►
From YouTube: Distribution Team Demo - Dec 10, 2020
Description
The Distribution team demos using gitlab pages with a in-progress MR deployed within the gitlab-helm chart. With self-signed certificates.
A
Welcome
everyone
to
the
demonstration
for
this
approximately
thursday
december
10
2020,
basically
on
your
time
zone
and
today
we're
going
to
be
demonstrating
the
new
gitlab
charts
feature,
which
is
a
merge
request
that
has
been
submitted
by
balu.
It
is
if
you're
wanting
to
follow
along
at
home.
A
It
is
in
the
charge,
repository
merge,
request,
one
six,
seven,
seven
and
just
for
giggles
paulo-
and
I
decided
I
just
do
it
live
here
for
the
first
time
is
the
best
way
to
do
a
demo
to
see
if
something
is
straightforward
is
to
just
see
how
bad
it,
how
bad
you
trip
trying
to
look
at
it
for
the
very
first
time,
so
I'm
gonna
go
ahead
and
share
my
screen
all
right.
A
A
So
this
is
my
basic
self
sign
and
we're
going
to
install
this
without
pages
support,
and
while
that
runs
so,
let's
just
say,
I'm
a
customer
and
what
I'm
gonna
assume
is.
I
have
no
idea,
I'm
using
the
charts,
because
someone
said
kubernetes
kind
of
like
the
the
good
old
cartoon
right,
the
dilbert
cartoon,
so
once
they're
too
ready,
so
I'm
gonna
try
it.
So,
according
to
this
documentation,
I
need
to
bring
a
wildcard
certificate
and
I
saw
that
in.
A
A
A
A
A
A
B
A
A
E
B
A
A
A
D
A
D
A
A
E
A
A
A
A
A
E
A
Yes,
I
do,
I
did
it
the
first
time
I
typed
it
in
the
wrong
path,
brains,
all
right,
so
that
running
you
now
see
the
pods
are
running
so
just
to
recap,
where
we're
at
we
have
modified
and
adapted
the
generate
certificates,
wild
card
service
from
what
I
know
to
be
from
the
getaway
service,
but
what
someone
would
use
to
if
they
were
just
coming
into
this
new.
We
have
built
our
wild
card,
self-signed
secret
and
added
it
to
kubernetes,
and
now
we
are
trying
to
install
pages.
A
Seriously
doubt
there
are
any
actual
migration
training,
but
yep.
E
A
I'm
trying
to
go
just
back
to
the
main
page.
I
haven't
even
set
up
a
page
of
repo
or
checkout,
yet
I
understand
you
sent.
A
A
A
D
E
E
A
A
That
is
a
deep
should.
I
saw
something
similar
the
other
night
that
I
noted
when
testing
the
upgraded
engine
and
ingress
going
from
version
0.20
to
0.41.2,
where,
if
there's
a
deep,
if
there's
enough
of
the
change,
all
the
pause
for
the
web
service
will
just
stop
responding
as
opposed
to
rolling
over
like
I
would
have
expected.
E
E
E
E
Do
you
mean
dns
on
the
cluster?
No,
the
clusters,
dns
doesn't
matter.
F
I'm
looking
up
at
the
google
dashboard,
it
does
show
that
the
dns
got
updated.
So
it's
looking
good
on
the
google
side.
Okay,.
B
E
A
D
A
A
A
B
D
G
That
pod
is
the
task
runner,
not
the
gateway
burner.
Did
you
turn
off
the
runners
in
your
yammel
things,
your
value's.
G
A
A
A
G
G
E
A
E
E
G
A
A
B
E
E
E
E
Okay,
my
apologies.
We
need
to
check
the
configuration
inside
the
pod.
E
E
Okay
should
be
on
port
8090.
Http
2
is
yes.
I
do
not
see
the
configuration
that
explicitly
sets
https
because
pages
isn't
serving
https
itself
and
it
is
trying
to
hit
the
external
api
which
is
good.
E
E
H
H
F
Hey
robert,
when
you
get
a
moment,
try
to
dig
that
host
name,
you
tried
to
go
to
I'm
showing
in
the
google's
dns
dashboard
that
the
ip
is
still
something
bizarre.
Like
previous
installation.
A
So
let
me
pull
that
up
over
here
is
showing
this
on
the
light
on
the
screen
share.
Is
that
going
to
be
a.
A
So
because
I
use
our
cloud
native
domain
for
pretty
much
every
demo
and
every
test
will
be.
E
External
dns
should
do
a
configuration
cycle
which
should
regenerate
that.
However,
this
ui
does
not
auto
update
okay.
E
F
All
right,
let's
see
it,
did
get
updated,
has
a
text
record
so
externally
and
that's
created
correctly.
It
looks
like.
A
B
A
B
E
A
E
E
F
F
A
I
I
also
have
it
up
here
too:
I'm
gonna
stop
my
screen
share
and
let
somebody
screen
share.
First,
I
could
just
prove
to
the
posterity
of
the
video
that,
yes,
we
can
see
this.
A
A
So
I
think
that
overall
I'll
declare
this
in
working
most
of
the
issues
encountered
were
just
typical,
systematic
I'll
think
on
whether
or
not
that
would
be
something
that
we
could
translate
to
think
about,
whether
to
translate
back
to
things
that
we
can
document.
That
would
be
a
follower
to
this,
not
part
of
this
overall,
though
quite
successful
stuff
went
in.
Is
there
anything
else
you
want
to
kick
the
tires
on?
E
The
demo
at
hour,
I
would
like
to
know
about
an
application
feature
flag
previously,
now
it's
possible
that
recent
versions
of
pages
and
even
master
on
rails
have
changed
this,
but
for
the
last
month
you
actually
had
to
set
one
or
even
two
application
feature
flags
for
pages
to
serve
directly
from
artifact
storage.
E
A
A
A
A
E
A
E
E
C
E
Well,
that
fired
up
nice
and
quick
this
time
around.
I
am
not
actually
seeing
a
nice
way
to
get
all
of
the
feature
flags.
E
Yeah
you
get
them
one
by
one
using
feature:
capital
f,
feature,
dot,
enabled
question
mark
and
then
the
string
which
I
will
have
to
fetch.
You
said.
E
E
That's
enough,
actually,
interestingly,
it
looks
like
they
may
have
automatically
flipped
this.
E
E
Well,
that
for
giggles,
let's
increase
the
scale
of
the
gitlab
page's
deployment,
does
it
have
an
hpa?
E
E
E
No,
not
three
hundred
watch
the
cluster
die.
Okay,.
E
F
And
robert,
the
ttl
on
those
records
are
300
seconds.
So
by
now
you
should
be
able
to
get
there.
I
would.
E
Think
yeah
pro
tip
for
developers,
we
should
probably
document
how
to
set
the
ttls
with
external
dns,
so
that
they're,
not
five
minutes.
E
E
E
E
E
E
Why
does
this
matter?
You
might
ask
architecturally
speaking.
The
way
pages
used
to
work
is
that
you
would
actually
write
this
into
the
file
system
somewhere
and
then
pages
would
serve
from
the
file
system.
Hence
pages
couldn't
work
in
cloud
native
due
to
the
nfs
requirement
when
they
moved
to
doing
this
through
artifact
storage,
they
initially
took
it
and
instead
of
literally
taking
the
output,
bundle
sticking
in
artifacts
and
then
sidekicks
comes
along.
It
pulls
the
artifact
down,
extracts
it
and
writes
it
to
the
file
system.
A
E
E
G
E
E
E
It
doesn't,
we
don't
have
a
way
to
load
extra
tls
certificates
into
it
and
it
would
actually
turn
around
and
go.
I
can't
serve
https
on
this
because
I
don't
have
a
certificate
for
it
and
it
would
go
back
to
its
default
one
and
then
you'd
get
a
complaint
like
self-signed,
nginx,
ingress
or
kubernetes
ingress
doesn't
match
the
hostname,
so
we
actually
have
to
do
that
in
the
next
stages
to
actually
make
it
possible
to
serve
https.
E
E
A
So
I'm
fairly
convinced
that
it
works
so
I'll.
Take
a
look.
I
guess
the
next
step
for
me
is
the
reviewers
to
see
if
there's
any
naming
style,
other
things
from
our
decision
records
that
I
would
quibble
with
and
then
yeah.