►
From YouTube: 2021-10-19 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
october
19th
2021
rook
community
meeting.
So
let's
go
ahead
and
catch
up
on
upcoming
milestones
in
releases
1.6,
we're
not
sure
if
we'll
be
doing
another
patch
for
that
one.
So
there
likely
isn't
much
to
talk
about
on
the
board
here,
but
we
will
bring
it
up
real,
quick
yeah.
It
should
be
empty,
nothing,
nothing
to
talk
about
there.
It
looks
like
so.
A
B
Yeah
yeah,
I
took
a
look
at
the
back
ports.
There
have
been
a
couple
of
backboards
to
the
1.6
branch
in
the
last
like
month,
but
they
I
don't.
They
don't
warrant
a
new
release.
I
don't
think
or
nobody's
asked
for
those
upstream.
It
was
more
related
to
something
downstream.
We
were
doing
so
I'd,
say
yeah.
We
don't
need
an
upstream
release
unless
it's
requested
at
some
point
for
1.6.
C
B
D
C
A
C
B
B
A
Yeah
yep
well
that
sounds
good,
alright
cool
that
we
can
talk
about
then,
what's
upcoming
in
a
release
tomorrow
for
1.7,
which
is
the
latest
supported
release.
B
B
If
that
works
for
people-
and
I
didn't
get
a
chance
to
look
at
the
1.7
board
to
see
if
things
are
updated
there,
but,
let's
see
glad
it
seems
like
it's,
it's
just
things
in
progress
and
it's
just
kind
of
the
bi-weekly
release
with
nothing
real
urgent,
as
I
recall
for
tomorrow,
but
we'll
just
get
it
out.
There.
D
B
C
For
sure
I'm
finally
releasing
tomorrow,
but
if
at
any
point
in
time
we
have
to
release
in
your
pto,
I
guess
we
more
or
less
know
the
process
now,
because
it's
more
or
less
just
pushing
a
tag
anyways,
so
it
doesn't
really
require
much
but
yeah.
I
think
it
would
be
good
if
we
have
to
urgently
remove
something.
At
least
we
must
be
confident
how
we
do
releases
too.
D
B
Would
be
someone
it'd
be
nice
to
honestly
have
someone
closer
to
my
time
zone?
It
just
feels
like
everything's
happening
in
europe
and
then
we're
catching
up
on
that
in
the
us.
When
we
come
online
and
then,
after
things
finally
calm
down
around
here,
then
we
do
the
release
when
it's
a
quieter
afternoon.
A
lot.
That's
been
the
pattern
so
far,
not
that
it
needs
to
happen
that
way,
but
we
should.
We
should
just
get
a
few
of
us
on
a
call
and
do
a
release
together.
A
D
C
I
haven't
tried
the
new.
I
know
I
told
you
about
it,
but
now
get
up
has
does
the
releases.
Does
the
release
notes
for
you,
yeah.
D
C
D
A
D
A
So
yeah
github
is
really
coming
along
with
a
lot
of
a
lot
of
cool
features.
They're
they're
really
investing
and
moving
forward
things.
That's
pretty
cool.
A
Yeah
anything
else
on
on
1.7
or
let's
move
down
to
1.8,
coming
up
on
the
on
the
distance
horizon
still
more
than
a
month
away.
Yeah.
B
A
B
Support
those
are
the
big
ones
in
my
mind
that
and
then
there's
a
number
of
new
features
being
added,
as
always,.
E
Yeah,
I
think
we're
still
we're
still
rearranging
the
examples
right
that
the
example
manifests.
D
A
Yeah
speaking
of
community
topics,
jan
had
added
one
added
something
in
the
chat,
so
I'm
going
to
go
ahead
and
just
put
it
onto
the
agenda
here.
So
it's
oh
geez
that
didn't
copy.
D
A
There
we
go
all
right,
so
we
get
to
that
topic
today,
too.
Okay,
so
travis,
you
followed
up
on
operator,
kits
archiving
all
that
stuff
and
checking
dependencies
and
everything.
It's.
B
A
Okay,
that
sounds
good
and
then
so
bumps
got
a
couple
of
of
agenda
items
here.
Do
you
want
to
go
ahead
and
bring
those
up.
D
F
Is
published
and
we
can
we
can
use
the
tool
we
can
use
a
plugin
with
the
crew
command.
It
is
still
in
very
like
basic,
like
we
can
run
very
basic
chef
commands
and
we
have.
We
have
limited
that
to
our
staff
name
space
only
and
so
yeah,
but
basically
it
is
released
and
we
can
use
the
tool
with
crew.
F
I
need
more
feedback
on
that
and
I
will
be
working
on
the
issues
that
has
been
opened
like
to
to
make
the
tool
working
for
all
the
name
spaces.
So
yeah.
A
Do
one
thing
I
was,
I
was
thinking
about
with
the
great
work
on
this
too.
By
the
way.
One
thing
I
was
thinking
about
is
that
do
we
have
enough
insights
into
you
know
like
the
quick
start,
like
the
very
typical
super
easy
you
know,
install
with
the
manifests
that
we,
you
know
have
published
the
examples.
Do
we
or
is
that
common
enough
for
people
just
trying
it
out,
maybe
like
as
a
you
know,
sampled
like
getting
started
test
workflow?
A
Is
that
common
enough
to
like
add
a
command
for
that
so
like
cube,
control,
cube
control,
rook
ceph
install
and
it
basically
just
applies
demand
like
the
example
manifests.
Is
that
common
enough,
or
is
that
you
need
more
config
than
that
to
really
to
actually
be
successful?.
F
B
Well,
I
think
the
in
my
mind
this
tool.
We
want
to
be
more
of
a
troubleshooting
tool
and
kind
of
assist
you
in
things,
but
I'm
not
sure
we
in
my
mind
we
would
use
it
as
the
configuration
tool
to
update
crds
and
things.
I
think
we
would
it's
more
of
running
toolbox
commands
and
make
some
things
some
commands
simpler
if
we're
working
with
rook,
but
I
don't
know,
maybe
eventually
it
could
be
more
of
a
also
simplify
the
quick
start
example
sort
of
thing
that
should
be
cool
with
that.
If
we
wanted
maybe.
A
Yeah,
that's
totally
reasonable
if
that's
like
the
scope
that
we
want
to
keep
for
it.
That's
that's
not
like
a
big
deal,
just
kind
of
thinking
of
if
it,
because
at
the
end
of
the
day,
you
don't
want
it
to
result
in
a
a
experience,
that's
going
to
lead,
you
know
more
people
than
not
down
the
wrong
path.
A
Also
right,
if
there's
like
you
know
a
lot
of
config
values
that
people
need
to
change,
or
you
know
make
specific
for
their
environments
and
stuff
like
that,
so
you
wouldn't
want
to
end
up
with
if
it
does
streamline
things
and
make
it
like
hey
cool
like
I
just
have
to
install
the
rook
plug-in,
and
then
I
just
create
a
cluster
with
two
commands.
I'm
done
that
that
would
be
interesting,
but
it's
you
know
also
not
like
it's
not
really
a
production
sort
of
thing
and
you
know
going
down
the
path
of
setting.
B
A
All
right
and
then
updates
on
on
snick.
D
B
D
B
F
F
The
last
comment
I
had
if
you
open
that
link,
I
have
tested
on
my
fork
and
it
is
running
and
if
you
just
click
on
the
security
just
to
have
more
details,
so
yeah
see
like
the
tests
are
running
and
it
is
green.
So,
but
I
don't
know,
I
am
facing
some
weird
issues
with
rook
rook
paper
and
the
issue
is
coming
from
like
related
to
some
environment
vehicle
issues.
So
I
I
I
will
have
to
look
on
that.
B
F
C
C
D
C
F
I
just
copied
the
changes
in
my
fork
and
boost.
I
think
you
can
open
the
changes
here
from
here
too.
C
A
A
C
A
Okay
and
then
last
one
here,
for
you
is
the
bomb,
updated
controller,
runtime.
F
So
the
changes
are
there
like.
I
have
the
pr,
but
one
of
the
unit
test
is
failing
with
you
in
64
like
deep
copy
added
with
ui64,
and
I
tried
to
change
all
the
ui
64
we
are
using
in
our
code
to
in
64,
then
also
I
am
getting
the
same
test
like
the
same
unit
test
is
failing,
so
I
don't
know
why.
F
C
Did
you
capture
the
error
once
you
have
moved
to
using
n64
instead
of
viewing
64.
F
D
A
A
Okay
and
then
blaine,
I
see
you
typing
right
now,
so
I
don't
want
to
interrupt
you.
Do
you
have
an
update
also
on
the
helm,
chart
refactoring.
E
Oh
yeah,
I
mean
it
is
proceeding
well
sort
of
slowly.
E
It's
mostly
it's
not
my
highest
priority
item
the
yeah
here,
I'm
just
trying
to
cut
down
on
the
amount
of
like
duplicating
the
same
stuff
and
the
eventual
goal
is
to
really
stop
duplicating
the
same
stuff
and
use
the
helm
chart
to
generate
all
of
the
crds
that
we
have
in
our
example,
common.yaml
manifest,
so
that
we
don't
have
any
issues
of
like
skew
between
like
one
installation
method
and
another,
and
this
is
also
going
to
help
us
in
the
future,
doing
some
because
audits
of
the
the
rbac
to
make
sure
that
we
aren't.
A
And
since
you've
been
typing
in
providing
a
response
there
in
the
agenda,
doc
blade
do
you
want
to
go
ahead
and
verbally
kind
of
chat
through
what
you've
been
typing.
E
Yeah
I
haven't
looked
at
storage
os
or
open
ebs
in
a
while
to
do
a
comparison,
but
I
have
started
looking
at
their
documentation.
E
E
But
other
than
that,
both
of
them
like
act
in
a
in
a
different
space
than
rook
entirely
rook
is,
or
I
should
say,
rook
with
seth
specifically,
is
focused
on
actually
using
ceph
to
create
storage
that
kubernetes
applications
can
consume
rather
than
using
existing
persistent
volumes
or
potentially
just
raw
raw
disks
that
aren't
replicated
like
ceph
is
a
an
enterprise.
E
Rook
just
runs
ceph
in
order
to
give
like
the
kubernetes
cluster
the
option
of
having
a
like
a
way
to
have
pvcs
for
storage.
That
is
actually
replicated.
Storage
like
you
would
find
in
a
cloud
provider,
but
you
know
like
that,
is
actually
replicated
and
actually
has
performance
characteristics
that
you
can
rely
on
like
you
would
have
in
a
cloud
provider,
but
you
may
not
be
in
a
cloud
provider
or
alternately,
not
all
cloud
providers
or
cloud
provider
regions
provide
the
all
of
the
storage
you
want.
Some
regions
might
not
have
object.
Storage,
for
example.
E
I
know
I've
heard
users
complain
about
and
that's
one
of
the
reasons
they
use
rook.
E
E
E
And
yeah,
what
rook
does
not
do
that?
What
storage,
os
and
open
ebs
seem
to
do
is
consume
existing
cloud
provider
pvcs
and
expose
them
like
in
a
like
sort
of
consistent,
consistent
way,
given
whatever
their
semantics
are,
I
could
imagine
a
scenario
where
someone
wanted
to
have
rook
and
then
use
storage
os
or
I
guess
it's
not
on
dat
or
open
ebs
on
top
of
it.
B
E
B
E
E
E
Yeah
so
rook
can
consume
cloud
provider
volumes,
but
it
would
not
be
a
direct
pass-through.
It
actually
would
then
go
and
make
sure
that
whatever
was
on
those
volumes
was
actually
replicated
and
like
it
could
tier
them
to
use
like
fast
storage
for.
G
There's
actually
a
big
a
big
one:
storage
os
is
closed,
source.
E
G
Yeah,
so
they
have
multiple
storage
engines.
Okay,
the
the
the
common
one
they
that
you'll
see
used
is
open.
Zfs,
okay
and
you
know-
that's-
got
snapshots
and
rating
and.
D
G
Data
protection
storage
address
is
more
like
running.
The
back
end
is
more
like
iscsi
and
and
raid
traditional
raid.
G
G
G
Or
a
block
pvc
right
and
then
file
systems
around
top
of
that.
So
in
the
background
they
do
have
that
the
more
interesting
one
is
they
have
a
a
maya
store
which
is
an
nvme
over
fabric
back
end.
G
D
C
Yeah
right
now
we,
the
step
csi
driver,
just
uses
the
kernel
mount
the
kernel
rvd,
which
is
obviously
more
native,
but
iscsi
has
some
fronts.
I
guess,
for
instance,
windows,
nodes.
A
Yeah
thanks
for
all
that,
all
those
all
that
insight
richard
that's,
definitely
useful
to
share
here
and
have
a
kind
of
add
to
the
knowledge
base
here.
Anything
else
that
folks
want
to
add
to
this
particular
topic.
A
D
A
That
needs
to
be
done
for
for
risk
5,
I'm
not
super
familiar
with
the
requirements
of
that
arc.
D
C
E
A
D
A
If
that's
everything
it
looks
like,
we
have
an
action
item
for
folks
who
want
to
get
involved
in
running
the
release
tomorrow,
for
what
is
that
the
latest
1.7,
I
think,
but
yeah.
So
I
think
that's
that's
everything
so
thanks
everybody
for
joining
in
today.
Thanks
for
the
good
questions
and
the
the
responses
as
well
too
richard.