►
From YouTube: KubeEdge Community Meeting 20200311 (Pacific Time)
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
All
right
the
recording
is
on
so
hello.
Everybody
welcome
to
the
cuvee
community
meeting,
it's
the
pacific
time
today.
It's
20
20,
11
march
wednesday.
Okay,
so
we
have
a
few
items
on
the
agenda.
We
can
first
go
through
that
and
if
anyone
has
anything
to
discuss,
we
can
discuss
after
the
items
all.
A
Right,
okay,
can
you
see
my
screen?
Yes,
yes,.
B
A
All
right,
I
think
the
first
one
is
the
installer.
A
A
All
right,
we
have
discussed
something
offline
previously
about
the
installer,
so
we
definitely
need
to
improve
the
installer
during
this
release.
A
So
if
any
disaster
recovery
or
manual
reboot
that
may
need
to
manually
spin
up
the
process
again.
So
maybe
we
can
install
h
core
as
system
service,
so
the
process
can
automatically
come
back.
A
D
Hi
kevin
yes,
so
for
cooperage,
so
only
the
ubuntu
is
guaranteed
right
now,
right,
yes,.
E
A
Yeah-
and
we
have
a
pro
request
to
support
sender
west
yesterday,
but
I
think
it's
still
under
review.
B
One
question:
do
we
need
a
fabric
fedora
support,
because
it's
just
the?
Why
is
the
enterprise
version?
One
is
the
open
source
word
of
red
hat
right.
A
I
think
maybe
we
need
verify
the
two,
but
I
think
it's
okay,
if,
if
the
script
and
the
and
the
code
wrong
well,
we
can
just
use
one
implementation.
D
F
D
Yeah,
I
also
did
something
similar
before
you
need
to
define
a
spec.
So
in
spec
you
need
to
write,
write
some
dependencies
and
also
the
libraries.
So
it
will
be,
it
will
be
fine
and,
and
then
you
also
need
to
maybe
how
some
some
some
guys
from
the
community
to
review
the
specs.
So
once
the
spec
is
gone
merged
in
the
distros.
C
D
Can
we
can
just
how
the
I
install
the
package
by
some
some
rpm
commands.
B
A
Yeah
and
probably
also
snap.
D
B
A
B
Stores,
it
is
different
right.
Yeah
systemd
is
for
centos
a
woman
will
probably
a
system
control
or.
B
Yeah,
I
think
maybe
yeah.
A
So
for
the
for
the
cloud
core,
I
think
we
need
also
expose
it,
expose
the
actually
cloud
harvest
service
or
to
or
to
the
load.
A
A
So
for
the
vm
support,
would
you
like
to
give
the
background.
B
Yeah,
there's
a
few
more
like
the
nfv
functions,
the
the
technical
may.
You
want
to
use
kubernetes
as
a
platform,
however,
a
lot
of
their
a
lot
of
their
applications
still
in
the
vm
image,
and
also
they
claim
even
they
repackage
as
a
container
and
still
take
a
few
gig.
So
it's
uh-huh.
E
A
Yeah,
maybe
we
if
we
can
find
a
way
to
implement
as
a
runtime,
it
should
be
much
easier.
F
Is
it
like
providing
openstack
services
quite
very
similar
to
the
openstack
services
on
the
host
directly
like
we
can
run
virtual
machines
directly
on
the
host
on
which
the
cube
edge
worker
node
is
running?
Is
it
like
that.
B
Yeah
it's
it's
kind
of
about
this.
I
I
don't
want
to
use
the
openstack
to
compare
because
that's
a
huge
platform.
I
I
want
to
compare
with
more
like
a
replay
make
a
darker
engine
or
a
container
engine
compared
with
vm
provisioner.
F
B
I
don't
know
if
we
want
to.
It
should
be
the
same
as
how
do
we
manage
the
container?
Currently,
I
don't
want
to
make
a
difference.
It's
just
a
one
runtime,
so
it
could
be
a
container
format
or
the
vm
format.
It's
okay!
It's
it's!
It's!
I
think
in
the
in
the
the
ad
core
may
look
like
it's
just
a
process
or
application
running.
It's
no
big
difference!
That's
what
I'm!
Okay,
okay,.
A
Yeah,
I
do
have
a
question:
will
there
be
virtual
machines,
spin
up
or
the
applications
are
running
on
the
same
host
with
age,
core
process
running.
B
I
think
the
id
core
only
spin
up
the
only
provision
of
vm
and
probably
we
use
a
cloud
need
cloud
in
it
to
start
a
service
inside
the
vm.
A
So
maybe
maybe
convert
or
virtualet.
B
B
A
But
that's
fine.
I
think
it
depends
on
the
depends
on
the
investigation.
A
Okay,
by
the
way,
I
think
so
currently
for
the
install
solar
improvement
and
the
the
vm
support,
we
will
need
when
we
need
a
new
tracking
issue
for
this
to
work.
A
Yeah
we
can
have
one
umbrella
issue
to
collect
all
the
other
things
and
for
each
detailed
items
we
can
create
and
create
another
one.
E
Yeah,
this
is
jane.
I
I
just
did
this
because
we
actually
started
to
propose
to
send
a
proposal
to
a
crano
for
the
blueprint,
as
we
talked
about
here
before
so
the
product
that
the
ptl
will
be
in,
because
he
he's
very
knowledgeable
about
cube
edge
itself
and
our
cranial,
even
though
it's
a
cranial
project,
but
the
focus
is,
is
to
promote
kubej.
E
I
think
he's
the
right
person
to
to
be
the
ptl
for
this.
So
in
the
next
week,
or
so
we
will
present
this.
So
cranial
has
an
incubation
process
so
present
our
proposal
in
two
a
cranial
cause
conference
calls
one.
Is
they
call
the
technical
technical
call
and
then
the
second
is
the
tsc
meeting
for
approval.
E
Project
and
then
I
think
we
are
shooting
for
the
validation
and
sometime
before
november
30th
and-
and
I
think,
because
okay
now
has
two
official
releases
every
year
one
is
may
30th
the
the
other
is
november
30th.
E
We
want
to
be
listed
as,
as
in
the
november
30th
validated
blueprint
project,
so
we
will
once
we
have
the
blueprint
project.
We
will
just
follow
the
the
guideline
of
the
cranial
community.
E
We
may
post
weekly
meetings,
I
think
in
we
can
discuss
how,
whether
you
know
on
some
cases
we
can
comply.
You
know
the
ukrainian
quebec
meeting
with
this
one,
not
necessarily
yeah,
not
necessarily
all
the
time.
Just
from
time
to
time
when.
B
E
Yeah
and
also,
I
think,
the
china
mobile
they,
they
also
they're
interested,
so
they
they
participate
participated
in
our
project.
I
view
them
as
more
from
the
requirements
for
interviews
yeah.
I
don't
think
that
there
will
be
the
still
there'll
be
any
committer
or
something
like
that
from
from
china
mobile,
but
definitely
they
will
they'll
function
as
a
contributor,
which
means
a
cradle.
A
contributor
doesn't
necessarily
need
to
commit
code
on
the
or
any
any
very
technical
like
configuration
file.
E
E
And
as
we
mentioned,
I
think
initially
this
will
be.
The
first
release
will
be
ml,
inference,
offloading
use
cases,
but
in
the
future
there
will
be
more
other
use
cases
and-
and
also
we
may
swap
out
some
of
the
the
components
in
the
end-to-end
solution.
Like
say,
for
example,
the
first
release,
probably
we'll
we'll
be
using
the
intel
hardware,
but
somewhere
along
the
way
the
arm
hardware
may
be
swapped
in
so
things
like
that
would
would
be
possible
now.
E
Actually,
I
think
we
are
just
go
ahead
and
combine
the
next
item
so
for
the
first
release
there
are
some
work
related
to
using
cool
badge
to
install
the
the
ml
framework.
Here,
mostly,
is
tens
rt
or
or
just
the
tensorflow
serving
you
know.
I
don't
know
whether
the
community
had
any
experience
in
this
area.
If
there
is,
we
definitely
would
like
to
be
connected
to
whoever
had
experience
in
this.
E
Maybe
we
don't
have
the.
I
don't
know
whether
we
have
the
answer
in
this
meeting,
but
definitely.
B
Yeah,
I
think
we
even
can
create
some
survey
or
something
yeah
jen.
Do
you
have
plans
to
publish
bp
for
the
blueprint
song.
B
E
E
I
think
there
are
two
presentations
involved,
so
it
may
take
one
or
two
weeks
because
I
think
some
of
the
meetings
they
they
only
have
once
every
week
and
then
you
have
to
request
in
a
mess.
E
I
think
in
the
future,
in
in
this
weekly
call
we
we
definitely
should,
at
least
for
the
specific
time
it
will
give
a
status
update,
either
from
in
or
from
me
on
the
ukrainian
project
program,
and
we
will
ask
for
some
help.
E
And
also,
I
think,
there's
another
part
of
on
the
cranial
project
side,
which
is
the
api
white
paper.
I
think
I
just
briefly
now
I
I
can
give
a
little
bit
more
detail,
maybe
in
next
pacific
time
meeting.
So
a
cranial
right
now
is
trying
to
come
up
with
a
blueprint,
blueprint
project
family
called
the
public
public
cloud
interface,
so
the
public
cloud
there
they
they
are
trying
to
get
it
into
these
project.
E
Families
are
aws
with
kdi
and
and
azure
with
with
att,
and
also,
of
course,
alibaba
and
china
mobile.
I
definitely
think
huawei
cloud
should
be
considered,
but
I'm
raising
this
here
because
I
view
kubej
as
on
the
edge
solution
side
for
huawei
cloud
overall.
How
we
should
go
about
this,
I'm
just
giving
a
a
very
brief
heads
up
now,
but
I
just
let
you
guys
know
there
is
something
going
on
on
that
part
we
can
see.
E
E
A
I
think,
thanks
again
for
the
useful
information
I
think
it's
for
the
presenting
as
well
another
thing
so
for
cubby.
I
think
the
first
thing
is
that
we
want
to
make
it
a
vendor,
neutral,
open
source
project.
A
So
so
I
think,
first
of
all,
improv
letting
more
and
more
people
know
cuba.
Project
is
definitely
helpful
and
that
we
did
also
clarify
that
cuba
is
not
just
for
integrating
any
specific
vendor.
It's
it's.
It's
open
to
integrate
with
other
cloud
offerings,
and
there
are
also
many
people
asking
the.
A
Is
there
any
limitation
running
cubbies
on
the
currently
most
of
the
popular
public
cloud,
offering
so
in
design
and
by
motivation
to
which
we
will
not
have
any
limitation
on
this?
But
we
also
need
to
verify
if
there's
any
existing
issue,
either
cost
introduced
by
cubase
or
it's
caused
by
the
implementation
of
the
vendors
when
they
offering
the
kubernetes
service.
A
So
for
the
presenting
our
cloud
to
a
carino
community,
I
would
recommend
we
reach
out
the
product
for
the
team
of
quality
club.
E
Okay,
yeah,
that
makes
sense,
that's
actually
a
very
good
character.
So
then,
I
guess
in
this
in
this
weekly
call
we'll
focus
on
the
cranial
core
badge
project,
so
we'll
we'll
discuss
the
other
item
in
in
other
appropriate
meetings.
Okay,
yeah,
that's
a
very
good
clarification.
E
Yeah,
do
you
guys
have
any
questions,
because
sometimes
we
may
we
assume
things,
because
we
were
working
on
this
every
every
day,
I'll
I'll
start
to
post
some
links
into
the
slack
channel
of
cooperage?
Would
that
be
appropriate?
E
I
mean
the
the
blueprint
project
of
of
this
cooperage
stuff.
E
B
Hey
jane,
can
you
add
the
the
page
link
the
blueprint
page
to
the
this
doc,
so
the
other
people,
if
they
are
interested,
they
can
take
a
look.
E
B
A
A
E
C
A
C
A
Oh,
by
the
way,
I
think
one
thing
might
be
worth
mentioning
is
that
we
fixed
the
docker
files
so
for
components,
especially
the
cloud
core,
and
also
I
found
that
we
missed
some
images
on
docker
hub.
So
I
pushed
the
1.2.1
release
docker
image
to
download,
so
you
should
people
should
able
to
to
deploy
a
cloud
cover
as
deployment
in
kubernetes
now
and
there
is.
There
is
some
yamas
for
reference.
I
can
just
briefly
show
it.
A
So
for
the
next,
maybe
we
can
just
organize
this
yamas
and
created
like
a
turnkey
script
to
to
restore
the.
A
A
A
Okay,
basically,
I
think
the
other
overall,
the
steps
are
similar
so
for
the
configure.
E
A
We
just
updated
the
config
to
reflect
to
the
latest
component
config
api.
A
A
F
E
F
Currently,
what
I
am
doing
in
my
like
for
my
requirement
what
I
did
is,
like,
I
created
puppet
modules
to
deploy
edge
core
on
raspberry
pi's.
So
I
just
install
like
I
just
have
created
one
small
module
on
my
puppet
main
server
and
then
all
the
raspberry
pi
just
configure
edge
core
by
default,
with
all
the
configurations
and
everything.
F
So
it's
like
configuration
management.
I
don't
manually
configure
edge
core
on
or
like
install
edge
core
on
each
of
the
devices
just
say
command
that
okay
install
edge
core
on
this
device
and
it's
getting
installed.
A
So,
as
far
as
now,
I
know
there
are
people
in
the
community
still
looking
for
a
simple
way
to
get
the
things
installed
and
manage
it.
Maybe
you
can
give
a
demo
show
either
this
time
or
next
the
time
about
or
where.
B
C
D
B
C
F
Okay,
so
I
will,
I
will
see
what
I
can
show.
Let
me
give
like
give
me
some
time.
Probably
I
would
just
open
source
my
scripts
and
then
you
can
just
have
a
look
at
it,
which
would
give
you
more
idea
about
it.
Would
that
be
okay,
cool.
A
C
B
B
B
B
A
Yeah,
so
maybe
we
can
have
a
demo
section,
every
community
meeting,
so
just
open
for
anyone
have
any
anything.
Interesting
would
like
to
show.
We
can
use
that.
E
A
A
By
the
way,
I
think
maybe
you
people
already
know
about
that
cubicle
europe
is
posted
upon
to
july
or
august.
So
I
think
we
can.
We
can
come
back,
come
back
on
the
planning
that
will
probably
organize
a
face-to-face
meet
during
cubicle
europe.
B
A
It
too
late,
I
think,
that's
still
under
confirmation,
but
probably
will
there
will
be
additional
slots.
I
have
discussed
mentioned
about
this
requirement
with
the
foundation
guys
they
are
confirming
that
so
currently
the
the
thing
to
verify
all
these
is
really
depending
on
the
the
the
conference
center.
They
are
so
they
are
still
looking
for
a
new
place
due
to
the
new
time
so.
A
Maybe
we
can
just
wait,
wait
for
the
the
official
announcement
about
the
new
conference
date,
but
if,
if
it's
just
a
face-to-face
meet,
we
can
either
use
like
dinner
time
or
or
other
time.
It
should
be
much.
B
Easier
yeah,
I
remember
it
during
the
san
diego
one
that
tina
she
hosts
one
in
in
qualcomm
in
santiago,
but
it's
a
little
bit
too
far
from
the
main
campers.
So
I
hope
this
time
we
can
have
a
unconversed
one,
so
people
don't
need
any
mouse.
E
B
Oh,
I
talked
to
the
francis
from
leonardo
community
steering
committees
today
about
the
potential
to
make
a
to
contribute
to
the
nano
community.
That's
the
armed
community.
So
I
talk
to
him.
B
So
we
are
thinking,
maybe
we
can
build
a
the
added
computing
framework,
a
reference
framework.
There
too.
A
Yeah
by
the
way
regarding
leonardo,
I
heard
of
someone
from
working
in
the
la
nero
community
mentioning
that
the
leonardo
community
have
some
more
armed
resources
for
building
ci,
so
maybe
regarding
the
project.
Currently,
we
are
kind
of
a
lack
of
cr
resources,
so
maybe
we
can
use
some
of
there.
A
But
I
don't
remember
correctly
if
it's
just
a
64
byte,
I
think.
B
B
A
E
B
D
A
Yeah,
thank
you,
everybody
for
joining
the
cubah
community
meeting,
so
the
meeting
next
week
will
be
on
march
18th.
It
will
be
europa
time
next
week,
so
yeah.
Thank
you
again
for
everybody
joining
the
meeting.
See
you
next
time.