►
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
B
A
Like
weird,
you
don't
have
any
opinion
well,
this
is
this
week
in
Cassandra,
and
we
are
actually
talking
about
things
that
are
happening
this
week
would
be
pretty
cool.
What
do
we
got
with
us
here?
We
have.
Should
we
high
we've
got
Luke
Tillman
Patrick
McFadden.
We
also
have
Jeff
carpenter
joining
us
from
choice,
hotels,
who
is
the
author
of
an
o'reilly
Cassandra
book
that
we're
going
to
be
talking
about
today
of
the
definitive
guide
for
Cassandra,
which
is
pretty
cool.
A
We
had
Jeff
jirsa
on
a
little
while
ago
to
talk
about
this
time.
Window
compaction
strategy
is
finally
replacing
dated
compaction
strategy,
so
this
is,
this
is
a
pretty
big
deal
and
time
window
is
just
offers
fewer
knobs
to
turn
its
same
defaults
out
of
the
box,
and
it
turns
out
a
lot
of
people
are
using
this
in
production.
So
if
you're
thinking
about
going
to
30-
and
this
was
kind
of
your
hand
up-
you
know
this
is
this-
is
a
good
time
to
start
thinking
about
it
230.
C
Far
as
new
releases
go,
there
was
a
link
on
the
blog
post
to
the
the
Python
driver.
Just
got
a
new
release
as
well,
and
they
kind
of
rolled
out
this
new
feature.
Jeff
mentioned
it
when
we
were
talking
before
we
before
we
came
on
the
air
here,
I'm
called
execution
profiles,
which
is
kind
of
like
a
new
experimental
thing,
which
is
super
interesting
kind
of
a
way
to
like
to
group.
C
You
know,
like
settings
configuration
settings
that
you
might
you
know
want
to
run
so,
like
you
know
your
default
consistency
level
and
your
default
retry
policy
and
that
kind
of
stuff
and
then
have
and
then
use
those
to
execute.
You
know
statements
in
your
code,
so
kind
of
an
interesting
new
feature.
I
know
they've
been
kicking
it
around,
like
I've,
been
following
the
juror
and
the
c-sharp
driver
they've
been
kicking
around
how
to
implement
it
there
and
I.
Imagine
it's
coming
in
the
Java
driver
as
well.
So
definitely
check
that
out.
C
A
A
A
B
C
It's
it
sounds
like
really.
The
idea
is
they.
You
know
like
there's,
so
many
configuration
options
with
the
driver
and
it's
only
getting
like
more
and
more.
You
know
like
there's,
there
keep
adding
more
configurations,
and
so
this
is
kind
of
just
a
nice
way
to
like
group
configurations
together
and
to
make
your
your
experience
as
a
developer
easier.
You
know
kind
of
one
place
where
you
can
manage
the
the
profile
and
then
reuse
it
you
know.
C
So
if
you
got
to
go,
you
got
to
go
change
like
say:
maybe
you're
the
consistency
level
you're
using
for
rights
and
I
think
we'll
change
it
in
25
places
in
your
code.
You
know
you
go
one
place
where
you
created
the
profile
and
go
change
it
there
and
how
to
propagate
down
so
kind
of
just
a
nice
feature
for
developers.
Make
your
life
easier,
I!
Think
that's
awesome.
I.
B
C
C
A
Need
it
I
need
a
language.
Apparently
I
would
be
so
that
the
interesting
thing
about
cql
engine
is
that
the
original
way
that
we
wrote
it
was
you
effectively
passed
all
your
parameters
in
and
to
begin
with,
so
you
effectively
got
one
profile,
and
then
that
was
it
like
there
was
no.
There
was
no
point
in
which
you
would
like
change.
A
I
mean
you
could
change
consistency,
level
and
things
like
that
on
a
per
query
basis,
so
I
don't
see
any
reason
why
you
couldn't
pass
him
like
user
dot
objects,
top
profile,
dot,
get
or
whatever
and
get
back
a
single
object
with
like
a
profile
that
was
tuned
for
consistency,
level,
warm
with
or
cereal
cereal
I
might
be
the
better
one,
because
if
you're,
like
oh
I,
want
to
make
sure
I
have
you
know
I'm
using
LWT
to
do
rights
so
I
also
want
to
use
serial
to
do
Reed's.
A
C
B
A
A
What
answers
Luke
coach
wanting
to
be
thorough,
okay?
Well
so
the
other,
the
other
thing
I,
don't
seriousness,
I
in
the
rust
driver
and
I.
Actually,
in
the
show
notes,
we
gotta
link
for
this
tweet
that
Patrick
put
out.
If
you
can
retweet
this
thing
like
it
will
be
really
helpful.
D
A
D
B
E
B
B
A
B
So
I,
you
know
well
I
tweeted
that,
because
I
was
just
listening
to
you
rant,
which
is
kind
of
a
common
occurrence,
but
it
was,
it
seemed
like
it
was
a
little
more
heartfelt
like
it's.
You
and
I
were
talking
yesterday
about
the
just
it's
a
new
language
and
it's
got
a
lot
of
rough
edges,
and
this
is
one
of
them.
Yeah.
A
It's
just
it's
it's
difficult
because
you
know
getting
good
support
for
database
drivers
and
these
libraries
allow
you
to
actually
like
do
things
that
are
useful.
It's
like
database
drivers
and
like
web
frameworks
and
all
the
middleware
fun
that
that
you
want
in
a
language
like
you
know,
if
you're
looking
at
building
really
high-performance
massively
scalable
systems,
rust
is
looking
like
a
really
really
viable
alternative,
especially
with
all
the
you
know.
Safety
checks
built
in
a
compile-time
and
I
would
just
love
to
see
really
solid,
Cassandra
support
in
there.
A
So
that
was
why
I
was
looking
at
it
so
yeah.
If
anyone
can
retweet
that
or
you
know,
help
fix
that
bug,
I
think
I'd
be
really
really
cool
all
right.
Well,
moving
on
you
know,
I
won't
beat
the
dead
horse
of
my
anger.
With
with
the
rough
situation,
we
have
something
that
Patrick
I
know
is
near
and
dear
to
your
heart
cloud
native
deployments:
Cassandra
DS
and
Cooper
Nettie's,
your
google
fan
number
one
and
it's
the
shirt.
B
Is
the
back
story,
so
the
communities
thing
is
pretty
interesting.
It's
like
any
of
these
orchestrations
frameworks
that
are
vying
for
number
one
yeah,
mezzos
cooper,
nay
days
now
docker
swarm.
Who,
who
knows
what's
next,
but
it
is
the
thing
I
mean
when
we're
coming
coming
around
to
this
this
cloud,
world
and
containers,
etc.
B
Cuckoo-Bananas
mezzos,
Dockers
form
they're
about
deploying
a
lot
of
something
and
controlling
all
the
aspects
of
it,
and
so
this
blog
post
on
Cooper
Nene's
or
it's
not
really
a
blog
post.
It's
a
readme,
but
it's
pretty
new
an
arrow
is
a
blog
post
and
involved
with
it.
This
is
the
the
guys
at
Cooper
Nettie's
trying
to
deploy
a
massive
clusters.
So
just
recently
they
released
a
bit
about
putting
out
a
thousand
node
cluster
with
communities
which
I
think
is
pretty
epic
I.
That
was
no
deployment
there.
B
You
go
that
that's
a
high-water
mark,
that's
a
good
one!
If
you
need
more
than
that,
then
you
should
call,
but
it
I
I've
worked
with
these
guys
a
bit
there.
There
are
some
really
interesting
stuff
happening
in
this
project,
especially
with
seed
nodes
and
how
it
manages
a
infrastructure
coming
and
going.
B
That's
that's
probably
one
of
the
big
challenges
as
as
you're
deploying
in
the
cloud
or
using
something
like
uber
natives
potentially
have
nodes,
it
can
disappear
on
you
and
the
framework
is
built
around
this
idea
of
bringing
it
back
to
life
like
if
you
have
a
state
that
you're
trying
to
obtain
a
thousand
node
cluster
and
one
disappears,
it's
going
to
try
to
retain
that
stasis.
No,
no.
There
has
to
be
a
thousand
of
these,
and
this
this
blog
or
this
readme
is
really.
How
do
you?
A
Situation
where
you
have
a
network
partition
between
availability
zones-
and
you
know
in
your
thousand
up
closer-
you
lose
a
third
of
them
right.
Is
it
going
to
just
prove,
like
all
of
a
sudden,
take
that
three
seconds
of
I
can't
talk
to
the
other
machines
and
provision
300
new
machines?
Or
is
it
smart
enough
to
like
you
know
not
to
that
horrible
thing?
Well,.
B
I
any
anytime,
you
use
an
orchestration
framework
like
this.
I
always
tell
people
do
not
do
anything
automatically
and
it
should.
It
should
make
you
aware
of
the
state,
and
then
you
should
push
the
button
to
retain
that
state
and
for
just
those
reasons
we
haven't
gotten
to
the
point
where
we
have
AI
renting
our
infrastructure.
It's
not
aware
of
things
so
partitions
are
a
great
example
right
and
data
soon
Ellen
see
another
data
center.
You
don't
know
how
long
that's
going
to
be.
B
One
of
the
things
you
don't
want
is
that
if
let's
say
you
have
a
500
nodes
and
one
data
center
of
500
notes
and
another
and
there's
a
partition
event
that
they
both
think.
Oh
no
I'm,
missing.
Half
my
cluster
and
spit
up
500
nodes,
each
number,
it's
just
completely
pathological
behavior!
You
do
not
want
that
very
tragic,
tragic,
because
you'd
have
a
lot
of
nodes
that
mean
nothing
to
you
anymore.
That's
a
lot
of
money
to
this.
B
Even
auto-scaling
groups,
the
SGS
in
Amazon
I,
never
really
got
comfortable
with
a
SGS
working.
You
know
totally
automatically
and
I,
never
for
the
database,
never
for
the
database
and
for
even
for
web
servers.
You
know
bad
things
can
happen,
you
get
DeeDee
hasta
notes
and
you
have
a
thousand
in
genetic
servers
running
mm-hmm.
B
A
On
yeah,
anything
from
you
guys
any
thoughts
of
us
Luke.
Should
we
hi
Luke
Snyder,
don't.
C
B
A
Why
don't
why
don't
we
move
on
second
half
guess
what
we
got:
Jeff
carpenter
hanging
out,
no
video,
unfortunately,
but
that's
okay,
he's
still
with
us!
I
hope,
jerry-.
B
D
E
I'm
really
excited
about
the
new
book,
as
as
John
was
saying
before
you
know
there
was
this
Cassandra,
the
definitive
guide
out
there
since
2010
and
its
really
well
written.
I
think
that
first
edition,
but
it
was
obsolete
because
you
know
it
was
written
against
0.6
and
then
Evan
rewrote
it
against
0.7,
because
things
were
changing
so
fast
and
you
can
even
find
things
about
0.8.
So
it's
really
old.
It's
all
about
column,
families
and
super
column
families.
E
You
know.
Since
then
we
have
secondary
indexes.
We
have
materialized
views,
we
have
lightweight
transactions.
Now
we
have
sassy
indexes,
so
everything
has
changed
so
much.
It
was
pretty
close
to
a
total
rewrite
of
the
book
which
I
I
was
super
fortunate
to
be
given
the
privilege
of
doing
you
know.
I
evan
is
my
boss
at
choice:
hotels,
he's
our
chief
technology
officer
and
I
definitely
saw
people
come.
E
And
he
people-
you
know,
people
have
said
to
him:
hey
this
book,
helped
me
get
my
job,
I
learned
Cassandra
and
has
helped
me
get
a
new
job,
so
I
thought
wow
I
mean
if
I
have
the
opportunity
to
help.
Do
that
and
help
do
something
that
will
really
impact
people.
That's
what
I
want
to
do.
So
you
know
I
was
I
was
to
Cassandra
expert
when
I
started
writing,
but
the
great
thing
was
I
I
had
to
learn
everything
so
that
I
could
write
about
it.
So.
D
E
Be
be
at
home,
writing
on
the
weekends-
and
I
learned
things
about
you-
know
level,
compaction,
strategy
or
hey.
I
didn't
realize
that
no
tool,
TP
stats,
you
know-
gives
me
things
that
I
can
actually
use
and
plug
into
our
monitoring.
So
I
was
learning
things
all
the
time
and
then
bringing
them
into
work
as
we
were
going
through
our
growing
pains,
a
choice
started
to
use
Cassandra,
so
it
was
a
kind
of
a
mutually
edifying
exercise.
A
Very
cool
all
right,
yeah,
the
point:
6
information
not
generally
useful.
Today,
yeah
you
got
it
actually,
I
would,
I
would
say
almost
zero
words
in
that
book
would
be
helpful
if
you
were,
if
you
were
starting.
B
E
On
the
30
series,
so
I've
run
programs.
I
think
I
know
for
sure
that
Iran
thing
is
against
30
I,
don't
know
the
minor
version
number,
but
I
could
look
it
up.
So
it's
it's
quite
current
and
I've
been
tracking.
The
book
also
has
I
mentioned
some
of
the
features
that
have
come
in
since
three
dot
0
like
the
SAS
II
indexes,
for
example.
So
you
know
oh.
D
Don't
read
my
blog
post
yeah.
E
B
B
B
E
You
know
I,
think
you
cannot
stress
the
importance
of
the
data
model
and
done
as
many
times
as
you
think
you
have
it
right,
it's
just
when
you
get
it
into
production,
you
get
it
at
scale.
You
learn
things
that
that's
the
only
way
that
you're
going
to
learn
so
there's
major
components
of
our
data
model
that
we're
on
the
third
generation
of
the
schema
and
that's
okay.
D
B
A
You
don't
you
know
you
always
forget
about
like
one
thing
or
the
product
changes,
I
find
and
all
of
a
sudden
you're,
like
oh
great
I,
gotta
change
these
few
things.
Hopefully
materialized
views
makes
out
a
little
bit
easier
to
deal
with.
E
A
B
E
Well,
you
know
the
it's
funny
because
I
love
Cassandra,
so
you
know
I'm
a
major
advocate
for
it
within
our
organization,
so
you
know
I
kind
of
love
it
for
what
it
is
and
no
for
what
it
is.
The
complaints
that
I
get
all
the
time
is
I
can't
do
a
joint.
You
know
what
do
you
mean
if
I
want
this
data
I
have
to
have
to
put
at
my
table
and
I?
E
C
E
C
E
You
know
I
think,
there's
all
kinds
of
little
surprises
I,
you
know
I
think
as
great
as
the
datastax
documentation
is.
There
are
some
things
that
you
can
still
only
learn
by
going
and
down
into
the
source
code.
So
you
know
there
was
a
blog
post.
You
know
even
a
couple
weeks
ago
where
I
was
like.
What
do
you
mean?
There's
only
one
ring
I
thought
there
was
a
separate
token
ring
in
every
data
center,
not
yeah
that
was
Luke's.
Blog
post
sounds
like
you
know.
D
D
E
E
E
B
You
know
we
and
we
have
some
good
authors
out
there.
It
would
be
great
to
have
like
a
little
book
signing
and
you
know,
like
Robbie
Strickland
has
a
pretty
good
book
on
high
performance,
Cassandra,
so
I
think
it'd
be
great
to
have
you
guys
there,
because
you
I
mentioned
this
to
you
and
a
few
others.
You
know
this
you're.
What
you're
doing
is
the
part
of
the
Cassandra
community.
Yes,
this
is
a
commercial
book,
and
this
is,
you
know,
Riley
publisher,
but
this
is
something
let's
face
it.
B
This
is
not
something
you're
making
a
lot
of
money
doing.
This
is
a
labor
of
love
and
you're
you're
doing
a
great
service
for
the
community,
because
it's
true
I
mean
people
love
to
have
books
on
their
desk
when
they're
doing
stuff
and
or
a
PDF
that
I
can
search
and
look
through.
It's
just
that
referenceable
materials
yeah.
E
E
I
have
to
thank
you
over
6000
thinks
that
you
know
I
love
the
videos,
the
datastax
academy.
That
stuff
is
so
great.
The
you
know
the
online
documentation
that
you
guys
have.
You
know
I
learned
a
ton
from
that,
but
I'll
tell
you.
I
have
a
co-worker
said:
yeah
I
tried
those
watching
those
videos
I
just
can't
it
doesn't
do
anything
for
me.
I
just
I
need
to
sit
down
and
read
something.
So
it's
like
there's
so
many
different
ways
to
learn.
A
A
B
A
B
A
Very
cool
yeah
Jeff,
thanks
for
coming
on,
really
appreciate
it.
I
was
great
talking,
you
I
think!
That's
it.
Let's
shut
it
down
thanks.