►
From YouTube: OpenCrowbar Community Meeting 2014-07-30
Description
Review backlog and defect queue
Discuss capabilities for IPMI interfaces and Chef Metal integration
http://pad.opencrowbar.org/p/community-design-meeting-2014-07-30
A
A
A
All
right,
ok,
so
you
get
a
conversation
with
that.
Any
other
items
people
want
to
want
to
add
to
the
agenda
before
we
get
go
all
right.
Those
of
you
who
don't
know
me
when
science
comes
I
get
would
take
whatever
in
iterators
that
I
want.
In
that
case,
it
was
no
no
other
things.
Alright
cool,
so
I
created
a
new
new
etherpad
and
updated
the
last
meeting
to
actually
link
appropriately
to
this
one
so
going
through
the
things
that
we
were
planning
on
planning
on.
A
A
I
would
like
to
go
through
the
github
that
gives
me
more
structure.
I
haven't
been
really
plugged
in
right
now.
I
have
the
time
to
get
really
plugged
in
they're
going
through.
Github
is
a
good
thing
for
me.
Okay,
I
wish
there
was
a
time
stamp
on
this.
Let's
take
a
look
at
them
really
quickly.
So,
let's
see,
rearrange
the
bootstrap
recipe
and
start
editing
sent
to
f7
supports.
This
is
one
of
the
things
that
was
closed.
Victor.
You
were
working
on
this.
A
Yes,
that
rearranged
with
Shep
recipe
probably
caused
some
of
the
issues
that
people
are
now
talk
about
on
the
mailing
list,
with
the
create
repo
stuff.
So
I
need
to
revisit
some
of
that
and
start
adding
sinto.
A
some
support.
Pretty
much
just
adds
a
simple
at
seven
thingy
in
dr.
admin
and
hang
out
context.
Is
there
anything
in
particular
around
sent
OS
that
we
need
to
need
to
handle
I
could
notice
on
the
CentOS
7
stuff.
Some
of
the
some
of
the
networking
pieces
seem
to
be
a
little
different.
A
Did
you
not
yeah?
That
was
it?
I
haven't
gotten
that
far.
Yet
this
is
a
support
for
CentOS
7
of
the
client,
yet
yeah,
okay,
yeah
I'm,
did
bring
up
a
scent
to
Wes
image
and
done
through
it
a
little
bit
and
I
noticed
a
lot
of
the
standard
network.
Modification
tools
are
not
available
in
the
base
image.
You
actually
need
to
install
a
series
of
backported
packages,
but
we'd
have
to
take
a
look
at
that
a
little
bit
more
okay,
yeah,
that's
one
of
the
things
that
we
talked
about.
A
They
like
IP,
like
field.
If
config
and
route
tools,
those
are
deprecated
and
they
don't
think
it
installed
by
default
anymore.
Yes,
that
was
really
just
can
pick
yeah.
It's
really
annoying
ya
know
knows
I
see
at
her.
She
use
IP
and
love
it
yeah
love
you
get
used
to
the
old
new
way,
yeah,
it's
a
only
been
in
there
for
like
10
years.
Yes,
no
clears
it's
probably
more
like
15,
but
ok,
yeah,
that's
the
one
to
trip
me
across
and
then
I
started.
A
Looking
for,
you
can
actually
add
the
old
packages
and
make
it
work
again,
but
the
IP
command
does
seem
to
handle
them
alright
cool.
So
we
also
have
you
know
Travis,
which
is
one
of
the
one
of
the
items
that
we're
all
supposed
to
go
for
us
about
four
last
sprint,
so
Victor
you
had
opened
this
I
know
that
Rob
was
working
on
it,
but
I
don't
know
if
you
know
where,
where
things
actually
stand
with
the
Travis
CI
system,
I
believe
Travis.
These
days
is
actually
doing
its
thing.
A
If
you
just
go
back
out
of
issues
and
up
to
core,
you
should
be
able
to,
you
know:
click
on
a
random,
open,
pull
request
and
see
whether
or
not
Travis's
pasture
build
it.
It
should
actually
be
going
through
pretty
much
a
complete
BDD
run
at
this
point.
Ok
cool
and
so
that's
going
through
the
the
BDD
test
and
you
do
need
to
add
a
Travis
config
file.
Or
did
we
add
that
into
the
into
the
into
the
base
code
tree
so
that
people
pull
it?
That's
all
done
as
far
as
I.
A
As
far
as
I
mean
we
can
go,
take
a
look
at
it
that
should
all
done
and
if
every
pull
request
that
comes
in
should
be
running
through
travesty
I.
All
right.
That's
that's
an
awesome
milestone
you
know
and,
as
so
now
run
through,
the
BDD
test
will
continue
to
expand
those
and
the
next
thing
on
the
testing
pieces.
That
we'd
want
to
take
a
look
at
and
start
seeing
what
we'll
be
doing
for
the
integration
tests
or
any
modifications
we
want
to
make
to
Travis.
A
A
A
Construct
the
gym,
failing,
that's
another
bug,
fix
yeah
I,
don't
know
how
far
down
we
want
to
go
through
the
through
the
closed
issues,
we're
already
back
to
the
last
meeting
at
theft.
At
this
point,
ok,
Pro
kind
of
a
cool
right.
Well
that
makes
sense
then,
and
then
there's
a
you
know.
We
have
15
open
ones
that
are
currently
in
here,
so
we'll
continue
moving
those
attracting
them.
Are
there
any
anyone's
in
here
that
we
particularly
want
to
chat
about?
A
If
what,
if
robber
here
he'd
want
to
chat
with
off
a
lot
about
the
readystate
stuff,
I
haven't
had
a
chance
to
look
at
it
personally,
yet
the
rate
estate
Wizards
has
anyone
else
on
the
meeting
had
a
chance
to
poke
through
that
wizard.
I
believe
it
was
split
test
had
to
disable
it
yeah
I
know
just
the
stuff
I'd
been
working
on,
which
is
to
be
added
increasingly
horribly,
maimed,
available.
A
Note
managers
models
to
let
to
be
able
to
do
some
discovery
and
driving
of
the
node
manager
stuff
via
the
API
okay,
and
is
that
one
of
these
one
of
these
ones
here
when
these
open
the
open
cases
get
start
from
the
top
they're
can
stop
this
one?
Okay
and
there
it
was.
You
think,
I'd
be
dead,
okay,
alright
cool!
So
how
much
longer
do
you
think
this
one
here
goes?
Do
you
have
what
you
need
to
find
in
order
to
do
it
or
you
need
us
to
define
it
more
or
I'm.
A
A
A
When
my
brain
wasn't
quite
caffeinated
enough,
we
could
call
it
the
Oracle
available,
node
manager
for
the
sun
java
systems
available,
node
manager,
yeah.
I
know
it's
the
idea.
The
names
are
getting
to
that
territory
and
I
yeah
I
want
something
short
indistinct
as
I
keeps
leaving
call
it
terraformed,
that's
a
popular
project,
name
on
get
up
there
for
my
own.
There
we
go.
A
Thank
you
know,
because
that's
what
we
want
to
do:
terraform
Mars
for
requests
to
find
a
better
name.
Yeah
stole
my
look
slippy
much
there.
Alright
cool,
alright,
so
I
see
that
think
they'd
be
like
bigger
claw
hammer.
Something
like
that.
Ok,
so
there's
additional
other
than
the
practice
of
naming
is
not
great.
What's
what's
left
to
be
done
with
that,
and
are
you
planning
on
working
on
that
display
I'm
planning
on
working
on
it,
hopefully
getting
it
merged
the
sprint?
A
A
A
A
A
A
Right
0,
so
nothing
more
advanced
than
that
no
multirate
functions
I
mean
as
we're
as
we're
talking
through
this
right.
The
actual
raid
number
is
a
parameter:
that's
just
passed
into
the
into
the
Iraq,
so
you
could
pick
rate
five
as
well
as
raid
0,
or
there
are
specific
definitions
that
need
to
be
done,
for
which
disks
are
part
of
the
part
of
the
raid
array,
the
ladder
the
ladder?
A
Okay,
so
right
now,
you're
just
going
to
handle
raw,
which
means
you
don't
necessarily
need
to
pick
disks
and
a
raid
0,
which
is
just
a
strike
well
yeah
I'm,
either
going
to
either
set
all
the
disks
to
non
rage
and
if
POS
pics,
if
the
controller
supports
that
and
will
allow
the
Australis
position-
or
you
know,
take
the
first
disc
and
make
it
a
single
dis
grade
0.
If
you
can't
do
that,
okay,
yeah,
that's
what
I'm
it's
what
we
had
in
the
last
sprint.
A
A
A
Alrighty,
let's
see
any
other
open
issues
that
we
want
to
talk
about
before
we
go
through
the
list
items
that
we
had
from
the
meeting
two
weeks
ago,
I
had
started
work
on
a
gold
base,
crowbar
client
and
I
intend
to
take
that
up
again.
This
weekend
it
was
supposed
to
be
a
sort
of
a
demo
API
learning
tool
for
for
newbies
could
crowbar
and
do
a
little
video
to
or
some
other
kind
of
instructional
to
to
get
folks
up
and
running
with
it.
A
A
Well,
I
guess:
are
you
going
to
be
in
checking
code
into
the
base
to
handle
it,
or
is
this
something
that
you
don't
tend
to
do
an
outside
of
it?
Oh,
so
it's
not
I
would
my
suggestion
would
be
to
maybe
floated
on
the
list.
Yeah
I'm,
just
so
the
people
that
are
aware
of
it.
I'm
not
sure
that
an
issue
necessarily
needs
to
be
created.
Okay,
I
guess
the
code
bases
people
might
misrepresent
that
and
say
you
know
this
will
be
supporting,
go
or
something
to
that
effect.
A
A
Okay,
so
there's
the
disc
migration,
so
the
chef
metal
stuff.
This
is
something
that
Rob
kicked
out
to
the
to
the
community
and
it
can
bring
that
up
really
quick,
the
chef
metal
stories,
technical
details-
and
you
know
what
he's
trying
to
what
he
was
trying
to
accomplish
with
that
anybody
have
any
comments,
questions
or
anything
they
want
to
talk
about
it
around
the
chef
metal
story.
A
A
A
Ok
cool
so
apparently
that
so
there
weren't
pull
request
that
went
in
for
that
code.
I
haven't
been
following
the
the
github
tree,
I
think
I
had,
but
I
can
pull
pull
request
to.
Listen.
Ok,
alright,
how
functional
it
is.
We
don't
know
we'll
need
to
hear
from
Rob
on
that,
since
nobody
Charlie,
therefore,
but
I
thought
they
were
pull
request
out
there.
Only
little
trip,
ok,
alright.
So
the
last
thing
that
we
had
on
the
list
from
two
weeks
ago
was
to
create
a.
A
A
I
switched
up
the
car
p
and
build
process
to
make
it
a
little
bit
simpler
and
not
work
and
not
rely
on
our
somewhat
crazy
Ruby
build
process,
I
ripped
it
out
and
made
about
a
lot
simpler
and
changing
locations
where
the
RPMs
are
at
okay.
The
install
of
the
install
guides
been
updated
with
that
information.
Oh
excellent.
Thank
you.
You
just
pop
it
up
here.
A
Okay,
so
that
sounds
good
so
and
I
know
a
couple.
People
played
with
a
few
people
submitted
bugs
against
it
for
some
issues
with
with
his
version,
a
ruby
is
that
what
I
recall
yeah
so
I
bumped,
the
version
of
Ruby
to
Ruby
student,
one
and
I
ripped
out
the
I
ripped
process
that
we
were
using
for
building
Ruby.
A
It
was
trying
to
build
Ruby
packages
that
looked
that
were
structured
and
have
the
same
package
subdivisions
that
the
as
the
RPMs
that
the
Red
Hat
themselves
uses
so
like
it
was
splitting
out
rubygems
and
all
the
bacon
rubygems
into
their
own
packages.
But
for
what
we're
doing
we
don't
need
or
care
about
that.
So
right
now,
I
put
to
a
much
simpler,
Ruby
build
process
that
is
only
building
a
single
package
that
contains
all
of
Ruby,
including
the
baked
in
gyms,
that
we
need.
A
Thank
you
so
another
thing
that
came
out
through
the
list,
Tom
that
we
discussed
the
last
time
we
had
this
meeting
is
some
way
to
lower
the
impedance
for
people
to
actually
try
out
the
the
OPA
Kirko
dish,
and
one
of
the
things
we
decided
to
try
to
do
was
build
a
virtual
machine
that
someone
could
import
that
basically
handled
a
lot
of
the
base
setup
function,
so
that
someone
could
run
through
the
installation
process
with
Widow
crowbar.
So,
according
to
mr.
A
hershfeld,
which
is
this
one
here,
who
did
that
work,
he
put
together
everyone
and
it
is
now
stored
in
sourceforge,
so
people
can
download
it.
I
don't
know
if
we
actually
went
through
and
updated
any
links
or
anything
or
effectively
exposed
it.
People
who
learn
on
this
call
or
the
mailing
list,
but
if
we
decide
that
this
is
this
is
something
we're
going
to
continue
to
maintain.
You
need
to
work
with
people
find
it
useful
arm.
A
And
there's
a
series
of
tasks
that
you
need
to
do,
which
are
just
environmental,
mostly
around
the
networking
posts,
and
then
they
do
the
get
pull
to
get
the
latest
code
and
make
sure
you
dr.
bridge
and
theoretically,
that
will
give
you
what
you
what
you
need.
A
sledgehammer
is
not
included
this
in
the
conversation
that
I
had
with
with
Rob
about
it,
but
it
will
be
pulled
down
with
you
actually
do
the
install
the
other
thing
that
I've
been
thinking
about.
A
After
looking
at
all
the
hash
ecor
stuff
as
I
finally
notice,
the
existence
of
packer
petition
have
noticed
a
long
time
ago,
which
is
a
tool
they
can
actually
do.
Parallel,
builds
of
multiple
different
disk
images
for
different
for
different
kind
of
vm,
so
it
can
do
like
qemu
kbm
type
stuff.
They
can
also
do
virtualbox,
VMware
task,
even
amazon
cloud
instances.
I
was
thinking
of
trying
to
use
that
to
automate
a
pre
bootstrapped
vm,
which
would
paint
sledgehammer
and
all
the
necessary
stuff
to
actually
deploy
things.
A
Excellent,
all
right,
good,
all
right,
so
that
covers
what
we
planned
on
doing
so
I
have
what
is
kind
of
on
the
docket
for
next
sprint,
as
we
were
talking
around
these
pieces,
are
there
any
other
anything
in
there?
That
needs
to
be
changed
or
anything
else
that
we
want
to
want
to
try
to
tackle
on
the
next
week.
A
A
Look
at
it.
You
know
the
other
thing
that
I'd
been
thinking
of
doing
probably
not
next
sprint,
but
just
as
a
general
background
task
is
a
long
time
ago
we
ripped
out
support
for
nagios
and
ganglia
or
monitor
and
I
want
to
replace
that
with
using
I'm
playing
around
with
a
tool
called
another
tool,
compassion
for
call
compal
and
use
that
for
monitoring
and
possibly
more
awesome
stuff
further
down.
The
road.
A
A
Okay
and
what
type
of
benefits
does
that
give
us
above
not
give
the
ganglia
which
kinda
like
the
de
facto
small
deployment
standards
right
I
know
they
have
scalability
issues
and
a
whole
bunch
of
other
stuff
console
will
scale
almost
indefinitely
and
in
addition
to
doing
monitoring,
it
can
also
be
used
if
we
need
to
start
doing
services
that
are
themselves
clustered
and
we
need
to
start
handling
things
like
cluster
failover.
It
has
some
pretty
nice
stuff
baked
in
for
doing
that.
A
And
it
has
faked
it
service
monitoring,
which
we
confuse
to
monitor
just
about
anything
on
boxes,
in
a
way
that
scales
horizontally
almost
forever.
Oh,
ok,
then
we'd
be
able
to
integrate
it
with
the
with
the
deployment
path.
In
other
words,
it
has
easy
hooks
for
us
to
be
able
to
add
new
services
for
monitoring
and
new
monitoring
agents.
Very
nice,
much
less
of
a
pain
in
the
ass,
the
nagios,
oh
about
20
years,
less
of
a
pain
in
the
ass
yeah
yeah.
This
is
this
actually
a
number
of
those
things
that
we
could
use.
A
Just
so
you
know
anybody
who
remembers
the
history,
the
the
inclusion
of
nagios,
inc
and
leah
was
done
simply
because
it
was
very
easy
for
the
integration
arm,
but
it
turned
out
to
be
a
you
know:
pretty
important
and
differentiated
creature
for
people
who
are
evaluating
the
product
to
have
a
you
know,
a
basic
monitoring
system
built
into
the
into
the
deployment
engine,
mostly
because
it
actually
controlled
the
monitoring.
So
they
were
another
series
of
steps
that
needed
to
be
done
afterwards
to
to
ensure
that
you
were
monitoring
your
environment
correctly.
A
A
Can
the
nice
thing
about
console
is
like
all
the
other
half
orc
tools,
it's
written
in
go
compiles
to
a
single
binary
that
you
can
just
SCP
onto
a
system
and
it
will
work-
and
it's
almost
most
popular
thing
for
that
seriously.
It's
like
it's!
If
you
get
the
x86
64
binary
0
to
work
on,
basically
all
like
City
Hall,
the
linux
distributions
that
have
run
on
an
x86
64
platform.
Long
as
they're
kernel
is
a
not
positively
ancient
cool
yeah
go
ahead.
That
sounds
great,
no
puzzles
so
much
yeah.
A
Well
as
we're
in
the
in
the
weeds.
Here,
you
know
it's
a
good
time
to
add
these
sort
of
things.
I
think
Cynthia,
you
know
the
the
monitoring
stuff
again
was
somewhat
under
appreciated,
but
been
doing
a
lot
of
conversations
with
end
users
and
that
sort
of
stuff
explaining
how
that
worked.
Actually,
you
know
added
a
lot
to
the
project
yeah
you
know
explaining
to
the
customer
location
and
then
this
is
what
happens
so
they
just
loved
it.
You
know
part
of
the
reason
why
we
didn't
you
know
didn't
truly
expand.
A
It
is
once
you
start
getting
to
a
larger
cluster
deployment.
Typically,
your
user
is
going
to
want
to
integrate
with
you
know,
with
their
existing
monitoring
systems
that
they're
using
are
ready
to
monitor
their.
You
know
their
data
center
and
the
rest
of
their
functions
rather
than
incorporating
a
new
tool.
A
If
it's
a
real
pain
in
the
butt
for
data
center
managers
or
it's
a
lot
different.
They
are
testing
and
begging
for
for
things
like
console
and
hate
nagios,
and
they
haven't
already
ripped
it
out
and
put
something
like
sensory
or
console.
And
then
then
this
will
just
be
a
dream
for
them:
yeah
yeah!
That's
where
worthy!
Where
the
enterprise
is
just
like
well,
okay,
give
me
anything.
You
know
shaving
to
meet
ideas.
A
Oh
yeah
I
usually
have
found
that
you
know
that
there
was
you
know.
Anything
was
good
for
pilot,
but
once
you
got
to,
you
know
do
something
that
they
were
actually
going
to
roll
through,
like
their
help
desk
in
their
their
operation
center.
Then
it
needed
to
go
through
the
same
tools
that
they
were
using
for
everything
else,
but
but
in
any
case,
I
think
it's
very
valuable.
It
would
be
spectacular
to
you
know
to
chump
that
in
there,
even
if
it
isn't
fully
integrated
but
avid
as
part
of
peaches.
A
The
other
thing
that's
nice
is:
it
adds
a
series
of
very
useful
statistics
for
people
to
four
people
through
to
handle,
so
it
one
last
thing
and
then
I'll
get
up
off
of
this
topic.
This
is
effectively
going
to
handle
the
roles
of
both
nadia's
and
gangly
us,
or
both
monitoring
and
alerting
from
no
one
handle
alerting
at
all.
A
Okay,
alerting
still
something
else
like
that.
It's
like
there's
nothing
baked
into
console
build.
That
has
any
no
notion
of
like
how
to
email
an
administrator,
or
you
know
someone's
page
or
make
a
thing.
You
turn
red
on
your
monitoring
board
or
make
a
thing.
You
turn
red,
that's
where
I
think
we
should
also
consider
bringing
in
sensu.
A
A
Cool
all
right,
so
we
can
see
about
see
about
that
as
well,
but
console
first
and
then
we'll
look
at
the
alerting
functions
afterwards,
all
right.
Okay,
so
we
have
20
minutes
left
in
the
regular
scheduled
meeting.
Are
there
any
other
items
that
we
wanted
to
cover?
I
think
we
have
a
pretty
good
list
of
stuff
to
tackle
a
sprint.
It
was
actually
pretty
good.
Last
couple
of
spreads
I
wish
rob
was
here
to
talk
about
the
about
the
the
wizard
piece,
but
we'll
catch
that
next
week.
A
A
Hopefully
next
next
meeting
will
make
some
decisions
around
you
know
console
seems
fine,
I,
don't
scare
anything
but
we'll
see
what's
in
there,
and
hopefully
people
can
spend
a
little
bit
of
time
I'm,
taking
a
look
at
the
better
chef
metal
and
the
vm
that
was
created
to
make
sure
that
those
things
are
kosher
and
I'm
good
to
go.
Yeah
I'll
definitely
be
trying
them
out.
A
Alright
sweet
well
with
that,
unless
anybody
has
anything
else,
I'll
shut
down
the
meeting
and
we'll
all
be
able
to
go,
get
a
cup
of
coffee
or
something
before
we
do
something
else.
Their
status.
On
the
other
item,
things
Oh
to
talk
about
yeah
I
just
wanted
to
talk
about
the
terraform.
Oh
I'm,
sorry
I,
skipped
over.
Oh
yeah
go
for
it
please!
A
Well,
if
you
go
to
terraform
that
I,
oh,
they
have
created
what
looks
like
an
infrastructure
abstraction
system
and
and
they're,
claiming
that
they
have
their
metal
support,
not
sure
what
it
is
we
go
to
the
intro.
They
don't
have
bare
metal
support,
yeah
they're,
claiming
it
and
I
want
to
I
just
want
to
try
to
get
a
handle
on
where
terraform
begins
and
ends
partially
in
comparison
to
crowbar
mmm-hmm.
A
A
A
A
The
that's
the
very
high-level
executive
summary
for
people
here.
What
keeps
improper,
yeah
and
here's
the
fun
there's
what
it's
like?
Ok
cool,
so
I'll
take
the
action
to
go
through
that,
if
you
guys
want
and
I
at
the
next
meeting,
we
talk
about
it
since
you,
you
guys,
picked
up
a
series
of
other
things,
I'm
also
going
to
plan
on.
You
know
downloading
the
vm
playing
with
the
vm
a
little
bit
making
sure
that
that
effectively
works
the
way.
A
Someone
who
necessarily
doesn't
know
how
to
set
up
a
reasonable
Devon
environment
like
myself,
can
I
get
a
system
up
and
running
I
just
haven't
spent
the
time
to
figure
out
puppet
and
all
the
other
stuff
I.
Usually,
when
I
test
the
system,
I
try
to
test
the
system
as
a
as
an
user
as
opposed
to
developer,
so
pretty
familiar
with
doc
p.m.
install
and
that
sort
of
stuff,
but
but
in
any
case
so
I
also
try
to
give
that
a
crack.
This
week
as
well.
A
Right
where
there
were
there
any
other
items
that
we
wanted
to
wanted
to
go
over,
I
think
that
hit
the
other
items
advantage
of
merch
FL,
alright,
uhhhhh,
alright,
with
that
I'll
shut
it
down
guys
appreciate
the
running
every
time,
yeah
problem
at
all.
Now
that
I
have
my
I
had
access
to
all
of
Rob's
random
tools.
He
tends
to
try
to
use
ten
or
eleven
eighty
things
to
run.