►
From YouTube: 2018-07-31 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
A
B
A
Okay,
I
guess
that
must
be
a
new
thing
that
he
created
yeah.
We
are
not.
We
have
not
been
using
that
in
the
past,
nor
are
we
using
it
now
consistently.
So
maybe
we
can
have
a
little
discussion
about
that.
But
let's
talk
about
the
the
issues
that
we
have
here
to
finish
off
and
get
0.8.1
release
out
my
dog's
choking
on
something
over
here:
sorry:
okay!
So
let's
go
ahead
and
start
talking
about
1919,
then
mm-hmm.
B
Got
this
one,
so
thanks
for
the
feedback
last
night,
just
making
a
few
changes
based
on
that
and
then
that
I
should
be
in
and
hopefully
backboarded
before
long
here
in
the
next
few
hours
and
yeah
I,
don't
think,
there's
anything
else
open!
There
just
needs
a
lot
of
testing
every
time.
I
touch
it
as
far
as
upgrades
and
anyway
starting
the
OS
DS
reliably
is.
A
B
A
B
A
B
A
Yeah
we
do
not
have
an
integration
test
for
upgrading
across
versions.
I
didn't
I
did
not
expect
us
to
right
now.
I
am
happy
with
the
progress
that
our
upgrade
support
has
been
making
so
far.
You
know
we
have
continued
to
make
strides
towards
it
a
fully
automated
solution.
So
that's
that's
good
progress.
A
A
Interesting
yeah,
yeah
I
had
I
had
thought
fairly
confidently
that
the
only
time
and
orchestration
would
get
kicked
off.
There
was
if
they
knowed
the
set
of
nodes
changed.
No
one
node
was
added.
Remove.
So
that's
interesting
to
learn
that
here.
Do
you
have
any
concerns,
then
about
deployments
being
updated
too
frequently
or
when
we
don't
want
that
behavior
to
happen
or
because
of
the
you
know,
kubernetes
behavior,
of
only
really
kind
of
updating
the
deployment
object
when
something's
actually
changed.
You
know
when,
within
the
cabinet
api
that
that's
not
a
concern.
B
I
think
it's
not
a
concern
I
think
initially,
when
doing
this
automatic
upgrade
I
was
thinking.
Well,
it's
only
do
this
specific
version
if
it's
from
ODOT
800
just
to
make
sure
we're
doing
we're
not
too
aggressive,
but
after
our
discussion
it
I
feel
better
about
yes,
because
kubernetes
doesn't
upgrade
the
pod
spec
unless
it
actually
changed.
It
won't
get
restarted
too
often
and
updating
the
CRD
is
not
a
common
operation.
A
Don't
I,
don't
particularly
think
I
do
either
with
the
observations
that
we've
made
from
this
week
with
testing
and
trying
it
out.
I
do
so.
That's
that's
kind
of
another
example,
though
of
you
know,
we've
made
strides
or
progress
towards
automatic
upgrades.
We've
also
made
strides
in
progress
along
the
way
here
with
supporting
you
know,
dynamic
cluster
updates
to
write
when
properties
change.
The
operator
makes
that
desired
state
of
reality,
and
so
we've
made
progress
on
both
of
those
endeavors,
even
though
neither
one
is
officially
supported
right.
A
A
Sorry
folks,
hey
Alexander,
relax,
yes,
I
was
pointing
at
this
here
and
moving
my
mouse
around
gesturing
thusly
and
not
showing
that
to
anybody.
So
yes,
this
issue
here
is
in
the
same
1945
pull
request.
Yes,
okay,
that's
cool!
So
both
of
these
are
covered
here
and
then
this
is
also
in
review
about
improving
our
the
Flex
volume.
A
Documentation
and
I
saw
some
back
and
forth
from
from
you
and
from
Travis
and
now
and
on
that
and
I
want
to
take
a
quick
look
at
this
as
well
Alexander
right
after
this
meeting,
I
believe
and
but
it
looked
like
it
was,
it
was
a
good
change.
So
far
so
I,
don't
think
I,
don't
think,
there's
anything
they
were
like
worried
much
about
here.
A
C
B
A
B
A
D
D
Basically,
the
issue
is
saying:
is
the
bonds
were
being
mounted
without
the
group
set
to
the
group
in
the
security
context,
and
not
group
writable,
so
I
submitted
a
PR
to
shown
it,
but
then
I
think
rude.
Fs
pointed
out
that
flex
volumes
do
that
already.
The
last
like,
if
you
like,
follow
the
code
flex
volumes
after
it
calls
the
actual
device
plug
in
it.
D
It
Jones
inch
mods
the
the
directory,
so
I
was
playing
around
with
it
a
little
bit
more
and
I
found
that
on
my
system,
if
I
set
it
to
I,
don't
know
you
guys
are
probably
are
familiar
to
your
project.
There
is
like
a
mountain
device
which
mounts
the
volume
to
somewhere.
That
rip
knows
about,
and
then
there's
the
mat
mount
command,
which
then,
by
mount
that
directory
to
where
kubernetes
expects
it
to
be
to
mount
it
into
the
pod
or
two
for
the
pod
to
look
for
it.
D
If
I
do
it
Jone
inch
mod
on
it
in
the
mount
device
function,
then
the
end
permissions
are
what
I
expected.
But
if
I
do
it
in
the
mount
one
where
it's
at,
which
is
where
it
by
mounts
it
in
their
permissions,
don't
take
effect
and
obviously
because
because
the
flex
volume
is
doing
the
mod
and
shown
after
the
mount
is
called
I
assume
that
it's
hitting
the
same
thing
on
my
system.
But
then
route
FS
spun
out
the
local
environment,
which
did
not
reproduce
this
so
I've
been
trying
to
get
I
was
using
open.
D
Three
nine
he's
his
environment,
310
I've
been
trying
to
get
a
310
cluster
spun
off
to
see
if
something
weird
there,
but
it
makes
literally
no
sense
to
me
because
I
can
do
it
from
the
CLI
fine
after
the
bond
mount
and
the
and
the
permissions
take
effect
and
the
go
function
is
literally
just
doing
syscalls
so
I'm
doing
it.
I'm
gonna,
see
if
it
reproduces
on
my
system
with
wind
pretends
installed,
but
I
am
totally
lost.
I'm
like
what
the
possible
cause
of
that
to
be.
D
B
D
D
B
C
B
D
B
D
Yeah,
but
once
once
we
figure
out,
what's
going
on,
the
PRF
should
be
fairly
minimal.
We
just
have
to
make
sure
that
we
have
to
have
to
do
whatever
is
required
to
make
it
so
that
the
work
that
kubernetes
is
doing
actually
works,
but,
like
the
so
I
don't
know
if
waiting
herbs,
I
don't
know
we
can
before
that,
but
ignore
the
poor
request.
I
have
open
right
now,
it's
totally
useless.
D
D
I'm
not
sure
if
it's
a
race
condition
issue,
then
it
probably
would
affect
anything
that
has
the
security
context
enabled
I
would
imagine
if
you
have
a
security
context
that
requires
you
to
run
as
a
specific
user
and
then
email
for
things,
not
group,
writable
and
not
owned
by
group,
then
or
not
with
the
proper
group
set
I.
Imagine
it
would
affect
anything
I
think,
there's,
probably
not
that
many
kubernetes
installations
that
have
that
stuff
enable
at
the
moment
but
I,
don't
think,
there's
anything
like
super
specific
to
open
ship
that
should
cause.
C
Are
you
sure
that
it's
aroma
see
your
sister,
because
I
think
on
a
normal
test
that
I
can
without
a
problem
set
FS
group
which
I'm
not
sure
if
it
works
on
block
search,
but
at
least
it's
working
on
file
system,
so
yeah
I?
Think
if
you
guys
run
us
useless
you
would
you
still
need
to
manually
set
FS
group?
It's
not
like
copies
well.
C
C
D
C
B
A
A
Thanks
for
your
efforts
on
this
Fabian
sounds
like
it's
I
hope
you
have
some
hair
left
after.
A
B
A
B
A
C
A
A
A
C
C
A
C
A
A
Okay,
so
that
is
closed
and
done
and
I
trust
you
Travis
this
time
this
time,
alright,
so
that
is
everything
from
0.8.1
I
think
we
discussed
all
the
issues
and
we
have
a
plan
for
all
of
them.
So
let's
keep
following
up
on
the
ones
that
are
in
review,
get
those
all
merged
and
backported,
and
then
we
will
reconvene
for
making
sure
that
we're
ready
for
the
release
process
with
the
release
criteria
that
we've
defined
recently,
it's
not
good
all
right,
I
think
we
have
an
item
for
the
0.9
themes
and
priorities
later
on.
A
C
This
kind
of
sympathy
question
on
what
convenience
version
we
support
because
kind
of
what
bassam
said,
which
we
support
the
last
two
versions
that
client
go
support
and
the
problem
with
clint
garrison
is
kind
of
supporting
like
every
cuba
newest
version,
except
like
one
or
two.
So
I'm
not
sure
if
you
want
to
do
that.
C
A
I
think,
historically,
that
was
what
we
had
started
with,
is
maintaining
the
same
compatibility
matrix
that
client
code
does,
but
I
think
we
have
diverged
from
that,
and
I
we're
not
gonna
go
back
on
on
that
divergence.
We
are
not
going
to
re-implement.
You
know
third-party
resources
and
stuff
for
kubernetes
one
about
six
in
previous
we're.
Definitely
not
going
to
do
that.
So
do
is,
does
anybody
know?
Do
we
I
have
an
official
statement
anywhere
in
our
documentation
or
I?
Read
me
or
anything
that
says
what
specific
versions
of
Cabrini's
we
support
right.
A
C
A
Yeah
so
I
think
then
so
far.
If
our
documentation
says
that
we
support
1.7
and
newer,
which
is
true-
and
we
don't
say
anything
about
matching
client
go
formally,
then
I
think
we're
probably
good
here,
and
we
could
just
stop
saying
that
we,
you
know
in
an
informal
context,
that
we
support
the
same
matrix
because
we
do
not,
and
we
should
I,
don't
think
we
should
either.
C
C
B
A
C
May
be
putting
it
like
that,
but
if
most
people,
you
know
I'm
not
sure
when
the
survey
was
from
company's
version,
but
I
say
at
the
time
of
the
survey
there
was
like
1.9
out
and
if
they're,
like
you
I'd,
say
10%
on
one,
that's
one
of
eight
or
something
or
one
of
them,
I
wanna,
submit
and
the
rest
is
kind
of
on
holiday.
Don't
109!
There
isn't,
like
you
know,
I,
don't
see
a
reason
why
somebody
wouldn't
keep
the
kubernetes
up
to
date.
A
C
A
So,
for
me,
I,
don't
know
if
that's
I
wouldn't
say
that
that's
gonna
new
miracle
feel
like
range
I.
Don't
think
it
would
should
be
the
policy
that
we
said
I
think
it
needs
to
be
taken
into
account.
You
know
what
the
specific
features
are
and
what
the
user
needs
are,
but
I
think
for
sure
that
are
the
policy
should
be
that
when
there's
a
new
version
released,
we
support
it
right
away.
A
A
Wouldn't
we
take
those
dependencies
on
new
features
of
kubernetes
that
results
in
dropping
support
for
older
versions,
so
I
wouldn't
be
I
would
not
be
comfortable
with
saying
the
latest
three
always
that
that
does
not
make
me
comfortable
and
think
it
needs
to
be
in
a
more
case-by-case
basis,
depending
on
what
the
features
and
dependencies
are.
If
that
makes
sense,
sure.
A
A
Yeah
that
could
be
fair,
but
yeah.
Your
your
progression
there
from
three
to
four
to
five
while
you're
talking
about
it
there
it
is
you
know,
makes
me
that
kind
of
reinforces
my
thinking
about
that.
A
numerical
range
doesn't
make
sense,
you
know
a
hard
number
range.
Does
it
make
sense
to
me?
I
think
it
needs
to
be
about
what
features
came
in
what
versions
and
what
our
users
are
using.
B
C
A
Yeah
I
think
what
Blaine
I
believe
it
was
was
saying
of.
You
know
the
minimum
number
of
versions
that
we
support,
that
that
sounds
reasonable
to
the
I.
Think
that
you
know
when
we,
the
you
know
the
general
principles
in
this
discussion
here
we
can
kind
of
agree
upon,
but
then
I
think
when
it
comes
time
to
actually
remove
support
for
something
that's
gonna
have
to
be
a
specific
conversation,
then
we'll
have
in
the
future.
If
you
know
why
I
would
be
removing
1.7
what's
driving
us
to
do
that,
you
know
who?
A
A
E
What
like
I
can
read
the
code
and
I
can
but
having
some
documentation
that
talks
more
high-level
about,
what's
going
on
with,
would
be
really
useful
and
I
think
like
there
are
some
go
files
that
I
saw
that
had
go
doc,
style
comments
above
functions
or
methods
or
structs
or
whatever
that
were
we're
helpful
in
navigating
the
code.
I
think
that's
something
that
we
should
consider
making
part
of
future
editions
and
future.
A
Think
in
general
blame
that
you
know
I
would
definitely
be
very
supportive
of
that.
You
know
the
the
general
desire
to
have
new
contribute
contributors
or
developers
come
into
the
codebase
and
be
able
to
kind
of
understand.
You
know
how
things
are
connected
or
why
things
are
done
by
looking
at
the
code
itself
and
the
documentation
for
those
functions
and
such
like
I.
Definitely,
you
know
believe
in
that
very
very
heartily
we
had
done.
A
B
A
E
Yeah
I
I
definitely
realized
that
what
is
pre-existing
it
would
be
a
pretty
pretty
huge
effort
to
go
and
and
generate
go
doc.
Comments
for
everything
that
does
exist,
but
making
some
effort
to
make
sure
that
things
exist
for
for
new
things
and
for
things
that
are
getting
modified.
A
fair
bit
is
I
think
a
good
way
to
make
a
good
a
good
effort
for
the
future
without
slowing
down
too
much
and
and
doing
something.
That
would
be
a
lot
of
work
now.
Yeah.
A
Iii
definitely
support
that
blame
under
other
people
can
chime
in
if
they
have
strong
opinions
on
that.
But
otherwise
you
know
if
we
do
or
for
all
in
agreement
on
that
here,
then
you
know
opening
if
you
want
to
go
ahead,
Blaine
and
open
a
pull
request
to
add.
That's
that
an
item
there
to
the
PR
checklist
template.
That's
under
I
think
that
under
the
github
folder
in
the
Brooke
repo,
then
that
would
be
a
welcome
PR.
Unless
anybody
else
disagrees.
That
sounds.
B
A
A
In
Travis,
but
what
happened
to
you
and
I
man?
We
used
to
be
like
this.
Our
c-sharp
days
are
hardcore
style
cough.
You
know,
believers
like
we'd
like
we
were.
We
used
to
be
a
hundred
percent
compliant
with
all
that
stuff,
and
then
we
got
to
meet
switched
over
to
go
Lang
and
we've
become
lazy,
slobs
well,.
A
B
So
let
the
last
meeting
we
talked
about
potentially
closing
issues
automatically
that
are
stale
and
so
I
found
this.
You
know
this
bot
that
we
can
enable
it's
pretty
straightforward,
has
a
couple
of
configuration
options,
but
unless
anybody
is
in
disagreement,
it
sounds
like
we're
her
based
on
our
last
discussion,
we're
ready
to
go
ahead
and
enable
this.
This
bot,
which
basically
means
after
90
days,
issues
that
haven't
been
touched,
no
comments
or
anything
would
be
closed.
If
someone,
if
there's
a
need
to
keep
it
alive,
there's
a
couple
of
key
words.
B
A
Yeah
of
all
that
Travis
I
do
have
two
quick
questions,
though
one
is
G.
What
do
you
know
the
behavior
of
the
this
bot,
the
first
time
it
gets
run?
If
does
it
automatically
close
everything
and
right
away?
This
hasn't
been
such
a
ninety
days
without
any
warnings
or
how
was
what's
the
first
run
experience
Oh.
C
A
A
A
E
C
C
C
He
has
an
issue
with
my
style
or
not
an
issue
but
yeah.
So
basically,
we
chosen
I
kind
of
had
it
about
some
changes.
In
my
pull
request
in
one
of
my
progress
where
yeah
like,
for
example,
if
they're
so
spaced
unnecessarily
at
the
end
of
a
line,
it
will
get
removed
and
stuff
like
that
and
yeah.
That's
the
first
point
or
removal
of
unnecessary
spaces,
and
we
kind
of
agreed
that
we
should
well.
C
C
C
And
stuff
where
you
can
enable
it
certain
things
that
we
want
to
code
to
look
like
like
simple,
just
very
simple,
like
style,
guys
like
as
I,
said,
removal
of
unnecessary
spaces
or
a
new
line
to
the
end
of
a
file
or
something
like
that
is
without
a
problem
possible
and
stuff.
Like
we
s
code,
atom
and
even
probably
like
all
editors.
C
B
C
A
Yeah,
this
definitely
sounds
like
the
type
of
thing
that
you
can't.
You
know,
100%
enforce
right,
so
I
think
your
attitude,
Travis
of
you
know
the
more
the
more
of
the
key.
You
know
core
contributors
or
the
you
know,
people
that
are
making
commits
if
they
have
those
settings
set,
then
that
makes
the
you
know
less
less
often,
for
you
know
massive
corrections
to
have
to
be
done
later
on
right,
mm-hmm,
so
yeah
Alexandre.
If
you
share
what
those
settings
are
and
the
dev
guide,
then
I
will
update
my
settings
to
in
compliance.
A
C
A
A
B
I
think
it's
I
mean
we
already
established
a
pattern,
even
with
my
PRA.
Now
that
we
can
do
it
get
a
long
ways
with
all
the
other
demons,
yeah
I
think
we
should
move
that
up
ODOT
night,
it
might
be
90%
automated,
but
maybe
there's
still
a
couple
of
manual
steps
left
to
see
as
our
first
school
of
automation,
right
right.
A
Yeah
I
think
that
that,
as
a
goal
in
0.9
is
fine,
but
then
you
know
the
fully
automated
for
sure
needs
a
week.
I
don't
think
we
can
declare.
You
know
version
1
until
without
that
right
here
in
the
back.
Maybe
that
makes
complete
sense
to
me.
Is
there
any?
Has
there
been
any
start
on
a
CSI
plugin
or
any
thinking
about
that
for
a
step
specific
one,
not.
B
A
Okay,
there
was
some,
so
it's
actually
the
second
step
back.
There
are
some
features
that
didn't
that
were
not
in
0.8.
Right
like
this
is
one
I
can
think
of
for
adding
and
removing
specific
discs
to
the
cluster
CRT
that
one
that
one
I've
still
think
it's
a
valuable
feature
and
want
to
see
in
0.9,
mm-hm.
A
B
A
B
B
F
Yeah
I
just
checked
it.
It
was
in
blameless
but
yeah
you
you're,
probably
right,
but
you
were
waiting
for
it
to
get
back.
Okay,.
A
A
F
Let
me
just
so
the
module
is
is
in
the
default
value
of
the
manager.
Initial
module
setting,
which
means
for
a
newly
created
cluster
you'll,
get
the
balance
of
module
automatically.
If
you
upgrade
from
an
older
version
of
Ceph
that
didn't
have
it,
then
you
would
have
to
manually
enable
it
thank
you
and
that
people
can
potentially
change
that
setting.
If
you
want
it,
enabled
it's
good
to
send
the
command
to
the
cluster,
to
enable
it
okay.
B
A
F
The
the
module
or
so
I
think
the
balance
module
is
one
of
the
ones
that,
in
addition
to
being
enabled
or
disabled,
it
has
modes
of
operation.
So
it
may
be
that
by
default,
the
module
is
enabled
in
a
sense
that
it's
running
in
South
Manager
and
you
can
send
commands
to
it
and
things
like
that,
but
it
might
not
actually
be
doing
the
work
of
that
background.
Balancing,
oh
until
somebody's
like
please
use
the
following
policy
policy
for
multi,
balancing
I,
see.
A
Okay,
that
that
makes
sense
to
me
then
John.
Thank
you,
a
big,
a
big
item
here
that
is
not
captured
as
some
more
progress
in
features
with
some
of
the
other
operators
for
other
storage
fighters
like
cockroach
TV.
The
big
one
is
the
secure
mode
right
now,
it's
installed
in
an
insecure
mode,
so
I
need
to
do
some
certificate
management
stuff
like
that.
So
that's
a
big
one
another
than
one
Travis
I
could
think
of.
Is
you
know
some
of
the
thinking
that
had
been
done
on
other?
A
What
is
what
is
the
right
terminology?
I'm
looking
for
here,
like
dynamic
provisioning
of
other
types
of
storage,
you
know,
like
buckets,
was
one
that
Tony
had
started
some
design
work
on
from
our
databases.
Like
you
know,
other
types
of
storage,
besides,
you
know
block
file
and
object.
So
that's
something
of
note.
I,
don't
know
of
exactly
where
it's
gonna
fall
on
the
roadmap,
but
that's
something
that's
not
really
captured
here.
I,
don't
know
exactly
what
specific
progress
needs
but
need
to
be
done
with
Mineo,
but
there's
probably
stuff.
C
A
C
E
I
have
I
guess
an
update
on
running
without
privileged
containers.
It
at
least
first
Steph
I
I,
still
think
it's
a
good
idea
to
try
to
plan
for
a
future
where
privilege
containers
are
aren't
necessary.
I
think
the
immediate
future
looks
like
we
will
still
need
to
have
privileged
containers,
but
the
the
biggest
security
hole
that
I
see
in
the
containers
is
having
read,
write
access
to
/dev,
so
limiting
that
to
read-only
access
to
/dev
when
like
when
that's
absolutely
necessary
and
otherwise
trying
to
give
containers
access
only
to
the
specific
discs
they
need.
E
C
E
A
A
F
Yeah
I
think
it's
just
a
couple
of
other
thoughts
on
that
some
some
people
are
interested
in
being
able
to
do
things
like
blinking
the
LEDs
on
their
devices.
I
am
not
completely
familiar
with
what
the
wiring
is
in
in
the
OS
for
doing
that,
but
it
would
be
focusing
to
keep
in
mind
if
you're,
clamping
down
on
the
access
to
devices
that
that
the
containers
have,
and
also
things
like
getting
a
smart
data.
F
So
you
know
in
a
scenario
where
the
OSD
is,
let's
say
on
a
an
LD
m
lv,
but
to
get
the
smart
data
it
needs
to
be
able
to
trace
back
to
the
underlying
physical
device
and
query
the
smart
status
on
that
I'm
a
little
bit
fuzzy
on
what
the
right
answers
when
T
of
this
actually
is.
So
that's
just
a
couple
of
things
to
paranoid.
E
C
Admin
needs
to
go
to
the
Machine
by
SSH
or
something
and
do
that
stuff
or
even
smart
data
you
normally
in
a
container
world
of
having
you
know,
one
processor
per
container.
At
least
you
would
have
an
extra
but
export
or
something
running
on
that
node,
which
is
a
completely
separate
pot
or
at
least
a
separate
container,
which
I
would
have
different
different
access
rights
to
the
disks
life-debt.
Yes,.
F
Ideally,
we
would
have
external
tools
that
provide
all
of
this
device
management
stuff.
The
trouble
is
that,
depending
on
whether
we're
running
in
kubernetes
or
running
in
on
a
bare
metal,
environment
or
anywhere
else,
we
would
usually
find
that
that
stuff
either
doesn't
exist
or
even
if
it
does
exist,
it
would
be
different
right.
F
So,
since
quite
recently,
we
we
now
have
like
the
code
in
SEF
for
actually
marking
an
OSD
out
when
it
when
it
has
a
smart
indication
of
failure,
and
that
wiring
is
a
lot
simpler
if
we
get
the
small
data
from
the
OSDs
and
it
just
works
everywhere,
and
so
that
stuff
is
gonna
be
unsafe
anyway.
So
if
something
like
Brooke
implemented
its
own
disk
failure
prediction
and
had
words
to
tell
stuff
about
that,
it
would
end
up
being
a
little
bit
redundant.
F
A
Let's
go
ahead
and
wrap
up
this
conversation
because
were
five
minutes
over
time
now
we
could
think
that
the
rest
of
that
offline,
I
think
so
Travis
do
you.
You
took
a
couple
notes
on
the
roadmap
stuff
and
you
have
a
couple
of
other
items
on
your
plate
right
now.
Do
you
want
to
just
send
me
that
over
and
then
I
can
take
a
stab
at
updating
the
roadmap?
Sure
yeah.
A
A
Sounds
good
no
I
could
follow
up
on
finalizing
the
roadmap
and
getting
a
PR
open
and
all
that
stuff,
okay
cool.
So
then
we
I
believe
that's.
It
was
the
end
of
the
discussion
here.
The
agenda
items
here,
sorry
for
keeping
everyone
a
little
bit
over
today
was
there
anything
else
that
needs
to
be
discussed
here
before
we
break
off
I.
C
Have
one
small
thing
to
mention:
I
just
saw
a
dent.
The
fossa
facade
check
is
currently
failing,
yeah
well,
I,
don't
see.
What's
what
a
problem
is
we
because
it
says,
like
a
flick
projects,
maintain
access
required
and
one
unless
into
pencil
the
dependencies
main
maintenance
maintain
access
required,
which
I
well
I
can
see
that
on
from
salsa
it
would
be
good
if
someone
takes
a
look
who
faces
to
that.
I
stay
has
at
least
one
issue
from
a
scan
from
last
four
hours
Lester.
So
that's.