►
From YouTube: CNCF Harbor's Community Zoom Meeting - June 28 3, 2023
Description
CNCF Harbor's Community Zoom Meeting
A
All
right
we're
up
and
recording
hello.
Everyone
today
is
June
28th
and
that's
the
official
community
meeting
for
project
Harbor
with
that.
Please
follow
the
code
of
conduct
and
just
be
nice
to
each
other.
My
name
is
Olin
vasulov
and
I'm.
The
community
manager
for
project
Harbor,
I'm,
saying
that,
because
I
can
see
some
new
faces
like
correct
me
if
I'm
wrong,
but
that's
David,
yeah
on
on
the
call.
So
with
that
said,
I'm
gonna
share
the
community
meeting
notes
linked
to
the
chat.
A
So
please,
if
you
feel
like
it
all
yourself,
audio
Tropics
for
discussion
and
yeah
and
let's
see
what
we
have
there:
okay,
I've
added
a
few
stuff,
and
then
we
have
debate
and
then
Andrea
and
Stefan
and
then
volume
all
right.
So,
okay,
sorry
I
added
my
stuff
first,
so
I'm
gonna
go
first!
Maybe
it's
not
that
that
cool
but
yeah,
that's
chronologically
added
I'm,
not
sure
what
I
should
give
me
a
sec
share
this
one
alrighty.
A
A
A
I'm
not
sure
if
you
have
jaolin
on
the
call.
No,
but
we
have
young.
B
A
Yeah,
yes
I'm
here,
oh
I'm,
sorry
I
I
missed
to
see
you
on
the
list.
Sorry
for
that,
okay!
Hi!
No
problem!
So
do
we
know
because
I
suppose
someone
from
the
VMware
China
team
will
travel
to
to
Shanghai.
Do
you
happen
to
know
who's
gonna
travel?
If
anyone
is
going
to
travel.
C
Yeah
I
I,
suppose
we
we
have
I
mean
as
many
as
oh,
the
team
members
can
travel
to
sharks,
but
at
least
we
will
have
to
Nintendo
to
member
for
this
10
Nintendo
track
station
right.
So
at
least
we
will
have
two
and
for
others,
I
think
we
will
try
to.
A
Okay,
in
that
case,
can
you
DM
me
the
names
or
whoever
is
going
to
go
to
and
give
the
the
maintenance
track
to
China
to
drop
me.
A
C
A
A
So,
in
that
case,
if
you
think
I
can
help
with
submission
of
the
maintenance
track,
just
just
drop
me
a
line
I
can
help
out.
I
can
submit
that
I
I
can
work
with
the
team
of
structuring
everything
like
the
the
agenda
and
the
the
slides
and
everything
that's
required
for
China.
A
On
the
other
side,
on
the
other
side,
I'm
Gonna
Fill
in
the
form
for
the
kiosk,
for
both
events
for
China
and
for
North
America,
yeah,
yeah,
okay,
I've.
A
That
depends
on
the
team
and
who's
going
to
be
there.
Last.
A
It
with
the
young
one
and
I
think
it
was
very
successful
in,
in
my
opinion,
so
I
think
it
makes
sense
to
make
it
again
like
the
project
media
on
site.
A
Yeah,
yes,
so
we're
going
to
submit
that
right.
Okay,
we're
gonna,
submit
that,
because
for
Chicago
as
well
as
far
as
I
know
me
and
Vadim
are
going
so
I
think
we
can
hold
a
project
meeting
there
as
well
and
and
kiosk.
But
that's
like
we
have
a
little
bit
of
time
on
this
one,
and
my
last
topic
is
for
the
update.
You
know:
I've
been
shooting
these
kind
of
short
videos
for
the
project
update
on
the
keynote
sessions.
A
This
time
it's
a
little
bit
different.
We
have
to
provide
them
a
little
bit
in
advance
and
when
I
say
in
advance,
I
mean
three
months
which.
A
Awkward,
but
they
wanna
make
it
like
all
videos
to
be
the
same,
so
you
saw
that
project
updates.
Someone
is
shooting
over
Zoom,
some
other
like
me,
or
trying
to
make
it
a
little
bit
more
professional
in
terms
of
video
quality
and
audio
quality
and
stuff
like
this,
so
keep
calm
is
asking
us
to
submit
that
in
advance
and
that's
10
of
December.
A
That's
like
two
months
before
the
conference,
so
it's
highly
possible
that
we
have
a
lot
of
changes
until
the
conference,
but
still
we
have
to
collect
some
some
topics
that
we
want
to
cover
on
the
on
the
update.
So
if
someone
wants
to
something
and
some
new
new
stuff
that
we're
gonna
add
in
the
290
release,
which
will
be
like
few
days
after
the
end
date
of
10th
of
December,
please
shoot
them
in
in
into
the
channel
or
DM
me.
A
What
exactly
you
want
to
be
added
into
the
project
update,
so
I
can
shoot
that
like
latest
possible
moment
for
me,
it's
like
the
end
of
next
month,
so
end
of
July
before
I
go
on
vacation,
so
we
can
I
can
send
it
to
the
team
and
to
go
a
few
interactions
of
fixing
stuff
or
something
like
that.
So
keep
that
in
mind
for
the
next
two
one
two
weeks
and
if
you
want
something
very
specific
to
be
added,
just
drop
me
a
line.
A
There's
no
project
update
on
on
keynote
stage
plant
for
now,
yeah,
not
sure.
What's
the
reason,
I
I
think
it's
time
constraints
as
the
new
way
of
producing
that
kind
of
videos
and
audio
stuff
is
very
lengthy,
so
they're
not
going
to
to
be
able
to
to
do
that
sometimes
so
yeah,
but
that's
for
the
rest
of
the
folks
who
are
not
aware
of
what
you're
doing.
A
That's
like
our
time
of
45
to
60
seconds
time
on
stage
to
shine
and
bright
on
the
keynote
session
and
say:
hey
we're
here,
we're
doing
this
and
that
and
joining
us
or
wherever
you
can
check
one
of
this.
If
you
haven't
my
beautiful
faces
on
that
video,
so
you
can
join
me
another
four
or
five
minutes
seconds
so
yeah.
That's
that's
the
whole
story
behind
these
updates,
it's
like
to
show
to
to
the
Cloud
native
Community
what
we've
done,
where
we're
heading
and
yeah
how
to
join
us.
A
A
That
is
also
fine.
Okay,
welcome
to
the
club
super
happy
to
see
you
contributing
on
GitHub.
So
can
you
first
tell
there
was
something
about
yourself
and
then
I'll
stop
sharing
and
if
you
can
do
what
you
discussed
in
the
pr.
E
Sure
yeah,
so
I'm,
David,
I'm,
lead
engineer
at
plural
I,
think
the
best
way
to
quickly
describe
what
we
do
is
it's
kind
of
like
an
app
store,
but
for
cloud
native
applications.
That
makes
it
very
easy
for
people
to
install
like
complex
application
Stacks
in
like
a
production,
ready,
Manner,
and
we
do
also
have
things
like
Auto
upgrade,
and
things
like
that
and
in
the
process
of
adding
Harbor
to
our
catalog
I,
ran
into
I
think
a
an
issue.
E
A
lot
of
people
are
familiar
with,
and
that
is
that
irsa
on
eks
doesn't
work
for
Authentication
and
so
I
I,
dug
into
that
a
bit
more
creative
PR
that
fixes
it,
which
I'm
expect
or
I,
expect
it
to
be
quite
a
controversial
PR,
but
I
thought
I'd,
create
it
anyway
and
kind
of
highlight
it
to
the
group.
E
So
I
guess
I'll
share
my
screen.
E
So
for
those
that
do
not
know,
this
is
the
pr
that
I'm
talking
about
and
what
this
does
is.
It
adds
two
patches
to
the
the
registry
Builder,
one
that
updates
the
AWS
SDK
and
one
that
fixes
the
way
that
the
SDK
is
is
loaded
so
that
irsa
works.
E
The
the
pr
for
the
first
patch
here
is
that
fixes
the
way
the
SDK
is
loaded
is
quite
small,
knock
super
interesting,
the
other
patch
that,
sadly,
was
required
to
get
this
to
work,
which
is
updating.
The
SDK
is
obviously
a
huge
change
right.
So
what
I
did
is
I
just
I
thought:
I'd,
try
it
anyway
and
see
if
it
worked,
and
it
does
work
so
I
can
just
quickly
show
it
actually
working
or
first
not
working
and
then
with
the
image
I
built
that
actually
works.
E
So
I'm
not
just
trying
to
push
a
container
image
to
to
Harbor,
and
you
can
see
here
in
the
logs
of
the
registry
that
it's
getting
access
to
nod
logs
to
AWS.
E
E
Is
this
PR
something
that
people
want
to
merge
or
not?
One
of
the
or
I
think
a
couple
of
the
topics
here
are
obviously
Harbor
relies
on
on
the
Upstream
distribution
project.
E
Upstream
distribution
are
only
going
to
add
security
fixes
to
the
current
release,
Channel,
and
neither
of
these
PRS
are
considered
to
be
a
security
release,
so
they
most
likely
won't
get
added
to
like
a
release
that
can
be
cut
easily
V3
of
distribution
for
those
not
familiar
with
it.
That's
on
the
current
main
branch,
it's
very
different
from
the
current
release.
E
Branch,
like
their
current
release,
Branch
isn't
even
using
go
modules
and
and
Main
is
it's
not
really
sure
when
V3
is
going
to
be
released
now,
I've
been
trying
to
work
with
some
of
the
maintainers
of
distribution
Upstream
to
get
like
the
last
few
things
that
need
to
happen
to
get
V3
released
because
I
think
that'll
make
everybody
happy
since
maintaining
a
release.
Branch,
that's
so
different
from
the
current,
like
main
branch,
I
think
probably
has
caused
like
the
the
activity
to
kind
of
die
down
a
bit
on
the
on
that
project.
E
E
A
Thank
you.
The
floor
is
open
for
this
cushion
I
personally
I,
like
that.
We
can
have
that.
Finally,
fixed
the
fix
is
very
hacky.
Obviously
I
think
Pete
was
also
joining
that
conversation
through
the
pr
stuff.
So
maybe
you
can
share
some
some
of
your
thoughts.
F
Yeah
yeah
so
hi
everyone
so
yeah.
My
name
is
Pete
I'm.
Like
a
platform
engineer
working
for
big
financial
services,
company
called
Fidelity
in
the
the
south
of
the
UK.
F
We
run
a
massive
Cloud
Foundry
we're
currently
building
a
big
kubernetes
platform,
which
you're
probably
going
to
switch
to
from
cloud
Foundry
and
we'll
and
we're
using
Harbor
it's
working
great
for
us
and
that
you
know,
thanks
for
all
your
contributions
and
making
this
awesome
open
source
software,
but
yeah
one
problem
we're
having
is
that
as
soon
as
you
start
to
use
static,
S3
credentials,
a
thousand
alarms
will
go
off
in
the
security
teams.
You
know
and
I've
been
in
a
lot
of
meetings.
Saying
hey.
F
We
really
want
to
use
this
awesome
Harbor
thing
and
we
have
to
cut
all
these
non-compliance
exceptions
for
Harbor,
so
definitely
in
financial
services
and
I
expect
for
all
large
Enterprises.
It's
going
to
be
a
really
great
help
to
support
irfa
and
yeah.
F
I
would
also
personally
make
my
life
easier,
because
there's
a
lot
of
like
security
teams
that
I've
had
to
like
so
yeah
it's
safe
and
here's
all
the
mitigations
we're
using
to
using
these
static,
S3
credentials
in
a
safe
way,
and
you
know
yeah,
so
I
just
want
to
express
our
support
as
a
user.
For
this
feature.
A
Yeah,
thank
you
and
on
the
opposite
side,
I
I
saw
young
one.
You
added
some
command
that
you
don't
like
the
the
pr
and
or
not
the
pr
itself,
but
the
way
how
we
fix
that
thing.
G
Yes,
actually
personally,
I
I'm
against
that
this.
This
change,
because
that
so
the
main
reason
is
that
the
the
distribution
like
Harbor
is
another
cncf.
G
So
the
first
of
all,
we
should
contact
to
the
maintenance
from
the
distribution
Community
to
contact
to
them
on
the
awesome
to
release
a
new
patch.
So
then
Harbor
can
pick
up
the
latest
patch
into
Harbor
as
a
binary
and
and
actually
since
the
the
distribution
is
another
open
source
project
hosted
by
since
that,
we
Arbor
does
not
want
to
maintain
a
forked,
a
co-base
or
a
modified
code
base.
I'll
use
the
hike
away
and
to
to
introduce
some
PRS
from
the
Upstream,
especially
for
the
SDK
upgrade.
G
G
First
of
all,
we
should
raise
a
a
problem
in
the
Upstream
community
and
we
should
have
certain
answer
from
them
whether
they
will
have
this
PR
or
not,
whether
they
will
continue
to
maintain
the
release
to
the
A
or
or
whether
they
will
have
a
new
manner
or
major
release
in
the
4cm
months
or
years
yeah
so
yeah.
This
is
the
main
reason
from
me,
but
another
problem
is
that
Harbor
has
several
dependencies
other
than
distribution.
G
We
have
a
lot
of
other
dependencies
like
bego
or
Theory,
or
the
the
components
that
we
deprecate,
the
Char
Museum.
Normally,
yes,
we
actually
do
not
want
to
maintain
too
many
South
Coast.
G
We
just
want
want
to
focus
on
the
hardware
itself
right,
so
so
we
we
I
can
give
you
some
example
like
whatever
you
date
for
three
week
or
or
Pico.
G
Harbor
relies
on
these
two
components
to
do
scanning
or
or
do
the
website
a
framework,
and
when
we
find
some
problem,
no
matter
is
the
cve
or
security
problem
or
or
any
feature
requests
or
park.
We
always
conduct
through
the
container
and
ask
them
ask
them
to
help
us
to
address
the
issues
and
ship
another
patch
with
this
Wars,
and
then
we
can
Pump
Up
the
new
newer
version.
I
think
this
is
a
normal
process
for
us.
G
If
we
decide
to
a
fork
the
these
two
PRS
into
the
harbor,
so
for
me
what
what
is
the
long-term
maintenance
principle,
so
today
we
can
accept
this
2prs,
but
and
tomorrow
someone
came
to
us
that
and
asked
us
to
merge
several
other
PRS
and
what
are
we
gonna
do
and
and
if
these
PRS
have
some
conflict?
So
how
can
we
resolve
them
right?
So
so
we
so
in
this
case
we
eventually
need
to
Fork
the
Upstream
source
code
and
into
Harbor
repository,
and
then
we
can.
G
We
have
to
take
either
to
maintenance
code
base
and
but
I
I,
don't
think
at
least
it
makes
sense
for
art.
For
me
because,
like
I
mentioned
earlier,
distribution
is
another
project
that
hosted
by
cncf.
Harbor
and
distribution
are
in
the
same
community,
so
we
should
do
some
collaboration
rather
than
we
we
Fork
them.
So
it's
called
right.
E
B
D
E
Can
solve
this
Upstream
potentially
within
the
people
we
have
here,
or
we
can
like
find
the
other
solution
Upstream,
which
is
not
to
merge
these
PRS
into
the
current
release
Channel,
but
to
put
our
hands
together
and
work
through
the
not
a
lot
of
things
that
are
left
open
for
V3
to
be
able
to
be
released.
G
Yeah
I
think
we
first
off.
First
of
all,
I
think
we
should
raise
this
problem
into
the
distribution
community
and
and
no
matter
through
the
the
GitHub
issue
or
over
the
slack
Channel
I
think
you
can
adjust
this
set
up
the
message
in
the
section
also,
then
I
can
ping
the
maintenance
from
the
Distributors
group
actually
I'm,
one
of
them
and
I
I
have
already
discussed
this
problem
with
them
and
they
are
not
willing
to
accept
this
PRS.
G
We
just
discussed
some
issues
on
the
release
on
the
on
what
we
have
and
what,
whether
we
can
cut
a
new
release
on
2.9,
but
but
but
everything
and
has
no
concrete
answer
so
far,
but
but
I
think
we
you
can
reach
this
problem
in
in
the
distribution
community.
So
then
we
can
start
the
the
discussion
around
us
on
and
I
can
involve
more
maintenance
to
to
give
you
a
response,
and
so
then
we
can
decide
whether
what
what
can
we,
what
we
can
do
next
yeah
yeah.
E
I
think
that
the
big
problem
is
the
Upstream
like
distribution.
Community
is
not
active
and,
like
that's
I,
think
the
problems
here
right
so
there,
or
at
least
it's
not
very
active
one
of
the
maintainers
middle
C-
is
quite
active.
I've
been
discussing
a
lot
of
things
with
him
and
I
also
proposed
the
idea
of
potentially
cutting
a
V3
Alpha
release
of
distribution,
so
that
we
have
something
that's
cut
off
the
current
main
branch
and
then
more
breaking
changes
could
still
be
added
to
it
in
that
Alpha
stage.
D
G
And
I
totally
understand:
yeah
yeah
yeah
yeah
go
ahead.
I.
A
Was
going
to
ask
David
if,
if
that
Community
is
not
active
in
terms
of
like
people,
not
contributing
or
people
are
just
too
busy
to
do
other
stuff?
What's
your
feeling,
because
I'm
not
like
attending
any
of
the
the
meetings
for
distribution?
Oh.
E
I'm
not
attending
the
meetings
either.
I
might
be
just
like
a
naive
like
bystander
in
this
case.
So
it's
just
my
observation
and
I
don't
mean
to
offend
anybody
if
it's,
if
it's
incorrect,
but
over
the
past
couple
months,
I've
been
trying
to
get
some
of
the
open
milestones
for
V3
merged
and
and
some
other
things
that
have
come
in
and
it's
difficult
to
get
people
to
actually
review
the
PRS.
B
E
So
if
you
have
like
for
the
milestone
for
V3,
there's
still
22
open
issues,
one
of
my
PR's
I
think
closes
three
or
four
of
those.
Some
of
them
can
probably
just
be
removed
altogether,
so
there's
there's
like
maybe
10
or
or
so
things
that
need
to
get
merged
before
V3
should
be
ready,
but
to
get
those
things
merged.
We
need
multiple
maintainers
to
be
able
to
review
the
PRS.
B
A
So
can
we
can
we
leverage
your
role
in
that
team,
somehow
to
to
drive
that
communication
I'm
super
happy
to
jump
in
and
and
discuss
that
thing
with
the
other
team
I'll
say,
but.
A
Because
if
we
okay,
my
is
a
like
a
cider
outside
of
that
thing.
I
I
see
that
we
right
now
that
we're
discussing
we
have
a
solution
which
actually
fixes
the
hardware
issues
which
are
caused
by
the
main
distribution
and
you
as
a
representative
of
both
sides.
Maybe
you
can
help
out
with
the
with.
A
H
A
To
say
that
I
share
your
your
thoughts
on
merging
something
that
deviates
our
standard
way
of
doing
things
that
that
can
cause
like
a
cascading
other
cascading
issues
like
this,
which
we
cannot
handle
properly
so
I
mean
I'm
in
favor
of
fixing
it
Upstream
first
and
then
try
to
even
if
that
will
delay
a
little
bit
some
functionality.
As
someone
pointed
out
in
the
into
the
pr
that
those
are
like
Enterprise
level,
Readiness
of
functionalities.
G
A
If
you
have
someone
for,
if
you
have
contacts
with
someone
from
that
team,
please
Loop
me
in
in
some
kind
of
males
and
I
can
find
our
way
through
the
cncf.
Maybe
you
can
we
can
look
up
for
support
from
the
cncf
as
the
the
on
top
if
if
the
project
suffers
from
that's
the
reason,
why
I
ask
you
David
about
this
stuff?
If
the
project
suffers
from
insufficient
resources,
one
of
the
one
of
the
tasks
of
cncf
is
actually
to
fix.
That
thing
is
an
issue
long
term.
A
So,
okay,
if
we
can
identify
stuff
like
this,
like
we
had
that
issue
with
the
notary,
for
example,
not
review
one.
A
B
E
I'm
frankly
surprised
that
distribution
is,
it
doesn't
see
much
activity
or
maintenance,
since
it's
being
used
by
so
many
people
like
it's
used
by
Docker
Hub,
it's
used
by
Harbor
I
think
it's
also
being
used
like
for
the
GitHub
container
registry,
like
it's
it's
being
used
as
a
base
registry
for
a
lot
of
big
services
from
companies
with
big
amounts
of
money,
but
somehow
it's
just
not
actually
being
maintained
very
well
and
I'm,
not
sure
why
that
is,
but
it's
just
yeah
I.
Just
like
surprise
me.
A
No,
it's
a
it's
a
great
conversation
to
have
with
with
the
team
itself
who
is
behind
it
and
and
also
with
cncf,
if,
if
that
particular
project
needs
more
intention
in
time
in
terms
of
like
resources
to
work
on
it,
that's
the
role
of
cncf
actually
to
fix
that
kind
of
issues.
So
yeah.
Thank
you
for
that.
So
young
one,
sorry
once
again,
if
I'm
putting
something
in
your
plate
to
to
work
on
it,
but
you
have
my
support.
B
A
This
one
I
I
suppose
David
will
also
jump
in
Happily
to
drive
this
as
Europe
author
of
some
of
these
fixes,
I
think
you're
interested
in
that
to
get
merged
so
yeah
all
right.
Thank
you
very
much.
A
Anyone
wants
to
share
something
else
on
that
topic.
A
I
The
call
yes
right-
yeah
I-
am
here
yeah
thanks
for
having
me
again
so
last
week,
I,
oh
last
week,
two
weeks
last
meeting
I
presented
its
proposal
about
introducing
City
events
and
so
for
those
who
are
not
there
as
the
event
is,
the
proposal
is
about
adding
a
new
format
for
cloud
events
in
the
new
framework
that
was
introduced.
I
Sorry
for
the
cloud
events,
notifiers
and
City
events
is
a
project
within
the
continuous
delivery
Foundation
that
aims
to
standardize
the
format
of
events
in
the
cicd
space,
so
I
have
the
The
Proposal
PR
up
and
thanks
for
the
reviews
on
that.
I
I
The
pr
itself,
it's
relatively
small,
the
the
big
number
of
files.
I
guess,
is
the
what
comes
in
in
Vendor.
But
if
you
exclude
the
vendor,
it's
a
relatively
small
change,
so
the
framework
that
is
available
on
the
hardware
side
allows
to
just
define
basically
a
new
format,
and
so
that's
pretty
straightforward.
I
And
I
have
it
running
here,
so
this
is
my
Harbor
instance.
Hopefully
you
you
can
see
the
screen.
Fine,
so
I
created
a
CD
events
project
here,
where
you
have
some
images
and
I
wanted
to
show
what
what
it
looks
like,
as
you
may
expect.
If
you
create
a
new
webbook,
it
looks
exactly
as
before,
but
there
is
a
new
Option
here
on
the
cloud
events.
I
There
is
also
City
events
available
and
there
you
have
to
pre-provisioned
here
one's
called
CD
events,
which
uses
a
City
events
format
and
points
to
a
request:
bin
URL
similar
for
the
cloud
events,
I
have
it's
using
Cloud
events
and
it's
pointing
us
to
the
same
request,
pin
URL
and.
B
I
I
D
B
I
Yeah,
so
it
was
successful
and
you
can
see
the
payload
data
here
and
if
I
switch
to
the
request,
pin
page
yeah,
you
can
see
there
are
two
new
events.
One
is
point
two
three
CD
events
endpoint
and
you
can
see
this
is
the
seed
event,
so
the
type
is
Dev
I,
don't
know
if
this
is
readable
to
have
cd
events,
artifacts
published
one
zero
one
one
and
the
subject
is
in.
I
We
use
this
package
URL
format,
so
it's
an
Ocio
Arbor
and
then
with
the
URL
that
points
to
where
the
image
goes
and
yeah.
So
all
all
the
details
from
the
original
message
are
here:
you
can
see
in
custom
data.
We
have
the
the
list
of
resources,
the
digest
tag,
resource
URL
and
Repository,
and
then
from
City
events.
We
have
the
context
and
subject
part
so
one
one
thing
that
I
wanted
to
highlight
is
that
in
Harvard,
this
resources
is
actually
an
array
a
list.
I
So
we
in
the
events
I've
seen
there
is
always
one
resource,
but
I
imagine
it
is
possible
to
to
have
multiple
ones
and
so
because
of
that
in
City
events.
Today,
we
support
one
resource
per
per
event.
This
is
something
we
might
change
in
future.
So
what
I've
done
in
the
initial
implementation
is
to
look
at
the
First
Resource
use
that
resource
for
the
subject
of
CD
event
and
then
put
the
complete
list
in
the
custom
data.
I
I
A
Thank
you.
Anyone
any
questions
to
this
one
or
Andrea.
Do
you
need
anyone
to
review
your
PR
for
from
the
for
the
proposal
or
have
you
created
the
pr
with
the
code
submission
already
into
the
main
Repository.
I
Yes,
a
very
big
and
great
to
to
merge
the
The
Proposal.
First
I
I
went
ahead
and
created
a
PR
in
the
in
the
main
repulatory,
so
that
I
could
show
the
demo
but
yeah.
Of
course,
I
can
update
that
and
change
that,
based
on
the
any
feedback
on
the
proposal,
but
I
wanted
to
go
ahead
and
demonstrate
it
to
the
community.
Here.
A
G
Yeah,
thank
you.
We
will
take
time
to
review
the
proposal
and
then
review
the
pr.
Thank
you.
Thank
you.
So
much.
K
Yeah
yeah,
just
a
small
proposal
or
I,
think
it
could
be
one.
K
B
K
So
in
this,
in
this
migration
guide,
you
you
suggest,
to
install
a
local
hardware
instance
set
an
replication,
replicate,
charts,
migrate
them
to
oci
and
replicate
them
back
for
our
teams.
This
is
not
a.
K
This
is
a
little
bit
of
an
oversized
solution
and
what
we
are
done
is
we
provide
a
small,
really
really
small
and
and
and
simple
script
with
a
loop
that
extracts
all
the
charts
from
a
project
and
pushes
them
again
as
a
as
a
bulk
as
an
oci
artifact
in
the
project
back
again
and
I
just
want
to
provide
our
simple
script
as
a
contribution
to
the
community.
K
If
this
is
some
in
some
of
interest
of
you
and
since
this
documentation
is
in
a
GitHub
Wiki-
and
this
does
not
support
pull
requests
as
far
as
I
know,
I
just
have
our
script
as
a
gist
file.
To
my
personal
account,
you
can
have
a
look
at
it
and
you
can
just
give
me
a
signal
if
this
is
okay.
If
this
is
creating.
K
H
B
H
H
Thank
you
yeah.
We
have
the
same
need
for
our
customers
and
I
didn't
talk
about
it
for
now,
because
I
developed
it
I
developed
another
script
like
yours,
but
in
golang
that
do
the
same
thing,
but
a
little
bit
more
complex.
H
You
you
give
you
you
give
to
the
scripture,
URL
user
and
password
and
it
it
fetch
all
projects
from
the
other
API
and
then
it
it
gets
also
the
app
chart
from
the
API
yeah.
Then
it
pull
it
and
push
it.
So
it's
pretty
like
your
script
and
it's
in
go
long,
so
I
I
wanted
to
open
source
it
to,
but
I
didn't
do
it
for
now,
but
yeah
I
I
found
that
the
solution
that
is
proposing
the
wiki
is
a
bit
complex.
Maybe
so
simple
solution
would
be
nice.
K
Great
cool,
so
how
can
we
proceed
from
this
point
on.
K
Or
Yen
I
think
you
would
like
to
say
something
or
you
wanted
to
say
something.
G
Oh
yeah
I
just
want
to
confirm:
does
this
Great
support,
post
Docker
compiles
and
a
ham,
sharpest
installation.
G
Doing
this
I
mean
that
if
you
go
to
the
Wiki
page
for
migration,
we
we
mentioned
that
the
the.
If
your
Hardware
was
deployed
by
through
a
Docker
compose
you.
Can
you
just
follow
the
stack
one,
two
three
file?
But
if
the
harbor
is
deployed
through
the
Hampshire
yeah.
K
G
K
K
C
K
For
feedback,
how
can
we.
J
Proceed
is
this:
is
it
a
Wiki?
So
if
it's
a
Wiki
I
would
say
you
can
just
go
ahead
and
change
the
documentation
there.
If
this
is
okay,
I.
K
J
K
Add
it
to
the
to
the
end
to
the
bottom
of
the
of
the
description.
If
this
is
okay,.
D
B
J
A
A
So
people
can
actually
be
aware
of
this
one
not
to
yeah,
because
if,
if
you
stick
it
in
the
wiki
I'm,
not
sure
who
is
reading
that
Wiki,
except
if,
like
this
group
of
people
that
we're
on
that
call.
A
K
K
A
Anyone
on
on
that
topic,
we
have
10
minutes
left.
So
if
nobody
has
questions
about
this
one
or
any
suggestions
by
the
way,
just
one
thing
Thomas,
you
you
mentioned
you've
created
something
right
on:
go:
link,
yep
yeah.
Do
you
care
of
adding
your
stuff
as
well
there?
If
you
think
it's
going
to
be
useful
for
others,.
H
Oh
yeah
I
like
but
I,
don't
know
where
to
put
the
code
of
this
so
I
I
for
now,
I
asked
my
company
to
to
create
a
new
project
inside
it
to
to
put
my
my
script,
but
I
don't
know
if
we
can
put
it
directly
inside
the
arbor
repository
I.
Don't.
A
Know
is
it,
is
it
ovh,
proprietary
thing
or
IP
kind
of
stuff
or
no.
H
I
don't
know
I
developed.
It
took
my
time
at
ovh,
so
I,
don't.
A
Yeah,
at
least
that's
that's
the
policy
yeah
at
VMware.
If
you
develop
something
that
solves
business
problems
and
and
are
not
developed
under
the
open
source
license
or
IP
of
the
company,
but
I
think
it
makes
sense
for
stuff
like
this,
because
Stefan
wants
to
share
that
script
and
you
as
well.
Maybe
we
can
create
like
a
additional
repository
into
the
organization
like
of
scripts
or
I.
A
Don't
know
some
I,
don't
know
how
to
call
it,
but
like
a
useful
tools
or
Joy
whatever
and
with
less
restrictions
of
what
we
add
there,
which
is
with
disclosure
that
it's
like.
That's
how
we
do
it
or
that's
how
I
I
did
it
and
just
to
share
to
share
like
collaboration,
kind
of
kind
of
stuff
for
scripting
and
some
some
Snippets
of
stuff
yeah.
H
Okay
I
check
with
my
manager
for
to
be
sure
yeah.
A
G
Stefan,
you
can
just
create
a
PR
against
to
the
tool
folder
and
under
the
source
code.
We
have
a
two-folder
in
a
source
code
and
there
are
a
lot
of
scripts
and
utilities
scripts.
You
can
just
some
epr
to
that
to
there.
So
then
we
can
review
your
script
and
maybe
we
can
just
add
a
link
in
the
wiki
to
the
script
once
we
we
have
accept
appear.
K
K
It
I
can
I,
can
I
can
open
MPR
and
edit.
There
would
be
also
nice
place
for
the
Golan
code.
I
think.
Yes,.
A
My
idea
was
to
avoid
adding
such
scripts,
which
are
not
like
widely
Community
adopted.
I
would
say
it's
like
okay.
Do
you
know
why
just
an
idea
that
I'm
explaining
why
I
I
propose
to
have
a
different
repository,
less
restrictive
one
like
just
for
collaboration,
because
people
can
easily
understand.
Oh,
that
script
is
in
the
main
repository,
so
it
should
be
always
working
and
it's
like
Stephanie.
It's
not
working.
A
Fault
and
then
we
start
like
a
cascading
problems,
I'm
just
trying
to
avoid
that
kind
of
situation
if
yeah,
especially
in
the
in
the
that
super
sensitive
area
of
OCR
charts
but
I,
agree
yeah
but
yeah.
Let's,
let's
make
it
to
the
main
repositories
suggested
and
see
how
it
goes
and
if
we,
if
we
think
that
there'll
be
a
need
for
more
scripts
around
like
database,
optimizations
or
I,
don't
know
ready
kind
of
stuff.
We
can.
We
can
trigger
another
repository
and
and
work
it
out
there.
A
J
J
I
would
like
to
highlight
again
a
bit
the
topic
of
the
garbage
collector.
You
know
and
we
are
running
a
few.
You
know
terabytes
I,
don't
know
how
many
terabytes
there
are
and
the
problem
is
with
it
with
the
garbage
collection
is
that
it
fails
to
run.
If
you
need
to
you
know,
if
you
need
to
garbage
collect
a
few
things
or
if
it's
extend
exceed
a
certain
size,
so
I
think
it's
like.
J
If
it
was
one
terabyte
it
it
fails,
and
there
have
been
quite
a
few
Solutions
already
trying
to
implement
this
implement
or
find
a
solution
there
and
I've
summarized
it
a
bit,
and
we
have
seen
quite
a
few
Solutions
quite
a
few
issues
over
the
time
and
the
building
you
know
subsequently
solved
one
after
the
other,
but
there
are
still
a
few
issues
that
prevent
a
reliable
GC
run.
So
currently
I
think
we
are
seeing
on
one
instance.
J
That's
you
know
about
like
three
terabyte
of
cleanup
is
the
problem
with
it
just
hangs
right
the
garbage
collection
just
hangs,
and
there
is
no
feedback,
no
information,
no
debug
message.
It
just
stops
working
and
there
is
no
any
feedback
loop
and
then,
after
the
24
hours
or
we
set
it
to
72
hours,
it
just
Idols
out
and
then
stops.
J
So
you
could
really
see
that
nothing
happens
there
right
and
it's
in
a
block
State
we're
trying
to
find
out
where
the
problem
is
but
I
found,
for
example,
already
a
few
issues
on
that
on
Harbor
on
the
Harbor
on
the
harbor
side,
so
I
know
people
already
have
a
few.
Quite
a
few
people
reported
the
problem
and
there
is
a
actually
an
interesting
proposal
from
you
know
from
from
a
few
community
members
trying
to
solve
this
issue.
Let
me
just
share
it
quickly.
J
Okay,
so
you
should
see
now
the
the
this
one.
You
know
this
is
a
quite
interesting
proposal,
because
it
IT
addresses
a
few.
A
few
issues
here.
J
Right
and
I
tried
to
reach
the
this
user
so
far,
not
not
very
successful,
but
he's
he's
trying
to
address
the
the
problem
on
multiple
levels
right
so
he's
trying
to
address
it
on
a
Docker
distribution
level
and
also
on
the
on
the
Harbor
level,
and
he
makes
a
few
quite
a
few
proposals,
and
it
seems
that
for
for
him
that
the
solution
works-
and
there
is
also
Danny
who
is
also
working
on
a
GC
solution.
He
made
a
few
proposals
to
to
Harbor
and
one
was
actually
accepted.
J
This
was
not
this
one
yeah.
J
Yeah
so
this
proposal
or
this
PR
it
does
already
solve
a
few
problems.
So
we
have
seen
these
problems
in
the
past
and
it
does
solve
it.
Definitely,
but
it's
not
kind
of
a
work
does
not
solve
all
the
problems
that
are
there
and
I.
Would,
you
know
see,
try
to
figure
out
how
we
can
address
the
problem
again
in
Harbor
and
finally
find
a
a
good
solution
for
the
garbage
collection
part.
J
J
So
this
is.
This
is
from
you
right
and
I
didn't
understand
what
the
the
solution
here
is.
Could
you
elaborate
a
bit
on
on
the
pr
here.
G
J
G
Yeah,
a
lot
of
users
from
Community
reported
the
performance
problem
on
the
garbage
collection
yeah.
So
then
2009
we
decided
to
introduce
the
parallel
Edition
in
the
garbage
collection.
So
this
PR
is
just
to
resolve
this
problem.
Yeah
and
and.
G
The
pr
is
not
quite
easy
to
understand
and,
and
we
should
narrow
down
the
the
parallel
just
in
the
sweep
phase
rather
than
both
Mark
and
sweep,
because
that
the
the
most
of
time,
the
garbage
collection
will
is
handling
the
filed
Edition
rather
than
the
the
database
status
change.
So
we
just
narrow
down
the
the
coaching
in
the
sweep
function
and
and
do
now
to
make
the
the
logic
to
be
more
complicated
and
easy
to
maintain.
G
So
you
can
just
I
I
will
leave
comment
in
this
PR
and
you
can
have
to
renew
my
PR
account.
J
G
B
J
G
Maybe
you
can
just
focus
on
the
a
garbage
collection.go
yeah
this.
H
B
G
J
J
J
J
Okay,
that's
it
for
my
side.
A
Thank
you,
madam
we're
four
minutes
over
the
hour.
So
let's
do
it
quick,
Thomas!
Your
index
is
about
the
talks
about
not
very
cosine
migration.
H
I
think
we
talked
about
it
two
two
Community
meetings
ago,
I,
don't
remember
who
talked
about
it,
but
we
we
say
that
it
could
be
useful
for
our
users
to
know
how
to
migrate.
So
we
I
know
that
we
can't
migrate
to
signatures,
but
at
least
explain
how
to
how
to
to
handle
the
notary
deletion.
D
J
H
B
G
So
because
that
we
cannot
migrate
the
signature
from
notary
to
cosine,
so
the
user
has
to
use
cosine
to
sign
the
artifact
again
and
leverage
the
code
size
nature
to
to
do
their
pool
policies.
Yes,
but
from
Harborside
we
will
provide
a
Wiki
page
to
describe
the
situation
and
then
the
help
user.
To
my
agree,
I'm
the
migration
that
I
mean
there
is
just
remove
the
signature
from
not
three
and
use
cosine
to
sign
them
again.
H
Okay,
and
can
we
provide
this
script
that
could
sign
also
old
images,
Maybe,
not.
G
We
may
maybe
just
provide
a
example
for
for
them
to
sign
one
artifact
with
cosine
and
they
can
repeat
the
steps
to
send
any
artifact.
They
want.
H
A
D
Yeah
I'll
be
quick,
so
at
the
kcd
Mumbai
I
met
a
person
from
UFC
cloud
and
they
provide
ephemeral
VMS
for
pull
requests
so
I
think
they
had
a
chat
with
the
harbor
maintainers
at
some
point
earlier
as
well.
So
the
person
asked
me
that
if
Arthur
would
be
interested
in
a
sort
of
that
now,
I
can
sort
of
facilitate
that
work
towards
it.
So.
A
You
can
check
the
the
PRS
and
the
issues,
it's
a
like
a
super
long
discussion
around
this
one
and
we
met
with
Vadim
me
and
buddy.
We
spoke
with
them
on
keep
con
artistam
actually
right
now
they
have
already
set
up
that
for
Harbor,
it's
actually
there's
a
PR
opened
and
it's
working,
but
they
postpone
that
until
they
introduce
the
home
process
of
installation
Harbor
because
they
right
now,
then
they
do
not
support
help.
So
that's
like
on
pause
for
now.
B
J
A
Also,
it's
very
useful
when
we
when,
when
we
do
UI
kind
of
stuff
ux,
but
if
you
want
to
do
the
core
development,
it's
super
complicated
to
access
logs
or
actually
to
debug
containers
and
such
so
it's
it
has
its
use
cases,
but
it's
not
solving
for
all
thanks
for
bringing
that
up
we're
keeping
contact
with
them
actually
every
month
checking
what's
going
on
there,
so
yeah!
It's
at
some
point.
We're
gonna
merge
that,
because
I
we
liked
it
very
much.
A
It's
super
useful
for
some
stuff,
so
I
think
we
will
have
it
just
let
them
finish.
The
implementation
on
home.
A
A
It
yeah
and
stone
you
wanted
to
introduce
something
I'm,
not
sure
how
how
long
that
will
be
I'll
see
you
with
us.
A
A
So,
okay
Stone,
can
we
postpone
that
for
two
weeks
please
then
I'm
or
or
I
can
and
then
we
you
can
stop
the
the
the
recording
afterwards.
Sorry
for
that
I
just
yeah.
K
A
A
Okay,
I'm
an
elite
now,
but
you
can,
you
can
do
the
you
can
do
the
the
the
introduction
and
I'll
check
it
afterwards.
All
right.
A
K
You
can
we
postpone
to
the
next
quickly
commit
me.
A
Sure
sure,
if
it's
working
for
you
even
better
I'm,
sorry
for
that
that
we
just
I,
think
my
my
notification
into
the
mailing
list
works
and
we
have
more
attendance.
So
thank
you
very
much
everyone.
No
please
DM
me
in
slack,
so
I
can
answer
your
question
about
the
the
Enterprise
support
that
you're
looking
for
I
can
give
you
some
ideas
all
right.
A
Thank
you.
Thank
you
very
much,
I'm,
sorry
for
cutting
it
off,
but
it
can
go
on
and
on
so
all
right.
Thank
you.