►
From YouTube: Ceph Orchestrator Meeting 2020-08-17
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
The
the
I
container
builds
are
broken
okay,
which
means
that
I
can't
merge
anything
right
now,
which
is
a
pity,
and
if
one
of
you
knows
if
david
galloway
is
going
to
be
online
today,
I
haven't
seen
him
today
because
it
looks
like
he
is
basically
the
only
one
having
access
to
to
the
to
quiet.io.
C
Maybe
we
can
comment
this
to
somebody
in
the
in
the
christina
venus
team:
okay,
david.
I
think
that
he's
in
the
cristina's
team,
so
maybe
somebody
in
the
in
this
team
can
help
us.
Okay,.
A
A
A
Then
we
can.
We.
I
tried
to
talk
about
it
three
weeks
ago
and
then
you
were
on
vacation
and
then
I
was
on
vacation.
C
Yeah,
well,
I
I
propose
start
with
with
the
refactorization
okay,
and
it
was
a
lot
of
lines
of
code
and
it
was
rejected.
Okay,
we
can.
We
can
try
to
to
do
the
same
step
by
step.
Okay
and
well.
I
can.
I
can
start
with
that
this
week.
Okay,
the
the
first
thing
that
I
wanted
to
do
is
just
to
what
to
to
finish
with
the
with
the
back
of
the
open
boards.
Okay
and
after
that
that
basically
separates
firewall
management
start
with.
C
Refactorization,
we
can't
do
that.
Okay,.
A
Okay,
the
other
refactorizations
that
were
pretty
critical
as
it
seems,
are
making
self
adm
a
proper
package.
C
A
And
properly
distributing
safe,
adm
binary
to
the
to
the
real
remote
machines
yeah.
How
that
ever
is
going
to
look
like
I'm
pretty
open
with
with
that,
but
it
just
needs
to
properly
work
and
the
the
thing
we
right
now
have
doesn't
really
work
in
the
long
run.
C
C
Well,
I
I
agree
with
you
that
it's
important
and
try
to
to
work
only
with
the
fadm
package
mode,
okay,
and
I
think
that
maybe
well
just
to
know
one
one
thing.
It
could
be
acceptable,
for
example,
to
install
fav
as
package
as
part
of
the
requirements.
C
For
example,
we
are
we're
having
a
check
in
this
moment
in
order
to
to
check
before
the
hosts
are
prepared,
the
the
minimum
requirements-
okay
and
what
we
can
do,
for
example,
is
just
to
to
launch
this
style
of
of
fedm
as
a
package
if
it
is
not
installed,
okay
and
to
have
installed,
install
it
in
every
node
favn,
okay,
with
one
version
in
in
in
order
to
check
what
what
is
the
current
version
and
be
able
to
manage
all
all
the
problems
that
we
can
have
with
the
managing
versions.
A
C
Okay,
I
will
take
a
look
okay,
but
as
maybe
as
part
of
the
the
factor.
A
A
A
Yeah
exactly
and
if
we
really
want
to
want
to
extend
on
its
functionality
when
it
comes
to
managing
stuff
by
on
just
deploying
containers
like
like
like
stuff.
That
is
not
really
important,
that
that's
that's
right.
Right
now,
in
the
in
the
job
of
self-ansible,
for
example,
right,
like
restarting
machines,
I
don't
know
we
really
have
to
properly
re-architect
yourself,
edm,
not
not
re-architecture
the
fadm
but
migrate,
the
architecture
to
something
that
is
much
more
into
developers.
C
A
C
A
It's
just
to
yeah
there's
something
that
really
prevents
us
from
from
doing
large-scale
python
refactorizations
and
just
have
to
make
those
things
go
away,
and
then
we
can
continue
with
proper
and
and
then
at
that
point.
Refactorizations
of
fadm
binary
will
be
much
easier
because
then
we
have
all
the
the
proper
tools.
A
A
To
to
the
to
the.
A
A
Yeah,
that's,
and
at
least
provides
something
that
is
similar
to
what
we
have
in
seven
ansible
right
now,
because
defensible
also
converts
the
latest
tag
to
the
digest,
which
means
that
we
are
not
inventing
something
new
here,
but
only.
A
Copying
over
the
solution
that
we
already
have
with
the
defensible
enters
fdm
and
that
I
think
a
good
good
way
forward.
Yes,.
A
A
Yeah,
any
any
high
level
feedback
would
be
great
if
that's
actually
working
or
I
mean
the
code
itself
doesn't
work.
Yet
it's
still
work
in
progress,
but
any
high
feedback
would
be
great.
A
Okay,
jack
yeah
put.
C
B
I
don't
have
some
sort
of
stone
opposition
to
it,
but
I
do
want
to
include
all
of
that
or
all
of
that
output
so
that
somebody
who's
never
seen
the
installation
can
see
exactly
what
happens
when
it's
being
added.
So
if
you
think
it's
okay
to
have
a
gigantic
page
that
has
a
big
scroll
bar
on
the
side,
I'm
cool
with
it.
A
No,
I
I
don't
don't
that
we
have
to
have
one
big
page
with
the
scrollbar
in
it.
It's
just
that
we
have
only
one
guide
for
getting
started
and
not
two.
B
B
I'm
I'm
not
trying
to
confuse
the
confuse
the
matter
with
an
extraneous
getting
started
guide.
A
Mean
you're
pretty
hefty:
what's
that,
what's
pretty
hefty,
they
they
overlap
pretty
big.
It's.
B
Deploying
a
new
cluster
is
a
little
bit
more
abstract
like
it.
Doesn't
it's
not
like
a
strict
trail
of
breadcrumbs
through
the
forest
it
doesn't
like?
It
doesn't
lead
you
from
one
thing
to
the
next
thing.
The
next
thing
it
requires
somebody
who
might
not
know
what
the
architecture
of
their
deployment
is
going
to
look
like
to
make
decisions
about
that.
B
So,
like
I,
I
understand
the
desire
not
to
have
a
like
a
second
procedure,
because
it
might
be
confusing,
but
I
think
that
the
the
install
guide-
it's
not
the
so-called
crisp
install
guide,
leads
too
much
up
to
the
first
time.
Reader.
B
So
I
will
this
right
here
is
chat.
B
Are
you,
let
me
ask
you
this
this?
This
might
actually
help
me
straighten
this
out
in
my
head.
B
Do
you
think
it's
a
good
idea
to
have
the
button
on
the
front
page,
which
I
will
make
a
bit
smaller?
I
I
take
your
point
to
have
the
button
on
the
front
page
link
not
to
this
so-called
octopus
getting
started
guide
but
to
the
set
adm
install
guide
that
already
exists.
A
I
I
mean
this
that
definitely
makes
sense
to
have
a
big
button,
not
that
big,
but
a
prominent
way
to
right
direct
use.
B
Us
to
do
you
want
it
to
point
to
the
the
deploying
a
cluster
guide
instead
of
the
octopus
getting
started
guide
that
I've
written.
B
Because
if
so,
then,
what
I'll
do
is
I'm
sorry?
B
I
didn't
mean
to
like
just
ask
a
question
and
start
talking
again,
but
I
will
go
through
the
deployment
guide
this
week,
like
on
the
sepia
cluster
myself
and
see
if
it
seems
like
there's
any
any
places
where
it
needs
to
be
explained
a
little
bit
more
for
a
beginner,
because
because
this
is
all
based
on
like
back
in
may,
when
I
was
when
I
was
deploying
a
three-node
cluster
on
some
intel
nox
like
there
were,
I
it
I
felt
like
the
the
cognitive
load
on
the
beginner
was
heavier
than
it
should
have
been
with
the
unemployment
guy
that
it
had
been.
A
A
But
but
I
don't
know-
and
it
just
feels
weird
to
have
this
thing-
one
for
different
audience
and
the
other.
B
Well,
the
thing
is
I:
I
don't
really
want
to
lose
the
step-by-step
guide,
because
it
reminds
me
I'll
use
an
analogy
here.
It's
a
labored
analogy
when
you're
setting
up
arch
the
first
time
and
like
if
you're
not
from
the
90s,
so
you
didn't
have
to
do
like
linux
installs
from
from
like
the
from
scratch.
B
So,
like
you,
you
just
don't
know
that
when
you
start
out,
so
you
need
like
a
guide
that
tells
you
that,
and
I
feel
like
the
deployment
guide,
that
we
have
sort
of
assumes
that
people
know
that
they,
you
know
boot
route
swap
and
home,
or
the
four
standard
partitions
that
you
get
and
like
in
in
the
install
set
adm
section.
B
It's
good,
I'm
not
saying
that
it's
bad!
It's!
It's
not
it's
good!
The
sentence
isn't
under
clean
there's
nothing
in
there
to
be
to
be
embarrassed
about,
but
it's
a
reference
like
it
has
several
different
methods
of
installing
separation.
Like
there's,
dnf
install
sephidium
zipper,
install
sefadm
or
you
can
just
like
batch
it
with
curl.
B
And
I
I
guess
here
I'm
assuming
that
the
reader
is
going
to
be
I'll,
just
be
blunt,
dumber
and
more
impatient
than
we're.
Thinking
like
I,
I
think
that
a
reader
will
not
necessarily
realize
on
first
blush
that,
like
zipper,
doesn't
work
in
red
hat
based
deployments
for
instance.
B
So
I
I
don't
like
I
I
I
get
your
point
and
I
think
that
for
someone
like
you
who
is
so
conversant
with
the
ceph
infrastructure,
like
of
course
like,
of
course,
the
deployment
guide
has
no
pitfalls
in
it.
For
you
because
you,
you
know
how
to
contextualize
everything,
but
the
deployment
guide
is
more
reference
than
it
is
tutorial
and
it
doesn't
hold
the
hand
of
the
reader
all
the
way
through
and
the
idea
behind
this
chris
getting
started
guide
is
like.
A
I
mean,
and
then
that's
actually
good.
We
have
with
with
this
install
guide
the
old
one.
It's
too
too
short,
which
one
the
old
one.
B
Sophie
not
the
deploying
any
stuff
cluster
page.
B
Yeah,
it's!
I
don't
think
that
it's
too
short,
I
think
that,
like
it
says
so
we
we
have
a
step
that
is
like
you,
you
fetch
safety
and
then
you
bootstrap
the
new
cluster.
So,
like
I,
I
know
what
that
is.
You
know
what
that
is.
You
know
what
it
is
better
than
I
do,
and
then,
after
that
we
have
the
thing
where
we
we
direct
people
to
enable
itself
cli.
B
So
they
don't
have
to
go
into
the
shell
every
time
and
then,
after
that
we
just
tell
them
add
some
more
host
to
the
cluster,
which
is
technically
correct,
like
there's,
there's
nothing
technically
wrong
with
it,
but
if
they've
never
set
up
stuff
before
like
how
will
they
know
how
many
hosts
that
they
have
that
that
we
expect
them
to
set
up
how
we
know
they're
supposed
to
go
to
three
like
it's?
B
It
is
written
more
like
a
reference
than
it
is
a
tutorial,
and
so
the
idea
behind
the
installation
guide
is
just
that.
It
is
a
a
step-by-step
handholding
tutorial
and
I
don't
think
that
we
have
that
level
of
of
detail
in
the
install
guide
as
it
stands
right
now,.
B
Like
it
tells
you,
you
know
to
deploy,
monitors
on
a
specific
set
of
hosts
stuff
or
apply
mon,
and
then
we
have
inside
the
angle
brackets
host
one
host
to
host
three.
B
But
I
can't
remember
the
exact
details
of
this
right
now,
but
I
remember
that
there's
something
funny
there
like
you
have
to.
You
have
to
do
that
command
a
certain
way
so
that
the
second
and
third
commands
don't
clobber
the
first
one
or
so
that
each
sub
command
doesn't
clobber
the
one
that
came
before
it,
and
I
think
we've
got
that
in
the
install
guide
and
not
in
the
sepham
deployment
guide.
B
B
A
Really
valuable
and
the
idea
that
the
user
really
has
to
know
less
about
self-adm
and
seven
general
is
really
the
preferred
way
to
go
forward.
B
Oh,
so,
do
you
wanna?
Do
you
wanna
like
replace
what
we
have
with
the
crisp
getting
started
guide,
because
if
so,
that's
going
to
take
a
look
like
I'm
not
opposed
to
that
either,
but
that's
going
to
take
some
work
because
the
like
the
the
crisp
getting
started.
Guy
just
has
everything
it
has
like
the
it
has
the
host
name
of
these
knucks
in
my
house
in
it,
and
all
this
stuff
that
seems
sort
of
would
seem
extraneous
in
the
context
of
something
that's
more
like
the
reference
of
the
deployment
guide.
A
A
One
called
one
called
deploying
in
yourself
cluster
and
the
other
one
called
installing
self-octopus,
which
is
basically
the
same
thing
and
do
the
exact
same
steps
and
the
first
one
is
a
bit
less
for
both
and
two
and
depends
on
too
much
existing
knowledge
and
the
other
one
is
better.
I
think
in
general,
it's
better,
but
it
misses
something
that
we
at
least
misses.
Rtw,
mrs
always
misses
some
additional
stuff
right:
yeah
deploying
ganesha,
for
example,
mds.
A
So
so
my
vote
would
be
actually
to
merge
those
guides
and
maybe
split
it
out
into
different
pages,
one
for
for
the
chris
getting
started
guide
and
the
additional
pages
for
adding
ganesha,
adding
lgw
and
stuff,
like
that.
So
having
a
a
common
page
to
get
the
self
cluster
running
and
then
adding
those
gateways
in
in
different
different
pages.
I
don't
know.
B
So
like
a
central
place
where
we
walk
people
through
the
basics
and
then
after
that,
it
it
like
branches
out
right,
yeah
stuff,
like
that.
Okay.
A
C
B
All
right,
I
will
I'll
see
what
I
can
do
with
it
and
then
send
you
a
pull
request
like
starting
to
bring
these
things
together,
and
maybe
maybe
we
can
get
this
done
by
this
week.
B
I
want
to
get
the
install
guide
into
some
sort
of
decent
shape,
because
then
I
want
to
go
through
the
dev
guide
thoroughly
and
and
like
clean
it,
and
then,
after
that
I
have
some
like
once
we
have
a
really
good
install
guide
and
then
the
dev
guide
has
been
looked
over
by
everybody
who
knows
all
the
specific
stuff
about
the
different
parts
of
it.
I
have
a
whole
chain
of
video
making
tools
so
that
we
can
make
video
tutorials
or
video
walkthroughs
of
things.
B
So
what
I'll
do
this
week
is
I'll?
Try
to
merge
these
two
different
pages
into
one
and-
and
I
guess
we'll
probably
have
an
awful
lot
of
conversations
on
on
github
this
week.
So
that
should
be.
That
should
be
a
good
good
thing
to
it'll,
be
good
practice.