►
From YouTube: Patrick McGarry @ Ceph Day Paris
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
A
So
we'll
talk
a
little
bit
about
the
community.
I've
got
some
some
stuff
about
current
dev
operations
around
giant
some
of
the
modules
orchestration
and
deployment,
as
well
as
some
of
the
discussions
that
we're
having
right
now
around
our
release.
Cadence
and
then
I
can
answer
any
questions
you
may
have
about
ceph
in
general,
or
the
community
and
I'll
do
my
best
not
to
flub
them
so
the
community.
A
So
the
question
I
get
asked
most
often
these
days
is
kind
of
what
is
life
like
in
the
post
acquisition
world.
Now
that
ceph
and
inktank
folks
are
living
life
under
the
red
hat,
I
have
to
say
it's
been
pretty
good.
The
the
nicest
thing
about
red
hat
is
they've
really
been.
You
know
obviously
doing
open
source
for
a
long
time,
and
so
they
kind
of
have
a
an
ecosystem
of
people
who
get
it.
A
So
the
very
first
thing
that
happened
once
we
got
into
red
hat,
was
everyone
inside
red
hat
was
told
to
leave
us
alone?
I
said
these
guys
are
doing
great
things,
just
let
them
keep
doing
those
things
we'll
figure
out
the
rest
later.
A
For
those
of
you
that
didn't
follow
the
acquisition
closely,
it
was
probably
the
fastest
one.
I've
ever
seen.
A
I
and
I
haven't
seen
a
whole
lot
of
them,
but
I
do
know
that
in
a
room
filled
with
10,
you
know
bankers
and
lawyers
and
everyone
when
this
thing
finally
got
signed,
they
all
kind
of
looked
around
the
room,
astonished
and
said:
I've
never
seen
an
acquisition
close
this
fast,
so
it
was
kind
of
fast
and
furious
there
for
a
little
bit,
but
thankfully
they've
kind
of
just
left
us
to
our
own
devices,
with
the
exception
of
obviously
giving
us
some
additional
resources,
and
now
now
some
time
has
passed
and
everyone
has
kind
of
become
acclimated.
A
We
are
starting
to
get
kind
of
folded
into
red
hat
a
little
bit
more
and
with
that
being
folded
in
a
little
bit
more
comes
obviously
a
little
bit
more
focused
on
the
red
hat
distributions.
As
most
of
you
know,
you
know:
sef
started
its
days
as
primarily
in
ubuntu
shop,
all
of
the
developers
developed
on
ubuntu.
A
Obviously,
that
has
grown
even
before
the
acquisition
and
now
that
we're
here
we're
doing
a
very
concerted
effort
to
make
sure
that
all
of
the
distributions
get
their
their
due
attentions
and
with
that
there's
obviously
a
fair
amount
of
time.
That's
being
spent
looking
at
the
kind
of
the
stack
of
red
hat
distributions-
and
I
don't
know
if
you
guys
have
been
following
the
whole
centos
thing
as
well,
but
with
that
they're
kind
of
looking
at
fedora
as
the
fast-moving
errant
little
brother.
A
We've
got
a
number
of
community
members
now
that
are
starting
to
push
for
kind
of
centos
as
the
the
way
to
go
with
sef.
Now
that
we're
under
red
hat,
that
said,
we
are
still
maintaining
all
of
our
ubuntu
and
sousa
coverage.
We
are
making
sure
that,
because
we
got
acquired
by
red
hat,
we're
making
extra
sure
that
the
I
think
it
was
about.
80
percent
of
our
users
were
on
ubuntu
before
the
acquisition.
So
we
want
to
make
sure
that
those
folks
don't
feel
like
they're
left
out
in
the
cold.
A
A
The
answer
is
we
don't
know
the
the
good
part
of
the
answer,
however,
is
neither
does
anyone
else
and
everybody's
willing
to
figure
it
out
it's
kind
of
up
in
the
air
right
now,
given
the
state
of
ceffes,
we
are
kind
of
very
easily
drawing
the
line
down
the
middle
of
the
road
and
saying
gluster
has
a
file
system.
That's
where
that
goes
right
now,
within
the
red
hat
ecosystem
and
saf,
has
the
object
and
block
stuff.
That's
been
maturing
very
quickly
under
the
umbrella
of
openstack.
A
So
that's
kind
of
the
easy
divide
division
right
now
now
that
said,
a
lot
of
effort
is
still
going
into
cfs
right
now
and
when
cefs
becomes
mature,
that's
going
to
take
that
whole
messaging
and
throw
it
right
out
the
window,
so
we're
still
kind
of
figuring
out
how
it's
going
to
land.
But
the
short
answer
is
red
hat.
It
really
is
doubling
down
on
both
of
them
to
make
sure
that
neither
community
feels
like
they're
getting
squashed
under
the
other
one,
so
yeah.
A
It
should
be
an
interesting
future
either
way,
and
obviously,
with
all
of
this,
you
know
we're
learning
things
in
both
directions:
red
hat
is
learning
some
things
from
us
and
and
we're
definitely
learning
some
things
from
them,
especially
how
to
scale
people
to
large
numbers.
You
know,
given
that
ink
tank
was
only
about
55
aboard
the
acquisition,
and
now
we
have
you
know
6
500
new
co-workers,
so
it's
been
kind
of
interesting
all
right.
The
next
big
thing
in
the
community
is
ceph
days.
A
The
saf
day,
events
have
been
really
great
for
us
in
the
past
because
it
allowed.
You
know
people
like
us
to
kind
of
get
together
and
discuss
what's
next.
What
do
we
need
out
of
ceph?
What's
it
doing?
What's
it
not
doing,
and
then
we
could
take
that
feedback
directly
back
to
our
developers
and
hand
it
to
them,
and
it's
been
a
really
good
driving
force
behind
kind
of
our
roadmap
decisions.
A
It's
I
tell
you,
it
was
the
ceph
days,
probably
more
than
anything
else
that
pushed
us
into
saying:
okay,
we're
a
startup.
We
need
to
focus.
What
do
we
focus
on
and
we
focused
on
openstack.
It
turns
out
that
was
the
right
answer,
but
now,
like
I
said,
we're
taking
the
focus
of
ceph
days
and
we're
shifting
them
a
little
bit,
we're
cutting
out
all
of
the
marketing
sales
portion
of
it
and
we're
trying
to
make
these
really
naturally,
community
events.
A
You
will
probably
see
much
more
developer,
focused
activities
at
these
things
in
the
future,
and
I
and
I
can
tell
you
for
sure
that
no
two
ceph
days
are
going
to
be
alike.
So
if
you
have
the
opportunities
to
attend
multiples,
I
would
highly
encourage
you
to
do
so.
Every
single
one
of
the
ceph
days
is
pulling
speakers
out
of
the
community
just
like
you've.
A
Seen
today
already
now,
a
lot
of
our
community
members
that
were
recruited
to
speak
here
today
are
becoming
less
community
and
more
just
members,
as
red
hat
goes
on
a
crazy
acquisition
spree,
but
in
the
future,
you'll
definitely
see
like
san
jose.
A
We've
got
people
from
you
know,
brocade
and
we've
had
people
from
citrix,
and
you
know
there's
all
kinds
of
people
out
there
that
are
developing
with
ceph,
either
in
enterprise
environments
as
a
user
or
are
just
interested
hobbyists
and
those
are
the
types
of
people
you'll
see
showing
up
at
these
events
so
definitely
spread.
The
word.
A
We've
got
one
coming
up
in
san
jose
a
week
from
today
we
have
ones
coming
up
in
new
york,
city
and
london
and
by
all
means,
if
you
have
a
location
where
you
would
like
to
host
this
f
day,
you
have
a
venue.
You
know
I
thanks
to
to
leak
for
finding
this
venue.
A
This
is
a
very
unique
venue,
but
it's
always
fun
to
go
kind
of
new
places
and
and
see
where
we
can,
where
we
can
kick
steph
around,
so
yeah
definitely
keep
an
eye
on
this
day
program
going
forward.
A
A
For
those
of
you
that
didn't
know
it
existed,
we
have
a
website
that
tracks
many
of
our
community
metrics
from
you
know,
irc
to
the
mailing
lists
to
code
commits
and
tracker
activity,
we're
doing
a
lot
of
a
lot
of
stuff
that
we
want
to
know.
We
want
to
know
what
the
health
is
of
our
community
and
in
true
ceph
fashion.
We
don't
want
to
make
any
of
this
clothes,
so
all
of
these
materials
and
and
information
that
we're
gathering
about
our
community
we're
immediately
sharing
back
with
our
community.
A
So
we
want
to
make
sure
that
that's
kind
of
accessible
to
everyone-
that's
out
there
and
a
lot
of
this
really
is
community
driven.
So
if
there's
something
that
you
find
interesting
or
a
stat,
you
think
is
missing.
Let
us
know
we
put
out
us.
We
put
together
a
statement
of
work
for
the
betergia
guys
who
are
the
ones
that
are
building
this.
A
For
those
of
you
who
don't
know,
betergia
is
a
really
small
little
company,
that
kind
of
went
nuts
with
a
metrics,
dashboard
and
they're
doing
a
metrics
as
a
service
they're,
the
ones
who
built
the
openstack
metrics.
So
so
all
of
this
stuff
is
run
by
by
the
baturgia
guys,
and
they
took
a
number
of
open
source
projects.
A
The
the
metrics
grimoire
cvs
analyst
and
a
few
others
and
kind
of
put
it
all
together
into
this
really
cool
tight
little
metrics
package
and
they'll
customize
it
any
which
way
you
want,
because
then
they
can
take
that
customization
and
turn
it
around
and
use
it
for
all
of
the
other
people
that
are
using
them
for
metrics.
So
it's
actually
a
really
cool
little
project
and
we've
seen
some
some
awesome
stuff
come
out
of
it.
That's
helped
help
make
some
good
decisions.
A
So
if
you're
interested
check
it
out,
if
you
have
any
feedback,
we're
always
happy
to
get
it.
The
the
latest
thing
that
we
added
there
was
the
the
all
history
button
over
there.
You
see,
you
can
now
break
it
down
by
quarters,
so
we
can
actually
go
back
in
history
and
kind
of
look
at
snapshots
of
what
was
happening
and
when
the
user
committee,
we
actually
just
had
a
meeting
what
about
a
week
ago,
a
week
and
a
half
ago,
something
like
that.
A
Pretty
recently,
the
user
committee
was
kind
of
our
first
baby
steps
towards
moving
our
governance
out
of
ink
tank
and
into
the
community.
A
Obviously
we
want
to
do
a
lot
more
of
that,
and
more
of
that
is
still
coming.
How
that's
going
to
manifest.
I
still
don't
know
yet
that's
one
of
those
big
question
marks
from
the
acquisition
that
hasn't
been
answered
yet,
but
the
user
committee
has
actually
been
doing
some
really
good
stuff
before
ceph
days
was
moved
more
into
a
community
effort.
A
The
user
committee
spent
a
lot
of
time,
organizing
meetups
and-
and
you
know
like
launch,
would
you
launch
this
with
firefly
as
as
a
part
of
the
work
up
towards
firefly?
I
think
it
was
yeah
right
before
firefly,
so
the
the
user
committee
launched
with
the
the
focus
of
kind
of
promoting
the
community
and
giving
them
more
of
a
solidified
voice
as
it
comes
to
you
know
things
like
self-development,
but
also
the
the
social
parts
that
go
along
with
that.
A
A
The
the
rains
have
been
handed
over
to
to
eric
is
eric.
Here
I
saw
his
name
on
the
list.
There
is
a
flight
problem.
Oh
he
got.
He
got
yeah
caught
by
the
our
friends
at
air,
france,
okay,
yeah,
so
he's
he's
the
new.
The
new
chair
of
the
user
committee
had
a
few
other
people
step
up
with
roles
in
there
still
more
roles
to
fill.
A
Obviously
I
will
continue
to
participate
in
that,
but
but
I
would
like
to
see
a
lot
more
of
a
separation
of
church
and
state
between
community
and
upstream
versus
product
and
red
hat,
and
I'm
going
to
do
a
lot
of
things
in
the
near
future
to
kind
of
delineate
that
and
make
sure
that
our
community
has
the
ability
to
kind
of
draw
lines
in
the
sand
and
say
no.
No.
This
is
what
we're
doing,
and
the
user
committee
is
obviously
a
great
first
start
to
do
that.
A
This
meeting
that
just
took
place
within
the
last
couple
of
weeks
addressed
a
few
things
and
another
thing
that
those
of
you
that
hang
out
on
the
mailing
lists,
a
lot
may
have
noticed
the
there
after
every
major
release
you
see,
a
list
of
people
of
here
are
contributors
by
commits,
and
it
kind
of
breaks
out
all
of
that
interesting
stuff
out.
A
That
is
a
function
of
the
user
committee.
There's
there's
a
rotating
schedule
now
of
folks
that
put
together
that
kind
of
snapshot
of
what
went
into
each
major
release.
So
that's
just
one
of
the
things
that
kind
of
is
a
part
of
the
the
day-to-day
operations
or
release
to
release
operations.
Anyway,
the
contributor
credits
and
one
of
the
big
discussions
that
came
out
of
this
most
recent
one
was
our
release.
Cadence
there's
been
some
rumblings
that
that
people
feel
like
the
cef
release.
A
A
Every
other
one
of
them
is
what
goes
into
kind
of
the
productized
polished.
You
know
the
ink
tank
seph
enterprise
is
what
it
used
to
be
called.
I
don't
know
what
it's
going
to
be
called
now,
but
every
other
one
of
those
was
the
really
kind
of
honed
polished.
This
is
your
production
and
the
ones
that
are
opposite.
A
The
in-between
releases
were
feature
releases
that
were
designed
to
kind
of
get
out
there,
not
quite
dev,
because
they
were
polished
and
released,
but
but
obviously
we
aren't
expecting
enterprise
production
deployments
to
be
upgrading
their
cef
cluster
every
three
months.
That
would
be
pretty
ridiculous,
so
I
feel,
like
our
communication
needs
to
get
better,
but
obviously
we
can
continue
to
kind
of
feel
that
out
and
see
where
the
release
cadence
will
fall.
Maybe
we'll
slow
it
down
a
little
bit.
A
But
for
now
we
feel
like
every
six
months,
is
pretty
reasonable
in
terms
of
feature
releases,
especially
because
there
are
so
many
new
features
that
are
coming
so
quickly.
You
know
the
erasure,
coating
and
tiering
stuff
having
cfs
that's
now.
Finally,
starting
to
show
up
on
our
horizon.
These
are
all
things
that
are
very
important
to
us
to
get
out
there,
because
it
means
huge
chunks
of
adoption
for
ceph
and
growth
and
the
ability
for
it
to
kind
of
continue
to
blossom
very
very
quickly.
So
obviously
the
meetups
continue
to
grow.
A
I
think
we're
up
to
about
30
meetups.
Now,
some
of
them
more
actives
than
others,
but
the
meetups
obviously
are
a
function
of
the
user
committee,
but
anyone
can
start
a
meetup.
It
can
be
two
guys
in
a
pub.
It
can
be
30
guys
in
a
conference
room.
It
doesn't
really
matter,
but
we
do
have
infrastructure
in
place
via
the
user
committee
to
provide
you
with
things
like
swag
and
logistics
and
there's
there's
help
that
you
can
get
to
start
your
meet
up,
publicize
it,
etc,
etc.
A
If
you
have
a
bunch
of
folks
that
want
to
get
together
and
talk
stuff
and
then
yeah,
we
then
the
rest
of
that
meeting
was
discussing
specific
use
cases
and
and
then
discussing
some
of
the
upcoming
work
that
was
coming
in
the
next
release
of
ceph.
So
the
user
committee,
it's
pretty
lightweight,
but
we
would
like
to
see
it
kind
of
start
to
grow
and
be
formalized
a
little
bit
more
if
you're
interested
in
participating.
I
definitely
encourage
you
to
do
so.
A
There's
all
kinds
of
places
where
folks
can
can
pick
up
and
help
with
that,
but,
like
I
said
it
is
our
first
step
towards
broader
governance,
and
that
is
coming,
whether
it's
an
independent
third
party
or
it's
just
bylaws
and-
and
you
know,
copyright
stuff-
that
we
write
up
as
red
hat
as
a
single
vendor
is
yet
to
be
decided.
But
it's
coming,
google
summer
of
code,
this
actually
just
wrapped
up
for
us.
This
was
our
first
year
participating
in
google
summer
of
code.
A
Does
anybody
not
know
what
google
summer
of
code
is?
I
said
all
right,
I
figured
most.
People
would
cool
it's
in
it's.
What
10th
year
now.
Is
that
what
they're?
Yes,
10th
year
of
google
summer
of
code,
this
was
seth's.
First
year
participating
we
applied
last
year,
didn't
get
accepted.
A
I
guess
they
just
didn't
feel
like
we'd
gained
enough
momentum
yet,
but
we
had
two
students
that
signed
up
for
for
chef
projects
out
of
or
two
students
that
we
were
able
to
accept
out
of
12
applications
that
we
had
for
people
to
work
on
ceph.
So
there
was
a
lot
of
interest
relative
to
our
size
and
available
slots,
but
we
had
two
pretty
cool,
pretty
cool
little
projects
that
worked
out
with
that
the
stuff
that
loic
was
talking
about
earlier,
the
reliability
modeling.
A
We
had
a
student,
a
young
lady
in
switzerland
that
was
working
on
some
of
that
stuff
and
extending
that
very
academic
it
was.
It
was
cool
to
see
her
end
right
up.
I
encourage
you
it's
on
the
on
the
seth
wiki
under
the
development
section.
If
you
go
look
there,
you
should
be
able
to
find
it
just
a
reliability,
final
report
or
something
like
that.
If
you,
if
you
want,
I
can
find
you
the
ur
afterwards,
but
she
did
some
really
cool
work
there
very,
very
academic.
The
other
one
was
the
wireshark.
A
Like
two
and
a
half
years
ago
it
was
written
and
it
just
never
worked,
it
wouldn't
compile
with
modern
wireshark,
and
so
we
had
a
guy
come
in
this
summer
he
wrote
a
brand
new
one
and
actually
works
pretty
well.
If
you
look
at
it,
you
can.
You
can
turn
on
the
desector,
plug
it
into
your
cef
cluster
and
watch
it
go
it's
it's
pretty
slick.
He
did
a
pretty
good
job
with
that.
A
We
will
be
applying
again
next
year.
So
if
you
are
hip
deep
in
ceph
code,
all
the
time-
and
you
want
to
be
a
mentor-
definitely
send
me
a
note.
Let
me
know
it's
always
nice.
When
we
can
get
because
we
had,
I
don't
know,
I
think,
25
or
30
suggestions
of
projects
that
people
could
write
up
and
apply
for,
but
it's
always
nice.
A
When
of
those
projects,
they
can
all
be
different
mentors,
so
people
can
kind
of
not
only
pick
what
they
want
to
work
on,
but
who
they
want
to
work
on
it
with
so
centos.
As
I
mentioned,
there's
been
a
lot
of
kind
of
excitement
and
effort
going
into
ceph
on
centos,
and
one
of
these
efforts
that
we
started
just
before
the
acquisition
was
getting
more
involved
with
them
on
a
more
formal
basis.
A
So
what
we're
trying
to
do
with
the
storage
sag
and
it's
still
very
much
in
its
infancy,
but
we're
trying
to
make
it
so
that
it's
very
easy
for
you
to
take
a
centos
machine
and
turn
it
into
a
storage
node
of
whatever
type
we've
we've
already
got.
Packages
put
together
for
both
steph
and
gluster,
and
we've
started
to
talk
to
like
the
extreme
fs
guys
and
the
oh.
Who
was
the
other
one
I
know,
but
there
was
another.
A
There
were
another
couple
of
distributed
storage
type,
applications
that
want
to
start
participating
in
this
stuff.
So
you
know
if
you,
if
you
know
any
of
those
guys
or
if
you
are
one
of
those
guys.
A
Let
me
know
the
more
the
merrier,
because
I
think
right
now,
all
of
the
the
people
that
are
driving
this
thing
are
now
internal
to
to
red
hat,
so
it'd
be
good
to
get
some
not
have
such
an
echo
chamber,
so
yeah
governance
yeah,
as
I
mentioned,
with
the
user
committee,
stuff
being
kind
of
the
first
step
towards
governance.
Those
of
you
that
have
heard
my
discussions
or
talks
previously
about
governance.
I
haven't
forgotten
about
you,
it's
just.
I
don't
navigate
political
water
so
well.
A
I
tend
to
be
a
bull
in
a
china
shop,
so
it
takes
me
a
little
longer
to
get
things
done.
Sometimes,
when
they're
big
seth
developer
summit,
we
got
another
one
coming
up,
it
hasn't
been
scheduled
yet,
but
the
look
for
the
call
to
blueprints
to
open
within
the
next
week,
or
so
the
the
firefly
giant
kind
of
release
timetable
got
pushed
out
because
firefly
was
so
big.
A
There
was
so
much
stuff
that
had
to
get
reviewed,
that
our
timetables
have
gotten
a
little
sloppy,
but
we're
going
to
tighten
that
back
up
the
next
cds
is
going
to
be
all
about
kind
of
the
next
iteration,
the
next
giant
release
and
and
moving
beyond
that,
because
we've
got
a
lot
of
blueprints
that
were
put
together
for
giant
and
then
there
was
the
interim
giant
hammer
release,
and
so
this
this
one
is
going
to
be
looking
specifically
at
hammer
and
collecting
all
of
the
stuff
that
got
left
out
or
or
didn't
make
it
into
giant
and
then
moving
forward.
A
So
if
you
have
any
development
ideas
desires
for
what's
going
to
go
into
hammer,
this
is
where
you
need
to
express
them
and
discuss
them
for
those
of
you
that
aren't
familiar
with
our
cef
developer
summit,
they're
100
online.
So
we
run
them.
They
used
to
be
google
hangouts.
Now
it's
a
tool
called
blue
jeans:
it's
an
open
source
tool,
it's
video
conferencing
and
basically
it's
it's
set
up
so
that
everybody
gets
a
chance
at
sage.
A
So
if
you
have
a
feature
or
a
bug
fix
or
some
you
know
sizable
chunk
of
work,
that's
going
to
go
into
the
next
release.
We
ask
you
to
write
a
blueprint.
Blueprints.
Are
it's
a
form
type
of
setup
on
our
wiki?
You
just
go
in
you
say
you
know,
I
want
to
add
whizbang
feature
a
you
give
me
a
couple
of
sentences
about
what
it
is
who
owns
it.
Ostensibly
it
would
be
you
and
then
maybe
a
couple
of
work
items.
It's
really
lightweight.
It's
not
meant
to
be
an
onerous
process.
A
It's
basically
everybody
who
actually
takes
the
time
to
fill
out
a
blueprint,
gets
a
30
minute
session
with
sage
to
discuss
their
upcoming
work
and
he
can
give
them
feedback
on
okay,
here's
how
you
should
do
it
or
keep
in
mind
you're
going
to
have
to
think
about
this
other
stuff.
That's
in
or
I
just
committed.
A
You
know,
3
000
lines
this
morning
as
sages
want
to
do
and
and
he'll
he'll
be
able
to
give
you
an
idea
of
what
kinds
of
things
you'll
run
against,
but
but
yeah
the
call
for
blueprints
will
be
opening
soon.
So
if
you
have
any
any
development
work,
that's
going
into
ceph,
keep
an
eye
out
for
that
on
social
media
and
the
ceph
blog
is
where
it'll
hit
all
right.
So
the
development
stuff
giant,
as
I
think,
has
been
mentioned
a
couple
of
times.
A
The
tree
is
frozen
as
of
nine
september,
so
we're
in
feature
freeze,
now
and
and
squashing
bugs
and
and
having
large
fights
in
conference
rooms,
and
things
like
that,
as
we
work
towards
the
release
of
giant
dot
85,
which
was
released
within
the
last
couple
of
weeks,
includes
some
some
pretty
cool
stuff.
That's
going
to
be
kind
of
the
banner
features
for
giant
chief
among
those
is
some
of
the
support
groundwork
for
the
rdma
stuff.
A
A
A
Additionally,
we've
also
seen
some
improvements
to
the
civit
web
based
front
end
for
the
rados
gateway,
which
is
really
exciting
to
me,
because
every
time
I
wanted
to
play
with
the
object
stuff,
I
had
to
stand
up.
You
know
apache
and
go
through
all
of
the
machinations
that
it
took
to
get
the
the
cgi
stuff
going,
and
that
was
way
too
much
overhead.
But
now
that
we
have
this
civic
web
stuff,
I'm
pretty
excited
to
see
how
that
works,
and
one
other
thing
that
was
kind
of
small.
A
But
I
felt
like
it
was
worth
mentioning
is
the
the
developers
you
know
sage
and
some
of
our
other
core
developers
are
really
starting
to
take
some
time
and
gather
user
feedback
from
the
lists
and
people
they
talk
to
about
kind
of
how
to
admin
a
cluster
and
how
to
make
that
better.
A
You
know
some
of
that's
going
to
come
with
the
gui
stuff,
but
as
the
gui
stuff
grows
and
becomes
easier,
they
want
to
make
sure
that
the
cli
kind
of
keeps
pace,
and
so
there's
there's
some
new
things
like
being
able
to
show
what
your
osd
is
blocked
by
type
things.
A
There's
also
we're
also
looking
at
a
few
other
things
that
are
oh,
it
was
the
other
one
dimension
now
that
I'm
under
the
lights.
What's
that
auditing
yeah
yeah,
some
of
the
auditing
and
metrics
and
stuff,
like
that,
just
just
being
able
to
pull
out
information
from
your
cluster
manage
day-to-day
things
as
well
as
making
you
know
things
like
preparing
for
for
dr
and
those
kinds
of
things.
A
Those
are
all
getting
a
lot
better,
so
you
know
keep
the
feedback
coming
they're,
definitely
listening
and
spending
a
lot
more
time.
Thinking
about
how
to
make
life
better
for
your
admins
and
so
the
next
one.
The
dot
86
release
is
probably
going
to
be
more
or
less
a
giant
release
candidate
and
then
we'll
continue,
squashing
bugs
right
up
until
giant
flies
out
the
door.
A
A
few
things
I'll
call
out
for
each
of
the
individual
modules
itself,
rbd
with
with
giant,
is
going
to
have
client-side
caching
turned
on
by
default.
This
was
something
that
we've
been
looking
at
for
a
long
time.
There
were
some
just.
I
won't
get
into
the
whole
stack
of
decisions,
but
originally
it
was
decided
that
we
didn't
want
to
do
that,
but
now
that
we're
kind
of
maturing
a
little
bit
more
and
a
lot,
a
lot
of
our
participation
from
folks
is
from
the
openstack
world,
and
they
really
want
this
client
side.
A
Caching,
stuff
and
a
lot
of
people
don't
remember
to
turn
it
on.
In
fact,
that
was
kind
of
the
crux
of
the
whole
thing
when
there
were
some
performance
benchmarks
that
came
out
a
while
ago,
a
little
grenade
that
was
thrown
over
the
wall
from
our
friends
in
gluster
and
and
kind
of
the
whole
crux
of
the
of
the
argument
was
the
fact
that
they
didn't
turn
on
client-side.
Caching,
and
it
was
you
know,
it
was
just
a
bad
setup.
A
The
other
thing
I'd
mention
about
rbd
is
rbd
was
in
the
news,
the
the
eucalyptus
stuff.
So
now
you
can
do
block
storage
with
eucalyptus,
backed
by
sef.
There
was
a
guy
who
wrote
up
a
great
blog
entry
on
it
it
I
just
thought
it
was
interesting
to
call
out
because
of
the
acquisition
of
the
eucalyptus
stuff,
so
always
fun
to
see
it
in
the
news.
A
Rgw.
As
I
mentioned,
the
civic
web
front.
End
stuff
is
very
exciting
for
us.
We're
also
looking
at
some
of
the
the
bucket
limitations
trying
to
decide
kind
of,
because
we
can
do
both
amazon,
s3
and
and
openstack
swift
natively
into
the
raido's
gateway,
sometimes
there's
decisions
that
will
affect
one
or
the
other.
So
there's
a
lot
of
of
thought.
That's
going
into
right
now
kind
of
how
best
to
handle
those
use
cases,
especially
as
it
relates
to
the
bucket
limits
and
things
like
that,
cfs
all
right.
A
This
is
the
most
exciting
part
about
all
of
this.
All
you
know
the
the
acquisition
hand
waving
the
biggest
excitement
for
me
and
and
I'd
imagine
for
sage
two
is
the
fact
that
a
fair
amount
of
resources
are
being
brought
to
bear
on
cefs
by
us
now,
if,
if
any
of
you
know
about
the
history
of
seth,
you
know
that
this
is
where
it
started.
A
The
trilabs
grant
was
to
build
a
better
luster
and
in
doing
so
kind
of
the
the
focus
of
sage
at
the
beginning,
before
he
kind
of
took
over
the
whole
thing
was
to
build
a
a
horizontally
scalable
metadata
server.
That
was
his
whole
deal,
and
so
this
is
very
near
and
dear
to
his
heart.
Having
saffs
not
be
or
be
the
only
part
of
the
platform,
that's
not
you
know,
production
ready
or
supported,
or
whatever
I
think
was,
was
making
him.
A
It
was
burning
him
up
inside
a
little
bit,
and
so
it's
nice
to
get
back
to
the
the
roots
and
get
cephes
to
where
it
needs
to
be.
That
said,
I've
I've
kind
of
got
to
start
backpedaling,
because
you
know,
while
we
say
it's
not
production
ready.
That's
that's
a
support
claim
more
than
anything
else,
because
there
are
a
lot
of
people
that
are
using
cfs
in
production
and
doing
quite
well
with
it.
A
You
know
it's
it's
just
a
matter
of
what
your
use
case
is
and
and
how
you're
handling
things
like
your
metadata,
but
for
those
of
you
that
are
involved,
you'll
know
that
the
core
team
of
ceph
developers
is
about
nine
guys,
there's
a
bunch
of
other
stuff
that
goes
on.
You
know
the
calamari
and
the
deployment
and
orchestration
stuff
and
there's
a
lot
of
other
stuff
that
goes
on
around
it.
But
it's
about
nine
guys
in
our
core
team
and
three
of
them
now
are
assigned
to
cefs
full-time.
A
So
it's
pretty
exciting
for
us
to
see
ceffes
actually
start
making
some
progress
in
addition
to
the
heroic
and
herculean
efforts
that
have
come
in
from
the
community,
including
like
wow,
I
think
there's
17
guys
in
china
now
that
are
working
on
cefs
full-time.
A
So
it's
it's
nice
that
we
can
start
to
match
the
effort
from
outside
and
even
more
than
that,
we've
started
a
kind
of
dog
foods
ffs.
So
it's
to
the
point
now
where
we
on
all
of
our
qa
infrastructure
internally,
we
actually
use
cfs
and
that's
kind
of
helped
us
to
sand
off
some
of
the
rough
edges
that
were
immediately
apparent,
so
feedback
is
always
encouraged,
especially
on
cfs.
It's
come
a
long
way
in
a
very
short
amount
of
time,
since
we've
kind
of
devoted
these
resources
to
it.
A
So
if
you
want
to
spin
it
up
and
take
another,
take
another
spin
to
the
the
garden,
I
know
sage
would
welcome
any
feedback
that
comes
in
on
the
lists
deployment
orchestration.
I
know
sebastian
spent
a
fair
time,
a
fair
amount
of
his
time
going
through
some
of
the
ansible
stuff.
So
I
won't
belabor
this
point
too
much.
A
A
Chef,
puppet
ansible,
salt,
juju,
all
viable
options,
crowbar
I
left
out
which
pete
mentioned,
but
we
have
available
recipes
for
all
of
them.
The
most
the
two
most
mature,
I
would
say,
are
probably
chef
and
puppet
and
obviously
those
are
the
two
most
mature
in
terms
of
solutions.
Now,
if
you
like
the
raggedy
wild
west
edge,
you
probably
are
more
happy
with
things
like
ansible
and
salt.
It's
interesting
to
note
that
sef
deploy
is
actually
largely
debased
on
and
utilizes
stuff
from
ansible.
A
So
I
know
alfredo's
a
big
fan
of
ansible
each
has
their
own.
You
know
pros
and
cons,
but
definitely
if,
if
you're,
if
you're,
spinning
up
a
cef
cluster-
and
you
have
thoughts
of
moving
towards
production,
anything
else,
I
would
start
using
the
orchestration
and
deployment
stuff,
because
it's
it's
well
worth
well
worth
using
and
we
talked
about
the
release
communications,
definitely
going
to
get
that
going
better,
that's
pretty
much!
All
I
had
for
my
stuff.
Do
we
have
any
questions.