►
From YouTube: 20181204 sig cluster lifecycle
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
Hello
today
is
Tuesday
December
4th
2018.
This
is
the
standard
sequester
lifecycle
call
I
will
start
with
a
PSA
that
Robert
Bailey
wanted
to
mention
to
folks
as
a
reminder
that
there
is
a
zoom
security
vulnerability.
If
you
haven't
seen
it
there's
a
link
in
the
meeting
notes.
Please
update
yours
in
clients.
A
Most
folks
should
have
seen
an
update.
You
can
do
that
by
basically
saying
zoom
said
check
for
updates
and
your
the
the
link
inside
the
doc
has
details
on
the
exact
conversions.
You
should
be
looking
for.
The
reason
you
want
to
do
this
update
is
because
you
don't
wanna,
have
your
meetings
pwned
from
hackers
that
can
spread
bad
things.
A
That's
one
update.
The
second
update
I
wanted
to
do
is
I
was
hoping
other
folks
from
other
sub
projects
would
be
here
today
to
discuss.
Updates
I
would
like
to
give
during
the
community
updates
during
the
contributors
summit.
I
will
share
my
screen
here
so
currently,
there
are
separate
subsections
that
I
have
in
the
dock
and
we'll
take
a
look
at
the
participants
here.
Members
is
there
anyone
from
mini
cube
or
the
call.
A
A
C
A
A
This
is
great,
I
will
I
will
definitely
add
details
here.
I.
Might
this
goes
into
the
2019
section
too?
So
what
I'm?
Looking
when
I'm
looking
for
it
in
the
doc
is
there
is
a
past
present
future
of
sequester
lifecycle
and
what
I'm
trying
to
do
is
give
a
brief
update
of
what
happened
in
2018
and
then
down
below
I'm
also
trained
to
talk
about,
what's
what's
coming
in
2019,
for
some
projects,
if
they
have
details
that
they
want
to
share,
so
it
looks
like
he
did.
Both
that's
awesome.
A
A
Do
but
I
will
get
to
it:
okay,
yeah!
There
is
a
cups
update
here
too,
as
well.
I,
don't
know
if
there's
notable
call-outs
that
you
want
to
mention
with
regards
to
bundles
I,
do
have
a
couple
of
slides
in
here
referencing
that
too
as
well.
So
you
know
obviously
the
the
interesting
things.
What
happened
in
2018
with
cops
as
well,
as
you
know,
potential
roadmap
items
for
2019,
and
if
there
are
details
you
might
need
to
share
with
about
bundles
there's
a
separate
slide
where
they
also
give
a
shout-out
for
your
talk,
awesome.
D
A
A
You
need
to
do
go
to
the
kubernetes
board.
So
if
you
go
here,
sorry
I
just
have
fast
links
on
the
night
at
it.
If
you
gave
me
the
kubernetes
board,
you
need
to
basically
print
yeah.
If
you
create
a
new
issue.
They'll
have
like
you
know,
details
on
repository
creation,
perfect,
thank
you!
So
much
yeah
and
just
link
to
the
mailing
list,
stuff
and
you're,
ready
to
vote.
A
All
right
so
I
think
those
are
all
the
sub-project
information
I
wanted
to
get
in
there.
Everyone
should
have
access
to
make
comments
if
you
want
to
add
more
details,
I'm
trying
to
give
a
time
for
every
major
sub
project,
that's
part
of
the
cluster
lifecycle,
so
there's
Guba,
DM
cluster
API,
cops,
poop
spray,
mini
cube
and
then
again,
conversely,
down
below,
what's
what's
coming
in
2019
talk
about
bundles
Etsy,
D,
management's,
Road,
metric
rubidium?
How
angry
I
am
about
test
apparatus
and
soapboxing
about
what
we
need
to
do
to
fix
it.
A
A
Know
if
you've
been
talking
with
Ben
I
physically
feces,
on
the
call
with
his
work
that
he's
been
doing
on
kind
but
I
think
the
the
deduping
of
some
of
the
apparatus,
for
you
know,
sort
of
cluster
stand-up
for
CI
integration
tools
or
just
basically
one-stop
developer
tools
would
be
super
duper
helpful
to
kind
of
like
unite
the
clans.
There
yeah.
E
A
Well,
I'm
gonna
sit
down
with
you
to
you
and
Robert
probably
later
on
this
week
and
go
through
the
other
slides
in
more
detail.
This
is
more
the
broader
update
for
the
sig,
but
I
I
need
to
talk
about.
There's,
there's
three
separate
talks
at
coop
con
that
are
sig
focused
yeah
and
one
is
the
intro
which
are
going
to
copy
a
lot
of
slides
from
China.
A
One
is
the
deep
dive
which
I'm
also
going
to
copy
a
lot
of
slides
from
China
I
already
have,
and
then
there's
the
cluster
API,
one
which
I'm
sure
there's
enough
detail
and
information
that
folks
can
come
on,
but
I
need
to
talk
with
Robert
this
week
it
sit
down
together
with
everyone
to
try
and
get
those
slides
cribbed
here
to
cribbed
and
the
other
talks.
My
my
intent
for
this
talk
is
basically
they
give
a
global
style
update.
A
F
A
A
A
A
That's
pretty
much
all
I
have
with
regards
to
global
updates.
Are
there
other
things
that
folks
wanted
to
discuss
as
potential
roadmap
items
for
or
discuss
in
general
about
koukin
I
know
that
we
have
a
couple
of
birds
of
the
Feathertop
sessions,
which
I
think
would
be
highly
useful
together
and
together
in
a
room
to
talk
about
several
topics?
Probably
I
don't
know
if
it
makes
sense
to
sort
of
pre
pre
vet
the
conversation
now
to
try
and
get
there
a
lot
of
things
to
discuss
what?
If
folks
thinks
they're.
F
A
Global,
like
higher
level
roadmap
of
of
integration
and
features
that
we
want
to
call
out
like
I,
think
at
CD,
a
DM
and
bundle
management
across
every
tool
would
be
a
highly
highly
useful,
leveraged
piece.
The
puzzle,
yeah
right,
I,
think
what
we're
starting
to
do,
which
is
the
right
approach,
is
to
have
these
sort
of
layered
toolbox
of
tools
and
then
use
them
across
all
of
the
other
components
so
that
we
don't
have
this
duplication
of
effort
over
time.
I
think.
A
A
A
Absolutely
I
opened
up
a
tracking
umbrella
issue
in
regards
to
packaging
yesterday
and
looped
in
both
the
sig,
as
well
as
some
other
people
who
are
interested
in
working
on
it.
My
packaging,
do
you
mean,
like
yam
undead
packages
or
yep
yep
yum
and
dead
packages,
as
well
as
the
other
assorted
things
that
we've
talked
about,
fixing
that
I
didn't
know
what
he
meant.
That
community
is
something
some
manifest.
A
A
G
A
Ideally,
all
of
the
artifacts
can
be
published
from
the
KK
repo
as
part
of
a
build
target
which
is
just
release.
I.
Think
we've
wildly
overcomplicated
what
we
currently
have
and
I
know
that
and
we've
also
have
constraining.
A
We
have
conflating
requirements
that
we've
created
for
ourselves
that
I'd
like
to
break
down
and
solve
this
problem
in
most
other
smaller
packages
that
you
do
releases
from
the
the
idea
of
a
release
is
simply
a
one-time,
very
minor
change
that
you
do
to
the
repository
like
for
most
of
our
projects
that
have
do
it's
just
an
annotated
tag.
You
you
annotate
a
tag
and
all
the
other
mechanisms
go
from
there.
A
The
the
actual
mechanisms
that
currently
exist
for
a
lot
of
they're
releasing
are
wildly
over
complicated
in
my
opinion
and
but
I
think
we
need
to
have
a
coming
to
terms
conversation
with
this
thing
release
because
we
we've
sinned.
Questor
lifecycle
has
borne
a
lot
of
burden
over
many
releases,
with
some
of
the
issues
that
existed.
A
A
F
A
So
I
don't
know,
there's
too
many
dependencies
currently
still
at
this
point,
and
there
are
at
this
point
it's
been
beneficial
to
rev
them
I
think
over
time,
once
we
have
a
well
factored
sort
of
componentized
version
of
the
world
where
we
can
actually
tease
the
dependencies
apart,
then
that
might
be
more
possible.
Last
I
heard
from
a.
A
Last
I
heard
from
an
actual
breaking
out
the
repository
thing.
It
is
essentially
stalled,
so
like
API
machinery
was
going
to
be
the
pilot
case
for
a
lot
of
stuff,
and
a
lot
of
that
work
has
stalled.
I,
don't
exactly
know
where
the
future
is
going
to
go
with
the
idea
of
breaking
out
into
multiple
federal,
repos
versus
doing
sort
of
the
mirroring
apparatus.
F
H
A
F
A
A
Yeah
as
a
PSA
2
for
folks
who
weren't
aware,
we
created
a
survey
for
kuba
DM
as
part
of
the
ga
release
to
try
and
get
more
feedback
on
what
are
the
most
high-value
areas
and
that's
just
good
general
usage
information
that
folks
would
like
us
to
focus
on
and
we'll
try
to
once
we
have
detailed
roll-up
information
from
that
survey.
We
will
report
back
out
to
the
broader
sig,
and
here
are
the
common
things
we
got
back
from
the
survey.
A
F
F
They
should
have
so
with
this
we're
first
gonna
target
a
code
like
pure
code
refactor
with
the
kept
linked,
but
then
also
start
to
change
the
behavior
of
the
components
to
be
consistent,
so
to
have
the
same
set
of
endpoints
for
those
components
who
run
HTTP
use
Cobra,
you
maybe
not
use
Cobra
use
logging
in
a
standardized
way.
Every
component
should
support
component
config.
The
component
config
type
should
be
in
a
good
place
that
allows
for
rendering
etcetra
etcetra
etcetra.
F
So
the
first,
the
cap
that
I
have
here
is
a
draft
of
so
that
I'm
gonna
still
resolve
the
comments
there,
but
it's
like,
as
is
approved
by
sync
architecture.
I
present
the
digital.
In
the
last
meeting,
the
last
Thursday
mm-hmm
and
after
after
I've
updated
and
we
finalize
the
code
refactoring
our
code
moving
speck
or
cap,
we're
gonna
implement
that
in
the
workgroup
components
or
workgroup
whatever
that
myself
and
Mike
dolphin
and
Stephanie
malsky
are
gonna.
Gonna
lead
I.
A
F
Yeah
yeah
absolutely
already
on
the
the
only
thing
that
we
so
like
got
full
approval
from
sega
architecture
other
than
some
bike
shipping
and
that
it
can't
become
cold
case
that
io
components
it's
a
library.
So
we
have
to
like
lib
components
or
components:
labor
whatever
and
well,
we'll
just
with
that.
But
the
idea
in
itself.
A
That
was
all
about
the
earlier
part
of
you
know
doing
the
what's.
Next,
one
of
the
things
I
want
to
discuss,
possibly
during
the
bath,
is
doing
a
sort
of
like
we
mentioned
a
global
2019
sort
of
planning
type
of
thing
that
would
cut
across
everybody
right
like
bundles,
something
I
would
cut
across
everybody,
sort
of
general
state
of
cou
medium.
You
know
integration
and
how
to
simplify
things
like
libraries
for
phases
right,
that's
a
way
to
better
integrate
things
well
defined
libraries
for
phases.
A
F
Do
you
see
the
class
life
cycle
Road
map?
Yes,
okay,
cool,
so
I'll
just
pick
this
up
later,
but
all
content
here
is
what
we
did
for
what
we
like
just
had
a
brain
dump
and
and
wrote
down
4018
and
we've
like
already
implemented
a
couple,
a
couple
of
these
which
I'll
remove
later.
But
you
said,
with
regards
to
cubed
M,
go
code,
better
structure.
F
J
F
D
I
can
say
four
four
cops
like
I
think
the
big
goal
is
sort
of
continue.
What
will
contain
2018,
I
think
which
is
trying
to
make
it
less
of
an
island,
and
you
know
using
things
like
vADM
and
cluster
bundles
and
hopefully
QA
DM
and
just
you
know,
that's
a
less
of
a
monolithic
or
they're,
nothing
in
more
of
a
Canary's
bundle
of
open-source
pieces
that
are
shared
yeah.
F
D
J
A
H
H
A
D
F
Right
so
so
they
wouldn't
be
you
you
can
ask
for
like
okay,
please
I
know
that
I'm
changing
stuff
that
can
be
I,
don't
know
dangerous
and
tricky.
So
now,
I'll
request
that
it
runs
for
jakie
instead
of
it
always
Fujiki
or
whatever,
like
I,
think
we
can.
There
are
many
of
the
Google
jobs
right
now
that
there
are
exceptions
so
like
some
scalability
ones
and
more
specific,
like
there
are
testing
real
stuff,
but
it's
in
redundant.
F
A
If
you
want
to
do
it
yourself
and
host
it
from
the
comedian
repository
as
just
in
standard
markdown
seems
like
a
reasonable
thing,
but
I
don't
think
we
should
be
publishing
it
to
the
wild,
because
the
the
problem
that
I've
seen
in
a
lot
of
the
communications
that
we
get
on
slack,
is
basically
I
blowed
off
my
own
foot
with
a
shotgun.
Please
help
me
fix
it,
and-
and
the
short
answer
is,
don't
use
a
shotgun.
You
know
as
a
cooking
utensil
or
something
like
that.
A
H
F
F
F
A
Like
yeah
provider,
specific
things,
they're,
not
agnostic
or
general-purpose
across
providers,
well
I
think
we
should
we
could
possibly
link
to
them
and
I.
Think
that's!
Okay,
like
you
know,
if
there's
you
know,
federated
cloud
provider
integrations
I,
think
leading
to
it
makes
a
ton
of
sense,
but
I
don't
think
they
should
be
hosted
or
maintained
in
the
making
a
repo
right.
F
So
I'm
thinking
like
we've,
discussed
this
earlier,
but
okay,
so
we
have
leased,
notes,
they're
building.
This
is
not
good.
This
doesn't
explain
enough
and
doesn't
have
the
right,
I
think
links
either
to
the
community
and
and
contribution
stuff.
Anyway,
we
have
cubed
in
here.
I
would
be
so
we
have
cops
and
cubes
prior
here,
I
think
we.
What
we
could
do
is
like
move.
If
they
don't
like,
you
bet
him
have
a
directory
of
stuff.
They
at
least
it
could
be
top
level
here,
as
mini
cube,
PKI
certificates
is
seems
pretty.
A
F
F
H
This
probably
sick
dogs
that
is
owning
everything
because
we
like
proposed
for
them
to
and
I
organized
the
setup
folder
and
we
didn't
receive
like
any
response,
because
the
evolution
of
the
self
father
is
like
super
chaotic
right
now.
For
instance,
cube
ATM
is
under
the
independent
folder
for
some
reason,
I
don't
know
why
I
mean
the
whole.
The
whole
self
Oldham
needs
a
restructure
and
I'm,
not
sure
why
I
suppose
less
I
go
should
be
the
owner.
H
F
I
think
I
think
that's
a
distinction
between
who
owns
the
directory
structure
and
merging
things
and
who
owns
the
actual
content
and
what
I
mean
is
that
where
they
own
it,
so
the
actual
content
in
everything
under
set
up
so
we
might
have
to
or
we
have
to
collaborate
with
them,
of
course,
to
to
merge
stuff
to
reorder
metadata
and
folder
structure.
Yes,
but
it's
still,
we
that
are
owning
the
content
of
this
thing.
So
that's
what
I
mean
that
like?
F
If
we
see
that
okay,
this
reference
is
a
lot
of
old,
unmaintained
stuff
that
haven't
been
getting
comets
for
years.
Then
we
go
ahead
and
remove
this
dock.
We
promote
the
cube
spray
and
cops
things
up
here
top
level
we
go
ahead,
remove
the
custom
cluster
from
scratch,
because
nobody's
gonna
have
time
to
maintain
it
and
right
now
it's
more
harmful
than
good.
We
revisit
together
with
say,
class
like
with
sig
scalability
building
large
clusters,
section
I,
don't
know.
If
we
is
there
like
thing
is
here:
we
should
we
should
check,
etc.
F
F
Things
like
this
is
this:
actually,
because
we
deprecated
the
years
ago,
we
deprecated
the
cluster
directory
sure
it
might
still
get
your
production
ready
cluster
using
cheeky.
But
cheeky
is
the
cluster
director
plus
other
things
right,
so
do
we
recommend
to
anyone
to
actually
use
get
case
that
I
Oh,
for
example?
Oh.
H
F
E
F
E
A
H
A
Was
told
that
if
sink
docks
creates
a
heat
map
with
a
checklist
that
they
Confederate
this
work
out,
because
apparently,
there's
pressure
with
inside
of
pressure
being
put
to
bear
contributors
to
the
CN
CF
that
they
need
to
actually
execute
and
throw
it
on
versus
doing
the
pay-to-play
thing
so
that
that
might
be
a
good
thing
to
curate
ahead
of
time
is
just
having
a
meta
issue
with
all
of
the
details
that
we've
discussed
here,
broken
down
into
checklist
forms.
So
that
way,
third
party
can
do
the
work.
H
H
Lot
of
problem
is
that
we,
if
we
move
stuff
around
and
like
rename
like
documents
and
move
documents
into
new
folders,
like
the
appropriate
folders
they
belong
to,
like,
for
instance,
key
medium
should
not
be
in
independent.
If
we
do
that,
we
are
going
to
break
like
backwards
compatibility
well
now,.
H
F
F
A
Discussion
as
well
as
this
needs
to
be
brought
up
with
steering
about
deprecation
policy
there
do
you
have
a.
We
do,
have
a
process
that's
lightly
outlined
with
that,
basically,
how
to
move
this
into
an
etic
and
people
can
always
bring
it
back
from
the
Attic
right
like
there's,
nothing,
there's
nothing
that
prevents
folks
from
you
know
like
dusting
stuff
off
and
what
is
what
is
old,
is
now
new
again,
but
that
way,
it's
clear
to
the
community
that
these
things
are
not
actively
being
maintained,
I
preferred
not
to
call
it
graveyard.
We've
actually
had.
F
Yeah
right
so
here
we
have
yeah,
it
was,
it
was
called
retired.
So,
for
example,
doesn't
make
sense
to
move
something
there
I
don't
know
about,
but
like
revisit
the
state
of
our
own
sub
probe,
let's
see
and
also
we
could
have
liked
that
let
the
sub
projects
should
check
in
with
with
this,
in
this
cool
like
once
a
month
or
something
it's
also
a
doable
thing.
A
F
A
It's
the
course
the
kernel
right
like
you,
you
have
a
good
version,
a
version
of
the
kernel
and
then
you'd
start
to
like
your
distro.
Basically
bundles,
like
you
know,
your
basic
tool
chain
would
be
things
like
at
CD
and
and
required
add-ons
like
core
dns,
then
outside
of
that,
you
know
it's
basically
the
the
federated
list
of
packages
that
folks
want
to
curate
and
gets
in
the
weird
distributions
based.
Oh.