►
From YouTube: Heptio Ark Q+A 20181009
Description
Join us here and on Slack for an open question and answer session with the Heptio Ark Developers. All users and expertise levels are welcome! https://kubernetes.slack.com/messages/ark-dr
Full notes at: https://github.com/heptio/ark-community/blob/master/meetings/2018/qa/1009/README.md
A
Who's
gonna.
Actually,
before
you
start
his
reading
emails,
we
started
alright.
B
B
B
Something
that
we
also
do
is
we
kind
of
basically
pay
attention
to
the
slack
channel
throughout
the
month
in
between
office
hours,
and
then
we
do
kind
of
lean
out
questions
that
we
think
are
reusable
and
useful
for
the
community
and
then
kind
of
reiterate
them
here
in
a
more
high
and
with
settings.
So
the
way
this
works
as
we
do
this
once
a
month.
So
if
he
at
any
point,
if
you're
using
art,
can
you
feel
like
you're
getting
stuck
or
anything
like
that,
you
can
always
join
this
call.
B
You
can
actually
join
the
zoom
directly
if
you
want
we'll
just
invite
you
and
we
can
walk
through
your
problems
or
just
answer
your
questions.
So
we
have.
We
have
some
announcements,
so
shout
outs
and
that
will
get
into
questions.
Has
anyone
joined
since
we
started
here
that
needs
to
be
introduced
themselves.
B
A
B
If
someone
asks
a
question,
it
would
like
intercept
them
and
it
will
send
them
a
whisper
that
has
the
answer.
So
we
use
this
in
the
kubernetes
channel
a
lot
for
like
hey
what
what
are
the
dates
for
cube
con
right?
That's
like
a
question.
People
always
ask,
and
it's
you
know
so
I'm.
In
a
lot
of
the
help
channels,
we've
actually
started
to
file
in
a
lot
of
focal
questions
and
the
more
questions
that
it
has.
Obviously,
the
smarter,
smarter
it
gets.
B
D
The
person
who
did
the
spot,
his
name
is
very
glad,
lor,
perk
and
super
helpful
person
and
he
actually
added
the
ability
to
filter
by
work
space.
So
now,
I
can
answer
that
question
before
I
couldn't
because
it
was
pretty
hard
to
separate,
but
any
I
also
wanted
to
add
a
couple
more
things
about
the
usefulness
of
these
parts.
It
doesn't
send
you
as
a
user
as
a
as
a
user
of
the
channel.
E
More
thing
on
that
it,
it
is
a
bot
but
there's
also
some
other
slack
integration
that
it
uses.
So
you
don't
have
to
type
slash
vocal
and
then
your
question,
you
can
just
ask
it
in
the
channel
which
some
people
were
confused
by
but
yeah.
If
you
don't
know
hopeful
evening,
you
know
folk
like
this.
It
should
be
able
to.
B
Okay,
moving
on
cuz,
we
got
a
first
question
in
the
YouTube
chat.
Thank
you.
Moody,
sada,
we'll
get
to
your
question
here
in
a
minute.
Quick
shout
out
to
cubed
XCOM,
which
mentioned
arc,
is
one
of
your
top
ten
cool
things
to
install
in
your
cluster.
So
thanks
for
the
shout
out
there
and
there's
a
new
repo
carlesha,
did
you
make
this
this
repo
dude?
Who
wants
to
talk
about
this
art
community,
repo,
real,
quick,
yeah.
D
It's
really
a
joint
effort,
I
just
it
was
me
who
pushed
in
this
stuff.
It's
it's
done
very
much
in
the
format
of
the
TG
ki
repo,
and
so
we
have
a
place
to
put
the
notes
for
the
Q&A
sessions,
which
is
one
of
these
ones
that
we're
doing
right
now
and
they
are
design
spec
sections
which
are
awesome
and
we
so
we're
going
to
have
one
of
one
of
each
each
month.
D
So
every
two
weeks
we're
going
to
be
having
one
of
it
and
it
has
an
index
without
of
the
sessions
and
you
can
browse
then,
once
you
get
to
a
page
corresponding
to
one
of
the
online
meanings
you
will.
You
will
get
all
the
contents
for
that
session.
So
you
get
the
announcement
and
blinks
everything
is
going
to
be
typed
up
in
there
and
also
the
link
to
the
video,
which
is
the
most
important
one
and.
B
D
Oh
sorry,
okay
very
relevant.
The
issue
list
in
that
repo
is
a
gray.
It's
not
the
only
place,
but
it's
a
great
place
for
you
for
people
to
add
request.
Questions
request
less.
If
you
want
to
first,
do
a
walk-through
of
a
particular
design
aspect
of
arc
or
Q&A
questions,
so
you
can
just
easily
add
an
issue
there.
Yeah.
B
And
I
went
ahead
and
posted
the
link
to
the
notes
in
in
the
slide
channel,
and
what
we'll
do
is
at
the
end
of
this
we'll
just
commit
the
whole
thing
to
github
under
this
repo
and
we'll
just
have
one
repo
with
all
the
notes.
If
you
ever
miss
a
show
or
anything
like
that,
you
could
just
go
in
there
grab
the
playlist
from
the
YouTube
channel
I've
created
a
playlist
for
all
of
these.
B
C
0.9.0
I
was
a
collection
of
bug,
fixes
I
think
we
had
six
different
bug
fixes
and
you
can
check
out
the
release
on
github
and
see
the
change
log
for
the
details
of
those
we'd
like
to
call
out
that
four
out
of
those
six
bug
fixes
were
actually
contributed
by
nan
hep
do
employees
by
community
members.
So
we're
really
excited
to
help
from
folks
who
are
using
our
canned,
see
bugs
and
can
contribute
fixes.
So
we
hope
to
keep
having
those
inputs.
B
Some
great
and
then
lastly,
Andy
and
Carly
she
awaits
wait
frantically
we'll
be
attending
Cuba
in
Seattle,
so
reach
out
to
us.
If
you
want
to
meet
them,
we'll
have
a
booth
there.
If
you
just
want
to
arrange
to
come
talk
to
us
or
you
have
any
questions,
feel
free
to
ping.
Us
in
the
channel
we'd
be
more
than
happy
to
to
talk
to
you.
Alright,
so
quick,
shout
outs
before
we
get
to
the
questions
here,
thanks
to
the
following
github
users
for
contributing
bug
fixes
into
the
nine
point.
B
Nine
point:
seven
release,
I'm
gonna
butcher,
this
one
but
Emily's
Zorich,
MWI,
ecz,
zürich,
mark
TC,
saw
the
hexa
and
Tim
or
Ryman
know
that
I
did
I.
Put
your
those
total
I.
Just
the
last
one
is
Timo
Oh,
Timo,
Riemann,
okay,
I,
get
that
and
note
to
myself
try
to
find
try
to
find
the
real
names
instead
of
trying
to
vocalize
their
github
handles.
Alright,
we
ready
for
some
questions.
We've
already
got
one
from
moody.
B
Sata,
oh
and
Ross
would
like
to
say
that
he's
also
going
to
be
a
key
concept
interested
in
talk
to
that.
So
I'm
gonna.
Ask
Moody's
question
and
then
I
have
a
related
question
to
it
that
we
got
before
so.
Moody
asks
how
much
art
for
art
doesn't
have
a
way
to
expose
prometheus
metrics.
Is
there
an
alternative
way
to
trigger
alerts
on
backup
failures?
That's
kind
of
related
to
a
question
that
we
had
over
the
month
is
how
can
I
get
Prometheus
metrics
for
our
periods
so
consider
that,
like
a
two-part
question,
so.
A
The
home
chart
for
arc
is
community
maintained.
We've
reviewed
the
initial
pull
request
to
add
it
to
the
charts
repository,
but
we
are
quite
appreciative
that
community
members
have
put
in
a
lot
of
time
and
effort
to
create
it
and
maintain
it.
So
I
would
recommend
that
if
it
doesn't
support
an
easy
way
to
set
up
the
metrics
or
set
of
Prometheus
that
you
could
open
up
an
issue
in
the
helm,
charts
repository
and
specify
for
marker,
that's
what
you're
looking
for
and
we
can
try
and
provide
some
guidance
there.
A
It
should
also
be
possible
to
add
annotations
for
the
arc
deployment
and
the
pod
specification
after
helm
has
put
it
in
there.
I'm,
not
super
sure
on.
If
that
would
be
maintained
across
helm
version
upgrades
as
you
pull
in
newer
versions
of
Arc.
But
probably
the
best
thing
to
do
is
open
up
an
issue
in
the
charts,
repo,
okay.
B
A
E
A
If
there
are
additional
metrics
that
you're
interested
in
seeing
please
let
us
know,
there's
one
open
issue:
I
think
it's
number
84
but
we'll
double-check
and
get
the
right
link
in
place
for
metrics
in
general.
At
some
point,
I
would
like
to
close
that
one
out
and
open
up
new
issues
for
individual
metrics
suggestions
as
they
come
in,
but
that
would
probably
be
the
best
place
right
now
to
go.
Take
a
look.
B
A
B
So
this
is
not
the
first
time
someone's
asked
about
the
helmet
chart.
If
there's
anybody
out
there,
who's
willing
to
help
us
out
with
a
helmet
I
know
that
we
can
give
advice
on
what
we'd
like
to
see,
and
maybe
you
can
read
I'm
thinking
of
a
person
who's
using
the
help
chart
right
now
and
I've
not
asked
him
how
he
exactly
he's
set
it
up,
but
I
know
he's
using
Prometheus,
so
maybe
that's
a
place
where
we
can
help
find
people
that
are
used.
B
You
know
trying
to
set
it
up
the
same
way,
and
so
you
can
get
that
goodness
in
one
place.
So
thanks
very
much
moody
for
your
question
feel
free
to
ask
a
follow-up.
If
you
have,
if
you
have
any
looking
forward
to
seeing
if
there's
any
other
prometheus
metrics,
you
would
like
us
to
add.
Okay,
any
other
comments
on
prometheus
or
arc
or
anything
move
on.
B
C
Absolutely
so
so,
we've
decided
to
snap
to
a
two-week
cadence
for
bug-fix
releases
and
so
the
the
way
that
we
envision
this
working
is
that
we
have
a
kind
of
a
standing
every
other
week,
release
scheduled
and
that's
on
Thursday
afternoon.
So
we
did.
We
did
one
last
Thursday,
so
the
next
one
is
scheduled
for
next
week
on
Thursday,
we'll
basically
monitor
over
those
two
weeks,
whether
there
are
any
new
bug
fixes
coming
in
and
whether
they
are
relatively
easy
to
to
backport
into
the
previous
version.
C
We've
set
up
a
new
label
in
github,
which
you
can
see
in
the
notes,
is
called
back
port
candidate,
and
so,
as
things
come
in
and
we're
thinking
about,
possibly
back
porting
them,
we'll
we'll
use
that
label
to
identify
them,
and
then
you
know
day
or
two
before
we
get
to
that
release.
We'll
we'll
make
a
call
as
to
whether
we're
we
actually
have
enough
fixes
to
to
cut
a
release.
But
if
there
are
things
that
that
are
back
portable,
then
then
you
should
expect
a
release.
So
so
look
for
the
next
bug
fix
release.
C
Next
week
we
are
also
working
hard
on
version.
10
we've
been
working
on
that
for
a
couple
of
months
now
and
I
think
we're
we're
getting
close
to
wrapping
it
up,
but
we
still
have
a
few
weeks
of
work
to
go,
so
we
don't
have
a
specific
date
for
that.
But
I
think
you
know
towards
the
end
of
October.
Beginning
of
November
is
when,
when
that
one
should
be
expected
and.
A
B
A
B
A
That
is
if
it's
something
that
looks
like
it's
going
to
be
a
breaking
change,
then
it
either
has
to
wait
until
2
point
or
it
has
to
get
into
1.0
or
we
pull
the
community
and
see
what
acceptance
is
for
a
breakage
in
the
1.0
a
line,
but
we're
really
trying
to
figure
out
what
are
the
must-have
features
and
fixes
for
a
1.0
and
then
what
could
come
in
1.1
or
1.
Point
2
vs.
breaking
changes
for
like
a
not
a
redesign,
but
something
significant
enough
to
call
it
2.0
right.
B
Alright,
any
other
questions
on
releases
please
feel
free
to
toss
them
into
the
Select
channel.
Alright,
any
any
comments
on
releases
before
I
move
on
from
our
last
year's
panel.
Wayne
is
strangely
quiet
today,
alright
next
question:
what's
the
procedure
for
using
Azure
and
then
I
get
I
got
this
link
to
an
azure
config
file
here.
E
Feel
this
I
think
we've
had
a
few
questions
in
the
past
month
about
setting
up
with
Roger
and
running
running
mark,
particularly
with
aks
they're
managed
kubernetes,
offering
the
link
there
is
our
documentation
on
how
to
how
we
suggest
you
run
arc
with
a
natcher.
Probably
the
biggest
caveat
is
azor.
Has
a
concept
of
resource
groups
that
it
places
things
in
and
with
aks
there
is
a
new
resource
group
generated
for
you
that
you
have
to
point
arquette
and
I.
C
Yeah
I
did
recently
get
a
PR
merge
to
just
make
some
tweaks
to
the
Azure
documentation
to
hopefully
make
the
the
resource
group
nuances
clearer,
so
I
think
at
this
point
it
should.
If
you
follow
those
instructions,
it
should
be
pretty
clear
as
to
to
what
you
need
to
do,
but,
as
always,
if,
if
you're
run
into
issues
setting
up
or
running
it
and
I
reach
out
to
us
on
slack
or
through
github
yeah.
B
B
A
B
A
In
whatever
terminology
the
the
storage
provider
uses
and
for
larger
snapshots
or
on
slow
days,
it
would
take
sometimes
more
than
30
seconds
to
have
a
rave
all
um--
and
we
we
just
removed
the
this
30-second
wait,
because
kubernetes
will
repeatedly
attempt
to
attach
the
volume
to
the
to
the
node.
Even
if
the
status
is
not
ready,
so
it'll
just
keep
trying
until
it
is
ready
and
then
things
proceed.
So
this
was
just
an
artificial
weight
that
we
had
in
there
that
it's
hard
to
get
right
for
all
the
use
cases.
A
That's
not
an
art
thing,
so
arc
will
be
a
disc
and
then
set
up
a
cooper,
Nettie's
persistent
volume
to
point
at
that
disc
and
then,
when
you
or
when
arc,
creates
a
pod
or
deployment
and
replicas
that
create
a
pod.
That's
referring
to
this
PV.
Ultimately,
the
scheduler
will
stick
the
pod
on
a
node
and
the
couplet
will
or
sorry.
The
controller
manager
will
repeatedly
try
to
attach
it
until
it's
ready.
Okay,.
C
Steve
yeah
I
can
talk
a
little
bit
about
this,
so
in
the
latest,
kubernetes
release,
1.12
and
alpha
feature
for
supporting
volume.
Snapshots
was
released
as
part
of
kubernetes.
It's
an
out
of
tree
feature.
It's
it's
an
alpha,
so
it's
still
experimental
and
liable
to
change.
We've
been
sort
of
tracking
it
and
then
keeping
an
eye
on
it
to
see
if
it's
something
that
we
would
ultimately
want
to
integrate
with
arc
and
I
would
say
at
this
point,
we're
still
not
sure.
Given
that
it's
it's
definitely
an
alpha
feature.
C
We
do
want
to
take
some
time
to
experiment
with
it
and
kind
of
see
what
the
what
the
pros
and
cons
of
it
are
really
relative
to
our
approach
over
the
next
few
weeks
and
also
kind
of
see
how
it
how
it
moves
along
towards
stability.
So
I
think
you
know
overall.
The
message
here
is
that
we
need
to
wait
and
see,
but
we're
definitely
aware
of
it
definitely
going
to
experiment
it
with
it
and
then
see
what
makes
sense.
Oh
and.
B
And
then
just
kind
of
a
general
question
when
things
land,
so
in
kubernetes
things
go
alpha
beta
and
then
production
or
GA
do
we
as
an
arc
generally
keep
track
of
also
features
coming
in
and
then
like
when
they
turn
beta
and
you
kind
of
take
a
more
ship
like
how
do
you?
How
do
you
triage
all
the
incoming
alpha
features
like
like
so.
A
Volume
snapshots
definitely
is
in
the
same
domain,
given
that
art
does
volume
snapshotting,
so
that's
definitely
one
that
we
as
Steve
said
have
our
eyes
on
and
we
will
play
around
with
it.
We
have
some
time,
but
for
other
alpha
and
beta
API
is
there
largely
just
things
that
we
are
able
to
backup
and
restore,
and
sometimes
they
may
require
some
special
handling.
Maybe
we
need
to
order
how
we
change
the
order
in
which
we
do
a
restore.
A
B
B
So
I
think
probably
the
question
the
person
is
asking
and
I
think
this
is
related
to
the
question
before
is
who
do
I
talk
to
as
far
as
getting
a
feature
request
in
so
I
would
say,
file
an
issue
on
it.
Let
us
know
about
it,
you
know
and
we're
happy
to
help
advise
people
if
they
are
looking
to
implement.
B
E
This
so
there's
been
a
there's,
been
interest
in
having
an
arc
on
a
cluster
and
having
a
little
bit
of
self-service.
So
say,
you've
got
a
namespace
for
a
given
team
and
they
people
want
to
be
able
to
make
their
own
backups
without
having
that
group
able
to
backup
and
restore
anyone
else's
stuff
right
now.
Arc
is
not
really
configured
for
that.
You
can
use
kubernetes
service
accounts
to
limit
what
an
art
installation
can
manage,
but
right
now
that
would
likely
need
an
art
deployment,
/
namespace
and
we
haven't
really
tested
that.
A
Fairly,
confident
that
we
can
run
multiple
arcs
in
a
cluster
having
them
scoped
to
each
namespace
I,
don't
think
there
will
be
any
issues
there
where
there
will
be
an
issue
is
with
persistent
volumes
and
anything
else.
That's
cluster
scoped,
so,
depending
on
what
privileges
are
granted
to
the
service
account
that
arc
is
running
as
if,
for
example,
the
service
account
can
do
anything
it
wants
in
namespace
X.
So
it's
basically
full
admin
on
that
namespace.
A
You
still
have
to
potentially
give
it
access
to
cluster
roles,
cluster
role,
bindings,
persistent
volumes
and
there's
no
way
to
really
segment,
which
ones
you
have
access.
The
service
account
has
access
to.
So
we
are
kind
of
in
the
back
of
our
minds,
trying
to
figure
out
if
there's
some
way
to
have
a
singular
arc.
Installation
like
we
do
now,
but
still
have
sufficient
security
so
that
non
administrative
users
can
come
in
and
request
backups
and
restores
without
having
a
privilege
escalation.
A
B
It
sounds
like
that'd,
be
really
useful
for
teams
right,
like
a
development
team
and
I
just
got
a
bunch
of
names
today,
I
don't
even
know
like
I,
don't
know
I
could
just
imagine
myself
going
what's
everyone?
What's
everyone
doing,
you
know
what
I'm
gonna
back
all
this
up,
just
let
George
go
and
create
stuff
yeah.
E
And-
and
we've
also
heard
people
want
this
in
like
a
multi-tiered
way,
so
we
have
a
a
you,
have
a
cluster
administrator
who
can
backup
everything
right
and
they'll
do
like
reggae
backups
of
the
cluster.
But
maybe
this
team
wants
to
create
a
new
namespace
that
they're
working
on
to
move
it
over
without
bugging
in
the
cluster
admin
and
a
little
bit
more
self-service.
Yeah.
B
Exactly
so
that
sounds
awesome
with
that
more
questions.
If
you
have
them
please,
we
are
monitoring
hash
art,
dr
on
the
kubernetes
slack,
as
well
as
YouTube.
If
you're
just
joining
us,
please
feel
free
to
ask
any
questions
other
than
that
we've
run
through
all
the
questions
this
month.
So
we're
just
waiting
for
live
questions
here
and
while
that's
happening,
we
can
maybe
talk
about
the
next
design
topic
for
the
next
session.
B
So
do
we
have?
We
were
talking
a
little
bit
about
this
before
the
stream
started.
Do
we
have
any
design
topics
that
we
want
to
discuss,
or
are
we
still
pretty
much
head
down
in
point
10
and
volume
snapshots,
because
we
can?
We
can
just
punt
that
and
then
just
go
straight
to
the
next
Q&A,
which
would
be
on.
C
So
I
think
you
know
given
where
we
are
in
o
dot
n
I
think
we've
pretty
much
nailed
down
all
the
design.
Questions
for
this
release,
sure
and
I
think
we
are
gonna,
be
pretty
heads
down
on
just
coding
and
trying
to
close
out
o
n,
so
I
think
from
the
you
know
from
the
core
dev
team.
We
don't
have
anything
that
we
need
to
talk
about.
That
said,
if
there's
you
know,
if
there's
a
topic,
that's
interesting
to
someone
in
the
community
we'd
love
to
hear
about
it.
Ok,.
B
So
I'll
do
this
I'll
just
basically
say:
hey
everyone's
heads
down.
Unless
anybody
wants
a
specific
session
for
anything
and
we'll
kind
of
keep
it
penciled
in
well
for
sure
do
this
next
QA
on
6,
November
and
depending
on
what
it
that
I
don't
know
where
the
next
I
don't
look
too
far,
I
think
that
might
land
around
as
we're
getting
to
November
in
December
the
design
sessions,
because
there
later
in
the
month,
I
might
start
to
run
into
people's
holidays.
So
we're
just
call
it
call
it
when
we
see
it
I
guess
so.