►
From YouTube: 2020-04-21 Rook Community Meeting
Description
No description was provided for this meeting.
If this is YOUR meeting, an easy way to fix this is to add a description to your video, wherever mtngs.io found it (probably YouTube).
A
Okay,
the
recording
has
started-
and
this
is
the
April
21st
2020
look
community
meeting.
We
will
start,
as
always
with
our
milestone.
Checkup
I,
don't
think
we
have
anything
planned
for
1.2,
since
we
have
1.3
out
and
we'll
be
doing
some
patch
releases
on
1.3
to
continue
addressing
issues
that
come
up
in
in
driving
the
quality
and
stability
in
that
one.
B
A
C
A
C
A
B
B
B
A
C
Right,
let's
see
the
one
issue
I
see
I'm
blocking,
or
at
least
their
pot
n
to
infinity.
Oh
I,
think
that
wouldn't
block
through
release,
so
they
so
just
to
be
clear.
The
any
issue
that's
tagged,
41.3
we'll
currently
be
put
automatically
in
the
blocking
release,
column
and
then
just
kind
of
as
a
default,
and
then
it
can
be
moved
over
if
it's
not
blocking
the
next
patch
release.
So
I
don't
think
that
that
one
would
block
the
patch
release
at
all.
It's
just
just
nice
to
have
41.3.
A
C
A
Right:
okay:
let's
move,
then
on
head
to
the
community
topic
section,
so,
first,
a
quick
update
on
graduation.
We
confirmed
with
the
technical
oversight
committee
that
we
do
indeed
have
to
author
a
due
diligence
document.
That
was
a
new
requirement
for
the
incubation
stage.
That
was
not
there
when
we
achieved
when
we
were
accepted
into
incubation
back
in
2018,
but
the
technical
oversight
committee
has
confirmed
that
all
projects
do
you
need
to
do
this
retro
actively
for
a
graduation
if
they
did
not
have
to
do
it
in
an
incubation?
A
A
For
it
and
we
need
to
drive
that
process
and
the
storage
sake,
you
can
give
us
some
support
with
if
they
need
zoo
but
I,
don't
know
how
productive
that'll
be,
since
a
lot
of
it
is
answering
the
questions
with
from
a
expertise
and
debating
knowledge
of
the
project,
which
is
us
so
where
we'll
have
to
like
work
into
to
accomplish
this.
Unfortunately,
I'm.
A
Yeah
I
think
I've
seen
I've
seen
a
kind
of
a
quite
a
range
in
length
of
due
diligence.
Documents
from
various
projects
aiming
point
earnest
at
helm
is
being
a
good
recent
example
and
I
think
there's
just
like
six
pages
or
so.
There
apparently
is
a
lot
of
duplicate
or
repetitive
questions
within
the
template.
So
if
you
address
a
topic
in
general
and
then
the
thought
the
question
comes
up
again,
you
don't
have
to
re-enter
it
or
reformulate
it
or
anything
like
that.
So
I
expect
you
know
like
a
five
or
six
page.
A
A
We
had
one
project
out
for
like
on
the
docket,
for
students
to
submit
proposals
to
for
updating
the
NFS
operator
and
converting
its
controller
runtime
and
adding
a
lot
of
new
features
and
functionality
and
quality
adjustable
quality
issues
there
related
to
performance
and
stability,
and
things
like
that.
So
we
got
three
proposals.
A
From
students
on
the
NFS
operator
there
and
Rohan
Ashish
and
I
all
reviewed
all
of
them
and
conferred
together
and
they
all
came
the
exact
same
conclusion
on
one
of
those
proposals
being
the
strongest
of
the
bunch.
So
we
officially
selected
that
one
as
our
choice
for
the
one
that
we
want
to
mentor
and
take
on
for
google
Summer
of
Code
this
summer
here.
C
A
C
C
Link
say
to
slack
I,
put
a
question
in
the
cockroach
channel
a
few
days
ago
and
just
thought
I'd
bring
it
up
here
so
I.
A
few
weeks
ago,
I
watched
a
little
presentation
from
cockroach
guys
on
how
to
sort
of
a
getting
started
with
cockroach
and
deploying
it
and
stuff
and
the
there
their
perspective
on
deploying
cockroach.
C
Is
that,
and
they
said
twelve
minutes
into
the
presentation
that
basically
we
use
held
in
shards,
we
don't
need
operators
like
that
are
considered,
so
the
I
guess
the
question
everything
up
here
is:
if
we
think
that
the
cockroach
operator
is
important,
how
can
we
push
them
to
get
involved
or
convince
them
of
it
and
the
goal
being
to
get
the
cockroach
operator
out
of
alpha
state
right?
It's
been
been
a
for
a
while.
Well
like
a
few
other
operators
we
have,
but
that
one
since
I
heard
them
talk
about
it
like
that.
C
A
Yeah
so
my
take
on
this
is
that
when
you
know
we
were
talking
to
the
cockroach
folks,
you
know
the
end
of
the
year
2019
last
year.
You
know
one
of
the
big
important
one
of
the
important
things
for
them.
Is
that,
yes,
that
they
are
fairly
this
fairly?
They
have
a
lot
of
confidence
in
the
cockroach
application
layer
itself
around
its
self
healing
in
self
management
and
scalability,
etc
that
it
has
built
into
the
application
layer,
logic
itself,
which
kind
of
removed
some
needs
for
operator
logic.
Perhaps,
but
they
were,
they
were
interested
in.
A
You
know
continuing
the
effort
on
the
Brooks
cockroach
cockroach
operator,
they're
stumbling
block
to
it
was
that
it
was.
The
effort
was
still
based
on
the
old
rock
operator
kit
and
not
using
controller
run
time.
So
it
was
interesting.
I
think.
Maybe
it's
like
a
week
or
two
ago
there
was
a
community
I
think
it
was
a
community
effort,
I'm,
not
sure
who
the
the
the
contributor
was,
that
opened
a
PR
to
convert
the
cockroach
operator
to
control
the
runtime
based
implementation,
but
that's
a
PR
out
there.
A
That
would
be
nice
to
actually
get
in
and
have
that.
You
know
the
cockroach
operator
be
completely
controller,
runtime
based
and
then
re-engage
with
cockroach
to
see,
if
that's
something
that
they're
more
interested
in
and
could
be
contributing
to
and
participating
in
and
taking
ownership
of,
because
it
now
is
controller
runtime
based
and
their
biggest
impediments
has
been
addressed.
So.
C
B
A
A
C
Gonna
start
talking
about
flaw:
it's
nothing
yeah!
Hopefully
a
lid
I.
Just
put
these
here's
kind
of
FYI
I,
don't
know
that,
there's
any
huge
design
discussion.
We
need
to
have
more
concern
about
them,
but
the
first
one
is
admission
control
or
so
this
is
basically
when
you're,
creating
the
see
ours.
Some
upfront
validation
before
the
CR
is
actually
committed,
so
we
can
reject
if
the
settings
are
bed.this.
The
admission
controller
does
require
a
lot
of
extra
kind
of.
C
What's
the
right
word
extra
things
to
get
it
configured,
so
it
seems
like
a
pain
to
to
deploy,
because
you
need
a
certificate.
It's
got
a
you
know:
Web
API,
the
the
requests
come
through
to
validate
excetera,
but
it
it
is
a
really
nice
thing,
I
think
to
be
able
to
validate
the
settings
for
those
who
will
be
able
to
do
those
extra
steps
to
set
up
the
admission
control
tomorrow
in
the
Ceph
huddle,
we're
planning
it.
We'll
probably
have
a
demo
on
this.
A
I'd
say
got
a
500
error,
so
that's
a
github
is
down
yeah,
that's
right
is
there
any
one
of
the
things
that
I
had
noticed
about
mission
controllers
is
that
I
was
hoping
that
there
was
more.
You
know,
support
in
like
upstream
frameworks
like
the
controller,
runtime
or
operator
framework
or
whatever
to
make
that
eat
the
process
easier
of
getting
a
admission
controller
with
its
web
via
a
web
api
and
your
certificates
and
everything
into
the
cluster
is
there?
A
C
I
mean
it
just
doesn't
seem
like
there's
much
support
and
kubernetes
itself
for
to
make
it
easy
to
bring
it
up
like
you
have
to
get
your
own
certificate
and
things
so
for
the
default
or
just
to
get
going
sure.
You
use
a
self
sensor
to
install
it
if
you
want,
and
it
won't
be
something
we
can
install
or
configure
by
default,
it'll
be
this
kind
of
additional
step.
Oh
you
want
this
to
validate
your
see
ours,
you
know,
follow
these
steps.
Alright,.
C
A
Yeah,
it
sounds
like
one
of
those
things,
that's
a
great
idea
to
support
better
experiences,
but
it
is
difficult
to
get
going
with
much
like
CRD,
versioning
and
migrations
and
stuff.
You
have
to
write
a
book
for
that
as
well.
To
do
conversions,
it's
the
same
problem
as
an
admission
control,
essentially
right.
C
He'll
charge
support,
yeah
I,
don't
seem
in
this
meeting,
but
having
an
upstream
discussion
on
that
issue
about
how
we
support
multiple
clusters
from
basically
using
a
hell
chart,
because
it's
a
real
pain
with
the
Ceph
operator.
At
least
you
know,
cradle
our
back
in
a
different
name
space
to
run
your
cluster
in
another
namespace
from
the
operator.
So
it's
an
interesting
idea
anyway,
to
use
a
helmet
chart
to
be
able
to
create
the
are
back
for
you
and
iterate
over
that.
C
A
A
It's
their
little:
what's
it
called
even
has
a
name:
I
forgot
it
octa
cat.
That's
the
name.
Octopus
goes.
C
Apparently
they're
having
issues
building
arm,
so
there
is
a
two
dot
one
dot
Oh
CSI
release
out,
but
until
there's
an
arm
release
I'd
like
to
I,
don't
think
we're
gonna
get
the
default
for
the
rook
except
operator
to
install
anyway,
so
waiting
for
the
arm
issue
to
be
resolved
to
make
it
the
default
again
in
work.
Oh.
A
A
A
Yeah
yeah
yeah,
so
we'll
follow
up
on
staffing
that
then.
A
A
A
A
C
B
B
Yeah,
perhaps
I
can
prepare
something
for
next
time
or
the
one
after
like
for
sure.
It
would
be
nice
to
do
some
kind
of
a
retrospective
once
I'm
done
with
the
Southwest
OCR,
which
might
be
around
two
to
three
weeks,
I'm
hoping
so
by
then.
Maybe
I
can
doing
this
time.
Do
a
little
retrospective
on
how
it
went
and
yeah
I
guess.