►
From YouTube: 2019-07-10 :: 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).
B
A
E
A
E
Well,
my
first
idea
is
to
go
to
order
Orchestrator
host
poster
pinch,
because
we
already
have
a
poster
page.
Around
Nashville
and
I
have
a
feeling
we
might
have
a
thread
or
two
kinds
of
of
hosts.
One
is
one,
is
Austin's
already
managed
by
staff
yeah.
A
A
E
A
E
D
E
E
D
E
E
A
E
A
F
A
A
Your
turn
to
double
your
pull
request
for
the
answer.
Orchestrator
I
can
run
tautology
on
that
on
that
pull
request,
I
think.
C
Sorry,
you
can't
brah
I!
Can
oh,
yes,
okay,
perfect,
okay,
if
no
I
I,
think
that
was
probably
the
best
thing
that
we
can
do
is
to
to
leave
it
in
the
current
status
and
well
III.
Don't
know
if
he
did
I
think
that
maybe
it
could
survive
in
upstream
branches,
but
I
do
not
expect
any
any
new
effort
in
this
path.
A
A
A
F
Yes,
I'm
not
really
working
on
anything
right
now:
I
guess,
I'm,
mostly
working
on
understanding,
rook
and
the
kubernetes
working
I
have
a
cluster
up
and
running
with
the
Bedrosian
self
in
it
I'm
playing
with.
But
it's
not
quite
working
out
having
some
problems
with
flannel,
say:
yeah,
that's
working.
D
Not
a
huge
amount
right
now,
I
had
other
unrelated
downstream
stuff
too,
that
needed
focusing
on
for
a
while,
but
I
need
to
get
back
to
the
the
ice
skating
and
Ganesha
URL
here
that
I
have
had
sitting
around
for
some
time.
I
did
whenever
it
was
a
couple
of
weeks
ago.
I
did
go
through
and
test
your
cache
inventory
cash
thingy,
and
that
works
fine,
that
that
pier
as
at
least
I've
tested
tested
that
against
deep
sea.
So
from
my
perspective,
that's
good
to
go
in
and.
D
Thus,
if
volume
drive
groups
PR
thanks
for
your
comment
on
that
about
the
rationale
for
putting
that
in
CFO,
Liam
I
think
that's
completely
correct
and
sensible.
The
only
other
place
you
could
possibly
think
of
doing
it
is
maybe
it
may
be
like
a
library,
but
then
all
the
all
the
orchestrators
would
have
to
import
that
so
I,
don't
really
see
the
sense,
I
think
so.
I
think
it's
kind
of
the
right
place
anyway.
D
I
haven't
finished,
reviewing
the
code
yet
so
I'll
make
some
comments
on
that
a
little
bit
later
and
yeah
I
think
that's
it
familiar
for
the
moment.
A
Yes,
have
a
look
at
that,
it
might
make
sense
to
add
drive
groups
to
the
safe
dashboard.
E
A
E
A
G
A
Anyway,
yeah
I'm
I'm
really
happy.
Okay,
then
thanks
for
your
time
and
then
next
meeting
us
on
Monday,
hey.
D
E
A
E
A
G
D
A
A
D
E
B
A
B
A
B
A
E
B
F
E
B
A
D
A
F
A
Inverters
and
maybe
a
tariff
environment,
or
something
like
that,
and
then
it's
actually
doable
to
do
that
to
test
your
core
castrato.
A
B
F
D
E
B
B
H
A
F
B
F
A
F
B
F
A
B
A
D
B
F
D
B
Test,
if
it
succeeds,
then
the
test
continues
and
the
this
work
unit
script
using
the
git
clone
operation
is
problematic.
Now
that
now
that
the
the
QA
it's
been
this
way
for
a
couple
years
now
that
the
QA
subtree
of
the
repo
is
integrated
into
the
main
stephanie
code
used
to
live
in
a
separate
bit,
Reaper
Theory
before
remember
that
and
called
CEFCU
a
suite
and
now
there's
no
separate
repo.
B
H
B
H
B
H
H
B
One
of
the
problems-
actually
that's
the
other
part
of
the
problem,
yeah
it
I
I,
wrote
a
long
comment
in
one
of
in
my
PR
that
but
the
answer,
the
short
answer
to
the
question
is,
for
example,
I'm
implementing
this
test.
That
runs
the
Ceph
dashboard
end-to-end
tests,
which
are
part
of
the
SAP
dashboard
code.
B
It
it
packages
those
into
an
RPM
together
with
a
script,
the
batch
script
and
and
then
I
instead
of
using
the
work
unit
task,
because
it
has
this
cylinder
blown
operation
in
it,
which
is
which
is
problematic,
I
I,
simply
install
the
RPM.
I
include
the
RPM
in
the
in
the
additional
packages
for
the
install
task,
though
its
installed
and
then
I
just
run
I
just
run.
The
script
basically
make
it
simple.
I
just
issued
issue
a
command
on
the
on
the
node.
B
H
Operation,
I,
don't
think
we're
not
going
to
do
that
with
I,
don't
think
like
unless
we
want
to
package
up
a
separate
like
QA
saying
that
doesn't
get
distributed
to
users
ever
right,
because
it's
all
just
like
internal
tests.
It's
all
internal
testing
scripts,
it's
not
something
that
anyone's
ever
gonna
run
on
their
own.
They
expect
to
like
have
a
cluster.
They
can
destroy.
B
H
Mean
I'm
not
fundamentally
against
anything
with
packaging,
because
I
never
have
enough
confidence
to
make
those
decisions.
I,
I'm
I,
don't
I,
don't
love
the
idea
of
my
nature,
though,
just
because
it's
all
like
packaging,
you
expected
installed
somewhere
and
distribute,
whereas
I
I
guess
this.
This
would
be
a
question.
I'd
be
more
worried
about
the
impact
on
the
infrastructure
and
what
like
the
people?
Who
do
that
again
care
about.
H
Packaging,
you
know
and
I
care
about
being
in
the
context
of
the
upstream
like
shamans
I,
guess
probably
it
would
be
a
very
small
package
in
one
really
at
any
load
but
I.
Don't
know
that
and
I
mean,
and
it
would
be
a
lot
more
one-time
use
packages
that
would
like
go
up
for
every
every
PR
and
then
disappear.
Every
equation
go
away.
H
H
F
B
Know
and
it
pulls
in
like
a
hundred
megabytes.
So,
however
much
you
know,
you
know
it,
maybe
maybe
the
the
clone
operation
could
maybe
there's
a
way
to
clone
just
the
subtree
of
the
clone
and
with.
E
B
Smaller
that
would
help
in
that
and
then
the
other,
the
other.
The
other
issue
is
that
in
our
infrastructure
we
packages
that
are
that
are
built
in
the
in
the
build
service,
and
so
we
don't
have
any
Shama,
and
you
know
that's
guaranteeing
that
the
sha-1
that
we
built
is
the
same
as
the
sha-1
that
we're
cloning
in
the
work
working
at
Texas.
That's
a.
B
H
H
Any
any
thoughts
or
updates
no
I
haven't
working
on
that
for
a
while.
I
would
like
for
it
to
happen
at
some
point:
hey
I'm,
not
sure
if
I'm
gonna
be
able
to
or
not
and
I
had
several
you
know
threads
on
the
mailing
list
about
it
and
they
just
concerned
that
gives
no
one's
ever
gonna
succeeded,
so
he's
been
putting
forward
in
alternate
strategies
like
building
work
tests
in
any
communities
and
just
doing
it
separately.
I
guess
we'll
see
how
it
turns
out.
Yeah.
B
And
that's
that's
another
another
thing
where
there
you
know
this
corpus
of
tests
that
assumed
that
they're
running
in
the
sepia
environment,
with
the
set
tasks
and
and
the
github
mirror,
and
it's
kind
of
part
of
that
and
then,
if
it
but
but
I
guess.
If,
if
we
you
know
if
a
book
is
gonna
be
deemed
fundamentally
incompatible
and
we'll
have
to
think
outside
of
the
box,
yeah.
H
I,
don't
think
I
said
anyone
relieved
once
it's
just
like
you
know.
People
have
over
the
years.
Several
times
said:
oh
we're
going
to
like
clean
it
up,
so
you
don't
have
to
use
the
staff
Tasker.
Then
no
one
ever
actually
finishes
it.
So
David's
gotten
tired
of
waiting,
isn't
sure
this
that
it
can
be
done.
B
B
G
G
G
Don't
know
it's
probably
critic
to
myself
rather
than
to
anybody
else,
but
I
think
I'm,
trying
to
like
swallowed
too
many
PRS,
and
my
hunch
is
that
we've
got
about
maybe
200
plus
PR
this
time
around
and
it's
very
difficult
to
like
pull
them
all
together
at
the
end.
So
essentially
what
happened
was
that
we
discovered
couple
of
issues.
One
is
related
to
defensible.
Another
is
to
I
think
something
like
on
G.
G
G
G
G
The
way
we
self
deploy
I'm,
not
sure,
if
you
guys
like
aware
of
that,
then
maybe
you
know
anybody.
We
don't
have
any
like
warm
body
responsible
for
self
deploy
and
it's
like
its
homeless,
and
you
know
we're
trying
to
hold
it
and
to
maintain
and
I
guess
like
on
one
hand,
people
don't
want
to
drop
it,
but
on
another
hand
it
doesn't
get
enough
love
and
that's
what
happens.
I.
G
G
E
B
B
G
B
B
B
A
What
would
you
call
it
as
I
say,
choker,
straighter,
classic.