►
From YouTube: Kubernetes Kops Office Hours 20180216
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
A
No
the
other
thing
that
obviously
has
happen
is
I'm.
Gonna,
do
a
pass
over
all
the
PRS
make
sure
everything
there's
something
that's
sitting
there
that
really
needs
to
be
into
1/9.
I
know
there
are
some
of
those
I've
been
working
on
that,
but
yes,
they
keep
building
up
and
I
know
there
are
some,
but
if
anyone
has
anything
that
they
really
think
is
important
for
customer
9.
That
they're
aware
of
please
speak
up.
Otherwise,
we'll
do
a
cops.
1
9
alpha
1
theta
1
this
weekend
right
what
your
idea.
A
Yeah
we
got
out,
we
got
our
cops
181
last
Monday
or
maybe
was
that
Friday
I,
don't
recall,
but
the
big
thing
in
cups
181,
along
with
something
all
the
network
providers
which
seems
to
be
perpetual,
was
adding
a
taller
ignoring
the
new.
He
said,
yeah
no
file,
which
is
added
in
1/9,
the
cops
1
9.
So
we
now
have
a
downgrade
pass.
A
You
will
have
to
downgrade
to
181.
If
you
choose
to
do
that,
though
cops
181
companies
doesn't
matter
good.
Well,
that's
seen
on
contentious,
which
is
good
news,
always
I.
Think
Chris
love
had
a
bunch
of
things
on
the
agenda.
I,
don't
know
if
he's
here,
he
is
not
yet
and
I
think
the
questions
around
phases
and
lifecycle.
It's
definitely
best
to
wait
for
him.
We
are
generally,
he
put
it.
There's
a
housekeeper
housekeeping
item
on
the
agenda,
which
is
that
we're
generally
trying
to
organize
our
calendars
a
little
bit
better.
I.
A
Think
that
they
are,
the
documents
are
owned
by
individuals.
Obviously,
I
don't
know
who
owns
the
documents?
The
calendars
are
owned
by
individuals
and
I.
Think
we're
gonna
try
to
figure
out
a
you
know,
an
ownership.
That
is
a
group
account.
If
we
change
calendars,
we
will
try
to
make
it
seamless,
I,
don't
know
exactly
how
we're
gonna
do
that,
but
I
don't
want
to
I.
Imagine
what
we'll
do
is
well.
If
we
choose
counters,
we'll
double
run
the
calendar
for
a
while.
A
B
I'm
new
to
cops,
not
super
new
to
kubernetes
generally,
but
as
far
as
implementing
it
at
a
company
I
am.
We
have
a
pretty
specific
use
case
where
we
want
to
maintain
point-to-point
connection
between
all
of
our
clients
and
our
servers.
The
way
it
currently
exists,
but
still
have
kubernetes
orchestrate
the
starting
and
stopping
of
things
we're
in
AWS.
B
So
we
have
some
particular
networking
things
that
we
need
to
get
in
place
for
that
to
work,
and
there
is
a
branch
that
was
emerged
in
the
last
month,
or
so
they
that
allows
cops
to
spin
up
clusters
that
use
this
CNI
plugin
that
was
developed
by
an
Amazon
engineer
anyway.
The
short
version
of
it
is,
you
know,
I'm,
just
starting
to
get
into
it,
starting
to
learn
about
cops
and
use
it
for
the
first
time
and
I've
just
been
having
some
issues
with
getting
it
to
work
period.
B
B
C
A
B
B
D
D
D
11.81
there
we
go
so
that
it
launches.
If
you
look
in
your
rerun,
a
Debian
should
be
Debian
yeah.
If
you
look
in
the
daemon
log
files,
you'll
be
able
to
see
the
error
note
up
if
it's
still
bailing,
if
it
is
and
that
make
file
change,
didn't
fix
it
ping,
they
should
have
instructions
on
how
to
do
a
makeup
load
and
that's
what
you're
going
to
do
is
post
your
own
note
up.
Ok
should
get
you
by
there,
but
yeah
I,
post,
sometimes
post
some
like
with
the
gossip
cluster.
B
B
A
B
A
That
will
actually
be
the
big
thing,
I,
think
and
or
one
of
the
big
things
in
cups
or
nine
is
I.
Think
we
can
turn
off
the
alpha
feature,
gate
for
GCE
and
say
yes,
it
actually
is
now
perfect
ready
to
create
clusters
without
a
few
different.
But
yes
and
I
hope
that
the
I
hope
it's
just
the
great
issue
and
as
the
upload
issue
to
make
sure
that
yeah
the
versions
queue
I,
think
we
I'll
do
I've
heard
a
lot
of
people.
A
Have
this
problem
and
I
think
we
need
some
better
docs
around
here's,
how
you
upload,
both
at
the
same
time
right
because
it
is
we
try
to
we
try
to
make
it
so
that
it
will
just,
but
we
broke
I
broke
that
when
I
added
this
key
set
llamo
file
word
it
no
longer
just
works
as
long
as
the
versions
are
in
sync,
it's
fine,
but
if
you
just
build
cop
CLI
and
you're
running
the
previous
note,
that
does
not
work
god.
But
now
we
loaded
a
note
up.
A
B
A
Wonderful
and
Chris
I
think
you
had
two
topics
on
there
around
the
phases
in
the
life
cycle
we
talked
about
one
on
out,
for
we
seems
like
everyone's,
maybe
just
the
normal
bumping
of
docker
and
all
that
sort
of
stuff,
an
image
building
and
doing
a
pass
over
all
the
outstanding
PRS
make
sure
nothing
is
critical,
but
then
we
can
roll
with
one
nine,
four
one
or
beta
when
the
P
on
how
we
feel.
But
you
had
a
two
comments:
around
life
cycles
and
phases
yeah.
So
we've.
D
Been
talking
on
PRS,
exactly
how
you
want
the
lifecycle
overrides,
torque
and
I
think
I
finally
understand
and
what
understand
what
you
wanted
for
you
and
correct
me.
If
I'm
wrong,
you
want
tasks
of,
say,
security
group
type
or
whatever
the
past
we
have.
You
want
the
lifecycle
override
to
just
be
for
that
task
type.
A
Well,
I
mean
I.
Think
phases
are
a
group
of
of
this
sort
of
thing.
Know
that
think
that
the
you
have
some
use
cases
where
people
want
to
like
drill
down
even
more
beyond
the
phase
when
I
was
suggesting.
The
next
step
is
to
allow
you
to
change
the
mode
by
by
the
type
of
rule
and
I
guess
eventually
might
be
by
the
type,
a
name
of
the
actual
task
that
we're
actually
running
so
the
people
they
don't
know.
A
The
idea
is,
you
know,
cops
is
sort
of
like
terraform,
where
we
build
a
model
of
the
cloud
resources
that
we
want,
and
then
we
synchronized
the
model
with
social
like
kubernetes.
We
synchronize
the
model
with
the
real
world
with
the
actual
cloud,
and
to
do
that,
you
know
when
you,
when
you
do
a
cups,
update
cluster
or
without
yes,
you
see
a
preview
of
all
those
things.
A
That's
gonna
go
and
create
or
modify,
and
the
phases
idea
is
the
idea
that
in
more
complicated
enterprise
companies,
it
may
be
that
one
person
doesn't
have
the
responsibility
or
the
authority
to
do
all
those
tasks.
It
may
be
that
one
group
has
to
create
the
IAM
policies,
and
another
group
has
to
create
the
network,
and
another
group
has
to
can
break
the
the
compute
resources
and
so
phases
try
to
split
it
them
all
along
the
obvious
division
lines.
A
I
think
the
phases
we
have
right
now,
our
security
network
and
cluster
or
compute
or
cluster,
which
is
sort
of
the
compute
type
things
and
I
think
Chris
who's
been
driving.
All
of
this
work
has
found
a
use
case
where
the
security
team
is
split
between
I
am
Rollo's
and
security
group
things
and
jolts.
Are
there?
It's
it's
not
so
much
well,.
D
There's
multiple
things,
so
it
goes
back
to
the
PRS
that
we
have.
We
have
one.
That's
an
I
am
here
where
you
bring
your
own.
I
am
profiles
and
the
other
is
a
security
group
profile
or
security
groups
right.
So
I've
got
about
three
or
four
PRS
in
flight
around
this
and
fortunately,
unfortunately,
you
can't
use
those
to
PRS
independently
unless
we
had
another
phase.
A
D
D
D
D
A
So
you
want
both
or
we'd,
have
a
phase
which
were
a
flag
which
was
called
I,
think
they
call
it
phases
right
and
you
could
do
a
phase.
Okay
and
you
could
define
you
could
either
not
pass
it
at
all,
in
which
case
you
get
the
default.
We
create
everything
for
you,
it's
one
user
at
they
have
permissions,
you
could
say:
phase
equals
security
and
we'll
do
the
security
rules.
Phase
equals
network
would
be
the
network
phase
equals
cluster.
We'll
do
that
and
there's
a
natural
there
was
the
three
ones
that
you
run
in
succession.
A
If
you
are,
if
you
have
the
sort
of
obvious
split
where
you
have
a
security
group,
a
network
group
and
a
cluster
code,
if
you
have
a
just
a
security
group
and
a
everything
else,
group
you
can
do
phase
equals
security
and
then
phase
equals
network
comma
cluster.
If
you
have
a
more,
if
your
yet
more
complicated-
and
you
split
your
security
group
right,
then
what
I
think
we
do
is
I
think
we
have
phase
equals
security,
comma,
minus
security
group
security
group
rule
something
like
that.
D
Don't
like
it
secure
your
rule
or
I'm
running
sorry
if
this
is
so
complicated
and
so
ever
taking
up
so
much
time
on
this
guy's
so
phase,
you
want
to
remove
the
security
rule
task
or
you
give
running
face
cluster
and
you
don't
want,
and
you
want
to
set
the
life
cycle
for
this
phase
or
the.
So
that's
what
I
think
we're
confused
about
or
I'm
confused
about.
So
we
don't.
A
Yes,
I
mean
I,
think
I
think
we
want
to
do
like
so
the
phase
flat.
The
phase
is
a
way
to
set
a
set,
a
group
of
lifecycles
on
a
group
of
tasks,
their
design,
a
facet
right
and
the
facet
that
we
are
defining
it
on
is
then
the
grouping
which
we've
defined
right.
We
seem
to
think
there
are
these
three
teams
yeah.
Those
three
teams
might
not
correspond
to
what
you
have,
and
so
you
might,
you
might
want
a
way
to
define
it
based
on
the
facet
of
the
task
type.
A
We
might
want
to
find
it
based
on
the
facet
of
the
task
name
unpleasant,
but
gives
you
the
ultimate
control.
I
would
hope
that
what
we
can
do
is
if
we
have
it
based
on
the
face,
the
predefined
phases-
and
we
have
it
based
on
the
task
name.
The
type
of
the
tasks
right
based
on
the
predefined
phases
and
the
type
of
the
task
that
is
sufficient
I
would
hope
that
there's
no
one
crazy
enough
out
there.
That
has
a
different
team
managing
security
groups
in
one
zone
than
in
another
zone,
but
so
you're.
D
Thinking
I
am
firms
and
security
groups.
There's
there's
also
the
use
case
or
we've
got
these
insane
PRS
coming
in
where
we're
gonna
have
to
override
stuff.
That
cops
expects.
You
know,
no
matter
who,
what
permission
the
person
has
okay,
so
we
want
both
by
task
type
and
by
phase
I
have
to
PRS
that
do
that.
D
The
only
question
is
getting
them
to
work
together
course
is
gonna,
be
interesting,
I
think
they
will
work
together,
though
their
question
is
to
CLI
fly,
so
I
think
we
can
talk,
we
can
argue
about
the
CLI
Flags
offline,
cuz
I,
don't
want
to
have
a
CLI
flag.
That's
super
complicated
I'd.
Rather
this
is
earth
that
has
everything
in
it
and
I'm
somehow
parsing.
It
I'd
rather
have
like
three
different
flags
for
this,
but
we
can
talk
about
that
in
the
PRS,
because
I
think
it's
complicated
enough.
D
The
last
item
that
I
had
on
the
agenda
was
the
how
the
phase
overrides
were
working
because
I,
don't
think
the
validate
and
not
fail
is
working
in
cycles,
so
I
think
because
we
have
the
warn
exists
or
something
or
so
so
what
I
need?
What
I'm
looking
for
in
functionality
life
cycle
is
that
we
validate,
but
we
keep
on
going.
We
fail
the
validation,
but
we
allow
the
user
then
saying
user
to
keep
on
going.
How.
A
D
The
bug
I
think
there's
a
I,
don't
believe
it
behaves
in
the
manner
that
you
just
described.
Okay
and
that's-
that's
fine!
That's
just
a
fog
right,
because
I
think
you
I've
found
an
issue
where
you
said
that
we
need
to
return
a
error
type
to
the
executor,
so
that
it'll
pass
and
I
can't
find.
Of
course,
I
can't
find
that
issue.
Okay,
let
me
take
up
that
issue.
I've
taken
up
enough
time
on
this
call,
let
me
take
up
that
issue
and
I'll
ping.
You
back
with
that
issue.
D
A
F
Do
want
to
bring
up
something
your
friend
so
Chris
mentioned
an
open,
PR
related
to
life
cycles
for
the
external
I
am
stuff.
That's
my
PR
and
I
just
wanted
to
mention
that
it,
the
functionality
of
it,
is
still
works
without
life
cycles.
It's
just
that
defining
it
will.
You
know
it'll
apply
to
change
all
in
one
phase
in
our
specific
use
case
where
it's
all
applied
in
a
CI
environment.
In
one
step
it
works
perfectly.
Fine,
that's
awesome!
That's.
A
So
my
I
work
my
way
backwards
each
morning
and
I
get
as
far
as
I
get
so
you'll
notice
that,
like
a
newer
PRS,
are
being
reviewed
much
faster
than
the
older
PRS,
but
I
think
I
think
it's
on
the
second
page.
Now
we
get
through
that
first
page,
that's
when
I
figure,
maybe
the
weekend
will
be
quieter.
Yeah,
one-seven-one.
C
A
C
A
C
A
Might
have
different
things,
I
just
don't
know
what
the
correct
word
is.
It's
the
naming
problem,
but
I
also
just
merge
it,
and
then
we
can
I
mean
I.
Think
at
some
stage
we're
gonna
do
a
networking
split
out
of
API
objects,
which
will
be
a
almost
definitely
involve
a
new
API
version,
and
so,
if
we
decide
we
in
two
months
to
come
up
with
a
much
better
name,
which,
let's
be
honest,
won't
happen,
then
then
we
can
yeah.
C
D
So
to
add
some
background,
this
is
allowing
a
user
to
specify
more
add-ons
or
more
channels
which
goes
back
to
our
channels
program
that
our
channel's
binaries
of
use
to
a
cluster
spec.
So
what
the
user
did
was
they
added
an
array
of
add-ons
of
additional
channels
and
I'm
wondering
if
we
should
have
like
literally
channels,
objects
that
are
defined
in
the
cluster
spec
or
not
not
in
cluster
spec,
but
as
a
top-level
object.
D
H
H
H
Exactly
and
I,
and,
to
be
honest,
I
haven't
looked
at
this
since
I
got
back
from
holiday,
so
I've
got
a
got
to
dig
into
it
again,
but
that
was
my.
That
was
my
initial
impression
that
there's
still
some
more
so
I
think
was
this.
The
one
I'd
called
to
you
and
Justin
on
just
know:
it
was
a
different
I,
tagged
Justin
on
it.
Okay,
we're.
I
A
A
When
you
upgrade
cops,
you
may
expect
to
get
an
upgraded
version
of
weave
one
is
available,
but
at
the
same
time,
once
you,
if
you
go
and
upgrade
your
weave
directly
using
the
add-on
manager
on
the
cluster,
it's
very
unclear
who
wins
at
that
point
with
cops
wants
to
upgrade
and
you
must
upgrade.
There
is
a.
There
is
a
partial
solution
which
is
to
a
sort
of
an
overlay
type
concept.
A
So
if
your
customization
was
effectively
modifying
some
parameters,
adding
another
on
big
map
or
changing
the
memory
resources,
that
would
be
fine
if
you
did
it
using
the
overlay
way.
If
your
change
was
to
apply
a
different
version
of
weave,
then
it
becomes
non-trivial
to
know
who
should
win
you're.
A
Caching,
caching,
overlays
yeah
I
mean
I
think
that
we
are
we're,
not
I,
think
that
being
the
antennas,
that
would
be
a
base
version
for
each
version
of
leaf,
so
relieved,
one
two
three
four
right
and
then,
if
cops
wants
to
go
to
version
two
but
you've
already
upgraded
weave
to
version
3
making
up
numbers.
Obviously
what
how
do
we?
How
do
we
reconcile
those
two?
Presumably
you
or
your
you
in
because
your
version
is
bigger,
I
guess.
H
Yeah
I
mean
then
there's
also
the
the
odd
case
of
wanting
to
downgrade
right,
maybe
so,
like
I,
think
being
able
to
being
able
to
just
say:
look
we
want
to
manage
this.
We
want
to
override
whatever
cops
is
doing
and
we'll
take
responsibility
for
testing
our
clusters
and
and
everything
else.
That's
that's
more.
The
level
that
we're
it
could.
A
H
D
See
I
think
that's
outside
the
functionality
of
channels.
I
think
that's
I
think
we
have
two
things:
I
think
we
have
channels
and
I
have
a
manager
which
should
just
be
one
list
of
add-ons,
and
then
we
have
bootstrap
builder,
which
should
be
able
to
take
whatever
a
user
specifies
and
then
create
that
one
list
of
add-ons
I.
D
Don't
know
it's
messy,
but
I
think
we
just
need
to
start
somewhere
and
we
got
permission
to
do
that.
You
know
we
got
blessing
to
do
that
kind
of
sorta,
a
coupon
we've
already
exposed
the
channel.
So
technically
we
could.
You
know
the
channel
is
already
exposed,
so
somebody
could
build
their
own
channels
already.
D
A
A
We
avoid
having
the
channels
thing
as
a
concept,
but
rather
say
you're
able
to
apply
a
manifest
and
manifest
yeah
I
did
that
work
for
you,
Robin
for
at
least
partially
I
mean
we
gotta
solve
the
conflict
issue,
but
at
least
for
when
you're,
adding
stuff,
which
is
seems
like
additional
channels,
is
what
that's
about
with
that.
With
that
be
sufficient
to
just
have
a
list
of
manifests,
because.
A
A
D
A
A
Ie
has
a
list
of
well-known
things,
I
think
so
you
can
say
I
think
it's
literally
dashboard
right
and
you
would
write
the
word
dashboard
in
there
and
I
think
they
have
some
additional
options
that
I
recall,
but
then
dashboard
gets
expanded
out
to
the
known
manifester
channel
behind
the
scenes,
so
that
isn't,
that
is
an
API.
We
can
work
forever
right
because
we're
not
tying
to
a
particular
scheme.
Outra
format,
it's
a
string
right,
so
it
would.
It
would
then
get
expanded
to
a
channel
or
a
manifest,
but
it
wouldn't
be
visible.
A
G
A
And
that's
deliberate,
the
reason
being
that
so
challenges
I
can
tell
you
how
it's
basically
a
wrapper
around
it's
an
atom
manager
similar
to
the
current
atom
manager,
where
the
atom
manager
in
bash
that
GCE
uses
right.
It
has
a
API
type.
I
would
like
for
that.
Api
type
to
be
I
had
I
had
hoped
we
could
get
that
API
type.
Sort
of
more
aligned
with
upstream
it
seems
like
doctrine,
has
a
richer
concept,
which
is
Brian
grants,
declarative
application
management
approach,
which
is
pretty
good,
I,
think
I
think
it
makes
a
ton
of
sense.
A
The
idea
is
that
you'll
have
a
master
manifest
or
an
official
manifest,
and
then
you'll
have
additional
patch
files
which
will
get
it
on
top.
So
if
you
want
to
and
then
the
idea
being,
you
can
upgrade
the
the
base
manifest
and
your
patches
will
still
be
layered
on
top
and
the
idea
is
that,
then
you
know,
we've
can
upgrade
one
two
three
four,
you
as
an
operator
can
apply
your
own
changes
like
I.
A
Think
there's
a
peer
connection
limit
that
just
went
in
so
I
want
to
set
the
peer
connection
limit
separately,
based
on
my
own
class,
when
I
set
the
memory
resources
based
on
my
own
cluster
and
even
at
NIMH
upgrades,
your
changes
will
still
be
applied
on
top
and
you
can
still
customize
it
so
that
I
can
adopt
that
approach.
Then
I
think
it's
time
to
then
we
can
document
it
but
channels
right
now.
A
I
A
A
master
manifest
which
actually
links
to
child
manifest
files
and
each
one
of
those
was
a
semantic
version
version
associated
with
it.
They
get
stored
as
actually
annotations
on
the
cube
system,
wrong
space,
which
is
a
thing
and
then
and
that
way,
if
you
go
and
change
your
weave,
if
you
go
in,
if
you
couldn't
change,
your
weave
manifests
you're
using
weaving
means.
Your
weave
manifests
on
your
running
cluster
on
a
GC
cube,
Apple.
Five
minutes
later
it
gets
reverted
when
the
add-on
manager
goes
and
tweaks
it.
A
If
you
do
it
on
cups,
it
will
not
be
reverted
until
you
reapply
until
you
upgrade
weave,
whereupon
it
will
be
reverted,
and
so
the
add-on
manager
concept
is
intended.
So
the
declarative
application
management
is
intended
to
separate
out
your
changes
from
the
upstream
changes
so
that
there
is
actually
an
answer
to
how
do
we
allow
users
operators
their
customized
system
add-ons,
while
at
the
same
time
allowing
us
to
upgrade
system
add-ons?
I
D
D
A
D
A
A
Probably
change
it
at
yeah
room
around
there,
where
we,
where
we
go
through,
we
run
all
our
tasks
that
are
available
to
run
we
fork/join
them.
Then
we
look
through
the
errors
which
we
got.
If
any,
and
we
currently
say
any
any
non-nil
error
is
considered
the
real
error,
and
now
we
want
to
say
if
there
is
not
mail,
then
check
to
see
if
it
implements
this
interface.
Err.
A
D
A
B
C
A
Or
ponder
that
one
thing
that
occurred
to
me
is
remember:
it
is
if
you,
if
you
have
multiple
flags,
you
tend
to
lose
the
border,
not
careful.
If
I
have
multiple
Flags
I
tend
to
lose
the
order.
If
you
have
like
a
and
flag
B
and
the
user
types
a
a
equals
one
I
equals
to
B
equals
three,
you
will
get
a
equals.
One
and
two
B
equals
three,
but
you,
if
you
type
B,
equals
three
equals
one
I'm
losing
your
example.
That's
whatever
that's
like.
A
A
D
D
I
can
I'm
gonna
have
to
take
a
string,
Val
and
then
parse
it,
which
scares
the
bejesus
out
of
me
so
I'm
saying:
let's
just
leave
faces
alone
and
then
have
an
override
flag
or
life
cycle
overrides
or
faces
overrides
or
whatever
the
heck.
We
want
to
call
it
and
we'll
throw
all
the
overrides
in
there
and
then,
if
it's
and
then
I
can
have
a
naming
for
a
phase.
If
it's
a
phase
name,
then
I'll
pull
it
out
and
make
a
phase
override.
D
A
D
But
we
have
to
know
what
phase
were
running
is
one
thing,
and
then
we
have
to
know
what
lifecycle
we're
playing
with,
but
a
phase
just
sets
life
cycles.
A
phase
is
setting
what
group
of
life
cycles
were
running
or
running,
which
is
what
I
said
it
doesn't
set.
What
the
life
cycle
actually
is.
No,
it
just
sets
just
sets
life
cycles.
It
sets
what
life
cycles
are
run
or
what
it
says.
Actually
it
doesn't
even
set.
It
says
a
group
of
light
well.
D
C
D
D
I'm,
really,
let
me
let
me
implement
two
flags.
You
can
take
a
look
at
it
and
then
mess
around
with
it
and
we
can
decide
from
there
because
I
just
I
think
it's
getting
nasty
for
a
user
to
have
to
parse
two
flies,
as
it
really
kind
of
comes
in
my
head,
that
okay,
a
users
going
to
set
the
phases
to
run
the
way
that
it
is
to
find
the
code
or
the
user
is
going
to
set
the
face
and
the
overrides
or
a
user
can
set
the
overrides.
D
D
We
think
in
very
different
ways,
so
you
know
having
two
people
approach
the
same
problem
that
think
quite
differently
actually
gets
a
better
result.
So
I
I
can
do
both
and
we
can
take
a
look
at
both
cuz
I
got
two
different
PRS.
A
A
C
I
correct
me:
if
I'm
not
crazy,
but
at
one
point
dog
land
was
the
name,
then
they
changed
it
to
go
land
and
I.
Just
was
like
I
rage.
Quit
I
was
like
I'm
done.
They
renamed
this
thing.
I
can't
upgrade
it
I'm
out
of
here.
What
else
exists
and
I
was
like
people
did
say.
Bs
code
is
good
and
I've
been
missing
out
on
that
tools
like
I
didn't
install
the
tools.
There's
none
of
the
IDE
stuff
work.
Now
it's
great
yeah.