►
From YouTube: Cloud Foundry CAB call April 2022
Description
The topic for the call was Korifi (Cloud Foundry on Kubernetes).
Participants included several key community members, some folks from the Working Group, and the Cloud Foundry Foundation representatives. The important part of the call was a demo of the Korifi project, where it was installed on a Google Kubernetes cluster.
Watch the call to learn the status of the project and also take a peek at what the future direction holds!
A
Hello,
everybody
thank
you
for
joining
us
for
the
cap
call
in
april
2022.
A
Definitely
has
been
a
year
of
a
lot
of
progress
in
terms
of
what
the
community
has
achieved,
and
so
I
wanted
to
spend
some
time
during
this
call
just
chatting
generally
about
one
specific
project
which
is
named
kurifi,
which
is
our
third
iteration
of
putting
cloud
foundry
on
kubernetes.
A
So
there's
a
few
things
that
I
wanted
to
showcase
today,
some
of
it
technical
so
we'll
be
doing
like
a
walk
through
of
how
to
do
the
installation.
Hopefully
the
installation
works
okay
and
we're
able
to
push
an
app
on
to
and
do
like
the
complete
life
cycle,
but
also
I
wanted
to
talk
generally
about
where
the
project
stands.
A
A
If
you
know
I,
I
don't
want
to
sound
like
a
public
service
announcement,
but
if,
if
you,
if
at
all
you
see
something,
say
something
so
there
might
be
some
missteps
that
I
make,
that
might
be
some
parts
of
the
read
me
or
the
documentation.
That
is,
you
know
missing,
or
that
might
be
some
errors
that
you
spot.
So
just
if
there's
something
please
let
us
know.
Obviously
I
will.
I
have
my
notes
nearby
for
changes
to
some
of
these,
but
whatever
you
are
able
to
add
to
that
is
much
appreciated.
So.
A
A
A
A
vision
for
how
cloud
foundry
should
be
presented
on
kubernetes
and,
what's
the
best
way
to
you,
know,
add
value
to
something
as
popular
as
kubernetes
and
make
a
definitive
push
for
the
wonderful
cf
developer
experience
eric.
Do
you
want
to
add
some
information
as
to
what
the
document
is
and
quickly
summarize?
Some
of
these
points,
sorry,
to
put
you
on
this
part,.
B
Oh
hey
ron,
yeah
yeah
sure
I
can
speak
to
this
a
little
bit,
so
this
is
a
document
that
we
drafted
with
some
folks,
primarily
from
sap
and
ibm,
really
in
the
first
half
of
last
year,
laying
out
what
we
thought.
The
productive
pathway
forward
for
cloud
foundry
and
its
concepts
on
top
of
kubernetes
would
be
yeah,
especially
based
on
a
lot
of
what
we
had
learned
from
cf
for
kate
and
cube
cf.
B
And
so
you
know,
I
think,
the
document's
still
there
there's
a
little
bit
of
still
commentary
and
discussion
on
it
from
when
we
were
discussing
it
in
the
community
last
year.
But
you
know
what
we
really
settled
on
is
you
know?
B
We
think
that
there's
still
a
lot
of
value
in
these
developer,
abstractions
that
we've
built
up
in
the
in
the
cf
community
over
the
past
many
years,
not
just
applications
and
routing,
but
things
like
service
brokers
and
service
findings,
so
continuing
to
present
those
interfaces
and
integrating
those
with
the
evolution
of
technologies
like
build
packs
and
service
brokering
systems
still
seems
very
valuable,
but
wanting
to
integrate
that
more
more
thoroughly
with
a
lot
of
the
resources
that
we
see
coming
from
kubernetes
and
ultimately,
making
that
more
modular
and
extensible
as
an
architecture
in
ways
that
both
cube,
cf
and
c4
gates
really
weren't
designed
to
be
you
know.
B
Qcf
is
really
lifting
the
entirety
of
the
whole
bosch
based
architecture
onto
kubernetes
and
even
cf
for
kate's.
You
know
it
was
taking
more
steps
to
integrate
kind
of
the
back
end
with
kubernetes
systems
and
taking
advantage
of
what
it
could
do
around
workload,
placement
and
different
routing
systems,
but
it
was
still
fundamentally
like
taking
that
entire
cf
domain
model
and
kind
of
lifting
it
onto
kubernetes
and
running
it
there,
and
you
know,
I
think,
both
of
those
ways
were
they
were
kind
of.
They
were
a
little
bit
like
sugar
highs.
B
In
terms
of
you
know,
we
got
a
lot
of
activity
and
productivity
very
quickly,
but
then
we
started
hitting
more
significant
technical
walls
or
integration
walls
trying
to
integrate
more
deeply
or
more
thoroughly
with
kubernetes
technologies.
B
So
that
really
motivated
us
stepping
back
and
saying
like
what
should
this
overall
architecture
start
to
look
like,
and
you
know
how
should
we
take
a
more
productive
path
at
developing
that,
even
if
it,
in
some
cases,
it's
going
to
take
longer
to
iron
out
because
of
the
modularity
and
interfaces
that
we
need
to
build
into
it?
And
so
this
really
spurred
development,
starting
in
the
second
half
of
last
year
into
what's
now
what
we're
now
calling
the
kariki
project?
A
It's
part
of
the
vision
document
and
a
bunch
of
other
discussions
as
well,
but
yeah
here
we
are
and
let's
go
through
sort
of
the
steps
that
have
been
outlined
and
try
and
get
this
installed
on
like
a
gke
cluster.
Now
this
one
of
the
historic
difficulties
that
have
been
that
the
community
has
been
very
vocal
about
is
cloud
foundry
in
the
bosch
sort
of
version
as
well
has
not
been
the
easiest
to
you
know,
try
out.
A
So
I
think
there
are
some
projects
like
bosch,
light
and
other
things
that
try
and
make
bosch
available
on
like
systems
that
require
lower
compute
resources
and
things
like
that.
A
But
often
more
often
than
not,
I've
heard
people
say
that
you
know
it
can
be
a
little
difficult
to
just
have
like
a
a
trial
version
that
runs
on
low
resources
on
a
laptop,
and
it
particularly
affects
the
ability
to
advocate
this
to
like
a
bunch
of
different
people,
saying
you
know
just
try
this
out
on
your
machine
and
it
will
work,
and
so
one
of
the
big
changes
that
I've
observed
with
this
project
is
from
the
get
go.
You
have
the
ability
to
try
this
out
on
a
kind
cluster.
A
Now
that
was
true
of
cf
on
kids
cf4
gates
as
well,
but
the
the
focus
hasn't
been
as
much
as
it
is
right
now
so
kind
clusters
are
first
class
citizen
and
local
development.
Workflows
are
supported,
as
you
know,
as
a
first
first
class
developer
experience
for
this
project,
and
so
it's
not
recommended
that
you
get
like
a
big
fat
kubernetes
cluster
in
order
to
install
this,
but
the
installation
steps
are
independent
of
what
kubernetes
provider
that
you're
using.
A
So
this
should
definitely
work
on
other
machines
and
things
as
well.
So
if
you
have
time
clusters
go
for
it,
if
you
have
some
other
preferred,
kubernetes
providers
go
for
it.
The
experience
should
necessarily
be
the
same.
So
with
that
let's
get
started,
so
the
prerequisites
are
very
a
very
predictable
list.
I
guess
so.
You
obviously
need
cube
cpl.
A
I've
got
that
installed,
yes
and
then
compatible
kubernetes
cluster.
So,
like
I
mentioned,
this
is
a
gke
cluster
that
I
have
it's
reasonably
big,
a
container
registry
that
you
have
right
access
for,
so
today
we
will
be
making
use
of
docker
docker
hub
for
this,
then
again,
you
know
it's
it's
expected
to
work
with
like
other
container
registries
as
well.
You
obviously
need
the
cfcli,
and
so
I've
got
that
installed
as
well.
A
A
The
next
step,
obviously
is
to
clone
this
ripper.
So
this
is
the
official
cloud
formula.
A
So
kurifi
stands
for
a
peak
like
a
mountain
peak
and
we
decided
to
go
greek,
keeping
in
mind
that
kubernetes
and
pacquiao,
and
all
of
these
projects
are
taking
advantage
of
some
of
the
words
there,
and
so
the
community
came
together
and
voted
for
the
name
last
month,
so
among
the
different
ones
that
were
shortlisted.
A
I
think
we
had
like
the
least
amount
of
trademark
was
so
one
of
the
others
was
pharaohs
which
stands
for
lightning,
which
is
like
very
favored
among
the
community,
but
unfortunately
this
I
think
like
about
eight
projects
or
that
that's
already
named
for
us,
I
don't
recall
the
other
one
which,
but
then
the
meaning
stood
for
like
a
bridge
between
the
heaven
and
the
earth
so
which
is
also
like
a
very
interesting
name
for
a
project.
That's
you
know
trying
to
bind
apps
between
the
communities.
A
So
anyhow,
here
we
are
kurifi
has
been
shortlisted.
A
And
so
again,
this
is
the
official
sort
of
repo
on
github.
So
if
you
want
to
follow
along,
if
you
want
to
test
this
by
yourself,
obviously
you
know
where
to
hit
them.
So
the
first
step
is
to.
A
Clone
that
I
have
in
the
interest
of
saving
time
sandbox,
so
this.
A
A
Instructions
say
like
if
you
invented
a
cluster,
you
must
first
install
set
manager,
so
there's
two
options
for
installing
search
manager
that
are
available
so
there's
a
dependencies
folder
that
comes
within
the
repo,
so
all
of
the
recommended
versions
of
each
of
the
different
tools
that
you
need
are
available
here.
So
obviously
you
can
either
apply
that
directly
or
you
can
fetch
from
like
the
source
and
apply
that
so
I'm
just
you
know
I
I
prefer
doing
like
the
dependencies
that
come
with
it.
So
let
me
do
that.
A
Some
of
some
of
these
steps
is
longer
than
the
others,
so
I
will
try
and
fill
the
silence
with
my
commentary
as
best
as
I
can,
but.
C
C
So
when
you
create
certain
resources
on
it,
the
kubernetes
api
is
going
to
call
out
to
these
web
hooks
and
it
wants
to
do
that
over
tls.
So
we
have
to
get
searched
somehow
and
start
manager
provisions.
Those.
B
A
A
Installation
steps
recommend
is
to
alter
these
secrets,
for
which
image
registry
that
you're
going
to
use
so
there's
like
a
bunch
of
different
images
that
get
created
that
get
copied
over
and
other
things
as
part
of
the,
as
part
of
both
the
app
creation,
workflows
and
as
part
of
the
installation,
workflow
and
other
things
as
well.
So
the
installation
and
other
parts
require
access
to
like
a
registry,
the
container
registry,
and
so
it's
recommended
that
folks
alter
the
tag
parameter
in
this
camel
so
which.
A
Hey
this
is
so,
like
I
said,
we're
using
docker
here,
so
that
was
like
a
previous
tag
that
I've
replaced.
A
A
A
A
A
The
envoy
and
control
get
created
here
and
those
parts
will
allow
us
to
apply
the
right
kind
of
interest.
Was
that
static,
which
you
can
then
make
use
of
to
access
the
cluster.
A
If
folks
have
questions,
you
can
just
check
those
out
make
use
of
chat
and
I
don't
think
people
are
strangers
to
each
other.
So
in
this
step
we're
going
to
be
able
to
find
that
public
ip.
A
A
A
Yep
yeah
thanks
tim,
I
this
is
unique
just
as
much
so
we
move
on
to
creating
a
namespace
for
ide
and
then
we
need.
D
A
D
A
This
specific
set
of
instructions,
I
think,
point
to
like
an
older
version.
So,
let's,
let's
install
the
1.0
version
just
by
way
of
background
hierarchical
namespaces,
are
sort
of
the
tool
that's
going
to
help
with
arms
and
spaces,
and
so
what
we'll
do
is.
A
A
It
is
going
to
assist
with
orbs
and
name
spaces
and
maintain
that
hierarchically
as
a
one-on-one
mapping.
So
these
are
again
some
of
the
changes
that
are
very
apparent
from
the
start
between
what
we've
observed
with
like
previous
iterations
of
cfn
kubernetes.
So
there
wasn't
like
a
very
good
one
to
one
correspondence
between
all
the
workspaces
and
all
of
these.
Previously
there
was
one
I
think,
cf
workloads,
name
space
that
had
everything
this
time.
A
A
lot
of
effort
has
gone
into
recreating
the
same
kind
of
mapping
between
kubernetes
native
ways
of
organizing
these
namespaces
and
the
cloud
foundry
method
of
organizing
these
namespaces.
A
The
other
area,
where
there's
a
lot
of
change
and
what
has
become
very
equivalent
is
native,
is
also
the
authentication
and
our
back
we'll
get
to
that.
You
know
in
due
course,
but
I
just
wanted
to
mention
this,
so
this
also
like
a
like.
I
mentioned
a
plug-in
that
needs
to
be
installed
locally,
so
we're
going
to
do
that
as
well.
A
Yeah,
I'm
I'll
show
you
what
the
mapping
looks
like
so
there's
a
hash
and
it
basically
makes
use
of
like
a
prefix
to
that
hash.
But
the
hash
is
basically
what
differentiates
different
arts
and
different
spaces
that
we.
C
A
Sequence,
there
is
an
optional
step
here
for
installing
service
bindings.
Obviously,
if
anything
is
prefixed
with
optional,
I'm
going
to
skip
forward.
A
So
let's
see
ask
us
to
create
like
a
root
name,
space.
A
There's
a
couple
of
things
that
have
to
be
matched
so
like
the
kpac
image
tag
is
similar
to
the
one
that
we
did
before
just
apply
the
docker
url
here.
This
is
also
applying
the
docker
url.
A
So
there's
a
make
deploy
command
that
basically
does
the
final
deployment
again
some
of
these
config
changes.
I
did
before
the
call
just
to
the
interest
of
time.
A
A
D
A
C
A
C
C
C
A
C
That's
that's
the
one
that
is
used
to
template
out
that
contour
configuration.
C
So
this
one,
if
you
point
to
your
new
domain,
I
think
it
will
work.
A
Just
so
I
change
it
only
at
one
place
right.
C
Yeah
there's
we
should,
if
you
actually
want
a
push
to
be
functional,
we
have
to
change
it
in
that
other
file
too.
But
if,
if
you
just
make
deploy
right
now,
it
will
at
least
let
you
target
the
api
and
we'll
see
if
everything
else
is
set
up.
C
So
this
this,
that
file
is
used
for
the
actual
routing.
The
other
one
informs
the
api
about
what
its
base
url
needs
to
be
for,
like
redirects
and
stuff
that
it
gives
to
the
cfcli.
A
A
A
A
A
So
we're
there
now
yeah,
that's
that
was
a
fun
install
process.
I
guess.
C
You
want
cf
push
to
work
that
original
config
file
you're
at
you
need
to
change
all
those
example.coms
to
refer
to
this
domain
and
then
you'll
also
need
to
set
up
a
cf
domain.
Unless
you
want
to
push
with
the
no
route
flag,
but
otherwise
it
will,
it
will
fail
and
say
you
need
a
domain
to
be
able
to
create
a
route.
A
A
A
I
think
the
important
parts
so,
if
folks
are
able
to
see
this
question
from
brett
he's
he's
asked
if
there's
any
awareness
of
kubernetes
that
crosses
over
into
the
cf
experience.
I
think
the
big
places
are
in
auth
and
then
it's
been
sort
of
become
more
closely
resembling
the
kubernetes
experience
for
both
space,
the
name
spaces
and
how
we
organize
arms
and
spaces
in
general.
So
I
have
not
seen
any
other
places
again
tim.
Do
you
want
to
sort
of
take
this.
C
Yeah
sure
so,
like
I'll
just
focus
on
off
the
answer
is
yes,
so
the
shim
is
not
doing
any
auth
on
its
own.
It's
actually
delegating
everything
to
the
the
kubernetes
api
and
it's
backed
entirely
by
kubernetes
custom
resources
so
like
the
cli
is
talking
to
the
shim
which
implements
the
cloud
foundry
apis
based
on
those.
C
But
you
don't
have
to
use
that
like
you,
could
use
cube
cuddle
directly
and
make
like
a
cf
app
resource,
cf,
build
trigger
staging
and
everything,
and
that
inherently
requires
that
you
have
a
user
that
is
somehow
recognized
by
the
kubernetes
api.
But
the
kubernetes
like
off
system
is
very
extensible
too.
C
So
there's
like
projects
that
have
like
off
proxies
that
can
like
delegate
to
external
loss
systems
and
stuff
so
like
theoretically,
you
could
hook
up
a
uaa
or
something
that
kubernetes
is
delegating
to
for
your
user
management,
it's
just
with
with
karify
and
the
shim.
We
are
we're,
not
caring.
How
that
happens.
It's
just
like
as
long
as
the
kubernetes
api
trusts
your
user
and
can
apply
it's
it's
our
back
rules
against
it
then
then
stuff
will
work.
C
Cool
and
there's
like
benefits
to
this
too
so
like,
rather
than
cloud
controller,
being
the
like
sole
gatekeeper
of
of
off
here.
All
the
decisions
are
made
by
just
like
regular
kubernetes
are
back
so
like
we
have
a
space
developer
role
at
a
cf
admin
role
that
we're
using,
but
this
makes
it
much
more
extensible
where
people
can
make
custom
roles
that
apply
to
the
resources
they
want
without
having
to
like
go
down
like
months
long
effort
of
updating
cloud
controller
to
acknowledge
that
that
role
exists.
A
Cool
any
other
questions
that
folks
have.
We
can
take
them
here.
If
not,
you
know,
cf
gates.
Dev,
I
believe,
is
the
right
slack
channel.
You
know
so
feel
free
to
post
there,
although
I
think
that
will
soon
have
to
be
renamed.
A
Generally
speaking,
the
curry
free
release
is
what
we
are
looking
forward
to
the
most
this
year.
So
you
know
it's
it's
a
very
key
project
in
terms
of
its
utility
overall
and
for
you
know,
positioning
cloud
foundry
as
an
important
part
of
the
kubernetes
ecosystem
for
for
the
future,
so
yeah.
Thank
you.
Everyone
for
attending.
Thank
you,
tim
for
all
of
the
support.
A
Tomorrow,
so
I'll
share
a
link
to
that
as
well.
So
if
there
are
folks,
you
feel
might
be
interested
in
learning
a
little
bit
more
about
this.
You
know
feel
free
to
share
that
as
well,
and
this
is
obviously
going
to
be
like
a
big
pr
and
marketing
push.
So
hopefully,
you'll
hear
a
lot
more
about
the
whole
project,
but
yeah.
That's
it
for
now,
thanks
everybody
thanks
ashley
for
recording.