►
From YouTube: OpenStack Austin Meetup Oct 10 2014
Description
OpenStack Austin Meetup Oct 10 2014
Topic: RedHat RDO
Audio starts at 1:52
A
The
new
one
which
just
came
out
and
I
haven't
taken
it
yet
we
actually
have
a
red
hat
exam
a
course
and
an
exam
for
red
hat
openstack.
So
if
you
want
to
be
able
to
demonstrate
to
potential
employers
that
you
have
subject
matter,
expertise
with
openstack
certainly
experiences
the
best
way.
No
one
is
going
to
tell
you
hey,
take
a
class
and
you're
going
to
go,
be
an
expert
like
back
in
the
microsoft
days,
I'm
at
mcsc.
You
should
pay
me
70
000
a
year.
What
were
you
doing?
A
Last
week
I
was
a
bicycle
repairman
and
yes,
I
taught
that
class
and
the
dude
that
was
in
there.
I
was
like
this
is
my
last
class.
Thank
you.
So
definitely,
if
you,
if
you
want
to
take
some
training,
we
can
help
you
with
that.
So
we're
going
to
talk
a
little
bit
about,
and
this
is
super
super
quick,
I'm
going
to
blast
through
it,
it's
a
little
bit
of
marketing
fluff
and
then
I'm
going
to
show
you
red,
hat's
implementation
of
openstack.
A
A
John
terpstra,
the
dude,
who
literally
wrote
the
book
on
samba.
Ladies
and
gentlemen,
no
pressure
at
all
having
him
in
the
front
row,
so
you
guys
are
probably
already
familiar
with
the
components
of
openstack.
You
know:
you've
got
everything
from
compute
the
compute
networking
nodes,
you've
got
image,
service,
object,
store,
networking
volume,
service,
identity,
management
with
keystone
and
then
the
dashboard
with
horizon.
A
These
are
the
components
which
we
have
currently
packaged
and
are
part
of
the
red
hat
distribution
of
openstack
I'll,
show
you
what
that
is,
and
I
I
hear
this
crap
all
the
time.
It's
a
cloud
operating
system.
I
don't
know
that.
I
necessarily
agree
with
that,
but
if
it
makes
your
boss
happy
that
you
go
with
it
because
you
know,
classes
are
all
driven
by
trade
racks,
so
I
mean
it's
kind
of
like
an
os
and
that
it
needs
x86
hardware
and
it
needs
operating
environments,
hypervisor
services
and
so
on
uses
existing
code.
A
Libraries
for
functionality,
openstack
runs
on
linux,
fantastic
synergy
between
linux
and
openstack.
They
work
really
well
together.
It's
a
fast
operating
system.
It's
a
cool
cloud
environment,
red
hat's
version
of
openstack
is
optimized
for
and
integrated
with
red
hat
enterprise.
Linux
now
red
hat
has
actually
been
involved
in
openstack
since
2011.
A
we
did
not
publicize
it
because,
frankly,
we
were
involved
in
a
bunch
of
other
cloud
projects
as
well.
Red
hat
fosters
and
contributes
to
upstream
projects
all
the
time
we
try.
You
know
we
don't
really
make
big
announcements
about
it,
but
we
do
it
all
the
time
this
is
kind
of
an
eye
chart,
so
I'm
not
going
to
go
into
all
of
them,
except
to
say
that
you
know
our
very
first
developer
started
doing
this
with
diablo
when
essex
released.
We
were
the
number
three
contributor
to
essex,
we
weren't
officially
part
of
the
openstack.
A
Actually
there
wasn't
even
a
foundation
at
that
point
when
the
openstack
foundation
kicked
off
at
the
beginning
of
this.
This
year
we
were
the
number
two
contributor
to
folsom
and
when
grizzly
released,
we
were
actually
the
number
one
contributor
to
grizzly
a
lot
of
folks
don't
realize
that
we
have
been
very
heavily
involved
in
openstack
for
a
very
long
time.
We
see
a
ton
of
value
in
it.
It
is
absolutely,
in
our
opinion,
the
the
number
one
infrastructure
as
a
service
offering
out
there
in
the
open
source
community.
A
A
As
soon
as
openstack
became
a
foundation
as
opposed
to
something
that
was
being
run
primarily
by
a
commercial
entity,
we
were
able
to
really
start
participating
and
assisting
with
governance
and
release
models,
and
things
like
that.
So
the
question
is:
how
do
we
get
from
the
upstream
open
source,
openstack
distributions
to
red
hat's
version
of
openstack,
and
this
is
a
a
fairly
important
slide.
So
I'm
going
to
spend
a
little
bit
of
time
on
this.
A
If
you
know
anything
about
red
hat,
you're,
probably
familiar
with
red
hat
enterprise,
linux
right,
that's
kind
of
our
flagship,
it's
the
operating
system,
we've
been
doing
linux
since
93..
Everybody
knows
about
red
hat
and
linux
well,
upstream,
from
red
hat
enterprise.
Linux
is
the
fedora
linux
distribution
and
the
community
that's
built
around
it.
So
we
take
all
these
thousands
and
thousands
and
thousands
of
projects
that
are
out
there
in
the
community.
A
A
A
So,
as
I
said,
we
we
love
the
open
source
community.
We
contribute
a
heck
of
a
lot
of
code
to
it
freely
and
gladly.
Then
what
we
do
is
we
package
that
in
rpm
format
we
put
it
out
at
redhat.com
rdo.
I
think
it's
no
red
hat.com
openstack,
it's
packaged
in
the
community.
It's
facilitated
by
red
hat,
it's
really
aimed
at
architects
and
developers.
So
you
guys
are
probably
great
candidates
to
use
rdo
freely
available,
not
for
sale.
There's
no
support
around
it,
except
for
community
support.
A
A
Eventually,
what
happens?
Is
we
get
red
hat
red
hat
enterprise,
linux
openstack
platform?
So
this
is
hardened
certified.
It
has
a
longer
life
cycle
than
six
months.
We
do
a
six
month,
release
cadence,
but
we
offset
it
by
a
couple
of
months.
So
when
grizzly
comes
out
or
when
havana
comes
out,
I
should
say
we
won't
have
red
hat
openstack
havana
for
probably
two
months
after
the
upstream
community.
A
One
comes
out
because
it
takes
a
couple
of
months
for
everything
to
stabilize,
to
be
hard
and
to
be
certified
and
basically
to
make
sure
that
we
don't
sell
a
solution
that
bus
that
has
because
that
happens,
and
people
get
really
really
cranky.
So
we
don't
want
to
do
it
and
that
will
be
supported
by
red
hat
the
upstream.
You
guys
probably
know
this
source
code
only
released
every
six
months.
A
Two
to
three
snapshots,
including
bug
fixes
rdo,
does
follow
the
upstream
cadence,
but
we
just
package
it
for
red
hat
and
derivatives
and
then
the
openstack
platform,
like
I
said
initially
it's
going
to
be
a
one
year.
Life
cycle
we'll
eventually
put
that
out
to
a
two
two
year:
life
cycle
right
now,
because
it's
moving
so
quickly
because
upstream
is
moving
so
quickly.
It
just
doesn't
make
sense
for
us
to
try
to
support
it
longer
than
one
year
because
nobody's
going
to
keep
it
production
longer.
A
A
So
if
there
is
a
data,
corrupter
or
a
security
exploit
or
anything
like
that,
we're
not
going
to
say
yeah,
you
know
you
gotta
wait
for
a
couple
of
months
when
I
say
it's
all
set
by
a
couple
of
months.
What
I'm
saying
is
when
havana
comes
out
upstream
right,
right,
it'll,
probably
be
about
two
months
before
red
hat's
version
of
havana
comes
out,
but
bug
fixes
and
things
like
that
should
come
out
very
cool
and
for
the
for
the
year
cycle
it
will
have
those
as
well.
A
A
A
I
think
the
initial
run
is
like
hundreds
of
servers,
but
we're
going
to
go
to
like
tens
of
thousands
of
servers.
So
what's
a
typical
10,
000
node
rollout,
look
like
I
haven't
done
one
yet,
let's
find
out
so,
but
I
mean
basically,
the
components
are
designed
to
be
modular
and
you
can
run
different
components
on
different
machines,
so
keystone
and
horizon
are
typically
going
to
run
on
the
same
machine
because
the
keystone
is
your
identity
service
and
it
kind
of
manages
everything.
A
That's
going
on
within
your
openstack
environment
and
horizon
is
the
web
ui
that
allows
you
to
manage
all
of
that,
but
your
compute
nodes,
you
can
have
one
or
many.
So
that's
you
know,
that's
a
rack
of
nodes
and
then
your
image
service,
object,
store
and
volume
service
can
run.
On
the
same
boxes,
or
they
can't
run
separately,
so
it's
really
just.
However,
you
want
to.
However,
you
want
to
deploy
it
now.
A
How
many
folks
have
not
yet
actually
installed
openstack?
Is
anyone
still
fairly
new
still
poking
around
at
okay?
So
this
is
fairly
straightforward.
I'm
going
to
blast
through
it
pretty
quickly
the
big
thing
to
remember
about
cloud
computing
in
general,
whether
it's
openstack
or
any
other
cloud
computing
solution.
Is
you
have
traditional
I.t
environments
where
you
have
these
systems
of
records
that
are
state
full,
typically
authoritative,
permanent
they're,
important
you
back
them
up.
A
So
as
an
example,
traditional
environments,
you've
got
big
stateful
vms
you've
got
oracle
databases
or
application
instances
or
whatever,
where,
if
you
lose
that
you're
in
trouble,
these
life
cycles
are
going
to
be
in
years,
you're
going
to
scale
these
virtual
machines,
bigger
as
the
load
gets
bigger,
they're
not
designed
to
tolerate
failures
of
the
vm,
so
you're
going
to
work
really
hard
to
make
sure
that
vm
stays
up
and
is
healthy.
All
the
time
and
they're
going
to
be
typical,
typically
strong,
slas
around
the
performance
on
that
machine
cloud
is
not
like
that.
A
Cloud
is
typically
going
to
be
a
whole
bunch
of
small
stateless
vms.
One
application
may
actually
go
across
multiple
vms
or
there
may
be
multiple
components
of
an
application
to
a
vm.
Your
life
cycles
may
be
as
low
as
hours
as
as
high
as
months,
but
they're,
typically
impermanent
environments,
and
your
scale
is
scaled
out
to
not
scale
up
and
if
a
vm
dies
doesn't
matter,
the
application
in
a
cloud
computing
environment
should
be
designed
to
tolerate
such
a
failure.
In
fact,
when
I
talk
to
my
customers
about
well,
how
do
I
re-architect
my
applications?
A
I
say
write
your
application
as
if
20
of
your
nodes
are
going
to
fail
every
year
and
they're
like,
but
I'm
like.
Hopefully
that
won't
happen,
but
if
it
does,
you
want
to
be
prepared
for
it
make
it
stateless
make
a
you
know.
Put
it
put
your
put
your
instructions
on
a
message
bus
so
that
you've
got
acknowledgement
so
that
you
know
when
a
new
machine
spins
up.
I
can
pick
it
up
and
take
care
of
the
job.
A
A
All
of
my
big
1450
customers
that
I
deal
with
all
doing
smaller
proofs
of
concept,
except
for
this
big
telco,
are
doing
smaller
proofs
of
concept
where
they
may
have
three
nodes
or
50
nodes,
or
something
fairly
small,
that
they're
kind
of
poking
it
out
and
they're
going
to
put
their
their
cloud
stuff
out
there
and
start
putting
devops
into
place,
which
is
a
whole
nother
set
of
challenges,
which
is
an
awesome
thing
to
get
involved
in
and
does
everyone
read
mandatory
reading?
Have
you
heard
the
phoenix
project?
Everybody
everybody
all
right.
A
And
I'm
sure
you
guys
have
all
heard
the
analogy
of
pets
versus
cattle
yeah.
I
don't
need
to
go
over
this
okay,
good.
I
tell
this
story
when
I'm
up
north
as
a
southerner
and
I
talk
about
you
know
you
got
the
you
got.
I
always
use
the
herd
of
cattle
right,
so
you
got
a
herd
of
cattle,
you
got
20
20
ranch
hands
or
you
know
cowboys
who
are
riding.
A
You
know
20
000
head
of
cattle
across
and
if
cow
gets
sick,
you
put
a
bullet
in
it
and
keep
going,
and
I
do
that
up
north
I'm
from
texas.
That's
how
we
do
things
all
right.
So
if
your
workloads
are
cloud
enabled-
and
you
are
ready
to
build
a
cloud
on
red
hat
platform-
marketing
marketing
marketing-
now
I'm
gonna
again,
I'm
gonna
go
past
this.
A
If
and
this
this
actually
gets
into
another
technology
that,
if
you're
interested
in
cloud
management,
let
me
come
back
and
talk
to
you,
but
I'm
going
to
break
out
of
this
and
I'm
going
to
talk
now
really
about
the
specifics
of
red
hat's
openstack.
So
I
I
was
fortunate
to
go
to
a
technical
services
conference
in
bangkok
and
deliver
a
very,
very
brief
overview
on
essentially
for
internal
technical
resources
on
how
to
get
up
and
running
on
red
hat
openstack.
A
So
I'd
like
to
share
that
with
you
guys,
this
is
a
very
basic
installation.
This
is
a
very
basic
explanation.
A
lot
of
you
guys
are
probably
gonna
look
at
this
and
go
cool,
but
for
anyone
who's
new
to
openstack,
who
hasn't
done
it
before.
I'm
gonna
talk
about
a
couple
of
things
that
I've
that
I've
done
to
get
this
to
work
on
my
laptop,
so
I
can
play
with
it.
A
Bang
on
it,
get
familiar
with
it
start
to
learn
the
technologies
and
concepts
without
having
to
have
massive
data
center
grade
infrastructure
that
I
have
to
have
to
do
this
stuff
on.
So
this
is
designed
to
go
on
a
laptop
now
when
you
build
your
laptop,
it
can
be
super
simple,
like
you
know,
when
I
first
did
this,
I
just
did
a
real,
simple
boot
root
and
swap
nothing
fancy
at
all,
very
simple
installation
or
you
can
get
kind
of
complicated,
which
is
perfectly
fine.
You
could
actually
have
a
separate
volume
group
for
sender.
A
A
A
Then
I
created
a
physical
volume
and
then
I
created
a
volume
group,
but
I
didn't
make
logical
volumes.
So
that's
what
I
installed
on
this
laptop
down
underneath
now.
You
can
do
a
super,
simple
installation
of
red
hat
enterprise
linux.
When
I
did
my
kickstarter,
I
just
did
app
base,
so
it's
just
the
core
functionality
to
get
the
thing
up
and
running
no
graphical
user
interface.
Super
super
simple.
In
my
case,
I
registered
my
systems
red
hat
network.
A
If
you're
going
to
be
doing
this
off
of
rhn,
then
you
need
to
register
to
the
rel
parent
channel
red
hat
enterprise,
linux
server
and
then
also
the
openstack
3.0
child
channel.
That
is
grizzly.
Here's.
What
that
looks
like
and
I'm
sorry
it's
kind
of
an
eye
chart.
This
is
a
little
bit
small,
but
but
that's
what
it
looks
like
on
rhn
and
when
you
do
yum
repo
list
you'll
see
that
it's
got
the
red
hat
enterprise,
linux
and
the
red
hat
openstack
channels,
you
install
the
red
hat
installer.
A
The
installer
that
we
use
to
install
openstack
is
called
pacstack.
Anyone
worked
with
devstack
yeah
same
concept,
same
concept,
so
the
one
we've
got
is
called
openstack
packsack,
so
you
young,
install
it.
That's
what
that
looks
like
and
then
you
can
use
it
to
create
an
answer
file.
There's.
Actually,
a
number
of
ways
you
can
do
it,
you
can
pass
a
really
really
long
nasty
ugly,
horrific
command
line
with
you
know
everything
on
the
command
line,
or
you
can
put
it
into
an
answer
file
and
read
it
off.
A
A
A
This
is
volume
management.
This
is
the
defaults
it
is.
It
is
turned
on
by
default.
If
you
want
to
have
kind
of
the
equivalent
of
amazon
ebs
elastic
block
storage,
then
you
want
to
turn
on
nova
as
well.
So
you
make
sure
that
that's
on
now
anyone
worked
with
neutron
got
scars
to
prove
it
right
now
for
an
all-in-one
on
a
laptop,
I
recommend
against
turning
on
neutron.
Just
do
regular
nova.
A
Networking
neutron
is
really
cool
and
it's
very
flexible
and
it
allows
you
to
set
up
vlans,
so
you
can
have
vlans
between
physical
hosts
or
to
do
gre
encapsulation,
which
is
a
really
cool
way
of
doing
networking
without
having
to
set
up
vlans
for
a
standalone
instance
on
a
laptop,
totally
unnecessary
and
will
actually
complicate
things
to
the
point
where
it
gets
kind
of
frustrating.
So
I
I
say
to
turn
it
off
for
a
standalone
make
sure
that
horizon
is
installed.
A
Unless
you
want
to
do
everything
from
the
command
line
and
that's
perfectly
reasonable,
you
can
do
everything
from
the
command
line.
I
like
to
have
a
graphical
user
interface,
because
it
makes
for
much
better
demos
trying
to
do
a
demo
of
something
in
front
of
a
room
full
of
folks.
Here,
I'm
going
to
type
this,
not
terribly
sexy.
A
A
Yeah,
one
of
the
things
to
remember
is
that
you
can
put
this
on
a
dedicated
partition.
That's
what
I
did
in
that
second
example
of
the
kickstart
file,
the
partitioning
section
of
the
kickstart
file.
I
do
that.
I
prefer
to
do
that.
If
you
don't
do
that,
then
it
will
it'll
actually
create
a
loopback
file
which
is
great
for
demos.
Don't
try
to
use
it
production,
it's
it'd,
be
slow.
A
Does
it
change
the
number
of
replicas
automatically
by
default?
It
does
single.
It
doesn't
do
any
replicas
because
we're
doing
a
one-on-one.
If
you
did
you
is
that
based
on
pac-stack
we'll
do
that
by
default,
or
is
that
because
you
did
a
gen
and
you're
doing
a
stand-alone
by
default?
Well,
no
pack
stack
by
default.
A
If
you
don't
have
an
answer
file,
if
you
just
do
packsack
and
hit
enter
it'll,
ask
you
yeah
it'll
prompt
you
for
the
number
of
replicas,
where
you
want
to
put
them
and
you
can
actually
go
out.
You
can
get
super
super
complex
with
pacsac
and
actually
it's
it's
kind
of
funny,
because
you
know
how
developers
are
right.
Oh,
I
wouldn't
use
that
thing.
I'm
going
to
do
it
all
myself
and
those
of
us
who
are
in
the
field
who
just
need
to
roll
this
stuff
out
quickly.
I'm
like
pax
acts
great.
A
So
it's
really
funny
because
some
folks
are
like.
Oh
I've
never
used
backstack
and
those
of
us
are
like.
I
want
to
get
the
same
world
out
quickly.
Packstack
works
pretty
well
and
in
the
answer
file
you
can
actually
do
things
like
put
multiple
servers,
and
so
you
can
say
I
want
these
four
machines
to
do
my
compute
nodes
and
I
want
to
you,
know,
switch
on
these
nodes.
I
mean
you
can
get
pretty
much
physics.
A
Yeah,
it's
actually
really
easy.
In
fact,
if
you
look
at
the
docs
at
rdo,
you
know
the
stream
the
instructions
there.
The
quick
start
is
basically
what
I'm
describing
right
here
and
then
the
very
bottom
of
the
quick
start
is,
and
if
you
want
to
do
multiple,
here's
how
to
do
it,
it
just
says
you
go
back
and
you
edit
the
the
file
and
add
additional
servers,
run
it
again
and
it'll
go
log
into
those
systems.
A
If
you
run
it
the
first
time
you
put
it
in
your
file,
it
asks
all
these
questions
and
then
at
the
end
it
says
if
you
want
to
update
just
run
this
command
cause.
It
takes
an
answer
right,
yeah
yeah
and
the
answer
file
that
it
does
like.
If
you
spend
a
lot
of
time
kind
of
sweating
through
all
the
details
and
figuring
it
out.
It's
really
nice
because
now
you've
got
an
answer
file
and
you
can
rerun
it
multiple
times.
A
In
fact,
I'll
show
you
how
to
do
that
just
a
minute,
all
right
so
and
then
do
you
want
to
have
the
administrative
tools,
the
command
line,
administrative
tools,
including
the
rc
file
that
sets
environment
variables.
I
always
say
yes,
because
there
are
some
things,
especially
if
you
are
doing
neutron
networking
which
absolutely
have
to
be
done
for
the
command
line.
You
can't
do
it
through
horizon
comma
separated
list
of
time
servers
make
sure
your
time
is
right.
A
You
will
have
a
very
bad
day
if
your
systems
are
really
significantly
off
in
time
because
they
saw
you
know
you'll
you'll,
you'll,
just
get
all
kinds
of
flaky
results.
I've
had
things
like
you
know,
vm,
just
wouldn't
start,
because
it
was
waiting
for
the
time
to
match
what
was
on
the
other
team.
They
were
off
by,
like
five
minutes.
Consider
pulling
my
hair
out
trying
to
figure
out
why
the
hell,
this
thing
won't
start.
A
A
All
right
and
then
also
if
you
want
to
do
an
installation
of
nagios
to
monitor
your
environment.
You
just
say
yes
here
and
it'll,
actually
go
through
instead
of
modulus
for
you
as
well,
so
for
each
one
of
the
following
services,
you
can
actually
define
the
host
upon
which
the
service
will
run.
So
you
can
get.
You
can
actually
get
very,
very
complex.
You
can
actually
install
services
across
multiple
physical
systems
but
cupid
for
messaging
amqp
for
messaging,
keystone,
identity
management,
glance,
sender,
nova,
etc.
A
Those
are
all
available
to
be
installed
on
multiple
machines.
This
is
what
that
looks
like
now,
because
I'm
doing
this
all
on
one
machine.
All
of
these
services
are
all
running
on
the
same
ip
address,
but
in
a
lot
of
cases
like
you've
got
swift
storage
hosts
anytime,
you
see,
plural,
where
it
says
hosts.
You
can
actually.
A
A
I've
seen
everything
from
just
make
sure
you
have
really
good
backups.
I
know
to
we're
just
going
to
go
with
it.
I've
I've
seen
people
talking
about
my
sql
replication,
just
using
mycycle's
native
replication,
there's
not
a
there's,
not
a
clear
standard.
Yet
that's
you
know.
Openstack
is
really
cool,
but
it's
still
got
a
lot
of
sharp
edges
and
it's
it's
really
still
a
very
niche
product
like
everyone
that
I
talk
to
is
interested
in
it.
A
But
then,
when
I
show
them
what
some
of
the
limitations
are,
and
I
do
the
pets
versus
cattle
analogy-
I
really
freak
out
about
shooting
the
cow
they're
like
oh
you
mean
I
can't
put
oracle
on
this
like
it's,
not
a
replacement
for
vmware,
and
so
then
you
have
to
have
that
whole
conversation
again
and
say
you
really
need
to
understand.
This
is
an
environment
that
is
actually
kind
of
designed
for
failure.
I
know
it
sounds
crazy,
but
you're
not
going
to
put
big
oracle
date.
Databases.
A
Yes,
yes,
yeah,
I
mean
you
can,
but
the
design
goal
originally
and
actually
kind
of
how
it's
progressing.
That's
all
stuff
that
people
are
putting
on
after
the
fact
and
that's
perfectly
acceptable-
I
mean
I
hope
I
mean.
I
hope
that
it
does
become
something
that
you
can
just
stop.
I
can
go
back
to
my
customers,
like
here's,
the
red
hat,
openstack
mug,
just
put
that
on
your
desk
when
you're
dude
comes
out.
A
A
Yeah
there's
a
lot
of
cool
stuff
going
on
there
upstream
or
there's
a
lot
of
stuff
going
on
in
the
community.
We
will
only
support
our
platforms,
but
if
you
want
other
platforms,
I
strongly
encourage
you
to
open
a
bugzilla
like.
I
really
really
encourage
you
to
open
a
bugzilla.
If
there
are
platforms
you
want,
because
we
don't
know
that
that's
valuable
unless
customers
tell
us
so
I
personally
would
like
to
see
a
supportive
platform.
I
did
not
say
that
you
did
not
hear
that.
A
Yeah
you
could
well,
I
know
community,
I
know
upstream.
Can
I
don't?
I
don't
do
that
yeah
I
mean
it's.
It's
core
yeah
yeah.
I
mean
yes,
we're
using
qemu
for
sure.
But
as
far
as
like,
are
you
talking
about
doing
things
like
nested
virtualization,
so
you
can
do
that
if
you
do
a
19
or
fedora
20,
you
can't
do
it
on
rail.
Yet
actually
you
won't
be
able
to
do
it
on
rail
until
real
7.
they're
not
going
to
back
toward
that
functionality.
What's
that?
A
A
I'm
gonna
talk
to
you
later
about
doctor
yeah
yeah,
you
seen
docker.
We
had
the
last
meet
up:
oh
yeah,
yeah,
okay!
Well,
you
just
mentioned
docker
and
we
recently
had
announcement
that
open
shift
and
docker
partnership
yeah,
so
doctor
is
awesome
so
anyway,
this
is
what
it
winds
up.
Looking
like
you
can
set
passwords
for
all
of
these
various
services.
If
you
don't
want
to
set
a
password
just
let
the
pacsac
generate
one,
you
know
pretty
good
passwords,
they're
long
and
complicated.
A
Now,
if
you
did
not
create
a
volume
group
for
sender,
you
can
let
pax
stack
create
one
and
the
way
you
do
that,
is
you
just
leave
that
as
create
cinder
volume
set
to?
Yes,
if
you
created
your
own,
then
set
that
to
no,
if
you're
going
to
let
it
create
it
set
the
size.
Otherwise,
just
whatever
the
size
of
the
volume
you
created
was.
If
you
did
set
yours
up,
make
sure
you
set
that
to
no.
A
For
the
the
flat
this,
the
terminology
with
openstack,
really
kind
of
irritates
me,
there
are
two
networks
that
you
need
to
be
aware
of
for
for
anyone
who
had
an
install
yet
there's
two
networks
that
you
need
to
be
aware
of
at
minimum.
There's
an
internal
private
network.
If
you're,
if
you've
used
amazon
ec2,
you
know
when
you
spin
a
virtual
machine
up,
it's
actually
got
ip
addresses
that
are
rfc
1918,
like
not
public
addresses.
We
do
the
same
thing.
Openstack
does
the
same
thing.
A
A
It
even
works
most
of
the
time,
but
sometimes
I've
actually
run
into
weird
issues
where
in
an
all-in-one
I
spin
up
a
vm
and
it
picks
up
an
address
from
an
external
dhcp
server.
If
I'm
buying
the
loopback,
I'm
not
sure
how
that
works,
but
it
does
and
it's
kind
of
bad,
so
be
aware
that
that
might
happen.
Yes,
that
was
that
was
a
bug.
A
Oh
good,
okay,
two
weeks
ago,
okay,
awesome
I've
been
on
the
road
solid
for
two
weeks.
We
were,
we
were
waiting
for
it
and
openshift
yeah.
We
have
proxies
yeah,
it
was
really
cool.
Doing
a
live
demo
in
front
of
a
group
bigger
than
this
like
and
you'll,
see
that
the
virtual
machine
comes
up.
Crap
wrong
address,
hang
on
just
a
second
yeah.
So
what's
that
people.redhat.com
cameron
and
I'll
put
that
up
at
the
end,
I
will
give
out
business
cards
and
my
my
people
pages
on
here
as
well.
A
So
so
anyway,
that's
what
that's
out
there.
So
that's
that
internal
network,
you
can
set
the
range
set
cidr
so
set
the
subnet
mask.
I
I
actually
bring
it
down
a
little
bit
because
it
does
like
this
massive
2048,
node
or
2048
address
space.
I
crank
it
back
down
to
a
class
c
cidr
now,
there's
another
network
that
you
need
to
be
aware
of,
and
that
is
your
public
network.
A
Now,
it's
confusing
because
I'm
using
private
ip
addresses
for
my
public
network,
but
this
is
you
know
when
you're
in
a
lab
environment
or
you're
behind
behind
the
firewall.
That's
probably
gonna
be
the
case
now
the
thing
that
that
is
odd,
about
the
way
that
you
define
that
public
network
is,
you
need
to
use
cidr
to
define
it.
A
You
want
to
put
those
in
a
different
range
from
what's
being
handed
out
to
your
physical
machines.
That
makes
sense,
because
if
you're
mapping
addresses
you
get
ip
conflicts
and
you'll
have
a
bad
day,
and
somebody
from
I.t
will
call
you-
and
you
don't
want
that.
So
I
set
the
floating
ip
address
range,
for
instance
172,
31
100.159.27,
so
I
I
carved
my
subnet
down
into
a
range
of
only.
A
I
think
I
did
32
yeah
for
that.
I
used
this
before
I
did
32.
yeah,
so
so
I
got
32
addresses
that
are
available
to
be
passed
through
to
the
private
ip
addresses
on
the
machines
and
the
way
that
you
do.
That
again,
is
you
say
my
floating
range
is
blah.
Blah
blah
and
you've
got
to
put
that
that
slash
whatever
you're
doing
to
set
the
range.
A
If
you
want
a
little
bit
better
performance,
what
you
can
do
is
you
can
say:
there's
the
ip
address
slash
the
device
as
it
exists
under
slash
dev,
so
don't
put
slash,
dev,
slash,
sdb1
just
put
slash
sdb1
or
if
you're,
using.
If
you're
using
volume
groups,
then
it
would
be
slash
vg,
my
computer,
slash,
lvmylb,
okay,
but
just
leave
the
dev
part
out.
A
You
can
pretty
much
leave
the
rest
of
it
alone.
In
this
example,
because
we're
using
red,
hat's,
openstack
distribution,
you
don't
need
to
enable
epel
extra
packages
for
enterprise
linux,
which
is
the
fedora
project
for
packages
for
rel
and
since
austin
scientific
and
all
those
other
respins.
So
and
then
you
can
leave
the
rest
of
it
pretty
much
alone.
The
rest
of
it
has
to
do
with
whether
or
not
you're
registered
to
rhn.
A
A
While
we're
waiting
or
while
we're
doing
this,
I'm
going
to
do
exactly
what
we
just
said,
I've
got
the
answer
file
and
I've
filled
it
all
out.
I
put
all
my
ip
addresses
and
stuff
like
that.
This
is
going
to
take
a
little
while,
so
what
I'm
going
to
do
is
pack
sack
oops
young
dash
y
install
openstack
dash
stack.
A
A
Answers.Txt,
oh
typo,
thank
you
yeah.
I
love
live
demos,
absolutely
answer
god
someday!
I'm
gonna
learn
how
to
type
I've
only
been
doing
this
for
20
years
there
we
go
all
right.
So
it's
going
to
ask
me
for
the
root
password,
I'm
going
to
put
that
in
all
right,
so
it's
going
to
run
through.
Let
me
get
back
over
into
the
slides,
because
this
does
take
a
little
while,
and
I
want
to
glass
through
this
as
quickly
as
possible.
A
A
A
A
Yeah,
it's
a
little
it's
a
little
exciting.
Sometimes
I
was
going
to
run
sack
on
this
machine
and
it
burst
into
flames,
and
I
went
so
I'm
going
to
bring
a
second
laptop
with
me
and
show
them
on
the
rel
6.4
laptop.
So
so
this
is
what
it
looks
like
right.
So
you
go
through
and
it'll
ask
you
for
the
passwords
for
each
of
the
machines.
It'll
go
through
and
run
through
all
these
puppet
jobs
and
we
use
puppet.
A
It
goes
through
and
it
runs
all
these
public
manifests
installs.
The
appropriate
rpms
makes
all
the
settings
pokes
all
the
entries
into
the
my
sql
database.
You
know
it
does.
It
builds
the
tables
and
dumps
information
into
them.
Sets
up
all
the
networking,
it'll
actually
install
a
newer
kernel.
If
you're
running
rule
6.4
it'll
run
a
it'll
install
an
openstack
kernel
that
includes
the
capacity
to
do
great,
which
we're
not
using.
A
We
didn't
define
neutron
net
or
quantum
is
what
it's
called
here,
but
it's
actually
been
updated
to
neutron,
so
we
won't
have
any
of
that
stuff
on
there.
But
eventually
what
ends
up
happening?
Is
you
get
a
screen
that
looks
like
this
it'll
say
we're
completed
your
your
rc
file,
which
is
where
all
your
like
administrative
password
and
the
urls
to
to
make
rest
api
calls
to
manage.
A
The
system
are
in
that
rc
file
and
then
it'll
set
up
your
nagios
information,
and
you
may
get
a
message
that
looks
like
this:
hey
a
kernel
package
with
net
namespace
support
has
been
installed.
You
need
to
reboot,
so
you
reboot.
If
that's
the
case
and
then,
as
we
mentioned
it,
it
saves
the
setup
log
file
and
it
does
create
an
answer
file.
If
you
just
ran
it
from
the
command
line,
it
got
prompted.
It's
really
really
a
very
smart
installer.
A
So
when
you
get
done,
you
can
look
in
the
keystone.
Rc
admin
file
and
you'll
see
the
administrator
password
right
there
and
it's
one
of
these
random
crazy
strings.
Unless
you
went
in
and
changed
it,
so
you
then
open
up
a
web
browser,
go
to
the
machine
and
you
fill
in
admin
and
that
password
there
are
a
couple
of
steps
that
you
need
to
take
to
get
your
environment
finished
up.
You
need
to
upload
a
system
image.
Now
we've
got
a
rel64
system
image
on
our
hn.
It's
really
easy
to
find.
A
I
mean
there's
a
ton
of
images
out
there
for
virtualization
for
kvm
virtualization
and
you
can
use
any
of
them,
for
if
you
do
it
off
rhn,
that's
what
it
looks
like
there's
kvm
guest
image
you
can
download
it
and
then
what
you're
going
to
do
is
in
the
web.
Ui
you're
going
to
go
click
on
the
project
and
choose
images
and
choose
create
image,
so
click
on
project
choose
images
and
snapshots.
Click
on
the
button
that
says
create
image.
A
If
you're
going
to
populate
this
box,
give
it
a
name-
and
that's
the
name
of
the
image-
remember
that
that
image
is
going
to
be
used
to
spawn
up
multiple
instances.
So
I
generally
recommend
you
give
it
kind
of
a
generic
name.
I
called
it
real
6q
cow
2,
give
it
the
path
to
the
image,
the
format
of
the
image
and
go
to
town.
You
can
leave
these
alone,
the
minimum,
disc
and
minimum
round
ram.
One
thing
you
do
want
to
do
is
make
sure
you
check
the
box.
A
That
says
this
is
a
public
image
that
means
that
anyone
who,
logs
into
the
interface
it's
got
an
account
and
use
this
image
to
spin
up
a
guess
when
you
do
that,
it
actually
stores
it
remember
glance
is
image
storage.
So,
if
you
look
under
bar
lib
glance,
images
there's
a
funky
named
file
in
there.
If
you're
in
file
again
it
you
can
see
it's
a
qemu
image.
So
when
you
upload
your
image,
it
gets
stored
in
glance
storage
after
a
few
minutes,
you'll
see
that
it's
active
and
it's
ready
to
go
now.
A
Have
you
oh
keep
here
we
set
up
key
pairs
and
actually
what's
cool
is
when
we
launch
this
image,
we
actually
use
the
sfs
to
poke
the
key
pair
into
it.
So
we
actually
mount
the
image
poke
in
the
ssh
keys.
So
you
need
to
create
a
pair
of
ssh
keys.
The
way
that
you
do
that
is,
go
into
project
access
and
security
key
pairs
and
then
choose
create
a
key
pair.
Give
it
a
name
anything
you
want,
and
it'll
save
a
pen
file
to
your
local
file
system.
A
A
So
now
that
we've
defined
access
for
ssh
and
we've
defined
what
our
vm
image
looks
like,
we
can
go
spin
up
an
instance
and
the
way
that
we
do
that
is
go
into.
Actually
you
go
into
projects
instances
and
choose
launch
instance:
you'll
be
presented
with
the
dialog
box.
That
gives
you
a
couple
of
options
you
want
to
do.
A
A
So
the
next
thing
that
you
want
to
do
is
just
make
sure
that
it
came
up
cleanly
and
it
booted
correctly.
The
way
that
you
do
that
is
you
go
into
the
console
to
make
sure
that
it
booted
up.
So
you
go
to
the
machine,
you
click
the
drop
down
right
there
and
choose
console
and
you'll
see.
I've
got
a
machine.
I
got
a
virtual
machine,
crazy
thing
about
this,
and
I
don't
know
if
the
upstream
openstack
is
like
this.
You
can't
type
in
this
console.
A
A
You
have
to
do
something
crazy.
You
have
to
click
the
link
that
says
show
only
the
console.
Then
it
passes
through
keystrokes.
I
don't
know,
that's
a
bug
or
if
it
does
it
in
upstream,
doesn't
do
it
upstream
as
well.
Okay,
good,
because
I
was
sitting
there
going
to
the
end:
is
it
okay?
So
this
is
not
a
full
vnc
session.
It's
just
display
only.
A
All
right
so
now
you've
got
the
machina,
that's
great,
but
it
can't
do
anything
yet,
or
at
least
you
can't
get
to
it
from
the
outside.
Yet
so
you
can
associate
a
floating
ip
address.
Remember
we
defined
that
range
of
floating
ip
addresses
and
the
way
that
you
do
that
is
you
click
on
the
instance.
You
choose
associate
floating
ip
now,
one
thing
to
be
aware
of
because
we're
using
nova
networking,
because
this
is
a
really
really
simple
installation.
When
you
choose
associated
floating
ip,
it's
going
to
go.
Okay,
I'm
associating
the
next
one
available.
A
You
don't
have
any
choices,
what
it
is
or
anything
like
that
it's
just
going
to
assign
the
next
one
available.
If
you
do
more
advanced
networking
neutral
networking,
you
can
actually
get
really
granular
as
to
how
you
want
to
do
it
and
what
you
want
to
assign
and
some
pretty
cool
stuff
like
that.
But
again,
this
is
really
just
kind
of
a
quick
demo
of
how
that
stuff
all
works.
So
when
I
click
on
okay,
it
just
grabs
the
next
available
one.
So
I've
got
the
172
31..
A
A
You
have
to
open
up
ports
to
go
from
the
public
address
to
the
private
address.
The
mapping
is
there,
but
there's
no
ports
that
are
opened
up
and
so
the
way
that
you
do
that
is
fairly
straightforward.
You
go
into
the
project,
go
to
access
and
security
click
on
security
groups.
You
can
either
add
a
new
group
or
you
can
just
edit
the
existing
one.
There's
really
good
arguments
on
both
sides
of
the
house
as
to
whether
you
should
edit
the
existing
one
or
add
a
new
one.
I'm
not
going
to
fight
that
fight.
A
You
do
whatever
you
want.
So
in
this
case
I
chose
add
a
rule
and
it's
fairly
straightforward.
I
go
and
I
choose
the
protocol
tcp
or
udp
or
even
icmp.
There's
all
kinds
of
protocols
there
choose
the
port
and
then
choose
how
you're
going
to
define
the
source
or
you're
going
to
do
it
from
an
individual
host
or
you're
going
to
do
a
subnet
or
a
network.
In
this
case
you
choose
cidr
and
just
do
0.0.0.0,
and
that
says
I'm
opening
up
port
22
tcp
from
anywhere
to
this
instance.
A
A
Because
it
also
protects
yeah
from
each
other.
Yes,
yes,
so
the
the
public,
the
public
is
just
it's
just
passing
through
right.
Right
did,
I
say
differently.
Yeah.
I
think
you
said:
that's
fine,
okay,
my
apologies!
That
was
incorrect.
If
I
said
no,
I
spoke
incorrectly.
Thank
you.
So
yes,
that
is
correct.
So,
but.
A
The
like
some
some
people,
or
sometimes
what
I
wound
up
doing
is
if
I'm
like,
I'm
gonna,
do
iv
tables
rules,
I'll
just
say,
open
up
from
1
to
65
535,
which
is
actually
kind
of
stupid.
But
if
you're
not
running
nfs
and
you're
not
running
anything
else's
network
listening,
then
you
can
instead
of
having
two
layers
of
security.
You
can
do
everything
with
ip
tables
rules.
But,
yes,
you
are
correct.
That's
actually
on
the
guest
itself,
all
right!
So
now
we're
going
to
ssh
in
we've
opened
up
port
22..
A
So
we're
going
to
ssh
in
remember
that
ssh
is
real
cranky
about
permissions
for
pen
files,
so
you
got
to
do
chmod
600
on
it
and
then
you
could
do
ssh
root
at
host,
name
or
ssh
cloud
dash
user
at
host
name.
They
both
work-
and
I
was
under
the
impression
that
you
weren't
supposed
to
be
able
to
ssh
in
is
root
or
it
was
supposed
to
be
a
separate,
but
it
works.
It
works
for
root
or
cloud
user.
That's
your
image!
Yeah!
That's!
Basically,
yeah!
It's
it's
what
the
image
but
see.
A
I
thought
that
our
image
was
hardened
and
I'm
sorry
I've
been
plugging
the
guys
on
the
rdo
group
and
I'm
like
supposed
to
be
able
to
actually
work
the
cloud
yeah.
That's
right.
The
cloud
user
is
busted
right
now.
The
cloud
user
setup
is
there's
something
broken
about
it.
You
know,
do
you
remember
what
it
was?
A
Yeah
yeah:
we
need
to
fix
that
all
right
cool.
So
then
you
can
get
into
some
of
the
fancier
stuff,
like
add
a
volume,
so
you
can
attach
a
volume
like
similar
to
amazon
ebs
and
the
way
that
you
do,
that
is
you
go
into
project,
go
to
volumes
and
choose
create
a
volume.
It's
real,
simple!
You
just
go
in
and
choose
the
size
so
give
it
a
name,
give
it
a
type
and
give
it
the
size,
and
actually
you
don't
even
need
to
give
it
a
type
just
leave
it
leave
it
blank.
A
So
I
created
a
five
gig
volume
and
clicked
on
create
volume
and
it'll
it'll
show
that
it's
working
for
a
little
while
because
it's
actually
carving
out
that
volume
and
it's
carving
that
volume
out
of
the
sitter
volumes
volume
group
and
then,
when
you
get
done,
it'll
be
done.
It'll
it'll
show
that
it's
available
and
you
can
choose
edit
attachments.
A
Then,
when
you
do
that,
it's
fairly
straightforward,
when
you
click
on
edit
attachments,
you'll
have
a
list
of
machines
that
you
can
attach
it
to,
and
then
you
can
tell
it
what
device
name
you
want
to
give
it
and
it
will
ignore
it
because
it
will
give
it
the
next
device.
That's
available,
it'll
be
devv,
whatever
the
next
block,
whatever
the
next
disk
is
available
so
and
when
you
get
done,
it'll
take
a
few
seconds.
It'll
show
that
it's
attaching
and
eventually
it'll
say
that
it
has
attached.
A
So
it
is
attached
on
slash
dev
vdc,
ignore
that
it'll
just
be
whatever
the
next
block
device
is
available,
is
and
if
you
go
and
look
at
the
console,
you'll
see
all
of
the
the
traffic
that
gets
flashed
out
to
the
screen.
Saying
hey,
I
got
a
new
block
device
and
I'm
ready
to
go.
It
is
stored
under
cinder
volumes
volume
group.
So
if
you
go
and
you
do
lv
display
you'll
see
this
really
crazy.
Uuid
name
that's
assigned
to
that
volume.
But
this
is
for
this
is
persistent.
A
A
A
A
A
A
A
So
so
on
the
machine,
this
machine
down
here
it
got
done
and
it
said
that
it
installed
successfully
kept
the
log
file.
Now
it
did
install
a
newer
kernel
for
some
reason,
with
this
latest
spin
of
red
hat
openstack
we're
not
getting
the
message
that
says
you
need
to
reboot,
but
you
do
need
to
reboot
because
it
does
install
the
kernel
that
does
network
namespaces.
A
What
does
it
use
networking
spaces
for
network
namespaces
is
awesome.
It
doesn't
use
it
for
nova
networking,
but
if
you
have
quantum
or
neutron
as
it's
called
now,
network
namespaces
is
really
cool.
It'll
actually
allow
you
to
do
some
really
neat
things
like
when
you're
defining
tenant
networks
or
networks
for
for
different
organizations
within
your
openstack
installation.
You
can
actually
assign
out,
like
192.168.0.0.24.
A
A
It's
pretty
crazy,
but
it's
very
cool
see
if
we
came
back
up.
You
know,
I
think
the
single
boot
with
the
lid
closed
so
we'll
find
out
all
right.
Would
you
call
it
support
for
sdn
enabled
routers
as
well
or
is
it
just
using
the
virtual
interface
say
again,
I'm
sorry
software
configurable
routers
right?
Does
it
provide
support
for
that
or
it's
just
accomplishing
everything.
A
Everything
that
we're
doing
right
now
is
using
is
using
internal
networking
or
open
v
switch.
So
if
you
have
hardware
that
does
like
that's
open
or
not
open
switch
but
sdn
capable,
I
don't
believe
that
we
have
anything
that
talks
to
those
routers,
because
we've
we've
really
tried
to
make
it
so
that
everything
is
self-contained
in
software
right.
So
it
is
an
easy
way
to
extend
that
patch
is
cheerfully
accepted.
A
A
A
I
think
so
that
sounds
right.
Will
it
work
on
centos
rdo
will
work
on
sometimes
right.
So
I
guess
my
best
question
is
without
a
subscription
to
rhythm,
so
on
this
machine
that
I
actually
just
did
right
here.
I
actually
synced
the
repos
locally,
because
I
wasn't
sure
if
I'd
have
wi-fi
absolutely
you
need
a
repo
and
you
can
you
can
sync
the
rdo
repo
to
anything
it'll
work
on
synthos
or
scientific
or
red
hat
enterprise,
or
you
know
any
other
response
or
anything
like
that.
A
A
A
A
A
A
A
Nice,
one
and
honestly,
I'm
going
to
reboot,
because
I've
tried
to
restart
swift
services
and
there's
like
18
gazillion
of
them
right
now,
though,
what's
that
what's
going
to
be
doing,
I
have
no
idea.
I
haven't
figured
that
out
yet
I
don't
know,
but
I
haven't
set
up
anything
yet
I
know
I
know,
but
it
it
happens.
A
A
No,
no!
It's
a
super
super
simple
services.
Yeah!
It's!
It's
not
setting
up
a
puppet
infrastructure.
Okay,
so
it
looks
like
we're
good
there,
all
right
cool.
So
now
again
what
you
guys
saw
me
do
already,
I'm
just
going
to
do
again.
Let
me
get
the
web
browser
over
this
screen
right
here,
so
I'll
go
to
host237.tc.redhat.com.
A
A
A
And
there's
my
there's
my
image
file
again.
If
I
run
file
against
that
it'll
there
we
go
come
back
and
it
says
there
it
is
qemu
image
so
and
you
notice
that
in
the
background
it
says
success
so
now
I've
got
that
image
file
up
there.
So
the
next
thing
that
I
want
to
do
well,
there's
a
couple
things
that
I
could
do
at
this
point.
A
A
So
if
you
have
a,
if
you
have
a
key
pair
that
you
already
use
that
works
just
fine,
you
can
upload
it
as
well.
That's
fine
and
then
I'm
also
going
to
go
ahead
and
I'm
just
going
to
edit
the
existing
security
groups.
I
know
it
makes
sense
to
build
one
put
a
new
one
in
place.
That's
perfectly
acceptable,
but
for
now
I'm
just
going
to
say
I'm
going
to
allow
tcp
port
22
to
connect
from
anywhere
so
click
on
add.
A
A
A
A
You
have
to
have
allowed
icmp
if
you
want
to
do
pain
and
actually,
while
that's
happening,
let's
do
that
because
again
this
hope,
they're
gonna,
you
can
hang
out
yeah.
You
should
be
able
to
ping
out
yeah,
there's
no
algorithm,
but
if
you
want
to
allow,
if
you
want
to
ping
back
in
then
I
do
minus
one,
because
I
I
think
it's
silly
to
block
icmp
I
I
guess
I
think
your
question
is
more.
Are
you
routing
out
your
fixed
network?
Yes,
okay!
A
Oops
you
say
again:
well
I
guess
what
interfaces
is.
Has
the
route?
Is
it
the
public
one
or
is
it
on
fixed,
it'll
it'll?
Actually,
it
should
go
out
through
the
through
the
public
interface.
Hang
on
a
second,
it
spins
up
a
bridge
for
you
based
on
what
what
or
it
defines
a
bridge
for
you,
based
on
what
you
put
in
the
back
in
the
private
right
and
then.
A
A
A
A
The
boarding
enables
right,
I
think
the
rest
of
it
is
fine,
so
you
just
set
that
and
then
your
private
id
address
can
matter
and
when
you
start
floating
address,
oh
wow,
oh
did
we
do
that
and
does
pax
like
do
that
or
did
because
the
falls
for
rail
is
off,
so
I'm
assuming
packs
that
did
it.
I
don't
I
don't
know,
but
the
end
result
is
or
should.
A
My
puppet,
while
it
still
can
be
managed
manually,
it
is
definitely
not
recommended
yeah
ignore
that,
because
puppet
doesn't
leave
any
infrastructure
behind.
So
yes,
it
did,
it
went
through
and
it
fixed
all
that
stuff
for
you
and
actually,
if
you
look
at
that,
it
turned
down
it,
it
did
a
bunch
of
stuff,
it
did
a
bunch
of
stuff
on
the
bridging,
so
yeah,
that's
actually
really
impressive.
A
A
So
I'm
going
to
log
back
into
that
thing
and
let's
do
ping
at
100..
I
think
my
laptop
is
60.
and
there
you
go.
It
works.
So,
yes,
you
can
hang
out
and
once
you
set
up
rules
you
can
ping
in
as
well
so
yeah,
it's
pretty
cool
stuff.
I
mean
I
I'm
I'm
digging.
What
I'm
I'm
still
learning
I
mean
openstack
is
still
so
new
that
there's
still
a
lot
of
sort
of
you
know
sharp
edges
and
stuff.
If
you
want
to
go
further,
you
can
go
into
volumes.
A
A
A
Let's
do
that
so
there
you
go.
It
created
that
volume
four
gigs,
it's
been
a
long
week,
so
now
we
should
actually
actually
be
done
so
now
I
can
edit
the
attachment
I
can
say
attach
it
to
that
instance.
I
can
call
it
slash,
dev
slash
whatever
I
want.
It's
still
going
to
be
ddb
because
that's
the
next
available
block
device.
So
I
attach
it.
I
think
it
depends
on
your
guest,
always
I
I
know
I
yeah
it
does
work
because
I
did.
I
tried.
A
A
So
that's
it.
That's
I
mean
that's
a
super
super
basic
overview
of
red
hat's
implementation
of
openstack.
Any
questions.
Yes,
I'm
thinking
about
key
management.
Is
there
any
way
to
redefine
the
set?
I'm
sorry
previously
here,
a
set
of
key
pairs
like
20
or
30
or
whatever,
for
it
to
use
as
a
tool,
so
whatever
it
spins
up
in
the
license,
picks
the
next
one
or.
However,
it's
named
key
mapping
or
something
like
that
that
way,
you're
adding
a
lot
bigger.
A
No
there's
nothing
built
in
that
I'm
aware
of,
but
that
I
mean
that
shouldn't
be
difficult
to
do
you
can
generate.
You
know
you
can
do
ssh
keygen
multiple
times,
it's
all
like
yeah,
but
I
would
do
yeah,
but
I
would
the
common
pattern
seems
to
be
like
keeper
key
pair
per
attack.
Yes,
there's
generally
a
tennessee's
like
project
under
themselves,.
A
Right
so
yeah,
it
actually
took
me
a
while
to
kind
of
wrap
my
head
around
the
idea
of
tenant
networking
and
what
is
a
tenant
and
and-
and
I
use
project
interchangeably
because
you
know
it's
not
confusing
enough,
as
is
but
what's
that
yeah
domain
yeah.
So
so
it
was
described
to
me
the
most
common
or
that
a
common
use
case
for
a
tenant
network
would
be
a
customer.
So
maybe
I've
got
a
customer.
That's
an
internal
customer
like
a
business
unit
and
they
say
I
need
you
know.
A
I
need
to
be
able
to
do
a
burst
because
we
got
a
christmas
rush
coming
or
I'm
going
to
be
doing
big
number
crunching
and
I
need
to
so
that
might
be
one
tenant
network
and
you
give
them
their
own
user
accounts
and
they
could
log
in
and
spin
up
vms
to
their
heart's
content,
based
on
what
limits
you
set
and
then
another
tenant,
maybe
another
business
unit,
or
maybe
another
customer
externally
or
whatever.
But
once
I
finally
got
that
I
was
like
okay,
that's
what
that
means.
Yeah!
A
It's
often
a
little
bit
deep
tonight
because,
with
the
security
groups,
they're
a
firsthand
thing
yeah.
So
if
you
want
to
be
able
to
set
security
groups
between
your
website
and
if
you
have
a
database,
you
actually
want
to
have
them
in
separate
tenants
or
you
just
use
tables
on
the
guests
themselves.
Yeah.
A
So
was
this
helpful?
Okay
cool?
So
if
you
take
no
other
message
away
from
this,
but
but
this
it
is
that
we
are
absolutely
committed
to
openstack
as
a
company.
Red
hat
is
committed.
We're
not
messing
around
there's
a
reason:
we're
never
going
to
contribute
to
grizzly
we're
working
like
crazy
to
do
as
much
as
we
can
to
make
havana
really
stable.
A
We
see
this
as
incredibly
valuable
and
we're
not
we're
not
committing
all
this
code,
because
we
want
to
dump
our
chests
and
talk
about
how
awesome
we
are
we're
committing
this
code
because
we
are
committed
to
the
extreme
community.
Don't
get
me
wrong,
there's
a
little
thumping
of
chess
going
on
there.
No
no
argument
there,
I'm
proud
to
work
for
red
hat,
I'm
proud
of
the
work
that
we're
doing,
but
but
we
absolutely
see
this
is
valuable.
A
We
see
the
community
as
being
incredibly
valuable,
I'm
not
kidding
when
I
say
if
you
want
something
done,
please
open
up
a
feature,
request
or
open
a
bookzilla.
You
know
plugzilla.com
help
us
to
make
it
better.
It
doesn't
get
better
without
your
input
all
right.
Thank
you
very
much.
I
appreciate
it.