►
From YouTube: Foundational Infrastructure Working Group [Mar 24, 2022]
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
B
Okay,
welcome
all
to
the
foundational
infrastructure
working
group
meeting.
B
C
Yes,
let
me
also
just
get
it
running
out.
C
The
way
it's
used
on
our
site
seems
like
a
bug
on
other
people's
side,
and
so
the
idea
was
that
it's
adjusted
on
our
side,
but
it's
not
exactly
us
who's
using
this
and
yeah,
it's
complex
surface
fabric
right
yeah
exactly
to
for
them
to
change
it,
and
that
means
we're
currently
blocked
in
bumping.
C
The
director,
which
is
not
a
good
thing,
also
considering
jeremy
and
other
things
and
just
stay
up
to
date,
and
that's
why
yeah
there
was
a
flag
as
an
option
discussed,
but
I
can
see
that
it
adds
well
some
complexity,
which
is
not
good,
but
still
what's
what
are
options
from
your
perspective,
if
we
cannot
unblock
this
on
our
side,.
B
C
B
B
Yeah,
I
I
don't
know
what
what
more
to
add
I
mean
like
than
at
least
I
mean
it,
wouldn't
get
prioritized
on
the
vmware
side,
making
this
thing
right
so
that
it
would
be
a
pr
against
the
bosch
director
to
add
this.
That
would
exactly.
C
B
By
that's.
B
Your
team-
and
that
would
mean
you
still
have
like
a
prioritization
and
effort
issue,
and
then
we
would
have
the
open
source
discussion
around
it
like.
If
that
is
an
acceptable
thing
to
go
forward
with
like.
I
cannot
give
any
guarantee
there
right
that
it's
an
acceptable
thing.
I
think
we
have
had
this
discussion,
which
at
least
goes
into
the
direction
of
it
not
being
desirable,
but
I
mean
like
maybe
the
implementation
turns
out
to
be
not
that
bad
and
it's
okay-ish,
but
I
don't
know
like
it's
not
yeah.
I
guarantee.
B
C
Understood
yes,
unless
it
brings
us
in
into
the
position
that
we
have
to
fog
anyway,
maybe
then
but
okay
and
who
could.
C
Decide
would
we
need
to
discuss
it
well,
the
approvers
list,
I
guess,
which
is
mostly.
C
B
Yeah
I
mean
it's
the
approvers
that
decide
in
the
end
during
well
yeah,
and
if
it's
like
a
controversial
thing,
maybe
we'll
want
to
add
some
more
approvers
to
have
like
consensus.
B
Yes
like,
like,
I
would
say
like
if
we
can
find
like
three
people
that
would
agree
with
this
of
the
approvers,
then
I
think
it
would
be
acceptable,
like
I'm
not
gonna,
like
I
don't
value
my
personal
opinion
like
higher
compared
to
like
the
other
approvers,
but
at
least
there
must
be
consensus.
Okay,.
C
Then,
let's
do
it
like
this.
I
can,
of
course
internal
discussion
will
continue
anyway.
I
can
see
totally
that
we
don't
want
to
have
backwards
compatibility
if
we
don't
consider
it
to
be
good
for
the
complexity
of
the
product.
C
So
I
can
see
that
we
want
to
avoid
to
have
this
flag
and
I
would
try
to
yeah
to
talk
with
the
previous
outside
of
this
meeting,
but
that
we
can
work
out
something
this
can
I
don't
know
in
the
next
one
or
two
weeks
and
yeah
if
it's
clear
that
it's
not
possible,
then
it's
not
possible
right,
but
then
we
can
take
it
out
of
this
meeting
yeah.
Maybe
I.
B
A
Yeah
yeah
and
then.
E
I've
made,
as
concerning
the
semantics
of
stock
vms,
are
definitely
compatible
with
your
requirements
and
properly
implemented.
They
would
solve
the
your
problem
because
you
it
would
be,
but
it's
an
io
to
run
and
around
on
a
stop.
I
think
it's
a
softly
stopped
the
end.
Isn't
it
yes,.
C
E
E
If
by
bush
deploys-
which
is
not
the
case
currently,
so
you
may
run
old
code
when,
when
you
do
that,
it's
not
consistent
actually.
E
So
yeah
yeah,
I
think
we
just
need
someone
to
do
that.
E
There
wouldn't
be
no
flag,
it
would
just
be
a
clarification
of
the
semantics
of
what
it
what
it
means
for
a
stop
vm
and
when
it
comes
to
being
updated
or
not
when
it
comes
to
running
around
on
it,
just
just
clarifying
that
and
in
making
sure
the
implementation
is
properly
aligned.
With
the
the
semantics
we
we
decide.
B
Okay,
so
benjamin
you're,
suggesting
that
the
change
would
be
that
a
softly
stopped
instance
would
still
be
getting
applied,
changes
but
then
stay.
I
think,
that's
not
technically
feasible,
because
an
applied
spec
will
probably
also
start
the
service
instances
or
they.
E
E
D
B
E
B
E
Out
yeah,
maybe
the
applied
spec
message
could
contain
a
new
option
for
not
running
and
not
starting
the
jobs,
because
possibly
this
could
be
investigated.
B
Yeah,
but
that
will
be
like
quite
a
big
scope,
because
I
would
need
agent
changes
as
well.
Yeah.
G
C
Okay,
I
think,
for
today
we
had
enough
discussion
with
that.
I
will
take
it
with
me
again
and
we'll
see
this
might
come
back
or
not,
but
yeah
thanks
for
the
open
discussion.
C
H
A
comment
just
maybe
we
can
have
somehow
a
decision
on
our
side
how
to
proceed
with
backward
incompatible.
Changes
in
such
cases
maybe
have
a
grace
period.
I
don't
know
something
here.
B
E
B
B
What
I'm
proposing
yeah
so,
but
then
you
would
indeed
need
to
solve
the
underlying
issue,
which
is
like
you
won't
run
out
on
out
of
date,
code
potentially
and
that
that's
the
bug
you
want
to
shield
people
from,
because
that's
really
unexpected,
behavior
and
really
difficult
to
debug.
E
Yeah,
actually,
we
would
delay
the
occurrence
of
possible
errors
when
running
old
code.
The
the
instant
with
with
the
instance
would
be
updated
with
new
bits.
It's
only
that
the
code
would
not
be
run,
so
you
wouldn't
you
would
not
have
the
feedback
of
whether
it's
running
or
not.
You
would
delay
that
until
you,
you
run
the
jobs,
but
it
still
makes
sense.
I
mean
if
we
see
the
pro
the
problem
this
way.
C
I
mean
okay,
I
think
the
the
question
of
beyond
is
valid
right
because
for
one
it's
future
or
back
in
the
end,
it
doesn't
matter
it's
a
breaking
change.
So
to
say-
and
at
least
there
needs
to
be
some
time
for
people
to
adopt
and
and
the
flag
could
be
an
option
if
you
decide
to
have
it
in
for
a
certain
amount
of
time
yeah.
But
I
admit
it's
a
risk
that
it's
then
postponed
and
postponed
to
get
rid
of
it
like
with
all
the
applications
right.
C
But
if
you
properly
deprecate,
even
if
it's
a
bug
and
you
figure
out
that
there
are
use
cases,
people
are
using
this
bug,
then
a
proper
amount
of
time
is
a
valid
request.
From
my
perspective,
yeah
yeah
sure.
B
So
and
that's
that
could
be
used
as
an
argument
to
get
a
fleet
a
flag
in
right.
Yes,
so
that
I
would
just
probably
look
at
a
pr
that
would
just
rever
have
the
the
behavior,
the
the
small
change
that
was
made
to
fix
this,
like
restore
the
old
behavior
with
the
flag,
and
then
we
can
have
a
discussion
about
that.
B
Yeah
I
mean
the
better
solution
would
be
would
be
to
fix
the
underlying
issue
like
what
benjamin
is
proposing
so
but
yeah
that
could
be
worth
a
spike
to
like
at
least
investigate,
but
if
that's
possible,
because
that
I
think,
would
be
less
that
wouldn't
be
like
any
resistance
against
that.
I
think
that
would
be
just
like
yeah
a
proper
way
forward
right
and
then
we
don't
have
that
problem
of
deprecating
a
flag.
We
don't
need
a
flag
because
we
fix
the
underlying
issue
yeah
but
yeah.
I
don't
know
that
would
be
yeah.
C
C
But
but
at
least
that
I
mean
the
deprecation
time
frame
for
such
things,
I
think
that's
a
valid
discussion
thanks
for
bringing
that
up
there
yeah
and
I
will
take
it
with
me
also
regarding
maybe
considering
creating
a
pr
traveler
discussion
of
the
flag
and
the
time
frame
depending
on
how
expensive
it
would
be
for
us
and
then,
let's
see,
yeah.
C
Yeah
change
by
the
way
from
your
perspective,
does
it
make
a
difference
whether
it's
just
in
the
api
and
or
also
in
the
cli?
C
C
Yeah,
yes,
because
the
proposal
in
the
in
the
bug
was
to
actually
obviously
live
like
one
api
effect,
but
you're
right
having
a
configuration
flag
for
the
director.
F
F
Let's
look
at
this,
what
we
have,
I
won't
be
discussed.
B
Yeah
this
one
I
reviewed,
but
I
also
had
a
call
with
this
person
today,
because
that
was
scheduled
anyway.
It's
from
vmware.
B
B
This
one
came
in
so
this
is
for
the
bosch
deployment
resource,
it's
just
better
logging
or,
if
you
have
better
error
handling.
So
if
there's
an
error
in
marshall,
it's
so
that
it's
easier
to
to
add
or
to
trap
some
types.
I
think
that's
a
good
improvement,
so
I've
approved
that
and
I'll
ask
you
a
moment
to
look
at
it.
B
E
B
G
B
For
the
test
to
be
okay,
yeah,
that's
right!
Yes!
So
now
we
can
assign
reviewers
sure
I'm
happy
to
take
a
look
at
this.
B
B
This
one
also
came
in
with
changes.
It's
updated.
I
haven't
looked
at
it
so
three
days
ago,
yeah
this
one
just
needs
to
review
now.
I
will
also
assign
myself
here.
B
B
Yeah,
so
the
structure
should
now
be
different,
because
since
then
I
did
a
vr
to
add
support
for
stem
cell
variants
right
so
that
we
have
like
a
better
way
to
structure
this,
and
then
this
vr
has
now
been
updated
to
use
that
variance
br
three
days
ago.
So
this
could
get
a
review
view.
B
I
don't
know
felix.
Would
you
have
time
for
that
or.
C
Yeah,
I
can
look
into
that
next
week.
Okay,.
B
Yeah
felix
are:
are
you
all
still
using
what
version
of
postgres
are
you
using
within
sap
13?
So
the
latest?
Oh
certain?
Third
later
so
this
is
when
we
talk
about
deprecating
stuff
yeah.
I
thought
I
asked.
D
Then
9.4
is
just
not
available
anymore.
You,
you
can
still.
D
And
that's
still
that's
already
still
possible
because
the
9.4
migration
is
only
supported
until
bosch
271x.
D
And
the
same
is
with
postgres
10,
but
I
didn't
even
start
with
postcrest
10,
because
watch
deployment
only
moved
to
postgres
13
like
for
the
last
two
months.
I
think
maybe
that's
a
bit
too
early,
but
yeah.
C
And
is
postpress
10?
Is
it
still
supported.
B
You
mean
by
postgresql,
yes,
people.
I
don't
know.
C
Yeah
so
basically,
I
brought
these
questions
up
because
we
have
some
scanning
ongoing
of
entire
artifacts
and
had
rightfully
complaints.
Oh
this
postcards
version.
Do
you
consider
updating
it
and
of
course
we
can
say
we
don't
use
it,
but
I
was
just
wondering
whether
the
community
should
carry
it
around.
D
F
D
D
Fell,
quiff
and
you
just
get
a
message:
go
to
go
to
earlier,
but
use
an
earlier
boss
to
do
the
actual
upgrade
and
then
move
on.
D
B
B
D
But
yeah
I
would
it's
just
stated
in
the
release
notes
and
so
that
people
still
can
go
to
the
old
one.
If
they
experience
issues
they
can
see.
Oh
this
one
has
a
new
gem
and
we
just
need
to
go
to
an
older
for
all
our
openstack
versions.
We
can
downgrade
the
cpi,
it's,
maybe
not
the
nicest
thing
to
do.
D
B
B
Well,
maybe
it
is
maybe
I
just
looked
wrong
with
the
versions.
Maybe
if
I
go
here.
G
E
0.1
has
nothing
to
do
with
1.1
no.
B
Yeah
no!
This
is
wrong.
Wow,
there's
just
backboarding
stuff.
I
think
okay
now
never
mind
we'll
just
wait
for
the
review
here.
B
Yeah
this
this
is
related
to
the
work
constantine
has
been
doing.
Apparently
that
broke
the
windows
agent
stuff
for
the
firewall
broke,
built
another
environments,
just
in
w
added
fix.
This.
B
F
B
No
problem
yeah,
I
don't
know.
B
This
is
actually
like
a
broader
question
right.
It's
these
things
are
now,
so
this
is
coming
from
their
work
to
start
supporting
rel
again.
G
E
E
C
B
That's
it
quickly
go
over
open
issues.
F
D
B
F
E
F
C
E
B
Will
be
open
for
pr
or
open
for
contribution
right.
B
B
It
basically
means
that
your
blob
store
has
already
been
like
corrupted
or
something
right,
and
then
you
need
to
do
an
fix,
upload
release
dash
fix
right,
so
I
mean
something
else
already
went
wrong
and
then
this
is
like
an
improvement
to
recover
from
that.
But
I
mean
like
just
having
a
snippet
here
is
already
like
quite
useful.
So
if
people
run
into
this
and
find
this
done.
E
E
This
kind
of
corner
case
are
are
related
to
not
updating,
stop
it
stop
instance
stopped
instances,
and
so
the
links
in
the
database
with
the
package
and
jobs
actually
used
by
the
instance
are
updated,
but
they
do
not
reflect
what's
exactly
on
that
in
in
there
and
yeah
we,
we
have
corner
case
on
there
in
this
area.
F
B
Yeah,
I
did
something
wrong,
I
think
with
the
steel
bots
or
I
I
reconfigured
the
steel
bot
and
then
at
some
point
it
also
started.
Oh
actually,
this
is
in
the
working
group.
So
I
think
that's
correct,
but
I
did
fix
some
update
logic,
so
it
didn't.
It
was
not
looking
after
all
our
repos
and
now
it
is
so
that's
why
we're
getting
a
bunch
of
still
issues
on
that
sim
on
that
date,.
B
I
know
he
did
or
I'm
fairly
certain
that
he
didn't,
because
I
read
an
internal
update
on
this
or
without
any
details,
but
it
would
be
nice
right.
No,
I
hear
about
that.
B
Yeah,
I
think
that's
it
there's
nothing
else.