►
From YouTube: Kubernetes SIG Storage - Bi-Weekly Meeting 20210617
Description
Kubernetes Storage Special-Interest-Group (SIG) Bi-Weekly Meeting - 17 June 2021
Meeting Notes/Agenda: -
Find out more about the Storage SIG here: https://github.com/kubernetes/community/tree/master/sig-storage
Moderator: Xing Yang (VMware)
A
Hello,
everyone
today
is
june
17th
2021.
This
is
the
kubernetes
storage
week
meeting.
So
today
we
will
first
go
over
1.2
planning
spreadsheet.
So
the
next
deadline
is
code
freeze,
which
is
a
july
8th,
so
it's
actually
coming
and
then
I
think,
there's
one
topic
at
the
end.
It's
about
smb
csi
driver
all
right.
So.
A
B
Yeah,
so
the
the
css
pictures
got
merged
and-
and
we
have
yeah
this
work
in
prague-
progress
like
chang,
I
think,
has-
is
doing
the
kubelet
changes
required
in
kubelet
and
I
have
like.
I
have
split
the
items.
There's
two
cards.
Basically
one
in
this
azure
file,
csr
driver
and
one
is
cubelet
so
yeah
we
should
get
it
done.
B
Update
yeah,
it's
gonna
remain
in
beta,
there's
no
update
to
this
itself
because
it
was
blocked
on
the
next
item.
B
Yeah
this
one
is,
I
started
working
on
it.
I
wrote
I
made
the
api
changes.
I
have
some.
I
have
like
quite
a
bit
of
it
written
down.
Actually,
I
was
not
sure
how
to
split
this
into
this
into
other
cards,
because
it's
kind
of
dependent
on
each
other,
like
until
you
get
the
api
change.
You
can't
do
the
quota
thing
or
you
cannot
make
the
changes
into
the
controller
so
yeah
I'll
see
if
I
can
split
it
and
maybe
like
but
yeah
until
the
api
changes
are
there
like.
B
A
C
Yeah,
if
humble
is
online,
I
gave
him
ping
and
he
said
he
would
we'll
be
taking
a
look
at
it
again.
I
think
he
will
get
back
to
me
after
the
fs
group
change.
If
this
is
still
outstanding,
I
can
take
a
look.
A
A
Thank
you,
one
is
the
issues
related
to
assuming
volumes
amount
points.
This
is
a
bug
fix.
Do
we
have
a
mukahid
or
gene
here.
D
D
I
believe
this
race,
and
but
his
cpr,
like
is
big,
needs
some
time
to
to
be
merged
and
I'll
follow
up
on
that,
and
then,
after
that,
we
might
see
the
previous
com
reverted
pr
again
to
see
whether
that
can
be
yeah
merge
again
to
solve
this
issue.
D
A
E
I
think
for
these
items
I
think,
there's
just
continuing
you
know
more
bug
fixing
this
cycle,
but
that's
about
it.
Okay,.
A
F
E
It's
it's
a
bug
when
it's
a
bug,
if
you're
using
the
what
is
it
the
volume
attached
limit
the
volume
node
limit,
it's
not
accounting
for
it.
C
E
A
Minor
thanks
michelle
okay,
next
one
starting
over
filler
domain,
so
this
one's
still
depending
on
the
next
one,
which
is
one
group.
I
have
not
got
a
chance
to
update
this
one
yet
so
I
need
to
do
this
one
soon
and
next
one
is
a
csi
auto
retreat,
move,
iscsi
driver,
don't
finish.
G
A
Thank
you
and
the
next
one
is
mo
samba
sif's
scissor
driver
to
ga.
So
if
we
oh,
we
don't
have
any
water.
This
looks
like
andy
and
julie.
It's
not
somewhat
here.
A
Does
anyone
know
michelle
or
is
andy
and
julie
here?
Maybe.
H
A
A
Do
you
have
any
any
update
on
this
item?
Are
you
working
on
this
with
andy.
H
Oh
yeah
yeah:
this
is
the
item
I
entered
in
the
mist.
It's
like
provide
a
update
on
the
smb
csi
driver,
ga
plan,
so
the
rough
update
is
wait.
The
linux
part
is
quite
stable.
Now,
so
andy
have
already
cut
a
release
in
the
market.
Ga
and
for
the
windows
part,
the
things
two
main
things
are
left
one
is
we
needed
to
testing
and
the
switch
to
the
csi
proxy
ga
version,
and
another
thing
is:
there:
is
a
windows
node
register
issue.
H
H
Separately,
linux
pad
is
already
g8
yeah.
Oh.
A
H
Nd-
and
you
should
nv
oh
but
yeah
andy-
I
I
will
also
help
him.
You
know
bring
the
smb
csi
driver
to
ga.
A
Okay,
next
one
is
the
center
deprecation
notice
for
flux
volumes.
We
talked
about
this
last
time,
so
I
I
drafted
a
message
michelle.
If
you
can
take
a
look.
I
think
I
also
I
shared
with
you
michelle
and
also
matt
sounds
good.
Thank
you.
A
No
update,
okay,
I
believe
he
was
saying
last
time
after
last
meeting,
he
was
saying
that
he
was
not
clear
on
what
to
do
next,
so
so
christian,
I
love
christian
yeah.
G
A
F
F
We've
been
going
back
over
some
of
the
original
design
decisions.
The
community
has
changed
a
little
bit
and
we're
basically
trying
to
make
sure
that
all
of
the
design
decisions
that
were
made
and
cozy
because
because
this
design
stretches
back
like
a
year,
you
know
based
on
the
feedback
from
the
api
team,
that
it
all
still
makes
sense
and
so
we're
sort
of
re-litigating.
Some
of
the
you
know
old
arguments,
but
we're
coming
to
the
same
conclusions
more
or
less.
It's
just
a
slow
process.
A
Thank
you
yeah,
and
there
is
a
meeting
design
meeting
on
cozy
right
after
this
meeting.
So
if
you
are
interested
join
that
meeting
next
one
is
the
change
block
tracking.
I
don't
really
find
this
on
on
call.
I
don't
think
we
have
got
any
update.
I
think
he's
still
working
on
the
documentation.
Part
of
the
ct.
A
The
next
one
is
a
new
repair
on
the
access
mode.
Chris,
do
you
have
a
day.
E
Know
the
status
of
this
yeah.
He
has
a
couple
pr's
in
flight.
I
I
think
he's
making
good
progress
on.
F
E
D
Yes,
so
mauricio
and
iron
group
had
a
meeting
yesterday
and
we
talked
about
it
already
and
you
did
all
the
work
had
a
few
new
prs
merged,
there's
one
last
pending
which,
where
should
you
want
to
yeah
talk
about
it?
Since
I
think
merkel
is
here
too.
I
Sure
so
I
think
that
we're
done
with
all
of
the
changes
that
we
had
because
of
the
api
review
and
those
changes
will
be
part
of
a
new
beta
release,
workload
b1
beta
3
and,
at
the
same
time,
I
added
a
few
integration
tests
to
make
sure
that
we're
not
breaking
backwards,
compatibility
or
work
or
maintaining
backwards
compatibility
with
the
scenario
and
so
far
we
haven't
seen
any
problems.
I
have
already
integrated
it
with
a
gcpd
csi
and
it
was
working
fine
with
the
migration
tests.
I
So
we
decided
to
also
bump
the
b1
beta,
3
versions
to
b1
and
as
part
of
the
rc,
we're
going
to
release
both
apis
and
that
that
will
help
us
test
b1
directly
in
pdcsi
and
eventually
we'll
cut
the
through
ga
probably
later
next
week.
I'm
not
sure.
I
think
that
things
are
looking
good
for
for
not
csi
approximately.
D
So
yeah,
if
no
more,
like
other
issues
about
very
likely
like
at
the
end
of
next
week,
we
will
give
some
time
for
driver
to
test
it.
Yeah,
okay,.
E
Looks
like
they're,
not
here
yeah,
I
don't
think,
there's
much
update
here.
Okay,
I
think
there's
a
couple
of
bug
fixes
that
are
being
tracked.
I
think,
but
but
nothing
super
major.
A
Okay,
thank
you,
and
next
one
is
deprecate.
Entry
drivers.
E
A
A
Next
one
is
csm
operation
vsphere.
I
don't
have
additional
update
on
that
other
than
the
the
pr.
I
think
it's
still
still
out.
It's
still
being.
A
Update
and
the
next
one
is
the
gc
csma
brand
gce.
B
A
Okay,
I
think
it's
yeah,
it's
a
file.
Okay,
thank
you
so
sees
my
green
gce.
Do
we
have
any
update
on
this.
E
Yeah,
I
think,
basically
they're
still
trying
to
debug
performance
issues.
I
think
right
now
it's
pointing
to
some
sort
of
problem
either
in
the
api
server
or
the
client
go,
but
it
seems
like
yeah
there's
some
performance
regression
between
120
and
121.
A
E
Yeah,
as
in
like
the
the
informers,
the
informers
on
the
side
cars,
they
seem
to
be
a
lot
slower
for
some
reason:
they're
still
trying
to
debug.
What's
going
on
there.
E
E
There
was
also,
I
believe,
a
bug
fix
that
was
outstanding
regarding
the
way
that
they
do
topology,
I'm
not
sure
about
if
that
has
merged
yet
jan,
do
you
know.
E
This
is
the
the
ebs
migration
topology
bug
where
they
were
putting
both
topology
labels
on
the
pv.
J
A
Okay,
so
that's
that
one
and
then
it's
this
one.
Okay,
this
is
the
set
fs
and
seth
rpd
says
migration
is
okay.
I
think.
E
Humble
sent
out
a
email
to
kubernetes
dev,
I
think,
seeing
if
I'm
trying
to
gauge
community
use
of
the
entry
stuff
driver,
I
think
yeah.
I
think
we'll
have
to
give
give
it
some
time
for
people
to
respond
and
and
see
what
we
get.
E
A
Thank
you,
okay,
and
I
think
next
one
we
already
talked
about
this
is
the
this.
Is
the
topology
label
from
beta2ga
deviant
has
a
pr
that
is
being
reviewed.
A
And
control
body
mode
conversion
between
source
and
target
pvc
yeah,
so
ron,
I've
just
started
to
take
a
look
and
see
if
he
can
come
up
something
on
this.
A
Okay,
I
think
it's
probably
still
the
same
as
before
the
cap
is
not
merged
goes
through
some
concerns.
A
And
next
one
is
the
one
that
we
co-own
with
sig
off
user
id
ownership,
incoming
maps
and
secrets.
A
So
I
is
this
okay
come
on.
What
is
the
status
of
this?
Is
this
the
one
that
got
got
merged?
This
is
a
different
one.
Maybe
not
this
one.
It
was
there's
one
that
that
they
got
a
extension.
Maybe
I
don't
know
if
this
one
is
a
different
one.
A
A
A
Okay,
thank
you.
Next
one
is
the
non-graceful,
no
shadow,
so
I
think
the
band's
cup
did
not
get
murdered.
James
have
some
concerns
and
he
also
have
some
comments
on
the
cap,
which
I
still
need
to
address.
So
this
is
some
additional
testing.
F
The
csi
community
meeting
next
wednesday,
where
hopefully,
we
can
hash
out
the
issues
with
james
that
james
had.
F
A
Okay,
yeah
so
other
than
that,
so
I
also
did
more
testing
on
1.21,
which
I
think
previously
only
did
that
on
21
the
problem
still
exists,
so
we
still
have
this.
Basically,
when
you
shut
down
one
node,
the
the
workload
is
not
going
to
move
to
another
node
that
that
that
is
up
so
with
the
fix.
A
A
Moving
on
to
next,
which
is
a
enable
username
space
in
kublet,
so
you
already
get
shifted.
How
much
have
you
been
following
up
with
this.
B
Sorry,
the
in
the
name
space,
one
right,
I'm
sorry.
A
B
A
E
Michelle
do
you
know
the
status
of
this
I've
seen
updates
on
the
pr,
so
I
I
think
he's
actively
working
on
this.
Okay.
A
So
the
next
one
is
the
world
expansion
for
safer
set
so
just
check
with
shalini.
She
said
that
she
will
try
to
update
this
next
week.
A
And
next
one
is
excuse
container
notifier
for
application
snapshot.
We
are
still
waiting
for
signal
to
review
and
hopefully
get
it.
A
Approved
next
one
is
yuto's
mount
split
to
a
new
ripple.
Do
we
have
a
stringy
here
today.
E
A
Thanks
michelle
and
the
next
one
is
prioritization
and
volume
capacity.
Oh,
do
we
actually
are
we
doing
anything
on
this?
E
You
there
was
still
some
leftover
changes
for
from
the
alpha.
I
just
I
think
I
just
saw
the
pr
got
rebased,
so
I
think
I
think
he's
working
on.
E
A
E
Stay
on
track.
Yes,
I
just
spoke
to
shihong
yesterday
and
he
said
he's
going
to
start
working
on
this.
Okay.
A
Sounds
good,
I
think
that's
all
we
have
on
this
spreadsheet
going
back
here
and
looks
like
we
don't
have
additional
items
because
I
think
julie.
We
already
talked
about
this
one
right.
The
last
item.
Oh.
H
A
Okay,
all
right
are
there
anything
else
everyone
want
to
discuss
here.