►
From YouTube: SIG Cluster Lifecycle 2022-02-22
A
B
C
Awesome,
hey
everyone,
I'm
hoping
my
video
is
coming
through.
My
name
is
vipin
mohan,
I'm
a
principal
product
manager
at
amazon
web
services
working
closely
with
shape
blue
and
rohit's
team
on
a
classified
pay
provider
for
cloud
stack.
So
I
look
forward
to
talking
to
you
all
peter.
Do
you
want
to
go
next.
D
F
I
should
say
hello
guys,
I
don't
know
if
that's
all
the
aws
guys,
though
I'm
giles
syrah,
I'm
from
shape
blue
also,
I'm
also
been
working
on
the
the
cappy
provider
with
with
rohit
and
the
team.
I'm
actually
just
here,
just
to
sort
of
observe
the
process
really
and
just
to
understand.
What's
going
on
and
how
this
goes
today,
so
I'm
gonna,
I'm
gonna,
look
if
people
don't
mind
but
happy
to
contribute.
If
you
want
me
to.
A
B
Sure,
thank
you
so
just
to
give
a
background
before
this
call,
we
had
joined
the
weekly
cappy
meeting
and
from
that
we
started.
You
know,
pitching
this
new
provider
for
apache
cloud
stack,
so
apache
cloud
stack
is
infrastructure
as
a
service
cloud
computing
platform,
it's
an
open
source
project
under
the
asf,
and
we
presented
this
idea
to
initially
see
guidance
from
the
cappy
sub
project.
B
You
know
what
to
do
and
where
to
kind
of
start,
so
we
started
with
a
mailing
list
thread
and
from
that
we
subsequently
joined
the
weekly
cluster
api
provider
call
and
from
that
we
got
guidance
from
fabrizio
yasin
and
mike
to
kind
of
you
know,
put
together.
This
document
raise
a
github
issue
for
the
repository
migration,
and
here
we
are
sort
of
you
know,
to
pitch
the
new
provider
to
the
sig
and
get
the
guidance
and
support
and
eventual
approval
of
migrating.
B
This
repository
under
the
sig
just
to
give
an
update
on
where
this
product
currently
stands.
So
we
have
so
far
tested
two
versions
of
cloud
stack
cloud,
stack,
414
and
416
versions
against
three
main
hypervisors.
We
support,
namely
vmware
zenserver
and
kvm.
As
part
of
this
work,
we
have
already
submitted
some
of
the
patches
to
the
image
builder
project.
So
a
lot
of
you
know
things
like
support
for
centos
7,
rel
8
in
rocket
linux,
8
were
added
by
our
team.
B
As
part
of
this
work,
the
capy
provider
is
functional,
has
passed.
A
lot
of
the
testing
that
we
have
done
has
also
passed
a
set
of
end-to-end
tests
that
we
sort
of
did,
and
one
of
my
colleague
have
also
run
this
thing
against
the
conformance
test,
which
we
believe
has
also
passed.
So
that's
where
the
product
currently
stands
at,
and
we
are
basically
here
at
the
group
to
to
say
that
this
is
the
repository,
and
where
do
we
go
next?
B
How
do
we
get
this
thing
migrated
to
to
the
sig
as
part
of
the
due
diligence?
We
have
tried
to
get
all
the
licensing
done
so
any
even
from
the
very
beginning
we
have
adopted
the
apache
license
version
2..
We
have
ensured
that
all
the
license,
all
the
files
have
this
proper
header.
All
the
clas
of
various
authors
have
been
submitted
myself
peter
and
I
think
joe.
We
have
also
tested
when
we
raise
a
pull
request
to
communities
community,
the
bot
kind
of
says
that
we
have
passed
the
easy
cla
and
whatnot.
B
A
G
G
We
just
need
to
do
a
little
bit
of
do
the
legends
like
when
we
get
new
providers
with
that.
Like
you
know,
there
is
one
stamped
approval
for
each
one
of
the
cloud
providers
or
you
know,
methods
of
operating
but
other
than
that
just
requirements
to
have
active
maintainers
on
it.
Otherwise
you
know
or
proceed
to
archive
the
repo.
B
Vince
apologize,
I
think
I
should
have
introduced
myself
properly.
So
apart
from
being
a
principal
engineer
at
shea
blue,
I'm
also
a
long-term
committer
and
bmc
member
of
the
project.
I've
been
working
with
the
apache
cloud,
stack
open
source
project
since
the
last
10
years,
so
I've
kind
of
am
also
representing
you
can
say
the
cloud
stack
pmc,
which
is
the
project
management
committee,
that
oversight
releases
of
the
apache
cloud
stack
infrastructure
as
a
service
cloud
computing
software.
A
A
A
A
It
will
be
a
conflict
of
interest
with
this
provider
that
we
are
in
today.
So
you
know
vsphere
aws
providers.
We
want
to
avoid
this
conflict
of
interest,
so
it
might
be
nice
to
document
it
somewhere,
but
up
to
you,
if
you
want
to
do
it,
I
think
I
have
a
couple
of
questions.
Reading
this
the
working
progress
migration
proposal
sure
so
do
the
quest
api
maintainers
think
that
cap
c
should
be
the
name
reserved
for
this
provider.
A
B
B
A
A
B
B
G
Yeah,
that's
a
lot.
I
would
start
with
a
couple
of
folks,
especially
because
it
seems
like
nobody
here
has
org
access
yet
or
cognized
membership,
and
this
couple
of
looks
like
I
should
you
know
be
sponsored
by
someone
after
they
just
contribute
a
little
bit
like
in
other
repos,
as
well
or
in
class
trippy
itself.
G
You
could
have,
I
mean
I
would
just
select
two
or
maybe
three
folks
max
for
now
and
then
you
could
have
everybody
else
as
a
reviewer
and
then
you
know
over
time
as
the
contribution
go
up
like
you
can
promote
people
as
approvers
and
that's
kind
of
like
how
we
operate,
but
I
maintainers
and
approvers
have
to
be
active
as
well.
That
is
a
requirement
for
that.
A
It
could
act
like
a
bottleneck
at
the
beginning,
for
you
with
a
couple
of
people,
but
just
you
can
assign
you
can
add
to
the
progress,
is
the
most
active
people
and
then
slowly
scale
the
owner
as
well?
We
actually
do
perform
yearly
checks,
whether
the
waterfall
is
up
to
date
and
there's
a
on
a
related
note.
We
started
collecting
soup
project
updates
every
year,
so
you
have
to
provide
like
a
report
to
us.
Is
your
owner
follow
up
today?
It
is
your
contributor
same
md
file
up
to
date.
A
B
H
G
H
H
I
A
I
A
A
D
D
J
A
H
Let
me
say
in
the
cncf
going
into
the
cfccf
is
like
saying:
we
want
our
project
to
grow
in
a
in
a
safe
space,
which
is
public,
shared
stuff
like
that,
and
if
you
want
your
project
to
grow,
you
have
to
also
make
visible
your
test
senior
so
making
these
in
in
the
same
dashboard
that
other
folks
are
using
is
part
of
making
your
project
part
of
the
of
the
same
story.
So
it
is
up
to
you.
It
is
not
required
for
migration.
I
I
actually
don't
there
is
a
report
on
how
much
working
group
kate
sanfra
is
spending,
but
my
understanding
is
that
most
of
the
testing
itself
is
not
migrated
over
to
that
account.
So
it's
not
included
and
the
general
support
of
aws
is
also
not
included
in
in
that
bucket
either.
So
no
I
don't
I
don't
have
an
answer
is
that
is
the
obviously
short
it's.
A
H
B
Just
a
quick
one,
so
from
what
I
understand,
the
immediate
action
we
need
to
take
is
to
trim
down
the
onus
list.
Anything
else
we
need
to
do
before
you
know.
This
thing
is
is
ready
for
for
acceptance,
and
the
second
part
of
the
question
is:
what
is
the
mechanics
of
you
know?
Getting
a
repository
transfer?
Is
it
that
somebody
will
clone
the
existing
repository
or
do
we
use
have
to
use
github
transfer
transfer
of
ownership
to
the
to
the
sig
org.
A
B
B
D
A
H
B
Sure
absolutely
something
we
can
work
on
and
maybe,
as
part
of
in
the
next
week,
we'll
add
it
to
the
github
issue
or
the
proposal,
a
youtube
video
with
a
demo
of
cloud
stack
and
how
the
provider
can
be
used.
One
of
the
things
that's
still
pending
on
us
is
integration
of
cap
c
with
cluster
ctl.
So
right
now
our
mode
of
you
know
working
on
this
is
directly
and
not
via
cluster
ctl.
B
B
B
A
A
Which
is
a
very
good
way
for
us
to
collect
subproject
information,
because
we
have
so
many
different
sub
projects
and
we
are
adding
more
and
more
so.
I
really
like
the
survey
idea
because
it
can
serve
surface
governance
problems
and
the
survey
is
pretty
much
done
fabrizio.
I
think
20
su
project
is
the
right
number.
Do
you?
Are
we
missing
anyone
from
costa
api
at
this
point.
A
H
A
H
Yeah,
I
think,
let's,
let's
stick
to
the
summary
and
and
then
eventually
we
take
the
action
item
to
investigate
if
there
is
something
that
does
not
reason
it.
For
instance,
if
I
look
at
this
question,
it
seems
to
me
that
having
some
project
in
the
community
not
holding
a
regular
office
hour
is
something
that
is
not,
let
me
say,
in
line
with
the
community
spirit,
so
I
think
that
we
have
to
reach
out
to
these
40
percent
of
project
and
ask
why
or
propose
them
to.
A
To
be
clear,
community
biddings
are
optional,
they
are
not
mandatory.
In
fact,
recently
we
dropped
the
cube
spray
office
hours
with
agreement
from
the
maintainers
because
they
said,
there's
no
point
of
holding
a
meeting.
If
I'm
the
only
person
that
joins
the
meeting
every
time,
the
discussion
is
happening
on
slack,
which
is
acceptable
in
my
book
at
least.
H
A
H
A
A
A
A
I
A
Yeah,
if
you
have
someone
who,
who
you
trust
with
the
youtube
playlist
key
slash
link,
you
can
just
add
them.
So
you
know
maybe
delegate
responsibility.
You
don't
have
to
be
the
only
person
that
does
this.
If,
like
you
have
five
people
in
the
meeting,
you
just
tell
someone
to
record
up
to
their
youtube
and
then
we
can
add
it
to
the
playlist.
I
K
A
Yeah
and
also
bobby
bobby,
is
saying
that
the
youtube
credential,
basically
out
authorizing
someone
to
do
something
in
a
playlist
is
so
sketchy
right
now,
just
missing
api
features.
So
once
you
add
someone
with
credentials,
you
can
no
longer
remove
them
from
this
privilege.
It's
pretty
bad,
but
yeah.
It's
just
I
mean
probably
a
single
playlist
you
can.
A
You
can
get
keep
the
playlist,
but
as
long
as
someone
takes
the
video
records
and
uploads,
it's
like
that's
90
of
the
work
you
just
click
add
to
the
playlist,
which
is
pretty
simple
and
the
automation
around
this
whole
upload
process
is
so
sketchy.
It
doesn't
work,
so
we
have
to
continue
doing
it
manually.
Unfortunately,
for
now,.
H
A
Yeah,
I
really
like
ad
hoc
meetings
and
in
fact,
instead
of
the
regular
meetings
one
day,
we
can
perhaps
switch
some
super
projects.
Have
these
ad
hoc
slots
somehow,
for
instance,
in
cuba,
then
nowadays
we
just
don't
have
that
much
agenda.
So,
instead
of
the
regular
call,
we
can
have
like
a
monthly
ad
hoc
meeting
and
a
particular
time
instead
of
having
the
regular
call.
A
A
A
H
A
A
A
F
A
I
Perhaps
that's
something
we
should
like.
I
always
think
it's
great
if
we
can
learn
from
each
other.
So
if
someone
is
doing
a
great
job,
like
mini
cube,
looks
like
they're
doing,
maybe
we
should
all
copy
their
guide
and
modify
it
or
something
like
that,
but
like
we
can
open
issues
and
all
the
projects
to
like
consider.
A
H
H
I
remember
down,
and
there
are
people
asking:
oh,
we
need
a
new
contributor
for
this
area.
We
are
under
stuff
and
in
this
other
area,
and
so
I
think
that
we
have
that
might
take
away
from
this.
Is
that
maybe
that
in
our
sub
project,
there
is
not
a
common
understanding
on
how
to
on
what?
What
are
the
tools
that
you
can
use
to
grow
your
community
and
we
can
work
to
to
to
grow
this
common
understanding.
A
A
A
A
All
our
file
updates.
Ideally
this
is
a
number
that
we
should
like
get
to
zero
privilege
and
dim's
road
tool
to
run
on
the
repositories
that
you
maintain.
It's
called
maintainers,
but
the
maintainers
too
can
give
you
like
a
few
warnings.
If
somebody
in
the
owner's
file
is
like
a
stereo
entry
and
actually
I
I
ran
the
tool
against
some
other
files
and
the
communities
community
seeks
yamaha
when
I
was
able
to
clean
some
warnings,
but
basically
yeah
in
the
future,
we
would
want
to
get
this
like.
A
A
The
leads
of
a
particular
project
how
to
better
educate
their
people
that
are
climbing
the
the
ladder
and
it's
also
very
difficult.
I
I'm
I'm
I'm
going
to
admit
that
I'm
struggling
with
that,
but
I'm
doing
it
for
a
couple
of
people
in
kubernetes
and
I
managed
to
get
one
of
them
to
approve
her.
The
other
one
is
still
reviewer.
I
think
they
still
need
to
focus
on
particular
details
to
become
a
better
reviewer.
At
that
point,
maybe
they
can
become
an
approver.
A
A
H
Reviewer,
oh
in
cluster
guide,
it
is
a
continuous
work
to
to
keep
the
community
working.
So
we
are
continuously
reviewing
contributor
guidelines,
all
the
guidelines
testing
all
the
developer
tooling.
So
we
we
do
our
best
to
get
new
people
to
forks
come
in
then
every
cycle
that's
been
every
minor
release.
H
After
every
minor
release.
We
ask
people
if
they
want
to
step
in,
and
then
we
consider
basically
people
that
want
to
step
in
as
a
reviewer
or
as
a
maintainer
last
cycle.
We
did
also
given
that
cluster
api
code
base
is
getting
bigger.
We
define
we
divided
the
code
base
in
several
owner
files,
which
are
documented,
and
so
we
are
trying
to
make
it
easier
for
a
person
to
step
in
by
be
becoming
as
owner
only
of
a
small
part
of
the
project
instead
of
the
entire
project.
H
A
H
So
the
fact
that
at
this
meeting
there
are
they
are
not
so
participating
by
some
project
that
this
project
does
not
come
here
and
share
concern
about
the
contributor
base,
or
how
do
they
do
things
is,
is
kind,
is
kind
of
a
senior
that
we
are
not
providing
them
a
good
guidance
and
yeah.
This
is
something
that
we
should
reflect
on
and
and
see.
If
we
as
a
sig,
we
can
provide
a
better
guidance
around
this
project
and
not
in
on
every
aspect,
not
only
on
the
c
clusterize
cycle.
A
F
A
Projects
and
we
can
actually
go
and
ask
some
of
the
super
projects
that
replied
hey.
I
don't
have
reviewers
to
join
the
call
and
that's
how
we
can
start
a
conversation
with
them,
but
if
they
don't
have
any
feedback
for
us
or
any
questions,
it's
actually
not
mandatory
for
a
super
project
to
join
the
status
updates
that
we
do.
I'm
really
not
a
fan
of
that
unless
there
is
something
that
is
actually
to
be
discussed,
something
important
and
we
are
actually
discussing
to
drop
this
entirely.
A
A
And
actually,
I
can
actually
extract
some
of
these
in
the
annual
report.
I
asked
whether
we
have
anything
we
want
to
add
here,
but
nobody
contributed
comments
and
I
can
actually
take
some
of
these
responses
here
and
add
them,
because
we
have
a
lot
of
sub
projects
and
we
added
only
three
entries
here
about
dokkerishim
kosriyan
kivarium.
A
H
After
three
minutes
yeah,
my
takeaway
from
this
part
is
that
there
is
a
lot
of
good
things
going
on
great
things
going
on
in
the
project,
but
we
don't
give
them
visibility
internally.
Externally
these
are
lost
again.
We
lose
opportunity
because
we
can
learn
each
other.
We
can
share
experience
or
stuff
like
that.
So.
A
A
H
Yeah,
the
lack
of
mythic
in
persons.
I
think
that
is
something
right
around
the
my
point
is
that
if
there
is
a
project
that
is
willing
to
be
part
of
the
seed
and
to
be
part
of
the
kubernetes
sea,
but
it
is
not,
it
does
not
find
important
to
basically
be
part
of
a
story
or
of
a
narrative
of
the
idea
of
a
secretary
cycle
that
we
are
pushing
forward.
There
is
something
strange,
so
we
are
not
an
hosting
service.
A
Well,
frankly,
we
are
not
unique
in
this
with
this
problem,
because
we
have
other
I've
seen
other
projects
so
projects
of
other
six
and
these
so
project
not
necessarily
report
every
time
there
is
a
sick
call
or
on
a
mailing
list.
So
they
they
don't
give
status
updates
to
the
main
group.
They
have
their
own
bubble
and
if
there
is
something
they
will
obviously
go
to
the
to
the
sig
leads,
but
it's
not
that
they
report
or
communicate
with
the
sick
leads
on
a
daily
basis
or
a
weekly.
A
Yeah,
I
think
this
is
it's
a
complicated
problem.
This
is
not
something
that
we
can
solve
in
six
months.
It's
going
to
take
time
for
us
to
figure
out
how
to
properly
adapt
to
a
model
where
the
two
projects
are
more
communicative,
and
I
agree
that
we're
currently
like
we
are
becoming
a
hosting
service.
A
A
A
I
No,
I
mean,
I
think,
that's
great.
I
think
thank
you
for
organizing
the
survey
to
everyone
that
did
it
and
to
responding.
I
think,
like
a
lot
of
those
questions,
are
also
valuable.
Just
to
get
us
to
think
about
them
and,
like
think
about
ways,
we
can
encourage
get
the
products
sub
products
to
think
about
these
things
and
and
like.
If
there's
anything,
we
can
do
to
help
that's
great.
I
But
yes,
as
you
say,
like
you
know,
riding
docks
is
always
a
a
good
idea,
and
always
a
good
thing
to
do
with
updating
the
owners
is
another
thing
that,
like
I,
am
certainly
guilty
of
not
doing
on
a
regular
basis
and
is
also
a
good
thing
to
do
so,
like
I
think
the
the
poll
is
valuable
in
terms
of
being
thought
provoking
as
well.
The
survey
is
valuable
in
terms
of
being
thought-provoking
as
well.