►
From YouTube: Kubernetes Kops Office Hours 20180928
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
And
everyone
it
is
Friday
September
28th.
This
is
cops
office
hours.
I
am
your
moderator,
just
in
Santa
Barbara
I
work
at
Google.
There
are
a
couple
of
items
on
the
agenda,
but
if
you
do
have
anything
you
would
like
to
add
to
the
agenda.
Please
do
so
and
do
write
your
name
or
that
looks
like
it.
One
has
written
their
name
in
it
in
his
list,
I
think
otherwise.
I
will
kick
off
on
to
relatively
quick
updates
and
which
I'm
sure
around
people's
minds,
which
are
that
the
work
for
111
is
continuing.
A
I
think
we
had
to
do
a
rebase
on
the
kubernetes
111
branch.
I.
Guess
we
do
that.
Every
time
we
read
update,
cops
to
the
new
releases,
the
best
part
that
we
update
to
Carre's
111
and
there
were
some
changes
in
acquired.
So
that's
been
I,
guess
the
big
item,
but
otherwise
I
think
we're
getting
closer
to
releasing
cops,
111,
alpha,
I,
guess
and
then
we'll
go
through
the
normal
stabilization
validation
process.
So
hopefully
there
we're
trying
not
spending
significant
new
features
into
cops
111,
so
the
it
should
be
hopefully
about
two
weeks.
A
A
This
have
stoled,
which
is
accurate,
I,
think
that
there
is
hope,
which
is
if
we
move
that
to
see
our
DS
I
think
it
will
make
our
processes
all
bit
easier
and
may
actually
unstow
run
block
the
the
cop
server
effort
so
particularly
with
cluster
API.
Moving
to
CR
CR
DS.
There's
not
a
lot
of
point
in
us
not
also
moving
to
see
our
DS.
It
shouldn't
affect
the
SIA.
The
usage
of
the
cops
command-line
tool,
the
cop
CLI
at
all.
A
The
only
difference
would
be
that
we
might
be
able
to
run
on
a
server
or
thank
you
can
run
on
a
server
without
even
and
do
that
pretty
easily
Mikey
are
these
are
actually
pretty
cool
they.
You
can
register
a
CRT
against
your
community
server
and
get
an
extra
type
on
there,
and
you
don't
have
to
do
any
you.
You
don't
have
to
do
anything
and
they
appear
as
like.
A
Real
API
objects
to
the
clients
mostly,
and
that
is
increasingly
true
with
each
version,
and
there
is
a
pretty
strong
worth
stream
happening
in
communities
to
make
sure
these
comparable
in
in
functionality
to
what
you
would
get.
If
you
did
through
the
full
API
machinery-
and
that
wasn't
the
case
when
we
started
along
the
API
machinery
or
API
server
route,
cop
server
routes,
so
I
think
there's
there's
that's
a
good.
A
That's
a
good
path
and
it
looks
like
cluster
API
has
proven
the
way
and
is
going
to
move
to
see
or
disease
and
enjoy
the
benefits
of
that
yeah
I,
don't
know,
has
any
any
thoughts
or
questions
on
those
two
items.
Otherwise
Elena
I
see
I
have
once
again
failed.
You
I
apologize.
I
am
going
to
I'm
I'm,
definitely
gonna.
Do
it
from
you
I'm
guessing.
This
meeting
is
not
going
to
carry
on
for
much
longer,
but
we
will
see.
Is
that
important
yeah.
C
A
I'll
take
a
look
too
I
am
so
sorry,
I
realized
I
was
telling
me
on
mute
as
if
I
might
have
been
talking
over
people.
I
apologize
for
any
I
was
talking.
I
saw
later,
you
go,
I
said
if
I
spoke
over,
I
spoke
up
or
anyone
please
please,
please,
please
go
back
and
speak
up.
I
apologize
for
my
spoke
over
can
I.
B
A
There
so
there
is
a
there
is
a
weird
error
about
like
watch
failed
to
sync.
Is
that
the
one
you're
talking
about
yeah,
so
that
one
is
odd?
I
have
a
meta
fear
so
that
was
introduced.
There
was
a
peer
that
went
in
that
turned
what
was
previously
not
an
air
condition
into
an
error
and
specifically
that
errors
I'm
not
fully
familiar
with
what
what
it
means,
but
it's
something
about
like
the
watch
you're
doing
a
watching
it
nothing
fully
ready
the
cache.
A
Let
me
put
it
ready
and
became
a
hard
error
instead
of
a
sort
of
software
earth
God
written.
What
I
don't
know
is
whether
that's
because
we're
using
n
CD
to
still
in
so
we're
now
testing
113,
which
will
remove
support
from
one
SAT.
It
hasn't
yet
done
so,
and
so
I
want
to
get
us
able
to
run
n
city
3
there.
So
we
can
actually
say
this
happens
with
that.
Cd
3
or
this
doesn't
matter-
that's
83,
because.
A
B
B
A
B
A
Your
tests,
exactly
so
I
think
we
should
continue
to
test
at
cd2
with
cops
and
kubernetes
110
111
m112
for
the
foreseeable
future,
but
one
at
the
moment
the
planet
record
is
113
is
going
to
remove
that
support
for
a
city
too,
so
we
and
it
hasn't
yet
happened.
My
hope
is
at
that
PR
I
think
that
PR
is
only
against
the
113
branch,
so.
A
Yeah
yeah
I
I,
don't
yeah
I,
don't
I,
don't
have
any
other
way
than
other
than
to
I.
Guess
I
gotta
talk
to
the
two
Wojtek
who
wrote
it
and
figure
out
if
he
knows
what's
going
on,
but
I
wanted
to
try
to
get
a
CT
3
running
because
I'm
pretty
sure
it's
gonna
be
related
to
SAT
3,
either
that
or
it's
just
a
timing
issue
which
yeah
they're
always
they're,
also
hard
to
track
down.
So
I
would
like
to
be
able
to
narrow
it
down
to
2
XE
2
versus
C
3
as
well.
D
D
D
D
A
I
think
so
I
think
the
the
concern
is
that
the
PR
is
quite
large
and
Elena.
Helped
me
understand
your
use
case
previously.
I
mean
the
other
thing.
That's
coming
is
integration
with
the
cluster
API
or
the
machines
eBay,
which
should
effectively
take
a
lot
of
this
code
out
of
cops
across
all
providers,
and
so
the
that's
that
is
a
tempting
Mike
yeah
yeah.
B
It's
gonna
happen,
but
it's
not
happy
I.
Think,
like
I
I've
been
waiting
on
that
as
well,
but
I
think
it's
it's
not
gonna
happen
as
fast
as
we
wanted
to
so
I
think
we're
gonna
have
to
move
forward
with
this.
Like
we
I
I
think
we
we
should
figure
out
some
way
to
at
least
if
this,
and
even
if
we
don't
you
know,
want
to
envision
it
being
here
for
the
long
term.
I
think
I
am
inclined
to
agree.
B
A
Yeah,
it's
it's
actually
really!
So
it's
yeah!
It's
pretty
nice
in
that
at
least
four
nodes
with
cluster
API.
You
can
more
or
less
add
nodes
to
an
existing
cluster
wherever
you
find
it
like
more
or
less
with
like
very
little
integration,
so
that's
nice
and
sort
of
from
the
spot
instance
point
of
view
and
cost
savings.
I'd.
Imagine
the
majority
of
your
costs
come
from
notes,
rather
than
from
the
masters,
but
I
agree
Eric
that
we
should
probably
like
merge
this
and
look
at
at
some
stage.
A
B
A
A
B
A
D
B
Let's
get
your
PR
taken
care
of,
let's
get
it
done
today
soon
and
I
will
take
a
look
at
it
after
this
call
and
give
you
any
final
comments:
we'll
try
to
get
a
merger
system
as
possible
and
then,
as
as
cluster
API
sort
of
solidifies
and
and
it
is
better
to
find
them
which
is
going
to
work
on
that.
But
and
then
we'll
we'll
have
you
refactoring
exponents
and
everything
else,
but
it
should
be.
Hopefully
it
will
be,
will
have
sort
of
a
template
because
it
should
be
relatively
similar
for
cloud
providers.
B
Reaches
so
I,
this
is
a
sort
of
a
meta
topic
or
it's
a
sort
of
a
half
topic.
We
are
in
discussions
with
Paris
and
the
other
like
not
sure,
with
what
the
sig
is,
but
it's
like
community
involvement
we're
trying
to
get
an
intern
in
an
out
Ritchie
funded
intern,
and
so
we
are
in
communications.
We
may
there's
a
possibility
that
we
are
coming
into
it
a
little
bit
too
late,
but
we
have.
B
We
have
something
like
we
already
have
sort
of
a
draft
proposal
in
there
and
I'd
like
to
I
was
I
was
thinking
about
making
sort
of
a
again
outreach
or
a
community
in
directory.
In
our
repository
where
we
could
have
things
like
this
like
which
would
serve
as
sort
of
a
way
to
refactor
some
of
the
development,
you
know
getting
involved
documentation
as
well
as
make
it
more
accessible
to
the
community.
So
hopefully
we
are
not
too
late.
That's
that's!
B
Tvd
I,
actually
pinged
Paris,
just
a
few
minutes
ago
and
I'm,
hopefully
getting
at
a
response.
So
I'll,
let
everybody
know
when
I
hear
something,
but
if,
if
you
hear
me
asking
about
it
or
anyway,
Chris
Chris
has
been
really
helpful
with
that
Chris
and
Mike
have
been
super
helpful
with
that
as
well,
so
we've
been
sort
of
pushing
that
forward
and
anybody
that
is,
if
you
see
us
looking
for
for
help
on
that,
we
would
very
much
appreciate
it.
A
C
A
This
would
show
you
a
diff
of
kubernetes
objects
and
we
so
sorry
coop
cuddle
day,
for
she
would
different
than
kubernetes
objects
or
Cooper.
Don't
like
I,
don't
remember
the
syntax,
but
the
it
will
show
you
different
occur
brains
objects,
so
it
would
show
you
in
our
case
clusters
and
instance,
groups
and
I
think
secrets.
So.
A
A
Tasks
as
kubernetes
api
objects
and
then
so,
in
other
words,
so
today,
for
example,
you
change
a
security
group
and
we
have
a
task
called
security
group.
I
think-
and
we
show
you,
the
diff.
We
made
security
group,
a
kubernetes
api
object
or
c
rd.
Then
we
could
use
cout
cuddled
if
to
show
you
that
I
honestly
make
a
ton
of
work
to
do
it
right
now,
but
that's.
B
A
It
would
be
a
incredibly
fun
fact
session.
You
like
take
our
tasks,
convert
it
into
a
CR
D
and
then
take
the
task,
reconciliation
logic
and
put
it
as
a
controller.
Yes,
I
think
it
would
actually
work.
It
would
be
interesting,
but
it
would
be
it's.
It's
actually
really
interesting,
like
meta
suggests
that
you've
made
there
Elena,
which
is
like
yeah,
we
should
yeah
like
it.
Is
there
yeah
it's
some
it
is,
it
would
be,
it
would
be
great.
A
B
A
The
I
guess
the
chat,
the
challenge
is
gonna,
be
we
want
cuffs,
continue
to
run
without
a
without
a
server.
If
we
don't
need
to
require
a
kubernetes
api
server
in
order
to
run
the
cops
command
line
tool,
maybe
we
get
always
like
running
it
embedded
or
something
which
is
sort
of
what
the
cost
di
does
like.