►
From YouTube: Antrea Community Meeting 01/29/2020
Description
Antrea Community Meeting, January 29th 2020
B
B
D
A
C
D
D
C
D
Used
the
harbor
submission
as
a
as
a
baseline
to
see
make
sure
we
had
all
the
issue
areas
covered
plus
the
TOC
has
a
template
that
goes
through
and
physically
says.
Every
proposal
needs
to
have
this,
and
this
and
this
in
it
and
so
I've
ensured.
There's
there's
two
missing
pieces:
I
Illustrated
in
the
email
that
I'm
still
mopping
up
in
regard
to
the
architecture,
the
competitive
landscape
and
the
and
the
list
of
dependencies.
So
once
we
get
those
mopped
up
today,
we
should
have
a
submission.
B
And
I
ever
I
mean
I
went
through
your
document.
I
mean
it's
not
like
a
review
video
extremely
carefully,
but
I
have
only
one
question.
There
are
some
features
that
we
that
work
on
the
project
know
what
they
mean.
If
somebody
external
has
to
read
it,
they
may
get
confused.
I
gave
you
an
example.
Let
me
look
at
your
document.
Relate
it.
Oh,
this.
E
B
D
A
great
question,
so
what
I've
seen
done
on
some
of
the
other
proposals
and
I'd
already
mentioned
it
on
the
roadmap
as
well?
We
might
want
to
put
all
of
the
so
I'm.
Also
updating
the
roadmap.
I
have
to
PRS
I'm
submitting
today.
I
submitted
a
the
issue
yesterday
for
all
the
label,
stuff
and
I've
made
all
the
changes
to
contrib.
So
that's
gonna
be
one
PR
that
submitted
here
in
just
a
couple
minutes.
It's
ready
to
go.
The
other
PR
I'm
working
on
is
an
update
to
our
roadmap,
dot
MD.
D
B
D
Other
than
that
we
had
our
some
internal
discussions
around.
You
know
what
we
want
to
do
in
terms
of
feature
sets
and
I
want
to
make
sure
and
get
that
you
know
out
to
the
community
and
and
start
involving
more
community
members
in
discussing
that
roadmap
as
well
so
I've
taken
all
of
the
input
and
I'm
like
I,
said
updating
that
road
map
about
MD
as
the
basis
for
that
and
we'll
be
submitting
that
later
today,.
B
F
B
B
C
B
C
D
C
D
D
B
B
D
D
B
D
That's
gonna
be
required
for
a
lot
of
you
know:
users
that
are
you
know,
part
of
a
enterprise
or
some
other.
You
know
organization
that
has
an
open-source
compliance
policy
right.
There
they're
gonna,
want
to
know
what
all
open-source
licenses
are
used
plus
if
we
submit
to
the
CN
CF
they're
also
required
that,
as
part
of
their
process
doesn't.
B
B
Okay,
it
worked
all
right
so
so
far
in
the
payload
we
have
six
issues
two
bugs
and
for
announcements.
Some
of
them
are
moved
the
urban
mode
from
zero
to
zero,
in
particular,
Prometheus
and
unsee,
GL
and
I.
Think
this
one
also
so
moved
from
zero
to
zero
right.
They
want
to
group
out
supposing
the
logs
we
had
already
to
do.
You
put
zero
yeah.
C
B
Oh,
the
one
is
the
one
for
name,
the
port
that
we
already
discussed
about.
So
no,
no,
what
I
think
the
status
this
one
should
be
easy
right.
The
first
one,
the
cleanup
action
about
ountry
deletion
right,
yeah
I,
was.
C
C
B
F
D
B
E
D
So
I
think
it
would
be
useful
to
I
can
go
into
the
issue.
If
you
want
to
go,
be
send
me
the
list
that
you
have
currently
and
then
I
will
also
review
that
and
get
with
anybody
who
wants
to
consume
that
and
try
to
figure
out.
You
know
make
sure
we
have
a
comprehensive
list.
We
may
not
attack
them
all
in
this
first
issue
here,
but
at
least
I
want
to
start
capturing.
What
are
all
the
Prometheus
metrics.
B
All
right,
so,
regarding
exposing
the
logs,
we
know
that
this
is
not
really
a
feature
and
announcement
that
has
to
go
into
power
into
the
product
is
more
likely
an
infrastructure
announcement
and
we
are
trying
to
leverage
some
internal
infrastructure
so
that
we
can
host
these
logs
somewhere
and
make
them
accessible
for
the
entire
community.
But
from
the
latest
updates
that
I
got
there
has
been
any.
B
B
F
F
F
B
Right
and
I
don't
think
that
that's
not,
it
is
not
only
used
by
Neutron
issues
by
all
the
OpenStack.
She
is
right
for
okay,
okay
cool
anyway.
This
is
an
option
that
we
can
consider
if
the
finger
straddles
for
too
long
and
because
I
also
know
that
you
know
for
pushing
the
logs
to
public
routes,
there
will
be
budget
considerations,
so
basically
we
need
to
see
who
was
willing
to
fund
it
and
it's
not
necessarily
feasible.
There.
D
D
Let
me
let
me
take
this
and
if
you
Salvatori
want
to
provide
me
like
who
I
need
to
talk
to
regarding
the
existing
CI
stack,
that
we
have
for
for
OpenStack
and
whatnot
and
sharing
I'd,
be
happy
to
go
talk
to
them
and
work
with
puja
to
figure
out.
You
know
a
budget
and
where
we
might
be
able
to
acquire
space,
whether
if,
if
we
go
into
a
de
so.
B
D
E
B
D
A
C
B
D
B
B
D
B
That
we
probably
spaced
it
to
include
a
few
more
fixes
I,
you
know
honestly,
I,
don't
know
what
are
the
other
community
members
currently
working
on
if
they
have
spare
cycles
or
not
to
add
something
else,
I
know
that
they're
working
on
everyone
is
working
on
some
feature.
It
depends
on
whether
color
you're
thinking
of
adding
some
features
or
some
fog
fixes.
If
it's
about
bug,
fixes,
perhaps
somebody's
already
working
on
it.
If
it's
about
features
that
maybe
we
need
to
see
our
video
feature
is
to
see
we
can
fit
it
in.
D
D
You
know
any
any
types
of
bugs
that
may
impact
production
so
that
we
can
start
so
so,
for
example,
any
bugs,
if
there's
any
additional
logs,
that
we
think
we
might
need
to
support
a
production
release
if
there
are
additional
metrics
that
are
going
to
be
required
to
make
a
large
cluster
stable.
Those
are
the
types
of
issues
that
I'd
really
like
to
see
in
this
next
release.
B
C
B
D
The
other
things
that
I
think
would
be
useful
for
this
release
and
we
may
be
pulling
in
some
additional
help,
not
just
the
core
committers.
We
need
to
come
up
with
a
matrix
that
really
outlines
the
amount
of
testing
and
demos
that
we
have
done
on
different.
You
know
public
cloud
OS,
you
know
platform
combinations
right
so
that
we
can
start
understanding
and
prioritizing
what
we're
going
to
support
from
as
a
core
community
and
then
and
then
what
are
some
of
the
optional
combinations
that
will
be
supported
by
the
extended
community.
D
So
I'm
gonna,
I'm
gonna,
submit
a
prioritization
of
of
that
testing
and
then
get
your
feedback
on
how
we
go
about
essentially
certifying
that
that
on
tree
is
working,
for
example,
AWS
ec2
right.
How
do
we?
How
do
we
ensure
over
release
by
release
that
we
haven't
broken
the
functionality
of
on
tria
on
a
kubernetes
cluster,
with
cube
ADM
in
AWS,
ec2,
and
so
I
just
want
to
put
that
matrix
together
and
figure
out?
How
do
we
ensure
that
we
have
the
appropriate
testing
in
place
release
over
release
to
keep
that
stability?
D
B
All
right
so,
following
up
on
what
Cody
just
said
about
the
bugs
and
stabilization
I
would
like
to
host
one
of
the
most
exciting
things
that
one
can
do
a
meetings,
a
bug
scrub.
We
have
only
eight
issues
open
marked
as
bugs,
which
means
this
could
be
also
a
quick
thing.
The
goal
will
be
to
assess
all
all
the
open
bugs
and
assess
the
seriousness.
B
E
F
C
C
Is
no,
it
wasn't
our
back
issue,
basically
in
a
no
test.
We
deal
in
country
and
we
recreate
entry,
a
success
come
clean
up
stuff
and
that
that
breaks
the
one
of
the
entry
communities
api's
and
then
because
of
that,
the
test
time
space
cannot
be
deleted
because,
as
part
of
the
deletion
it
like
any
married
sample.
Yes,
I
found
a
workaround
which
was
like
to
configure
some
TTP
timeout
parameters
in
Linux
and
that's
specific
to
kind.
So
the
workaround
is
good
enough.
C
I
wasn't
able
to
figure
out
why
it's
only
opening
in
Canyon
that's
in
an
actual
cluster.
But
yes,
there
is
a
workaround.
So
I
don't
know
if
this
is
the
best
long-term
solution,
but
it
works
fine.
For
now
we
haven't
had
like
any
issue
with
the
Karem
end-to-end
testing
like
since
it
was
merged,
then
so.
B
C
It
already
applied
as
part
of
our
scripts
I,
don't
know
because
I'm
like,
if
you
look
at
every
project,
they
have
like
I,
mean
and
could
be
talked
about
it.
Last
week,
rate
issues
become
stale
when
they're
not
like.
There
is
no
update
in
a
bit
of
time.
So
in
this
case
we
know
we
have
a
solution.
We
know
it's
not
the
perfect
solution.
I,
don't
know
what
it.
D
C
E
C
Tcp
timeout
to
five
seconds,
and
so
basically
now
because
everything
runs
on
one
perfectly
fine,
when
everything
is
running
and
and
after
that
basically
TCP
gives
up
close
the
connection.
As
you
need
a
new
one.
It
will
way
stablish
a
new
one
if
you're
just
deleting
the
namespace
and
it's
just
gonna
go
so.
D
B
That
antonin
already
say
that
the
issue
will
anyway
go
still
if
nobody
sees
it
in
a
while.
An
open
issue
is
more
visible
than
a
closed
issue.
It
is
an
interesting
issue,
I
have
to
say
and
I,
don't
know
enough
about
the
kind
to
be
able
to
understand
what
could
be
the
difference
in
in
the
way.
The
connection
is
after
the
way
the
connection
is
established.
It
ideally
I
will
look
at
some.
B
C
C
In
entry
areas,
but
I
think
the
only
workaround
we
really
find
is
how
we
set
up
obvious
or
some
Linux
networking
week
and
I.
Don't
think
it
has
anything
to
do
with
purity.
I
wish
I
found
the
root
cause,
which,
in
my
opinion,
is
the
TCP
connection
should
get
reset
immediately,
because
there
is
no
longer
any
mapping
for
it
in
the
contract
table
I,
don't
know
where
it's
not
acronym.
B
That's
that's
right
anyway,
so
for
the
purpose
of
our
goal,
it's
not
something
that
we
want
to
target
with
it:
zero
for
zero.
The
next
one
I
fix
a
little
bit
more
interesting
qbp,
a
server
error,
error
logs
caused
by
system
and
three
at
Anzu.
Vmware.Com
antonin
did
you?
Do
you
have
more
details
on
this
yeah.
C
B
C
B
A
A
B
C
I
think
this
is
very
old,
yes
and
I.
Think
in
a
way,
it
may
have
been
the
issue
that,
with
issue
324,
we
talked
about
it
so
to
work
around
that
we
actually
were
skipping
the
test
on
time.
Okay-
and
we
assigned
this
one
to
me,
I
think
I
will
just
try
to
run
the
test
like
a
couple
times
now
that
we
have
the
workaround
and
if
he
was
indeed
the
same
route
cause.
If
you
can
just
like,
we
enable
the
test
for
the
kind
make.
B
Sense
make
sense.
Okay,
then
we
already
discussed
about
issue
181,
let's
move
to
152
and
three
agent
must
update
the
br8
bridge
if
the
data
path
type
has
changed
now.
This
is
one
of
the
oldest
one.
There
I
believe
we
have.
We
have
with
the
project.
Is
this
something
that
we
want
or
can
target
2
0
4
0,
or
is
this
bug
still
valid?
The
first
thing.
C
C
C
B
B
B
D
So
the
only
other
thing
I
was
going
to
quickly
mention
I
was
going
to
share
with
the
team
the
website
proof.
So
you
can
take
a
quick
look
at
that
I'll
post,
that
into
the
entree
a
channel
today,
so
that
you
can
give
me
any
feedback
on
anything.
That's
glaringly
wrong
or
any
or
needs
to
be.
Added
additionally,
has
mentioned.
I'll.