►
From YouTube: 2019-12-16 :: Ceph Orchestration Meeting
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
B
Yes,
I've
missed
the
last
couple
of
weeks
with
conflicts,
but
here
I
am
alright.
She
was
start
with
work,
then
yep,
so
with
Brooke
planning
on
a
1.2
release.
Hopefully
in
the
next
couple
of
days
there
we're
just
finishing
up
the
last
few
fixes
there
I,
don't
think,
there's
anything
related
to
orchestration
around
that
that
we
need
to
talk
about
here.
B
Looking
good
one
change
just
last
week
that
I
wanted
to
point
out
is
that
so
in
the
past,
Brooke
has
had
a
couple
of
scenarios
where
it
automatically
removed
OS
D's
in
particular.
If,
if
you
specified
your
basically
your
devices
in
this
in
the
cluster
CR,
if
you
removed
them
from
the
cluster
CR,
then
we
would
go
ahead
and
remove
the
OSDs
automatically
like.
B
If
you
removed
a
whole
note
from
the
CR,
we
would
remove
that,
but
there
were
a
few
cases
where
people
unintentionally
remove
things
from
the
CR
and
then
they
lost
data
or
they
had
to
go.
Do
things
to
restore
things,
so
we
don't
want
to
remove
things
automatically
based
on
desired
State.
So
now
I
removed
that
code
that
deletes
those
DS
automatically
and
documented
steps.
A
B
B
A
A
B
B
A
C
Mm-Hmm,
no,
not
really
really
looking
forward
to
to
seeing
Elly's
PR
so
that
we
can
actually
because
I've
been
hearing
about
that
PR.
For
like
two
weeks
now
and
I
know
he
has
a
document
on
his
local
branch,
but
I
really
want
to
see
the
design
so
that
we
can
like
more
actively
I
guess,
comment
on
it,
but
yeah
nothing's.
Nothing.
We
need
to
add
to
that
yeah.
Oh.
C
A
C
C
A
C
A
A
And
also
my
envisioned
that
the
future
Roxy
I
would
solve
that
problem
for
me,
but
it
turns
out
to
be
kind
of
still
a
long
way
ahead.
So
right
now
I'm
hopeful
for
to
test
for
suggestions
how
to
actually
integrate
that
manager
module
into
any
that
I.
We
have
already
because
I
never
have
the
time
nor
the
knowledge
to
integrate
that
into
yeah.
B
To
test
that
module,
we
basically
need
communities
and
rook
running
great
yes,
I
mean
it's
there's
not
really
like
a
unlocking
framework
or
something
for
no
the
modules
so
yeah
that
that's
certainly
a
work
item
to
consider
on
the
Brookside
CI.
We
definitely
pull
that
in
and
well
we
enable
the
module.
We
don't
do
much
more
with
it.
I
guess.
E
Well,
I
I
think
that
even
we
haven't
got
the
any
person
checking
what
is
happening
with
module
in
the
money
one
day
the
manager
model
for
rock.
So
we
don't
have
tests
and
we
don't
have
a
I,
don't
know,
but
it
seems
that
honey
bodies
is
taking
a
look
to
this
model
and
is
trying
to
to
check
all
day
or
the
last
features
that
we
are
integrating
in
intro.
Okay,
so
maybe
maybe
is
but
I
would
be
happy
to
help
with
that.
E
Okay
or
if
you
want,
if
everybody
is
sorry
I
can
I
can
start
to
take
a
look
over
that.
Okay,
another
thing
is
what
kind
of
integration
test
we
are
going
to
have
with
a
top
manager
model,
because
because
it
it
implies
to
to
use
things
that
I
think
that
we
don't
have
him
theology
or
the
current
continuous
integration
system
for
poor,
safe,
so
yeah.
E
Yes,
there
is
yeah
what
is
what
I
think
so,
the
problem
there's
the
real
problem
about
the
continuous
integration
is
with
that:
okay,
I
think
that
we
have
something
in
downstream
with
you,
seem
proud
and
open
safe,
but
I
don't
know
if
it
is
possible
to
integrate
this.
This
kind
of
test
in
this
way,
because
this
thing
is
downstream
and
second
thing-
is
that
probably
what
is
something
completely
new
that
would
guide
time
in
order
to
to
make
it
work?
I.
B
Yeah
one
thought
I'm
having
we've
had
this
thought
of
they
let
the
ability
to
run
the
rook
CI
against
a
recent
or
a
custom,
set
image
right.
So
if
we
could
get
if
somebody's
got
a
PR
or
can
or
we
got
the
latest
doctor
push
that
image
or
whatever
if
we
can
run
our
rook
CI
against
that
specific
thing,
maybe
on
the
rook
side
of
it,
we
can.
You
know,
we've
already
got
kubernetes
you've
already
got
the
rook
module
enabled
no
can
we
run
it
from
the
rook
side?
B
Maybe
it's
like
a
daily
test
that
and
maybe
we
can
modify
it
when
we're
running
PRS,
to
point
it
at
a
specific
image,
or
you
know,
maybe
at
least
initially
you
have
a
quick
way
to
get
off
the
ground
before
we
go.
Do
this
whole
set
of
work
like
maybe
there's
something
on
that
rook
side,
we
can
do
to
get
something
sooner.
A
A
A
E
A
Oh
Sh
you're
mainly
concerned
about
prices
for
its
moment
last
three
mm-hmm,
like
your
s3
test,
pull
request.
D
D
D
B
A
D
A
A
D
That's
just
looking
at
Borden
I
shouldn't
meant
the
things
that
are
sort
of
above
the
octopus
line
of
prometheus
is
in
maybe
prepare
hosts
were
taking
out
of
scope.
D
A
D
A
A
A
Right,
maybe
we'd
see,
need
some
more
fine-grained.
Looking
people
I
said
it
shouldn't.
Look
anything
at
all.
What
right
yeah.
D
Yeah,
okay:
they
think
about
that.
Probably
if
we
have
like
a
lock
helper,
where
you
just
give
it
the
names,
some
strings
of
things
to
lock
or
it's
like
the
cluster
FS
ID.
If
it's
like
an
RM
or
something
like
that,
guard
locks
and
I,
don't
know
like
we're
doing
a
ploy
and
it
delete.
At
the
same
time,
they're
all
gonna
collide.
A
A
D
D
Yeah
yeah:
okay,
that's
an
easy
one,
I'm
wondering
if
the
in
the
background
the
orchestrator
should
periodically
check
the
hosts
just
to
see
if
they're
up.
D
A
A
D
D
Yeah
I
just
haven't
had
time
to
look
at
it
I'd
like
to,
but
I
really
want
to
get
the
five
on
three
stuff
rich
this
week.
So
yeah
we
come
back
to
it
later.
Okay,
I,
don't
know
if
somebody
knows
actually,
maybe
we
can
somebody
if
points
at
the
rook
code
that
deploys
the
kanesha
and
like
what
it
actually
does.
I,
don't
know
how
complicated
it
is
like
we
have
to
do
to
run
Ganesha
properly.
C
D
C
It's
something
that
we
I
think
we
I
was
looking
at
this
with
young
like
couple
of
months
ago
become
he
had
because
he
had
questions
around
deepest
forcep
containers
and
then
we
asked
I
forgot.
I
forgot
his
name,
the
guy
who
wrote
cognition,
augmentation,
well,
yeah,
yeah,
yeah,
Jeff,
I
didn't
hear
and
we
asked
them.
Why
do
you
do
the
bus?
If
you
don't
like
if
every
single
containers
are
all
isolated
and
they
can't
really
do
enjoy?
Like
my
PC
connections?
Anyways-
and
it
was
like-
okay,
I,
don't
know,
but
apparently
it's
working.