►
Description
Sept 1, 2020
Status Updates
carlisia
pto last Friday
addressing code reviews
going to review/test PRs for 1.5
nrb
Planning for Q4 w/ VMware
Some planning for v1.6
reviewing PRs
v1.5.0 delayed (in discussion topics)
Welcoming Dave!
Community support this week
On vacation next week, will transfer responsibilities over to folks prior
bridget
Finally transitioned over to the Velero project
Getting up to speed
Reviewing PRs for 1.5
Discussion Topics
nrb: v1.5.0 has been delayed. We’ve got a list of release blocker PRs that we’d like to get through, and will not add more code changes unless there’s a critical bug/security fix. The PRs are in relative priority.
Timeframe - ~2 weeks
Testing will be more automated starting in v1.6, but we don’t have a defined e2e test suite right now, just whatever is in make ci
nrb: VMware will be having a company holiday on Friday, September 4 and Monday, September 7 will be US Labor Day.
carlisia: do we want to halt the Velero server in this case: “Invalid value for --disable-controllers flag provided” (https://github.com/vmware-tanzu/velero/blob/976484553062262324cc1304607ea00841262db8/pkg/cmd/server/server.go#L822)
Action Item: Carlisia to open an issue to discuss this and document the issue
phuong:
quick update on the backup cluster resources
Able to back up clustorerole and pv in the context of a namespace, don’t need to modify the selection logic
backing up PVC in parallel
Taking backups of pods in logical group in a serial manner could lead to drift
This is parallelization w/in a backup job, not parallelization of all jobs
Currently this would be VolumeSnapshotter plugins - may not be a resource contention issue
This is the main issue here - use case is for Cassandra
How would this work with the CSI plugin? Since it’s BackupItemAction, the interface is way different
At the Kubernetes API server level, would we want to specify parallelism per resource type?
CRDs may need this, too
Need to be careful to not overflow resource limits
Don’t want to overwhelm writing to the backup tarball
Once we get the body, that could be funneled to the tarball serially
Action Item: Phuong to make an issue to continue discussion
sseago: posted design doc for RestoreItemAction “wait for additional item to be ready” (https://github.com/vmware-tanzu/velero/pull/2867)
To add timeouts, we could use https://velero.io/docs/v1.4/custom-plugins/#plugin-configuration
brito-rafa: will work with jenting on supporting multiple API Groups (https://github.com/vmware-tanzu/velero/issues/2551)
Contributor Shoutouts
@a-mccarthy: point contributors to style guide
@a-mccarthy: add hugo default TOC
@a-mccarthy: add new table shortcode
@a-mccarthy: update docs to match style guide
@a-mccarthy: Update release checklist to include more info around blog posts and r…
@c3y1huang: [velero] Fix helm 3.3.0 lint issue
A
Hello,
everyone
and
welcome
to
yet
another
velour
community
meeting
and
open
discussion
today
is
september
1st
2020,
the
1st
of
september.
We
are
heading
into
fall
here.
I
hope
you
all
had
a
lovely
summer
and
are
getting
ready
for
a
little
chillier
weather
if
you
live
somewhere
where
it
gets
chilly.
B
Hello,
so
how
do
pto
last
friday-
and
I
have
two
pr's
one-
is
a
refactor
of
the
backup,
storage
location
controller
and
I
addressed
well.
Actually
I
wanted
to
refactor
the
refactor,
and
that
is
done
is
ready
for
a
last
for
another
review,
which
hopefully
will
be
the
last
one
that
takes
in
pretty
good
shape
now
and
I'm
addressing
the
code
reviews
on
my
status.
B
D
Hey
so
this
past
week
I
was
doing
a
lot
of
planning
exercises.
I
know
I've
been
saying
that
for
a
while
there
was
some
planning
exercises
within
vmware
that
I
was
taking
part
in,
and
there
were
some
discussions
that
stephanie
and
I
had
for
the
1.6
time
frame.
There's
also
some
of
that's
going
to
happen
later
today,
and
I
think
some
of
that
will
be
shared
at
the
commune
meeting.
Next
week
I've
been
reviewing
some
pr's.
D
A
lot
of
the
pr's
that
have
been
coming
through
have
been
documentation.
Pr's
fixing
up
some
problems
that
we
missed
with
our
transition
from
jekyll
to
hugo.
So
thank
you.
A
ton
abby
for
fixing
those
up.
That's
been
very
helpful.
As
folks
may
have
noticed.
We
did
say
that
the
one
we
talked
a
little
bit
about
v
1.5.0
being
released
yesterday,
and
that
did
not
happen
I'll
we'll
go
into
that
a
little
bit
more
in
the
discussion
topics.
D
D
E
Yeah
such
fun
yeah,
so
hey
thanks
for
the
introduction
and
yeah
so
I'll
be
joining
the
team.
We've
been
I've
been
working
with
kubernetes
from
the
vsphere
side
for
a
couple
three
years
now,
I've
been
doing
the
plug-in
for
the
valero
plug-in
for
vsphere
was
one
of
the
projects
we
started
over
there
in
the
storage
area
and
also
previously
did
the
we
called
first
class
disc,
which
is
the
disc
infrastructure
for
valero.
So
I'm
looking
forward
to
or
not
further
for
kubernetes.
E
D
F
Hi
so
so
nolan
introduced
me
a
couple
of
weeks
back,
but
I
think,
as
I
mentioned,
then
I
was
still
stuck
across
two
different
projects,
because
I've
been
working
on
solid
boy
up
until
this
point.
During
last
week
I
finally
transitioned
over
full
fully
to
blair,
so
I've
done
the
handover
of
sauna
boy,
so
I've
just
been
spending
the
last
week
kind
of
getting
up
to
speed.
D
D
D
Also
at
the
bottom.
The
number
2804
I'd
like
to
get
the
restore
hooks
work
from
andrew
reed
in
get
that
included
and
then
some
documentation
from
ashish
on
restore
hooks
and
then
the
top
one
is
from
me
for
just
kind
of
our
release
process.
Automation,
standpoint
for
the
beta.
We
did
like
a
little
ride,
along
with
most
of
the
team
to
see
how
the
beta
release
process
worked
and
we
ran
into
some
bugs
with
that
tag
really
script.
D
So
we
decided
to
take
a
little
step
back
and
say
this
isn't
quite
ready.
Let's
give
it
a
little
bit
more
time
to
bake,
let's
get
some
of
these
refactorings
in
and
test
those,
and
let's
also
get
these
restore
hooks
in,
so
these
are
kind
of
the
critical
ones
that
we'd
like
to
get
in
and
then
look
at,
maybe
cutting
an
rc
and
then
from
there.
If
we've
got
critical
bug
fixes
that
or
security
fixes
that
need
to
go
in.
We
can
look
at
that,
but
these
these
are
right.
D
My
other
hope
is
that
the
release
doesn't
block
on
me
taking
vacation,
but
given
given
the
testing,
I
don't
I'd,
also
don't
want
to
put
the
rc
out
and
then
immediately
gotta
release.
Right
dave
is
helping
us
out
with
getting
automation,
set
up
automated
testing.
This
was
a
big
theme
for
our
one,
six
planning
that
we
want
to
get
more
automated
testing
and
more
end-to-end
testing
setup.
D
This
may
not
start
in
the
open
just
in
in
interest
transparency.
This
may
not
start
as
an
open
source
thing.
The
results
may
be
fed
into
prs
and
run
as
nightly
jobs,
but
they're.
They
may
not
be
on
open
source
infrastructure
yet,
but
that's
something
we're
looking
at
in
the
1.6
time
frame,
but
we're
going
to
start
with
what
we
excuse
me
start
with
what
we
have
so
my
hope
is
not
much
more
than
a
couple
weeks,
but
I
don't
have
a
definitive
timeline
there.
D
G
D
Yeah
and
that's
fair
again
like
I
said,
I
think
I
think
the
the
timeline
for
the
beta
to
the
the
previous
timeline
was
pretty
short
and
given,
given
what
valero
does.
I
think
that
was
pretty
aggressive,
so
we're
going
to
broaden
that
a
little
bit
and
get
a
chance
to
get
these
these
refactorings
in
and
then
test
those.
H
Refactorings,
I
have
a
question
in
regards
to
testing
for
prs.
I
have
the
restore
hooks
pr
out.
I
was
wondering
how
much
end
to
end
testing
should
I
do
on
that
before
merging
and
how
much
should
I
do
after
merging,
because
I
just
I
compiled
from
branch
and
you
know
make
sure
it
works
one,
I'm
not
sure
if
there's
a
full
qa
suite,
I
need
to
be
doing
against
that.
D
That's
part
of
what
we're
going
to
be
working
on
in
one
six
is
getting
a
full
q,
fuller
q,
or
that
wasn't
great
that
wasn't
great
wording.
We're
going
to
start
building
out
our
end-to-end
test
suite
in
one
six.
So
there
is
not
a
great
suite:
that's
there
right
now
we
have
our
unit
tests
that
run
in
the
the
make
ci
target.
D
So
if
you've
included
new
unit
tests
or
integration
tests
in
your
go
definitions,
those
will
run
but
they're
they're
not
running
against
a
live
cluster
and
and
a
lot
of
the
testing
that's
been
done
through
valero's
lifetime.
So
far
has
been
manual.
So
with
one
six
we're
going
to
try
to
take
that
and
systematize
it
more,
maybe
more.
Automated
the
v
sphere,
plug-in
team
has
jenkins
installs
internally,
that
that
do
that
they
spin
up
the
cluster
they
run
through.
E
I
D
D
So
if,
if
it's
passing
make
ci
right
now,
that's
what
we've
got
and
then
manual
testing
is
kind
of.
Historically
it's
kind
of
been
we
stop
the
development
work
and
test
manually.
So
that's
where
we're
at
going
forward
we're
going
to
try
to
fix
that
and
so
from
an
engineering
point
of
view
that
that's
been
our
big
ask
for
getting
time
to
do
that
in
the
1.6
q3
q4
time
frame.
A
E
B
B
D
D
D
B
D
I
don't
know,
I
don't
know
if
anybody
on
the
call
is
really
using
this
think.
The
original
use
case
was
there
was
someone
that
was
using
it
to
basically
start
up
an
instance
of
valero
per
controller
using
the
disable
controllers
flag,
and
that
also
gets.
I
think,
the
old
restore
restore
only
flag,
behaviors
mimicked
by
disabling
the
appropriate
controllers.
A
G
Then
she's
still
able
to
back
up
the
cluster
resource
that
being
used
by
the
specific
namespace
that
we
backed
up.
So
because
of
that
including
the
pvc,
the
cluster
and
the
role.
I
forgot
exactly
the
type
of
pv
resource,
the
tire
plus
the
resort
that
ce
is
interested
in,
but
these
are
all
backup
within
the
context
of
the
namespace.
So
we
don't
need
the
enhancement
anymore,
at
least
for
what
we
are
trying
to
do.
So
that
is
the
update
on
that
item
and
we
also
have
another
questions
for
the
community
as
a
whole.
G
So
the
idea
was
that,
right
now
we
back
up
everything.
You
know
one
by
one
in
serials
like
one
backup,
for
example,
I,
if
you
run
a
part,
we
back
up
a
pot
finish
everything
under
the
path,
the
pv
pvc
at
the
bottom,
and
then
it
moved
to
the
next
part
and
so
on
so
forth.
So
there's
there
are
six
situations
when
we
want
to
back
up
these
all
of
the,
for
example,
if
we
using
a
hook
to
executing
a
snapshot
for
that
specific
pvc
to
backup
it
right.
G
If
you
want
to
back
up
a
bunch
of
them
that
relate
a
bunch
of
related
port,
for
example,
all
the
parts
that
belong
to
an
application,
a
database
application,
we
want
to
snapshot
them
as
close
as
possible
to
each
other.
If
we
do
it
in
serial,
for
example,
we
have
20
parts
right,
you
know,
then
these
will
be
like
you
know
the
first
part
and
the
last
part
may
be
too
long.
You
know
too
long
to
distance
apart.
So
can.
E
G
G
That
is
one
of
the
the
the
example
of
parallel
scientists
to
improve
the
performance.
If
we
involving
you,
know
such
a
thing,
so
that
is
just
what
I
I
have
a
question
here,
so
I
know
that
right
now,
we'll
be
doing
you
know,
cereals
and-
and
I
know
that,
but
I
I
I
wonder
if
we
you
know
any
plan
to
you,
know
parallelize
them
and
so
on
and
so
forth.
E
Yeah,
so
I
think
at
the
very
low
level
you
know
at
the
snapshot
level,
there's
really
not
a
resource
contention
problem,
but
we
do
have
to
be
careful
about
overloading
the
control
plane.
So
it's
going
to
depend
on
which
cloud
platform
we're
running
on
how
many
snapshots
we
can
issue
in
parallel,
for
example,.
D
So
what
I'm
wondering
if
this
this
would
be
something
that
we
would
need
to
design
such
that
you
could
say
similar
to
our
plug-ins
currently,
but
such
that
you
could
say:
okay,
four
pods
and
four
pvcs
do
parallelism
for
cluster
role
binding!
Do
parallelism
for
this
crd,
which
we
don't.
Maybe
we
don't
understand
it
yet,
because
people
can
define
their
own
arbitrary
stuff
do
parallelism.
G
D
D
At
a
high
level,
I'm
I'm
not
seeing
a
reason
we
wouldn't
want
this.
I
think
it's
a
matter
of
implementation,
details
and
design
because,
like
this
is
so
on
on
the
the
one,
I
think
one
of
the
big
things
we
want
to
look
at
in
one
sixth
time
frame,
at
least
from
a
design
perspective,
is
concurrency
at
the
back
up
and
restore
job
level,
so
being
able
to
run
multiple
backup
jobs
and
multiple
restore
jobs
at
a
time.
D
But
then
this
is
one
level
down
which
is
within
the
storage
backup
or
I'm
going
to
put
restore
to
the
side
right
now
within
the
backup
job.
I
want
to
pull
stuff
from
the
api
server
in
parallel
and
I
want
to
be
able
to
make
volume
snapshots
in
parallel.
If
my
storage
system
will
let
me
so
this
is,
this
is
basically
trying
to
get
to
levels
of
concurrency.
H
E
E
And
then,
in
terms
of
like
resources
and
stuff,
so
the
pods
themselves,
we
don't
actually
snapshot
anything
except
the
pod
definition
right.
So
that
should
not
be
so.
A
real
concern
here
is
with
getting
the
storage
as
close
as
possible
to
everything,
or
is
there
actually
something
going
on
with
the
pods
themselves.
G
It's
meant,
the
main
thing
is
the
the
data
like
the
pvc
being
used
by
that
part.
Let's
just
say
I
have
a
cluster
of
20
part
belong
to
cassandra
right
and
these
data.
They
are,
you
know,
a
certain
data
on
each
of
them
and
we
want
to
snapshot
them
all
in,
like
almost
like
a
cluster-wise
snapshot
like
20
pvc
being
snapshot
approximately
in
you
know
very
short
period
of
time
for
each
other.
So
that
is
the
the
the
scenario
that
I
concern
right
now.
E
D
E
D
E
D
Yeah,
so
it
sounds
like
the
the
actual
resource
we
care
about
here
isn't
necessarily
pvc,
it's
pv,
it's
it's
the
it's,
not
the
the
claim.
It's
the
the
fulfilled
claim
that
that
points
to
the
actual
disk,
so
you
say
pods
I
mean
we.
We
can
grab
anything
from
the
kubernetes
server
in
parallel.
I
think
that
that
one
I'm
a
little
less
concerned
on
it's
it's
the
pv
and
it's
associated
volume
snapshot
are
here
exactly.
D
G
G
D
D
E
E
D
Yeah,
so
I
think
this
is
this
is
a
good
discussion.
I'm
gonna
request
that
we
get
a
another
issue
for
this,
because
that
will
help
us
track
it,
and
I
think
this
that
would
be
useful
for
stephanie
as
well,
because
I
know
she's
been
trying
tracking
this
kind
of
high
level
stuff
as
feature
requests
so
fong.
D
If
you
could,
if
you
could
file
an
issue-
and
we
could
take
that
and
and
work
off
of
that
and
coordinate
there
and
maybe
produ
at
some
point-
produce
a
design
document
for
that,
because
I
think
there's
there's
going
to
be
little
special
cases
like
this
csi
plug-in
that'll
trip
us
up.
So
I
I
in
general
makes
total
sense
to
me.
G
I
Yeah,
so
I
I
linked
the
the
issue
here.
Basically,
this
comes
out
of
a
rather
linked
pr
for
the
design
doug.
This
came
out
of
an
issue
we
I
I
entered
several
months
ago,
where
we
had
a
failure
case
in
our
environment,
where
we're
restoring
image
stream
tags,
which
is
an
openshift
crd
and
basically
one
of
the
ways
that
you
can
create
an
image
stream
tag
is
with
where
one
tag
references
another.
I
So
there's
a
reference
from
one
to
another,
the
same
resource
type
and
so
their
initial
plugin
had
a
addition.
We
passed
the
additional
items,
slice,
including
one
of
these-
the
item
needed
to
be
there
and
we
found
that
there's
a
race
condition,
and
sometimes
this
works.
Sometimes
it
doesn't
because
the
way
the
code
works
there
you
go
and
we
return
the
additional
items.
Valero
then
restores
those
items
runs,
restore
item
all
the
way
through
and
then
continues
on
to
restore
the
current
item.
I
But
you
know
because
of
some
delays
with
kubernetes,
sometimes
it
doesn't
show
up
as
available
yet
so
when
we
try
to
restore
that
item,
it's
not
there.
So
I
entered
an
issue
about
that.
We
had
some
discussion
from
the
linked
issue
here,
where
basically,
the
back
and
forth
was
that
you
guys
were
interested
in
solving
this.
It
wasn't
really
a
priority,
but
you
know
I
suggested
that
I
read
a
design
doc
for
how
we
might
implement
this,
and
it's
been
on
the
back
burner
for
a
while.
I
I
finally
got
back
around
to
doing
that
and
it's
basically
a
proposal
to
add
an
additional
method
to
the
restore
item,
action
plug
plugin
that
will
allow
valero
to
after
calling
the
restore
and
all
the
additional
items
it
and
then
running
with
it
with
a
timeout.
To
basically
say:
hey,
are
these
items
ready
and
the
timeout
part
would
be
similar
to
what
we
do
with
the
crds
and
a
couple.
Other
cases
where
we
want
to
you
know
try
this
method.
I
You
know
until
we
hit
a
timeout
and
then
once
the
timeout
is
reached,
you
know
we
either
we
may
or
may
not
fail,
but
that's
so
that's
just
kind
of
the
basic
idea
here
I
can
go
in
more
detail
now
or
I
can
just
let
you
guys
look
at
it
offline
and
comment
on
the
pr.
I
think
I've
put
the
design
together
reasonably
close
to
what
nolan
and
I
had
originally
discussed
in
terms
of
what
was
desirable,
and
you
know
anyway,
I'm
certainly
open
to
making
updates.
I
Well,
that's
one
thing
I
should
call
out:
there
are
a
couple
of
different
options.
I
propose
because
this
would
require
a
change
to
the
restore
item.
Action
interface
and
you
know
the
easiest
thing
to
do
would
just
be
to
add
a
new
th
this
new.
If
you
scroll
down,
you
see
what
it
proposed
there,
but
I
forget
the
name
of
it
now
because
other
there
right.
So
the
our
additional
item
is
ready.
You
just
add
that
function
to
the
interface
because
of
the
backward
compatibility
issues.
I
If
we
do
that,
then
everybody
that
has
a
restore
item
action
plug-in
once
they
update
this
is
gonna,
have
to
go
and
add
that
and
it'll
be
a
no-op
and
a
lot
in
most
cases,
certainly
and
our
plugins.
We
only
have
one
plugin
where
we
would
need
this
right
now.
The
alternative
I
proposed
was
to
have
an
optional
interface
that
we
could
implement
and
then
valero
itself
would
check
to
see
if
it
implements
that
interface
and
then
call
it
if
it's
there.
I
The
the
second
case
is
more
backwards
compatible,
but
it's
you
know
from
a
design
point
of
view.
It's
not
necessarily.
I
mean
some
people
don't
like
this.
They
say
it's
not
good,
for
you
know
laying
style
reasons,
because
you're
checking
kind
of
at
runtime
whether
it
influences
interface,
but
those
are
two
basic
approaches
that
would
accomplish
the
same
thing.
I
From
my
point
of
view,
I
could
do
either
one
I'm
guessing
that
the
core
team
has
would
probably
have
an
opinion
as
to
which
would
be
preferable,
and
so
that's
the
biggest
thing
I
need
feedback
on
is.
Basically
we
would
need
to
come
to
a
decision
as
to
which
of
those
two
approaches
we
want
to
take,
and
I
could
remove
it
from
the
dock
and
then,
of
course,
the
rest
of
it
is
just
all
along
the
way.
I
You
know,
suggestions
for
improvements,
things
that
I
might
have
missed
or
might
have
gotten
wrong
here
in
terms
of
what
you
know
based
on
your
goals,
another,
I
guess
you
should
call
it
another
issue
to
think
about
is
the
question
of
timeouts
and
what
what
whether
we
should
make
that
configurable.
You
know
how
to
handle
that
in
terms
of
how
long
we're
going
to
wait,
because
you
hit
a
certain
point
where
you,
you
know
we're
done
waiting
and
then
even
there
we
have
a
decision.
I
If
you,
if
you
hit
the
timeout
and
the
wait
still
hasn't,
come
and
says
hey,
this
thing
is
ready.
You
know
we
could
fail
at
this
point
and
say
hey
this
time
it
wasn't
reached.
This
is
the
required
thing.
Let's
fail
this
this
item
or
we
could
say,
look
we've
hit
the
timeout,
but
that
doesn't
guarantee
it's
not
here.
Now.
Let's
just
you
know
best
effort
was
was
was
put
out.
Let's
try
anyway
and
go
on.
If
the
item
itself
fails,
then
we
still
have
a
failure.
I
You
could
even
go
more
complicated
than
that
and
you
know
make
it
configurable
at
the
restore
level.
I
don't
know
that
that's
warranted,
but
you
know
again,
those
are
the
kinds
of
things
I'd
be
looking
forward
from
the
quarantine
feedback
to
see
where
we're
going
with
this.
My
hope
is
to,
and
obviously
we're
not
talking,
one
five
here,
we're
talking
some
future
point,
but
basically
one
once
we
are
in
agreement
on.
You
know
approval
on
this.
I
D
D
I
D
J
Hello,
everyone-
this
is
going
to
be
quick,
just
a
quick
item
to
validate
that
genting
is
offering
to
help
on
that
ticket.
That's
something
that
has
been
dormant
for
a
couple
months.
I'm
gonna
be
reach
out
him
to
work
together
on
that
one.
I
know
that
some
just
refresh
some
people
and
put
some
context
is
basically
complete.
The
work
to
make
valero
support
multiple
api
group
versions.
J
So
this
is
a
very
high
level
proposal.
So
again,
I'm
going
to
reach
out
him
and
see
we
can
work
together.
He
can
code
the
contestant,
something
like
this.
One
of
the
things
that
didn't
kind
of
like
hold
this
development
on
the
little
core.
A
little
bit
is
the
concern
that
exchange
gonna
work
with
the
plugins.
J
D
J
A
This
was
added
yesterday,
so
we
now
have
a
new
badge
here
in
the
readme,
which
is
the
cia
best
practices
saying
that
we
passed
the
best
practices
for
for
the
cii
badge.
What
does
that
mean?
Well,
if
you
click
this
you'll
get
to
a
link
talking
about
how
belero,
as
a
project
is
following
best
practice,
best
practices
based
on
best
practices
that
have
been
created
by
the
core
infrastructure
initiative.
A
E
A
A
You
can
take
a
look
at
the
website
here
and
see
some
of
the
projects
that
you
like
and
enjoy
and
see
if
they
have
gone
through
this
same
process
as
well,
but
this
is
essentially
showing
what
we
do
and
how
we,
how
we
work
as
a
an
open
source
project,
how
we
handle
reporting
of
issues,
how
we
handle
automated
testing,
how
we
essentially
run
the
project.
So
this
is
something
that's
fun
and
interesting.
So
we
now
have
this
new
badge
here
that
we
can
enjoy
in
the
readme
and
then
yeah.
D
D
We
has,
or
two
of
these
abby
has
been
doing,
a
lot
of
work
around
our
style
guide
to
one
get
us
a
style
guide
and
point
people
to
two
to
update
the
existing
dots
to
conform
to
it.
I'm
sure,
there's
stuff
that
could
probably
still
use
work
but
even
getting
started
here,
has
been
a
huge
improvement.
So
thank
you.
There
there's
also
been
some
broken
links
throughout
the
site.
D
D
That
has
been
very
helpful,
and
not
only
is
she
fixing
them
she's
fixing
them
in
a
way
that
is
improving
or
maintaining
accessibility,
so
that
is
very
much
appreciated
and
then
finally,
she
has
been
working
with
me
to
make
sure
our
release
checklist
is
documented
in
public,
so
this
release
checklist
was
something
that
was
on
a
confluence
page
inside
of
vmware,
and
now
more
of
it
is
on.
I
think.
Actually
I
think
this
is
all
of
it
is
on
our
release.
Instructions
page.
D
D
Yeah
and
then
I
had
a
window
open
with
this
person's
name,
and
this
next
one
is
chin,
yeah
hwang
from
susa.
It
looks
like
fixing
some
linting
issues
on
helm,
3.3
on
our
helm
chart.
So
thank
you
very
much
for
finding
these
and
fixing
them
up
making
sure
they
pass
the
linting
tool
and
keep
clean.
So
thank
you
very
much.