►
From YouTube: OpenCrowbar Community Planning Meeting 2014 06 04
Description
http://pad.opencrowbar.org/p/community-planning-meeting-2014-06-04
Agenda:
For Review Stories
OpenCrowbar community website, will take pulls since this is github!
List Documentation Changes Needed
https://github.com/opencrowbar/core/issues/144
Review general (website content)
* "Just Enough Orchestration" very short description of Crowbar.
* http://pad.opencrowbar.org/p/alternatives-review
* http://pad.opencrowbar.org/p/ready-state-strengths
* principles content (reposted by Rob last week)
100 Node Challenge
Assuming incremental PXE boot (not 100 all at once)
Hardware config TBD (assume 1st article available as loan before install)
A
Hello,
everybody
welcome
to
the
June
4th
open
crib,
our
community
planning
meeting.
This
is
an
hour
by
hour,
equally
cycle.
This
is
our
planning
meeting.
Next
week
we
do
off
off
weeks.
We
do
the
design,
design
topics
and
set
up
this
agenda
in
our
in
our
last
meeting,
I
think
with
a
lot
of
great
material.
What
I
would
suggest
we
do
is
talk
through
the
stories
that
are
already
up
and
then
the
missing
missing
pieces
we
can
just
fill
in
and
type
collaboratively
on.
The
join
me
oops.
A
A
So
some
that
sound
reasonable
and
then
the
other
things
we've
actually
done
a
lot
of
work
on
this
ready
state
strengths,
wiki
page.
So
if
you're
worth
looking
at
that
and
talking
about
that
a
little
bit
altera's
review
and
what
we
should
also
do
is
talk
about
the
principles
content
which
I
reposted
last
week,
you
so
that
might
be
something
that
people
can
review
talking
about
crowbar
up
and
Kirk
community
website.
We
had
an
action
item
for
reviewing
in
this
defect.
A
C
I
took
a
look
at
effectively
online,
continuous
integration,
servers
and
services.
A
couple
of
the
criteria
that
I
made
up
for
myself
that
we
didn't
really
document
was
that
it
was
an
online
service
I,
be
that
it
would
support
the
community,
in
other
words,
what
we
didn't.
What
I
didn't
want
to
do,
although
I
looked
at
the
number
of
these,
the
customers
or
offering
CI
services,
but
they
were
doing
him
on
a
per
seat
license
basis
in
order
to
support
the
community
I
believe
that
it
should
probably
be
open.
C
Otherwise
we
would
have
to
purchase
a
you
know
an
unbeknownst
number
of
seats
in
order
to
handle
it.
It
also
needed
to
be
able
to
support
the
gating
concept
that
we
were
looking
for
and
what
that
basically
is
is
prior
to
code
being
committed
into
the
into
the
main
liner
or
production
branch.
We
would
want
it
tested
via
the
CI
system
and
have
the
CI
system
effectively
float
as
to
whether
it
past
the
the
integration
and
unit
tests
prior
to
us
actually
merging
it
in,
and
then
the
merge
would
be
done
manually
by
an
engineer.
C
So
those
are
kind
of
the
main
criteria.
I
looked
at
a
god
like
10
or
15
of
these
things,
the
main
deficiency
I've
seen
on
these
ones
that
are
out
there
other
than
the
fact
that
there's
a
number
of
projects
that
have
kind
of
gone
dead
is
that
they
only
support
a
single
operating
system
and
one
of
the
one
of
the
things
that
we
wanted.
Crowbar
to
be
able
to
do
is,
or
the
CI
system
to
be
able
to
test
the
integration
across
a
number
of
operating
system
platforms.
C
So
I
think
we're
effectively
going
to
have
to
forget
initially,
unless
we
want
to
extend
one
of
these
one
of
these
tools
arm
the
majority
of
them
use
ubuntu
and
the
online
ones
in
particular,
either
host
their
own
series
of
servers,
basically
virtual
machines
that
do
the
testing
or
effectively.
You
know
you
on
as
a
basis
in
which
the
party
is
using.
It
actually
needs
to
needs
to
fund
those
amazon
instances
anyway,
make
a
long
story
short.
C
What
it
looks
like
is
travis
CI
appears
to
be
the
the
quickest
and
easiest
one,
given
the
fact
that
it
integrates
well
with
with
the
git
repositories,
and
it
would
effectively
handle
a
abun
to
CI
system.
It
can
easily
be
extended
to
handle
the
Erlang
unit
tests
that
are
built
into
crowbar,
as
well
as
deploy
prerequisite
such
as
the
Postgres
database
that
we
need
and
other
such
items.
So
so.
A
C
F
C
C
E
C
D
C
Case
a
pretty
easy
to
pretty
easy
to
work
with.
We
could
probably
pull
it
off
in
pretty
short
order,
at
least
having
the
unit
tests
that
are
built
into
the
system
run
as
a
initial
gain.
Additional
smoke
functions
would
require
some
some
more
development.
I
think,
unless
we
already
have
the
scripted
summers,
that's.
A
A
A
D
E
B
E
D
F
A
D
A
Mike
tomorrow,
so
I'll
touch
base
with
him
yeah
p,
but
it's
about
this
and
whatever
it
takes.
One
things
that
would
I'm
hoping
to
do
is
connect
some
of
this
docker
work,
so
the
disk
mounting
stuff
I'd,
probably
we're
not
going
to
do
inside
ocker,
but
all
the
other
readiness
prep
would
be
interesting
to
do.
Inside
of
docker
containers,
I'm
going
to
free,
say
that
same
thing
for
OpenStack,
because
I
think
people
would
like
to
see
us
able
do
that.
Yeah.
C
A
F
C
Well,
it
sounds
like
you
know
when
I
hear
things
like
I
think
people
would
like
it.
What
that
makes
me
think
is:
there's
a
higher
level,
epic
for
like
developer
access
or
ease
of
development
or
some
something
to
that
effect.
We
could
actually
take
a
more
holistic
approach
to.
B
A
Let
me
take
a
second
to
write.
It
I,
don't
want
to
hate
today,
I'm
just
going
to
try
and
hand
off
yeah
I'll
type
that
later
okay,
so
you
weren't
saying
to
break
the
Hadoop
readiness
down
into
into
smaller
stories,
although
I
think.
D
C
I
wasn't
particularly
talking
just
about
the
Hadoop
one,
but
you
know
there
are
things
that
we
think
would
help
the
community
and
help
additional
developers
come
up
to
speed.
It'd,
be
good
to
kind
of
collapse
them
holistically
as
opposed
to
picking
around
the
edges
and
that's
what
I
meant
by
a
buy
it
and
epic,
potentially
potentially.
A
A
F
F
On
the
on
demand,
a
dip
readiness
program
I
was
more
as
traditional
requirements
that
actual
yes
and
but
so
we
need
to
back.
The
tricky
part
is
I'm
trying
to
focus
initially
on
getting
those
running
for
her
get
specific
whales.
Without
that
again
too
much
distress,
specifics,
which
is
a
little
tricky.
F
There
was
good
talk
in
life
that
I've
been
the
team
of
Bloomberg.
They
have
done
a
number
of
chef
recipes
but
deployed
up
from
the
pure
Apache
big
time.
I
had
this
one
Tory's
drunk
without
being
district,
and
if
I'm
gonna
go
back
and
look
at
some
of
the
things
they
did,
I
haven't
had
a
chance
to
look
at.
Their
click
only
may
be
able
to
leverage
some
of
the
stuff
they've
done
as
well.
F
Yeah,
that
sounds
good.
I
think
the
first
step
is
to
to
start
sketching
out
the
bar
clamp.
We
we've
more
or
less
done
this
in
the
past,
so
there
is
some
existing
dell
code.
That
does
a
lot
of
this
that
really
needed
to.
They
really
need
some
refactoring
to
simplify.
It
was
overly
complex
when
it
was
split
out
from
the
distro
specific
stuff,
so
those.
A
A
A
It's
a
matter
of
carving
the
Sabbath
its
own
gig.
So
the
idea
is
that
you
can
use
WS
man
to
enter
interact
with
red
bias.
There
is
a
gem
that
does
WS
man
calls
and
then
bring
that
in
and
then
I
documented.
A
couple
of
the
items
like
crowbar
can
talk
to
the
BMC
network
directly
to
the
crib
or
one
proxy
thing
should
be,
should
not
be
needed,
which
was
adding
a
lot
of
complexity
and
some
of
the
alkene
lon
configurations.
G
A
C
Have
the
design
doc,
so
we
can
either
publish
that
to
the
list
for
comments,
which
is
probably
an
appropriate
thing
to
do.
Although
the
design
doc
is
in
reviewing
it
it's
more
of
how
can
we
make
what
crowbar
one
does
better,
but
it
does
define
a
series
of
int'l
libraries
and
stuff
that
would
need
to
be.
You
know
done
in
this
particular
this
particular
design
infrastructure.
A
B
C
B
C
A
G
A
D
A
B
C
B
A
Would
I
would
set
as
an
initial
objective
that
it
does
nothing
but
build
actually
does
nothing
but
the
yes.
When
there's
when
there's
a
commitment,
that's
our
code
at
all,
it's
just
gates
and
then
we
can
add
complexity
from
there.
Just
just
does
the
vote,
because
that's
really
the
key
function
right
and
then
we
can
add,
add
complexity
well,.
C
I
think
the
a
we
can
discuss
about
our
need
to
tie
this
off.
I
think
the
initial
thing
would
be
to
get
it
to
actually
build
right,
not
the
not
the
gating
system,
but
actually
do
the
dead,
build
functions
in
some
sort
of
basic.
You
know
no
op
test
arm
and
you
know
tying
it
into
to
a
gating
function.
I
think
is
going
to
require
some
some
tree
manipulation,
but
not
to
figure.
A
A
Alright,
that
brings
us
to
the
openstack
story,
and
so
this
one
I
tried
to
cut
that
I.
Think
the
biggest
thing
to
talk
about
with
this
is
what
the
target
is,
which
is
that
we're
trying
to
drive
to
make
it
possible
to
use
the
chef's
up.
Streams
is
the
way
I
wrote
this
and
I
was
hoping
to
have
some
some
discussion
to
make
sure
that
this
is
reasonable.
C
I
think
that
you'll
find
that
there's
a
number
of
things
that
need
to
be
modified
in
the
upstream
cookbooks
in
order
for
them
to
work
in
a
tool
such
as
crowbar
right,
there's
just
the
office
ke
shins
that
need
to
be
put
in
place.
I
think
we
have
identified
a
number
of
those
over
the
past
couple
of
months
and
fixed
work.
Look
at
least.
A
B
A
The
when
I
talk
to
people
about
the
cookbooks
they
jump
really
quickly
into
you
know.
I
want
open
stacks
of
neutron
with
this.
With
that,
you
know
with
this
configuration,
which
is
we
want
to
drive
to
that.
My
hope
is
that
what
we
want
to
do
is
show
people
in
the
upstream
community
that
they
can
use
crowbar
to
do
that,
and
then
we
can
have
it
as
more
of
a
community
thing
that
we're
not
we're
not
trying
to
make
Neutron
work
necessarily
in
as
our
first
step
right.
A
A
G
I
am
not
a
hundred
percent
sure
I
mean
there.
Definitely
do
they
definitely
had
code
to
deploy
neutron
as
far
as
a
nova
network
I,
don't
know
you
know
in
the
crowbar
one
cookbooks
there's
there
was
actually
some
work
done
to
remove
support
for
Nova
network,
but
it
wasn't
completely
done
so.
As
a
result,
those
cookbooks
are
in
a
really
messed-up
state
in
regards
to
Nova
network
support.
B
E
H
E
G
H
E
H
No,
that's
a
true
state.
There's
no
desk
know
that
it
doesn't
perform.
It
has
certainly
has
perform,
especially
with
the
MTU
size
and
stuff.
You
have
to
fix
all
that,
but
alright
I'm,
okay
ones,
you
don't
think
it
needs
switch.
Laporte
I
just
thought
we
needed
switch
support
to
have
VX
line
as
well
and.
E
H
D
E
A
Good
I
believe
that
the
more
explicit
we
are
with
this,
the
easier
it
will
be
for
people
to
come
in
and
collaborate
once
at
you.
Let's
demonstrate
the
base
actually
just
Keystone
and
database
and
messaging
with
the
on
top
of
the
docker
containers.
I
think
would
be
enough
for
people
to
start
trying
to
trying
to
help
get
this
get
this
running
and
then
and
then
what
I?
So
we're
going
to
we're
going
to
assume
are
documented
some
expectations.
A
Controller
nodes
right.
Obviously
this
requires
the
hardware
pieces,
but
I
thought
is
worth
documenting
and
from
a
networking
perspective,
getting
networks
and
teens
all
that
stuff
going
as
I'm
trying
to
try
to
make
sure
we
have
some
cleared
that
technical
and
then
this
is
the
story.
There's
a
prereq
story
that
Victor's
been
working
on.
D
G
E
B
G
E
A
Think
that
it's
important
to
acknowledge
that
Chris,
you
are
your
earlier
statement,
was
correct.
We
did
push
back
on
using
chef
server
and
in
the
first
in
the
earlier
pass
through
this
and
and
have
since
revised
that
position.
You're,
not
you're,
not
crazy.
If
it
was,
it
was
the
position
and
the
point.
A
F
G
E
E
Out
to
be
a
gigantic
pain
in
the
ass
to
not
use
a
chef
server,
while
using
burke
shelf
with
multiple
centrist,
cookbook
sources,
because
Burks
files
don't
compose
I,
can't
cuddle
everything
up
into
I
can't
bundle
up
single
tarballs.
It
contains
all
of
our
cookbooks.
Without
going
through
major
contortion.
There.
G
Okay,
yeah.
That
sounds
like
a
reason:
alright,
cool
thanks.
A
A
A
All
right,
and
then
we
added
this
other
story
about
documenting
using
containers
for
workload,
developments
so
I
just
that
this
one's
pretty
short
and
sweet.
So
I
wrote
a
couple
of
sentences
for
that,
but
I
do
think
this
will
make
it
easier
for
people
to
develop
the
do
for
openstack
workloads,
and
so
it's
it's
a
the
reason
to
work
on
you.
If
you're
using
the
up
streams
for
OpenStack.
This
should
make
it
faster
and
easier
to
do
real
test
all
right.
A
A
A
A
Is
I'd
rather
have
him
on
the
call
before
I
announce
people's
names,
but
so
I
have
somebody
who
hasn't
hasn't
been
active
in
the
community,
who
has
a
fresh
perspective,
who
has
a
lot
of
community
management
experience
and
a
lot
of
DevOps
experience,
and
they
were
interested
in
helping
pull
together.
The
material
actually
build
a
community
website.
This.
A
D
A
A
And
I
think
a
lot
of
this
content
is
really
helpful.
Actually
is
the
community
website
content,
and
so
what
I
would
ask
is,
if
you
get
a
chance
first,
if
you
want
to
learn
more
about
how
we're
shooting
crowbar
and
how
we're
describing
to
the
community,
these
pages
are
really
helpful
for
that.
If
you
find
the
formatting,
not
that
helpful,
I
actually
did.
A
We
you
know,
post,
post
Doc's
are
very,
very
welcome,
and
so
we
have
the
readystate
strength
so
I'm
going
to
move
a
plan
to
move
that
one
into
that
same
area.
Nes
on
this
would
be
appreciated.
I
think
it's
really
helpful,
because
Ruby
state
is
starting
to
be
accepted
as
a
term
we're
starting
to
hear
this
a
little
bit
more
industry,
I,
think
being
able
to
say
crowbar,
delivers,
ready,
state
and
people
understand
what
that
means
really
helps
the
other.
The
other
component
with
this
is
the
alternatives.
A
D
A
A
H
A
This
up
and
alternatives
review
at
the
same
time.
It's
because
the
more
I
talk
to
people
about
this,
the
more
we
are
sort
of
fitting
in
between
razor
and
interval.
So
ansible
does
they
describe
themselves
over
here,
simple
orchestration,
but
they're,
very
they're,
a
DSLR,
really
a
chef
and
puppet
competitor
where
crowbars
position
is
a
extension
of
chef
and
puppet.
So
people
can
understand
the
comparison
with
razor
or
cobbler,
where
we
provision
equipment
provision,
provision,
stuff
and
then
I
understand
that
there's
an
orchestration
gap
in.
A
A
We
we
have
some
developed
tooling,
that's
with
some
os's
and
things
like
that.
That
aren't
accurate.
We
can
clean
those
up
the
principal
sections:
I
did
the
late,
binding
I
need
to
add
ready
state.
Pretty
straightforward.
I
did
a
whole
bunch
of
oh
there's
a
deploy
guide
that
I'd
love
to
see
somebody
walk
through,
but
we
need
to
be
helpful
that
screen
snapshot.
They'll,
show
you
that
and
then
I
have
graphics
and
the
principles
god
I
used
in
the
blog
post,
but
didn't
link
in
the
principal
submit
their
actually.
There
Oh.
B
A
And
I'd
gone
through
and
actually
walked
the
steps
of
doing
a
deployment
with
menus,
and
things
like
that.
This
actually
should
end
up
being
split
into
multiple
documents
with
screen
snapshots
or
something
like
that
to
tell
people
the
steps,
but
all
right,
the
document,
all
the
information
is
all
there.
We
just
need
to
keep
running
through
it,
helping
people
on
your
sent.
You
know
helping
people
navigate
through.
A
The
eye
the
number
one
thing
that
we
need
is
people
hanging
on
Doc's,
that's
going
going
through
them
finding
if
there's
a
shoes
adding
to
them
I
would
really
encourage
people
to
just
just
too
little.
If
you
find
tweaks
little
tweaks
I've
been
working
with
decoy.
Is
my
son
and
he's
been
finding
little
things
in
the
docs
and
knitting
patches
and
fixing
him,
which
is
exactly
the
we
need
all
right.
So
I
just
did
that.
A
So
I've
been
talking
to
a
couple
of
people
about
what
we
want
to
do
sort
of
to
show
show
off
the
capabilities
of
and
crowbar
because
we
completed
the
release,
we
got
things
going.
We
did
a
whole
bunch
of
lab
testing
on
it
and
we
haven't
sort
of
gotten
a
in
the
wild
sizable
deployment
done,
we
have
done
is
fallen,
but
so
what
I'd
like
to
do
talk
to
a
couple
people
about
this
is
to
actually
find
a.
B
A
That
is
interested
in
having
a
hundred
node
has
a
hundred
notes
to
do
a
deployment,
and
several
of
us
are
willing
to
take
vacation
time,
show
up
on
site
and
get
crowbar
to
do
100
no
deployment
as
part
of
the
system
right
now.
If
somebody
wanted
to
do
it,
we
wouldn't
have
written
bias
capabilities.
Our
preference
would
be
to
say.
A
The
system's
look
like
what
are
what's
the
configuration
and
what's
the
hardware
manufacturer
and
I'm
getting
that
ready,
because
we
can
discuss
as
a
group
what
that
is,
but
then
actually
do
100
no
deployment
and
and
get
that
sort
of.
You
know
that
not
checked
off
on
our
belt
I
think
we
could
do
you
know
a
smaller
deployment
would
be
an
interesting
thing
also,
but
it's
much
more.
I
think
the
hundred
node
milestone
considering
crowbar
ones
limitations.
B
A
A
A
D
E
E
D
E
Though
well,
when
you
can
make
it
up
work
first,
totally.
A
But
I,
if
somebody
comes
to
us
with
dell
gear,
then
we
need
to
make
sure
it's
here.
Somebody
sticks,
supermicro
or
open
computer,
so
know
that
we
need
to.
You
know:
we
need
to
know
what
that
is
and
be
able
to
skip
so
I'm,
assuming
if
raiden
BIOS
configuration
is
in
the
delivery,
then
we're
going
to
have
that
time.
Adjustments
as
a
consequence
yeah,
but
the.