►
From YouTube: OperatorHub Update with Daniel Messer (RedHat) from Operator Framework SIG June 18 2019
Description
Community OperatorHub Update
Daniel Messer (RedHat)
Operator Framework SIG June 18 2019
A
Okay,
all
right
so
for
those
of
you
who
don't
know
me
on
a
call
and
those
who
see
the
recording
later
and
Daniel
Messer
and
the
product
manager
in
the
old
shift
business
unit
at
Red,
Hat
and
I'm,
looking
over
parts
of
the
operator
framework,
specifically
the
SDK
and
the
operator
lifecycle
manager,
so
also
as
a
little
side.
Project
of
mine
I
do
see
and
look
after
operator
hub
that
IO,
which
is
our
community
registry
for
operators
targeting
the
water
company.
This
audience
and
we've
seen
quite
some
traction.
A
Actually,
since
we've
learnt
this
just
as
of
today,
we
actually
crossed
50
operators
on
that
particular
website,
which
is
pretty
amazing
and
we've
got
a
couple
of
interesting
new
additions
which
I'm
going
to
share
with
you
here
in
a
second
we're
also
going
to
talk
about
a
couple
of
new
features
that
we
have
already
released
on
the
side
and
that
we
are
planning
to
publish
in
the
next
one
or
two
months,
which
should
hopefully
make
the
process
of
packaging.
Your
operator
should
you've
already
started
to
write
one
or
like
update
your
system
operator.
A
Much
easier
parts
of
the
slide
did
not
try
any
loading
or
show
anyway.
Sometimes
Google
doesn't
like
me,
so
yeah
no
icons
for
us
today,
that's
kind
of
weird,
but
that
totally
looked
different
in
Google
slides
when
I
edited
it,
but
we
can
just
take
a
look
at
the
side
itself
right.
So
now
we
have.
We
have
a
couple
of
new
interesting
additions
from
across
the
board.
One,
a
couple
of
ones
that
I
wanted
to
pick
out.
Not
all
of
them
is
definitely
Luke.
Luke
is
like
an
overarching
storage
project
on
kubernetes.
A
So
that's
basically
running
your
self
defense
torch
on
kubernetes
itself,
and
there
are
multiple
flavors.
If
you
will,
one
of
which
is,
and
probably
most
widely
known
so
far
is
rousseff.
So
we've
added
group
self
to
operate
a
half
that
IO
has
been
packaged
and
has
been
tested
with
or
operated,
lifecycle
manager,
mini
cube
and
various
other
kubernetes
flavors,
and
you
also
see
next
to
it.
Luke
H
of
s,
which
is
another
flavor
of
rock,
so
rock,
is
just
really
at
a
generic
platform
for
managing.
A
Such
a
platform
in
a
consistent
and
production-grade
fashion
has
has
that
built
in
so
for
some
who
have
maybe
been
around
a
bit
longer.
You
probably
have
heard
around
container
native
storage
in
the
freedom
for
freedom,
5
timeframe,
where
we
had
this
like
little
component
called
Hikari.
That
would
usually
sit
outside
on
sometimes
inside
of
the
cluster,
and
would
you
know
do
that?
A
Do
that
sort
of
reasoning
for
you,
given
that
you
had
already
set
up
a
in
this
case,
cluster
deployment
right,
and
that
was
basically
what
it
did.
It
didn't
quite
look
after
the
lifecycle
of
the
cluster
installation
itself,
so
you
still
need
to
know.
You
know
how
cluster
works
and
you
know
all
the
kings
and
details
around
how
to
do
that
on
kubernetes
and
open
shares
and
hey
Kelly
would
be
there
to
basically
provide
an
API
for
cluster
now.
A
Ruk
completes
that
picture
in
a
much
more
realistic
sense
and
basically
provides
a
storage
lifecycle,
management
and
storage
management
on
the
on
the
platform,
and
it
does
it
in
a
generic
enough
way
so
that
it's
not
mostly
it's
not
only
chef
who
is
able
to
use,
make
use
of
that,
but
also
HFS,
which
is
another
highly
efficient,
distributed
storage
layer
on
kubernetes.
We
have
another
storage
solution.
A
You
see
sources
kind
of
a
reoccurring
theme
here
in
there,
which
is
port
works
which,
which
has
been
out
there
for
quite
a
while,
but
nowadays
at
least
the
operator
and
they've
actually
also
tested
and
certified
it
with
with
openshift
and
as
part
of
that,
they
also
provided
an
upstream
version
of
that
for
penguin
and
its
usage.
So
port
works
very
similar
to
two
rooks
F.
Basically
contain
arises.
The
storage
engine
itself
runs
at
the
ports
on
kubernetes
uses.
A
Another
thing
that
I
wanted
to
point
out
is
is
down
here.
It's
a
spinnaker
operator
by
biopsy
MX.
So
there
are,
there
is
actually
an
existing
spinning
or
operator.
That's
the
optimum
excursion
that
provides
some
added
value
and
and
some
additional
features
that
you
don't
get
with
the
you
know,
pin
the
distribution
of
spinnaker
itself,
but
a
spinnaker
is
quite
a
popular
see,
a
CD
or
a
specifically
CD
solution
for
for
kubernetes,
because
it
has
been
written
with
kubernetes
in
mind.
So
it's
very
easy
to
you
know:
implement
continuous
delivery
on
on
the
cluster.
A
With
that
and
with
the
offering
you
now
have
the
ability
to
run
the
spinning
or
operator
on
the
cluster
itself,
and
you
don't
need
to
be
an
expert
in
how
to
set
up
spinnaker,
which
can
be
quite
involved,
especially
given
all
the
configuration
options
that
you
are
allowed
to
choose
from.
You
would
just
use
the
operators
to
do
that
and
you
get
a
fully
featured.
Could
this
delivery
version
a
continuous
delivery
solution
on
kubernetes,
which
would
allow
you
to
do
things
like
cannery
deployments
and
automated
canary
deployment
so
pretty
exciting?
A
One
thing
we
were
particularly
proud
of
because
we
were
together
on
this-
is
still
Verdes
of
the
elastic
cloud
operator
on
both
OpenShift
and
kubernetes.
So
you
are
now
able
to
basically
get
your
elasticsearch
and
Cabana
instances
out
as
a
matter
of
just
you
know,
creating
another
object,
so
the
elastic
search,
not
one
operator,
hop
that
I
own
as
well,
is
packaged
and
ready
to
go
wherever
operator
lifecycle
manager
is
as
present,
there
are
some
more
exciting
ones
that
I
didn't
even
know
existed
before.
A
So
there
is
something
called
FF
DL,
which
is
an
operator
done
by
IBM,
which
is
a
fabric
for
deep
learning.
So
that's
something
that
I've
been
meaning
to
try
for
some
time
now,
but
it
didn't
quite
find
the
slots
in
light
and
I
had
my
dad
looks
very
promising
on
hot
walk
straight
and
install
FF
DL
on
kubernetes.
So
that's
like
a
collection
of
deep
learning
frameworks
right
there
for
you
to
leverage
kubernetes
ability
to
you
know
quickly.
A
Schedule
worked
out,
so
it
cannot
work
towards
as
part
of
the
that
model,
training
and
then
also
later
on.
The
serving
one
thing
that
was
featured
at
koukin
in
its
own
session
was
actually
aqua
security.
So,
if
you
ever
wanted
to,
like
you
know,
do
little
pen
testing
on
your
own
cluster.
So
don't
do
this
with
your
coworkers
cluster
or
your
production
cluster.
But
if
you
wanted
to
do
pen
testing
against
the
kubernetes
cluster
and
test
it
against
a
variety
of
known
best
practices
and
and.
A
Security
precautions,
you
could
do
that
so
aqua
basically
gives
you
a
set
of
tests
that
it
will
run
against
your
cluster
and
it
will
check
for
API
availability.
You
know:
unsecured,
endpoints,
unauthenticated,
endpoints,
missing,
SSL
encryption,
all
these
kind
of
things
and
yeah
it's
it's
available
as
an
operator,
and
you
can
quickly
install
that
on
your
cluster
and
get
a
report
of
which
kind
of
security
vulnerabilities
or
like
missing
best
practices
or
the
same
were
found
on
your
cluster.
So
pretty
exciting.
A
I
think
the
last
one
I
wanted
to
quickly
highlight
was
also
stream,
Z,
Kafka
I.
Think
since
we
last
spoke,
it
was
already
in
the
queue
and
available
no
beef,
but
not
on
operator
half
so
stream.
Z
is
an
opinionated
wager
on
Kafka
clusters
on
kubernetes
or
open
ships,
and
that
operator
basically
gives
you
the
plain,
vanilla,
edgy
version
available
on
your
site
or
on
your
on
your
cluster
and
as
you
can
see,
you
can
be
pretty
granular
of
the
resources
that
they
wanted
to
control
right.
A
So
even
things
like
topics
or
users
or
connectors
are
are
exported
next
post
by
this
operator
and
allows
you
to
run
things
like
Kafka
at
scale
at
the
scale
of
your
cluster.
This
is
what
I
kind
of
wanted
to
highlight
as
stuff
that
has
come
in
we
either
recently
or
some
time
ago
already.
Fortunately,
we
have
seen
a
quite
an
amount
of
uptake
in
terms
of
these
submissions.
A
On
the
other
hand,
we
weren't
quite
always
able
to
you,
know,
process
and
review
them
in
time,
which
is
why
we
spend
some
effort
actually
on
streamlining
this
particular
process
a
little
bit.
So
if
you
now
open
a
pull
request
on
your
on
operate
up
at
I/o,
you
would
actually
get
to
see
like
a
little
checklist
of
things
that
you
should
consider
and
think
about
and
check
off
before
you
kind
of
submit
your
PR.
A
So
these
are
basically
general
things
of
like
leading
our
contribution
guidelines,
which
we
have
now
put
in
place
where
we
basically
explain
to
you.
We
are
to
contribute
what
and
how
to
package
you
operator
what
that
bundle
format
is
all
about
and
what
you
need
to
put
in
there
in
order
to
make
all
em
happy
and
then
also
what
you
contribute
in
which
directory
and
what
platform
you're.
Targeting
with
that
right
and
then
we've
also
revamped
our
testing
Docs.
B
I
ask
a
quick
question:
I'm
this
is
Andre,
so
one
thing
we
were
discussing
actually
just
this
morning
is
what,
if
there's
there's
document
changes.
So
what
if
we
find
a
typo
in
the
description
or
something
wanna
fix
that
real
quick
I
mean
that
would
that
run
through
very
quickly,
I
guess,
because
that
wouldn't
necessarily
require
a
retest
of
any
of
the
code,
because
the
code
wouldn't
actually
change.
It's
just
documentation,
oh
for.
A
Sure
yeah,
we
would
love
to
do
it
for
you
to
find
PRS
against
any.
You
know
things
that
you
find
missing
or
like
wrong
there
so
yeah.
Definitely
this
is
you
know
by
no
means
just
owned
by
a
couple
of
people.
This
is
really
out
there
for
the
community
to
embrace
and
to
to
refine.
So
if
you
find
stuff,
you
know
just
give
us
a
PR
I.
B
A
And
yeah
use
this
check
this.
If
you
want
to
submit
or
update
you
operator,
just
to
be
sure
that
you
have
like
I'm
done
through
all
the
things
we
do
all
tests,
we
test
all
this
as
part
of
the
CI,
so
you
might
have
seen
some
changes
in
the
CI.
Actually,
that
is
executing
now
much
much
more
tests,
so
you
see
here
it
is
actually.
A
A
Rigorous
set
of
checks
and
in
place
that
will
basically,
you
know
to
the
linting,
make
sure
you're
not
colliding
with
any
kind
of
you
know,
operators
that
are
already
in
the
open,
just
catalog
they
will
deploy
before
all
them.
They
will
check
your
metadata.
They
will
check
that
you
haven't
modified
more
than
one
operator
at
the
time
and
as
part
of
the
new
PR
guidelines.
A
If
you
go
here
on
the
side,
the
upper
right
hand
corner
you,
go
to
the
contribute
button
and
there's
now
a
new
menu
item
that
says
preview
your
operator
and
what
we
mean
by
that
is.
You
can
basically
upload
your
CSV
and
you
would
basically
see
how
these
how
this
will
pop
up
at
the
side.
So
I
can
just
do
that
here.
Real,
quick
and
upload.
A
Yes,
ruk
chef
CSV
on
there
and
you
would
see
how
that
would
look
on
the
side
right.
You
see
that
tile
view
here.
The
list
view
here
and
you
would
also
see
how
the
descriptor
page
would
be
rendered
right.
So
that's
a
nice
way
to
kind
of
spot
any
kind
of
markdown
errors
or
anything
that
you
have
forgotten
in
here,
as
well
as
if
and
how
your
custom
visas
definitions
pop
up.
A
If
they
have
proper
examples,
so
that's
not
only
valuable
for
Pratap
that
I
know
but
can
also
be
used
for
overshift,
because
the
operator
hub
in
OPA
shirome
shift
is
visually
very
close
to
to
what
we
have
here
on
operator
after
I/o.
It
doesn't
display
all
the
data,
but
this
place
most
of
it,
so
you
said
as
a
low-hanging
fruit
as
a
check
to
see
if
your
CSV
is
correctly.
This
will
also
yield
any
kind
of
beard:
llamó,
gacho's,
formatting
errors,
indent
errors
and
whatnot
right.
A
So
much
better
to
see
it
quickly
like
this
than
to
wait
for
the
pipelines.
We
basically
do
its
thing,
and
you
know,
as
I
said:
that's
up
here
in
the
upper
right
hand,
corner
and
contribute
and
then
preview.
Another
thing
that
we
have
starting
to
put
in
place
and
are
looking
to
release
very
soon
as
a
baton
will
give
it
off
your
feedback
on.
This
is
a
way
to
visually
edit,
your
CSV,
basically
starting
from
scratch
right.
A
So
we
all
know
that
you
know
coming
up
of
the
CSV
can
be
a
little
bit
of
a
steep
hill,
even
with
the
examples,
because
there
are
so
many
feels
like-
and
you
probably
already
have.
Your
existing
cuban
Etta's
manifests
like
deployments,
service
accounts
or
back
and
so
on.
So
what
we
can
do
in
a
the
next
release
of
operator
hop
that
io
is
basically
on
going
to
contribute
menu
again
and
there's
another
item
that
says
package
to
operator
and
that
will
give
you
a
visual,
a
visual
CSV
editor.
A
A
Let's
take
this
one
and
it
would
start
to
populate
all
the
fields
in
the
editor
which
are
down
here.
So
we
basically
walk
you
through
all
the
steps
to
create
a
complete
and
high-quality
CSV,
and
this,
in
this
view,
I
and
there's
validation
on
each
step
as
you
go.
So
you
see,
for
instance,
here
is
the
cluster
service
version
of
rooks
F
referenced
a
couple
of
theories
right
and
in
order
to
you
know,
build
a
pallet
bundle
that
you
can
then
submit
as
a
TR.
We
would
also
need
those
Tod
files
right.
A
So
that's
already
telling
me
I'm
missing
something.
So
I
can
basically
upload
that
right
away,
don't
need
to
upload
this
package
mo
or
the
old
o
CSV,
and
you
see
it.
It
validates
that
automatically
and
it
basically
also
pre-populates,
are
the
editor
with
that.
So
I
can
now
start
to
go
into
the
editor.
Look
at
all
the
things
can
input
their,
maybe
I
want
to
change.
This
play
name
change
the
description,
the
one-liner,
as
well
as
the
more
elaborate
description
on
the
main
page.
So
you
see
here
we
have
a
markdown
editor.
A
We
give
you
a
visual
Lee
which
previews
just
to
make
sure
your.
You
have
the
right
tools
at
hand,
so
really
make
sure
operator
look
good
on
both
operator
hubs,
and
all
of
this
contains
validation
right.
So
it's
it's
much
harder
to
to
do
mistakes
here
or
to
miss
something
you
have
visual
inputs
for
almost
every
aspect
of
the
CSV.
You
know
you
can
set
your
capabilities
level.
You
get
an
explanation
of
what
it
is.
A
So
I
personally
find
this
much
more
satisfying,
much
more
guiding
than
you
know,
sitting
in
front
of
the
amel
editor
and
trying
to
think
of
anything
that
I
could
add.
There
you
know,
when
you're
done
with
this
kind
of
stuff,
you
always
click
all
set
with
operator
metadata,
and
then
this
is
kind
of
the
first
section
that
you've
completed
and
as
such,
you
would
walk
through
the
rest
of
the
editor.
So
here
we
would
ask
you
to
specify
all
the
CR
these
that
your
operator
owns.
You
would
be
able
to.
A
You
know
these
two
series,
edit
those
give
us
the
list
of
resources
that
are
behind,
so
we
are
able
to
populate
DUI
in
overshift
correctly.
With
this,
you
would
be
able
to
provide
us
with
examples
or
even
SPECT
descriptors,
to
make
a
custom
UI
for
your
operator.
So
all
of
this
is
now
much
better
documented
in
a
much
more
centralized
place
and
also
always
paired
with
a
visual
preview
of
what
you're
going
to
get
when
you
submit
this.
A
So
if
I,
for
instance,
miss
something
here
right,
let's
say:
I'm
busy,
not
specifying
a
name
that
would
be
an
error
right
and
if
I,
if
I,
try
to
run
with
that.
That
would
also
be
you
know,
kind
of
sticky
in
the
editor,
and
it
would
stubbornly
refuse
to
generate
a
bundle
out
of
this.
So
hopefully
it's
much
easier
and
much
more
straightforward
in
the
future
to
create
a
vetted
CSV
bundle
and
basically
package
you
already
existing
operator.
A
A
Regular
kubernetes
on
gentleman's
rolls
bol,
binding
service
accounts
and
and
NCR
DS,
and
we
do
it
actually
already
support
series.
But
what
you
really
want
to
do
is
you
want
to
take
these
standard
manifests
tips?
Probably
every
operator
developer
has
already
because
they
tested
it
on
their
local
cluster
and
also
the
standard
manifests
that
the
SDK
will
output
and
we
will
convert
as
much
of
that
as
possible
into
a
CSV.
So
as
all
the
fields-
and
you
just
need
to
you
know,
like
it's
indicated
here,
do
a
review
of
you
know
review
your
deployment.
A
This
is
a
fully
fledged
amble
editor,
so
this
will
contain
syntax,
highlighting
syntax
check
so
hopefully
much
harder
to
do
to
you
know
to
have
arrows
and
indices
V
and
yeah.
That's
that's
pretty
much.
What
we're
working
on
right
now
and
hopefully
be
able
to
release
that
Christian
as
a
beta
and
yeah
keep
Q
keep
being
tuned
in
for
that
I
think.
That's
all
I
had
any
any
questions
so
far.