►
From YouTube: Harbor Community Meeting 20190605 - Americas Time zone
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,
hello,
hello,
everybody
welcome
to
another
harbor
cncf
meeting.
This
is
the
second
meeting
of
the
day
we
already
had
the
harbor
asia
pacific
time
zone
meeting
earlier
today.
So
I
know
that
we
don't
have
a
ton
of
attendance
on
on
this
call
since
almost
25
people
on
our
earlier
call.
But
I
really
appreciate
you
guys
taking
the
time
to
come
and
meet
with
us,
doesn't
chat
about.
A
So
let's
get
started
here.
So
our
agenda
for
today
is
to
talk
a
little
bit
about
the
hardware
version
1.9
plan,
we'll
talk
about
the
four
or
five
big
commitments
and
a
couple
of
smaller
ones
that
we're
trying
to
first
cost
and
second
trying
to
kind
of
slot
into
the
1.9
release,
so
we'll
make
some
community
updates,
and
then
it's
open
for
discussion
later
on
and
it's
the
same
agenda
that
they're
using
for
the
asia
pacific
time
zone
as
well
so
starting
a
little
bit
of
the
wonderline
plan.
A
So
if
you
go
to
this
link-
and
these
slides
are
available
on
our
community
section
as
well,
you
will
see
okay
wants
to
verify
things,
but
you
can
open
it
here.
A
You
will
find
the
project
board
for
hardware
and
at
the
high
level
nathan
is
already
aware
of
some
of
these
items,
since
he
was
participating
in
the
maintainer
track
that
we
had,
where
we
kind
of
discussed
our
big
investments
made
sure
that
everybody
was
in
line
with
them
as
well
as
made
a
couple
of
calls
to
action,
so
that
folks,
that
have
bandwidth
and
resources
available
are
able
to
come
in
and
contribute.
Obviously,
the
same
call
goes
here
as
well.
A
If,
if
anybody
is
interested
and
has
time
to
come
and
help
us,
please
feel
free.
So
so
I'm
gonna
go
through
some
of
the
the
swim
lane
is
called
1.9
commitment
proposal.
A
The
reason
why
it's
still
a
proposal
is
because
we're
adding
t-shirt,
cost
sizing
for
each
item
and
then
based
on
that
we'll
figure
out
how
many
of
them
we
can
realistically
tackle
in
the
1.9
time
frame
we're
trying
to
ship
1.9
before
the
end
of
august,
so
that
we
can
have
a
secondary
release
before
cubicon
us
in
san
diego
going
in
november.
So
that
gives
us
two
relatively
short
releases
under
three
months.
I
think
it's
doable.
A
I
think
it's
gonna
allow
us
to
innovate
at
a
much
faster
pace,
but
also
the
the
side
effect
is
that
we
cannot
make
huge
investments.
We
have
to
take
them
small
by
small
chunk
size,
so
the
first
one
is
the
ability
to
implement
quotas
per
project.
That's
something
that
nate
does
feel
strongly
about,
and
you
know
kind
of
use
his
guidance
expertise
on
this
project
as
well,
and
that's
basically
giving
you
the
ability
to
do
a
couple
of
things.
A
We'll
start
simple,
we'll
start
giving
you
the
ability
to
limit
the
number
of
images
and
tags
in
a
single
project.
So
it's
just
a
straight
number.
You
say
I
can
only
have
10
images
here
once
that
limit
is
reached.
Unless
someone
deletes
an
image,
you
cannot
upload
another
one
and
the
second
one
is
the
ability
to
limit
a
project
on
a
per
size
and
we're
gonna
make
the
size
calculation
very
simple.
In
v1,
it's
gonna
be
the
total
virtual
size
of
all
your
images
and
your
repositories.
A
A
It
will
allow
people
that
use
hardware
as
a
content
provider
to
be
able
to
leverage
these
limits
so
that
no
single
project
runs
away
from
them
without
any
card
drills
and
later
down,
the
road
will
be
able
to
enhance
that
work
and
say
if
you
want
the
true
physical
on
disk
size,
that
the
single
project
is
occupying
will
be
able
to
do
that
as
well,
but
that
is
out
of
scope
or
version
one
along
with
that
and
now
open
it
up
for
questions.
A
A
very
very
complimentary
feature
is
the
retention
policy
so
being
able
to
add
the
retention
policy
for
images
is
going
to
be
very
important
and
that
retention
policy
is
going
to
give
us
a
couple
of
things.
How
many
days
all
the
images
do
you
want
to
keep
how
many
tags
per
repository
do
you
want
to
keep?
So
it's
not
only
a
date
driven
but
also
number
driven,
but
then
you
might
have
to
put
some
exceptions
to
these
processes.
A
So,
for
example,
don't
delete
the
last
remaining
image
or
don't
exclude
from
policies,
images
or
certain
labels
like,
for
example,
your
production
images.
So
if
you
have
a
production
image,
you
can
attach
the
label
to
it
and
that
can
be
excluded
from
the
retention
policy,
so
those
are
some
things
that
will
allow
an
operator
or
an
administrator
of
a
project
to
limit.
You
know
if
they
have
a
cicd
process
that
creates
an
image
every
day
or
multiple
per
day.
You
don't
want
to
keep
them
all.
A
You
want
to
keep
the
ones
that
matter
the
most,
the
ones
that
you
tagged
probably
and
then,
and
then
this
feature
will
allow
it
one
other
thing
that
goes
very
well
with:
that
is
the
ability
to
bring
import
labels
from
dockerfile.
So,
if
you're
building
your
images
using
your
own
ci
cd
policies
and
you're,
creating
metadata
on
the
images
in
the
form
of
labels,
today,
hardware
completely
discards,
those
so
we're
gonna,
be
able
to
import
them
now
and
you'll
be
able
to
attach
policy
to
that
policy.
A
Like
the
the
policies
I
mentioned
earlier,
around
retention
and
policies
around
replication.
All
of
those
policies
will
be
able
to
be
used
as
part
of
this
label
incorrectly.
Any
questions
so
far
on
the
first
three.
A
All
right
cool
moving
on
excuse
me,
the
next
one
is
and
this
this
is
going
to
be
a
huge
work
item.
I
don't
know
if
we're
gonna
be
able
to
fix
it
in
one
single
release,
but
our
goal
is
to
continue
making
progress
with
it
today
we're
a
repository
of
cloud
native
artifacts
like
container
images
and
home
choice,
but
nothing
else.
We
want
to
expand
the
concept
of
hardware
to
include
all
the
cloud
native
artifacts
that
could
include
things
like
scripts.
A
Very
importantly,
the
cnav
bundles
the
cloud
native
application
bundle,
which
is
a
new
project,
that's
getting
a
lot
of
steam
in
the
cloud
native
community
and
it
allows
you
to
portray
a
distributed
app
in
configuration
files.
We
want
to
be
able
to
support
that
and
then
also
operators
are
becoming
a
big
thing
in
kubernetes,
so
we're
going
to
be
able
to
become
the
repository
for
operators
as
well
now,
in
order
to
be
able
to
do
all
that
work,
it's
going
to
require
us
to
make
some
changes
today.
We
support
our
client.
A
A
Now
the
advantage
of
hardware
is
not
only.
Can
you
store
all
these
images,
but
you
can
enforce
our
policy.
Replication
will
work,
retention
policies
will
work,
our
bike
and
authorization
will
work.
Login
will
work
why
this
is
a
portable
work.
So
all
of
these
other
policies
that
you
have
in
our
world
will
work
perfectly
with
this
work.
A
The
only
thing
you
wouldn't
get
is
maybe
static
analysis
coming
with
claire,
but
everything
else
will
work
and
as
a
developer
of
cloud
native
apps,
all
the
resources
that
you
have
and
that
you
need
for
your
apps
to
be
able
to
be
deployed
in
production
will
be
available
and
hosted
in
hardware
and
have
a
single
place
to
go.
Look
for
them.
A
All
right
moving
on
the
application
working
group
would
like
to
continue
making
some
improvements
to
the
application
think
about
along
the
lines
of
adding
more
application
targets.
So
today
we
support
docker
hub.
We
support
docker
registry.
We
support
huawei
cloud
registry.
We
want
to
add
support
for
additional
registries,
maybe
aws
azure
gcp
registry,
maybe
some
other
on-prem
registries
as
well.
This
is
not
going
to
be
a
lot
of
work
because
most
of
these
other
registers
are
docker
client
compliant.
A
A
Webhooks
was
a
story
that
we
initially
were
targeting
for
one
debate:
it
ended
up
not
making
beans
so
now
we're
gonna
revamp
that
work
and
add
support
for
webhooks
in
1.9,
alex
zhu.
One
of
the
new
members
of
our
harbour
community
is
actually
cataloging
a
list
of
all
the
supported
web
hooks
that
we're
going
to
have
with
this
story
and
once
we
have
that
we'll
publish
that
to
the
community
request
feedback
and
then
we'll
continue
executing
on
the
story.
A
The
next
item
is
whitelist
cves,
so
this
is
something
that
came
up
a
couple
of
times
so
far
by
different
customers
and
users,
and
I
think
that
there's
a
lot
of
need
for
us
to
be
able
to
define
a
global
list
of
cds
that
you
can
exclude
from
policy.
So,
for
example,
today,
if
you
set
up
a
policy
that
says
in
my
project,
if
I
have
a
high
severity
cv,
do
not
allow
an
image
to
be
pulled,
but
that's
a
black
and
white
policy.
A
The
answer
is,
maybe
not
yet
so
an
alternative
is
you
add
it
to
a
white
list
for
a
short
period
of
time,
allow
production
images
to
still
be
pulled
and
then
later
on,
when
you
have
a
fix
for
the
vulnerability,
you
can
patch
your
images
and
then
force
back.
The
original
security
static
analysis
points.
A
And
that's
it!
That
was
the
last
item
that
we're
trying
to
tackle
like
I
mentioned
small
release.
So
I
think
even
these
five
or
six
items
may
be
too
much
for
a
single
three
month
release,
but
we
will
cost
them
or
slot
them
in
and
start
working
on
them
immediately.
There's
a
couple
of
other
process
item
things
that
we're
trying
to
look
for
for
1.9
things
like
the
security
disclosure
process
and
penetration
and
securities
testing
by
third
party.
A
A
All
right,
moving
back
to
the
slides
kubiconiu
was
in
barcelona
a
few
days
ago.
We
released
hardware
debaters
so
congrats
to
the
folks
that
contribute
there
nate.
I
know
you
are
your
commentator
so
congrats
to
you
on
on
this,
so
it
was
an
excellent
release.
It
was
very
well
received.
We
had
the
community
reception
on
may
20th.
We
had
approximately
30
35
folks
there,
all
of
them
very
interested
in
hardware.
A
We've
got
a
tremendous
feedback
on
hard
work,
both
at
our
two
sessions,
the
intro
and
the
deep
dive
we
presented
there,
where
we
were
where
we're
going,
and
some
of
the
new
features
that
we
have,
including
some
demos
folks,
were
lining
up
after
our
talks
to
talk
to
us
and
bring
their
own
perspective
into
hard
work.
We've
had
a
lot
of
folks
ask
us
for
new
features.
Some
of
them
are
in
our
roadmap
and
into
the
project
file
that
I
mentioned.
A
So
overall,
very
well
received
hardboard
is
gaining
a
lot
of
mind
share
within
the
cloud
native
community
and
I
think
we're
going
to
write
that
momentum
into
our
next
couple
of
releases
and
eventually
getting
graduated
in
cncf.
So
go
from
incubating
to
graduated.
A
To
the
picture
with
jomeda,
I
don't
know
why
I'm
not
in
the
picture.
Maybe
I
was
talking
to
someone
when
these
guys
all
took
it,
but
you
know
joe
was
there,
he
was
he
as
a
kubernetes
co-founder,
he
kind
of
talked
at
the
beginning
kind
of
kick-started
the
community
reception
and
talked
about
his
view
of
the
world
and
how
hardware
plays
a
key
role
there.
So
it
was
very
empowering
to
see
joe
kind
of
validate
and
solidify
our
ideas
about
how
to
take
cardboard
and
continue
advancing
our
vision.
A
A
Excellent,
so
all
right,
so
we
have
a
couple
pictures
here.
You
know
these
are
the
session
yeah.
I
think
this
is
the
intro
session.
So
look
at
the
bottom
right
picture.
That's
when
our
session
was
over,
how
many
people
were
around
us
and
the
line
just
did
not
laid
out,
so
we
had
to
be
kicked
out
of
the
room
eventually
so
tremendous
amount
of
interest
in
hard
work.
I
think
there's
a
lot
of
folks
are
paying
a
lot
of
money
to
some
of
our
competing
products.
A
All
right-
and
you
can
view
the
sessions
online
if
you
want
so
there's
the
internet,
the
test
session
are
available
on
youtube.
We
will
also
participate
in
the
cncf
answer
bar.
This
is
a
place
where
maintainers
of
different
projects
can
can
sit
there
for
office
hours
for
a
few
hours
every
day
and
folks
who
just
come
in
there
and
kind
of
talk
about
some
of
the
use
cases
and
needs
henry
to
the
picture
with
a
horrible
user
from
turkey.
A
So
if
you
notice
henry,
it
was
very
key
at
identifying
people
that
were
interested
in
hardware,
taking
some
pictures
of
them
and
continue
promoting
our
our
agenda
and
that's
it.
That's
all
the
slides
from
cubicle
europe.
Cubicon
shanghai
is
late
june.
We
will
have
a
community
reception
for
harbor
again,
where
dragonfly
hardware
and
tikv
are
gonna
present.
Some
of
them,
like
dragonfly,
does
have
a
demo
as
well.
They're
gonna
demonstrate
their
p2p
solution.
A
We
have
some
lightning
talks
and
some
sessions
as
well.
So
I
think
that's
gonna
be
continuing
the
momentum
of
copycon
eu
in
shanghai
and
then
and
that's
gonna
propel
us
for
a
1.9
release
as
well.
A
That's
it
anything
that
anybody
wants
to
bring
up.
A
B
Sure
no
worries
I
actually.
I
had
a
related
question,
so
I
was
looking
into
some
of
the
database
persistence.
Stuff
is,
are
the
database
migrations
auto-generated
by
the
orm
or
are
you
guys
hand
rolling
migrations?
Because
I
tried
to
run
some
tests
and
it
complained
with
an
error
saying
it
was
looking
for
like
migration
scripts.
A
I
believe
we
would
generate
those
manually.
Do
you
mind
asking
that
in
the
hardware
maintainers
chat,
though
today.
A
B
A
A
C
Can
you
send
me
an
invite
I'll,
try
to
tell
you
and
I
brought
it
up-
we
we
have
meetings
every
two
weeks,
the
china
one
was
last
night
and
they
definitely
would
like
a
presentation
geared
towards
running
harbor
instances
on-prem
at
edge
locations
yeah.
So
I
won't
be
able.
A
To
have
a
presentation
just
yet
maybe
a
discussion.
First,
I've
asked
the
team
as
a
result,
you
know
that's
something
that
was
on
my
mind
already
for
edge
scenarios,
so
I
had
asked
it
before
you
and
I
even
chatted,
but
they
haven't
had
time
yet
to
do
validation.
So
what
I
asked
the
team
to
do
was:
can
you
deploy
hardware
at
two
distributed
locations
and
use
some
top
some
networking
tool
to
increase
the
latency
start?
Dropping
packets
and
increase
decrease
the
reliability
and
see
how
well
it
does.
A
C
And
then
we
can
go
into
just
indicating
what
potential
issues
are
and
maybe
even
reaching
out
for
people
to
help
with
it.
A
Yeah,
absolutely
so,
I
think
that
it's
you
know
with
with
branch
scenarios
and
this
compute
being
pushed
down
to
the
edge,
whether
that's
for
telecom
or
retail
or
manufacturing.
I
think,
there's
going
to
be
a
big
need
for
images
are
being
there
out
there
at
the
edge
and
our
replication
capabilities
in
hardware
are
getting
better
and
better
than
that.
So
I
think
we
have
a
unique
position
in
here
to
actually
go
and
make
those
edge
case
scenarios
and
iot
as
a
reality.
C
A
Yeah,
so
so
absolutely
and
the
other
thing
by
the
way
is
we
are
trying
to
see
if
we
can
make
it
easier
to
update
claire
offline
so
that
if
you
are
in
an
air-gapped
environment
or
an
environment
with
such
highlight
incident,
that's
not
possible
for
you
we're
trying
to
look
into
that
as
well.
So
we
have
a
process
today,
but
it's
clunky.
So
maybe
we
can
make
this
better.
So
that's
another
thing
we're
looking
at.
C
Okay,
I'll
get
you
an
invite
for
that
within
the
next
24
hours.
Excellent,
thank
you
and
then
are
you
going
to
shanghai?
Am
I
going
to
see
you
there?
No,
the
one
of
the
teams
that
works
for
me
will
be
there
alex
joe
okay
yeah,
I'm
going
myself
so
I'll.
I'm
looking
forward
to
the
event.
A
It
just
didn't
work
out
the
travel
for
me,
otherwise
I
would
have
gone,
but
maybe
next
time,
okay,
cool
all
right
guys.
Well,
thank
you
all
for
attending
we're
a
short
group
today,
but
you
know
where
all
of
these
meetings
are
recorded,
so
folks
are
able
to
view
them
offline
and
we'll
talk
to
you
guys.
Next
time
really
appreciate
your
time.