►
From YouTube: Lightning Talk - Kubernetes and Ceph Integration: From Deployment to Production by Huamin Chen
Description
Lightning Talk - Kubernetes and Ceph Integration: From Deployment to Production - Huamin Chen, Red Hat
This talk presents recent status of Ceph and Kubernetes integration. Deploying Ceph Cluster on Kubernetes using DaemonSet significantly reduces the administrative overhead to get Ceph Cluster ready for end users. Using Ceph RBD Storage Class that is introduced in Kubernetes 1.4 brings flexibility and ease of use to end users.
About Huamin Chen
A passionate system software developer, Huamin Chen contributes to open source projects spanning from A to Z: Apache BigTop, Ceph, fio, Gluster, Kubernetes, Tachyon, and ZFS. Huamin Chen is currently employed by Red Hat. Follow him at http://github.com/rootfs
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
Alright,
let's
just
get
started
and
good
afternoon
everybody
and
good
morning
for
people
who
are
not
here.
My
name
is
Carmen
Chen
I
work
for
Red
Hat.
If
the
name
is
too
long
for
you
to
remember,
I'm,
also
known
as
a
loser,
that
some
kids
hop
on
Twitter
I
feel
fortunate
to
represent
a
host
of
brilliant
minds
to
work
on
kubernetes
and
staff
integration.
So
there's
a
lot
of
things
happens
in
the
background:
a
lots
of
interesting
projects.
A
So,
if
you
want
to
join
one
of
those
shaft
occurs
at
inspo,
kubernetes
and
a
bunch
of
other
side.
My
projects
you're
welcome
to-
and
maybe
you
can
talk
is
offline.
So
what
happens
in
this
talk?
It's
only
five
minutes.
We're
gonna
have
a
port
is
a
very
quick
overview.
What
is
going
on
so,
as
you
are
familiar
with,
staff
staff
has
a
multiple.
A
Are
nice
interfaces
to
integrate
with
your
different
use
cases,
so
staff
has
a
proc
storage
which
are
frequently
used
in
OpenStack
force
in
the
drivers,
as
well
as
the
file
system,
semantics,
which
is
becomes
a
kind
shinned
cut
into
your
production,
great
in
self
2.0,
which
is
case
released
this
summer
and
I.
Also,
staff
has
nice
as
three
compatible
objects,
interface
staff
readers,
great
way,
which
is
also
getting
a
lot
of
popularity.
When
you
use
in
OpenStack,
you
have
a
choice
of
Swift
and
self
readers
gateway.
So
that's
the
case
we
were
going
to
use.
A
So
what
are
using
primarily
on
kubernetes
is
the
staff
IBD,
the
readers
per
octave
I've
recently.
We
also
support
us
ffs,
so
the
voice
from
the
community,
as
well
as
from
the
end
user
for
service,
is
great
technology.
But
it's
really
hard
to
use
deploy
serve
is
a
nightmare
with
also
week
we
probably
cannot
cast
is
ready
to
production.
A
So
good
news
is
that,
with
a
puncture
for
hard
work
from
the
community
and
from
people
from
different
countries
or
companies,
whereas
myrrh
and
and
upon
some
people,
I,
don't
know
we
created
a
lot
of
technologies
to
facilitate
the
deployments
ourselves
cross
the
deployments.
So
one
way,
if
you
are
a
community
and
user,
you
can
deploy
staff
using
a
demon
set
if
you're
going
to
receive
parker
github
and
there's
some
people
created
a
repository.
A
So
you
cannot
use
the
bunch
of
scripts
and
images
staff
team
images
to
create
the
demon
set.
The
demon
status,
such
as
rent,
a
single
diamond
beauti,
main
image
staff
safety,
man
image
the
image
we
all
starts
self,
monitor
self
or
STS,
and
as
well
as
as
demons.
So
they
were
all
bangalore
into
a
single
image
and
through
the
deployments
or
diamond
size
deployments,
you
can
create
this
self
cluster
instantaneously.
A
We
also
support
self
deployments
through
the
expo,
which
is
more
also
for
compressed
these
environments.
You
can
create
a
cluster
using
self
ensberg
which
supports
para
mental
deployments,
as
well
as
the
continued
rise.
The
staff
used
a
function
of
our
different
images.
We
suppose
both
the
upstream
safe
demon
images,
as
well
as
the
has
supported
images,
are
safe
and
once
you
deploy
self,
when
you
create
part
and
creates
deployments,
you
can
consume
self
very
easily,
thanks
to
other
reasons,
great
progress
on
the
crust
to
the
storage
cross
and
dynamic
provisioning.
A
So
you
don't
really
need
to
know
the
details
of
your
staff
cluster.
Each
time
you
create
sub
heart
and
create
volume.
All
you
need
to
do
is
to
create
the
top
level
storage
crust
and
we
see
in
the
storage
class
you
just
embed
the
necessary
information.
So
the
next
time
when
you
create
the
crane,
you
can
just
remember
to
pick
about
your
crust
name,
and
then
you
get
the
Safa,
be
the
image
right
away,
the
kind
of
quick
sense
of
how
sorry
for
the
bad
presentation.
A
But
the
idea
is
that
this
is
a
storage
cross
definition.
So
I
was
34.
You
need
to
put
inside
into
the
storage
class
Authority
with
minimal,
so
you
need
to
have
the
IDs.
You
want
to
use
the
monitors
that
you've
already
created
in
the
South
cluster
and
a
bunch
of
other
things.
That
namespace
is
pretty
much
standard
and
the
secrets.
Why
do
you
have
to
keep
in
mind?
That
is
only
create
the
ki
BD
so
to
use
cell
phone
kubernetes?
A
That's
the
only
way
you
can
use
right
now
and
it's
only
supports
fib
the
image
one,
which
means
you
cannot
do
the
flattening
and
the
fancy
features
like
that
and
I
think
there's
working
progress
from
some
other
communities
from
the
community
and
developers.
That's
fine
debris
in
the
SAP
MBD
and
we
also
have
projects
called
a
QC
q,
mu
TCM.
You
could
it
be
the
pathogen
routine
technology
to
presents
IBD
as
a
local
block
device.
So
we've
seen
there's
our
brilliant
technologies.