►
From YouTube: #IndexerOfficeHours 80 - Launchpad 101
Description
We will be running weekly Indexer Office Hours in the โGraph Stageโ voice channel on Discord, every Tuesday between 5pm - 6pm UTC.
Link to Notion page - https://graphops.notion.site/Indexer-Office-Hours-80-528c639d605b47e7ba559d6bb74602a4
Link to presentation - https://docs.google.com/presentation/d/1Pd0RuH8Hwq3Y62fMycM6Z6yW4uYsOWNlGBTZmIh9dTE/edit?usp=sharing
___________________________________________________________________
๐ ๐จ๐๐ฒ๐ณ๐๐น ๐๐ถ๐ป๐ธ๐:
- ๐๐ผ๐ฑ๐ฒ๐
- http://codex.thegraph.com/
- ๐๐ฐ๐ฎ๐ฑ๐ฒ๐บ๐ - https://thegraph.academy/
- ๐๐ถ๐๐ฐ๐ผ๐ฟ๐ฑ - https://thegraph.com/discord
- ๐๐ผ๐ฟ๐๐บ -https://forum.thegraph.com/t/indexer-office-hours/1270
- ๐๐ป๐ฑ๐ฒ๐
๐ฒ๐ฟ ๐ฑ๐ผ๐ฐ๐ - https://thegraph.com/docs/indexing
- ๐ฆ๐๐ฎ๐ธ๐ฒ๐ฆ๐พ๐๐ถ๐ฑ ๐ง๐ฒ๐๐๐ป๐ฒ๐ ๐๐๐ถ๐ฑ๐ฒ - https://github.com/StakeSquid/graphprotocol-testnet-docker
- ๐ฆ๐๐ฎ๐ธ๐ฒ๐ฆ๐พ๐๐ถ๐ฑ ๐ ๐ฎ๐ถ๐ป๐ป๐ฒ๐ ๐๐๐ถ๐ฑ๐ฒ - https://github.com/StakeSquid/graphprotocol-mainnet-docker
___________________________________________________________________
A
A
And
welcome
back
to
another
episode
of
index
office
hours
today
is
the
11th
of
October,
and
this
is
index
office
hours
80..
If
anyone
is
new
to
index
office
hours,
this
is
the
Town
Square
for
indexes
where
ecosystem
participants
come
together
and
discuss
the
most
important
and
relevant
topics
for
indexes.
Today's
conversation
will
be
broken
down
into
four
segments
opening
remarks,
repo
watch
protocol
watch
and
we
have
a
workshop
that
we're
going
to
be
covering
with
Chris
and
Anna
all
about
LaunchPad.
A
We
love
participation
so
at
the
time
that
we
open
up
a
q,
a
we'd
love
to
hear
from
you,
either
in
Voice
or
in
chat,
so
please
feel
free
to
unmute
a
mic
and
join
the
conversation.
Just
as
a
reminder.
This
is
going
to
be
recorded
and
distributed
onto
public
channels
like
YouTube.
So
please
bear
that
in
mind
and
with
that,
let's
get
into
today's
conversation.
So
I'm
going
to
start
off
with
opening
remarks
and
give
a
shout
out
to
the
latest
GRT
IQ
episode
I'll
be
very
quick.
A
It
is
85
with
Tachi
who
is
the
CEO
and
co-founder
of
playground,
analytics
a
Data
Solutions
platform
built
upon
the
graph
that
provides
serverless,
on-chain
data
infrastructure
and
services
for
data
teams,
analysts
and
Engineers.
So
in
that
conversation,
Nick
and
Tachi
covered
a
variety
of
topics
cover
it
anywhere
from
clean
energy,
the
crypto
industry,
space
exploration,
UFOs
and
finally,
the
graph
and
web3
and
Nick
described
tachi's
descriptions
of
the
blockchain
ecosystem
as
the
best
you'll
ever
hear.
A
B
All
right,
hello,
everyone
and
thank
you
Abel
for
handing
this
over
to
me
all
right.
So
today,
on
Ripple
watch
we
have
quite
a
few
updates
and
let
me
just
mute
my
Okay
cool,
so
on
the
execution
layer,
client
for
Aragon.
We
have
had
the
new
release.
First
of
all,
notice
that
the
release
versioning
has
changed
from
from
having
a
long
release
version
to
something
like
2.27,
which
is
the
current
release
and
that's
more
in
line
with
Aragon
too.
B
B
C
B
B
B
B
I
think
yeah
upgrade
if
you're
not
currently
running
it
in
kubernetes
cool,
so
go
to
consensus,
layer,
clients,
no
new
release
on
prism.
We
have
had
a
new
release
for
Techo,
that's
version
2210
one.
It
is
a
recommended
upgrade
because
it
contains
quite
a
few
bug
fixes
included
one
that
was
unexpected
rejective
execution
due
to
full
task
queue.
B
And
an
issue
with
high
CPU
usage,
it
does
have
a
lot
of
breaking
changes.
I
will
not
go
through
them
because
they
are
very
specific
and
not
everyone
uses
Taco,
but
please
make
sure
to
check
the
release
notes.
The
reaper
watch
notes
on
the
agenda
on
a
name.
So
this
is
a
recommended
release
for
Kaku
on
Nimbus
no
new
releases
same
with
Lighthouse.
Moving
on
to
the
graph
stack,
we
have
had
a
new
release,
so
version
0.28.2.
B
Big
warning
with
this
one,
the
all
indexer,
are
advised
to
migrate
straight
into
version
0.28.2
and
entirely
bypass
0
280
and
0
28
1..
This
is
because
both
of
those
versions
have
a
bug
which
will
cause
some
graphs
to
stop
syncing
under
some
deployment
configurations
and
in
terms
of
in
terms
of
index
of
service
and
agent.
B
B
Okay,
indexer
service
and
agent
new
release
version
0.20.4.
This
is
a
pre-release.
It
has
a
docs
update.
It
has
fixed
a
bug
in
its
deployment
worth
allocating
towards
it
persists,
receipts
with
zero
fees,
and
it
changes
service
to
agent
in
the
agent
run
command
before
I
move
into
the
next
section,
which
is
the
new
graph
orchestration
tooling.
B
Same
with
Jim,
okay,
in
that
case,
moving
on
to
the
last
section,
which
is
graph
orchestration
tooling,
this
is
a
section
that
was
added
last
week,
so
we
have
steak
squid,
also
graph
protocol
test
net
Docker
and
graph
protocol
combined
Docker
have
not
had
new
releases
for
graph
of
Launchpad.
You
will
finally
see
it
today,
but
remember
it's
still
in
Alpha
State.
B
Data
on
the
helm,
chart
Ripple,
all
of
the
like
Nimbus
in
other
mind
and
all
of
the
recent
releases,
a
recent
changes
to
the
execution,
layer,
clients
and
consensus
layer.
Clients
have
been
updated
for
launched
path
core.
We
have
had
a
new
release
which
is
0.020
again,
that
updates
to
latest
Nimbus,
never
mind
and
icon
charts,
and
it
has
a
restructure
of
some
of
the
names
of
the
namespaces.
We're
using
I.
Don't
expect
anyone
to
fully
follow
the
last
section.
So
I
will
ask
at
the
end
of
the
workshop.
B
A
A
B
Awesome
well
welcome
everyone
to
our
much
awaited
intro
to
Launchpad,
as
Abel
said,
I
also
want
to
say
a
big
thank
you
to
Chris
and
the
groundhog
theme
for
helping
helping
work
on
this.
This
has
been
a
major
project
and
thank
you
in
advance
for
everyone
who
will
be
trying
this
project
and
supporting
us.
B
B
Launchpad
is
a
set
of
functionality
that
allows
you
to
set
up
host
and
network
abstraction
layers,
to
use
declarative
version,
control
infrastructures,
code
to
use,
containers
use
kubernetes
and
to
be
able
to
plug
in
whatever
host
you
have
come
in
whether
it
is
from
something
you're
hosting
in
your
garage,
something
you're
hosting
on
VMS
or
something
you're
hosting
on
public
Cloud
providers.
Now,
if
you're
new
to
indexer
office
hours
and
the
mips
program-
and
you
just
joined
us-
you
might
not
you-
you
might
be
completely
confused
in
that
case.
B
I
would
invite
you
to
go
back
to
some
of
our
previous
workshops
about
what
it
takes
to
run
an
indexer,
so
some
indexer
best
practices.
We
have
had
a
mips
workshop
on
infrastructure,
orchestration
and
some
things
to
think
about
before
deciding
how
you're
going
to
run
your
infrastructure.
Some
of
those
things
were
capex
and
Opex,
so
capex
being
Capital
expenditures.
Being
the
initial
cost
that
you
spend
on
your
setup,
which
applies
mainly
to
running
bare
metal
servers.
B
Only
and
then
we
talked
about
Opex,
which
is
the
operating
cost
which
can
be
higher
on
some
managed.
Cloud
platforms
can
be
very
low
on
running
only
on
bare
metal
and
it's
a
bit
of
a
mixed
bag
when
you
are
mixed
and
matching
some
VMS
that
are
renting,
with
some
bare
metal
or
with
some
public
Cloud.
So
just
to
turn,
to
give
you
a
refresher
as
to
why
we're
talking
today
about
this
project,
we
are
calling
launch
card.
Please
go
back
to
those
presentations
Okay.
B
B
So
for
for
this
setup,
we
will
be
using
a
task
file
as
the
main
tool
that
to
Define
and
run
your
tasks
on
your
local
machine.
So,
by
doing
that,
we
are
trying
to
abstract
away
some
of
the
complexity
of
running
all
of
the
other
tools,
so,
for
example,
Helm
file,
which
is
a
tool
for
defining
and
deploying
Helm
releases
ansible,
which
is
CLI
for
running
declarative,
provisioning
against
hosts
with
SSH
tube
CTL
help
and
SSH.
B
B
B
B
B
Graph
stack
and
then
we
have
gnosis
mainnet
and
it
may
not
as
well
as
the
graph
stack.
So
these
are
this
diagram.
That
sort
of
shows
you
the
different
components
now
I
bet
you're,
still
wondering
what
is
it
actually
well,
when
we
look
at
it
from
our
actual
structure,
point
of
view,
Launchpad
is
at
the
core
of
everything
is
two
separate
repos,
so
we
have
Launchpad
starter,
which
is
your
starter
repo,
and
once
you
have
cloned
it,
it
will
become
your
private,
instant
instance
of
infrastructures
code,
which
then
you
manage
yourself
now.
B
Launchpad
starter
provides
workflow
and
structure
for
mapping
infra
and
what
is
declared
in
there
will
be
acting
like
your
command
center
for
your
infrastructure,
aspartos
Launchpad
starter.
You
would
be
cloning
another
you
will
be
linking
as
I
get
some
model
another
repository.
This
is
Launchpad
core
that
contains
a
lot
of
task:
definitions
for
common
workflows
and
as
well
as
ansible
role,
definitions
and
some
upgrades
to
Launchpad
a
night
spell
definition
and
Helm
file
release
template
definitions.
B
So
if
you
think
of
Launchpad
core
as
the
place
for
templates
and
the
starter
repo
as
the
place
for
customization
that
be
a
good
way
of
looking
at
it.
In
addition
to
these
two
repos,
we
also
have
another
repo
which
is
Helm
charts.
Helm
charts
is
another
public
repo,
where
you
can
find
every
gun
another
mind
and
all
of
the
other
components
that
we
need
to
run
as
indexers
templated
and.
B
Made
kubernetes
friendly,
okay,
so
so
far,
you've
heard
a
few
things
about
Launchpad
and
hopefully
you
will
be
getting
a
night,
an
idea
of
what
it
is.
But
if
you
haven't
yet
for
today's
demo,
we
have
our
user
mary.it,
who
will
walk
us
through
a
full
demo,
but
before
I
do
that
I
would
like
to
pause
for
a
second
and
ask
if
there
are
any
questions
and
specifically
Chris.
If
there's
anything
you'd
like
to
add
on
top
of
what
I've
already
said,.
F
No
I
think
you
yeah
great
job
so
far.
If,
if
anybody
is,
is
kind
of
a
bit
unsure
there's
a
lot
of
new
stuff
here,
so
don't
don't
worry
if
things
are
not
totally
clear,
hopefully
the
demo
will
help,
but
this
this
is
a
somewhat
complex
stack
and
you
know
you
will
need
to
invest
some
time
in
understanding.
B
Thank
you
very
much
all
right,
so
our
user
Mary
has
recently
heard
about
the
graph
protocol
and
she
has
decided
to
give
indexing
a
go.
So
what
she
knows
is
that
she
needs
to.
She
needs
to
run
the
graph
protocol
indexer
infrastructure,
which
is
this-
the
application
components
that
you
see
on
the
right
hand,
side
under
the
index
infrastructure
Square.
B
So
she
knows
she
needs
to
run
all
of
this
components,
and
she
knows
that
the
indexer
hardware
requirements
just
to
get
you
started,
so
the
bear
bare
minimum
are
to
run
a
8
CPU
core
machine
with
16
gigabytes
of
RAM
and
quite
a
bit
of
storage.
So
she
has
a
machine.
It
is
an
Ubuntu
22
for
LDS
and
it
has
SSH
access
and
she's
been
looking
at
the
different
options
to
set
up
her
index
infrastructure,
and
she
recently
has
learned
about
Launchpad
and
she
has
decided
to
give
it
a
go.
F
B
Okay,
I
will
talk
through
this
first
okay,
so
in
terms
of
the
things
what
we'll
be
doing
to
get
started
so
for
Mary
to
get
started,
she
will
first
need
to
clone
Launchpad
starter
into
her
own
private
infrastructure
repo.
So
she
would
do
that
with
this
command
of
get
clown,
Launchpad
starter
then,
once
she
she
does
that
she
will
need
to
delete
the
origin
of
Launchpad
starter
so
that
it
has
its
so
that
Funtime
infra
has
its
own
origin.
Remote
set.
B
B
B
A
B
B
B
The
first
thing
you
will
see
here
is
that
a
prerequisite
for
Launchpad
is
to
run
at
least
one
host
running
Ubuntu
2204.
You
also
need
to
run
SSH
access.
You
also
need
to
have
SSH
access
using
public
key
Authentication
and
if
you
are
using
a
Mac,
Homebrew
must
be
installed
for
LaunchPad
dependencies
so
to
get
started.
The
very
first
thing
you
would
need
is
Task
file
installed
in
my
case,
because
I've
done
this
quite
a
few
times,
I
do
have
task
pass
file.
B
B
B
So
my
branch
is
up
to
date
with
origin
main,
but
this
origin
main
would
be
of
Launchpad
starter.
So
now
I
would
remove
your
region
of
Launchpad,
starter
and
well
add
actually
for
me,
because
this
will
be
an
ephemeral
repo
just
for
the
purpose
of
this
kid.
Just
for
the
purpose
of
this
demo,
I
will
not
be
adding
this
and
I
will
not
be
pushing
it
into
a
remote
repo.
So
I
have
done
all
of
this.
So
so
far
we
have
cloned
into
the
repo.
B
We
have
gone
into
the
new
folder
we've
created
and
we
have
removed
the
origin
of
Launchpad
starter.
Next
up
would
be
installing
the
Launchpad
Corsa
module
here.
It
is
suggested
to
use
task
Launchpad
setup.
But
if
you
read
this
comment,
it
says
this
will
run
two
other
tasks,
so
Launchpad
update
core
and
Launchpad
apps
I
personally,
don't
want
to
run
Launchpad
apps,
because
that
will
install
every
single
bit
of
software
and
need
it
for
Launchpad,
so
that
will
be
Helm
file,
unstable
Helm,
Cube,
CTL
and
a
few
others.
B
B
C
B
All
right
we're
back
on,
thank
you
for
bearing
with
me
so
so
far.
We
have
updated
the
two
repos
we
have
pulled
on
dependencies
in
and
we
have
installed
all
of
the
software
dependencies
that
are
needed.
Now.
We
are
gonna
update
the
configuration
with
our
host
definition,
so
to
do
that
under
my
fun
time
in
front
repo
I
have
an
inventory
directory
in
that
directory.
I
have
some
samples
on
how
to
run
on
how
to
set
up
your
inventory.
Your
answer
spell
inventory
with
hosts.
B
Vms
for
testing
on
your
machine.
So
if
I
do
CLI
list
here,
you
can
see
that
I
have
one
VM
already.
This
is
the
VM
that
I'll
be
using
for
this
demo.
So
this
is
the
IP
address
that
I'm
using
so
my
ansible
host
will
be
this
IP
address
the
portal
be
signing
up
with,
will
be
22,
the
user
I'll
be
signing
up
with,
will
be
k0s
and
the
ansible
user
will
also
be
k0s.
This
is
a
specificality
for
multi-pass.
B
C
B
C
B
B
Okay,
we
are
time
has
run
out,
so
I
will
take
that
back.
So
now
that
that
has
run,
which
has
a
screen.
Some
packages
you'll
be
able
to
see
all
of
the
code
and
ask
a
question
about
it
right
later,
but
the
next
step
would
be
to
install
k0s,
which
is
our
chosen
distribution
of
kubernetes
onto
all
of
the
hosts.
B
What
we
are
telling-
and
it's
about
to
do-
is
to
give
us
one
cluster
of
k0s
with
one
host
as
initial
controller,
and
that
controller
can
also
run
as
a
worker
in
k0s
and
just
kubernetes.
In
general,
you
have
the
concept
of
controllers
and
worker
nodes.
Usually
the
worker
nodes
are
the
ones
that
host
all
of
the
different
kubernetes
workloads
on
them
and
the
controller
nodes
are
the
admins
of
kubernetes
that
keep
eye
on
the
state
of
the
node
and
just
the
overall
state
of
the
cluster.
B
B
B
So,
just
reading
the
comments
from
Matthew
how
customizable,
how
customizable
will
this
be,
for
example,
it's
a
very
simple
example:
what
if
you
want
to
have
the
user
pre-existing
by
existing
automation,
you
already
have
and
not
created?
You
can
have
that
that
it
works
like
that.
You
are
not
creating.
So
sorry
just
to
clarify
in
our
case
Matthew
here,
because
ansible
user
and
Main
users
are
the
same
user.
B
B
Okay
I
will
yes,
you're
right,
I
will
keep
on
going
and
I'll
get
back
to
your
question:
okay,
okay!
So
now
you
can
see
that
we
have
a
node
that
is
ready
with
the
role
of
control
plane,
but
it
is
also
no
worker
which
means
I
can
Deploy
workloads
on
it.
Another
thing
you
can
see
are
depth
task
will
install
this
on
your
machine
as
well,
which
is
K9s,
is
a
tool
that
very
neatly
organizes
all
of
the
different
kubernetes
resources.
B
B
I
will
not
be
doing
that
because
again,
ephemeral
VM.
So
going
back
to
quickly
going
back
to
our
presentation,
we
have
started
at
the
local
environment.
We
have
clone
some
repos.
We
have
applied
some
ansible
tasks
with
applied
base
and
apply
k0s.
We
have
installed
the
kubernetes
cluster
with
apply
k0s.
B
B
A
B
B
Okay,
we'll
wait
for
all
of
these
to
run
for
the
purposes
of
this
demo.
We
will
not
go
into
too
much
detail
of
each
of
the
decisions
of
of
every
single
application
stack,
but
we
will
do
in-depth
demos
later
on
with
some
of
the
a
specific
customizations,
as
well
as
additional
design
decisions
and
how
we
got
there.
C
B
F
I
think
the
VM
might
really
be
struggling
under
load
to
anyone
who's.
Watching.
It's
worth
saying
this
entire
stack
is
being
installed
on
the
same
machine
that
Anna's
actually
developing,
on
which
to
be
honest,
is
is
a
bit
of
a
mistake.
We,
as
stake
machine,
has
suggested
it
would
have
been
better
for
the
demo
to
provision
a
remote
machine
but
yeah,
it's
I
think
as
it's
starting
everything
up.
It's
probably
struggling
and
the
the
kubernetes
API
server
may
not
be
responding
with
all
of
this
information.
C
B
B
B
B
Okay
and
then
you
can
go
into
the
logs.
You
can
also
set
up
more
detailed
queries
as
well,
so
this
is
for
logs,
so
we
can
see
we've
installed
Loki.
We
can
look
at
the
different
logs,
we
can
create
additional
alerts
and
then,
once
all
of
the
other
sorry
so
going
back
to
the
to
here,
so
we
have
installed
all
of
these,
and
once
we
get
to
the
next
stack
the
now
since
my
net,
we
will
be
able
to
have
dashboards
when
those
are
syncing.
B
B
So
we
have
grafana
here
we
have
grafana
overview.
For
example,
we
will
make
sure
to
add
more
dashboards
as
well.
Graphana
is
persistent.
What
that
means
is
that
since
I've
imported
a
new
dashboard,
the
logs
dashboard
that
you
can
find
here
now
and
if
I
restart
the
Pod
of
grafana,
this
dashboard
will
still
persist
by
virtue
of
persistence.
B
B
B
B
B
B
B
B
Okay,
so
once
this
is
becoming,
as
this
is
becoming
ready
going
back
to
our
Milestone
alert,
so
we
are
currently
deploying
gnosis
mainnet.
Once
we
have
deployed,
gnosis
may
not
will
be
able
to
deploy
the
graph,
the
indexer
and
then
the
graph
may
not
stack.
However,
as
we
now
knows,
this
needs
to
sync
with
data,
so
so
for
the
purposes
of
this
we
will.
We
will
stop
the
demo
here
and
continue
with
the
demo
at
a
later
date.
B
This
will
be
a
two-part
demo,
so
just
a
quick
recap
of
some
of
the
things
we
have
done
before
we
look
at
the
workload
again,
so
we
have
created
a
new
repo
from
Launchpad
starter.
We
have
linked
Launchpad
core
some
module,
and
these
are
the
two
main
components
of
of
your
new
infrastructure.
Then
we
have
set
up
the
client
and
with
that
we
have
updated
inventory
with
the
right
IP
address.
We
have
installed
dependencies
after
that.
B
B
So,
let's
quickly
talk
about
updates
and
how
you're
gonna
keep
up
with
this
project,
but
also
how
you're
gonna
push
on
updates
so
for
core
upgrades,
because
Launchpad
will
launch
five
core
will
keep
on
being
updated
with
new
components
and
new
version
updates.
You
would
run
the
following
commands,
so
you
would
again
run
that
task.
B
Launchpad
update,
call,
which
is
which
is
runs,
a
get
full
sum
module,
and
then
you
would
get
add-in
commit
and
push
your
changes
to
your
local
to
your
fun
time,
infra
in
repo,
that
is
for
core
updates,
then
for
starter
updates,
you
it's
a
bit
more
involved
than
corrupt
rates.
You
would
again
start
with
a
task.
The
task
Launchpad
pull
Upstream
starter.
B
B
A
Yeah
sure,
just
a
little
bit
about
the
session,
so
it's
it's
gonna
be
kind
of
like
office
hours
in
the
truth.
Sense
where
you
can
come
with
any
sort
of
question
around
launch
pad
or
if
you're
dealing
with
an
issue
you
can
come
and
ask
about
that
issue
and
either
Anna
or
Chris
can
help
out
with
that
issue
or,
if
you're,
just
interested
in
future
releases
And
discussing
Launchpad
in
general.
A
B
F
F
So
if
you
wanted
to
use
the
core
definitions
to
make
five
releases
of
Aragon
with
your
own
release,
names
and
service
names
and
everything
else,
you
can,
you
can
also
add
custom
tasks.
You
can
add
custom
ansible
configuration
and
you
can
add
you
know,
releases
and
namespace
definitions
completely
outside
of
what
is
covered
in
core.
F
So
you
can
mix
kind
of
nodes
with
heterogeneous
capabilities,
some
full
nodes,
some
archive
nodes,
some
Trace
nodes
and
it
will
intelligently
route
and
load
balance
requests,
and
you
can
actually
also
specify
nodes
that
are
external
to
the
cluster.
So
for
those
of
of
you
that
have
already
synced
a
nethermind
archive
node
as
part
of
phase
zero,
you
don't
have
to
start
all
over
again.
You
can
actually
use
that
node
in
the
context
of
launch
at
by
configuring
proxy
D
to
point
at
it.
That's
something
that
we
can
cover
in
tomorrow's
Launchpad
office
hours.
F
If
you
already
have
a
kubernetes
cluster
you'd,
obviously
just
skip
all
of
that,
and
you
would
only
run
the
releases
tasks
which
basically
assume
that
the
kubernetes
cluster
exists
and
that
your
local
kubernetes
client
is
pointing
at
that
cluster,
which
is
pretty
much
the
state
that
you
will
be
in.
So
you
can
use
all
of
this
tooling
to
deploy
and
manage
the
stack
on
top
of
kubernetes.
B
Just
to
just
to
add
to
that
you
can,
you
can
absolutely
do
that.
I
will
say,
though,
if
you
need
to
do
that.
Please
do
let
me
know,
because
that
specific
values
that
need
to
be
overwritten
in
your
release,
names
just
to
do
with
the
cubeletal
location.
So
k0s
has
a
different
Cube
like
location
from
other
public
Cloud
distributions.
B
G
F
Gemma
asks
if
lots
of
indexers
use
this
are
there
concerns
around
decentralization
or
lack
thereof,
or
is
it
so
flexible
that
everyone's
stack
will
still
be
different
enough?
That
one
bug
won't
take
everyone
down?
I
think
this
is
a
really
good
question
and
you
know
I
think
it
really
highlights
the
dimensionality
of
decentralization
I.
F
G
F
Matthew
Darwin
again
asked:
can
we
script
everything?
So
nothing
is
done
manually,
I,
guess
this
is
in
reference
to
Anna
importing
the
the
grafana
dashboard,
which
is
obviously
not
a
declarative.
Workflow,
that's
a
great
thing:
I.
We
wouldn't
recommend
that
you
import
dashboards
that
way,
because
you
obviously
don't
have
that
change,
that
state
reflected
in
your
git
repo
and
that's
the
whole
point
of
of
the
declarative
workflow.
F
Cool.
A
big
big
Jim
has
said
this
is
so
cool.
Thank
you,
big
Jim.
What
is
happening
when
the
services
are
coming
up?
Are
the
images
being
downloaded
and
run
like
Docker
or
our
Sim
or
or
our
mtvms
being
spun
up
and
configured
so
this
in
this
is
much
closer
to
Docker.
In
that
sense,
obviously,
doctor
like
doesn't
really
have
any
notion
of
host
abstraction,
there's
DACA
swarm,
which
is
obviously
a
separate
kind
of
deprecated
product
now,
but
a
Docker
itself,
just
one
runs
on
one
machine
right.
F
As
you
add
more
subgraphs
and
there's
this
kind
of
decoupling
between
your
application
workloads
and
the
underlying
infrastructure
that
runs
things
okay
and
then,
finally,
is
there
a
way
to
skip
installation,
slash
configuration
of
any
components?
Let's
say
I
already
have
graph.
Let's
say:
I
already
have
my
kubernetes
cluster
with
grafana
and
Prometheus.
C
F
So
you
could
either
just
comment
out
those
releases
in
the
kind
of
declarative
release
manifest
and
Skip
them
in
this
tool
or
you
you
could
yeah
just
not
apply
that
namespace
or
but
there's
no
requirement
that
you,
you
know,
need
to
run
everything
whatever
you
want.
Matthew
Darwin
asks
I
assume
we
can
control
which
load
runs
on
which
kinds
of
hosts
yes
most.
F
Certainly,
although
it's
worth
saying
that
that
is
a
feature
of
kubernetes
and
not
a
Launchpad
but
yeah
kubernetes
is
you
know
infinitely
flexible
in
that
respect,
in
terms
of
being
able
to
kind
of
annotate
nodes
with
labels
and
then
kind
of
Define
rules
for
how
this
the
set,
the
scheduler
will
take
workloads
that
you
give
it
and
allocate
them
to
hosts
based
on
those
rules.
So.
C
F
D
A
lot
of
the
work
you
do
in
the
early
stages
of
this
will
be
learning
how
to
use
things
like
it
properly,
because
it's
really
important,
especially
for
when
you
start
wanting
to
pull
the
updates
from
what
graph
Ops
are
doing
into
your
own,
your
own
repos
and
that's
really
basic
stuff
right.
But
it's
critical
in
order
to
be.
D
It's
absolutely
critical
that
you
start
I'm
learning
that
that
it's
critical,
that
you
start
with
an
IDE
rather
than
just
hacking
away
at
a
maybe
at
a
terminal
or
something
like
that
and
using
Vim
Nano
or
whatever
your
chosen
editor
is
to
edit
things
in
your
repo.
You
really
need
to
be
running
an
IDE.
E
E
That
being
said,
you
know,
for
those
that
have
a
hard
time
using
git
I,
really
really
recommend
that
you
take
some
time
to
get
to
learn
how
to
use
it
either
from
the
CLI
or
from
whatever
your
development.
Environment
of
choice
is,
but
really
the
things
you
need
to
know
are.
You
know
how
to
co,
how
to
push
how
to
pull
how
to
commit
now.
What
a
commit
message
is
how
to
roll
back
if
you
screw
something
up,
but
it's
really
not
that
hard.
E
D
E
F
F
F
That
is
a
great
question.
Two
things
to
highlight.
The
first
is
that
our
Helm
charts
also
a
link
one
in
the
zoom
chat.
Now
the
Aragon
chart
linked
to
just
a
doc
section
there
on
restoring
from
an
external
snapshot,
so
the
helm
charts
already
support
this
you.
You
could
obviously
have
that
you
know
internal
to
the
cluster
via
some
backup
system
or
served
via
https
or
whatever.
F
F
G
F
Will
I
CFS
is
pretty
good.
No,
this
is
single
host
CFS
Matthew,
so
this
would
basically
by
using
the
the
ZFS
local
PV
controller.
You
know
your
your
Aragon
or
blockchain
node
State
can
obviously
then
be
snapshotted
and
cloned,
and
you
know
their
workflows,
then,
for
spinning
up
copies
of
the
node.
That
way
you,
you
are
asking
many
questions,
but
that's
not
a
bad
thing.
There
yeah.
Basically,
there
are
actually
implementations
of
ZFS
over
the
network,
which
we
are
looking
at,
but
more
for
you
know.
F
G
C
F
Essentially,
what
I
would
recommend
your
path
is
there?
Is
you
provision
your
host,
the
way
that
you
want?
You
know
that
might
be
you
just
install
Ubuntu
and
use
Launchpad
host
space,
or
you
know
you
do
your
own
provisioning
part
of
at
once.
That
is
done.
You
then
provision
butterfs
right,
like
do
all
of
your
butter,
FS
config
and
mount
your
butterfs
volume
at
some
path
that
is
consistently
available
across
your
hosts
and
then
you
can
use.
Oh,
is
it
it's
B
tree
FS?
It's
not
butterfs!.
F
C
E
E
Assuming
you
set
up
the
I
o
channels
correctly,
if
you
were,
if
you
were
to
spin
up
a
large
SATA
array,
using
something
like
a
like
a
NetApp
filer,
or
something
like
that,
you
may
get
decent
enough
performance
to
run
an
ethereum
node,
but
really,
if
you
were
to
hook
it
up
to
a
32
gig
per
second,
I
o
Channel
Through
fiber
channel
and
distribute
that
across
a
bunch
of
different
files.
You
can
get
some
really
cool
performance
out
of
that
for
just
a
tidbit
from
the
old
days
from
my
previous
life.
F
E
F
We
we're
yeah
we're
running
more
pleb
Hardware,
so
details
on
ZFS,
yeah
I
think
we
should
end
soon-ish.
So
we
probably
can't
go
super
deep
into
that
now,
but
to
Cohen's
question
of
is
the
ZFS
integration
deeper
than
just
you
know,
putting
flat
files
on
some
ZFS
Mount
Mount
Point,
and
the
answer
is
most
definitely
yes,
so
each
kubernetes
workload
and
and
these
obviously
extend
to
the
workloads
and
configuration
that
are
available
in
Launchpad
each
one.
F
You
know
basically
specifies
what's
called
a
persistent
volume
claim,
and
it
essentially
just
says
I
want.
You
know
a
volume,
this
big
and
of
these
characteristics
and
with
the
ZFS
integration
that
PVC
will
actually
result
in
creating
a
zfsz
bowl
and
on
a
workload
by
workload
basis.
You
can
customize
all
the
ZFS
parameters,
everything
from
record
size
to
compression
to
deduping
to
to
a
time
whatever
you
need,
so
you
can
really
tune.
You
know,
ZFS,
for
each
workload.
A
Okay,
well,
if
there
aren't
any
other
questions,
then
this
might
be
a
good
opportunity
to
wrap
I
want
to
say
thank
you
to
each
and
every
one
of
you
for
joining
us
today.
This
has
been
a
very
informative
workshop
for
me.
I
hope
it's
been
the
same
for
you
and
we
look
forward
to
having
you
guys
at
the
next
ioh.
As
as
a
reminder,
this
will
be
recorded
put
onto
YouTube
and
if
you
have
any
questions,
please
head
over
to
the
Discord
channel
for
Launchpad
and
ask
your
questions
there
cool.