►
From YouTube: Demo of the GitLab Runner Operator for Kubernetes
Description
In this video we provide an overview of how to install GitLab Runner on a Kubernetes cluster using the GitLab Runner Operator available on OperatorHub.io
A
Perfect
so
hi
there.
So
today
we
are
going
to
talk
about
to
have
a
little
demo
on
how
to
install
the
gitlab
runner
operator
on
a
kubernetes
cluster.
So
I
kind
of
have
two
slides
to
kind
of
help
us
when
we
will
be
going
through.
So
what
we
will
cover
today
is
how
to
install
the
operator
on
the
kubernetes
cluster,
the
vanilla
one.
A
Then
we
will
talk
about
how
to
customize
the
configuration
when
installing
the
operator
and
in
the
example
I
will
use,
I
will
be
setting
it
in
with
termination
grace
period.
Second,
so
we
can
see
how
it
it's
taking
in
consideration
on
the
runner,
which
is
then
created,
and
I
also
add
some
useful
links
which
are
for
how
to.
I
will
be
myself
following
during
this
demo,
and
I
want
to
thank
georgie
because
he's
the
one
who
makes
sure
that
the
how-to
is
is
available,
and
I
believe
it's
really
well
written.
A
So
if
we
go
back,
the
first
link
we
will
be
using
is
on
doc,
gitlab.com
runner,
install
operator
html
install
on
kubernetes
during
the
demo,
I'm
using
a
gta
cluster
set
on
gcp.
So
I
made
sure
to
have
everything
already
set
up,
so
we
won't
have
to
wait
the
creation
and
all
those
things
while
we
are
presenting,
so
the
prerequisite
is
kubernetes
version
1.21
and
the
certification
manager.
A
Why
do
we
need
the
certification
manager
because
it
actually
helped
to
handle
all
the
things
related
to
the
certificate
the
operator
will
use
to
create
the
manager
and
the
runner
without
it,
the
installation
just
failed.
A
So
if
it's
really
mandatory
to
have
it
already
installed,
when
we
open
the
operator
hub
audio
website,
there
is
a
documentation
already
as
this
thing
here,
which
we
we
say
exactly
what
I
was
saying
about
the
prerequisites-
and
we
have
a
simple
comment
to
install
this
specific
version
of
the
certificate
manager
during
the
demo,
I'm
using
two
terminals,
one
to
have
to
tap
my
comment
and
the
second
one
will
be
mostly
a
key
line:
s
to
see
the
results
on
the
cluster,
so
we
can
see
exactly
what
is
happening
so
as
it's
written
in
the
documentation.
A
Okay,
just
in
case
the
user
has
a
script
to
have
all
those
things
automate.
It's
really
mandatory
to
make
sure
that
everything
is
installed
of
operator
just
won't
work.
A
Everything
need
to
be
installed,
but
it
just
won't
work.
So
that
was
the
first
step
to
install
the
certificate
manager
and
then,
as
it's
written
here
in
the
documentation,
we
will
click
on.
We
will
select
install
still
on
the
operator
website
and
follow
the
instructions
to
install
the
olm
and
the
operator.
A
A
A
I
can
be
wrong,
so
don't
take
it
100
for
for
true,
but
the
workbook
is
mostly
used
when,
during
the
installation
of
the
operator,
the
user
explicitly
enables
the
whiteboard
the
operator
can
be
installed
on
two
modes
with
a
web
fork
enable
of
the
weapon
not
enabled
in
our
case
I
I
feel
I
think,
by
default,
it's
visible,
so
I
don't
think
this
one
is
actually
used,
but
the
certificate
injector
is
used
instead
to
create
the
certificate
and
make
them
useful.
A
Yeah
the
olm
for
this
one,
I'm
not
100
sure,
because
during
my
test
I
I
didn't
really
interact
a
lot
with
it,
but
for
what
I
the
main,
the
main
advantage
I
found
with
it.
So
I
will
just
talk
about
this.
One
is
with
it.
It
allow
us
to
have
operator
installed
on
a
namespace
and
being
able
to
handle
a
runner
and
stands
on
another
another
namespace.
That
was
personally
the
main
advantage.
A
A
You
cannot
do
anything
else
outside
of
this
namespace
and
the
operator
lifecycle
manager
actually
have
during
the
the
implementation
of
during
the
support
of
your
operator
on
the
vanilla
cluster,
it
actually
helped
us
to
be
able
to
handle
the
ability
to
have
a
manager
and
staff
and
still
be
able
to
launch
a
runner
instead,
let's
say
on
the
default
namespace
without
without
having
the
process.
Failing
of
the
time.
A
My
pleasure
so
then,
just
now
we
install
the
operator,
lifecycle
manager
and
now
we
will
go
ahead
and
just
install
the
runner
operator
itself.
So
I
go
back
to
the
terminal.
A
Okay,
I
don't
know
if
you
can
read
during
the
installation
it's
kind
of
it
kind
of
checked
the
the
phase
so
once
it
was
filled
and
then
it
succeeded,
but
basically
it's
supposed
to
work.
It
might
take
a
bit
longer
depending
on
the
cluster
and
the
resources
which
are
available.
But
it's
supposed
to
work
so
now
I
will
install
the
operator
itself.
A
A
And
you
can
see
the
operator
the
version
and
everything
associated.
So
basically,
what
we
did
right
now
is
using
the
documentation
available
on
gitlab,
gitlab.gitlab
docs
and
the
operator
up.
We
were
able
to
install
all
the
to
install
all
the
prerequisites
and
the
operator
itself
so
now.
The
second
step
will
be
to
actually
create
an
instance
of
runner,
so
it
will
be
able
to
pick
up
job
and
to
do
everything
that
should
be
done
as
it
was
written
in
the
documentation
to
be
able
to
do
that.
A
We
need
a
runner
registration
token,
because
we
will
have
to
create
a
secret
which
will
be
used
to
register
to
the
runner,
to
the
to
the
git
lab
the
club.com
of
the
self-managed
instance
of
gitlab,
and
we
will
also
need
a
yaml
to
create
the
runner
itself.
A
As
I
was
saying
here,
just
using
those
two
examples
may
leave
some
user
confused
because
by
default
we
just
said
the
gitlab
url.
We
we
run
a
secret
runner,
the
registration
token
and
then,
if
a
user
wants
to
do
something
particular
it
doesn't
really
happen
to
know
exactly
how
to
do
so.
In
addition
to
those
five,
I
will
also
add
a
config
map
which
is
actually
used
to
help
customize
the
configuration.
A
A
The
link
is
also
on
the
slide.
I
was
sharing
at
the
beginning,
it's
kind
of
described
the
crd,
the
custom
resource
definition.
If
I'm
correct,
so
it's
actually
help
us
to
know
what
are
the
properties
which
can
be
set
when
creating
an
instance
of
runner.
So
we
can
see
we
can
set
the
azure
for
everything,
cache
related
on
azure,
the
build
image,
the
certification,
the
clone
you
ahead
and
so
on
and
so
forth.
A
A
A
So
basically,
what
I've
done
here
is
I
can
I
create
another
config
command
where
let
me
see
if
I
can
zoom
somehow
yeah,
can
you
read
what
I
like?
Okay.
B
A
Yeah,
if
I
wanted
to
set
another
another
property,
another
parameter,
I
would
have
just
copied
recipe
cpu
limit
and
copy
it
and
paste
it
in
the
config
map,
configmap
yammer,
but
for
our
example,
it's
the
only
one
I'm
setting
right
now,
then
we
have
runner
secret,
which
is
actually
saving
my
my
registration
token
and
the
last
one
will
be
the
gitlab
runner
itself,
where
we
actually
create
the
instance.
So
what
I
do
here
is
I
set
the
gitlab
url
to
gitlab.com
for
build
image.
A
I
will
be
using
alpine
for
the
token
I'm
telling
him
to
take
it
from
the
from
the
secret
which
is
created
too,
and
which
has
this
name
for
the
configuration
configuration
terminal.
I'm
asking
it
to
use
the
config
map
which
is
created
and
which
has
this
name,
gitlab,
runner,
configtomo
and
some
tag
and
some
environment
variable.
A
And
we
have
everything
running
here,
so
we
can
go
and
ssh
on
the
computer
and
check
if
the
configuration
has
actually
been
updated-
and
we
can
see
here
that
I
have
my
termination
price
period
set
to
five
by
default
is
three
thousand
six
hundred
an
hour.
So
now
we
can
see
that
it
has
been
set
to
five
and
yeah
everything
else.
A
A
A
B
A
A
We
then
set
a
determination
break
area
to
five
seconds
using
a
conflict
map
where
the
conflict
terminal
has
been
stored,
and
then
we
also
made
use
of
of
crt
to
know
exactly
which
parameter
we
are
able
to
set
in
our
use
case.
So
I
don't
know
if
you
have
some
question,
I
can
try
to
answer
based
on
what
I
know.
B
Actually,
romeo,
that
was
quite
clear
and
and
very
succinct.
I
just
wanted
to
clarify
one
thing:
there's
lots
of
documentation
or
lots
of
references
to
open
shift,
but
this
is
about
running
our
operator
on
a
vanilla
kubernetes.
B
B
So
maybe
we
we
need
to
put
that
just
to
the
the
top
level.
A
B
Yeah
yeah
that
way
end
users
aren't
confused.
If
you
look
at
the
at
the
beginning
of
it
in
the
description,
it
does
say
in
kubernetes
or
open
shift
clusters,
but
yes
yeah.
I
just
want
to
make
sure
folks
are
not
confused
and
so
the
other
thing
capability
level.
I
know
this
autopilot
is
not
checked,
but
that's
talking
about
basically
kubernetes
level,
auto
scaling
right
and
we're.
A
B
Care
of
auto
scaling
within
the
the
runner
executor
itself
is
that
right.
A
B
Yeah,
so
I'm
assuming
as
long
as
we
are
going
to
manage
that
ourselves
within
the
kubernetes
executor
that
we
won't.
We
won't
ever
see
this
autopilot
capability
level
checked
within
the
within
operator
hub,
but
that
might
change
in
the
future
right.
B
Yeah
well,
this
was
this
was
excellent.
Did
you
share
the
the
slides
in
the
meeting
invite.
A
A
Yeah
perfect,
we
added
also
here
in
the
meeting
and
and
yeah
I
mean
I
will
make
sure
to
have
it
available
for
everyone
who
actually
wants
to
access.
B
Nice
very
nice,
and
I
I
am
now
convinced
to
go
ahead
and
install
k9s.
I
I've
looked
at
it
before,
but
I've
I've
never
I've
never
played
with
it.
So
that
looked
honestly.
A
I
I
find
it
useful
because
me
too,
at
the
beginning,
I
wasn't
really
sure
if
I
needed
and
let's
say
I
want
to
see
the
secret
instead
of
having
to
pack
or
I
just
have
them
here,
but
for
sure
there's
too
many
of
them.
If
I
want
to
see
in
the
specific
namespace
I
just
hit
one
and
it
is
supposed
to
filter
yeah.
So
I
can
see
my
secret
here
and
same
for
configmatic.
I
just
have
to
type
complete
map.