►
Description
CNCF Harbor's Operator Meeting
A
A
So
with
the
project
I've
been
trying
to
get
someone
to
say
something
but
I'm
quite
confused.
So
let
me
show
you
something
can
I
share
my
screen,
I
hope
I
can
share
my
screen.
A
Let's
see
because
oh
my
God
I
can't
share
every
single
window.
A
Whatever,
let's
try
this
sure,
can
you
see
my
screen?
Yes,
yes,
yeah.
So,
what's
confusing
to
me
is
under
projects
in
projects
classic
there
was
a
release
1.4
plant,
yes,
and
there
were
some
people
working
on
it,
but
those
people
like
young
one-
they
don't
respond
for
like
two
months
now
and
I
have
no
idea
and
and
bit
SF
is
also
one
of
those
people.
I
have
no
idea
if
they
intend
to
continue
working
on
the
project
or
not
or
what's
the
idea
here.
Okay,.
B
Because
so
it's
my
company,
that
that
will
is
the
arbor
operator
first.
B
The
people
that
worked
on
it
isn't
working
in
my
company
anymore,
okay
and
after
that
we
released
it
in
to
the
cncf.
It's
VMware
that
started
to
work
on
it
that
that
to
the
project
and
and
developed
it,
so
we
were
still
in
contact
with
them
at
first,
we
were
in
contact
with
Stevenson,
so
that
was
the
leader
of
Arbor
and.
A
B
Followed
our
ballpark
project
2.,
but
you
know
VMware,
has
been
bought
by
another
company
yeah
last
year,
I
think
yeah,
and
from
that
moment
the
new
company
wanted
to
to
keep
a
lot
of
people.
So
they
stopped
the
development
of
rubber
operator.
So.
A
B
Developers
that
that
we
are
working
on
it
were
moved
on
other
projects,
okay,
so
on
the
VMware
side,
nobody
have
time
now
to
work
on
on
it.
So
so
they
all
abandoned
the
project
and
yeah
I
was
in
contact
with
Yan
at
first
and.
A
B
A
A
Okay,
so
here,
because
now
the
bigger
problem
is
when
they
did
whatever
they
were
planning
to
do
they.
They
did
a
thing
which
is
causing
us
a
lot
of
issues.
They
created
V1
beta1
of
the
cids
and
a
V1
Alpha
3.,
so
there's
currently
V1
Alpha,
3
and
V1
beta1,
and
this
is
where
our
issue
comes
in,
because
both
of
these
are
massive.
A
So
if
we
ever
want
to
upgrade
from
V1
Alpha
3
to
V1
beta1,
this
is
the
correct
thing
to
do
right,
but
right
now
the
problem
is
now
we
have.
We
want
beta1
and
we
want
Alpha
three
and
they
are
both
like
I
have
not
tried
to
create
a
diff,
but
from
what
I
can
see
they're
very
similar
there's,
not
a
lot
of
difference
between
the
two
and
this
causes
us
to
put
all
the
crds
in
the
helm
chart,
and
now
we
have
the
big
Cid
in
there
twice.
A
So
we
have
it
for
we
don't
know
if
I
want
and
we
no.
We
want
Alpha,
Four
and
V1
beta1
yeah,
and
this
causes
this
ham
chart
to
be
just
way
too.
Big
and
I
have
no
idea.
A
What
to
do
because
I
don't
know
what
they
were
thinking
when
they
did
this
and
I
also
don't
know
how
to
make
the
crd
smaller
now,
because
I
looked
at
it
like
there's
a
lot
of
options
in
the
cids
and
I,
don't
know
if
any
of
them
are
unused
or
not,
and
now
we
are
in
the
situation
where,
okay,
we
have
this
massive
Cid
in
two
versions.
A
A
B
Think
that
in
the
v1.1
version
there
was
always
the
one
beta.
B
So
maybe
we
can
delete
the
one
alpha
three,
but
do
you
think
it
will
give
us
some
time,
but
what
happen?
If
later
we
decide
to
release
a
new
version,
we'll
have
the
same
problem
so.
A
A
B
A
And
it's
not
clear
how
to
make
them
smaller.
There
was
supposed
to
be
a
change
away
from
notary
that
would
remove
a
bunch
of
cids,
but
it
would
add
other
ones
so
that
doesn't
help
and
I
was
thinking
about
what?
What
can
we
do
here
like
we
can
remove
the
cids
from
the
harm
chart,
but
then
there's
no
way
to
install
via
Helm
anymore,
because
you
would
have
to
have
some
weird
thing:
do
you
know
if
you
internally
your
company,
if
they
use
the
helm
chart
or
do
they
use
something
else.
B
Yeah
we
are
using
the
M
chart,
but
I
think
we
are
installing
the
crg
apart.
So
okay,
so
if
we
remove
it
from
the
charts,
it
will
not
impact
us.
Okay,.
A
So
this
is
then
what
I
would
just
suggest.
We
would
do
something
like
we
remove
the
CID
from
the
handshot
completely
we
say
the
helm
chart
only
manages
the
deployment
and
the
CID
needs
to
be
managed
separately.
It's
not
going
to
be
very
popular
I.
Think
many
people,
but
I
don't
under
like
I,
have
no
idea
what
else
we
can
do
at
this
point,
because
at
least
I
don't
understand
which
options
in
there
could
be
removed
and
I
like
I,
worked
with
the
operator
quite
a
bit,
but
we.
B
A
Crj
some
have
another
Hampshire
dedicated
to
the
cids,
some
of
them
to
stop
giving
him
chats.
Some
of
them
like
this
is
a
rare,
ish
problem
like
not
too
many
people
have
such
a
big
CID.
Okay,
and
that's
also
because
if
you
look
in
here,
we
have
like,
for
example,
if
I
go
to
the
Chart
Museum
type.
A
But
this
is
like
even
less
popular
I
think
to
split
up
the
the
the
project
a
bit
more
because,
if
you're
seeing
when
we
when
we
deploy
right,
we
have
multiple
components
you
can
deploy
like
you.
Can
you
can
just
deploy
The
Hub
operator,
but
you
can
also
run
it
with
mini
or
whatever,
and
then
your
CID
is
controlled
through
the
hardware
operator
CID.
Instead
of
you
know
just
directly
creating
the
menu
or
stuff
or
like
having
some
abstraction
or
the
the
mini
or
configuration
as
part
of
the
hardware
CID.
B
Yeah,
of
course,
yeah
on
our
side.
We
don't
choose
also
operator
well,
we
use
the
zarando
pass
gas
operators,
but
we.
A
A
B
A
Happened-
and
this
is
really
a
problem
because
I
had
to
follow
the
project
to
see
the
pull
requests
and
we
have
to
have
two
approvers
to
merge
a
pull
request
right
now.
So
if
one
of
us
creates
a
pull
request,
I
don't
think
we
can
match
it,
because
only
you
can
approve
my
pull
request
and
then
we
need
to
approve
us.
We
can't
merge
the
pull
request.
B
B
Yeah,
why
not
when,
when
the
pull
request
is
undone
by
one
of
us,
we
can.
We
can
tell
that
we,
we.
A
A
A
A
A
It
looks
pretty
good
to
me
so
essentially
what
this
person
did
was
like
you
know
how
Harbor
operator
only
creates
the
hardware
instance,
but
doesn't
configure
it
yeah
this
person
added
configuration,
so
they
added
a
pull
request
of
interacting
with
the
harbor
API
to
configure
Hardware
projects
through
Hardware
operator,
which
is
something
we
wanted.
Also
we
as
a
company
okay-
and
you
review
this
for
request
and
also
check
from
your
site
if
it
makes
sense.
A
A
A
And
then
there's
this
one
that
is
currently
active.
This
is
the
upgrade
controller.
Runtime
version.
Yeah
I
will
try
to
look
at
this
later
today
or
like
tomorrow,
but
this
person
is
trying
to
upgrade
the
controller
runtime
I,
don't
know
if
it's
gonna
work
because
they
are
creating
or
they're
updating
in
a
huge
step.
So
we
were
in
control
our
runtime
version
where's
the
file
yep.
A
B
A
yeah,
so
it's
it's
a
really
big
jump
and
now
the
problem
where
I
couldn't
really
help
them
is
they
run,
make
generate
to
regenerate
the
files,
but
it
created
a
div
of
like
14
000
likes,
and
that
does
not
seem
correct
to
me.
A
A
Something
is
definitely
weird:
I
will
try
to
look
at
it,
but
if
you
can
also,
if
you
have
a
quick
idea,
that
would
be
great
otherwise
I
will
get
to
it.
Probably
this
evening.
A
Yeah
no
problem.
Okay,
then,
let's
do
it
like
this.
Let's
just
do
like
a
to
maintain
a
situation,
because
otherwise
I
don't
think
we're
gonna
we're
gonna
get
very
far,
and
if
the
VMware
people
are
not
working
on
it
anymore,
then
we
should
also
I
would
try
and
come
up
with
something
what
we
do
for
1.4
yeah,
because
then
I
will
just
restructure
this,
because
it's
like
none
of
this
makes
a
lot
of
sense
to
me,
like
I.
Try
to
read
through
these
updates.
A
B
B
I
think
that
the
two
well
one
of
the
two
issues
in
progress,
the
second
one
yeah
yeah,
it's
a
subject
that
we
that
a
colleague
of
mine
in
my
company
took,
but
she
never
took
time
to
work
on
it
and
now
he's
not
working
my
team
anymore.
So,
okay,
so
we
can
move
it
in
the
to-do
to.
B
B
Yeah
I
think
so
he's
in
China,
so
yeah,
probably
okay,
well
I
think
we
can
do
is
1.4
without
its.
This.
A
A
It's
like
what
is
this
I
have
no
idea,
but
yeah
I
will
do
that
clean
up
and
then
maybe
in
two
weeks
up
until
or
like
hold
on,
will
I
be
there
in
two
weeks.
I
can
try
and
come,
but
I.
Think
no
I
should
be
there
in
two
weeks
again.
A
So
I
will
make
sure
we
do
the
changes
of
like
the
structure
of
the
project
within
the
next
two
weeks.
Remove
some
of
the
maintainers
clean
up
all
these
things
we
just
discussed
and
then
maybe
figure
out
how
we
go
forward
about
this.
A
So
once
the
projects
are
cleaned
up
and
we
kind
of
know
what's
even
in
here
yeah,
then
we
can
kind
of
check
what
we
do
next,
but
I
would
say
that
we
do
that
in
two
weeks
and
for
now
we
focus
on
these
open,
pull
requests
to
at
least
support
people,
making
changes
because
I
feel
really
bad.
I
like
I,
became
a
maintainer
of
the
like
contributing
like
three
pull
requests,
and
now
people
are
making
really
complicated,
pull
requests
and
I'm
like
I,
don't
know
if
this
makes
any
sense,
sometimes
yeah.
B
I,
never
I
think
I
never
opened
the
pull
request,
or
maybe
a
simple
one
and
now
I'm
reviewing
pictures.
So
it's
it's
kind
of
hard
yeah.
A
B
I
have
no
official
time
on
it
for
now.
Well,
for
the
to
explain
to
you,
we
are
using
the
rubber
operator
to
deploy
the
Registries
of
our
customers,
but
when
I
so
I
I
arrived
on
the
project
like
almost
two
years
ago
and
when
I
arrived,
we
were
we
were.
We
had
a
lot
of
legacy
and.
B
Deploying
Registries
in
2.0,
so
we,
which
was
always
two
years
ago,
very
old
version,
so.
A
B
A
B
From
the
end
of
summer,
we
will
have
to
do
on
our
side.
We'll
have
to
do
something
to
to
have
at
least
the
1.4
release
working
so.
A
B
That
time
I
will
it
will
be
easier
for
me
to
to
to
have
some
time,
because
if
the
release
is
not
here,
we
I
can't
separate
Arc.
Also.
A
B
B
A
Gonna
be
a
bit
hard
with
just
the
two
of
us,
but
let's
see
yeah
from
my
side,
there's
also
people
that
are
interested
in
contributing,
but
the
problem
I
had
so
far
is
like
it's
very
hard
to
convince
people.
If
you
tell
them
hey,
there's
a
there's
existing
contribute
maintainers,
but
they
never
answer.
A
So
I
need
to
do
some
advertisement
internally.
That
people
contribute
a
little
bit
more
because
we
also
use
server
operator
and
we
worked
with
another
company
to
have
a
project
that
kind
of
worked
towards
managing
Harbor
with
Harbor
operator
on
our
clusters
on
some
of
our
clusters,
and
we
had
to
do
a
lot
of
fixes
on
Hardware
operators.
So
if
you
look
here,
there's
like
a
time
here
where
I'm
doing
and
all
of
these
changes
so
from
November
December
and
to
January
and
March.
All
of
these
changes
were
related
to
that.
A
I
think
this
was
really
difficult
because
we
couldn't
get
anyone
to
approve
our
changes
and
we
fixed
a
lot
of
things
already
like
here
to
uploading,
upgrade
spillion
and
client
go
to
like
the
latest
versions,
so
I
think
2.6
of
Harbor
should
just
work
now
from
Maine,
but
we
need
to
test
them.
A
I
think
2.6
just
works
already,
but
we
would
have
to
test
it
and
I
don't
know.
Do
we
have
capacity
to
test
internally,
our
operator
or
how
does
it
work
for
you
like?
If
you
wanted
to
create
I,
don't
know
an
environment
where
you
test
your
use
cases.
B
Yeah
I
can
have
those
the
environment,
but
I'm
not
really
I
didn't
test
it
really
before
well,
I
yeah.
It
will
take
some
time
to
me
to
test
it
that
yeah
that
I
can
work
on
it
because
I
know
when
I
test
it
I
tested
in
almost
my
only
one
month
of
my
company
with
all
my
stuff,
I
need
to
have
a
smaller
environment
just
test
the
operator.