►
From YouTube: OCB: Helm Update with Martin Hickey (IBM) and guests
Description
OpenShift Commons Briefing
Helm Update with Martin Hickey (IBM) and guests
A
B
Hello,
everyone
and
welcome
to
another
openshift
commons
really
excited
today,
we're
changing
it
up
and
we
have
martin
hickey
here
from
ibm
who's
going
to
talk,
give
us
a
helm,
update
and
brad.
Thank
you.
So
much
for
joining,
I
wasn't
sure
if
you'd
be
able
to
join,
brad
is
also
from
ibm
and
we
have
andrew
block
who,
hopefully
most
of
you
know,
he's
a
red
hat
consultant
and
he
wrote
a
book
on
helm,
which
is
nothing
small.
So.
B
Your
name's
on
it
your
name's
on
it
so
today
you
know
since
we're
changing
it
up,
we're
also
going
to
you
know,
throw
away
the
slides,
so
we
don't
have
to
force
you
to
go,
watch
some
slides
and
I'm
going
to
start
with
martin
and
martin.
If
you
want
to
just
kick
it
off
with,
although
we
just
lost
martin
where'd,
you
go
there.
C
D
Thank
you
very
much
for
having
me
on
board
karina
and
to
all
the
the
openshift
comments.
Folks,
yeah
we'd
have
done
a
slide
deck,
but
we're
not
we're
just
going
to
chat.
I
think
it's
bugs.
D
I
think
they're
I'm
going
to
be
put
on
the
spot
here,
because
somebody
has
written
a
book
that
they
want
about
a
topic.
So
I'm
going
to
be
quizzing
andrew
about
stuff
that
I
should
know
about
in
hell.
So.
A
E
D
You
say
thanks
very
much
for
having
me
on
board.
I
my
name
is
martin
hickey.
I
work
for
ibm
in
the
open
technology
group.
I
spend
a
lot
of
time
out
in
open
source
communities
and
I'm
a
maintainer
in
the
helm
community.
So
it's
great
to
get
on
board
here
to
have
a
chat
about
helen.
How
would
you
like
me
to
kick
off
karina
any
topics.
B
D
Okay,
I
think
that's
probably
the
best
place
to
start.
I
suppose
helm
is
it's
a
lot
of
things
to
a
lot
of
people.
If
that
makes
sense,
it
usually
comes
up.
He
usually
gets
the
the
name
of
the
package
manner
for
manager
for
kubernetes,
and
I
suppose
for
me
looking
at
from
from
my
point
of
view
and
from
talking
to
people
over
the
last
few
years
and
also
in
the
community.
D
It
really,
I
just
was
one
of
its
core
things
that
it
always
keeps
on
board
is
allowing
somebody
new
coming
to
kubernetes
to
be
able
to
get
to
know
kubernetes
or
to
be
able
to
install
their
app
in
absent
kubernetes.
I
think
everyone
on
this
call,
even
though
we
work
on
kubernetes
communities
in
different
parts,
kubernetes
is
difficult.
D
It's
big,
it's
large,
its
apis
are
just
increasing
every
every
release,
the
amount
of
resources
that
are
out
there.
So
one
of
the
key
things
we
always
look
at
is:
is
that
person
starting
out
wanting
to
see
how
does
kubernetes
work?
If
I
I
don't
know,
if
I
put
a
mongodb,
if
I
deploy
mongodb
into
into
my
cluster,
what's
going
to
come
up,
can
I
take
a
look
at
what's
it
what's
in
underneath
the
heat
underneath
the
hood?
So
it's
it's
that
simplification.
D
I
think
it
brings
to
start
with
and
that's
on
one
side
and
then,
if
the
other
side
is
the
person,
that's
controlling
the
cluster,
that's
deploying
their
stuff
out
into
their
cluster.
That's!
What's
an
example
I
have
is
lately.
I
was
installing
issue
into
a
cluster
and
I
used
this
to
use
helm
charts
to
do
it
and
I
started
using
helm3
the
latest
version
of
hell
and
I
looked
into
it,
my
goodness
there
must
have
been
10,
10
12,
different
crds.
D
They
were
using
every
object
you
could
think
of
under
the
hood
all
the
different
kubernetes
objects.
They
had
serious
charts.
You
know
what
I
mean
really
really
big
sharks.
They
know
what
I
mean
it
was.
It
was
amazing
to
see
it.
So
these
are
the
kind
of
scales
we
look
at.
The
other
side
of
the
10
is,
is,
I
suppose,
just
two
types
of
of
of
particular.
D
Maybe
users
of
of
helm
what
I
talked
about,
one
is
the
consumer,
so
the
hem
can
nicely
package
up
all
your
manifest
files
and
put
it
into
a
logical
container.
Like
do
you
know
what
it's
similar
to
to
a
folder
or
a
directory
that
we
have
in
an
operating
system
and
that
somebody
can
just
take
that
chart
and
go
handle
install.
D
They
don't
know
they
don't
have
to
know.
What's
in
inside
those
template
files
or
the
manifest
files,
whatever
you
want
to
call
it,
and
they
can
just
install
that
and
then
you
have
the
person
who's
constructing
the
the
chart.
So
they
are
the
person.
Then
that
knows
all
right.
What
are
the
particular
objects?
I
want
to
put
out
there
so
that
I
can
get
that
up
out
there.
So
I
think
that's
what
it
brings
to
a
lot
of
people.
D
You
know
it
can
be
for
the
person
starting
out
or
it
can
be
the
person
that's
really
experienced.
Also
too,
I
think
the
beauty
of
it
is
that
extraction
of
the
configuration
from
the
resources
you
want
to
deploy
or
the
kubernetes
objects
you
want
to
deploy,
and
I
think
the
beauty
to
that
is.
If,
for
example,
you
want
to
deploy,
I
don't
know
a
web,
some
web
server
like
nginx
or
apache,
etc,
and
you
want
to
maybe
deploy
three
versions
of
it
out
into
your
cluster.
D
So
I
think
that's
what
it
gives
you
gives,
that
level
of
abstraction
from
between
the
configuration
and
the
underlying
resource
objects,
and
also
the
ability
for
the
person
starting
out
or
the
newbie,
and
then
the
more
experienced
person
of
of
our
underlying.
I
suppose
cloud
operation
system,
kubernetes.
A
Yeah,
so
from
my
from
my
standpoint,
as
you
mentioned,
martin,
it's
a
package
manager,
you
install
you,
get
a
brand
new
laptop.
It
has
windows,
it
has
an
os
x
that
has
linux
the
first
thing
you're
going
to
do.
Is
you
want
to
install
things?
You
have
kubernetes
it's
vanilla
by
itself
or
if
you
have
an
openshift
container
platform,
it
has
some
services
that
are
coming
with
the
platform.
You
want
to
deploy
your
own
things.
A
A
Oh,
I
can
go
ahead
and
spin
up
a
pod
automatically
and
then
be
able
to
communicate
it
because
it
created
an
ingress
or
a
route
and
open
shift
that
that
that,
to
me
is
the
wow
factor
that
helm
provides,
because
otherwise
you
have
to
the
onboarding
and
ramp
up
period
is
just
too
great
for
many
individuals,
especially
with
kubernetes.
D
Yeah,
I
think
I
think
your
analogy
with
the
with
the
os
is
good,
and
even
if
we
come
into
your
venom,
around
fedora
and
the
red
hat
clients
and
stuff
like
that,
when
you
look
at
it
this
way,
there's
you
know
most
of
us.
If
we
put
when
we
put
an
os
on
our
laptop,
for
example
like
fedora
for
a
long
time
online
was,
you
know,
you're,
normally
just
a
consumer.
You
just
want
to
put
the
packages
out
there.
D
A
And
even
when
you
start
to
talk
about
how
we
go
ahead
and
share
charts,
you
know
think
about
yum
repositories
or
brew
taps
similar
concept
in
helm.
You
can
use
repositories
in
helm
to
be
able
to
share
charts,
and
especially
with
some
of
the
newer
implementations
regarding
the
oci
format.
We're
starting
to
see
docker
registry
or
container
registries
be
able
to
start
supporting
the
storage
of
helm
charts.
D
Yeah-
and
I
think
I
think
that's
great-
you
brought
that
up
andrew,
because
this
is
something
that
came
in
handy
and
we
brought
in
the
concept
of
you,
know,
experimental
features.
So
today
it
was
a
fully
fledged
feature
before
you
saw
it
in
a
particular
release,
but
with
the
idea
of
bringing
in
the
experimental
features
like
they're
doing
in
kubernetes,
it
allowed
us
to
bring
in
the
oci
registry.
D
Now
the
oci
registry
support
is
still
a
work
in
progress,
because
there's
a
lot
of
work
going
on
behind
the
scenes
with
the
oci
community
itself
and
the
concept
here
of
so
at
the
moment
we
still
have.
The
concept
of
here
of
oci
registries
are
for
storing
docker
containers,
but
the
idea
here
going
forward
is
what,
if
we
have
different
media
so,
for
example,
the
helm
sharks,
as
you
mentioned,
or
anything
else,
even
videos,
whatever
you
want
to
put
put
in
there.
D
So
the
idea
here
is,
you
know
the
ability
to
be
able
to
use
something
other
than
repository
and
the
registry.
The
ocr
interest
is
a
great
idea
because
you
already
for
most
people
that
work
out
in
kubernetes
you're
already
going
to
have
your
docker
containers,
your
images
out
there.
So
the
id
here
that
you
know
can
put
your
charts
in
there
as
well
brings
an
extra,
I
suppose,
both
today
or
to
go
to
your
your
your
plate.
A
And
I
work
with
a
lot
of
large
organizations
and
they
really
hate
having
to
add
more
and
more
tools
into
their
infrastructure,
so
being
able
to
reuse
existing
tools
to
be
able
to
take
advantage
of
different
technologies.
Now
brad,
you
haven't
had
a
chance
to
really
speak
your
mind.
Thus
far,
did
you
have
anything
in
particular,
you
want
to
you
know,
talk
about
regarding
helm,.
A
C
You
know
we're
not
going
to
let
him
off
the
witness
chair
yet
so
martin,
can
you
talk
about
the
the
the
the
enhancements
that
were
added
to
helm
version
three
compared
to
version
two,
and
and
did
those
help
reduce
the
the
friction
of
people
being
able
to
use
helm?
How
did
that
make
life
better?
Because
there
were
some
dramatic
changes.
D
Yeah,
I
think
we
before
I
kind
of
answered
that
question.
I
think
I'm
going
to
do
a
small
little
bit
of
history
on
this
to
see.
You
know
how
helmetry
came
about
and
why
penelope
came
about
so
to
start
with,
and
it
would
have
been
before
my
my
time
in
the
community
back
at
ford's
tenant
summit
in
it
was
about
early
2018.
D
D
Taylor
caused
a
lot
of
friction,
particularly
if
you
used
helms
straight
out
of
the
box,
because
taylor
ran
in
god
mode.
So
what
I
mean
in
that
situation
is
it
could
do
anything
he
wanted
in
the
cluster.
So
that
means,
if
you
were
on
any
command,
he
was
allowed
to
do
whatever
he
wanted,
so
that
was
just
a
big,
a
big
no-no
for
for
most
cluster
admins
etc.
D
D
So
if
we
go
back
a
little
bit
to
the
history,
they
know
how
helen
2
came
about
was
that
ended
up
being
a
collaboration
between
the
deus
that
had
that
had
basically
created
helm
and
google
and
google's
deployment
service
okay,
so
what
the
decision
was
made
was
to
bring
both
of
them
together
the
helm
version,
one
which
goes
called
helen
classic,
which
was
just
declined
to
only
architecture
with
the
deployment
servers
and
it
created
a
helm
2,
which
was
your
client
and
your
your
particular
service
in
the
cluster
so
fast
forward,
then
to
helen
tree.
D
There
was
a
big
push
in
the
community
then
to
bring
him
trio
so
getting
back
to
brad's
question
what
what
did
he
bring
to
the
table?
So
I
think
one
of
the
key
messages
coming
out
of
it
was
simple
simplification
and
security.
Okay,
so
the
idea
here
do
we
really
need
a
service
in
the
cluster
or
why
can't
the
helmet
line
talk
directly
to
the
kubernetes
api
service,
so
that
was
one
of
the
first
things
to
happen.
D
The
taking
out
of
of
this
service
tiller
and
what
that
required
then
was
quite
a
large
refactor
underneath
the
hood,
but
what
he
did
was
it
probably
put
the
code
basically
in
half,
okay,
so
removed
a
particular
amount
of
code
that
was
no
longer
necessary
or
deprecated,
and
it
also
brought
the
advantage
then
of
some
issues
that
users
were
having
around
grpc
issues
and
connection
issues
with
the
tiller
service.
D
They
of
course
disappeared,
because
now
you
were
talking
to
the
kubernetes
api
service,
exactly
how
cube
qctl
does
so
in
the
particular
kubernetes
way.
What
that
then
brought
was
the
security
factor,
then
that
kubernetes
provides-
and
this
is
around
your
our
back-
your
user
authentication
authorization
etc-
know
that
capability,
because
it
wasn't
there
when
helm2
was
being
created.
Our
back
was
probably
at
an
embryonic
stage
and
all
the
security
that
was
being
put
into
kubernetes.
It
now
was
able
to
be
used
by
helm.
D
C
Okay
and
you
know,
martin
last
time
we
were
actually
physically
at
a
kubecon
conference.
I
remember
it's
going
to
the
helm
talks
and
there
was
just
huge
rooms
of
people
that
had
interest
in
helm
and
so
there's
that
that
huge
amount
of
of
adoption
of
helm
but
at
the
same
time
we're
seeing
newer
models
for
for
deploying
applications
right.
C
D
Yeah,
so
I
I
think
one
thing
you
touched
on
there
around
the
hell
and
charts
is
very
important.
I
would
say
one
of
the
greatest
achievements
of
hell
has
been
the
helm
chart.
Okay
and
helen
chart
is,
I
don't
know
how
you
describe
it?
It's
it's
kind
of
like
one
of
those
inventions
like
velcro.
You
know
it.
Just
is
simple:
it's
useful
and
you
say
to
yourself
how,
in
the
name
of
goodness,
did
I
not
paint
that
before
that
man
did?
D
D
I
think
he
actually
patented
and
we
didn't
see
ventrac
about
10
years
ago
or
whatever,
but
it's
just
the
way
that
it
allows
you
to
give
a
standard
structure
or
format
of
your
your
app
or
your
apps
that
you
want
to
deploy,
and
I
think
that's
the
beauty
of
it,
and
I-
and
I
give
you
an
example
that
even
in
the
community
today
and
people
are
coming
in
and
asking
questions,
not
everybody
uses
every
aspect
of
the
helm
engine
and
what
I
mean
by
that
is
so
generally.
D
D
So
we
have
you,
get
that
little
10
layer
of
templating
on
top
of
your
kubernetes
manifest
files,
and
that
template
gives
you
this
ability
for
logic,
programming,
logic
in
there,
so
your
riffs
etc.
It
also
pulls
the
configuration
out.
So
what
people
sometimes
do
is
they?
They
run
the
command
and
this
command
came
in
as
a
an
id.
A
few
few
years
ago
by
community
member
and
the
maintainers,
it
was
before
my
time
at
the
time
said.
You
know
what
this
is.
D
So
with
that,
then
you
can
then
pipe
that
into
cubectl,
so
you
could
actually
do
your
install
with
cubectl
instead
of
using
hand
to
do
the
install.
So
hen
gives
you
that
flexibility,
that's
one
of
the
things
I
think
it
is
now.
If
we
go
into,
as
you
say,
you
mentioned,
operators,
etc.
So
we
didn't
get
into
conversation
about
day,
one
day,
two
etc
so
yeah.
When
you
look
at
operators,
operators
bring
a
lot
to
the
table.
D
They
bring
you
that
ability
where
you
can
put
that
complex
business
logic
into
the
operator
and
that
operator
can
manage
the.
I
suppose,
the
complexities
of
an
app
that's
out
there.
So,
for
example,
I
don't
know,
maybe
it's
a
database
that
needs
to
be
running
24
7.,
maybe
somebody
you
know
before
some.
Some
poor
person
had
to
be
up
all
night
in
the
middle
at
night.
D
You
know
making
sure
that
you
know
before
and
devops
that
that
the
particular
the
particular
environment
didn't
go
down,
whereas
now,
with
with
your
operators,
you
get
that
capability
that
it
can
basically
recover
from
that
that
it's
checking
that
database.
It
has
that
logic
in
there
and
even
I've
seen
today,
there
seems
to
be
operators
out
there
that
could
have.
I
don't
know
ten
thousand
lines
of
code
in
it.
You
know
what
this
you
know.
D
This
is
when
you
think
about
it
is
massive,
so
it
has
that
business
logic
into
it
and
where
hell,
my
support
comes
in
then
is,
is,
I
suppose,
around
day,
one
where
it
can
do
those
installations
in
it
can
install
your
your
operators
in
etc.
Like
that,
so
when
I
look
at
it,
I
think
operators
and
helm.
If
you
look
at
it,
you
know
they
can
exist
independently
and
they
can
coexist
together.
D
Okay
and
I
think,
as
these
technologies
evolve
going
forward,
it's
only
going
to
be
for
the
betterment
of
users
and
your
your
operators
and
your
devs
are
devops
people.
So
I
think,
there's
basically
a
great
synergy
there
with
board
of
boarding
technologies,
and
I
look
forward
to
how
they
evolve
as
they're
going
forward.
A
For
me,
one
of
the
you
know,
underutilized
features
of
helm
is
going
to
be
hooks
to
be
able
to
manage
the
lifecycle
of
application.
You
know
we
talked
about
operators,
they
can
really
manage
the
full
life
cycle
of
applications.
I
mean
I
love
operators.
They
are
foundational
to,
like
I
spend
most
of
my
days
in
the
openshift
world,
and
especially
regarding
the
the
operator
sdk
becoming.
You
know,
a
project
adopted
by
the
cncf
we're
going
to
see
even
more
with
operators,
but
those
of
you
who
want
to
just
use
helm
charts.
A
You
can
use
a
full
lifecycle
management
of
applications
using
hooks
to
be
able
to
inject
functionality
at
different
points
in
the
lifecycle
of
an
application
of
a
chart
itself.
Pardon
me
so
if
you
want
to
do
an
upgrade,
you
can
manage
database
upgrades
or
an
application
upgrade
or
perform
tasks
or
different
functions
at
different
points
in
time.
That's
another
big
benefit
of
helm
that
I've
seen
very
used
by
my
customers.
Quite
a
bit.
D
Yeah,
that's
actually
andrew
I'm
going
to
pick
up
another
one
if
that's
okay,
because
this
is
a
very
interesting
thing
what
you
said
and
I'm
going
to
go
back
to
it
again
help
settle
with
the
idea
of
being
able
to
install
your
app
into
the
into
the
cluster
in
in
the
time
that
it
came.
I
suppose
it
started
around
the
time
kubernetes
started.
It
has
always
wanted
to
keep
that
core
basic
model,
being
a
user
being
able
to
install
the
app
into
the
into
the
cluster
in
time.
D
Lots
of
you
know
great
new
advancements
from
coming
kubernetes
I
bring
in
the
great
crds
customer
sports
definitions
that
still
are
trying
to
be.
You
know
understood
properly
in
the
kubernetes
community,
because
there's
still
a
lot
of
work
to
do
with
them,
and
what
we've
found
over
time
is
that
you
can
make
helen,
do
so
many
things,
especially
on
hooks,
crds,
etc.
D
But
there's
certain
capabilities
that
that
helm
just
is
not
able
to
support
all
the
time.
And
we
always
talk
about
that
and
say,
and
it's
something
we
did
with
helm3
with
crds
is
that
we
removed
the
ability
to
update
and
delete
crds
in
helm.
So
in
helm3
now
it
will
install
crds
if
that
crd
is
not
already
in
the
cluster.
D
But
it's
not
going
to
do
the
the
rest
of
the
crud,
the
update
or
delete
on
it,
and
the
reason
behind
that
is,
I
suppose,
we're
trying
to
protect
people
that
maybe
are
using
helm,
but
maybe
don't
have
that
advanced
understanding
of
the
of
kubernetes,
as
you
mentioned,
andrew
with
some
of
your
customers
and
yourself,
and
what
we
mean
by
that
is
that
if
we
allowed
updates
and
deletes
of
crds
because
crds
are
cluster
wide-
and
I
ucrd
is
an
example
because
they're
cluster
wide
you
don't
know
what
app
is
using,
that
particular
crt.
D
D
But
I
I
don't
realize
then
that
somebody
else
in
a
different
if
it's
a
multi-tenant
cluster
may
also
then
use
that
crd
and
they
install
their
app
b
and
then
suddenly
I
say:
oh
I'm
finishing
up
with
my
app
and
I
go
delete
of
my
app
before
it
allowed
you
to
a
helm,
would
delete
the
crd
as
well,
but
you
can
imagine
what
happens.
D
So
there's
a
lot
of
things
like
that
that
we
get
requests
into
the
community
and
we
have
to
just
say
sorry,
look,
you
know
we
don't
we're
not
going
to
put
that
in
because
of
of
the
dangers
to
people
and
also
too,
we
found
as
well
at
conferences
like
people
come
up
to
us
with
amazing
things,
they're
doing
with
hell
and
we're
going.
Oh,
my
goodness,
you
know
what
I
mean
like
this
is
you
know
this
is
advanced
stuff
and
we're
kind
of
going.
We
wouldn't
even
have
thought
of
that
so
yeah.
B
A
I
can't
I
can't
say
I
have
any
favorites,
it's
the
ones
that
implement
certain
patterns
that
I
tend
to
focus
on
and
ones
that
can
be
reusable
and
then
there's
kind
of
this
concept
of
library
charts
the
idea
to
integrate
certain
functions
across
multiple
charts.
So
I
can
create
a
reusable
set
of
components
that
can
be
then
reused
in
other
charts.
That's
one
of
the
one
of
the
areas
that
was
brought
into
for
helm3.
Is
this
concept
of
library
charts?
A
D
Yeah
andrew,
that's
a
it's
a
funny
one,
because
library
charts
was
in
helm2
for
your
fyi,
and
it's
just
that
we
made
it
a
first-class
citizen
in
in
helm3
and
he
got
over.
The
idea
of
that.
You
know
to
think
about
a
helen
chart
is,
it
looks
like
a
chart
but
as
as
andrew
says,
what
it's
doing
is
it's
giving
you
common
common
code
that
you
can
use
across
charts
it's
following
the
dry
pattern,
the
reuse
pattern.
That's
that's!
D
Coming
out
there
and
one
of
the
key
parts
is
it's
not
renderable
so,
which
means
we
usually
take
a
chats.
They
render
down
and
you
can
install
them
into
the
cluster.
You
cannot
install
library
tracking
cluster.
So
what
we
want
to
do
in
hand
tree
was
just
putting
the
capability
into
the
core
because
we're
very
cognizant
of
backwards
compatibility.
D
D
Compatibility
with
some
versioning
so
when
we
went
to
the
when
we
basically
when
we
brought
that
capability
in
it
would
say
right
we're
going
to
put
the
logic
in
here
now
that
if
somebody
tries
to
install
an
upgrade
chart,
it's
just
basically
going
to
say
you
know
it's
an
overcharge.
You
can't
install
it
and
also
what
it
did
load
us,
because
we
didn't
want
to
break
the
back
as
competitive
it
is.
D
We
have
a
new
field
called
tight
inside
in
the
chat
yaml
and
that
type
field
defaults
to
application,
which
was
the
way
charts,
always
where
in
helm,
one
and
helen
two
that
their
application
that
can
be
installed
into
cluster,
and
now
we
have
the
one
called
library.
So
it
also
gives
us
that
capability
is,
if
we
add
more
logic
and
more
types
going
forward.
We
can
extend.
B
Thank
you
and
we'll
keep
going
sad
to
see
andre
leave,
but
we
still
have
austin
here.
I
don't
know
if
you
wanted
to
join
but
aj.
I
do
want
to
address
your
comment.
Aj's
asking
about
helm
in
openshift
and
red
hat,
so
helm
became
a
first
class
citizen
right
with
openshift
four,
and
it
was
tech
preview
in
4-3
and
now
with
4-4.
B
It
is,
you
know,
ga,
with
an
open
shift,
so
you'll
see
more
and
more
support,
but
getting
back
to
you
know
core
helm,
and
you
mentioned
so
martin,
you
mentioned
backward
compatibility.
B
D
Sure,
but
you
know
before
I
go
on
here,
there's
no
such
thing
as
a
free
dinner,
so
I
actually,
I
have
a
question
for
for
you
folks
and
maybe
it's
for
adrian,
I'm
not
sure.
So
I
I
thought
it
was
brilliant
that
she
made
he
made
hell
first
class
citizen
in.
I
think
it
was
four
to
four:
was
it
an
openshift
four
or
four?
D
B
D
B
A
new
gonna
see
who
I
wanted
to
pick
on,
I
don't
know
shubik
or
pre-drive.
If
you
wanted
to
jump
in,
I
know
you've
worked
it.
F
A
good
job
there,
so
I
think
marginal
question
is:
why
did
we
make
him
to
your
first
class
citizen
right
yeah?
So
I
think
from
a
technical
perspective,
I
was
very
happy
that
we
didn't
have
to
have
our
sre
teams
with
openshift,
dedicated,
etc,
have
troubles
with
filler.
To
be
honest
because
we
could
fall
back
on
the
you
know,
out
of
the
box,
our
back
model
and
open
shift
for
me
technically.
F
I
love
that
from
an
engineering
perspective,
I
found
it
supportable,
that's
one
and
then
from
a
user-based
perspective,
everybody
loves
charts
hand,
I
think,
from
a
user
perspective
that
was
there
so
because
users
loved
it
and
because,
from
an
entering
perspective,
we
could
support
it
see
mike,
and
I
and
the
different
folks
got
together
to
say:
hey,
we
have
all
the
pieces
in
there.
Let's
ensure
it
works
with
openshift
specific
apis.
F
We
haven't
been
confident
deployment,
config
and
ship
it
and
that's
what
we
did
and
then
the
next
step
was
that:
how
do
we
ensure
that
we
can
make
give
developers
a
good
experience
on
our
developer
perspective
in
console,
which
is
where
a
lot
of
work
actually
maintained
to
ensure
that
it's
awesome?
That
on
the
helm,
is
an
icli
and
we
also
shipped
a
build
of
that.
But
then
the
goal
was
to
also
ensure
that
if
somebody
is
a
ui
user
and
and
just
wants
to
visualize
how
to
you
know,
install
rollback
upgrade.
F
That
was
there
too,
and
we
saw
that
it
did
fit
into
our
developer
perspective
model
that
we
that
our
team
was
investing.
So
that's
how
the
whole
decision
about
going
with
him,
ken
and
martin.
D
Okay,
that's
brilliant!
It's
good
that
there's
that
came
aboard
I
was.
I
was
quite
happy
when
I
saw
that.
C
Yep,
so
so
I
I
have
a
follow-up
question
are:
are
we
seeing
now
that's
a
first-class
citizen
in
open
shift
4-4?
Are
we
seeing
requirements
for
helm
coming
from
from
openshift,
or
is
it
pretty
much?
No,
it
works
perfect,
as
is,
and
we
just
needed
to
make
it
a
first
class
citizen.
F
Yeah
good
good
question,
so
I
think
from
the
core
functionality
perspective
that
works,
as
is
there
are
some
things
that
we
have
proposed
upstream
and
we've
also
included
in
our
upcoming
release.
Is
that
the
ability
to
define
a
repo
configuration
at
the
on
the
cluster
today
you
can
do
a
you
know.
You
can
add
a
new
help,
repo
on
your
cli.
There
is
no
equivalent
of
that
at
a
cluster
level
for
an
admin
to
maintain
that's
something
that
we've
proposed
upstream.
F
We've
actually
gone
through
an
openshift
enhancement
proposal
process,
which
is
a
very
open
process
that
we
have
internally,
but
it's
public
and
we've
actually
exposed
the
capability
on
the
developer
console
perspective
behind
the
scenes.
It
uses
a
new
crd,
but
the
user
doesn't
have
to
know
that.
But
then
we've
tried
to
figure
out.
F
How
do
we
gradually
build
out
a
more
cluster
experience
around
some
of
the
things
which
were
meant
to
live
in
the
cli
forever
and
then
we'll
see
how
to
best
work
out
with
upstream
and
ensure
that
align
with
the
entire
ecosystem,
better
so
from
a
core
engine
perspective.
Things
work
amazing
from
a
few
experienced
thing
from
an
experience
perspectives,
we
are
adding
new
things.
D
Okay,
that's
that's
great
and
I've
seen
some
of
those
I've
seen
some
of
those
proposals
upstream
as
well,
so
we
we
look
forward
to
more
more
input
from
from
you
and
your
folks.
B
And
he
is
on
this
call,
I
don't
know
if
you
wanted
to
jump
in
too,
but
you
know
we
have
lots
of
helm
people
here
today.
It's
awesome.
D
B
Some
people
are
struggling
still
with
the
migration
from
two
to
three
and
just
wondering
if
you
could
talk
through
that,
I
know
you
were
instrumental
in
you
know
the
migration
toolkit
or
the
migration
tooling,
I
mean
so
if
you
want
to
talk
through
kind
of
what
you
did
and
why.
D
Okay,
I
think
yeah,
I
think,
that'd
be
a
good
place
to
go
so
to
just
kind
of
do
a
high-level
view
of
when,
when
we
talked
about
some
of
the
capabilities
that
went
in
for
and
helen
v3,
we
should
also
be
that
this
required
underneath
the
hood.
D
I
suppose
it
was
nearly
like
a
react:
re-architecture,
a
refactor
of
the
plumbing
and
all
the
underlying
way.
We
store
data
and
so
forth
when
we
made
that
change,
so
what
it
meant
was
and
two
of
the
key
features.
So
that
would
be
how
we
stored
our
local
configuration.
So
what
I
mean
by
that
is
your
plugin
information,
your
local
repository
list
etc
and
the
format
it
was
stored
and
how
it
was
stored
locally
and
then
also
the
release
information.
D
So
the
release
information
is
is
very
important
because
in
case
people
don't
know
the
release.
Information
is
when
you
do
your
home
ls
and
you
see
the
different
apps,
you've
deployed
or
different
charts,
you've
deployed.
It
shows
that
release
information
gets
stored
in
the
cluster.
So
what
we
did
with
that
tool
was
we
made
quite
a
bit
of
a
change
to
the
objects
under
the
hood
that's
stored,
so
what
it
meant
was.
There
was
no
longer
a
one-to-one
mapping
between
the
local
state
information
and
then
the
release
metadata
information
between
two
and
three.
D
So
today,
if
anyone
is,
you
know,
wants
to
play
along
with
helen,
you
can
have
him
too
and
helen
tree
managing
the
same
cluster,
okay,
managing
wraps
in
the
same
cluster,
because
they
are,
as
for
once,
a
bit
more
independent,
really
of
each
other.
Okay,
so
you're
not
going
to
walk
up
on
top
of
each
other
on
it.
So
what
that
means
to
us
was
we
needed
migration
here?
So
this
wasn't
a
simple
case
of
pull
down
your
binary,
get
rid
of
tiller
and
away
you
go
because
what
it
meant
was.
D
How
would
you
manage
your
previous
releases
that
you
deployed
with
helm,
and
I
don't
think
anyone
here
wants
to
just
dump
the
releases
that
they've
that
they've
deployed?
I
think
they
still
want
to
manage
them.
So
we
we
came
up
with
an
id
over.
I
suppose,
as
hentry
was
was
moving
along
and
and
this
issue
actually
came
up
at
kubecon
2018.
D
I
think
that
was
around
december.
I
think
that
here
and
somebody
brought
it
open
the
deep
dive.
We
were
having
a
conversation
afterwards
in
the
deep
dive
and
they
brought
it
up
and
it's
it
basically
put
in
as
an
issue
for
community
and
we
worked
on
it
then
so
the
way
we
worked
around
it
was.
We
decided
that
right.
D
We
need
documentation,
definitely
to
explain
it
because
there's
two
ways
you
can
do
this,
you
can
manage
the
migration
yourself
and
what
I
mean
by
that
is
you
decide
that,
but
every
new
release
I'm
going
to
deploy
I'll
use
helen
tree
and
in
the
previous
ones,
in
helm2
I'll
leave
it
there
and
I'll
manage
it
and
over
time,
then,
once
that
I
no
longer
need
I
delete
them,
and
eventually
I
come
to
a
day
where
you
know
what
the
amount
of
releases
I've
been
having
to
imagine
are
so
minuscule
now
I
can
get
rid
of
him
too.
D
Okay,
but
then
we
wanted
an
automatic
automated
way
of
doing
it,
because
some
people
want
would
want
to
migrate
across
in
situ
I.e,
bring
their
releases
across,
bring
the
configuration
across
and
start
managing
straight
away
with
helm3.
So
we
we
developed
the
plug-in
for
it
and
it's
called
the
hem23
plug-in
and
what
it
does
is
it
it
migrates
across
your
configuration
and
it
migrates
across
releases
and
what's
nice
about
this
is
when
we
started
initially
we
so
there's
two
nice
fits
to.
It
is
when
you're
migrating
across
your
release.
D
So
what
I
mean
by
revisions
is
that
every
time
you
upgrade
your
your
particular
release,
it
gets
a
new
revision,
so
some
people
might
even
have
100
revisions
of
a
particular
release,
which
is
probably
you
know
it's
going
to
fill
up
your
cluster,
but
so
what
it
means
you
can
see
what
you're
going
to
bring
across
you
can
tell
you,
can
test
it
out
and
then
you
can
bring
it
across
and
then
one
by
one.
D
Then
you
can
delete
those
releases
that
helm
two
was
managing
and
when
I
mean
delete
by
that,
it's
just
delete
into
helm2
metadata.
It
does
not
touch
the
helm3
metadata
or
the
deployment
objects.
The
resources
you've
deployed
into
the
cluster,
so
the
migration
doesn't
go
touch
doors,
it's
just
touching
the
metadata
itself.
D
D
We're
currently
not
accepting
any
bug,
changes,
no
bug
fixes
and
we're
only
looking
at
security
fixes
between
now
and
over.
So
it's
probably
a
good
time
to
check
out
the
migration
and
and
get
going
with
it.
B
E
Hey
yeah
sure
karina
you
know.
Most
of
my
customers
started
right
at
home.
Three,
I
think
they
were
kind
of
waiting
to.
You
know
wait
for
the
project
to
incubate
fully
and,
of
course,
now
it's
a
great
now.
It's
a
graduated
project,
actually
martin,
one
of
the
questions
that
I
had
if
I
could
just
kind
of
pivot,
so
I've
been
working
with
helm
for
the
past.
I
want
to
say
a
year
and
a
half
since
I
probably
home
2.7,
2.8
and
customers.
E
F
D
Yeah,
that's
a
good
question
andrew
because
it
it
does
come
up
a
bit.
I
think
the
thing
with
this
is.
D
We
try
to
avoid
bloating
the
capability
of
help
and
it
kind
of
goes
back
to
what
I
mentioned
at
the
start.
Is
we
really
want
to
keep
him
for
the
user
who's
starting
out
primarily
and
the
ability
to
do
those
things
without
you
know
over
complicating
health,
but
while
we've
allowed
with
helm,
is
the
ability
to
expand
health
so
by
putting
the
plug-in
framework
in
place?
D
D
So,
for
example,
the
two
to
three
plugin
gets
run
as
helm,
two
to
three
like
that,
so
it's
been
run
by
helm
itself
and
around
the
sequences,
there's
so
many
ways
to
do
that
that
we've
we've
never
taken
that
capability
in
or
we
found
capability
that
would
stand
alone
or
work
well
inside
the
core
base.
D
Okay,
so
that's
probably
one
of
the
reasons
why
we
do
it
and
then
there's
been
a
series
of
plugins
after
that
which,
with
capability
like
this,
we're
not
going
to
we,
we
see
that
may
not
work
inside
in
the
core
we.
This
is
where
also
the
use
of
the
go
sdk
comes
in
so
the
way
hand
works.
D
Is
you
have
your
home
cli
like
a
lot
of
go,
go
projects,
your
health,
cli
code
or
your
command
code,
and
then,
basically
you
have
the
library
under
the
hood,
which
is
inside
the
package,
so
that
library
is
publicly
available.
The
the
go
docs
are
out
there,
so
helmet
strength
itself
uses
that,
and
probably
for
a
lot
of
extensions
and
stuff,
like
that,
we
say
that
to
customers
as
well,
and
there
are
a
lot
of
customers
that
are
leveraging,
help
that
write
their
own
clients
or
their
own
extension
clients.
D
E
D
E
No,
I
think
that's
great,
I
mean
the
good
thing
about
helm.
Is
that
it's
so
flexible?
You
can
pretty
much
do
whatever
it
is
that
you
need
to,
for
example,
on
the
topic
of
secrets,
what
I've,
what
I've
done
with
my
past
couple
of
customers
is
build
in
a
vault
feature.
Just
using
you
know,
just
regular
home
capabilities.
You
know
build
in
a
bullet
feature
to
template
out
a
in
a
container
and
a
side
car
that
uses
vault
agent
and
then
use
that
mechanism
to
pull
your
secrets
in.
D
Yeah,
because
and
when
you
talk
about
the
extensions
like
that,
it's
it's
amazing
stuff
comes
in
from
the
community,
but
you
know
so
when
they
come
in
really
passionate
about
the
thing
they
found
or
the
thing
they're
trying
to
work
on
and
like
we
love
that
passion,
but
sometimes
we
have
to
kind
of,
I
suppose
fan.
We
have
to
cool
that
passion
down
a
bit.
D
You
know
because
because
maybe
you
as
a
customer
or
someone
else
really
has
this
id,
but
that
is
that
age
case
that
if
you
put
it
in,
if
you
start
putting
into
a
into
a
product,
it
will
just
blow
up
the
product
and
and
and
goes
away
from
the
core
capabilities
of
products.
So
you
know
as
part
of
the
two
to
three
plugin.
D
It
is
the
simplicity
of
it
where
you
decide
the
particular
tailor
instance
or
the
particular
release
you
want
to
to
you
protect
you,
you
specify
the
release
that
you
want
to
convert
over
okay.
Now
someone
came
to
me
one
day
and
I
went
you
know
what.
Why
don't
you
get
all
the
releases
that
are
in
the
cluster?
You
know,
and
I
went
okay,
that's
a
really
good
idea,
but
the
code
base
is
just
going
to
go
bananas
in
this
okay.
D
So
what
came
back
of
it
in
the
end
is?
Is
a
script
and
someone
came
in
with
you
know.
I
don't
know
who
came
in,
but
they
said
why?
Don't
you
just
script
it
so
just
you
know,
use
helmets
or
cube
cdl
and
find
out
what's
there
and
then
just
point
that
into
it
and
then
that
went
into
the
documentation.
Now
some
people
will
look
at
that
and
they
say
you
know
that's
a
fudge.
D
You're
fudging
me
here,
but
it's
not
it's!
It's
that
balance
between
keeping
your
product.
You
know,
concise
and,
I
suppose
mobile
against
making
it.
You
know
like
a
tanker
on
fire.
You
know
that's
been
attacked
by
pirates.
You
know
what
I
mean
that
kind
of
a
thing
it
just
becomes
too
top-heavy
eventually
so
yeah.
B
D
I
think
I
think
you've
lost
four
didn't
the
ball
didn't
fall
into
four
okay,
that's
a
very
good
question
yeah
and
maybe
we're
still
a
or
we're
still
recovering
from
tree
getting
out
the
door,
which
is
great
at
the
moment.
There's
nothing
concrete
on
four.
We
haven't
exactly
talked
about
it
or
or
come
up
with
any
concrete
ids.
So
that's
a
straight-up
answer
for
you,
but
I
will
say
because
we
keep
the
same
version
and
so
so
tight
and
we
don't
like
to
break
or
break
backwards
compatibility.
D
We
are
seeing
features
at
the
moment
that
are
coming
in
going.
You
know
what
that's
going
to
break
backwards,
compatibility
and
we're
putting
four
lx
tags
on
okay,
so
what's
happening
in
the
moment
is,
is
what
we
call.
I
suppose
a
natural
buildup
of
features
are
building
up
slowly
in
the
four
bucket
that
doesn't
exist
at
the
moment
in
our
own
minds.
So
I
think
I
suppose
this
is
kind
of
a
call
out
to
folks.
D
You
know
if
you
have
ids
here,
if
you
think
you'd
like
to
know
where
you
want
helen
to
go
or
what
what's
going
to
be
held
next,
which
is
this
was
held
for,
as
you
say,
it'd
be
a
good
time
to
come
in.
So
that's
probably
an
answer
that
tells
you
nothing
will
tell
you
everything.
At
the
same
time,.
D
Okay,
so
the
key
place
and
that's
what
everyone
says
it's
good
to
get
out
so
so
helm
has
an
org
under
github,
github.com
and
underhand
there's
a
series
of
yeah
it's
best
to
describe
the
snow
because
it's
confusing
when
you
first
go
in,
but
when
you've
been
in
the
kubernetes
ecosystem.
You
understand
this,
but
so
underhand.
Then
there's
a
series
of
different
projects,
helen
being
the
core
engine
chat's
been
where
the
charts
are
this
track
museum,
even
the
223
plugin
is
in
there
etc.
D
So
there's
a
series
of
projects
saying
underneath
that
so
I'd
say
first
of
all:
go
there
we'll
also
go
to
the
hem
community
community
page,
which
is
I'm
going
to
say
hello,
but
I'm
going
to
put
it
in
here
because
for
some
reason
something's
telling
me
I
might
be
wrong.
No
I'm
right
hello!
D
Let
me
just
put
it
in
last
thing:
we'll
do
is
give
out
fake
and
somebody
loses
their
job
because
it
goes
to
some
sort.
It
shouldn't
have
gone
to
so
no
it's
a
helm,
dsh
and
that
that
got
a
really
good
revamp
in
in
the
last
couple
of
months.
So
it's
it's!
Quite
it's
quite
nice.
Now
there's
a
lot
of
content
in
there.
So
once
you
got
in
there,
you'll
get
a
link
to
anything
from
there
from
the
docs
et
cetera,
like
that.
Well,.
D
And
also
sorry,
the
staff
channel
we're
under
the
kubernetes
slack
workstation
workspace
and
we
have
two
channels:
helm,
dev
or
three
channels:
hemdev,
helm,
users
and
charts.
So
if
you're,
asking
probably
open
kind
of,
I
suppose
open
questions
go
to
helm
users
and
if
it's
a
dev
specific
thing
go
to
help.
Dave
and
charts
is
obviously
charts.
B
F
B
D
A
lot
of
great
work
done
on
the
page
by
a
person
called
roland
flynn.
He
did
some
great
work
on
that
and
a
lot
of
community
members
and
maintainers
are
putting
input
as
well.
D
You
much
every
everything
should
be
there
linked
there
and
if
you
find
something
that
isn't
yeah,
please
please
raise
an
issue
against
it
or
push
a.
B
Pr,
I
know
we
only
have
a
couple
of
minutes
left.
Do
you
want
to
mention?
Did
you
I
know,
there's
been
a
lot
of
translation
work
and
that's
something
that
openshift
has
also
been
pushing
and
console.
Is
language
work
thanks
brad?
D
Yeah
that'd
be
good
yeah
because
there's
a
lot
of
nice
work
done
by
community
members
and
they're
doing
translation
work
around
our
documentation
at
the
moment,
so
big
shout
out
to
the
people
that
have
been
involved,
the
community
members
they're
translating
into
korean
japanese
and
chinese.
So
that's
ongoing
work-
that's
happening
bit
by
bit
by
bit
because
it
takes
a
bit
of
work
to
get
through
some
of
those
pages.
So
that's
great!
So
I
suppose
it
kind
of
shows
to
the
the
level
of
interest.
D
That's
out
there,
that
you
usually
see
that
in
the
project
when
when
the
uptake
is
is
is
is,
is
there
is
that
people
like
to
get
out
there
and
translate
into
their
different
languages?
So
it's
it's
lovely
to
see
that
I
will
not
be
translating
into
irish
by
the
way.
So
just
in
case
we
we're
just
too
small
of
them
of
a
number
of
people.
For
that.
D
B
B
D
Yeah,
it
was
lovely
to
have
people
jumping
in
asking
questions
and
things
like
that.
You
know
because,
especially
you
know,
people
put
me
on
the
spot
as
well.
Oh
answer
that
question
and
I
gotta
go
yeah.
I
don't
know,
but
no,
hopefully
it's
giving
people
some
insight
into
it.
You
know
on
and
all
your
listeners
and
so
forth,
so
yeah.
Hopefully
it's
been
of
value
to
people,
but
thank.