►
Description
Let's take a look at a UBI-based container and see how we can be sure it can build and run in any scenario.
How Red Hat Enterprise Linux (RHEL) users and admins can benefit their organizations and improve their careers by learning how to use containers, Kubernetes, and Red Hat OpenShift.
Learn more at https://red.ht/leveluphour
A
All
right
good
morning,
good
afternoon,
good
evening
and
welcome
to
another
episode
of
the
level
up
hour
here
on
openshift
tv
I
am
chris
short
principal
technical
marketing
manager
and
executive
producer.
Of
this
thing
we
call
open
shift
tv,
I'm
joined
by
the
illustrious
langdon
white
good
morning,
langdon.
How
are
you,
sir
good.
B
Good,
as
you
can
see,
from
the
reflection
of
my
glasses,
as
we
were
just
discussing.
B
I
bent
my
normal
glasses,
and
so
I
am
wearing
the
worst
ones
yeah
and
I
just
look
like
a
blinding
light
every
time
I
look
up
at
the
screen,
which
kind
of.
A
Well,
if
you,
if
you
haven't
watched
the
channel
recently,
you
can
tell
that
I
have
changed
locations
within
my
house.
I
am
out
of
the
unfinished
basement,
aka
dungeon
and
in
our
guest
bedroom.
While
the
basement
is
actually
gonna
be
finished
slowly,
but
surely
over
the
winter
and
spring
months,
so
yeah
we're
making
moves
here
in
the
short
household
as
it
were.
B
Yes,
although
I
do,
I
will
say,
I
missed
the
flag
so.
A
So
I
so
I
talked
to
my
wife
about
that
and
I
was
like
you
know.
I
kind
of
don't
want
to
put
the
flags
up
in
the
guest
room
because
I
feel
like
I
could
do
something
better
and
she
was
like.
Oh
thank
god,
it
looks.
D
B
Yeah
yeah,
but
the
the
the
plain
wall
is
still
kind
of
boring.
So
you
definitely.
A
A
D
A
A
B
Your
your
art,
or
you
know
things
that
you're
hanging
on
the
wall
will
end
up.
Being
you
know,
bodybuick
boat,
face.
A
A
Be
hanging
up
behind
me
and
it
will
remind
you
where
I
am
just
like:
the
flags
did
and
I'll
leave
it
at
that
you'll
see
it
when
it
gets
installed,
which
might.
B
I
do
I
will
say
I
like
christian's
soccer,
whatever
we
call
those
scarves.
A
Yeah,
if
you're,
if
you
tune
into
the
the
get-offs
happy
hour
at
3
p.m,
eastern
tomorrow,
you
will
see
his
his
pretty
cool
soccer,
setup,
yeah
and
yeah.
No,
that's.
B
Actually,
that's
right,
speaking
of
which
I
need
to
follow
up
on
the
devconf
swag
this
year
is
a
dev
conf
scarf
like
a
you
know
like.
B
Yeah,
you
know
because
we
could
go
then
we
could
go.
You
know
kind
of
non-gender
specific
right.
So
it's
one
less
piece
of
information
we
have
to
collect
from
everyone.
B
And
they're
relatively
small
and
easy
to
ship,
and
it's
something
that
you
could
wear
like
on
camera.
You
know
so
as
it
gets
colder
right,
so
we
thought
it'd
be
fun,
so
hopefully
they
came
out
well,
but
we
still
haven't
gotten
them
because
our
logistics
team
it
got
superseded
by
some.
You
know
real
corporate
thing.
You
know
so
yeah.
C
C
B
All
right
so
for
anyone
who
is
not
sure
where
they
are,
this
is
the
level
up
hour
where
we
talk
about,
why
containers
are
cool
and
why
you
might
want
to
use
them,
and
we,
let's
see
that's.
B
B
Be
familiar
with
them,
yeah,
yeah,
yeah,
exactly,
and
why
you
know,
since
I
don't
know
some
ridiculous
amount
of
time
ago,
I've
been
basically
using
them
exclusively
for
almost
everything.
I
do.
I
kind
of
make
containers,
sometimes
to
my
detriment
where
I
have
it.
B
A
B
And
christian
may
actually
be
here,
but
it
is
very
early
for
him.
B
B
So
you
can
also
find
us
on
discord
where
we're
usually
pretty
responsive.
A
A
Think
it's
10
or
it
might
be
nine.
Now
I
forget:
okay
with
the
time.
B
A
But
I
know
when
I
like
so
to
give
you
an
idea
of
my
morning
routine.
The
first
thing
I
do
in
the
morning
at
7
00
am,
as
I
send
out
the
like
schedule,
email
to
the
day
internally-
and
I
know
at
that
time
it's
the
day
all
around
the
world.
So
I
can
actually
say
like
happy
wednesday
to
everybody
right
like.
B
Interesting
so
yeah
so
find
us
on
discord
and
then
so
what
we're
going
to
talk
about
today
is
what
I'm
kind
of
referring
to
is
resilient
ubi.
B
And
so,
when
I
say
resilient,
I
don't
mean
it
kind
of
technically
traditionally,
but
basically
one
of
the
things
about
ubi
right
is
that
if
you
are
on
a
registered
railback
a
box,
you
kind
of
have
a
different
package
set
available
than
if
you're
not,
and
so,
if
you
want
to
make
a
container
file
that
will
work
kind
of
for
anybody,
so
ubi
isn't
really
intended
in
a
sense
right,
isn't
really
intended
for
kind
of
distributing
container
files.
B
So
that's
kind
of
what
is
important,
or
at
least
I
think
it's
important.
So
that's
what
we
kind
of
want
to
talk
about
today
so
but,
as
always,
show
notes
from
last
time,
and
just
by
way
of
apology,
the
show
notes
from
the
time
before
are
also
in
the
repo
now,
even
though
they
weren't
ready
in
time
for
the
show
last
time
so,
but
both
of
them
are
there
now,
please
feel
free
to
comment
or
whatever
on
the
repo.
B
If
you
think
there's
anything
missing
or
I
glossed
over,
something
that
I
shouldn't
have
this
particular
episode
I
actually
put
in
there's
a
lot
more
further
reading
in
a
sense
than
I
normally
did,
primarily
because
of
kind
of
the
topic
area
right.
We
were
talking
about
ai
and
ml,
which
we
kind
of
glossed
over
a
lot
of
what
that
is
in
this.
B
Well,
I
have
a
particular
bias
because
I
I
did
a
lot
of
ai
work
like
grant,
funded
work
in
college
and
then.
C
B
Actually
went
and
worked
at
a
r
d
center
and
worked
at
on
ai
there
as
well
sort
of,
but
what
was
I
going
to
say?
Oh
so
well
so,
interestingly
enough,
because
we
did
have
like
a
good
engaging
discussion
in
the
show.
Last
time
I
have
teed
up
somebody
to
come
and
talk
some
more
about
kind
of
machine
learning,
and
specifically
the
problem
like
or
what's
interesting
about.
B
Containers
right
is
one
of
the
things
we
like
about
containers
or
images
right
is
that
you
have
something
that's
very
replicable
right,
so
it
worked
on
my
machine.
Isn't
really
a
thing
because
you
can
say
here
is
literally
what
worked
on
my
machine
now
you
can
run
it
and
it'll
work
on
your
machine
in
science
of
any
kind
really,
but
in
data
science
in
particular
shipping.
A
A
B
Right
right,
so
we
so
yeah,
so
we
chatted
yesterday
and
hopefully
we'll
be
able
to
get
him
on
the
show,
not
next
time
but
the
one
after
and
then
next
time
we're
going
to
talk
a
little
bit
more
about
nexcloud.
B
But
this
time
I
will
have
abstracts
out
pretty
soon.
So
that
should
be
exciting
and
we'll
also
be.
Will
we
be
on
other
channels
by
next
episode,
so.
A
Your
next
episode-
yes
special
announcement-
the
there
will
be
some
shows
that
will
be
crossing
the
chasm
as
it
were
over
to
the
official
red
hat
youtube
channel,
which
is
a
super
big
deal
for
us
here
on
openshift
tv
right,
because
we
started
this
within
the
the
cloud
platforms
business
unit
and
the
our
corporate
friends
realized
that
like
this
is
a
good
idea
and
the
the
team
lead
for
that
team
is
actually
someone
went
to
high
school
with
anna
stark.
A
A
Oh
that's
cool
anyway,
so
we
will
have
the
level
up
hour
on
the
58
000
subscriber
red
hat
youtube
the
next
go
round,
so
they
are
super
excited,
I'm
super
excited,
and
tomorrow
the
in
the
clouds
with
red
hat
leadership
with
ashesh
mandani
senior
vice
president
of
cloud
platforms
here
at
red
hat,
that
show
will
also
be
on
the
the
main
red
hat
youtube
channel.
A
So
we're
super
excited
about
it
right
because
it
shows
that
what
we're
doing
here
is
is
engaging,
is
working
and
needs
more
eyeballs
right
for
lack
of
a
better
term.
So
yeah
it's
it's
it's
like
we're
growing
up
and-
and
you
know,
you're
you're,
our
lucky
guinea
pig.
I
mean
awesome,
awesome,
awesome
early.
B
Does
that
mean
you're
coming
full
circle
with
having
a
shesh
on
for
the
first
red
hat
channel?
Show
because
he's
the
head
of
cloud
platforms.
We
started
in
class
that.
A
Is
interesting
kind
of
but
like
so
yes
but
we're
having
a
a
very,
very
special
guest
on
the
17th
for
that
in
the
clouds
with
red
hat
leadership,
matt
hicks
will
be
joining
us
he's.
A
C
A
A
Yeah
pretty
much
you
know
and
like
that's,
how
we've
been
describing
it
to
people
when
we
talk
about
it
right,
like
it's
just
syndication
at
that
point,
when
we
stream
to
all
these
different,
you
know
channels
right
from
facebook
to
periscope
to
youtube
and
twitch
the
oh
does
that
make
us
all
t.
Yes,
you
all
be
officially
the
level
up
our
hipsters,
because
you
were
here
before
anybody
else.
B
B
So
yeah
all
right,
we
should
probably
I
don't
know
talk
about
some
stuff.
B
Great
yeah,
we
were
joking
around
last
time
actually
about
how
it
took
us
so
long
to
actually
get
to
any
tech
stuff.
So
what
was
I
looking
for?
Okay,
so
I
was
hoping
that
I
was
going
to
use
a
vs
code
for
the
show
today.
Fancy.
A
B
B
All
right,
so
let
me
know
if
you
oh,
I
have
to
actually
finish
sharing.
Let
me
know
if
you
can't
actually
read
that
that's
very
legible.
B
B
Yeah,
actually
you
know
I
don't
know,
but
I
am
curious.
It
is
a
good
font,
it
might
be
source
code
sans.
C
B
It's
an
open
source
font,
I
I
wanna
say
done
by
red,
hat
or
or
maybe
fedora
or
maybe
gnome,
but
I'm
pretty
sure
it's
an
open
source
font
and
I
like
it.
I
also
love
vs
code.
I
particularly
if
you
haven't
checked
it
out
and
you're
kind
of
on
the
more
advanced
front
on
the
container
world
check
out
the
remote
containers
stuff,
which
is
pretty
amazing,
not
open
source,
although
there
does
seems
to
be
an
talking
about
them
using
it.
Oh
is
it
a
google?
B
B
All
right
there's
something
there
anyways,
let's
go
ahead,
so
okay,
so
what
I
did
was,
let
me
actually
let
me
start
with
a
little
bit
more
context.
All
right,
so
here,
I'm
gonna,
throw
in
the
chat
here
is
the
repo
that
I'm
working
from,
which
is
actually
a
fork
of
christian
hernandez's
repo
on
his
helper
node
yeah,
helper
node
is
like.
C
A
Yeah,
so
we
actually
have
our
own,
it
has
its
own
short
code.
There
you
go.
B
Yeah
so
helper
node
is
this
thing
that
helps
you
with
openshift
and
he's
been
working
on.
I
think
maybe
kind
of
a
an
improv,
an
improvement
right
because
going
by
the
helper
node
dev2
right
so,
but
he
was
asking
me
some
questions
about
how
to
do
the
ubi
to
basically
the
problem
here
is
this
dhcp
server
so.
B
However,
available
in
the
rel
or
in
the
red
hat
or
the
rel
repos
right,
so
what
I,
when
I
looked
at
this,
so
there's
some
things
that
I
thought
this
would
be
useful
to
do,
because
we
could
kind
of
talk
about.
You
know
we're
getting
along
on
how
we've
done
containers
right,
so
I
kind
of
wanted
to
say:
okay,
let's
take
a
container
file,
that's
kind
of
in
the
wild
and
look
at
is
it?
Is
it
well
done,
or
should
we
or
can
we
do
some
improvements
to
it?
B
And
so
I
think
there
are
some
improvements
we
could
do
for
it.
However,
one
of
the
major
ones
I
wanted
to
do
is
not
as
much
of
a
problem
as
I
expected
so
we'll
get
into
that
in
a
second.
Actually,
you
know
what
I
should
have
done.
Sorry,
let
me
share
the
screen
differently.
B
Yeah
exactly
and
let
me
find
a
terminal
window
which
I
thought
I
had
opened
in
advance.
B
B
What
I
wanted
to
show,
first
and
foremost,
okay,
so
I
built
this
container.
So
we
look
at
podman
images
and
let
me
stretch
that
out
a
little
bit,
so
it's
a
little
more
legible
and
we
can
say,
let's
kind
of
jump
into
the
container
okay.
So,
first
and
foremost,
this
is
a
registered
sorry.
I
lost
the
twitch
stream
there.
It
is
okay,
so
this
is
a
registered
rel
machine.
B
I
can't
figure
out
when
subscription
manager
will
get
type
ahead
and
when
it
won't,
sometimes
I
get
it,
and
sometimes
I
don't
or
not,
type
ahead.
Tab
completion,
auto
completion.
B
Maybe
if
sudo
has
been
launched
recently,
I
don't
know
all
right,
so
just
checking
it
is
actually
registered.
So
that's
good
one
of
these
days.
Maybe
we
should
talk
about
why?
What
system
purpose
is
because
I
had
a
hand
in
that
being
or
occurring
so
podman
rm.
B
And
so,
if
we
haven't
talked
about
it
on
the
show
which
we
may
not
have,
if
there's
an
entry
point
on
the
container
you're
trying
to
run
the
only
way
to
bypass
it,
is
by
passing
this
entry
point
flag.
If
you
put
bin
bash
at
the
end
of
your
run
command,
it
won't
work,
it'll
just
get
ignored
and
entry
point
will
be
there.
So
this
one
actually
has
an
entry
point.
B
I
do
it
if
for
any
container
image
that
I
didn't
directly
create
and
know
inside
and
out
and
yeah,
so
jp
dave.
I
think
sub
tab
works
if
you're
root,
but
I
but
sometimes
it
works.
I
wonder
if
I
have
pre
if
I've
done
sudo
and
I'm
still
in
the
time
window
when
sudo
is
still
not
requiring
a
password,
it
also
works.
I
think
that's
what
it
is,
but
I'm
not
sure
so
going
back
to
entry
point.
B
B
It
all
right,
and
so
what
I
found
this
was
something
I
actually
didn't
realize.
But
if
you
do
yum
list
dhcp.
B
Weirdly,
if
you
notice
here
so
dhcp
server,
is
in
fact
coming
from
the
rail
repo
which
sorry
excuse
me.
First,
oh
you're,
fine.
B
Yeah,
so
dhcp
server
came
from
the
rail
repo
and
what
I
found
very
confusing
about
that,
so
just
to
make
this
a
little
more
understandable,
oh
no,
I
guess
it
is
broken
correctly.
So
if
you
notice
this
enable
repo
centos
base
os
the
repo
file,
which
is
this
one
here,
which
is
fine,
that's
weird.
I
wonder
why
this
is
showing
us
all
misspelled.
B
If
you
look
at
and
if
you
notice
the
reason,
I'm
not
showing
you
both
in
the
terminal
is
because
if
you
see
an
extender
file,
it
actually
dumps
the
repo
file
at
the
end.
So
it's
not
actually
in
the
image.
But
so,
if
you
look
at
the
red
hat
repo,
no
less,
there
is
also
no
priority
here.
B
So
I
was
like
what
so
this
can
makes
me
nervous
because
if
I
don't
know
how
it
works
I'll
and
I
like
containers
because
they're
replicable,
that
makes
me
unhappy
so.
B
So
to
clearly
state
the
problem,
basically,
is
that
I
have
two
repo
files
that
theoretically,
because
I'm
not
disabling
the
rel
repos
right
and
it's
doing
what
I
want
mind
you
or
doing
what
christian
wants
in
theory
that
it's
getting
dhcp
server.
But
we
have
two
different
repo
files
that
don't
aren't
giving
any
sort
of
indication
of
which
one
will
be
used.
B
A
C
B
And
say
if
the
rel
one
is
available,
then
you
know
use
the
rel
one
if
it's
not
available,
because
I'm
not
on
a
registered
rail
box,
use
the
centos
one,
and
so
we
can
do
that
with
priorities.
Oh
wait!
A
minute!
B
Oh
no!
I'm!
Oh,
oh!
No!
I
am
I
just.
I
was
actually
looking
at
the
wrong
block
to
show
my
point.
It
just
happened
to
have
a
very
similar
name.
So.
A
Now
nlhcm
points
out,
I'm
sorry,
I
forget
your
real
name.
You
totals.
I
forget.
Both
repos
have
enabled
zero.
If
I
saw
correctly.
B
Was
looking
actually
at
the
wrong
repo?
So
where
is
it
that's.
A
I've
given
up
and
I've
realized
that
the
world
just
has
to
be
searched.
Yeah.
C
B
That's
I
don't
know
what
that
is.
What's
e4s
e4,
okay,
here
we
go
so
this
is
the
the
real
one
we
would
be
pulling
from.
It
is
enabled
it
even
has
it
has
gpg
check
the
reason
the
centos
repo
is
still
being
used,
even
though
it's
in
enabled
equals
zero.
Here
is
because
in
the
container
file
you're
explicitly
in
your
energy
before
in
the
install
statement,
so
they
are
both
enabled
you
know
at
that
line
a.
B
A
That
jp
day
did
mention
something
in
chat
not
to
interrupt
you,
but
I
am
because
this
is
kind
of
relevant.
You
have
the
rel
repo
loaded
by
default
on
your
server
that
you
are
running
your
container
building
your
container
to
so
so
that
is
loaded
first,
then
you
tell
it
to
about
the
centos
repo,
so
that
is
what
it's
second.
Is
that
the
reason
why.
B
So
so
that
seems
logical
or
it
could
also
be
that
you
know
on
a
red
hat
box,
the
rel
repos
take
priority.
B
What
I'm
worried
about,
though,
is
I
don't
know,
I
don't
know
what
the
answer
is,
and
on
top
of
that
I
don't
know
if
in
because
I
think
this
is
a
rally
3.
I
don't
know
if
in
raleigh
4
it
changes
so
iq
and
joke
asks
about
no
subscriptions
needed
for
dnf
install
so
dnf
and
yum
on
rel.
Eight
plus
are
the
same.
Essentially,
it's
yum
is
just
three
letters
that
actually
call
dnf.
B
B
So
but
it's
not
a
complete
set
of
the
real
content,
so,
for
example,
dhcp
server,
apparently
is
not
there.
Most
of
the
databases
are
not
there,
but
apache
is
there
and
php?
Is
there
for
example?
So
it's
so
you
need
a
subscription
to
access
some
of
the
rel
content
on
ubi,
sometimes
and
the
way
we
get
it
as
I've
talked
about
before.
In
fact,
a
little
bit
today
is,
if
you're
on
a
registered
rail
box,
and
you
try
to
build
that
image.
It
will
find
the
rel
repos.
B
I
don't
know
exactly
how
it
works,
but
it
does
work
and
it's
interesting
so
long
story
short.
The
first
instinct.
B
Right,
so
what
I
looked
up
prior
to
the
show
was:
what
is
the
default
priority,
which
is
99,
which
I
thought
was
very
odd.
I
was
like
why
not
around
100,
you
know
maybe.
B
Yeah
so
so
lower
is
more
important.
So
if
we
make
this
priority
150,
it
will
take
it'll
prioritize
the
rail
repo
over
the
centos
repo,
always,
and
so
that
should
take
care
of
the
problem
via
priorities
and,
let's
see,
and
then
we
should
just
be
kind
of
done.
The
thing
is,
though,
what
I
also
wasn't
in
love
with
is
that
this,
because
so
this
is
kind
of
where
it
gets
a
little
bit
arguable,
but
let's
say
and
theoretically
this
is
never
going
to
happen,
but
let's
say
the
mirror
or
whatever
this.
B
B
A
The
os
manages
right,
like
I
don't
like
breaking
defaults
in
general,
so
I
tend
to
leave.
I
don't
know
like.
If,
if
you
want
to
package
the
entire
container
up,
then
yes,
you
should,
if
you
want,
if
you
know
that
you
are
declaratively
changing
things
inside
these
configuration
files-
yes,
they
should
be
in
the
container,
but
I
don't
like
to
steer
too
far
away
from
system
defaults
either.
So
you
know
like
I
would
be
of
the
mind
to
say:
okay,
your
workaround
is
prioritized.
I
would
actually
go
speed-lunking
and
figure
out
how
it
actually
worked.
B
So
the
thing
is,
I
might
do
that
if
I
was
confident
that
that
was
part
of
the
public
api
of
dnf
right
or
yeah,
and
that
it
won't
change.
That's
that's
my
bigger
concern
because
I
I
mean
to
to
christian's
point
in
the
chat.
It
works
right
now.
What
difference
does
it
make
right
all
right
so
and
and
so
there's
some
value
to
trying
to
figure
it
out
completely.
B
However,
my
you
know
this
is.
This
is
actually
one
of
the
reasons
why
the
letters
yum
still
exist
in
rel
is
because
the
vast
majority
or
the
vast
number
of
people
who've
written
scripts
that
wrap
yum.
They
consume
its
output,
they
consume.
You
know
they
change
its
input,
et
cetera,
et
cetera,
and
so
there's
a
de
facto
api
on
yum
that
nobody
ever
promised.
B
But
people
rely
on
very,
very
heavily,
so
you
know
so
I
don't
know
whether
dnf
has
declared
somewhere
along
the
way
exactly
the
way,
the
priorities
work,
and
even
if
I
figure
out
how
it
works.
I
I
mean
it's
open
source
right.
I
could
go
and
look
at
the
code.
There
is
no
guarantee
that
that
code
won't
change,
because
it's
not
part
of
the
api
does
that
make
sense
and
why
I
wouldn't.
A
B
Yes,
sorry,
so
if
you
build
in
a
registered
box
with
packages
that
require
licenses
and
then
ship,
it
is
still
kosher
with
rh.
I
believe
that
you
end
up
with
a
supported
image
that
you
can
then,
whatever
you
call
redistribute.
B
But
I
think
the
the
part
where
I'm
not
100
clear
is,
I
don't
know
what
our
redistribution
rules
are
like.
I
assume
that
if
you
want
to
redistribute
something
you
have
to
be
a
red
hat
partner
and
that
that
redistributed
thing
has
to
be
have
a
a
check
box.
That
says
you
know
this
is
partner
approved
or
whatever
the
term
is
there's
a
certification
process,
and
I
know
we
do
this
with
container
images.
I
just
don't
know
all
the
details,
so
that
would
be
something.
B
Actually
we
could
ask
maisini.
Maybe
when
he's
on
or
or
one
of
the
other
rel
folk
well,.
A
Yeah
as
long
as
it's
on
a
rail
box
you're
fine,
it's
like
if
you're
on
a
centos
box,
then
I
think
that's
where
it
gets
gray.
B
Right
right,
I
don't
know
because
well
so
if
the
image
platform
is
an
approved
platform,
it
should
just
work.
I
I
guess
where
I
would
get
shaky
is
like,
or
it's
probably
or
I
I
assume
it's
actually
supported,
but
if
you're
really
running
a
ubi
8
container
with
a
bunch
of
stuff
installed
running
that
on
rail
7.
B
That
should
also
be
supported.
I
believe,
because
that's
kind
of
part
of
the
intent
yeah,
but
so
that's
that
would
be
something
we
can.
You
know
I
can
follow
up
with
that
offline
and
maybe
bring
it
back
to
the
next
show.
Yeah
and
yeah.
B
We
might
get
an
answer
during
the
show,
we'll
see
that
would
be
cool
so
going
back
to
the
to
the
thing
yeah.
We
don't
run
it
on
ubuntu
boxes.
What's
that
yeah,
I
used
to
use
control.
B
Right,
so
what
I
was
so
to
kind
of
go
further
back
on
the
on
the
you
know,
kind
of
my
earlier
point
of
like
I
don't
really
like
carrying
the
actual
repo
file
here,
so
I'd
rather
get
the
actual
rpm,
and
so
in
order
to
do
that,
pod
man
run.
D
B
Rm
centos
surprisingly
centos
their
default,
or
so
their
latest
tag.
So,
if
you
just
say
bear
like
this,
when
you
do
a
poll
you
get
the
equivalent
of
not
bad
this,
so
you're
actually
defaulting
to
the
latest
tag.
B
Their
latest
tag
is
not
currently
eight,
I'm
not
sure
if
that's
an
oversight
or
just
the
current
status,
oh
shoot.
I
thought
I
already
pulled
this.
B
B
One
of
them
is
the
fedora
registry.
Another
one
is
the
centos
registry
and
another
one
is
the
red
hat
registry
and
I
think
docker
is
there
too,
but
I
think
the
priority
is
docker
is
towards
the
end
and
centos
is
in
there,
but
I
don't
think
the
rel
one
is
there.
I
don't
think
unreal,
it
has
it
at
all.
Sorry,
I
see
there's
a
question.
A
B
We're
trying
to
build
a
red
hat,
supported
component,
you
know,
or
a
supported
container
image
most
of
the
time.
You
know,
because
right
now,
the
the
container
image
that
I
just
ran,
that
is
a
fully
supported
container
image.
B
I
don't
think
christian
added
anything
else,
weird
to
it,
because
it
does
not
actually
use
the
dhcp
server
from
centos
right
it.
It
pulled
it
from
rel,
as
we
verified
with
a
dnf
list
so
yeah.
So
we
can
change
the
order
of
preference
by
using
priorities.
The
thing
is,
and
we
did
that
a
little
bit
here
by
just
adding
this
priority
flag.
The
thing
is,
I
was
going
to
investigate.
B
Can
we
use
the
real
centos
repo
from
centos,
rather
than
having
a
hard-coded
file
on
disk
that
we're
using
in
our
container
image?
And
so
in
order
to
figure
that
out?
We
would
say
well
here
we'll
use
yum,
and
I
think
people
don't
realize
that
you
can
do
yum
provides
on
any
file.
It
doesn't
have
to
be
a
binary,
so
we
can
say
you
know.
Yum
provides
this
repo
file.
B
But
now
we
need
oops.
What
I
forgot
was
we
need
the
path
to
it
so
which
would
be.
B
But
we
need
to
make
it
one
match
yeah.
You
know
the
the
ubi8
that
we
have
here
right
right,
so
I
had
an
approach
to
doing
this.
A
Okay,
jp
dade,
to
answer
your
question:
scott
mcbrian
has
chimed
in
oh
says:
if
you
pull
in
a
non-ubi
package,
then
it
may
not
be
redistributed
redistributable,
there's
a
gray
area,
we're
actually
actively
working
on
it
and
making
trying
to
make
it
much
easier,
more
inclusive
and
less
gray.
A
B
B
I'm,
of
course,
looking
at
the
centos
seven.
B
Right
right,
so
you
know
we
can
use
this
link,
which
is
a
hard
link
to
the
file
we
want
and
use
the
rpm
directly.
The
thing
is,
we
have
to
keep
it
up
to
date,
but
I
don't
know
I
like
I
said
I'm
on
the
fence
of
whether
you
want
to
have
the
repo
file
locally
or
you
want
to
use
the
remote
one.
B
B
It
won't
cache
so
so
my
inclination
is
to
do
it
that
way,
and
then
we
can
totally
garbage.
D
To
at
the
oh,
my
goodness,
I
can't
type
repose.d.
B
So
we
add
the
priority
on
to
the
end,
to
make
it
a
lower
priority,
even
though
it's
a
higher
number,
then
we
can
get
rid
of
this
statement
altogether
and
this
we
can
leave
the
enabling
set
up
correctly
or
the
same
way,
and
we
could
also
combine
these
statements.
B
So
we
get
more
caching
and
then
I
also
believe
that
this
pip
install
will
also
not
cache.
I'm
not
100
sure
that
I
didn't
try
it
for
sure,
but
because
in
a
pip
install
there's,
usually
compilation,
I'm
not
sure
if
it'll
cache,
so
you
know
the
caching
stuff
is,
you
know,
always
a
little
dicey.
So
I
I
tend
to
err
on
the
side
of
safety
in
my
mind
and
then
I'm
kind
of
on
the
fence
of
whether
we
remove
the
centos
repo,
the
the
reason
that
I
don't
want
to
remove.
B
That
is
because,
if
I
do
this
is
where
you
know,
if
you
ever
do
a
yum
update
in
the
container,
which
is
generally
bad
form,
but
sometimes
necessary,
you
won't
get
any
updates
to
that
particular.
You
know
anything
that
came
from
that
repo
right.
So
so
I
thought
we'd
play
with
that
and
see
if
that
worked,
and
but
we
have
to
let's.
D
D
B
And
let's
call
it,
I
don't
know.
D
B
No
worries,
okay,
so
let's
say
dhcp
new
and
minus
f.
D
A
B
Container
image
should
be
fully
supported
as
long
as
everything
you
installed
in
that
image
came
from
a
rel
repo
and
the
fact
that
you
built
it
in
your
r
d
lab
is
moot
to
the
point
right.
A
Like
if
you're,
indeed,
if,
if
you're
using
supported
rail
instances,
you'll
be
supported
across
the
board
right
in
rel
packages
and
so
forth,
so
on
real
image,
real
packages,
real
hosts,
you're
supported
it.
The
fuzzy
area
is
when
you
take
a
fully
supported
image
right,
like
you,
built
it
on
a
rail
box
with
relic
packages,
and
you
then
run
that
image
someplace
else
on
a
non-rail
system
aka
what
happened?
What
is
the
support
structure
there?
That's
the
fuzzy
gray
area
that
we're
trying
to
figure
out
right
exactly.
B
Why
it
happened
exactly,
but
basically
the
gpg
keys
are
separated
in
a
different
rpm
and
because
I'm
basically
doing
a
bare
rpm
install
it
didn't
love
that
idea.
So
I
think
you
can
just
put
multiple
commands
on
one
line.
Oh
actually,
I
should
also
do.
B
D
B
B
You're,
good,
okay,
okay,
no
worries!
That's
why
we
do
the
peer
programming
on
the.
A
A
B
B
We'll
just
comment
this
stuff
out,
because
this
takes
a
long
time
to
do
its
thing
and
that
way
we
don't
have
to
spend
the
rest
of
our
lives
waiting
for
the
build.
We
can
put
it
back
in
once.
We
know
this
stuff
works
much
better,
that's
my
kind
of
floating
time.
So,
let's
see
if
what
I
call
it
dhcp
new
and
I
took
out
the
entry
point.
B
B
I
have
not
pushed
it
yet,
but
it
will
be
there.
Okay,
no
sorry!
I
haven't
yeah,
it's
not
it's,
not
nice
enough
along!
Well,
that's
not
entirely
true!
Actually,
so
the
stuff
we're
doing
right
now,
we're
doing
live,
but
the
repo
that
we're
basing
it
on
is
this
one:
okay,
cool,
oh
and
make
sure?
Oh
actually,
sorry,
that's.
B
Well,
the
fork
is
okay.
It's
that
it's
a
different
branch.
It's
called,
oh,
no,
that
is
the
branch
yeah.
B
B
In
the
in
the
repo
itself,
so
we
need
it
to
be
centos
base
and
so
exit
out
of
there.
B
A
B
Oh
actually,
we
should
we
could,
while
we're
waiting
for
that,
we
can
do
points.
B
B
But,
as
you
can
see,
netherlands
hackum,
I
think,
is
what
we
decided,
that
that
means
is
definitely
getting
up.
There
definitely
pulling
ahead,
and
if
you
want
to
get
your
own
points,
you
can
follow
the
link.
That's
there
and
then
put
in
that
code
yourself
or
the
second
link
is
basically
a
deep
link
with
the
code
already
injected.
But
noah
friction
narendev.
B
B
They
are
all
on
a
youtube
playlist
if
you
are
so
inclined,
even
in
order.
No
less,
that's
a
that's
a
great
image
and
I
I'm
afraid
so
what
we're
gonna
do
next
time.
I
almost
did
it
in
time
for
this
time,
but
not
quite
is.
I
also
want
to
highlight
going
forward
the
new
people
who
have
submitted
for
internet
points,
so
they
may
have
a
lower
number,
but
right.
B
The
fact
that
they
just
submitted
some
points
recently,
and
I
was
going
to
do
it
today,
but
I
was
looking
at
it
and
the
data
seemed
weird,
so
it
didn't.
I
didn't
want
to
post
it
with
it
being
fair.
A
B
And
I
didn't
have
enough
time
to
really
dig
out
what
was
going
on.
So
congratulations,
everybody
so
excited
to
see.
People
are
still
collecting
our
sweet.
Sweet
internet
points,
keep
in
mind.
Netherlands
hackham
did
a
pr
and
that's
what
pulled
them
ahead
of
everyone.
D
A
B
That's
how
you
that's,
how
you
can
improve
your
points.
You
can
also
join
the
discord
server
and
get
some
points.
Excuse
me
and
get
some
points
that
way
you
can.
You
know,
file
an
issue,
you
can
file
a
pr,
you
can
go
back
to
past
episodes
and
that's
all
there
is
to
it
and
it
looks
like
our
build
is
done.
B
Hopefully
we
didn't
get
any
errors
that
time
yes,
so
that
looks
better.
So
we
got
an
actual
signing
key
installed
and
we
didn't
get
an
error
about
the
priorities.
Now,
I'm
curious
if,
like
which
version
of
the
rpm
we
got
so
we
can
say
dn,
I'm
really
mixing
and
matching
today,
dnf
list.
B
I
mean
they
really
are.
They
really
are
pointing
at
the
same
thing.
It's
just
I
try
like
I,
I
think
I've
explained
on
the
show
before,
but
I
try
to
use
yum
whenever
on
rel,
because
I'm
looking
for
bugs
and
just
in
case
there's
something
that
doesn't
work
right,
but
I
think
it
just
points
at
the
same
thing
and
I'm
not
sure
I
know
like
the
yum
like.
B
So
as
expected,
or
as
hoped,
we're
still
getting
the
rel
version
of
dhcp
server,
we
you
know
we're
using
the
official
repo
file
which,
like
I
said,
arguable
basically
pulling
anything
from
the
internet
and
putting
it
inside
your
container
means
that
it
has
potential
of
not
being
as
replicable
right,
but
we
will
get
updates
if,
for
some
reason
we
do
an
update
in
here.
So
the
last
thing
I
kind
of
wanted
to
show
was
what
happens
if
we
unregister
it?
Does
it
still
work?
B
So
hopefully
this
won't
take
too
long
see
it
still
doesn't
tab
complete
for
me
unless
I
want
it
to
or
don't
want
it
to.
I
think
it's
just
on
register
right.
This
is
not
a
common.
B
So
in
theory
it's
when
I
make
sure
okay
cool,
so
I
should
still
have
all
the
repos,
but
I
should
not
have
access
to
them.
So
if
I
do
oh
boy,
if
I
do
our
build
again
see,
I
was
curious
about
that.
So
it
still
used
the
cash,
so
I'm
still
getting
real
repos
or
real.
B
Right
packages
because
I'm
still
cached
so
let's
say
podman
images.
Oh,
what
did
I
call
thcp
new,
so
I'm
just
going
to
kill
the
whole
image
just
clean
it.
Theoretically,
I
could
just
kill
the
one
line,
but
I
don't
think
it's
gonna
actually
save
us
anything.
A
I'm
assuming
it
purges
every
bit
of
cash
out.
There
should.
B
D
Things
up
not
without
pseudo.
B
D
B
So
no,
I
didn't
make
any
changes
to
the
container
itself
or
the
container
file
itself.
All
I
changed
was
the
host
that's
doing
the
building,
so
it's
trying
to
build
it
pulling
dhcp
server
and
it's
going
to
say
you
know:
where
can
I
get
the
ecp
server
from
because
I'm
not
registered?
B
B
Oh
well,
that's
that's
clear
as
a
bell
they're
both
called
base
os.
So
I
believe,
let's
just
see.
B
Oh
well,
let's
just
do
it
this
way,
I
think
there's
a
verbose
that
would
actually
tell
us,
but-
and
so
if
we
look
at
bass
repo,
oh
I
don't
have
less
yeah
see
the
name
of
this
repo
is
base
os,
so
that
not
very
useful.
B
B
So
this
is
the
red
hat
universal
base
image
eight
base
os
versus
when
we
were
building
before
we
were
using
the
red
hat
enterprise,
linux
8
base,
os
and-
and
so
you
kind
of
see,
even
in
the
output
of
the
build
you'll
see
that
you
have
switched
repos
so
netherlands
hakum
did
that
answer
a
question.
A
We
shall
see,
but
I'm
reading
what
jp
said.
B
Yeah
I
was
too
yeah
getting
the
registrations
to
work
yeah
so.
B
You
know
you
don't
get
it
as
non-root,
but
that's
that's
expected
in
a
sense
yeah.
I
mean
it's
significantly
better.
Now
I
think
it's
a
lot
less
confusing
as
well,
because
I
mean
I
normally
like
in
non-kind
of
production
environments.
Right
I
mean
I
normally
can
just
say:
subscription
manager
register
dash,
auto
attached.
B
It
just
works
and
remember
actually
one
thing
I'd
point
out
for
anybody
who's,
not
a
regular
watcher
of
the
show
is
if
you
need
a
rel
subscription
for
any
reason,
you
can
get
free
ones
using
the
developer
subscription,
which
do
you
have
a
a
cool
hashtag.
I.
B
Okay,
but
go
to
developer.redhat.com,
and
it
should
be
yeah
very.
A
B
But
yeah,
oh.
B
So
what
is
the
package
difference
between
morel
base
os
and
ubi
base
os?
I
don't
know
off
the
top
my
head,
I
do
know
it's
documented
I'd
have
to
like
dig
for
it
to
know
where
it
is
exactly
but
kind
of
my
rule
of
thumb
in
the
back
of
my
mind,
is
kind
of
anything
related
to
front
end,
not
front
end
but
sort
of
front
end.
Web
development,
so
programming
languages
web
servers
that
kind
of
stuff
those
are
all
in
ubi.
B
A
A
B
So
that
was
pretty
much
what
I
wanted
to
show
you
know.
I
I'm
not
sure
I
think
something
I
read
somewhere
says
that
the
reason
christian
is
using
the
pip
install
of
ansible
is
because
he's
relying
on
some
very
current
feature
of
ansible.
That's
not
in
the
sense
the
rel.
B
Version
of
ansible
so
yeah,
so
otherwise
I
would
definitely
remove
that
because,
basically
pip
installs
again
are
you
know,
you're
pulling
something
from
the
internet
that
may
have
changed
since
you
last
touched
it
plus,
like
I
said
they
break
caching,
a
lot
of
the
time
because
they
recompile
and
all
that
jazz,
not
that
there's
anything
inherently
bad
for
it
per
se,
as
much
as
it
is
a
container
image
is
something
I
want
replicable
and
I
want
it
to
be
fast
on
rebuilt.
B
So
otherwise
you
know
you
can
investigate
the
rest
of
the
thing
for
yourself.
I
will
talk
over
with
christian
whether
he
wants
something.
That's
like
a
pr
based
on
on
these
changes,
but
these
are.
This
is
closer
to
how
I
would
have
done
the
container
image,
and
I
thought
it
was
a
good
exercise
for
the
show
to
kind
of
see
the
difference
in
approaches
and
what
the
trade-offs
are
you're
making
when
you're
taking
different
approaches.
B
Yes,
exactly
exactly,
maybe
he
should
get
points
for
being
the.
A
B
I
said
I
just
thought
it
was
a
good
exercise.
I
thought
it
was
an
interesting
comparison
of
two
different
approaches
to
the
same
problem
as
well
as
how
to
solve
the
problem
of
I
have
ubi
and
it
will
be
used
sometimes
by
people
on
rel
registered
machines
and
sometimes
not
right
and
and
but
that's,
I
think,
our
show
unless
anybody
had
any
more
questions.
But
I
don't
see
any
more.
A
I
don't
see
any
either.
I
hope
everybody
got
the
link
for
the
points
and
coming
up
next
on
the
channel
at
oh,
I'm
on
the
wrong
day.
Let's
get
back
to
the
right
day
here.
Well,.
A
Up
next
on
the
channel
at
11am
eastern
is
the
openshift
administrator's
office
hours
with
sully,
and
I
and
bring
your
openshift
questions.
We
will
bring
answers
so
right
right.
We
will
hunt
for
them
and
give
them
to
you
if
you
don't
know
them
off
the
top
of
our
head
right,
but
yeah.
So
thank
you.
Everyone
for
joining
stick
around.
You
know
subscribe
to
the
calendar
if
you
haven't
yet
I
will
drop
that
link
in
chat
here.
A
It's
just
red
dot,
ht
stream,
cal
and
you
know,
click
the
little
plus
sign
on
the
bottom
right
of
the
calendar
to
add
it
to
your
own.
That
way,
you
know
when
we're
on
the
air
and
what
we'll
be
talking
about,
and
until
next
week,
where
we'll
be
on
the
redhead
youtube
as
well,
we'll
catch
you
then
and
stay
safe
out.
There
don't
watch.
A
Well,
I
mean
yeah,
there's
that
yeah
you
could
watch
us
everywhere
and
just
pick
which
one
you
want
to
listen
to
and
mute
the
others.