►
From YouTube: Antrea Community Meeting 05/09/2022
Description
Antrea Community Meeting, May 9th 2022
Antrea live show on flow visibility: https://youtu.be/cBO4b7KFOxo
A
Perfect
good
morning,
good
afternoon,
good
evening,
welcome
to
the
distance
of
the
andrea
community
meeting
and
today
is
tuesday
may
the
10th,
or
if
you
are
on
the
western
side
of
the
atlantic,
it
will
still
be
monday,
may
the
9th
for
you
for
today.
Unfortunately,
we
don't
have
any
pre-booked
topic
on
the
agenda,
which
means
that
we
can
go
directly
into
open
discussion.
A
B
B
Whether
we
should
release
the
m
chart-
and
there
is
one
question
that
came
up
about
how
to
manage
crds,
because
crd
crds
are
special
kind
of
resources
in
helm.
So
I
opened
an
issue
about
this.
I
haven't
had
time
to
come
back
to
that
issue
and
the
decision
hasn't
been
made
yet.
But
this
is
something
we
need
to
resolve
before
we
can
release
the
chart
so
that
it's
user
facing.
B
I
think,
maybe
I
should
add
that
the
rational
for
moving
from
customized
to
elm
was
just
that
we
kept
running
in
a
bunch
of
into
a
bunch
of
limitations
with
customize
and
maybe
were
using
it
wrong.
I
don't
know
using
it
for
the
wrong
purpose,
but
the
amount
of
customization
that
we're
able
to
do
with
customized
compared
to
like
the
full
power
of
the
elm
templating
engine.
B
B
But
now
with
elm
we
can
pretty
much
do
everything
with
elm.
There
is
more
customization
available
with
elm
and
I
think
also
it's
possible.
We
don't
do
it
right
now,
but
it
would
be
possible
if
someone
decides
to
turn
off
a
feature
of
andrea,
for
example,
we
would
be
able,
with
elm,
to
easily
remove
like
the
r
back
permissions
which
relate
to
that
feature.
B
I
think
that
users
who
are
familiar
with
elm
and
that
use
elm
to
manage
the
software
running
in
their
clusters
will
appreciate
that,
because
there
is
kind
of
like
a
standard
way
of
setting
resource
budgets
for
pods.
You
know
resource
requests
and
resource
limits,
kind
of
like
a
standard
way
potentially
to
label
things,
and
do
things
like
this
so
another
reason
to
to
support
helm,
but
right
now
we
just
use
it
for
manifest
generation.
So
we
just
use
like
the
templating
capabilities,
really
cool.
A
B
No,
because
really
I
mean
the
if
you
do
a
diff
of
the
manifest
it's
going
to
be
pretty
big
just
because
elm
has
reorganized
things
differently,
but
if
you
really
do
like
a
yamo
semantic
def,
there
aren't
really
that
these
aren't
really
any
different,
significant
differences
and,
as
I
said
it's
just
using
it,
we
just
used
to
generate
manifest
right
now.
So
if
someone
was
like
using
cube,
cdl
apply
to
install
entry
before
and
they
use
cube,
ctl
apply
to
upgrade
to
entry
or
1.7.
It's
just.
B
B
C
B
D
C
C
B
I
think
I
should
mention
there
is
one
I
guess.
Significant
change
is
that
with
customize
the
entry
config
map
used
to
have
a
an
automatically
generated
suffix
added
to
it,
so
it
was
never
entry.
A
config.
The
config
map
was
called
entry
config
and
then,
like
eight
random
characters,
which
were
based
on
a
hash
of
the
config
maps
content.
B
Now
with
elm,
we
use
kind
of
like
a
different
mechanism,
so
the
entry
config
map
is
now
called
entry,
config
and
and
yeah
that's
and
and
to
ensure
that
when
the
engineer
config
map
is
updated,
the
entry
pods
are
recreated
m
uses
an
annotation.
Of
course,
that
only
applies
if
you
use
elm
to
install
entria
and
and
that,
if
you
install
it
from
the
manifest.
C
A
A
A
I
I
don't
think
so,
all
right,
perfect,
okay,
so
these
the
only
the
only
other
thing
is
the
final
thing.
That
ever
is
that
I
noticed
it
on
the
community's
luck
channel
that
john
john
and
lunch
announced
that
gk
now
supports
network
policy
for
windows
pods
on
andrea.
I
think
that
there
isn't
much
to
discuss
about
this.
It's
just
a
very
good
news.
I
yeah,
and
I
think
it's
great,
that
we
can
do
more
with
the
windows
on
gke
and
with
andrea.
A
A
D
D
A
A
A
A
You
know,
for
instance,
we
are
also
introducing
the
python
through
spark
spark
is
another
component
which
is
used,
for
instance,
for
network
policy
recommendation
and
sparks
jobs
that
we
are
writing
them
in
python
at
the
moment,
because
I
think
that
spark
apache
spark
does
not
yet
support
golem.
So
for
all
these
reasons
we
are
moving
all
the
all
the
flow
visibility
components
to
this
new
repository
with
the
help
with
john
chancello.
A
We
are
also
setting
up
ci
for
the
for
this
new
repository
and
the
migration,
I
will
say,
is
pretty
much
completed,
but
it
will
be
staged,
meaning
that
for
andrea
1.7
we
will
keep
existing
flow
visibility.
Functionality
in
we
will
keep
existing
flow
visibility
functionality
in
antria,
but
all
the
new
features
concerning
flow
visibility
and
I
think
for
one
will
be
added
to
the
tier
repository,
so
andrea
1.7
will
still
ship
flow
visibility,
but
with
the
same
capability
of
andrea
1.6.
A
So
at
the
moment
we
are
working
on
completing
the
migration.
We
have
a
prs
in
review
for
the
for
these
new
features
and
to
be
consistent
with
the
entree
repository.
We
are
also
providing
the
helm
charts
for
automating
deployment
of
of
tia
bits.
Yeah.
That's,
I
think
that's
that's
a
summary,
zero.
I
know
that
you
have
problems
with
your
microphone,
but
if
there
is
anything
else
that
needs
to
be
added,
please
let
me
know
in
the
chat.
A
Yes,
yes,
andrea
1.7,
we
will,
we
will
keep
them
in
andrea
1.7
and
I'm
I'm
working
with
the
team
to
find
the
solution
so
that
if
you
have
anterior
1.7
with
flow
visibility
and
then
you
deploy
tear,
it
will
acquire
ownership
of
the
flow
visibility,
resources
and
you
know,
and
basically
it
will.
It
will
not
create
more
resources
but
just
update
the
existing
bits
and
take
an
ownership
of
those.
D
I
said
one
thing
I
yeah,
but
I
just
just
look
at
them:
they
are
a
repo
and
they
found
some
documentation
about
full
visibility.
The
new
fluid
video,
I
believe
I
think
one
thing
I
feel
a
little
concealing
that
in
the
anterior
ripple.
We
also
have
the
lowest
beauty
dock,
but
I
think
that
one
probably
point
to
the
other
implementation
with
ear
kiss
diet.
Maybe
I
I'm
a
ratio?
Okay,.
C
A
My
thinking
is
that
I
don't
want
to
create
a
distinct
documentation
for
tea,
even
if
there
are
two
distinct
products
I
would
like
to
keep
for.
I
would
like
to
keep
the
documentation
in
a
single
place,
for
I,
as
I
believe
this
might
be
simpler
for
users,
but
you
know
your
feedback
is
also
very
welcome
here.
D
Okay,
I
I
think
you
think
so
I
think
it's
so
capable
that
you,
you
want
to
put
a
talk
in
the
sarah
ripple.
If
I
just
mean
that
today
we
have
two
copy
of
the
talk
now,
actually
I'm
just
looking
at
the
two
dogs,
I
actually
I'm
not
sure
the
relation
between
these
two,
probably
they
can
be
someone
laughing
or
not.
I
am
not
sure.
A
D
D
A
Yes,
you
so
I,
yes,
we
have
an
anterior
network
flow
visibility,
yes,
and
then
we
have
it
waved
there
yeah.
You
have
a
good
point.
I
believe
it's
at
least
a
little
confusing
I'll
figure
out
with
alan
and
the
rest
of
the
team.
How
we
want
to
tackle
this,
the
thing
is,
it
should
yeah
if
we
have
the
same
documentation
in
two
places:
yeah,
it's
definitely
confusing
for
users.
A
Yeah,
let's
say
that
you
know
what
we
are
doing
here
is
at
the
moment
we
have
pretty
much
moved
the
code
and
documentation
from
one
repository
to
the
other,
but
yeah
there
are
some
loose
ends
as
a
documentation,
because
I
believe
that
it's.
What
we
want
to
do
eventually
is
to
build
the
andrea
dot,
io
documentation,
both
from
and
three
repository
and
the
taiya
repository,
but
that
will
require
some
additional
changes.
A
Okay,
cool
thanks
for
thanks
for
the
feedback
and
yeah
I
was
just
you
know.
I
did
not
realize
that
because
I
was
just
looking
at
the
generated
documentation
not
at
the
documentation
that
we
have
in
the
code
repository
okay,
thanks!
Well!
Is
there
any
other
question
feedback?
Any
other
curiosity
about
this
new
repository.