►
From YouTube: 2022-08-09 Rook Community 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
All
right
the
recording
has
started-
and
this
is
the
august
9th
2022
rook
community
meeting.
So
we
will
hop
on
into
our
milestone,
checkups
and
milestone
planning
for
the
1.8
branch.
We
don't
have
any
particular
upcoming
plans
patch
release,
so
I
think
we
can
maybe
just
skip
right
on
over
that
for
1.9
branch.
We
had
a
release
back
on
july
26
that
is
captured
here
and
folks
can
catch
up
on
that,
and
then
we
have
the
1.9.9
coming
up
later
on
this
week.
B
Yeah,
the
the
patch
release
plan
for
this
30
days,
really
just
the
kind
of
on
the
cadence.
This
will
be
about
two
and
a
half
weeks.
I
guess
I
don't
know
of
any
specific
issues
to
discuss
really
for
1.9.
We
do
have
an
nfs
test
that
we
need
to
get
re-enabled.
B
Our
rackshift
is
working
on
that
and
then
the
other
two
in
the
to
do
column.
I
was
looking
at
them
they're
just
kind
of
not
urgent
for
1.9
just
nice
to
have
if
we
can
track
them
down
and
get
them
back.
A
All
right
sounds
good
travis
thanks
for
the
1.9
update
there
and
then,
let's
hop
into
the
big
one
here,
the
the
next.
I
always
want
to
say
major
release,
because
it's
like
a
meaningful
one,
but
technically
the
semantic
version
means
a
minor
release.
So
the
next
minor,
the
1.10
release
that
one
is
coming
up
over
the
next
couple
of
weeks
as
well.
So
let's,
let's
dive
into
that
one
since
that
one's
the
next
important
release
coming
up.
A
B
Yep
yeah,
the
first
one
that
I
put
in
the
blocking
of
these
called
hoping
to
get,
go
1.19
support
out,
and
our
pattern
has
been
to
support
the
most
recent
two
versions
of
go
so
1.18
and
1.19
is
what
the
p
are
there
in
progress
is
getting
close
to.
The
only
failure
left
is
it's
not
building
on
mac,
which
is
unusual.
B
I
don't
think
we've
ever
had
that
unique
failure
before
so,
hopefully
we'll
get
that
result
soon,
get
that
in
and
then
the
second
one
is
that
you
know
if
shabon's
working
through
getting
the
step
commands
working
with
a
rook
operator's,
sefax
key
reading
instead
of
client.admin,
I
think
we're
getting
close.
It's
just
well,
we
won't
know
until
it's
done
and
and
if
it
ends
up
not
making
the
timeline,
I'm
comfortable
with
delaying
it
till
1.11,
but
still
trying
to
get
in
for
now.
A
And
what's
the
scope
of
this
one
travis
is
this
like,
basically,
all
of
the
operations
that
you
know
the
operator
will
perform
for
like
provisioning
and
bootstrapping
and
all
sorts
of
stuff?
Or
what's
the
scope
of
this.
B
So
any
operation
that
the
operator
does
the
idea
is
it
would
use
this
new
key
ring
so
that
in
the
log
we
can
see
who's
that
the
operator
is
doing.
These
operations,
like
we'd,
have
another
key
ring
for
the
toolbox.
Whatever
operation
that
performs
we'll
see,
oh
the
toolbox
did
it,
and
if
it's
the
csi,
the
csi
driver
will
not
be
affected
it.
It
has
its
own
key
rings
that
it's
provisioning
volumes
with
we're,
not
touching
that
for
now.
B
A
Okay
yeah,
so
it's
more
for
like
auditing
and,
like
you
know,
understanding
like
the
who's
performing
the
operations,
as
opposed
to
reducing
the
permissions
that
the
the
rook
operator
has.
B
A
That's
very
reasonable
cool.
Okay,
I
mean
also
you're,
saying
you
also
feel
like
we'll
see
where
this
gets
to
in
terms
of
delivering
on
it.
Do
you
think
that
that's,
but
still
blocking
the
release,
though,
or
you
know
like
we,
wouldn't
hold
the
release
for
it.
B
A
B
I'd
say:
technically:
it
could
wait.
A
Okay,
so
got
it.
That's
understood,
understood
all
right,
any
issues
that
are
in
progress
that
needs
some
follow-ups
or
need
some
decisions
made
or
support.
B
From
my
perspective,
we're
fine
and
if
anybody
thinks
1.10
should
be
blocked
on
any
of
these
other
issues,
let's
get
them
in
the
blocking
release.
Column
too,
but
generally,
I
think
the
other
issues
could
come
in
a
patch
released
if
needed.
B
A
B
A
Nice,
so
then
do
you
think
that,
like
in
terms
of
cutting
up
the
beta
one
here,
you
know
doing
the
release
branch
and
going
along
with
the
release
candidates
here.
This
then,
is
expected
to
include
this
bump
here.
So
it's
gone
the
day
after
so
we'll
get
that
csi
3.7.
You
know
dependency
and
have
that
as
part
of
the
release
candidate
here.
B
A
Yeah
cool
that
sounds
fine
to
me
all
right,
so
we've
got
the
got
the
timeline
here
for
getting
the
1.10
release
out.
We've
got
a
plan
here
for
what
we'll
do
with
this
dependency
on
this
csi
plugin
update,
we'll
keep
on
executing
on
that
all
right
sweet.
So
we
can
move
on
then
from
milestones
and
releases
and
move
on
to
the
community
topics.
A
So
I
will
click
on
this
issue
here
about
the
set
manager
consistently
crashing.
B
A
Got
it
and
what's
travis
with
the
latest
1.9
release,
what
is
what
is
the
version
of
seth
seth
that
will
be
installed
by
default
or
use
by
default.
B
So,
unless,
unless
seth
has
another
patch
release,
it'll
be
1723.,
okay,
perfect
right,
yeah
and
then
yeah
as
ceph
releases,
its
patch
releases.
We
just
make
that
the
default
generally
for
the
latest
rook
release.
A
Yep,
that's
great
all
right
so
17.2.3,
which
will
be
functional,
perfect
all
right
and
that's
good
thanks
for
the
follow
up
on
that
from
the
last
meeting,
then
all
right
and
then,
let's
move
on
to
our
our
plug-in
our
control
plug-in
for
using
the
package
manager.
B
C
Yeah
sure
so
we
are
basically
adding
this
cuba
city
roxaf
health
command,
which
will
basically
get
the
status
of
overall
cluster,
for
example
like
what
is
the
calm
state
and
if
we
have
at
least
three
months
running
on
different
node
and
similar
for
the
osgs
and
also
we'll
check,
if,
like
the
pgs,
are
active
and
clean
and
or
not,
and
we
have
at
least
one
manager,
port
up
and
down.
We
are
like
in
future.
C
We
are
planning
to
add
more,
like
commands,
more
checks
like
that
inside
the
health
command.
So,
but
for
now
we
have
this
checks
only,
and
we
have
this
nice
cool
coloring
thing
which
will
give
user
idea
like
how
severe
their
like
warnings
or
adder
is.
If
it
is
a
green
or
blue,
it
is
like
kind
of
flogging
and
healthy
and
warning
is
kind
of
like
you
should
look
at
those
adders
android
means
cluster
is.
C
Yeah
so
once
this
is
like,
we
are
waiting
for
the
next
pr.
That
path
is
working
on
and
once
that
is
released,
we
will
have
a
new
plugin
release
and
then
we
will
add
the
documentation
on
the
rook
repo.
That
okay,
please
add
the
logs
from
this
and
like
this
command,
will
help
them
develop.
A
This
does
the
output
is
there?
Is
there
support
for,
like
a
you,
know,
machine
parcel
machine,
readable
output
like
a
json
or
output
from
from
this,
since,
and
also.
C
Yeah
we
can
have
a
json
output
too.
So,
if
you
so,
but
in
the
health
command,
we
don't
have
any
self
commands
or
any
check
that
prints
in
json
format.
But
if
we
have
something
we
can
add
that.
B
Yeah,
we
will
speak
on
an
interesting
question,
though,
because
we've
got
all
sorts
of
commands
printing
out
status
here
I
mean
you'd
end
up
with
a
bunch
of
different
json
blogs
if
we
output
in
json
so
anyway,
something
to
think
about.
But
for
now
I'd
say
it's
just
bash
person
cool
got
it
okay
and
then.
B
A
All
right
and
then
here's
another
feature
that
we
wanted
to
highlight
as
well,
for
the
plug-in.
B
D
Yeah
so
yeah
we
are
like
trying
to
add
a
debug
mode
for
the
rook
operator
and
also
launching
like
the
audio
operator.
So
what
does
debug
mode
does
is
like
whenever,
like
the
cluster
is
in
the
bed
state
where
the
monster
and
osd's
are
in
the
crash
loop
back
of
state
or
like
there
are
some
network
issues
or
some
thing,
and
we
are
we
want
like.
So
we
can
try
to
debug
the
cluster.
D
So
there
is
a
debug
command
that
has
been
added
with
this
crew
group
plugin
and
we
just
go
and
run
that
debug
command
and
debug
deployment
will
be
created
for
the
particular
mon
or
osd
that
we
have
passed
and
user
can
use
the
theft
tools
on
that
debug
deployment
like
there
are
various
theft
tools
that
we
have
the
self-object
store
tool
or
the
sub
blue
stores
tool.
So
basically,
what
does
these
tools
is
like?
D
D
If
we
say
it
in
easy
terms,
yeah,
it
also
does
some
more
hacks
yeah
like
like
stop
reconciling
over
the
particular
parts
like
like
first
currently,
if
we
are
using
these
theft
tools
directly,
we
need
to
make
the
operator
down,
but
in
this
case
we
just
go
and
skip
their
concerns
over
these
parts
that
we
are
there
and
the
other
reconciliation
can
be
done
as
it
is.
B
Yeah
I'll
just
highlight
that
this
will
be
really
nice.
It's
been
a
common
request
really
for
a
long
time.
As
far
as
hey,
I
need
to
run
this
osd
object
store
tool
on
the
osd
while
it's
down
well,
how
do
you
do
that
in
the
pod,
so
the
pod
is
running
with
the
osd,
but
this
yeah
this
just
creates
a
pod
exactly
for
that
purpose,
scales
down
the
original
one
and
creep
into
them,
but
anyway,
good
good
stuff.
A
Yeah
thanks
for
sharing
that
berth
awesome,
yep
thanks
all
right!
So
that's
everything
I
think
for
the
the
q
control
plug-in,
and
so
let's
move
on
to
kubecon
about
any
update
for
for
you,
red
hat
folks
being
able
to
be
in
attendance.
B
Yeah
I'll
start
with
that,
so
for
kubecon
travel
policy
allows
us
to
be
there
and
budget
and
everything
if
we're
speaking.
So
that's
not
the
challenge.
There's
other
challenges.
Something
came
up
for
me.
I'm
not
able
to
make
it.
Unfortunately,
so
I
think
others
can
speak
if
they
can
make
it.
But
if
we
don't
get
sufficient
speaker
attendance,
we'll
just
do
virtual
one
more
time
and
then
hopefully
the
next
one
we
can
make
it
work
where's.
I
think
the
next
one
is
amsterdam.
B
A
Got
it
well
yeah,
I
I
should
be
there
myself,
like
I've,
got
a
maintainer
track
talk
for
for
the
crossband
project,
so
I
should
be
in
detroit
as
well.
A
Awesome
and
then
do
we,
so
we
need
to
do.
We
need
to
send
an
update,
then
or
finalize
the
response
about
how
we
want
to
do
the
talk.
If
you
want
to
do
it
virtual
or
in
person
that
needs
to
be
finalized
with
with
the
planners
right
yeah,
it.
B
A
B
A
Awesome
all
right,
so
that's
everything
that
was
on
the
agenda
items.
Then
anybody
else
have
anything
on
that
they
want
to
add.
Before
we
adjourn.
C
Just
a
quick
update
on
the
pr
that
I
I
was
working
on,
that
was
in
the
blocking
section
of
conduct
night.
I
was
just
testing
locally
and
now
it
is
working
fine.
I
pressed
the
output
like
if
maybe
like
someone
can
get
some
idea
from
the
output,
so
I'm
using
the
client
dot
loop
operator,
clearing
and
not.
C
Yes,
so
I
first
had
the
cluster
with
master
version,
and
then
I
changed
the
version
to
my
local
and
yeah.
It
is.
A
All
right:
well
then,
if
that's
everything,
then
we
can
go
ahead
and
adjourn
for
the
week
and
follow
up
with
each
other
on
slack.