►
A
Hello,
everyone
and
welcome
to
the
distribution
team
demo.
For
today
it
is
june
3rd
of
2021,
and
today
this
is
the
first
kind
of
public
show
of
what
we're
doing
with
the
smoke
testing
of
source
installs.
So
just
a
little
bit
of
a
background
we
had
about
one
or
two
cycles
ago.
A
We
had
a
request
come
in
for
an
update
to
do
system
d
based
in
it
unit
files
instead
of
the
init
files
you
provide,
and
there
was
a
lot
of
trouble
in
testing
that,
because
you
really
don't
do
a
lot
with
source
distillation
or
the
documentation
when
I
started
working
on
that,
I
realized
something
as
of
that
period,
which
is
during
1311
cycle
giddily
and
for
one
of
our
other
pieces
did
not
build
from
source.
A
So
now
it
was
making
it
through
in
our
builds.
But
if
you
were
following
the
source
install
it
wouldn't
work,
and
we've
seen
this
before,
where
folks
who
are
using
source,
install
run
into
these
weird
issues
that
don't
work,
we
don't
really
have
a
infrastructure
thing
in
place
to
test
it.
So
this
is
kind
of
where
this
is
going
is
to
provide
a
way
to
just
smoke
test.
Do
our
install
instructions,
work
and
then
also?
A
Hopefully,
this
will
flush
out
things
that
may
pop
up
for
us
things
like
when
we
updated
to
a
new
version
of
go,
gitly
didn't
compile,
and
we
didn't
know
that
until
late
in
the
cycle-
and
we
had
to
back
it
off
right
at
the
end
like
the
day
before
release,
so
we're
trying
to
flush
out
problems
like
that
early.
So
this
has
kind
of
a
couple
of
things.
It's
used
to
be
used
for
now
to
do
this,
we're
trying
to
keep
as
simple
as
possible
for
a
proof
of
concept.
A
There
are
a
couple
of
tenants
that
I
have
tried
to
follow
with
this
first
separation
of
concerns.
Okay,
I
want
the
installation
to
be
something
that
can
run
in
the
pipelines
it's
going
to
run
in
a
container,
but
it
can
also
I'm
going
to
show
you
how
it
also
runs
out
against
a
local
virtual
machine.
So
this
will
allow
people
to
just
download
just
the
script
part
of
this
and
test
it
against
anything.
A
They
want
to
target
and
then
in
our
pipelines,
we'll
have
a
basic
check,
and
the
idea
is
that
this
first
there's
basically
there's
going
to
be
three
steps
to
this.
The
first
step.
What
I
have
done
today
in
the
pipeline
setup,
that
is,
I
can
set
up
two
containers
and
target
and
my
ansible
script
can
talk
to
it.
That's
it.
The
the
second
part
of
this
demo
will
be
showing
what
it
looks
like.
A
Actually,
I've
already
run
it
because
it
takes
about
15-20
minutes
to
compile
and
build
it
all,
but
I
have
a
machine
installed
locally
that
has
gitlab
running
from
source
and
then,
after
that,
after
all
this
is
proved
out.
A
The
next
thing
we're
going
to
do
is
I'm
going
to
work
with
with
nelia
and
I'm
going
to
have
her
help
me
put
qa
against
this,
so
we
can
run
qa
against
this
as
well
just
to
make
sure
that
our
source
install
instructions
all
match
up
right,
but
the
first
thing
we
need
to
do
is
make
sure
that
this
actually
works.
So
this
is
in
the
gitlab
source
installer.
I
will
put
this
into
the
document.
A
Maybe
I
will
find
that
document
the
notes
somewhere,
I
will
not,
so
I
will
put
it
into
the
zoom
chat
and
I'll
add
it
afterwards.
But
this
is
issue
one
and
what
we're
doing
here
is
put
it
in
the
slack
chat,
because
I
can
find
that
quickly.
A
There
we
go
so
what
this
is
going
to
do
is
this
is
going
to
spin
up
containers
and
you
can
see
here
in
the
pipelines.
This
is
from
last
night
and
there
are
a
couple
things
that
happen
for
one
it
will
pull
on
an
upstream
and
install
targets.
That's
all
it's
going
to
do
and
then
the
other
thing
that
this
does
is.
A
It
runs
a
test
connection,
so
this
will
only
run
this
is
building
the
controller
and
the
target,
and
they
only
run
if
I
change
those
images
to
both
separate
container
images-
and
this
is
the
connect
test
connection
run
now,
you
can
see.
All
I
did
was
an
ansible
paint
module.
Just
can
I
talk
to
it,
but
what
I've
done
is
the
instead
of
running
the
test
on
the
the
source
install.
A
What
I
have
is
the
pipeline
is
running
with
the
ansible
controller
as
the
main
and
what
I've
done
is.
I
have
actually
set
up
the
service
or
set
up
the
the
target
installation,
the
installation
target
as
a
service.
So
if
we
look
at
the
ci
file-
let's
see
here
so
we've
actually
set
this
up
as
a
service
right
and
what
this
does
is.
This
means
that
I
can
run
the
install
target
in
the
background
they're
on
the
same
item.
A
I
can
provide
it
dns
so
now,
I'm
not
messing
with
dns
and
setting
up
a
terraform
virtual
machine
shout
out
and
props
to
to
jason
jason
plum,
and
I
had
a
very
long
discussion
slack
about
the
pros
and
cons
of
doing
this
with
terraforming
up
a
virtual
machine
versus
doing
this
in
containers.
So
this
is
kind
of
a
proven
concept
of
this
as
well,
so
we're
flipping
the
script
right.
A
What
we're
testing
is
the
service
and
then
what
we're
running
is
the
controller,
and
so
this
basically
is
going
to
set
all
this
up,
install
public
key
and
run
this
test
connection.
Now
that's
what
this
first,
mr,
is
to
do
just
to
get
the
kind
of
framework
together
and
say
we
agree.
This
is
the
way
to
get
the
machine
up
and
I
can
talk
to
it,
but
obviously
it
doesn't
install
git
lab.
So
the
next
part
that
will
come
after
we
prove
this
out
in
the
pipeline
is
the
ansible
script.
A
A
A
It
does
all
the
last
bits-
and
this
goes
all
the
way
as
far
down
as
to
it-
creates
workhorse
and
the
sockets
the
last
bits
that
I
do
not
have
in
the
ansible
script.
I
have
in
this
real
quick,
so
I
was
trying
to
get
them
to
work.
This
was
a
thing
at
the
end
of
last
week.
I
was
just
trying
to
quickly
get
this
in
to
run,
and
all
that
runs
now.
A
On
there
we
go
so
this
is
my
local
instance.
This
is
not.
I
just
went
from
the
internet.
This
is
just
my
local
network,
but
I
can
do
anything
you
want
here.
So
this
proves
out
that,
yes,
I
have
a
source
install
when
I
first
did
this
work,
I
had
to
add
a
lot
of
manual
patches
for
1311..
The
hope
is
that,
after
we
do
this
we'll
be
able
to
catch
those
not
release
source
with
the
need
to
have
folks
go
in
and
manually
know
how
to
like.
A
It
was
things
like
patching
the
make
file
to
get
the
right
version
of
big
commands
because
they
changed
which
fan
line,
arguments
they're
giving
get
or
they
would
change
things
to
the
make
command
because
they
had
a
it
was
in
a
halfway
state.
So,
let's
see,
let's
do
my
usual
project.
A
A
I
can
connect
with
ansible
to
that
machine,
so
the
next
step
is
once
this
framework
is
approved
and
the
metro
has
already
been
looking
at
it.
Merge
that
take
the
ansible
bit
put
those
last
like
five
or
six
commands
on
merge
that
on
top
we'll
have
a
smoke
test
and
then
from
there
I
think
nellie
and
I
will
partner
up
to
put
some
qa
stuff
up
against
it
and
make
sure
we
can
bump
on
it
with
the
rest
of
our
pipelines.
A
But
this
will
give
us
a
base
smoke
test
for
that
install
and
then
the
next
step
after
that
is
we're
going
to
have
to
talk
to
doc's
team
and
figure
out
how
to
keep
this.
Here's
the
documentation,
here's
a
smoke
test.
How
do
we
keep
that
aligned
right?
Because
we
don't
want
that?
We
want
this
to
be
a
low
key.
B
B
No
in
the
in
the
com,
like
in
the
instance,
which
shows
what
version
is
running.
A
A
I
said
the
the
ansible
playbook
will
be:
playbook
will
be
up
for
review
after
we
merge
the
other
piece
just
because
I
want
to
get.
I
wanted
to
get
the
proof
of
how
we're
gonna
do
it
in
pipelines
with
vm's
done
first.
So
then
I
want
to
keep
the
mr
as
small
as
possible
so
that
whoever
you
know
this
is
kind
of
a
it's
birthing.
Mr,
so
there's
a
lot
there's
a
lot
in
there
already
to
just
stand
things
up.
A
I
wanted
to
keep
all
the
extra
answer
bits
out
so
that
we
proved
out
these
machines
can
talk
and
then
we
can
come
back
and
do
the
rest
just
for
iterations
sake,
because
nobody
likes
a
three
thousand
line.
Mr
at
least
not
no
review.
I
sound
when
your
viewers
do
so
yeah,
so
these
are
the
last
steps
to
do
the
thing
that
I
find
interesting
and
I'm
still
trying
to
figure
this
part
out.
A
A
A
Entertaining
is
a
word
I'm
trying
to
figure
out
how
I
can
go
about
proactively,
making
a
documentation
change
to
make
that
a
lot
easier
to
figure
out,
because
it
was
not
straightforward
and
there's
documentation
for
it
and,
like
I
found
two
places
and
when
jason,
you
know
why
you
look
at
this.
It
was
the
third
place
that
it's
documented
that
I
hadn't
found
yet,
and
that
was
what
had
the
kind
of
the
magic
sauce.
A
So
I'm
gonna
try
and
figure
out
how
to
make
that
more
discoverable
as
far
as
the
services
themselves,
it's
nice
to
have
that
direct
name.
So
that
is
very
handy,
especially
in
the
ansible
inventory
setup.
It's
not
bad.
I
don't
know
that.
I'm
using
it
the
way
it's
intended
so
we'll
see
the
one
thing
I
find
kind
of
off-putting
and
the
metroid
pointed
out
trying
to
use
the
protection,
so
the
containers,
the
way
to
make
I
make
it
run.
Is
I
like
to
install
ssh
into
the
container.
A
A
I
just
start
service,
and
then
I
tell
no
that's
how
that
container
stays
up
for
the
gitlife
container,
because
then
they'll
never
fail
right.
I
don't
want
to
depend
on
the
service
running
or
not
running,
or
you
know
whatever,
so
I
made
it
tell
null
and
I
find
that
to
be
kind
of
hacky,
but
it
works,
and
the
other
thing
is
when
reading
about
having
ssh
in
a
container.
A
A
lot
of
folks
say:
that's
a
really
bad
idea
to
have
ssh
running
inside
a
container
so
for
a
pipeline
for
testing
in
a
local
environment
because
it's
not
accessible
to
the
outside
world.
I
don't
see
it
as
a
huge
problem.
I'm
gonna
tap
once
I
have
this
done.
I'm
gonna,
tap
andrew
and
have
security.
Take
a
look
at
it
and
be
like
hey
I've.
A
Seen
up
and
down
on
this,
do
you
have
an
opinion
and
get
some
more
feedback
that
way,
and
the
other
piece
of
this
is
that
the
metro
has
suggested
trying
a
the
ansible
has
a
docker
connection.
You
can
switch
your
connection
type
from
ssh
to
parameco
and
they
also
have
a
docker
connection.
That
is
now
part
of
upstream.
A
I
tried
and
tried
and
tried
on
that
last
night,
hoping
to
try
and
work
that
in
today's
demo-
and
the
answer
was
no,
it
didn't
work
because
they
couldn't
get
it
to
work.
So,
if
anybody's
interested
in
pinging
on
that,
this
containers
use
docker
connection
branch
on
the
installer.
A
If
anybody
wants
to
take
a
look
at
that
see
what
they
can
figure
out,
I'm
open
to
suggestions,
I'm
going
to
move
on
to
get
the
rest
of
it
done
because
I
feel
like
this
could
be
an
iteration
down
the
road,
but
I
figure
if
anybody's
interested.
This
was
the
other
piece
of
this
services
thing,
because
then
we
can
drop
the
services
all
together
and
have
it
all
run
this
way.
A
D
It's
not
a
bad
idea,
so
when
it
comes
to
the
ssh,
I
understand
that,
like
generally
speaking
for
a
actual
service
container,
it
doesn't
make
sense
to
have
ssh
in
there.
You
really
don't
want
to
be
doing
that
unless
you
have
something
that
actually
needs
to
use
the
ssh
protocol
yeah,
but
for
what
we're
using
it
for
and
the
way
we're
using
it
right
for
right
now,
it's
the
best
available
tool
to
us,
so
it
makes
total
sense.
D
A
A
Well,
here's
the
thing:
if
they're
wanting
to
do
a
source
install
inside
of
a
container,
then
if
we
pick
well,
I
haven't
tried
doing
the
docker
local
install
on
a
docker
local
machine
without
being
inside
of
docker
and
docker.
Because
from
what
I
can
tell-
and
this
is
where
I
kind
of
stumbled
into
it
in
the
docker
community-
plug-in
for
the
docker
connection.
A
The
the
semantic
version,
verification
that
happens
at
the
start
of
the
cut
module
does
not
seem
to
honor
that.
So
it's
like
a
check
inside
of
ansible.
So
it
might
actually
be
an
upstream
bug,
but
it
does
show
up
when
it
runs
it.
So
I'm
not
sure
what
it's
doing
and
that's
I
said
that
merits
more
investigation
so
but
yeah
well,
I
said
most
folks,
I
think
doing
source
installs
they're
doing
in
virtual
machines
seems
like,
and
this
will
target
that
easily.
A
A
And
there
is
one
other
interesting
tidbit
that
might
be
an
oddity
that
we
saw
this
with
full
scale
automation.
I
think
I
probably
need
to
put
the
ssh
port
on
an
alternate
port,
because
once
gitlab
is
up
and
running
it's
going
to
go
to
part
22.