►
From YouTube: Kubernetes SIG Service Catalog 2019-10-21
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
C
Something
something
yeah
basically
I
put
that
point
on
our
agenda,
so
it
was
reported
by
one
of
our
users
that
we
have
problem
with
our
charts.
So
basically,
if
you
hit
a
home
search,
then
you
will
see
the
newest
version
of
the
circle
log.
When
you
is
execute
in
style,
then
you
install
the
old
one,
basically
I
debate
that
issue
and
you
basically
that
there
is
a
problem
on
the
home
side.
C
The
second
you
can
click
on
it
and
basically
there
is
a
probe
issue.
We've
kind
of
the
proposal
is
there
like
an
umbrella
issue
for
tracking
the
service
in
Scotts,
cascading
deletion
and
basically,
what
I
did
I
just
go
through
all
discussion
about
his
service
installation
and
and
also
few
issues
from
the
face
to
face
meetings
and
what
we
did.
We
also
discussed
that
in
our
team-
and
there
is
a
some
kind
of
a
summary
which
approach
we
can
with
we
can
choose
at
the
beginning.
C
It
will
be
like
a
Daffy
alpha,
so
still
have
a
user
car
and
validate
that,
and
we
can
remove
it
a
bit
if
that
thing
will
not
fit
into
their
flow.
But
basically
what
we
want
to
do
is
just
implement
that
at
the
Alpha
feature,
hide
that
through
caffeine
deletion,
a
feature
flag
and
and
release
it
in
the
next
version.
So
anyone
can
just.
C
Okay,
so
basically
the
first
option
was
to
reuse
the
Cascade
flag
from
the
cube
steel,
and
it's
really
connected
with
the
garbage
collector
from
the
kubernetes,
and
that
guy
allows
you
to
remove
your
object
with
two
options.
First,
one
is
a
foreground.
Is
there
really
is
the
thing
that
we
really
want
to
use
so,
basically,
before
your
services
that
will
be
deleted,
service
binding
needs
to
be
delicious,
needs
to
be
triggered
and
they
need
to
be
deleted
up
front
with
the
park
background
one
previously.
C
What
need
to
do
is
delete
the
main
object,
so
in
that
case
service
instance
and
then
the
service
binding
can
be
deleted.
So
is
the
different
scenario
and
in
our
case
it's
like
a
not
an
option
to
go
because
it's
not
fit
to
our
business
logic
right
to
fist,
delete
the
instance
and
then
take
care
of
this
service.
Binding,
collector
service
bindings
right.
B
C
If
you
will
go
with
the
keep
steal
cascade,
you
still
will
have
like
an
one
option,
just
to
always
have
cascade
set
to
for
true
right,
because
setting
it
to
false
will
break
our
our
flow.
Because
you
are,
you
can
implement
that
in
service
catalog,
but
it
will
somehow
rate
your
business
logic
right
just
to
delete
a
service
instance.
That's
the
first
one
and
then
take
care
of
the
connected
service
binding.
This
I
don't
see
any
logic
behind
up
in
in
service
in
communities.
C
A
No
other
way
around
this
I
feel
like
deleting
a
service
instance
and
getting
back
an
error.
Hey
the
service
instance
still
has
service
bindings
is
acceptable.
Behavior,
so
I
feel
like
pulling
the
rug
out
from
underneath
people
on
that
is
kind
of
shitty
and
I'm,
not
necessarily
certain
that
that's
a
better
default
behavior,
then
I
rather
I'm,
not
certain
that
it
is
a
better
default
behavior
to
be
like
hey,
delete
the
service,
10
cents,
and
it's
like
okay,
I'm
gonna
delete
these
other
service.
Spottings.
A
C
A
C
One
option
needs
to
go
with
the
cascade
because
you
GBL
does
not
have
something
like
the
generic
parameters
that
can
be
pushed
to
your
service
instance
to
any
kind.
So,
basically,
you
can
only
raise
the
Cascade
and
we
are
not.
We
are
not
able
to
reuse
it
because
of
the
contract.
On
the
background
strategy,
it
was
like
so.
C
B
C
But
basically,
a
lot
of
people
still
don't
have,
as
we
cat
should
be
some
kind
of
the
helper
tool.
But
the
most
scenario
that
you
want
to
execute
should
be
available
through
chip.
Cli
thing
right.
The
easy
cat
should
be
only
like
have
a
better
experience
right,
but
something
like
already
provides
something
like
that:
better
printing,
the
columns
with
the
additional
information
about
the
external
I'm
and
so
and
so
forth,
but
even
in
the
even
with
the
keep
still.
You
also
have
something
like
that,
because
you
have
additional
critter
comes
with
the
keeps
TL.
C
Additional
field
on
this
under
spec,
saying
Dilys
behavior
before
to
fail
on
bindings
but
can
be
said
also
to
force
cascade,
was
also
discussed
and
the
last
one
as
as
I
mentioned,
it
was
the
only
safe
from
cascade
deletion.
And
what
this
was
this
also
this
statement
from
the
last
face-to-face
meeting
handle
year
ago
on
July.
So
there
was
the
something
that
that
was
mentioned
already
by
Morgan
on
the
pre
request.
It
was
the
matrix
deletion
of
matrix.
C
So,
basically,
right
now
we
have
some
sky
not
as
I
describe
the
expectation
that
also
from
our
user
is
by
deleting
a
credits,
API
resource
results
in
total
deletion
of
the
resources
underlying
resources,
and
also
that
they
want
to
somehow
do
that
from
the
UI.
So
we
just
wanted
to
do
to
make
it
as
a
some
kind
of
the
native
feature
in
the
sense
Kellogg.
C
C
So
maybe
then
you
can
also
put
your
opinion
about
it
because,
as
I
said
at
the
end
of
that
issue,
there's
also
we
are
waiting
for
your
feedback
and
use
cases.
This
approach
happens
to
break
your
flow
here
because
release
it's
not
like
to
say
except
table
for
us,
so
I
mean
to
somehow
come
up
with
the
solution
that
everyone
will
be
happy.
A
C
Quite
easy
to
do
it
quite
problematic,
so
yeah,
so
so
I
can
just
modify
that
and
still
of
course
you
do
be
still
in
alpha
state.
So
we
don't
give
an
equal
quadrant
is
that
it
will
be
promoted
to
better
or
GA
its
kind,
some
kind
of
the
Alpha
so
just
waiting
for
the
feedback,
and
if
you,
if
it
will
be
good
and
we
make
it
normal
but
I
think
that
it
I'm
not
expecting
to
change
that
for
in
0.3
race.
I
think
that
it
will
be
maybe
in
0
or
R.
C
Okay
but
yeah,
but
at
least
we
need
to
study
something
because
we
will
just
want
to
use
that
future
future
in
our
releases,
and
it
can
be
also
enough
for
us
the
problem.
So
thanks
for
reviewing
that
yeah,
there
is
a
few
just
information
about
the
recently
purchased
merge
pull
request.
One
was
about
the
Travis,
so
we
only
master
branch.
We
just
executed
a
few
times
build
action,
dockable
action.
It
was
unnecessary,
so
it
was
chained
and
also
right.
Now
we
have
information.
C
What
kind
of
the
build
is
executed
on
given
uneven
stage,
so
you
can
see
already
something
like
hey,
I'm,
building
images
for
ARM
architecture,
for
MD,
64
and
so
on.
For
that
guy
just
improved
the
immigration
process.
We
just
fixed
the
retry
function
and
added
additional
error
wrapping.
So
right
now
we
also
know
the
context
of
the
error.
Previously,
it
was
quite
misleading
because
you're
getting
some
errors-
and
you
didn't
know
if
it
comes
from
the
create
action
or
a
deduction,
so
it
was.
A
C
And,
and
also
it
was
a
pre-press
from
the
community,
someone
just
updated
our
deployments
to
support
the
newest
British
version
and
they
also
asked
when
we
want
to
release
new
system
up
with
that
think
so
I
also
they
was
going
on
or
agenda
I've
seen
that,
after
after
merging
the
Cascade
deletion,
then
we
can
release
the
new
better
of
the
source
catwalk
but
yeah.
These
need
to
be
a
justice
with
with
our
test.
C
And
also
one
ish
I
just
created
an
issue.
The
good
first
issue
for
someone
from
the
committee
that
one
basically,
some
of
someone
from
committee
created
recently
an
issue
to
just
execute.
He
executed
static,
check
and
some
other
tools
to
just
improve
our
use
on
the
per
request
to
just
improve
our
source
code.
So
I've
just
merged
it,
but
as
a
follow
up,
I've
created
that
task
to.
C
To
clearly
say
that
we
need
to
somehow
clean
up
what
we
are
currently
doing,
because,
right
now
our
scripts
for
testing
are
spread
over
few
directories.
We
didn't
don't
have
creased
strategy
for
doing
that.
So,
in
my
opinion,
should
be
a
little
bit
refactored,
something
similar
to
what
what
operators
have
or
settlement
under,
because
it's
quite
clear
to
follow
and
and
and
then
I
think,
is
a
really
good
for
this.
For
someone
from
the
community,
oh
yeah
and
the
traumatic.