►
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
Structure
working
group
yeah,
the
board
is
working
again.
A
This
was
removed
somewhere.
This
addresses
this
issue.
This
issue
was
actually
already
addressed
in
a
PR
here,
but
it
could
be
that
more
releases
have
this
problem
right,
as
we
don't
know,
that's
why
this
PR
is
backboarding
that
feature
and
putting
it
behind
the
flag.
A
So
this
this
was
removed
in
Ruby
3.2,
but
since
we
are
using
the
the
Ruby
on
the
director
to
render
templates
and
some
templates
we
found
at
least
these
data
dog
runs
in
the
Wilds,
but
there
could
be
more
right
that
rely
on
this
and
for
us
it's,
we
cannot
ship
breaking
changes
this
way.
So
that's
why
we
would
like
to
have
an
option
to
enable
the
old
Behavior.
B
And
also
heard
that
the
application,
those
controller
apis,
the
CFA
PhD
here,
issues
with
the
retries,
the
HTTP
quantity
Ruby,
does
and
through
3.2
patient
one
version
of
Ruby,
which
worked
perfectly,
but
in
the
next
version
the
rocket
commit
decided
to
change
again.
The
behavior
of
the
http
foreign.
B
Doesn't
Implement
correctly
timeout
the
timeouts
you
set
for
the
connections
and,
for
example,
if
you
have
top
three
options
when
the
cloud
controller
talks
to
the
Diego
apis,
there
are
through
t32
API.
So
you
have
multiple
apis
and
if
the
dialogue
doesn't
work
properly,
then
it
cannot
write
the
other
apis.
You
just
sync,
with
the
one,
the
first
one
you'd
like
to
connect
them
in
12..
A
B
A
Yeah
but
Bank,
could
you
take
a
look
at
spr
it
got
assigned
to
you.
Yes,.
B
C
A
This
one
has
one
approval
and
I
just
paint
Chris
salsa,
so
with
this
merch
soon,
and
then
we
can
move
over
to
the
issues.
I
already
did
a
bit
of
triage,
so
there's
this
yeah
to
me.
This
just
sounds
like
dry
run,
right
that
we
already
have.
C
We
need
to
attract
this
issue,
but
we
could
not
reproduce
it.
We
do
some
last
checks,
but
yeah
we
will
close
it.
C
B
I
think
it
you
can
easily,
because
if
you
just
try
to
send
the
agent
somewhere,
it's
of
course,
revocation
which
is
not
reachable.
Somehow
it
works
the
agent
and
then
we
try
to
execute
parallel
tasks
and
they
should
timeout.
Actually.
So
just
yes,
yeah.
D
I
already
couldn't
replicate
it
locally
when
the
issue
existed
at
sap
like
a
year
two
years
ago,
because
locally
I
couldn't
replicate
it,
but
we
could
replicate
it
within
the
within
the
sap
environments
at
that
time,
yeah
so
I
I
don't
know.
D
C
Yeah
I
I
will
check
it
again
but
yeah
it.
B
The
issue
is
that
the
zoom
DNS
with
signed
URLs
the
core,
tries
to
reach
the
lobster.
D
B
Don't
know
you
can
I
mean
the
easiest
way
would
be
to
just
manipulate
the
IP
table
and
send
actions
to
the
agent
to
zinc,
with
DNS,
sent
five
to
ten
actions
and
then
tried
to
do
also
something
different
like
a
task
executors,
some
kind
of
task
get
VMS
or
something,
and
it
showed
this
invisible.
The
task
timing
count
because
the
agent
is
busy
to
see
the
DNS.
B
D
A
B
A
Okay,
we
will
hear
an
update,
I
guess
in
the
issue
and
we'll
see:
okay.
A
A
Okay,
then,
this
might
be
interesting
to
you
all
so
apparently,
there's
something.
A
Yeah,
so
this
line
is
important,
at
least
in
some
environments,
to
get
the
so
if
you're
reusing
the
IP
address,
which
you
do
often,
where
you're
recreating
an
instance
right.
Some
Network
components
won't
update
the
ARP
tables.
If
you
don't
send
this
arping
package
and
the
problem
is
that
to
set
up
the
Nets
firewall,
you
would
have
to
resolve
the
DNS.
A
A
A
But
that
then
should
fail
and
we'll
just
go
through
the
whole
bootstrap
of
the
agent
again
and
then
the
next
time.
It
should
probably
work.
D
A
A
B
The
new
Mac
address
is
related
to
the
the
IP
address,
so
updates
the
MAC
address
change
so
to
say
yes,.
A
Yes,
propagate
the
MAC
address
change
yeah
because
you
have
a
new
instance
with
a
new
Mac
address
that
uses
the
same
IP
address,
and
otherwise
it
has
to
wait
for
the
router
routing
components
to
like
the
router
to
update
Optical
router,
but
the
router
to
they
have
like
a
TTR
or
something
on
these
entries
right.
So
it
needs
to
expire,
and
that's
not
my
deal
but
yeah
I
think
we
find
found
the
root
cause
here.
So
there
will
be
an
incoming
PR
to
fix
the
make
that
change
should
be
pretty
straightforward.
B
A
Yeah
that
that
was
the
other
thing.
So,
if
you're
gonna
do
an
aha
version
where
you
want
to
be
able
to
recover
or
the
recover
and
basically
update
your
IP
address
through
DNS
for
all
your
instances,
then
maybe
the
better
solution
would
be
to
make
it
opt
in
that
firewall
and
that's
firewall
thing,
because
that's
firewall
is
not
going
to
work
in
its
current
form.
For
when
you
change
your
director
IP
address,
because
it
just
puts
in
an
IP
tables
Rule
and
it
doesn't
re-evaluate
it
based
on
the
DNS
change.
Yeah.
A
If
you
do
that,
if
you
use
a
DNS
entry
for
your
abortion
directors
and
then
the
Nets
firewall
iptables
rule
will
be
potentially
wrong
right,
so
it
it
gets
resolved
to
one
IP
address.
That's
currently
returned
by
DNS
and
then,
if
you
update
the
DNS
entry,
the
the
agents
have
already
configured
the
instances.
So
it
won't
reconfigure
that
firewall
rule.
A
So
maybe
you
want
to
make
it
opt-in
all
together,
because
the
feature
itself
is
not
that
necessary
right.
It
was
more
for
the
kubernetes
workloads,
basically,
because
on
CF
you
already
have
application
security
groups
that
would
limit
or
prevent
access
to
the
management
network.
At
least
that's
really
common.
A
To
do
that
basically
has
the
same
effect.
A
Yeah
something
to
keep
in
mind
when
you're
doing
h8
wash.
D
It
wasn't
me
just
a
mission
you
were
misusing
my
stuff.
It
was
not
designed
to
do
this
foreign
yeah,
I
I.
Looked
it
up,
I,
just
I
I,
it's
freakishly,
weird
that
once
you
do
it
the
first
time
it
changes
it
and
when
you
do
it
the
second
time
it
overrides
it
back
to
the
default.
It's
just
like
I
can't
find
the
proper
growth
battles
where
how
this,
how
this
happens,
but
it's
weird
yeah.
D
D
D
A
D
Also
so
also,
basically,
it
you're
also
prefixing
independent
packages,
but
the
dependent
packages
should
not
be
it's
not
something
that
that's
what
you
it's,
not
something
that
when
you
vendor
a
package,
does
it's
when
you
create
a
package
where
you
set
the
dependencies.
D
Yeah
yeah,
we
could
do
that,
but
I
understood
from
from
Constantine
that
it
was
someone
from
the
from
VMware
team
that
was.
A
D
B
A
And
so
this
is
the
problem,
so
did
they
have
a
health
Checker
package?
That's
supposed
to
be
vendored
in
other
things,.
D
C
B
D
A
Okay,
it's
this
is
this
Russian
I
would
say
open
for
contribution.
A
A
D
D
A
D
C
Currently,
he
derived
this
as
a
side
project,
because
we
couldn't
have
other
topics
planners
to
focus
on
that
in
Q2,
then
yeah.
We
want
to
bring
it
in
and
use
it.
D
D
B
Debunk
she
has
also
CPI.
If
you
want
someone.
A
B
A
Maybe
that's
the
point
because
I
believe
there's
to
figure
this
out.
There
was
some
like
mapping
that
was
done
to
like
the
port.
Binding
thing
was
not
documented,
but
it
was
something
I
found
in
the
underlying
Docker
library
or
something
it's
just
like
just
piping
it
or
putting
it
through
or
something
you
can
actually
use
quite
a
lot
of
things.
There.
D
A
C
B
For
contribution
yeah,
we
have
the
new
requesting.
B
B
A
Relaxes
what
did
we
decide.
B
Do
we
missed
something
in
the
pr?
The
old
repository
is
still
exist,
I
think
we
didn't
delete
it
from
here.
B
A
A
D
A
D
D
B
A
I
don't
understand
what's
going
on
here.
Maybe
that
works.
D
B
A
Okay,
thanks
any
other
business
from
my.
B
B
Cool
archives,
yeah
I,
see
that
some
people
are
joining
the
CPA
channel
from
time
and
asking
questions
there
and
no
one
cares.
A
Maybe
we
can
update
the
the
descriptions
of
the
channels
to
just
say
that
they
should
go
to
the
Porsche
main
or
the
boss,
Channel.
A
B
A
B
B
A
Yeah
I'm
happy
to
move
over
to
something
that
they
maintain.