►
From YouTube: Kubernetes 1.12 Release Team Meeting 20180723
C
C
B
B
Started
typing
a
few
names
and
of
attendees
in
the
minutes,
but
I'll
actually
I
think
I'm.
Gonna.
Leave
that
to
you
all
to
do
so.
I
can
really
think
about
the
words
coming
out
of
my
mouth
say
things
that
maybe
make
sense
all
right.
It's
a
it's
a
minute
or
almost
two
after
so
we're
gonna
go
ahead
and
get
started
I'm
guessing
this
will
probably
be
a
relatively
short
meeting
given
where
we
are
in
the
cycle,
but
wanted
to
go
over
basic
test
status.
B
B
So
we
will
sort
of
repeat
this
meeting
tomorrow
morning,
extra
early
and
we'll
see
what
we
get
for
additional
shadows
from
the
other
time
zones
calling
in
and
we'll
see
how
that
goes
for
the
next
couple
weeks
and
give
it
a
go
see
what
happens,
and
that
was
the
main
thing
that
I
wanted
to
share
with
folks
and
then
go
off
into
test
status,
features
and
issues.
So
I
see
yes,
Mohammed
is
on
so
can
we
start
with
you
with
the
status.
G
Yeah,
okay.
So
as
far
as
they're
concerned,
we
have
we've
heard
about
number
of
failing
tests
and
also
on
what
master,
as
well
as
Oh
agreed.
So
a
couple
of
issues
were
closed
and
some
of
them
are
being
tracked.
That
is,
there
are
peers
open
for
kind
of
fix
them.
I've
seen
a
bunch
of
failures
on
upgrade,
for
which
I
will
be
I've
already
raised,
issues
for
a
couple
of
them,
I'll
be
raising
more
and.
G
B
B
G
A
stinks
and
I'm
and
I'm
able
to
recover
the
people
like
I,
have
got
good
response.
Another
issue
that
already
raised
that,
mostly
because
of
carrying
around
sick,
custard
or
cycle
and
seg
networking,
and
there
are
a
bunch
of
them
on
sick
storage
as
well.
Otherwise,
everything
else
is
looking.
Okay
and
most
of
our
tests
are
around
DC
idk
failures.
F
G
B
B
We
could
do
it
tomorrow
or
next
week
as
well.
I
just
I
want
to
make
sure
that
folks,
broadly
across
the
team,
have
gotten
a
sense
of
test
grid
and
where
to
click
on
what
and
I'm
kind
of
thinking
out
loud
going
through
it,
because
it's
some
of
the
things
aren't
always
immediately
obvious.
We
can
hit
that
a
little
bit
later
so
yeah,
okay,.
B
E
I
think
I
put
all
my
notes
on
it
or
wrongly
break
they're
in
the
notes.
Those
are
all
mine
under
yeah.
I
should
have
put
that
under
issues
instead
of
feature
collection,
yeah,
I,
love,
more
yeah.
All
of
that
so
I'm,
seeing
when
I
compare
the
open
features.
E
That
whoa
would
what
is
happening
so
currently
I'm.
Only
seeing
like
four
open
bugs
two
of
them
need
the
milestone,
approval
and
I'm
not
actually
sure
how
that
works.
So
I
called
those
out
individually,
I,
don't
know
if
I
should
do
anything
there.
There
is
one
that
I
noticed
about
the
persistent
of
volume
creation,
which
I
think
should
go
away,
because
somebody
did
some
research
and
found
a
work.
E
So
I'm
going
to
follow
up
and
say:
hey:
can
you
either
just
post
a
workaround
or
figure
out
what
what
the
deal
is
here,
because
I
think
the
last
activity
was
seen
about
a
week
ago,
which
I
just
generally
get
the
feeling
that
everyone
is
just
still
kind
of
like
on
the
slow
uptake
on
most
of
these,
which
I
guess
is
fair,
given
where
we
are
in
the
release
cycle.
Yeah.
B
B
There
was
there's
some
conference
travel
that
folks
had
last
week
and
also
this
week,
so
I
think
there's
a
number
of
things
going
on
there,
and
it's
also
at
this
time
of
the
cycle.
It
seems
like
traditionally
there's
not
as
much
issues
and
then
really
starting
last
cycle
last
cycle,
I
think
was
the
first
full
cycle
where
issues
weren't
required
on
PRS,
but
definitely
there's
a
lot
less
issue
volume
going
on
and
it
picked
up
a
bit
towards
the
end
but
versus
PRS.
B
It
seems,
like
things
are
skewing
in
the
direction
of
PRS
alone
and
obviously
at
the
point
in
the
dev
cycle.
Right
now,
people
are
working
on
things
in
advance
of
PRS
on
the
milestone
part.
That
really
should
be
the
cigs
flagging
it
and
approving
it.
So
if
it's
not
just
being
the
the
sigelei,
the
appropriate
sigelei
there
and
say
like
hey,
is
this
targeted
for
for
the
milestone
and
don't
forget
to
label
and
approve.
E
And
then
I
was
looking
at
the
features
collection
as
well.
I,
don't
know
if
that
is
necessarily
my
job,
but
I
wanted
to
see
what
the
features
were.
E
There
is
still
a
little
bit
of
a
discrepancy
between
tracked,
yes
and
open
feature
issues
in
KK,
which
I
don't
know
where
the
discrepancy
comes
from,
it's
possible
that
some
of
those
issues
are
not
in
court
kubernetes
are
we
have
some
to
have
some
automation,
I
didn't
even
see
someone
there
weren't
even
like
mentions
and
some
of
the
features
issues,
so
it
was
really
to
track
down.
There's.
B
A
distinction
between
being
in
the
milestone
and
being
in
features
which
I
don't
think
I've
seen
anybody
declare
like
what
makes
something
a
feature
versus
a
bug-fix
and
like
historically
in
software
engineering.
I
think
this
is
something
we
do
bad
like.
Oh
it's
too
late
to
get
a
feature
in
well.
This
is
actually
a
bug
fix,
so
I.
E
B
H
B
H
Away
well,
then,
I'm
not
gonna,
be
able
to
give
too
much
of
a
log
updates,
considering
I'm
just
now
getting
ramped
up
on
everything.
That's
been
happening,
but
Steven
went
through
on
a
tagging
spree
last
week
around
Thursday,
so
I
believe
that
he
is
got
a
lot
of
things
that,
in
the
features
repo
tagged,
he
gave
me
and
I
just
started
it
about
two
hours
ago,
just
kind
of
doing
a
1.11
triage.
H
So
looking
at
a
bunch
of
the
things
that
were
milestones
and
1.11
that
didn't
actually
make
it
and
seeing
if
there's
any
current
updates
to
see
if
they
are
going
to
make
it
in
112,
so
I'm
currently
3/4
of
the
way
the
list
through
of
just
seeing.
If
there's
any
updates,
if
they're
going
to
have
any
sort
of
changes
in
released,
if
they're
gonna
be
in
stable
or
beta
or
anything
like
that
or
if
they're
just
gonna
hold
steady,
he
also
gave
some
leeway
to
help
the
other
guys
that
are
doing
this
as
well.
H
B
Those
are
specific
to
things
that
are
have
been
entered
into.
The
Kay
features
repo
as
issues,
but
don't
have
a
milestone,
I'm
assuming
that's
what
he's
targeting
correct
the
the
gap
that
Gwyn
was
noticing
between
KK
and
other
repos
increasingly
and
Kay
features
is
one
where
Steven
had
mentioned.
He
and
I
think
Jace
were
working
on
something
for
a
proposal
there,
but
especially
as
the
we
do,
this
splitting
of
the
mana
lift
it
becomes
fuzzy.
We're
features
are
defined
and
from
a
least
release
team
perspective.
B
H
That's
and
there's
some
changes
that
we've
introduced
into
that
as
well.
In
regards
of
being
able
to
see
what
something
is
moving
towards
something
because
we
used
to
be
able
to
see
like
it
would
just
say:
theta,
we
don't
know
like
what
does
I
mean
it's
moving
to
beta
or
it's
like
you
know.
We
didn't
really
know
what
it
meant
so,
there's
some
things
right
now
as
being
in
tracked
or
in
progress
or
graduating
and
there's
actually
another
great
tab.
That
was
just
created.
A
That's
just
a
historical
artifact
of
a
way
that
we
at
one
time
try
to
store
and
share
github
queries
so
that
we
all
were
from
the
same
I,
guess
toolset
and
so
that
someone
more
enterprising
than
ourselves
back
then
could
do
some
kind
of
historical
analysis.
Since
point
of
time
is
well,
it's
good
hubs.
Api
is
very
much
point
in
time
and
so
much
historical
yeah.
B
Later
in
the
cycle
and
I
I
think
probably
roughly
from
feature
freeze
onward
will
start
doing
that
and
the
the
burndown
minutes
here
and
that
Josh
had
carried
that
on
more
or
less.
We
didn't
quite
do
the
the
full
every
sig
leaderboard,
but
just
to
kind
of
keep
things
visually
condense
just
highlighted
what
was
what
each
time
and
yet
it's
it's
a
pain.
You
know
we
can
share
the
queries,
but
the
results
are
actually
what's
particularly
interesting.
So
having
the
link
to
the
query
is
useful,
but
then
also
having
the
human
condense
it
into.
B
Last
week
we
had
a
bit
of
a
hiccup
on
the
1.11
dot,
one
release
so
caleb
had
just
finished,
doing
a
brain
dump
to
Doug
MacEachern
on
branch
management
role
and
talking
about
how
some
of
the
artifact
creation
won't
be
too
much
of
an
issue.
But
there's
a
little
bit
of
a
google
or
specific
manual
kick
some
stuff
off
thing
and
it'll
happen
just
at
the
end
of
the
release.
But
then
one
dot
11.1
ran
into
the
same
issue
and
we
all
kind
of
realized.
B
A
Nothing
other
than
to
note
it's
actually
we're
kind
of
in
a
lie.
Guess
what
I
would
consider
kind
of
a
regressed
state
from
where
we
were
I
mean
there's
still
ongoing
discussions
internally,
but
certainly
it
would
be
nice
to
get
back
to
the
place
where
we
were
in
terms
of
cutting
there.
The
patch
releases
I'm,
hoping
that
we
won't
be
so
this
team
won't
be
negatively
impacted
by
that.
So
my
fingers
are
still
crossed.
A
E
B
This
is
actually
something
that
is
reasonably
well
documented,
I
feel
like
and
it
it,
but
you
wouldn't
have
seen
it
necessarily
because
there
is
a
specific
lead
role
for
it.
So
we
have
okay,
the
release,
branch
manager
and
they're
looking
at
managing
what
patches
are
going
into
the
release
branch,
especially
once
it's
been
effectively
forked
and
things
have
to
get
cherry
picked
back
into
it,
but
that's
only
up
until
release
and
then
the
roll
typically
hands
off
I.
B
Think
looking
back
through
the
list,
it's
usually
been
somebody
different
who
manages
the
patches
coming
into
that
buy
cherry
pick
back
ports
for
the
next
nine
months
or
three
three
releases.
So
we
have
somebody
in
that
role
from
the
111
release
team.
We
have
somebody
tagged
for
it
in
the
112,
but
as
of
today,
the
one
that
eleven
teams
person
handles
it
and
our
team
person
isn't
yet
into
their
position.
B
B
Yes,
this
is
something
that
should
be
backported
and
labeling
it
as
such
and
there's
a
set
of
labels
that
mark
things
as
cherry
pick,
candidates
and
cherry
pick
approved
and
then
on
a
on
the
cadence
of
a
couple
weeks,
or
so
the
the
given
patch
manager
for
the
one
at
eight
nine.
Well,
what
would
be
eight
nine
ten,
eleven,
currently
nine
ten
eleven
are
doing
dot
ports
and
putting
out
those
releases.
B
I
Apologies
I
was
trying
to
and
you
let
me
move
over
here
to
where
I
can
be
seen
and
heard.
So
this
was
just
something
that
we
kind
of
came
up
with
on
the
sigdoc
scheme,
because
actually
no
one
is
a
technical
writer
on
our
team.
Okay,
we'll
have
plenty
of
review,
but
we're
all
software
engineers.
So
we
figure
why
not
introduce
more
software,
so
we
wrote
a
little
serverless
nodejs
app.
I
I
I
So,
basically,
what
it
does
is
it's
it'll
periodically
pull
the
graph
girl
API,
looking
at
the
112
milestone
and
KK
specifically,
and
just
update
this
table
with
with
the
various
stat
I
status,
to
what
statuses,
obviously,
that's
the
plural
of
status.
Goods
Lord
I,
need
sleep
and
then
we're
using
this
to
sort
of
triage
and
figure
out
what
we
actually
need
to
go
bother
people
for
docs,
because
it
turns
out
it's
generally
more
than
what's
in
Kay
features
but
less
than
the
total
set
of
all
PRS
inside
of
a
milestone.
I
H
I
B
On
it,
does
it
so
like
just
on
picking
on
the
first
one
there
in
the
list,
five
six,
three
five
eight.
If
the
milestone
label
got
removed
from
that,
would
it
still
show
up
in
this
table
and
have
something
for
a
flag
for
a
few
days
or
a
week
or
something
saying
like
hey?
This
thing
disappeared,
yeah.
I
B
That's
one
of
the
big
gaps
is
that
you
we're
losing
things
potentially
and
then
I
think
on
bug.
Triage
I
particularly
saw
this
where
something
would
come
up
a
label
whether
intentionally
or
inadvertently,
gets
removed,
and
then
a
week
later
it
pops
back
up
was
like
hey
here's,
this
major
thing
that
nobody
did
anything
on,
but
it
wasn't
showing
in
our
queries
anymore,
so
that
right
and
if
that
gray,
space
of
persistence
in
there
is
important
for
us
to
do
good
risk
management,
I
feel
so.
I
I
guess
the
good
news
is:
if
it
doesn't
remove
anything,
then
the
risk
management
is
simply.
You
might
be
over
checking
on
PRS
that
are
now
no
longer
useful,
but
essentially
anything
that
comes
under
the
the
milestone
of
the
112
release
will
appear
in
this
spreadsheet
in
one
form
or
another,
and
then
as
it
progresses
through
the
milestone
anytime,
that
the
last
activity
of
the
PR
is
not
the
last
activity
in
the
table.
It
just
gets
updated
with,
with
whatever
the
relevant
details
are,
which
is
which
is
good
enough
for
our
use
case.
B
I
F
B
We
could
still
notice
if
something
came
back
up,
that
we'd
have
our
sort
of
triage
Diagnostics
that
are
still
that
we
could
look
back
on
notes,
wise,
hey,
I,
I
know
that
Josh
was
trying
to
build
some
automation
and
was
bumping
into
some
issues
with
github
specifically.
But
this
is
definitely
an
area
where
we
have
a
need
for
a
better
tool.
So
I
I'm
gonna
play
with
the
sum
over
the
next
week
or
two
curious
eyes
and
Mohammed
and
koi
he'll.
I
That
gets
merged
into
the
master
branch
of
the
github
repo
I
put
in
there.
It
gets
auto
deployed
to
the
service
app.
So
if
you
want
to
it'll
it'll
literally
start
working
the
next
time
you
merge
it,
it
snowed.
So
don't
kill
me
I
know:
I
tried
to
write
it
and
go,
but
as
it
turns
out,
there's
no
good
graph.
Ql
library
and
the
marshalling
and
unmarshal
of
of
JSON
from
graphic
eval
was
very
dicey,
so
I
just
decided
to
use
node.
F
I
So
I
can
that
the
the
actual
graph
QL
query
it's
hard
coded
into
the
that's
actually.
Another
reason
why
I
chose
notice,
because
I
really
wanted
to
use
graph
QL,
because
I
didn't
need
the
lion's
share
of
all
the
PO
requests.
I
just
needed
the
thing
I
needed
or
the
fields
I
needed.
So
here's
the
actual
query,
I'll
put
it
in
the
chat.
I
F
One
thing
I
have
is
for
for
features
or
any
new
change.
That's
going.
We
also
want
to
make
sure
that
test
coverage
is
there
right.
So
we
asked
the
authors
to
like
mention
that
is
just
needed,
for
this
is
not
needed
for
this
in
the
features
and
the
feature
bug
itself.
Is
there
a
way
to
pull
that
information?
I
know
it's
like
freeform
text,
there's
no
like
specific
field
where
they
would
add
that
information,
so
I'm
just
curious
how
you.
I
Could
parse
through
the
comments
and
then
search
for
as
long
as
it's
sort
of
like
a
semi-regular,
especially
I'm,
not
going
to
try
to
make
a
natural
language
processing
project
but
but
I
think
we
could
totally
if
there
was
like
a
somewhat
regular
thing
like
this
needs
tests
or
some
like
hyphenated
snippet
of
text
and
absolutely
it
could
parser
and
do
that.
The.
I
I
H
B
B
B
That's
on:
let's:
let's
do
that,
I!
Think
from
my
perspective,
it
looks
like
there
would
just
be
a
few
little
bits
of
additional
information
and
then
we'd
we'd
want
to
kind
of
agree
on
how
we
use
it
so
doing
the
triage
who
owns
it
and
updates
of
the
notes
or
the
the
human
editable
non
queried
fields
that
we
have
some.
So
it's
not
just
chaos
in
there,
but
right
yeah
this.
This
would
be
useful
and
I'm
presuming
so
do
you
have
some
way
of
controlling
access
to
it?
Yes,.
I
I
I
F
I
I
C
A
B
D
One
other
quick
thing:
just
well:
I'm
gonna
go
ahead
and
finally
follow
through
on
changing
priority,
failing
tests
to
kind
filling
tests.
There
are
a
couple
Doc's
I'm
gonna
have
to
update,
in
the
sake,
release
repo
I'm,
assuming
you're,
all
cool
with
just
sort
of
a
find/replace.
But
if
you
care
I
can
also
put
in
a
historical
hey.
This
used
to
be
so,
and
so
now
it's
so,
and
so
it's
get
the
histories
there,
yeah
alright,
so
email
out
to
akkada
later
today,
cool.
B
Stephen
augustus
also
has
a
similar
thing
out
there
around
some
of
the
the
features
labels
but
yeah.
So
continuing
I
mean
this
is
a
work
in
progress,
cleaning
up
labels
and
streamlining
processes,
making
things
more
obvious
and
understandable
to
new
users
as
well
as
us,
I,
don't
know
how
many
times
I
eat
that
one
wrong.
B
All
right,
well,
I,
guess
with
that
I
will
call
it
and
anybody
who's
able
to
attend
tomorrow
morning
will
kind
of
repeat
this:
a
bit
go
over
current
status.
Not
a
lot
will
have
changed
in
a
day,
but
share
the
the
things
that
we've
talked
about
with
the
folks
and
other
time
zones
and
see
what
other
additional
input
they
might
have
on
questions
and
processes
and
things
all
right
have
a
great
day.
Everybody
I
thank.