►
Description
Want to view more sessions and keep the conversations going? Join us for KubeCon + CloudNativeCon North America in Seattle, December 11 - 13, 2018 (http://bit.ly/KCCNCNA18) or in Shanghai, November 14-15 (http://bit.ly/kccncchina18).
OpenStack SIG Deep Dive – Chris Hoge, OpenStack Foundation & David Lyle, Intel (Intermediate Skill Level)
SIG OpenStack coordinates the cross-community efforts of the OpenStack and Kubernetes communities. This includes coordinating improvements to and documentation of the OpenStack cloud provider implementation in Kubernetes as well as supporting efforts to deploy OpenStack itself using Kubernetes. Attend this session to collaborate on the SIG's future plans. To review the agenda for the session, or to propose an item of your own for discussion, refer to the etherpad at https://etherpad.openstack.org/p/CPH-kubernetes-sig-openstack-deep-dive
About Chris Hoge
Chris Hoge is a Strategic Program Manager for the OpenStack Foundation. He's been an active contributor to the Interop Working Group (formerly DefCore), and helps run the trademark program for the OpenStack Foundation. He also works on collaborations between the OpenStack and Kubernetes communities. Previously he worked as an OpenStack community manager and developer at Puppet Labs, and operated a research cloud for the College of Arts and Sciences at The University of Oregon. When not cloud computing, he enjoys long-distance running, dancing, and throwing a ball for his Border Collie.
About David Lyle
David Lyle works on cloud technologies as a Cloud Software Architect in Intel's Open Source Technology Center. David is primarily focused on Kubernetes and OpenStack.
Join us for KubeCon + CloudNativeCon in Barcelona May 20 - 23, Shanghai June 24 - 26, and San Diego November 18 - 21! Learn more at https://kubecon.io. The conference features presentations from developers and end users of Kubernetes, Prometheus, Envoy and all of the other CNCF-hosted projects.
A
Thanks
everybody
for
coming
for
this
for
coming
to
the
cig,
OpenStack
deep
dive,
typically
in
in
in
the
past,
the
way
that
we
have
we
run
two
sessions.
The
first
one
has
been
kind
of
an
intro
to
the
working
group
that
was
two
days
ago,
and
that
was
more
of
a
formal
presentation
of
the
work
that
we've
done
and
the
work
that
we're
planning
on
doing
and
just
kind
of
an
introduction
for
folks
who
may
not
have
you
know
you
know
have
have
worked
with
the
open
with
it
with
cig
OpenStack.
A
Before
for
the
deep
dives,
we
tend
to
run
these
like
we
do
one
of
our
development
sessions
or
one
of
our
forum
sessions
where
it's
the
the
content
is
mostly
driven
by
what
folks
in
the
audience
would
like
to
talk
about
the
issues
that
they're
facing
the
things
that
we
are.
You
know
currently
working
on
and
and
and
kind
of
building
a
set
of
issues
that
we're
going
to
be
taking
forward
and
addressing
over
the
next
several
months,
so
I
think
so
so
we
typically
work
from
an
ether
pad
which
is
a
collaborative
document.
A
It's
it's
similar
to
Google
Docs
but
simpler
and
the
URL,
for
it
is
right
here
and
once
pardon,
that
is
a
small
K
and
converted
eg
add
is
case-sensitive.
The
URLs
and
the
ether
pad
I
can
erase
that
and
make
it
a
small
K
yeah,
and
you
know
that's
and
that's
where
you'll
be
able
to
find
the
agenda.
You'll
be
able
to
take
notes,
you'll
be
able
to
add
to
it,
and
it
will
be
a
working
document
for
the
for
the
coming
months.
Over
for
back
to
you.
A
A
Openstack
is
one
of
these:
the
OpenStack
provider,
the
AWS
provider,
the
GCP
provider,
but
a
cloud
controller
manager,
which
is
an
interface
between
kubernetes
and
the
cloud
provider
has
been
built
out
and
there
and
all
of
the
providers
are
expected
to
re-implement
the
features
of
their
upstream
providers
as
an
external
cloud
provider,
and
so
within
the
width
and
and
and
all
the
class
and
all
the
providers
are
working
on
this
right
now
so
OpenStack
is,
is
one
of
the
is
one
of
the
providers.
That's
pretty
far
ahead
of
the
game.
A
We
have
a
repository
in
a
cloud
provider
OpenStack
within
hosted
within
the
kubernetes
repository,
and
we
have.
We
have
end-to-end
testing
in
place,
we're
working
on
Doc's.
We
have
a
number
of
different
drivers,
for
we
have.
We
have.
We
have
a
flex
and
a
CSI
driver
for
cinder
in
there.
We
also
have
a
keystone
authentication
and
authorization
driver
will
probably
be
adding
a
Manilla
driver,
a
CSI
Manila
driver
in
there
too,
so
that
will
be
expanding
our
storage
options
for
those
people
who
want
a
synchronous
storage.
A
A
A
A
A
A
B
A
Another
another
work
item
that
that
were
pretty
heavily
involved
with
right
now
and
and
unfortunately,
neither
DIMMs
nor
Melvin
were
able
to
make
it
to
Kubek
on
this.
This
time
around,
but
they've
done
a
tremendous
amount
of
work
in
building
out
the
test
coverage
of
of
the
provider
and
also
the
the
various
drivers
right
now
we're
there's
active
work
on
having
having
the
test
run
for
the
for
the
CSI
driver
for
Synder,
and
this
is
being
done
in
open
lab.
A
So
if
you
aren't
familiar
familiar
with
an
open
lab
is
a
is
an
independent
organization,
that's
devoted
to
testing
free,
open
source
projects,
particularly
those
related
to
cloud,
and
so
that
covers
you
know,
OpenStack
projects
and
kubernetes,
and
so
right
now,
they're
they're
running
testing
for
things
like
terraform
and
go
for
cloud.
That's
where
the
CI
CD
is
happening
for
those,
but
also
for
the
the
signalmen
stack
is
also
taking
advantage
of
their
resources
for
testing
and
and
so
there's
work
that
we
need
to
do
on
that.
A
A
For
those
you
who
have
been
attending
or
who
aren't
attending
in
our,
how
many
would
be
interested
in
coming
to
this
sig
openstack
meetings
and
practicing
the
other
their
typical
zoom
meetings.
That
are,
you
know
like
the
community
meeting
here.
So
what
are
the
things
that
we
would
need?
We
would
need
to
cover
in
those
meetings
to
make
them
more
useful
for
the
participants,
because
I
know
that
for
us
it's
a
good
way
for
us
to
catch
up
and
get
a
sense
of
the
things
that
we
need
to
accomplish
over
the
next
few
weeks.
A
A
Okay,
so
so
one
suggestion
is
you:
can
it
can't
attend?
Because
it's
in
the
middle
of
the
night,
so
in
the
so
in
the
etherpad?
Would
it
be
possible
for
maybe
people
to
put
times
that
you
know
maybe
a
range
of
times?
You
know
either
in
UTC
or
with
your
time
zone
attached,
so
that
we
can
reevaluate
the
meeting
time?
Because
it's
pretty
it's
fairly
easy
to
change.
It's
just
a
pull
request
to
the
community
repository
and
because
we
would
be
happy
to
change
the
meeting
time
so
that
everyone
else
can
attend.
A
And
it's
it's
it's
difficult,
because
we
have
a
bunch
of
contributors
from
the
asia-pacific
region.
We
also
have
contributors
from
Europe,
so
there
may
need
to
be
a
tick-tock
thing
where
we
have
an
early
meeting
once
a
month
and
a
late
meeting
once
a
month,
so
that
we
can
get
a
little
more
coverage
across
all
of
the
time
zones.
B
Or
I
mean
it's
valid,
just
say
to
that:
perhaps
meetings
aren't
valuable
and
we
would
rather
resolve
things
with
issues
and
I
think
it's
time.
It's
nice
to
have
time
to
actually
discuss
things,
but
we
could
also
do
it
on
an
on
a
need
basis.
If
people
are
not
really
interested
in
tending
it,
we
won't.
We
don't
need
to
go
through
the
motions
just
to
go
through
the
motions,
but
if
it's
valuable,
we're
would
love
to
keep
them
going.
Yeah.
A
A
A
B
A
A
Typically,
the
way
the
the
community
has
looked
at
scaling
based
off
of
load
or
other
conditions
is
to
have
some
sort
of
time-series
database
installed
with
a
metrics
api,
but
traditionally
an
OpenStack
there.this
lammeter
project
has
has
kind
of
broken
up
into
smaller
projects.
It's
not
widely
deployed,
and-
and
so
there's
so
I-
think
that
if
you,
if
you
build
an
auto
scaling
component
into
the
OpenStack
cloud
provider,
I,
don't
think
that
you
can
necessarily
rely
on
something
like
heat,
auto
scaling.
You
know
because
he
may
or
may
not
be
available.
A
Ceilometer
may
or
may
not
be
available
or
any
of
the
other
tooling.
For
that,
so
does
anybody
have
any
experience?
And
this
or
do
you
have
to
have
people
come
up
with
their
own
solutions
for
auto
scaling
and
OpenStack
or
auto
scaling?
Where
there's
not
a
direct
metrics
API
that
they
can
query
to
to
use
as
a
trigger
for
this?
Because
if
there's
a
general
solution
and
and
it's
it,
it's
a
solution
that
we
could,
you
know
implement
as
part
of
the
provider
but
is,
is
more
generic
and
not
necess,
certainly
dependent
upon.
B
Because
and
I
mean
in
OpenStack
there
there's
at
least
there's
cynllun
that
wanting
to
do
as
auto-scaling
there's
there's
heat
and
then,
but
we
can't
guarantee
any
of
those
are
gonna
be
there.
So
we
want
to
try
and
build
a
more
bare-bones
approach
and
it
may
come
down
to
polling
at
some
point
to
get
enough
information,
but
that's
not
a
very
attractive
solution.
B
B
C
B
C
B
C
A
Yes,
yes,
so
and
I
don't
I
mean
I'm,
not
I'm,
not
in
time
and
I'm,
not
fully
familiar
with
how
the
auto-scaling
works
within
the
CCM
itself.
You
know
there
are.
There
are
interface,
you
know,
they're
they're
defined
interfaces
for
this.
The
interfaces
that
we've
implemented
are
the
load,
balancer
and
storage
interfaces.
B
C
A
A
B
A
A
A
Issues
a
four
cinder:
okay,
it's
volume,
attachments,
yeah.
How
many
people
are
you
just
using
cinder
within
the
yeah?
People
are
using
it
or
just
using
are
just
using
cinder.
Yeah
I
mean
yeah
load.
Balancing
is
what
is
one
of
those
things
that
I
don't
know
if
there's
anything
that
we
can
do
about
that
in
just
new
Tron
L
baths,
v2
has
been
deprecated,
so
it's
not
recommended
to
be
used
anymore
and
it'll
will
be
probably
depending
strictly
on
activity.
Octavia,
but
Octavia
is
not
widely
deployed
and
it
can
be
fairly
resource
intensive.
A
You
know
and
I
know
that
for
load
balancing,
if
I,
if
I
were
to
use
it
I
would
I
would
probably
just
grab
one
of
the
like
the
engine.
X
load,
balancers,
load,
balancers
and
just
add
that,
but
I
don't
know
if
anyone
hasn't
has
anything
to
add
about
the
load
balancing
in
the
provider
if
they
have
any
experiences
with
it.
C
A
I
mean
but
I
mean
Octavia.
Do
you
have
jet
plans
to
talk
David
I,
yeah,
yeah,
yeah,
yeah,
yeah
yeah?
It
has
higher
requirements
because
wants
Barbic
hand
for
four
keys
and
it
and
it
and
it
brings
up
an
entire
VM,
and
so
it's
not
using
you
know,
what's
a
whole
VM
do
to
your
load
balancing.
So
you
know
there
are
lighter
weight
ways
to
do
it,
but
you
know
so.
C
B
There's
been
I
mean
this
is
something
I'm
interested
in
as
a
company
built
accelerators,
obviously,
but
there's
also
been
from
a
few
of
the
drive-bys
in
the
meetings
interest
in
being
able
to
schedule
based
on
GPU
and
supporting
that
kind
of
scheduling
pass
through
and
just
trying
to
determine,
if
that's
a
wider
interest
or
if
that's
a
pretty
specific
interest,
at
least
one
not
in
the
room.
Okay,
that's
probably
a
little
bit
further
down
the
list
as
far
as
things
that
I
can
tackle,
but
I'm
you
know
yeah.
A
There
isn't
there
is
an
openstack
project,
that's
devoted
to
providing
GPUs
as
a
compute
resource
called
cyborg,
it's
still
in
early
stages
and
I'm,
not
sure
how
that
would
integrate.
But
if
it's
something
you're
interested
in
OpenStack
in
general,
that
might
be
a
good
place
to
start
looking
and
then
possibly
seeing
how
that
could
be
integrated
in
in
inside
of
this.
A
A
C
C
A
C
C
A
A
A
B
A
B
C
A
B
A
Know
what
go
for
cloud?
Is
it
it's
a
it's
a
it's
a
go
SDK
to
interact
with
with
OpenStack
clouds
it
was.
It
was
developed
at
Rackspace.
It
was
originally
Rackspace
product
project.
You
can
still
find
the
old
gopher
cloud
at
the
Rackspace
gate
repository,
but
there
is
a
it's
in.
Is
it
it's?
This
go
for
cloud
and
it
go
for
cloud
namespace.
No.
A
A
A
A
A
Yes,
yeah
there
are
so
within
the
within
the
within
the
cloud
provider
repository.
There
are
two
drivers
for
cinder,
one
is
a
flex
driver
and
the
other
is
a
CSI
driver.
Both
both
are
maintained.
Both
both
are
both
came
out
of
experimental
work
that
was
being
done
in
six
storage,
and
this
is
this
is
kind
of
the
new
home
for
them,
and
actually
some
of
the
some
of
the
original
authors
of
that
have
have
approached
us
and
said.
A
Oh,
we
want
to
be
come
back
involved
with
this
again,
so
DIMMs
has
been
working
on
getting
testing
in
place
so
that
you
know
so
that
so
that
we
can
be
sure
it
works
and
it's
tested,
but
it's
actually
it's
syndra
has
an
external
storage
solution
has
been
available
for
over
a
year
now,
yeah.
So
in
the
repository,
is
a
community
slash
cloud
provider,
cloud
provider,
OpenStack
yeah
this.
A
Yeah
yeah
was
that
was
that
you
that
brought
it
up
in
this
in
the
slack
Channel
earlier
today.
Someone
else
brought
it
up
and
said
that
they
have
a
driver
that
they
want
to
contribute
to
the
repository.
So
we're
we're
gonna.
Take
that
submission
that
you
know
pretty
much
sometime
in
the
next
few
weeks.
It's
going
to
start
rolling
in
and
and
we'll
have
that
there.
A
Okay,
well,
I
think
we're
at
time
thanks
for
everybody
for
for
coming.
If
you
have
any
more
questions,
we're
will
be
here
afterwards
and
sig
OpenStack
channel
everybody's
welcome.
If
you
want
to
contribute,
we
have
we
have
the
owners
file
set
up
and
you
can
email
us
and
we
can
get
you
hooked
up
if
you,
if
you
want
to
start
reviewing
code
and
you
know
contributing
back
so
thank
you,
everybody
for
coming
and
hopefully
we'll
see
you
at
the
summit.