►
From YouTube: Swarming in Docker
Description
A screencast tutorial on how to run NuPIC swarms in a Docker container.
A
Hi
everyone,
my
name,
is
Austin
Marshall
I'm,
an
engineer.
Elementa
I
contributed
to
me
a
pic
and
other
projects
and
I
wanted
to
demonstrate
how
you
might
use
docker
to
run
a
swarm
in
an
isolated
environment.
Dr.
is
pretty
nice
because
it
allows
you
to
create
reproducible,
isolated
environments
that
can
be
transient.
A
A
There
are
instructions
for
running
a
swarm
with
docker
and
today
I'll
be
using
a
core
OS
instance
that
I'm
running
with
vagrant
and
virtualbox
on
my
own
laptop
the
instructions
for
doing
that
yourself
are
here,
but
the
instructions
apply
to
any
environment
which
you
have
access
to
a
doctor
command
and
a
doctor
host,
and
so
let's
get
started.
The
first
thing
that
we
need
to
do
is
start
on
my
sequel
container.
A
A
Just
a
few
points
to
look
at
we're
going
to
call
it
new
pic
dash,
my
sequel,
we're
going
to
assign
the
loot
password
to
big
new
pic,
which
implies,
will
be
using
the
default
root
user.
We're
going
to
listen
on
port
3306,
we're
going
to
demonize
the
product.
The
process.
This
means
a
little
bit
running
in
the
background.
A
A
A
Before
I
got
started,
I
created
a
swimming
scenario
based
off
of
the
early
in
korean
freshman
under
the
simpler
capital.
So
I
have
a
csv
file
that
looks
like
that
and
that's
in
a
directory
called
my
form
and
the
file
for
that
CFE
implement
this
used
to
buy
a
small
underscore
test
by
CSP,
and
then
I
have
a
search
underscored:
def
Jason
file
that
consists
of
the
contents
of
the
second
here,
but
with
the
comment
removed.
So
it's
college
days.
A
A
There's
some
additional
swarm
parameters
in
there.
I
won't
get
into
the
details
of
the
swarm
parameters.
If
you'd
like
to
learn
more
about
that,
there's
more
information
on
our
wiki,
but
it
just
want
to
walk
through
the
mechanics
of
running
a
storm
and
docker.
So
now
that
I
have
my
swarms
set
up
again,
that's
on
a
local
directory
of
my
sworn
that
contains
these
two
files.
A
The
next
thing
I
want
to
do
is
actually
run
by
swarms,
and
in
order
to
do
that,
I
will
use
docker
to
run
the
new
pic
container
with
a
command
that
runs
that
swarm
and
so
I
adopted
this.
This
next
steps
command
here
to
match
my
environment,
and
I
will
copy
and
paste
that
from
my
notes
here.
So
if
we
look
up
over
there
real
quick,
we
can
see
that
you
just
make
that
real,
quick.
A
New
pic
has
a
configuration
screen
that
lets
me
specify
configuration
directives
as
environment
variables,
and
so
this
one
right
here,
which
is
an
entity
a
top
new
pic
cluster
davis
password,
is
overriding.
The
nuh
tek
da
cluster
database
password
configuration
director
that
we
would
see
elsewhere
in
the
code
someone.
A
It's
warm
just
looking
for
these
two
parameters
to
determine
the
password
in
the
host,
so
the
value
that
we're
going
to
pass
for
the
password
should
match
the
value
that
we
passed
from
my
super
root
password
in
the
earlier
commands
with
onto
my
sequel
container
and
the
host
is
going
to
be
my
sequel.
It's
going
to
be
my
sequel
because
we're
taking
advantage
of
this
link
command-line
argument
to
dr.
A
run,
which
tells
us
to
link
the
my
sequel
hostname
to
the
new
pic
dash,
my
sequel
container
that
we
created
earlier
we're
also
going
to
create
a
volume
within
this
container.
The
mountain
point
for
that
volume
within
the
container
will
be
/.
My
swarm
and
the
source
of
it
will
be
the
mice
form
directly
that
I've
created
locally.
A
A
A
So
I'm
just
going
to
copy
this
and
paste
it
here,
and
so,
let's
give
you
that
one
more
time
or
point
dr.
Rudd
specifying
the
to
environment
variables
to
specify
our
my
sequel
credentials.
We
have
our
volumes
set
up
so
that
the
local
directly
that
we
created
called
my
swarm
as
our
swamp
parameters
are
available
in
the
container
at
the
location.
/.
A
A
A
A
A
Now
we
about
a
bunch
of
extra
files,
I've
learnt
their
previous
days.
You
have
a
description,
duck
py
file,
the
permutations
duck,
do
I
file
the
pickled
model
that
we
dated
basically
the
model
checkpoint
and
then
the
report
that
was
reported
earlier
and
if
we're
looking
in
my
spawn
model,
0d
some
additional
files,
so,
for
example,
I
wanted
to
create
a
CLA
model
with
the
monograms
that
were
the
result
of
its
form
that
could
find
that
and
model
cram
stops
you
why
it's
like
that?
Ok.
A
It's
this
should
all
look.
No,
you
have
much
experience
working
with
the
dopa,
which
is
the
new
pic
online
prediction
game
like
and
it's
everything
that
you
need
in
order
to
start
a
model
and
potentially
pick
up
where
this
one
left
off
and
there
you
have
it.
You
have
everything
that
that
you
need
in
order
to
run
a
a
swarm
and
dr.
hope
you
enjoy.