►
From YouTube: CentOS Automotive SIG - March 8, 2023
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
There
we
go
alrighty,
hello,
everybody
from
wherever
you're
joining
from
this
is
the
monthly
call
for
the
Centos
Automotive
Sig,
see
we've
got
a
half
an
hour
today
and
not
a
whole
lot
to
go
over,
but
we
have
a
few
important
things
we
I
would,
let's
see
start
with
just
introduction
for
those
who
are
new.
If
there's
anybody
new
on
the
call
we
do
this
call
about
once
a
month.
A
It
is
very
conversational
and
not
very
well
organized,
so
please
feel
free
to
dive
in
with
any
questions
that
you
might
have
the
slide
deck
looks
long,
but
it
actually
only
has
the
three
two
slides
really
the
agenda
and
the
technical
status
that
we'll
go
over
in
just
a
second
and
then
the
rest
of
these
slides
are
just
backup
and
I,
usually
post
the
slide
deck
afterwards.
A
B
A
C
C
So,
as
you
may
have
noticed
a
little
while
ago,
there
was
a
blog
post
published
on
reddit.com
about
our
green
containers
and
cars,
and
one
of
the
aspects
there
that
we
were
looking
for
is
is
to
have
a
system
that
allows
to
deterministically
manage
services
across
multiple
systems
or
multiple
physical
systems
in
a
way
that
is
compatible
with
running
containers.
And
to
this
end,
we
came
up
with
we
designed
that
we
like,
which
is
based
on
systemd
and
so
G4,
put
the
link
to
the
blog
post
on
the
chat
there.
C
And
so
we
come
up
with
a
design
on
that.
We
likely,
if
that's
relies
on
systemd
system,
is
starting
the
services,
the
services
being
containerized
application
or
not
containerized
application,
and
then
we
have
on
the
top
of
that
something
called
earter.
Utter
means
Shepherd
in
German
and
it's
a
it's
a
client
client
server
relation,
basically
where
we
have
a
primary
nodes
that
that
is
being
instructed
to
start
services
on
remote
nodes,
and
then
it
connects
to
the
agenda,
connects
to
the
local
systemd
and
the
local
system.
C
The
start
and
sort
of
the
services
and
all
the
communication
is,
is
going
through
debus
and
then
so,
if
you
have
a
our
ID
is
that
there
is
on
the
top
of
it.
There
is
a
larger
application.
That
is,
that
would
know
the
the
ideal
state
of
the
systems,
and
that
would
just
basically
tell
you
to
you
know:
I
need
I,
need
you
to
start
service
a
on
not
one
and
service
B
on
node
two
and
then
here
it
does
the
communication,
underneath
that
and
you
can
find
it.
Jeff
has
mentioned
it
here.
C
H-I-R-T-E,
it's
apparently
failing
to
build
on
the
copper
right
now,
but
that's
something
we'll
look
into
it's
we're
getting
ready
to
watch
1.0,
release
that
we
want
to
blog
about,
and
it
will
contain
the
minimal
set
of
information.
The
minimal
set
of
features
to
to
start
stop
Services
on
on
remote
systems.
C
Currently
it
does
what
we
want.
It
works
the
way
we
want,
which
is
nice.
There
are
still
more
work.
We
want
to
do
to
it,
but
it
is
a
it
looks,
promise
it
looks
promising
and
and
the
performance
wise
it's
it's
doing
a
very
good
job
so
far
so
happy
to
to
take
any
questions,
but
I
don't
want
to
necessarily
to
spend
too
much
time
on
this
at
the
moment.
If
there
is,
if
there
is
a
the
need
or
the
wish
to
have
a
dedicated
data
presentation,
then.
D
C
We
can
just
have
a
you
know,
a
proper
slot
where
we
go
towards
the.
C
How
it
works,
how
it
communicates
or
you
can
use
it
or
what
are
the
assumptions
of
use
and
so
on.
A
C
Yeah
I
was
definitely
it's
definitely
meant
to
be
running
in
the
on
the
automotive
context
sent
for
for
the
other
409.
It's
not
yet
in
Authority
proper,
because
we
we
haven't
made
it
proper
0.1
release.
But
the
idea
is
to
to
add
once
we
do
a
proper
0.1
release,
then
we'll
package
it
for
for
Fedora.
B
I
would
say
that
there's
nothing
specific
about
Auto
SD
I
mean
if
you
basically
had
two
or
three
nodes
that
you
wanted
to.
You
know
coordinate
with
systemd
events
across
three
or
four
nodes,
and
this
would
also
work
for
that.
It's
not
you
know
it's
not
really
that
and
I
think
quarter.
West
had
a
mental
blocking
on
the
name,
Corner
West
had
a
project
years
ago.
That
did
something
similar
to
this,
where
you
were
relaying
system,
D
events
between
nodes.
C
Was
designed
yeah
I
agree
with
we've
done
there.
It
was
it
was
created
for
for
automobile,
but
it's
it
can
be
used
outside
of
it.
The
one
limitation
that
we
have
currently
for
for
really
public
use
I
would
say
is
that
we
don't.
We
don't
yet
have
anything
that
allows
to
do
authorization
or
encryption,
so
we
don't
have
any
TLS.
Communications
here
is
not
encrypted
because
for
the
use
case
we
have
in
mind.
Currently
the
encryption
is
handled
at
a
lower
level,
so
we
don't
have
to
to
care
about
it.
C
We
don't
have
to
deal
with
it
at
the
air
terrible,
but
if
we
open
this
to
to
a
more
public
infrastructure
where
the
encryption
is
not
taken
care
at
a
lower
level
than
having
some
simple
triangle
fan
here,
that
will
become
necessary.
So
there
is,
there
will
be
some
work
here
to
do,
but
in
the
the
fundamental
I
agree
with
them,
the
fundamental
is
applicable
to
basic,
practically
innocent.
A
C
May
be
worse
also
to
to
bring
forwards
a
couple
of
another
couple
of
blog
posts
that
were
written,
one
is
from
done
and
that's
interesting,
quadlets
and
the
other
one
is
from
eagle,
and
it's
it's
again
about
quite
a
little
bit
than
doing
with
doing
quadlets,
with
a
little
bit
more
meat
example
a
little
bit
more.
It's
a
multi-container
example,
and
so
quadlets
is
the
is
a
mechanism
that
was
recently
added
to
podman,
which
allows
you
to
easily
manage
container
to
start
container
from
systemd.
C
So,
basically,
you
create
a
DOT
container
file
or
dot
Cube
files,
and
that's
that
container
files
contains
very
simple
instructions
on
which
container
you
want
to
run
how
you
want
to
which,
which
container
image
do
you
want
to
run?
Which
commands
do
you
want
to
run
in
that
container
image
and
so
on?
And
then
once
this
is
done,
when
you
do
a
systemctl
Reload
or
a
systemcj
demon,
reload
or
real
demon,
and
then
or
basically
at
boot
time
this?
C
C
This
is
automatically
handled
by
by
quadlet
for
you,
so
all
you
have
to
to
take
care
is
a
small
container
file
which
is
fairly
fairly
self-descriptive,
and-
and
so
there
are
a
couple
of
blog
posts
here
with
a
simple
introduction
to
quadlet
and
a
more
a
more
Lancy
example
of
your
more
complex
software.
B
B
Her
it's
all
about
managing
system.
The
unit
files
across
you
know
one
or
more
nodes,
so
the
integration
of
the
two
allows
us
to
run.
Containers
under
her
is
is
really
what
the
goal
is,
but
I
see.
B
B
B
Need
to
mention
is
the
image
name
in
this
case:
we're
writing
the
image
name
with
an
exec,
but
and
then
you
can
wrap
it
with
all
the
standard
system-
D
type
stuff.
So
if
you
wanted
to
throw
this
into
a
specific
c
groups,
if
you
want
to
you,
know
launch
this
before
or
after
certain
applications
are
up
and
running,
you
can
do
all
the
standard
system
D
stuff,
but
underneath
the
covers
with
the
quad
lid,
systemd,
quadlet
and
podman
will
do
is
generate
a
much
more
complex
service
file.
B
B
Basically,
all
the
intelligence
that
we've
developed
over
the
last
few
years
of
figuring
out
how
to
best
run
podman
as
assistant
D
service.
So
if
you
scroll
down
a
little
bit
further
you'll
see
the
exact
stat
command
and.
A
B
Yeah,
so
you
can
see
that
the
command
is
pretty
complex
for
and
it's
taking
advantage
of
some
system.
D
things
like
the
perception
is
injecting
the
the
name
of
the
unit
file
into
the
container,
and
then
you
know
setting
up
some
other
features
underneath
it,
but
you
get
the
basic
idea
of
that.
Our
goal
here
is
to
make
it
real
simple
for
people
to
create
unifiles
that
are
running
containers
and
then,
as
we
get
smarter
or
better
about
building,
you
know.
A
B
Okay,
absolutely
I'm
planning
on
doing
a
whole
talk
on
I've,
submitted
it
to
Dev
cop
check
and
oh
excellent.
Okay,
all
the
places
to
basically
do
a
talk
on
containers
and
I
call
it
containers
which
would
cover
a
lot
of
these
features,
features
and
functionality.
Yeah.
B
Yeah,
so
more
in
depth,
one
for
her
too,
but
the
general
concept
of
things
that
we're
thinking
about
for
as
well
as
separation
of
Azle
B
from
the
qms
okay
next
standards.
Instead
of
that,
please.
A
B
A
B
Yeah,
the
problem
is
getting
toxin
to
those
type
of
conferences
or
yeah.
A
B
A
B
Shy
on
that
right,
especially
where
people
might
think
Carter
is
a
competitor
against
kubecon,
even
though
it's
not
intended
to
be
yeah
kubernetes
so
right,
so
that
would
be
potentially
an
issue.
A
Okay,
okay,
well
any
questions
from
the
group
on
on
these
containers,
because
we've
got
obviously
a
lot
going
on
with
containers.
A
So
the
other
thing
we
wanted
to
to
cover
here
is
the
external
compatible
images
as
an
open
source
project.
The
Centos
Automotive
Sig
is
very
conversant
with
both
Sophie.
The
Sophie
is
an
arm
initiative.
It
stands
for
a
scalable.
A
A
This
is
a
an
architecture
of
built
around
a
vision
for
or
software-defined
Vehicles
as
a
software
architecture,
and
the
work
going
on
in
Sophie
is,
we
feel,
is,
is
very
important
to
the
automotive
world
as
it
kind
of
expands
or
evolves
into
this
kind
of
a
framework,
and
one
of
the
things
that
you,
some
of
the
things
that
you
can
see
here
are
are
you
know
the
fact
that
this
is
all
Container
based
and
that's
why
we've
been
doing
putting
so
much
work
into
things
like
here
and
and
quadlets
and
so
in
service
to
the
work
here
we
Sophie
is
created
is
has
basically
two
major
tracks.
A
One
is
creating
a
a
set
of
reference
software.
A
A
But
the
specification
work
is
coming
along.
It's
due
to
be
published,
I
believe
in
the
next
week
or
so,
and
the
anything
closing
at
once
here,
the
so.
What
we
have
also
provided
here
is
a
a
version
of
Centos
that
employs
the
the
corresponds
to
the
Sophie
Vision,
the
Sophie
architecture,
vision,
and
you
can
get
to
that.
A
A
So
anyway,
if
you
scroll
down
with
the
sample
images,
you
can
see
that
there
is
a
Sophie
image
here
that
is
built
nightly,
that
that
in
which
Sophie,
also
or
in
which
Auto
SD
also
uses
has
the
same
features
as
as
ewall
is
the
the
reference
architecture.
You
understand
underneath
Sophie,
so.
A
That
we
work
with
the
Sophie
project
in
order
to
kind
of
further
this
image
for
this.
This
vision
we're
starting
to
do
the
same
thing
now
with
with
Eclipse
sdv,
Clips
sdvs.
Obviously,
the
software
defined
vehicle
working
group
underneath
the
clips
Eclipse
sdv,
has
I
believe
14
projects
now
that
are
all
focused
on
various
aspects
of
of
software-defined
vehicles.
A
There
isn't
an
overarching,
architectural
Vision,
yet
with
sdd
in
the
same
way
that
there's
with
Sophie,
because
the
project
is
essentially
brand
new
I,
believe
it's
celebrating
its
first
anniversary
this
month
and
next
month.
A
So
so
there
is
an
effort
underway
to
to
kind
of
gather
all
of
these
projects
together
into
a
landscape,
not
really
a
software
architecture,
but
more
of
a
landscape
so
that
people
can
see
how
they
act
and
interact
with
each
other,
and
we
also,
we
have
a
an
sdv
image
now
as
well.
It's
in
its
own
repository,
so
it
can
be
consumed
by
some
of
these
projects,
and
it
currently
supports
the
eclipse.
A
So
the
only
other
thing
that's
that
is
ongoing-
that
we
don't
have
a
a
presentation
for
yet
is
a
process
for
incorporating
new
hardware.
I
know
that
that
has
been
a
question
from
some
of
you,
or
at
least
from
some
people.
I,
don't
know
if
anybody's
on
this
call
been
asking
that
question,
but
we
are
working
on
it.
It
is
not
as
easy
as
it
sounds.
A
C
The
it's
it's
proving
to
be
surprisingly
hard
to
find
the
exact
right
person
hatred
to
talk
to
and
and
try
to
to
to
get
them
to
to
talk
with
us
to
see
what
what
the
process
would
look
like.
So
it's
it's
definitely
working
process.
We've
made
progress.
We
we've
identified
the
workflow
that
we
believe
would
be
the
the
suitable
one.
There
are
still
boxes
in
this
workflows.
C
That's
we
are
unsure
about
and
we're
trying
to
to
find
the
right
people
to
get
the
answer
to
these
questions,
and
so
you
know
some
people
also
sometimes
more
time
off
from
work,
which
also
then
delays
getting
answers.
So
we're
definitely
it's
definitely
work
in
progress,
It's,
not
something.
We
have
dropped
the
ball
on
and
I
I
hope
we
love
we'll,
have
answers
soon,
because
it's
a
it's
been,
we've
been
let
it
for
for
I,
don't
know
a
couple
of
months
by
now,
I
think
Phillip.
E
Yeah
a
question
about
Hardware:
what
is
your
vision?
Do
you
envision?
You
read
that
to
support
a
certain
amount
of
reference
Hardware
or
you
expect
the
community
to
provide
some
waterbot
package
and
have
a
way
to
include
this
Rosebud
package
in
in
Santos,
Automotive
Ripper.
C
I
can
take
it,
I
think
the
former
the
so
I.
C
We
are
providing
a
Linux
distribution
and
therefore
that's,
as
you
know,
a
query
runs
sets
of
packages
and
features
that
are
tested
together
and,
and
one
of
this
package
is
going
to
be
the
kernel
and
the
my
my
understanding
is
that
we
don't
want
to
have
a
kernel
per
board,
so
we
we
will
want
to
have
a
kernel
that
is
compatible
with
you
know
all
the
boards
that
are
of
interest
and-
and
that
would
be
it
so
I-
think
having
a
mechanism
where
people
could
expand
the
supports
of
the
kernel
for
a
specific
board.
C
I
am
not
aware
that
you
can
easily
do
that
out
of
three
without
you
know,
being
able
without
recombining
the
kernel,
but
if
you
recompile
the
kernel,
then
you're
no
longer
using
you
know
the
auto
SD
kernel
and
if
we're
speaking
about
the
product
and
you're
no
longer
using
the
ndq
OS
kernel,
that
is
the
one
that
is
part
of
the
product
and
supported
and
certified
and
orbit.
E
You
know
it's
it's
option
of
choice:
okay,
it's
obvious
that
for
some
other
Hardware
that
we
are
currently
using
in
Project,
it's
never
going
to
fly
because
we
don't
have
enough
resources.
So
we
have
to
skim
down
the
kernel
and
in
fact
we
have
to
skim
down
the
full
Linux
to
to
support
what
we
have
to
support
the
first,
the
first
thing
being
to
boot
in
less
than
five
seconds.
Okay,
so
that's
clearly
thread.
The
second
thing
is
you're.
E
Never
going
to
support
enough
Hardware
to
respond
to
the
community
need
okay,
because
everyone
is
going
to
come
with
some
kind
of
probably
SOC
reference.
So
CEO
is
there
not
that
many
SOC
or
NSS
and
XP
TI
Qualcomm?
You
know,
so
it's
pretty
easy
to
have
all
of
them,
but
people
will
come
with
their
own.
I
o
and
they
will
have.
One
will
have
three
cans,
the
other
one
would
have
to
and
say
say
one
gigabyte,
internet
and
so
on.
So
it's
it's.
E
It's
tough
to
respond
to
everyone
without
providing
people
the
capability
to
do
customization
on
their
Hardware.
C
E
I
I
understand
that
I,
don't
say
it's
an
easy.
It's
an
easy.
There
is
an
easy
answer,
but
it's
clear
that
if
we
look
for
yorkto
and
I'm
not
part
of
the
yukto
fan
club,
even
if
we
know
it
pretty
well,
it's
one
of
their
friends.
Okay,
it's
very
easy
to
support
a
lot
of
custom
hardware,
and
so
it's
something
we
have
to
address.
Otherwise
people
will
not
be
in
a
position
to
use
the
system.
You
know
even
for
development.
E
F
A
Agree
I'd
say
these
are
our
ongoing
questions.
I
mean
what
we're.
What
we're
really
doing
is
that
yakto
is
really
based
on
an
embedded
design
and
build
Paradigm
and
Centos
really
isn't
so
we're
kind
of
maybe
changing
that
Paradigm
a
little
bit
in
order
to
accommodate
some
of
the
features
in
in
the
software-defined
vehicle
world.
C
E
Yeah,
but
to
be
honest
in
Project
we
have
currently
with
customers,
they
have
small
Hardware
with
very
specific
features,
and
they
nevertheless
want
to
have
a
per
package
distribution
with
long-term
support
and
making
sure
that
you
know
the
QA
team
does
not
break
the
kernel
and
the
middleware
team
is
not
in
a
position
to
change
the
neural
parameter
and
so
on.
So
you
know
it's
not
black
and
white.
So
there's
a
lot
of
Gray
Zone
and
people
are
clearly
willing
to
leverage
the
high
level
capabilities
that
we
have
in
the
cloud.
E
Like
you
know,
containers
per
package.
You
know
very
fast
update
on
on
only
one
specific
feature
in
case
of
vulnerability
and
at
the
same
time
they
are
willing
to
have
cheap
Hardware
with
a
lot
of
functionality.
So.
F
A
E
I
have
another
question
on
you
know
more,
like
the
collaboration
with
the
external
organization,
I'm
surprised
that
no
one
came
out
with
the
covisa
VSS
framework,
because
we
see
this
as
very
important
feature
for
our.
D
E
Okay,
so
for
people
who
don't
know
that
DSS
is
a
is
a
normalized
way
to
represent
techno
that
you
extract
from
the
car,
which
is.
E
A
hard
subject,
because
we
we
have
a
lot
of
symptoms
that
we
have
to
deal
with
and
quick
scar
inside
ellipse
eclipses
is
one
of
the
projects
that
Leverage
The
VSS.
So
VSS
is
a
collaboration
in
between
coveza
coveza
was
used
to
be
Genevieve
in
the
past.
The.
B
F
E
You
know
if
we're
the
only
one
who
you
know,
we
need
the
VSS
I
understand
that
it's
not
included
but
I'm
just
surprised,
because
we
we
see
a
demand
for
VSS
like
we
see
a
strong
demand
for
container
or
cyber
security,
and
it's
part
of
the
Arts
object.
A
A
It's
just
we
just
haven't
gotten
to
it.
Yet
you
know
we're
we're
still
sort
of
working
on
some
of
the
more
basic
things.
A
So
we
are
at
time
so
I
wanted
to
say.
If
there's
any
other
questions
feel
free
to
jump
over
to
the
mailing
list.
Hi
Desmond.
F
Yeah
hi
my
successful
I'm
working
in
USD
India
in
embedded
projects
so
recently
I've
been
looking
over
through
your
Centos
Automotive
distro
and
I
have
a
major
doubt
that
is
I
need.
A
comparison
between
the
Center
is
normal.
Centers
stream
and
Centos
Automotive
could.
D
A
C
So
basically
yeah
the
first
one.
There
is
I
think
Santa
stream
is
the
public
in
development
version
of
the
product,
Rel
ative
stream
distribution
is
the
public
in
development
version
of
the
product
reduction
vehicle
is
and
Auto
SD
is
itself
a
derivative
or
Center
stream,
because
we
leverage
the
work
that
is
happening
inside.
E
C
The
main
difference
that
we
have
currently
between
sunsestream
and
the
anato
SD
is
the
the
kernel,
which
is
a
real-time
kernel,
which
is
enabled
which
is
allowed
to
be
built
for
ar-64
and
I.
Believe
the
center
stream
also
has
the
canola
RT,
but
I,
don't
think
they
build
it
for
a
r64,
and
it
has
a
few
patches
in
automotive
Sig,
not
in
other.
In
notosd,
we
have
a
few
patches,
I.
Think
the
Raspberry
Pi
support,
for
example,
is
better.
You
know
the
auto
SD
kernel
than
it
is
in
the
center
stream
kernel.
F
Yeah,
okay,
do
we
have
support
for
emx
nxp
board
an
xpim
export
imx8.
C
F
D
Properly
being
supported,
I
I've
seen
a
roadmap
for
some
time
this
year,
but
it
won't
be
our
team
doing
that
so,
I
don't
know
the
details,
but
I've
seen
I've
seen
being
called
as
priority
down
the
line.
A
Great,
what
I
was
going
to
say
is
definitely
if
you
have
any
further
questions
or
you
want
to
continue
on
the
discussion
jump
on
over
to
the
The
Matrix
Channel,
which
we
have
here
or
the
mailing
list.
There's
mailing
list
as
well
and
we
meet
once
a
month
or
so,
and
so
we'll
continue
to
do
that.