►
From YouTube: New User Training: 13 Spin Early User Program
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
So
this
one
is
not
about
quantum
mechanics.
This
is
talk
about
spin
I'm,
Tony,
Wildish,
I
work
in
the
data
science
engagement
group,
but
I'm.
Giving
this
talk
on
behalf
of
the
spin
working
group,
which
is
mostly
the
infrastructure
services
group
spin
I'll,
get
to
what
it
is
in
a
minute
if
lists
start
with
the
problem
that
we're
trying
to
solve
in
general,
in
a
complex
system
like
nurse
large
experiments,
large
collaborations,
you
do
more
than
just
log.
In
run,
a
few
commands
run
batch
jobs.
A
Typically,
people
run
all
sorts
of
services,
databases,
web
servers,
workflow
managers,
things
like
this,
and
not
just
things
that
run
in
the
background
as
batch
jobs,
so
also
in
distributed
collaborations.
There
are
quite
frequently
services
that
communicate
across
the
sites
involved
in
that
so
there's
all
sorts
of
things
and
where,
typically,
these
services
would
be
on
dedicated
hardware,
you
would
buy
a
node
for
running
your
web
server.
You
would
have
a
system
administrator
install
things
on
it.
For
you,
this
is
a
very
expensive
way
of
doing
things.
A
It's
slow
its
clunky,
the
alternative
which
a
lot
of
people
do
is
to
just
install
stuff.
In
their
home
directory
and
run
it
out
at
their
home
directory,
which
is
also
not
very
satisfactory,
because
it's
not
well
managed
it's
not
immune
to
things
like
down
times,
system
changes
stuff
like
that
and
spin.
The
whole
purpose
of
spin
is
to
give
these
services
a
proper
home,
so
spin,
the
name
down
there.
The
name
came
before
the
acronym
I
think
that
classic
example
of
just
flying
the
acronym
after
we
found
the
name
for
it.
Scalable
platform
infrastructure.
A
It's
at
nurse
k--.
If
you
know
what
Amazon
Web
Services
is,
you
can
think
of
this
as
AWS
for
nurse,
though
I
don't
think
Corey
would
like
it
missus.
Let
me
saying
that,
but
that's
basically
a
good
way
to
think
about
it.
It
provides
you
a
place
to
run
services,
get
access
to
resources
for
things
that
you
want
to
run
as
persistent
services
here
at
Nazca,
so
it
runs
docker
containers
in
a
manageable
environment.
It's
a
flexible
scalable
platform
and
it's
tightly
integrated
with
nurse
resources.
A
A
So
why
is
it
running
docker?
Well,
docker
is
a
very
good
technology
and
you've,
probably
all
heard
of
it.
The
the
the
great
thing
about
docker
is:
you
can
develop
a
service
on
your
laptop
and
get
it
working
the
way
you
want
and
from
there
you
can
then
deploy
it
in
minutes
onto
spin,
and
you
can
be
pretty
certain.
It
will
run
exactly
the
same
way.
No
dependency
hell,
no
I've
got
version.
This
and
you've
got
version
of
that
of
MySQL.
None
of
that
once
it's
on
spin,
if
you
find
it,
it's
suddenly
popular.
A
Your
paper
goes
viral
and
everyone
starts
hitting
your
website.
You
can
have
the
service
automatically
scale
out,
put
a
load
balancer
in
front
of
it,
and
then
it
will
just
scale
out
automatically
for
performance.
You
can
get
access
to
the
HPC
systems
here,
so
you
can
submit
batch
jobs
access
the
file
systems.
Things
like
this,
so
that
you
can
do
things
like
have
a
web
service
which
receives
incoming
data,
submits
a
batch
job
to
run
on
that
and
then
provides
the
data
back
through
the
user
who
submitted
the
sample
and
nurse,
manages
all
of
this.
A
A
A
You
can
create
a
new
database
web
portal
or
other
kind
of
application
on
demand,
you'll
be
able
to
just
log
in
to
spin
and
spin
one
up
the
way
you
want
to
you
don't
need
to
worry
about
any
of
the
details
of
the
underlying
infrastructure.
What
machine
am
I
running
on
you?
Don't
care
about
that?
You
get
a
URL
for
free,
you
get
a
stable,
DNS
name
all
this
sort
of
thing.
A
You
will
not
see
any
breaking
service
if
your
service
is
designed
correctly,
and
so
you
you
will
be
you'll
get
far
higher
uptime
than
any
of
the
individual
machines
would
plus
using
docker
images
is
really
very
useful.
Docker
is
very
widely
accepted
by
the
community.
There
are
a
large
number
of
standard,
dr
images
out
there
that
get
you
started
if
you
were
to
put
together
a
web
service
and
a
database
server.
You
just
look
for
the
official
apache
docker
image
in
the
official
MySQL
image
and
you're,
often
going
up
and
running.
A
A
A
It's
often
compared
to
a
virtual
machine.
It's
not
really
like
a
virtual
machine,
but
you
can
think
of
it.
That
way.
If
you
like,
it's,
like
a
mini
virtual
machine
that
you
can
carry
around
with
you,
it's
very
easy
to
learn
the
basics.
Sara
tons
of
good
documentation
out
there
and
lots
of
good
tutorials,
just
Google
docker
tutorial
and
you'll
find
one.
Somebody
somewhere
has
put
together
a
simple
tutorial.
It
does
exactly
what
you
want
and,
as
I
say,
there's
a
large
ecosystem,
lots
of
very
good
examples.
Now
there
to
start
with
already.
A
A
This
l2
I'm,
repeating
myself
a
bit
here.
This
gives
you
a
very
efficient
use
of
resources
as
well.
The
fact
that
the
containers
are
lighter
weight
than
virtual
machines
means
that
you
can
pack
more
of
them
into
the
machines
that
are
being
used
behind
the
scenes
on
this.
So
there's
really
a
great
deal
of
flexibility
on
the
nurse
side
in
making
sure
that
things
are
scheduled
and
shared
around
efficiently.
A
So
there's
a
lot
of
terminology
that
goes
along
with
this
container
image
service
and
stack
container.
An
image
tend
to
be
used.
Interchangeably
I
have
to
read
this
myself
to
get
it
right
and
I
had
to
look
it
up
to
make
sure
I
got
it
right
first,
so
the
image
is
a
stand-alone
software
package.
The
standalone
piece
of
software
that
contains
everything
you
need
think
of
this
is
the
the
DVD
with
the
installation
on
it.
So
it
contains
the
the
package
when
the
container
is
a
runnable
instance
of
that.
A
So
you
take
the
image
and
you
start
a
container
from
it.
So
the
service
is
typically
comprised
of
one
or
more
containers.
Web
server
database
server,
load
balancer-
something
like
this,
which
put
together,
will
all
provide
a
single
piece
of
functionality,
a
single
capability,
so
a
web
server
might
have
two
containers.
It
might
have
a
front-end
for
authorization
in
a
back-end
which
actually
does
the
request
processing.
A
You
can
separate
it
out
like
that,
and
then
you
can
have
these
two
developed
by
separate
people
as
long
as
they
agree
on
the
API.
It
works
nicely
or
you
can
have
a
database
server
with
a
single
container
or
more
than
one,
if
you
wish
put
all
these
services
together
and
that
forms
an
application
stack.
A
A
This
load
balancer
is
serving
a
web
service
which
has,
in
this
case
two
container
instances
running
so
it's
load
balancing
across
the
two
of
them,
and
these
two
web
services
are
both
using
a
key
value,
store
service,
maybe
MongoDB
in
one
case,
and
also
a
database
service
as
well.
So
both
of
these
web
service
instances
can
be
using
both
those
you
can
easily
put
together
any
sort
of
level
of
complexity
that
you
need
for
your
application.
A
If
we
were
to
just
give
you
a
bunch
of
virtual
machines
and
say
you
can
start
your
stack
there,
then
you
wouldn't
be
much
better
off
really,
then
you
would
be
just
trying
to
run
it
out
of
the
home
directories.
You
still
need
to
make
sure
that
things
can
be
scheduled.
They
can
start
on
when
you
want
them
to
start.
They
can
restart
after
system
upgrades
things
like
this,
that
you're
not
swapping
resources
because
you're
using
the
same
machines
as
everybody
else.
A
So
for
this
you
need
an
orchestration
service,
an
orchestration
tool
and
the
tool
we've
settled
on
is
called
Rancher.
It's
commercially
available
products,
we've
adopted
it
here
and
the
idea
behind
this
is
that
you
can
you
can
upload
your
application
stack.
You
can
just
give
it
a
list
of
docker
images
and
tell
it
how
these
are
plumbed
together.
A
This
service
uses
port
so-and-so
to
connect
to
that
service
and
just
simple
configuration
in
a
text
file
and
then
Rancher
will
make
sure
that
your
service
is
running
in
the
way
that
you've
described
so
underneath
it
you'll
have
a
bunch
of
nodes.
These
are,
in
fact,
virtual
machines.
In
our
case
and
these
nodes
will
be
running
application
stacks,
a
variety
of
different
applications,
tanks,
all
mixed
together
in
various
ways-
you
can
specify
how
this
mix
happens.
A
For
instance,
if
you
want
to
make
sure
that
your
database
back-end
is
not
only
load
balanced,
but
that
each
instance
of
it
is
on
a
separate
physical
host.
You
can
tell
it
to
do
this
and
it
will
then
spread
them
out
in
the
same
way
in
the
right
way
and
when
a
rolling
upgrade
of
the
operating
system
is
done.
These
services
will
be
stopped
one
by
one
in
a
way
which
keeps
your
entire
stack
running
and
will
then
bring
them
back
afterwards.
A
So,
yes,
it
manages
failover
service
ownership
and
then
the
all
than
the
details
of
management,
which
you
really
don't
want
to
care
about.
If
you
can
avoid
it,
so
the
philosophy
behind
docker,
it's
often
summarized
as
build
ship
run,
you
build
it,
you
send
it
somewhere
and
you
run
it
there,
and
it's
that
simple.
You
don't
have
to
install
deploy
patch
figure
out
how
to
make
this
installation
work
on
that
machine,
because
the
docker
file,
the
docker
image,
does
it
all
for
you.
A
So
on
the
left
there
on
your
laptop,
you
start
from
a
docker
file
the
text
file,
which
describes
your
your
service,
and
from
that
you
build
a
docker
image.
The
image
then
gets
sent
to
a
registry.
We
have
one
in
spin,
so
the
spin
registry
is
a
place
where
you
would
send
your
images
to
and
from
that
registry
your
application
stack
in
blue
there.
A
Your
application
stack
would
say,
I
want
this
image
from
the
registry
and
I
want
that
image
from
the
registry
plumbed
together
in
various
ways:
different
ports,
different
services,
different
environment
variables,
different
configurations,
whole
set
of
services
all
plugged
together,
and
that
forms
a
replication
stack
which
provides
your
your
total
functionality.
You
can
also
pull
images
from
docker
hub
if
you
wish,
or
from
elsewhere,
on
the
Internet
you're,
not
just
constrained
to
pulling
from
spin.
A
So
if
you
have,
if
you
want
to
use
a
stock
of
Apache
image,
you
can
just
pull
it
straight
from
the
dock
of
Apache
images.
Things
like
that
and
then,
within
this
your
your
stack,
your
application
stack
can
access
things
like
the
global
file
systems.
That
Lisa
was
talking
about
earlier
through
a
volume
mount
so
that
you
can
mount
deep
into
the
file
systems.
You
can
get
read/write
access.
If
you
need
it
or
through
API
is
in
SSH.
A
So
who's
using
it
so
far,
we
currently
have
a
number
of
early
adopters,
the
CID
division,
computational
research.
They
have
a
development
project
on
there,
which
is
soon
moving
to
production.
There's
the
ESS
dive
data
archive
collaboration
between
nurse
and
Ciardi,
that's
also
using
spin
and
at
the
joint
genome
Institute.
We
have
several
users
who
have
moved
the
number
of
applications,
their
web
portals
databases,
things
like
that
also
within
ask:
we
are
using
it
internally
for
a
number
of
things
license
servers
the
compilers
that
we
heard
about
earlier.
A
A
Pi
Tokyo
I
won't
talk
about
about
that
much,
but
that's
an
interface
for
looking
at
file
system,
performance
and
Bechet
or
performance,
and
the
multi-factor
authentication
that
we
also
heard
about
earlier,
that's
being
prototyped
in
ask
spin,
so
edge
services,
those
of
you
from
the
particle
physics,
communities
will
know
about
frontier
cash
and
CBM
FS
and
those
are
running
inside
spin
as
well
and
also
internally.
We
use
it
for
our
CFG
group,
use
it
for
their
own
documentation
and
the
shifter
registry
which
you'll
hear
about
ship.
The
next.
A
The
shifter
registry
is
also
stored
in
spin,
so
there's
all
sorts
of
edge
services.
If
you
want
to
run
something
which
is
not
an
interactive
command,
that
you're
waiting
for,
and
it's
not
a
patch
job
that
you're
going
to
run
and
on
its
own
and
come
back
if
it's
something
that
you
want
to
run,
that
runs
long
term
and
you
need
to
know
that
it's
constantly
running
you
should
be
thinking
about
using
spin
for
it.
A
So
the
current
status
spin
is
currently
in
pilot
mode,
so
internal
projects
for
mask
several
users,
such
as
a
joint
genome
Institute,
who
are
working
directly
with
nurse
staff,
and
it
will
soon
be
open
to
trained
users
so
soon
be
going
into
beta
mode,
so
that
people
who
have
had
some
training
on
it
will
be
now
to
use
it
directly
through
an
API.
So
next
question
is:
when
do
you
get
the
training?
A
The
first
spin
up?
I
did
not
choose
that
name.
The
first
spin
up
hands-on
training
session.
This
planned
for
mid-may
capacity
will
be
limited
at
first
because
we
want
to
learn
how
to
we
have
a
learning
curve
on
how
to
support
you
as
much
as
you
have
the
learning
curve
on
how
to
use
the
service.
So
there
will
be
an
application
available
on
the
website.
I'll
show
you
the
website
in
a
minute
and
just
watch.
The
weekly
newsletter
Rebecca
sends
around
a
very
nice
weekly
newsletter,
which
everybody
reads:
don't
they
yeah
she's
got
you
named.
A
So
what's
the
weekly
newsletter
for
announcements,
the
training
will
be
announced
in
there
and
then
go
to
the
website
and
sign
up
so
the
website
under
the
nurse
gov
for
users,
data
analytics
and
services
and
then
well.
It
fits
on
my
screen,
but
it
doesn't
fit
there
down
at
the
bottom,
you'll
find
spin.