►
Description
The 11th meeting of the CMS/Kubernetes SIG sees a short presentation from DDEV team member Shawn Stratton! Shawn attempts to give the group an improved structure by defining the problem scope and outlining approaches to help the broader CMS community work with K8s.
Catch up with the group on GitHub: http://bit.ly/338dXC5
A
So
kicking
off
another
meeting
of
the
the
massively
fun
cube,
sig
CMS
kubernetes
focus
group
that
we've
been
working
on
with
Drupal
a
little
bit.
I.
Think
that
you
know
you
some
of
the
motivation
for
creating
this
group
is
specifically
to
talk
about
some
of
the
challenges
that
we
face
with
hosting
CMS's
like
Drupal
on
kubernetes.
A
It's
got
its
fair
share
of
challenges.
We
have
a
pretty
strong
community
in
the
drupal
space
and
I
think
are
looking
at
ways
of
taking
what
we're
working
on
and
taking
what
we're
understanding
to
the
broader
world
in
some
capacity.
There's
a
lot
of
lessons
that,
for
example,
the
WordPress
community
or
the
type
of
three
community
or
the
symphony
community,
might
be
able
to
appreciate,
as
a
result
of
going
through
understanding
the
benefits
of
kubernetes
and
what
that
has
to
offer.
So,
that's
generally,
why
we
get
together.
A
There
is
a
github
resource
page
that
everybody
can
go
to
to
kind
of
see
the
latest
updates
with
the
cig,
and
that
is
github.com,
slash,
droid,
slash,
cig,
CMS,
saying
CMS.
To
be
specific,
I
mean
we
also
have
a
little
bit
of
interest
in
a
conversation
that
happens
on
a
regular
basis
in
the
drupal.org
slack
channels.
So
if
you
have
a
chance
to
join
us
in
pound
kubernetes
in
the
drupal.org
slack
channel,
that's
a
great
way
of
talking
to
us
as
well.
A
A
A
lot
of
different
pieces
of
advice
is
some
simple
questions
and
answer
types
of
sessions
have
happened
in
this
conversation
previously,
but
in
off
channel
conversations
we
basically
are
starting
to
come
up
with
maybe
a
little
bit
more
focus
for
this
group
and
want
to
rely
on
a
lot
of
the
learnings
that
we
get
from
the
kubernetes
community.
As
far
as
how
to
manage
these
groups,
how
to
make
these
groups,
as
applicable
as
possible
to
a
broad
range
of
interests
into
a
broad
range
of
people.
A
So
I
don't
want
to
kind
of
give
away
too
much.
But
we've
got
a
pretty
fantastic
set
of
proposals
that
I
think
are
coming
out.
Shaun
Stratton
is
going
to
walk
us
through
those
a
little
bit
and
share
some
of
his
ideas
that
he
has
for
possibly
unifying
some
of
the
effort
that
we
put
into
this
group
so
interested
in
hearing
those.
A
Normally
we
start
off
with
a
quick
round
of
introductions,
but
the
people
that
are
on
this
call
are
fairly
familiar.
They've
all
been
on
previous
calls
before
so
I
think
we
can
go
ahead
and
skip
that
part
and
Shawn.
If
you'd
like
to
go
ahead
and
share
your
screen,
I
assume
that
there's
yeah
cool.
So
if
you
don't
mind
I'll
hand
the
mic
over
to
you
and
looking
forward
to
hearing
what
you
have
to
say.
B
C
I'm
Shaun
Stratton,
this
presentation
will
be
actually
fairly
fast,
I,
usually
present
to
a
group,
that's
more
geared
kids.
So
the
fact
that
we
have
a
small
room
today
that
kind
of
negative,
but
in
general,
but
in
general
I,
just
want
to
do
a
quick
introduction.
Just
so
there's
some
ground
as
to
why
I'm
making
the
proposal
I've
been
a
PHP
developers,
technically
they
late
90s,
but
the
first
major
work
was
in
2001.
I
worked
as
a
PHP
developer,
primarily
until
2012,
which
point
I
switched
over
to
operations.
C
So
why
am
I
talking
today?
Back
in
November,
we
had
an
on
site
where
I
hadn't
obtained
t
be
part
of
the
group
and
during
the
on-site,
ty
I
noticed
that
you
know.
While
we
were
discussing
issues
in
general,
there
wasn't
a
really
a
big
driving
force
behind
the
sing
itself.
There
weren't,
for
example,
actionable
items
coming
out
of
that
meeting
there.
There
weren't
things
to
focus.
You
know,
effort
towards
a
table
goal,
so
that
was
kind
of
my
Drive
and
I.
C
C
C
There
isn't
really
a
good
definition
of
CMS
out
there,
because
at
the
base
of
it
you
know
shopping
carts,
a
CMS
or
a
framework
could
because
they
receive
us
so
I
create
a
list
of
what
I
consider
the
top
five
current
CMS
is
out
there
to
me,
that's
more
like
blog
platforms
and
then
static
website
platforms.
Things
of
that
nature.
My
list
at
the
moment
is
WordPress
Drupal,
Joomla
typo3
and
a
JavaScript
library
called
ghost,
but
even
within
that
it
was
really
difficult
to
kind
of
narrow
down,
but
the
current
landscape
of
those
five.
C
You
know
it
brought
some
interesting
challenges
for
the
community's
community
or
specifically
for
the
CMS
community
on
communities
out
of
the
list.
I
just
gave
out
of
the
out
of
those
five.
Only
four
or
holding
one
of
them
actually
had
its
own
official
docker
image
out
out
and
about
that
wasn't
matched
by
the
doctor
community,
for
example.
C
Wordpress
is
docker
image,
dropper
and
documents
they're
all
handled
by
docker
community,
so
they're,
darker
official.
When
you
go
to
you
the
the
docker
registry,
for
example,
they
will
be
marked
as
official
images
but
they're,
not
necessarily
guided
by
the
community.
That's
by
the
software
running
on
them.
Further,
none
of
the
projects
and
styles
offered
any
form
of
good
remains
to
climate
manifests
or
helm
charts,
or
any
tools
like
that
now
there's
a
third-party
company
named
Nami
that
does
I
should
provide
home
charts
for
four
out
of
five.
C
They
current
divided
on
charts
before
WordPress
goes
drupal,
joomla
and
ton
of
other
software
out
there.
But
again
it's
not
close
to
the
CMS
group
itself.
It's
not
close
to
that
community.
It's
not
management
a
community,
but
the
real
surprising
that
was
when
I
was
actually
looking
through
the
documentation
on
wordpress
or
through
the
documentation
on
drupal.
None
of
them
actually
had
any
active
sections
of
hey.
This
is
how
you
run
our
software
on
this
platform.
C
Now,
in
famous,
didn't
really
have
a
lot
on
cloud
platforms
either,
but
at
the
same
time,
magic
pen
ian
is
a
new
developer
in
finding
that
there's
no
documentation
of
how
to
run
it.
In
today's
and
today,
technology
I
also
found
a
couple:
long
learning
tickets
for
functionality,
especially
in
WordPress.
You
know
that
they've
been
out
there.
C
The
the
request
has
been
in
the
case
of
the
one
I
thought
on
WordPress,
for
example,
to
add
environmental
support,
some
function
that
I
don't
eat
your
call,
but
the
point
being
is
the
the
requesting
it's
been
on,
sir,
since
2017
nobody's
really
actually
focused
on
that
work.
So
it
almost
feels
like
go.
Let's
the
communities
communities
to
new
for
these
projects
to
really
take
seriously
and
then
the
final
one
as
you
as
leaders,
advocate
working
Danish
for
a
while
are
the
standard
application.
C
Architecture
is
what
Goodson's
twelve
factor
now.
That's
a
good
guideline
for
how
to
write
applications
of
deploy.
Here's
it's
not
necessarily
the
best,
but
at
the
same
time
it
kind
of
gives
from
the
underlying
underlying
requirements
that
we
have
you
know.
So,
for
example,
you
don't
write
to
local
file
system,
each
application,
triva
stateless,
you
use
environment
for
your
configuration,
etc,
and
none
of
these
CMS's
support.
You
know
even
the
basic,
the
basic
outputs
of
that.
C
Let's
say:
I
knew
nothing
about
WordPress
and
today,
I
got
the
you
know:
Kevin
Kenai
from
Ohio
said
you
must
deploy
WordPress
to
our
production
instance,
our
production
cluster.
Where
would
I
come
to
find
that
information?
You
know
we're
do
I
really
look
now.
The
obviously
the
community
has
stepped
up
significantly.
There
are
tons
of
blogs
about
everything.
One
of
these.
You
know
the
CMS
systems
out
there
and
how
to
run
them
on
kubernetes
and
recommendations.
Some
are
more
up-to-date
than
others.
C
It's
gonna
couple
that
you
know
probably
should
have
expired
two
two
years
ago
that
these
API
versions
and
their
deployment
examples
that
no
longer
or
no
longer
supported,
etc,
etc.
But
it
creates
an
interesting
challenge
and
that
challenge
is
compounded
by
the
fact
that
each
DCMS
systems
is
obviously
their
own
project
run
by
their
own
community
running
with
their
own
tech
stack,
whether
it's
the
technology
in
use
by
the
CMS
or
if
it's
the
technology
you
know
of
how
they
host
their
posters,
responds,
let's
get
or
mccurry
or
it
lab
versus
a
github,
etcetera,
etc.
C
So
it
seems
to
me
that
we're
in
a
position
to
address
some
of
these
challenges,
specifically
I,
think
that
we
are
in
position
to
where
we
can
at
least
as
members
of
those
communities
as
well
as
this.
This
SIG
and,
as
you
know,
journal
ambassadors,
for
you
know,
development
within
the
space
and
we
can
address
so.
My
proposal,
it's
multifaceted,
but
the
first
proposal
that
I
think
we
need
to
do,
and
this
would
help
a
lot
and
focusing
the
efforts
towards
you
know
what
we
do
to
adopt.
C
Is
we
needed
to
find
a
mission
for
ourselves
now?
Obviously,
here
I've
got
mission
statement,
that's
great,
because
that'll
help
other
people
that
join
the
CMS
saying
after
this
to
kind
of
understand
what
our
goals
are,
but
beyond
that
it's
just
actually
identifying
the
mission
and
then
find
how
we're
gonna
implement
it
angle
in
that
direction.
I
think
we
need
to
document
our
involvement
with
the
CMS's.
That
journey
means
we
can
pay
partnerships.
C
You
know
identify
people
that
you're
either
part
of
both
communities
or
somebody
from
one
of
the
CMS
communities
that
will
work
for
us
beyond.
That
I
think
is
quite
like
being
quite
easy
for
us
to
create
a
resource
package
just
to
kind
of
address.
Some
of
the
main
challenges
must
these:
let's
face
it.
Some
of
these
CMS's,
like
WordPress,
have
been
around
for
for
a
while,
and
it
becomes
a
little
challenging
tired
and,
if
I,
how
to
simply
update
some
things
without
breaking
backwards
compatibility.
C
So
it
would
be
beneficial
to
create
a
resource
package,
yeah
some
documentation
and
some
guidelines
for
how
to
humble,
with
these
CMS's
to
easily
onboard
antique
communities
and
then,
as
a
stretch,
goal
with,
if
you
feel
the
at
the
time,
and
if
we
can
do
something
a
little
bit
more
serious.
My
people,
where
I'm
beneficial,
create
a
certification
program
to
work.
You
know
we
can
give
another
the
end
user
or
the
CMS
opens
with
media
badge,
since
this
runs
on
kubernetes
that
we
know
from
the
communities.
C
But
this
doesn't
obviously
doesn't
go
far
enough
as
far
as
actually
working
on
addressing
some
of
those
shortcomings,
I
was
mentioning
earlier.
This
is
great
for
us
in
this
group,
but
I
think
the
the
real
power
here
is
the
partnerships
with
the
other
groups,
some
of
the
areas
I've
identified
for
you
know,
working
working.
We
have
the
e
to
the
CMS
groups
directly,
where
we
can
have
a
massive
impact,
because
that
people
would
be
documentation
now
when
it
comes
to
WordPress
and
even
I
believe
Drupal.
C
There
are
some
example
documentation
that
communities
itself
hosts,
but
it
would
be
good
to
create
reference
architecture.
You
know
this
is
your
service.
This
is
your
optional,
my
sequel
database.
This
is
how
you
configure.
This
is
how
you
configure
that
and
then
work
with
the
CMS
community
itself
to
create
documentation
that
can
be
integrated
into
their
actual,
hosted
documentation
on
how
to
run
their
system
within
communities
beyond
that,
as
I
mentioned
earlier,
most
these
products
don't
maintain
their
own
docking
containers.
They
they're
completely
hands-off.
C
They
they
ship
the
software
out,
and
then
you
know
the
doctor
community,
for
example,
will
turn
around
and
create
the
docker
image
file
and
then
go
from
there,
but
it
may
be
beneficial,
at
least
with
some
of
these
larger
ones
if
they
actually
host
or
if
they
actually
create
their
own
doctor-doctor
files
or
other
container,
when
containerization
technology
build
packs
just
working
with
that
would
allow
them
to
have
more
control
over
what
it
is
that
they're
actually
publishing
out
there
I.
Don't
you
know
again,
I
can't
speak
for
daily
here,
but
I.
C
Would
you
rather
upset
it?
There
was
some
in
let's
say
the
word
post
official
image
that
now
I'm
having
to
go
out
and
trying
to
communicate
with
three
different
groups
identify
where
the
bug
comes
from
how
to
resolve
it.
Get
it
to
the
poem
request
process
may
get
out
there.
So
I
see,
there's
some
significant
improvements
in
how
that
can
work.
I
just
think
I'm
working
with
the
open
source
projects
themselves
say
look
we
can
to
read
the
files
to
you,
but
it
would
be
beneficial
if
this
part
of
you
know
your
own
release
cycle.
C
You
know,
deployment
ya
know,
manifests
and
work
with
them
to
teach
about
how
to
use
customize,
for
example,
or
any
a
number
of
other
different.
You
know
different
tools
out
there
and
then
finally,
I
think
the
other
major
area
where
we're
we're
kind
of
lie
at
the
moment
is
advocacy
so
sig
memory.
You
know
we're
kind
of
in
this
unique
position
that
we
care
about
software
runs
on
kubernetes.
We
care
about
kubernetes,
but
we
don't
necessarily
do
a
good
job
at
you
know.
C
So
some
of
these
things,
you
know,
might
be
beneficial.
Just
somebody
going
through
and
saying
hey
as
part
of
that's
part
of
the
saying,
I'm
also
part
of
the
Drupal
community,
I
think
it's
thoroughly
into
containerization.
Specifically,
you
know.
Maybe
I
can
address
some
of
those
issues.
Maybe
I
can
help.
You
know
advocate
person
on,
for
instance,
it
should
just
be
prioritized
and
brought
into
the
workflow.
C
So
that's
that's.
Where
I
see
us
coming
from
just
eating
the
products
themselves,
you
know
whether
it's
in
the
form
of
advocacy
or
whether
it's
in
the
form
to
you
know
to
mean
documentation,
or
you
know,
guiding
through
some
of
the
technology
technological
challenges
of
adopting
kubernetes,
whether
it's
defining
the
architecture
or
defining
the
manifests
or
saying
hey.
This
is
how
we
package
manage
this
or
here's
how
we
use
templates
to
make
this
possible
etc,
but
at
the
same
time,
I
know
that
you
know,
as
as
Kevin
I
discussed.
C
C
You
know,
call
on
others
to
speak
up,
and
you
know
bring
up
your
own
ideas
today,
we'll
be
filing
a
ticket
on
the
on
the
the
cigs
git
repository
to
kind
of
talk
to
this
proposal
in
a
open
forum
and
I'm,
not
even
sure,
what's
in
the
write
process
would
be
to
adopt
it,
but
this
would
be
the
perfect
time
and
I'm
inviting
people
to
come
out
and
actually
make
comments
on
that
give
feedback.
You
know
mention
things
that
I
forgot
mentioned
the
areas
where
I
may
have
overshot
the
yeah.
C
C
The
examples
I
had
here,
something
that
I
was
kind
of
thinking
about
and
pointing
to
earlier,
is
maybe
we've
been
official
didn't
get
a
list
of
plugins
for
the
based
CMS
is
like
Drupal
or
WordPress
that
make
it
possible
to
run.
You
know
that
CMS
directly
on
communities,
I'm
thinking
of
things
like,
for
example,
uploading
files
to
you-
know,
block
storage
for
buckets
rather
than
directly
into
the
process
itself.
You
know
so
I'm
completely
open.
If
anybody
has
any
ideas,
please
reach
out,
please
yeah
advocate
for
your
thoughts
and
that's
yeah.
A
A
A
So
typically,
what
we
do
at
this
point
is
just
kind
of
open
it
up
to
any
recent
challenges
or
celebrations
that
anybody
from
the
group
would
like
to
share
any
news
that
they've
heard
of
in
the
community
that
in
the
kubernetes
community
or
the
Drupal
community
or
any
other
CMS
community
that
is
exciting
to
them.
At
this
particular
time.
A
I
know
some
of
the
things
that
are
very
interesting
to
me
right
now,
just
kind
of
the
progression
of
home
3.
You
know
home
presented
a
lot
of
good
opportunity,
a
lot
of
good
potential
and
I
felt,
like
it
kind
of
fell
apart
a
little
bit
coming
out
of
version
2
and
into
version
3,
so
I
I
personally
haven't
been
keeping
up
with
that
enough,
but
that
is
one
area
that
I
do
find
intriguing
at
this
particular
time.
I
think.
D
Shawn
has
worked
with
it
a
little
bit
more
I'm
I'm.
Also
in
the
curious
phase.
I've
I
have
scars
for
life
from
home
too
and
I
agree
like
the
promise
of
helm.
2
is
incredibly
solid
and
I
think
it's.
It
was
really
ambitious
and
did
a
lot
of
great
things,
so
I'm
curious
actually
to
see
what
helm
3
is
doing
with
the
new
architecture
that
they
have
I've.
C
Worked
really
quite
a
bit
I'm
unfortunate
I
think
there's
some
shortcomings
in
it.
Probably
not
a
blocky
nature
like
home,
tease,
tease,
tiller
model
but
I
mean
they're
they're,
definitely
challenges
with
some
little
rough
edges,
and
things
like
that.
Unfortunately,
where
I
skew
the
real
problem
is
the
community
support
round.
3
is
not
as
good
as
I
would
want
it
to
be,
and
that's
just
coming
from
my
own
experience
with
Fargo
and
and
also
with
terraform.
You
know
it's
building,
but
you
know
the
the
plugins
support.
C
The
big
difference
is
that
the
model
is
completely
changed
right.
So
originally,
this
is
true
for
terraform
I,
don't
know
about
others,
but
originally
there
were
two
protos.
Why
don't
you
the
rasterizer,
charts
and
completely
ditched
the
you
know?
This
is
managed
by
helm
yeah.
This
is
an
install
to
tiller,
so
till
here's
aware
of
what's
being
installed,
Center,
etc
or
you
you
know,
you
actually
managed
tiller
as
well,
and
the
problem
is
tiller
is
not
part
of
this.
C
D
I
will
add
this
I
generally
I
have
been
bitten
enough
by
tools
that
do
runtime,
rendering
of
manifests
in
like
especially
like
in
a
C
ICD
pattern.
I've
generally
really
now
try
to
steer
away
from
that,
and
the
reason
is,
is
it
becomes
a
lot
more
difficult
depending
on
how
sophisticated
your
template
is
or
your
templating
tool
is
interpreting
your
whatever
your
markup
or
whatever?
Is
it's
interpreting?
D
C
Yeah
I
mean
it's
nothing,
a
challenge.
I'm
not
gonna,
lie
about
that.
It's
it's
been
a
challenge
in
the
past,
especially
having
written
home
charts.
You
know,
the
the
the
btf
counter
per
minute
is
usually
fairly
high
when
you're
doing
that
and
at
the
same
time
there's
a
significant
amount
of
promise
still
and
bundling.
The
entire
project
templates
isn't
something
that
can
easily
be
modified
and
I'm
yeah.
The
problem
really
is
it's:
it's
six
in
one
hand
and
a
half
dozen
in
the
other.
C
D
C
The
biggest
challenge
here
and
I
mean
this
is
kind
of
least
where
I,
where
I
can
have
have
issues.
Is
it's
not
enough
to
know
how
am
I
not
enough
to
know
customize
I'm
not
know
JSON
it
every
tool
that
comes
into
play
eventually,
you're,
given
into
you,
have
to
build
some
working
knowledge
on.
It
really
becomes
a
major
challenge
because
over
time,
the
different
approaches
and
different
conventions,
etc,
they
all
run
counter.
C
So
when
you,
you
know,
when
you
focus
on
one
tool
for
too
long,
you
completely
forget
what
now
customized
specific
stuff,
or
you
know
the
helm,
specific
stuff
that
you
worked
so
hard
as
figuring
out.
So
that's
really
the
big
challenge-
and
this
is
you
know
when
we
were
in
the
container
orchestration
Wars,
and
this
was
kind
of
the
challenge
we're
addressing.
This
is
why
you
know
kubernetes
over
UK,
air
or
ECS
or
whatever
else.
C
D
D
It
seems
like
right
now
with
that.
What
what
the
community
is
doing
is
that
it
you
have
helm,
3
and
customizes.
Really
two
competing
projects,
the
terraform
module
I
would
say
would
be.
I
would
put
that
as
like,
like
a
second
here
in
that
Sun
know
that
I
think
that
necessarily
about
terraform,
it's
more
in
that
that's
not
like
in
a
module
of
a
much
greater
scoped
project.
C
Much
agreed
I
mean
the
thing
about.
It
is
even
even
with
Helmand
customized
they're,
not
technically
competing
projects,
I,
don't
get
album
or
like
I
looked
at
at
and
look
at
customized
more
as
I.
Look
at
no
command-line
tools,
like
maybe
art
and
and-
and
you
know,
sad
or
something
like
that,
the
the
point
of
it
being
is
that
they
all
have
different
focuses,
but
it'll
be
nice
to
see
something.
That's
cohesive
across
the
entire
board
yeah
and
you
run
the
risk
of
you
know.
C
I
advocate,
for
that
we
run
the
risk
of
having
you
know
like
rpm
and
apt
and
having
massive
competition
out
there,
but
I
just
don't
think
it's
properly
solved
in
this
morning,
I
mean
there's
no
way
of
saying
look,
I
want
word
pro
or
I
won't
WordPress.
Where
I
want
my
sequel
in
memcached,
you
know
sure
Helms
got
some
charts
there,
although
the
efficiency
and
how
long
they've
been
maintained
and
how
they're
there's
nothing
questionable
for
some
of
them.
But
again
it's
it's
a
discussion
about
yeah.
C
D
As
long
as
you
know
how
to
write
go
to
do
like
a
transformer
plug-in,
alright
generator
plugins,
actually
the
bigger
one
where
you
can
codify
that,
if
it,
if
it's
complex
enough
to
where
writing
it
and
go
or
in
Python,
makes
more
sense
or
it's
more
intuitive,
it's
it.
The
programming
language
fits
it.
It's
better
suited
to
describe
the
solution
to
the
problem
you're
trying
to
solve,
or
what
your
then
I
think.
That's
the
way
to
go
and
I
think
helm
has
been
trying
to
do
that
with
the
Lua
support.
D
D
D
A
Generally
speaking
as
we
move
forward,
you
know
this
is
going
to
be
a
relatively
short
meeting
on
February
12th.
We
have
a
gentleman
named
Nick,
Santa,
Maria
I,
believe
he's
from
Australia,
but
he
might
be
from
New,
Zealand
and
I'm
sure
that
I
will
receive
slack
from
getting
that
confused,
but
he
works
on
a
project
called
skipper,
which
is
a
very
interesting
approach.
A
Oh
sting
with
kubernetes
and
he's
going
to
give
us
an
overview
that
it's
potentially
possible
that
we
will
have
Florian
Laurette
I
believe
is
how
you
pronounce
his
name
from
wonder:
he's
going
to
give
us
a
session
on
migration
from
home
to
town
3.
So
that
seems
like
a
pretty
relevant
topic
today
to
today's
conversation,
we're
still
looking
for
a
presenter
for
the
February
26th
meeting,
so
if
anybody's
available
for
that
please
reach
out
and
slack
or
on
github
and
we'll
be
happy
to
accommodate
you
and
then
moving
into
April.
A
So
it
looks
like
we
still
have
a
little
bit
of
work
to
fill
out
March,
but
we
have
a
gentleman
named
Brad
Jones,
who
is
going
to
give
a
presentation
on
cert
manager
and
traffic
2.0
to
point
X
I
should
say
so
definitely
keep
an
eye
out
because
it
looks
like
we
do
have
some
content
starting
to
formulate.
That
should
be
interesting
to
most
folks
again.
Thank
you
all
for
your
time.