►
Description
Kubernetes Storage Special-Interest-Group (SIG) Volume Populator Design Meeting - 04 May 2021
Meeting Notes/Agenda: -
Find out more about the Storage SIG here: https://github.com/kubernetes/community/tree/master/sig-storage
A
There
we
go
okay,
so
hello,
welcome.
This
is
the
kubernetes
six
storage
volume,
populators
weekly
meeting
the
main,
the
main
thing
that
we're
trying
to
get
done
is
get
the
the
cap
merge
before
the
enhancements
freeze,
which
is
an
incredibly
short
deadline
for
the
122
cycle
I
reached
out
to
tim
and
he
actually
responded
back
and
said
that
I'm
I'm
in
his
queue
and
he
will
get
to
me
whatever
that
means
so
it
I
guess
it
sounds
promising
we
still
have
about.
A
I
think
what
nine
days
until
the
actual
freeze,
what
I
wanted
to
do
today,
let's
see
here
yeah,
I
didn't
update
the
indenter,
yet
I
wanted
to
go
through
the
actual
changes
to
the
cap
so
to
this
one,
if
you
can
see
my
firefox.
A
Yep,
okay,
so
what
what
timid
wanted
us
to
do
was
to
split
apart
the
the
the
original
cap
into
just
the
data
populators
and
the
separate
kep
for
the
backwards
and
compatible
change,
so
they
can
be
tracked
separately
and
I
actually
have
two
separate
pr's.
I
have
not
touched
this
particular
cup
since
the
121
cycle,
but
I
think
it's
still
all
all
good,
so
it
has
its
own
prr
reviewer.
A
I
guess
I
I
don't
know
when
the
when
the
pr
review
happens
relative
to
the
approval
from
api
reviewers
if
they
need
to
happen
in
parallel
or
what
but.
B
You
only
want
to
get
some
like
look
good
to
me,
then
I
think
they're
really
pretty
quick
right.
I
think
you
you
had
that
review
last
time.
Did
you.
A
Well,
we
did
for
the
for
the
cap
itself,
but
then
tim
had
asked
that
we
split
it
so
that
we
could
separately
track
the
backwards,
incompatible
change
and
that
I
don't
know
if
so.
This
is
new
content.
Let
me
scroll
down
there's
this
new
cap
called
reject
invalid
pvc
data
source,
which
basically
is
just
a
copy
paste
of
some
of
the
components
of
the
well.
A
A
lot
of
this
is
template
stuff,
but
it
links
to
the
volume
populators
cap
and
then
explains
that
we
were
going
to
have
to
break
this
existing
behavior
and
then
very
briefly,
cover
the
goals,
the
non-goals,
the
proposal,
the
risks
and
mitigations
the
test
plan,
and
and
oh.
B
Also,
this
one
actually
does
not
have
the
that
question
narrow.
Is
that
question
in
the
original
cap.
A
When
I
copy
pasted
the
kep,
I
did
not
copy
the
the
pr
section
of
it,
but
maybe
maybe
that's
what
I
need
to
do
for
david
and
then
get
david
to
review
it.
I
wanted
to
give
buy
in
from
from
you
guys
on
on
this
particular
part
of
it.
The
you
know,
I
don't.
A
It
the
let's,
discuss
it
right
now,
just
to
make
sure
that
we're
on
the
same
page
and
then,
if
you
want
to
take
a
more
detailed
look
at
it
after
the
meeting.
That's
fine.
D
Yeah,
I
I
guess
the
context
here
was:
we
wanted
to
make
sure
basically
how
to
handle
the
the
weird
rejection
acceptance
of
data
sources,
behavior
that
we
had
yeah.
A
C
A
Yeah
zero
line,
it
just
deletes
three
files
and
removes
five
lines
from
from
registry.
Core
persistent
volume
claims
strategy
dot
go.
It
basically
removes
these
drop
disabled
field
calls,
but
that's
all
this
kept.
That's
all
this
change
does,
and-
and
this
is
this
is
what
is
called
for
in
the
cap-
is
to
do
this,
so
so
we
need
to
just
get
formal
agreement.
A
B
A
I
don't
know
I
mean
I'll
I'll
go
through
the
questionnaire
question
by
question
and
I
guess
and
put
in
some
answers,
and
most
of
them
will
just
be
like
not
applicable,
not
applicable,
but
for
the
ones
where
they're
I'll
either
refer
back
to
the
other
cap
or
answer
it
in
this
one.
I
can
update
it:
okay
for
the
sake
of
the
prr
people,
to
make
them
happy
because,
because
yeah,
I
do
want
david
to
look
at
this
and
say
yes,
but
but
I
guess.
D
A
And
then
yeah,
and
then,
if
tim
approves
this,
then
I
think
the
structure
of
this
pr
is
that
it
creates
a
new
cap.
I
don't
know
if
I
guess
it
doesn't
delete
anything
out
of
the
old
cap.
No
it
just.
It
just
creates
a
separate
cap
for
the
for
reject
invalid
pvc
data
source
and
leaves
the
the
the
original
one
alone,
because
we
were.
B
A
The
kept
number
is
2546,
so
there's
probably
an
issue
2546
out
there
that
refers
to
this.
In
fact,
we
could
go
find
it
your
issues,
don't
ignore
user
input
on
this
one
yeah
2546,
so.
A
Yeah,
so
so,
if
we
can
get
this
one
merged,
then,
and
then
I
just
have
to
get
all
the
the
coding
done
for
the
other,
repos
and-
and
so
the
status
update
on
on
the
other
repos
is
that
I
haven't
really
made
any
progress
on
the
release
tools.
A
Changes
I
mean
I,
I
have
a
pr
that
updates
release
tools
and
that
got
merged.
I
just
need
to
do
the
release
and
and
if
there's
any
more
code,
review
comments
I'll
respond
to
those
and
and
update
them
and
of
course
we
still
need.
A
I
still
need
to
add
tests
to
one
of
the
repos,
but
you
know
that'll
be
an
incremental
thing,
we'll
get
some
like
one
basic
integration
test
and
then
get
get
a
proud
job
based
on
that
test
and
then
we'll
expand
from
there
for
the
for
the
controller
that
that
validates
the
data
source
is
the.
What
do
we
call
the
volume
volume
data
source,
validator
or
something?
A
But
but
yeah,
so
I
guess
the
immediate
thing
is,
if
you
guys
could
put
lgtm
on
on
this
other
cup.
A
Okay,
I'm
trying
to
think
if
there's
anything
else
we
need
to
discuss
today.
A
As
far
as
as
far
as
I
can
tell
that's
the
only
other
thing
we
could
do
with
the
data
sources.
If
people
decide
they
don't
want
to
do
this.
Fortunately,
the
the
volume
populator
stuff
could
could
move
ahead
without
this,
but
we
wanted
are.
A
Well
so
the
problem
is,
we
have
a
feature
gate
for
the
for
the
popular
for
the
any
volume
data
source
and
once
that
goes
to
beta,
then
we
will
already
be
changing
this
this
this
logic,
so
we
won't
be
dropping
the
disabled
fields,
but
what
would
happen
is?
Is
we.
A
What
would
happen
you
would
just
get
a
pvc
with
like
a
pod
in
it
and
it
would
just
never
bind
because
there
would
never
be
a
populator
for
any
core
object
type.
A
So
yeah,
I'm
gonna
have
to
go,
go
back
and
find
out.
What
is
the
thing
that
throws
the
air
if
you
have
actual
garbage
in
the?
I
think
that
there's
already
code
merged
that,
if,
if
you
have
a
core
object,
you
get
an
error.
A
If
the
the
thing
is
is
because
of
the
the
drop
disabled
fields
code
that
currently
exists.
It
routes
around
the
error
by
dropping
the
contents
of
the
field
before
it
gets
to
the
error,
checking
got
it
so
so
the
problem
is
is
when
we
the
feature
gate
flips.
A
A
Yeah,
I
think
that's
how
tim
regarded
it.
He
just
he
wanted
it
to
have
a
cap
wrapped
around
it
so
that
it
would
raise
the
visibility,
get
a
top-level
release
note
and-
and
I
guess
we
could
have
an
opportunity
to
if
if
we
approve
the
cap
and
we
merge
the
code
and
someone
screams
in
the
middle
of
the
122
release,
we
could
back
it
out
before
we
cut
122.
A
A
D
A
All
right,
so
I
I
I
don't
know
if
tim
is
going
to
want
to
schedule
a
meeting
when
I
get
to
the
front
of
his
queue
or,
if
he's
just
going
to
review
it
and
say
yes
or
how
that's
going
to
go.
But
if
I
do
hear
back
from
tim
and
he
wants
to
meet
I'll,
try
to
invite
the
two
of
you
and
the
other
regular
attendees
for
this
meeting
and
we'll
try
to
get
something
figured
out.
A
A
Yeah
and
I'm
sure
he's
very
busy,
and
I
I'm
sure
that
kubecon
and
the
short
short
schedule
is
not
making
it
easy,
because
I
know
that
there's
a
long
list
of
other
storage
caps
that
are
probably
trying
to
get
get
approvals
too.
I
know
that
the
the
cozy
one,
because
I'm
in
all
those
meetings,
they're
they're,
trying
very
hard
to
get
cozy
approved
for
122
to
finally
be
implementable.
D
Yeah
fingers
crossed
I'm
surprised
how
quickly
the
enhancement
deadline
is
this
cycle.
A
The
only
the
only
logic
that
I
can
imagine
is
that
they
took
a
week
or
two
to
decide
whether
they
were
going
to
have
a
three
or
four
month.
Release
cycle.
A
And
so
the
announcement
went
out
like
a
week
or
two
later
than
ordinarily
would
have
for
what
the
schedule
was
going
to
be
so
like
that
was
one
or
two
weeks
of
dead
time
where
people
could
have
been
working
on
caps,
but
but
they
hadn't
announced
the
schedule
yet.
So
I
think
that's
where
the
time
went.
But
but
I
I
don't
know
the
logic
of
the
release.
People.
A
All
right,
well,
that's
that's
all
I
have,
if
you
guys
would
please
yeah,
take
a
look
at
the
at
the
cap.
That's
the
the
main
thing.