►
From YouTube: Kubernetes SIG Cluster Lifecycle 20180116
Description
Meeting Notes: https://docs.google.com/document/d/1deJYPIF4LmhGjDVaqrswErIrV7mtwJgovtLnPCDxP7U/edit#heading=h.hmxea1qykdh
Highlights:
- Old meeting notes have been archived at https://docs.google.com/document/d/17J496IR2tXKw7k97fxwz2KUWOf9rpBD3pIEsmDiJQSw/edit
- Cleanup of the cluster directory is proceeding, including a PR out to delete all of the salt configs
- Check out the support for multiple clusters in kubeadm-dind
- PR & Issue triage should happen today
A
Hello
and
welcome
to
the
sink
cluster
lifecycle
meeting
for
Tuesday
January
16
2018
like
last
week.
I
think
we
should
start
up
start
off
with
some
follow-ups
from
last
week.
So
the
first
thing
I
put
on
here
was
sorting
the
backlog,
so
Tim
I
think
you
were
planning
on
sorting
the
backlog
for
the
cuvette
Henrico
and
we're
looking
for
help
for
sorting
and
triaging
issues
and
the
Camaro's
repo
and
had
a
couple
of
volunteers.
So
I
wanted
to
follow
up
and
see.
If
we
made
progress
on
that
or
if
we
should
keep
checking
in
I.
B
Need
to
poke
them
I
need
to
folks
into
to
take
sections
of
the
backlog
out
from
me
and
then
just
go
ahead
and
do
it
so
what
I
should
do
today?
What
I
will
commit
to
today
is
go
through
and
send
an
email
to
the
list
and
poke
the
people
that
had
said
they
would
volunteer
and
get
that
done
today.
It's
my
bad
for
not
getting
that
done
this
week.
A
Well,
that's
that's
why
we
remind
you
every
week
so
maybe
make
some
slow
progress
right.
The
other
thing
I
found
looking
through
last
week's
meeting
notes
was
we
talked
about
sending
out
a
survey
for
what
we
think
should
be
part
of
GBM
nga.
So
I
put
your
name
on
this
time
because
you
would
make
a
suggestion
to
send
it
out,
not
sure
we
actually
got
a
volunteer
to
send
it
out
or
not
so
I
wanted
to
see
if
you
remembered
more
clearly
than
I
did
or
the
meeting
notes
specified,
I
thought.
B
You
said
you
were
gonna,
send
it
out,
but
no
I
know
I
know.
I
did
say:
I
was
gonna,
send
it
out.
I
recommended
the
idea
to
make
sure
that
when
we
push
out
whatever
a
version
of
what
we
call
GA
is
that
it
meets
whatever
the
community
has
deemed
worthy
of
that
title
right,
because
we
don't
want
to
push
something
out
into
GA
and
have
somebody
say,
hey,
it's
missing
feature
XYZ.
A
Okay,
yeah
I,
don't
recall
volunteering
to
send
that
out.
I,
don't
think
I
will
have
time
to
do
it
this
week.
So
if
anybody
has
a
time
to
joe
mention
in
chat
that
this
group
is
taking
GA
more
seriously
than
a
lot
of
other
places,
humanities,
I
think
tim
has
mentioned
in
the
past.
That
is
on
purpose
and
then
we're
trying
to
be
be
pretty
cautious
about
marking
things
GA.
Maybe
that's
the
wrong
approach
in
the
communities.
A
Community
I
think
we're
we're
really
trying
to
make
sure
we
have
something
that
people
can
can
use
and
rely
on
here.
As
I
was
saying,
as
I
was
saying.
If,
if
somebody
wants
to
volunteer
to
put
together
a
survey
I,
don't
think
it
would
be
too
difficult,
we
could
send
it
out
to
communities,
Devon
criminals,
users
and
ask
folks
to
forward
it
belong
to
anybody
else
who
they
think
should
would
be
interested.
A
All
right
I'll
hit
up
next
week,
so
next
thing
on
the
agenda
was
I
have
been
chatting
with
Lucas
a
little
bit
off
line
and
we,
since
we
switched
the
cube
she
dived
in
focus.
Meeting
on
Wednesday,
is
to
have
a
new
agenda.
We
decide
to
create
a
new
dock,
to
attract
notes
going
forward
and,
as
we
were
doing
that,
and
it
was
the
beginning
of
a
new
year,
it
made
sense
to
sort
of
the
same
thing
with
the
stock.
A
So
what
I've
done
is
I
made
a
copy
of
the
stock
and
ripped
out
all
of
the
old
meeting
notes
from
this
talk
and
left
them
in
the
copy,
and
if
you
scroll
down
in
this
talk
not
too
far
down
to
page
seven
you'll
see
a
link
to
the
old
meeting
words
and
that
links
to
the
meeting
notes
from
2016
and
2017.
We
use
a
stock
for
2018
and
then
we'll
probably
want
to
cycle
about
once
a
year.
A
So
the
only
concern
with
this
I
have
is,
it
will
break
some
links.
So
I
know
that
all
of
the
the
YouTube
videos
that
I
upload
I
link
into
the
heading
of
the
dock,
for
where
that
meeting
took
place
and
those
links
will
break.
But
the
stable
link
for
the
meeting
notes
from
the
calendar
invite
and
so
forth
will
still
point
to
the
right
place.
So
we
may
want
to
do
it
slightly
differently
next
year,
when
we
rotate
and
actually
create
a
new
document,
the
calendar,
as
opposed
to
making
a
copy
for
the
archive.
A
A
So
if
you
guys
are
watching
the
PRS
go
by
we're
deleting
all
of
the
old
cubicle
providers
that
have
been
deprecated
for
a
long
time,
I
think
we
delete
four
or
so
of
them
in
the
last
week
and
the
mic
has
a
PRI
right
now
to
delete
all
of
the
salt.
So
all
the
salt
configuration
will
be
disappearing
here,
the
next
couple
of
days.
It's
blocked
on
merging
right
now,
because
we
use
that
some
of
some
of
that
stuff
internally
engine
key
make
some
internal
changes.
A
First
before
we
can
go
with
it,
but
once
that's
done,
we
should
be
able
to
delete
all
the
salt
code.
We're
left
with
a
couple
of
providers
that
are
used
for
end-to-end
tests
and
we're
we're
gonna
slowly
be
ripping
those
out
over
the
next
probably
release
or
two
is.
We
can
place
things
with
the
cost
or
API.
But
if
you
look
the
cluster
directory
now
it's
much
much
slimmer
than
it
was
a
week
ago,
which
is
pretty
nice.
A
This
is
also
sort
of
follow
up
I'm
trying
to
get
rid
of
get
Kas
IO,
which
is
basically
automating
the
install
via
queue
up.
If
you,
you
know
curl
pipe
and
we
still
use
it
in
a
little
bit
of
our
end
and
I
think
infrastructure.
So
we
didn't
fix
that.
But
once
that's
gone,
that's
another
thing
we
can
do.
C
Yes,
I
got
a
little
obsessive
last
night
and
I
pushed
it
I
put
in
a
PR
for
this
change
to
namespace
all
of
the
resources
that
are
created
by
the
branches,
a
DM
DM
cluster
project-
if
you
guys
haven't
used
this
before,
it's
really
useful
way
to
speith
your
local
or
a
remote
docker
in
docker,
multi,
node,
kubernetes
cluster
for
the
purpose
of
building
and
testing
code,
and
also
playing
with
things
like
scheduler,
bugs
and
etc,
and
yeah
I
could
have
found
that
one
cluster
wasn't
enough.
C
Sometimes
I
wanted
to
test
a
change
and
kind
of
like
try
a
different
permutation
and
compared
them.
So
this
was
yeah
just
adding
a
little
bit
of
name
spacing,
and
if
anybody
wants
to
kind
of
throw
some
rocks
at
it,
I
really
appreciate
that
so
not
in
kubernetes
work
it's
in
the
and
to
sorc,
but
I
mean
their
clas
like
really
easy
to
mess
with
appreciate
any
feedback.
If
anybody's
curious.
B
There's
been
a
long-standing
goal
or
objective
of
mine
to
kill
the
Koopa
scripter
II
that
a
lot
of
people
use
for
testing,
not
the
coop
script
or
the
local
cluster
up
with
d-ind
cluster
I.
Don't
know
if
you
have
any
bandwidth
recycles
for
that,
but
that
would
be
it
would
be
awesome
if
that
would
be
possible.
C
C
B
D
D
You
can
part
from
just
the
project
readme,
so
maybe
if
we
can
discredit
very
much
that
lives
in
the
reef
in
the
dot
sweeper
that
would
be
great
and
because
I,
we
also
have
already
long-running
issue
in
the
Cuban
DM
repo,
which
basically
tries
to
make
death
development
workflow
be
easiest,
so
we
can
once
we
have
the
docks
for
that.
I
think
the
current
might
be
FDIC
I'm,
like
functionality
is
that
we
can
once
we
have
this
episode.
C
A
This
actually
ties
in
really
well
with
the
last
conversation,
because
there
are
sort
of
two
cube
up
directories
that
I
was
old.
I
couldn't
delete,
one
which
is
being
used
by
some
cubes
for
users
for
local
testing,
because
cube
spray
was
deemed
to
be
too
heavy
weight
for
dev
testing,
even
though
they
use
cube
spring
prod.
A
C
Have
a
some
very,
very
considerate
use
cases
is
what
I
found
lots
of
different
ways
to
create
volumes
that
have
your
mounted
like
kubernetes
binaries,
and
things
like
that,
so
that
you
can
update
the
builds
binaries
and
keep
the
cluster
up
and
things
I
think
that
there
needs
to
be
a
lot
more
documentation.
It's
good
I
found
that
it's
it's
definitely
lacking.
It
seems
like
there's
quite
a
bit
of
tribal
knowledge
around
how
they
use
a
lot
of
tooling
so,
but
that's
something
that
we
can
contribute
and
I'm
sure
it
leads
go
well
received.
A
C
A
A
B
D
So
one
thing
which
maybe
it's
more
suitable
for
tomorrow's
implementation
meeting
but
Lucas
mentioned
that,
like
one
of
the
action
items
were
like
like
changes
in
that
came
out
of
cucumbers,
that
core
OS
specifically
think
that,
like
self-hosted
EDD
is
no
longer
stable,
which
we
you
know,
we've
had
rumblings
of
for,
like
a
while
now
so
it'd
be
good
just
to
try
and
like
identify
harkening
towards
that.
Because
of
coral
s
no
longer
lie
and
feel
that
it's
stable
and
probably
no
longer
innovative
aspects
of
cycles
to
maintain
it.
D
D
Yeah,
which
is
I,
mean
yeah,
it
kind
of
makes
sense,
so
we
should
probably,
like
so
I-
have
that
long-running
pull
request,
which
I'm
more
than
happy
to
close,
because
I'm
not
sure
I
have
the
scientists
to
keep
that
up.
So
if
we,
if
we
like
I
sort
of
calibrate
our
opinion
with
coral
s,
and
we
say
that
it's
not
really
not
only
production
that
maybe
not
deaf,
where
are
they,
then
I
can
go
ahead
and
close
that,
maybe
that's
it's
true
like
a
conversation
as
well
tomorrow,
yeah.
E
D
D
So
we
were
just
talking
about
caresses
aptitude
to
build
self
hosted
at
C
D,
and
if
it's
going
to
be
sunset,
and
so
we
could
have
probably
like
a
cure
for
discussion
tomorrow,
as
I
said,
it
might
be
useful.
Just
have
like
an
inside
to
our
opinion
of
what
coral
Isis
thinking
is
towards
self-hosted
SED
yeah.
F
A
D
We
don't
necessarily
need
to
speak
about
an
hour
late.
Maybe
are
you
free,
like
this
I
think
it's
this
time
tomorrow
for
the
for
the
implementation
meeting?
So
if
you
all
thanks
for
that
Ryan,
maybe
we
can
sort
of
get
more
into
depth
of
our
the
reasoning
behind
the
Sun
setting
it
up.
I
thought
it'd
be
cool
with
you.
You.
F
A
Alright,
I
don't
see
any
other
agenda
items,
so
I
think
we're
gonna
gonna,
call
it
and
have
a
nice
short
meeting
to
kick
off
the
week
for
folks
in
the
u.s.
that
had
a
holiday
yesterday
and
let
folks
catch
up
on
all
of
their
backlog,
but
I'm
sure
it's
built
up
over
the
weekend
with
all
of
the
activity
from
the
people
in
Europe
still
been
working
on,
kubernetes
only
been
off
vacation.
So
thanks.