you
in
contact.
If
you
have
any
questions,
you
know
with
the
product
team
for
agent-based
installer,
but
the
big
thing
to
me
is,
you
know,
use
case
wise
I
find
this
one
to
be
pretty
convenient,
for
you
know
instantiating
clusters.
This
may
be
how
I
now
deploy
all
single
node
openshift
clusters
from
now
on,
because
I,
don't
right,
I
run
one
command,
it
spits
out
an
ISO
and
now
I'm
good
to
go.
A
So
with
my
you
know,
my
agent
config
I'm
setting
a
static
IP
for
my
single
node
openshift.
So
this
may
become
my
preferred
method
because
it's
faster,
in
my
opinion,
even
than
using
the
assisted
installer,
you
know,
assistant
installer
I
still
have
to
go
through
and
click
a
bunch
of
stuff
in
the
GUI.
This
way,
I
have
one
yaml
file
that
I
keep
and
as
I
need
to
reinstall
things
or
redeploy.
I
can
keep
doing
that
so
Peter
install
Red,
Hat
openshift
on-prem
with
the
agent
based
installer.
Oh
okay,
yeah.
A
A
Yeah,
that's
exactly
what
that
is
released
seven
days
ago,
nice,
so
hopefully
that
URL
yeah,
so
that
YouTube
video
that
P
that
Peter
PC
louderbox
there
chatted
is
a
video
that
shows
exactly
what
we're
looking
at
here
right.
It's
the
demo
of
that
process
end
to
end
all
right.
So
thank
you.
Everyone
really
appreciate
you
joining
me
today.
I
hope
this
has
been
informational,
I,
I,
hope
it's
been
useful
again.
You
can
always
reach
out
to
me
at
andrew.sullivan
I'm,
going
to
turn
this
guy
off
now.
A
So
you
can
see
me
again,
so
you
can
always
reach
out
to
me
at
andrew.sullivan
redhat.com.
You
can
also
reach
out
to
me
on
various
social
media
at
practicalandrew.
There
is
also
a
slack
channel
on
kubernetes
slack
that
I'm
going
to
copy
link
I'm
going
to
publish
through
here
as
well,
so
you
can
always
join
the
kubernetes
slack
and
be
able
to
interact
with
us
there.
A
Both
Johnny
and
I,
are
on
that
slack
channel,
in
addition
to
a
whole
bunch
of
other
red
Hatters
and
Community
folks
and
all
kinds
of
other
things
that
are
going
on
there.
So
I
find
it's
a
great
way
to
interact
with
us
again.
I
will
be
better
I'm
I'm
tweaking
as
soon
as
we
end
our
stream
here,
I'm
tweaking
my
Twitter
settings.
A
So
that
way,
I
actually
get
notifications
for
things,
I'm,
not
a
Twitter
person,
so
I
don't
monitor
it
constantly,
but
apparently
I
was
not
getting
like
literally
any
notifications
for
the
last
couple
of
months,
so
I'm
gonna
fix
that.
So
at
practical
Andrew
is
a
good
way
to
get
me
there,
as
well
as
on
Reddit
I,
I,
frequently
haunt
both
the
openshift
and
red
hat
subreddits.
So
you
can
always
ask
questions
there.
If
you
have
anything
that
being
said
again,
I
won't
be
here
next
week,
but
Johnny
should
be
here.