►
From YouTube: Velero Community Meeting - Dec 8, 2020
Description
Dec 8, 2020
Status Updates
carlisia
Community support last week
Addressed PR reviews for the Tilt Setup (https://github.com/vmware-tanzu/velero/pull/3119)
Mainly did PR testing and reviews
bridget
Spent quite a bit of time on PR reviews
Experimenting with Delve/Tilt debugging
Working on initial steps for supporting multiple credentials
Mostly experimental work based on design doc
dsmithuchida
Working on Distributed Data Generator on Azure
nrb
PR reviews
Working on board processes
Discussion Topics
Rafael @brito-rafa and Frankie @codegold79 - Restore from multiple backed up API Group Versions. A WIP PR #3133 has been submitted that follows the corresponding design doc.
nrb - Restore progress and waiting for additional items?
ballpark of release timeframe of version 1.6 and/or 1.5.3 ?
Community Shoutouts
Swanand Shende for helping folks out in our upstream Slack
A
B
Hello,
so
last
week
I
was
in
community
support.
I,
the
team
did
a
really
great
job
giving
feedback
when
they
took
set
up
pr.
It
made
it
a
ton
better.
So
I
spent
a
lot
of
time
addressing
those
reviews
and
also
doing
review
testing
some
pr's
and
reviewing
them,
and
I
don't
know
what
time
when
that's
pretty
much
what
I
did
last.
A
D
B
I
think
we
can
fix
it
on
it,
I'm
dying
to
start
using
it.
I
appreciate
it
actually
because
I'm
really
like
just
yesterday,
I
made
a
huge
mistake.
Like
we
were
talking
about
bridget,
I
was
testing
a
pr
and
I
I'm
pretty
sure
I
did
the
shenanigans
to
send
a
fresh
image
for
from
that
branch,
and
I
didn't-
and
I
thought
the
guy
had
a
bug,
but
he
didn't
you
know
what
I'm
talking
about
it's
just
just
you
know
having
to
make
the
the
image
the
and
pushing
it
to
the
cluster
to.
C
Test
yeah
I'd
like
to
see
this
getting
merged
really
soon.
I
did
take
a
look,
so
I
know
that
there
was
a
we
discussed
in
our
backlog
meeting
yesterday
about
adding
support
for
the
rustic
restore
helper.
I
I
looked
today
when
valero
will
try
to
use
a
different
tag.
It
takes
a
tag
based
on
the
build
info
version,
but
because
there's
no
build
info
version
set
on
it,
it
should
just
use
the
latest
tag.
C
B
So
I'm
really
looking
forward
to
that
too,
because
I
don't
have,
I
don't
know
how
I
have
never
invested
time
to
figure
out
the
setup
for
the
debugging
on
on
the
gold
land.
I
can
do
the
basic,
but
I
know
there's
a
bunch
of
things
you
need
to
set
up.
So
it's
going
to
be
a
very
sweet
setup
for
us
to
have
this
in
the
debugger.
B
C
Hi
so
yeah.
I
also
felt
like
I
spent
quite
a
bit
of
time
on
pr
reviews
last
week
and
as
clearly
as
you
mentioned,
I've
been
experimenting
with
getting
the
delve
debugger
hooked
up
into
the
tilt
setup,
which
was
surprisingly
easy
to
do,
and
it's
very
easy
to
get
it
set
up
if
you're,
using
goland
or
even
any
other
ide,
where
you
or
just
delve
itself.
So
I
think
you
can
use
it
with
visual
studio
code
as
well
or
the
dell
cli.
C
C
Once
carly
says
pr
is
merged
and
then,
aside
from
that,
I've
been
starting
work
on
the
support
for
multiple
credentials,
not
necessarily
making
changes
to
support
multiple
credentials,
but
just
some
of
the
scaffolding
work
that
needs
to
happen
initially
before
we
start
introducing
more
sets
of
credentials.
So
it's
kind
of
experimental
work
things
that
I
said
in
the
design,
doc
that
I
thought
I
could
do.
C
A
C
Aws
aws
based,
but
what
I'm
working
on
at
the
moment
is
trying
to
get
away
from
specifying
the
environment
variables
in
the
deployment
and
instead
setting
them
at
the
point
whenever
we
launch
a
plug-in
process
so
that
you
could
start
different
plug-in
processes
with
different
credential
environment
variables
set.
So
at
the
moment
it's
kind
of
plug-in
agnostic.
But
that's
that's
one.
I'm
experimenting
with
at
the
moment
cool.
E
A
D
D
So
that's
to
help
alleviate
that
issue.
The
one
exception
to
that
is
the
helm
chart.
The
helm
chart
is
not
being
rolled
into
the
core
valero
repo
for
issue
reporting,
because
it's
it's
still
different
enough
and
there
may
be
other
helm
charts
in
that
overall
home
chart
repo
that
aren't
relevant
to
valero.
So
we
don't
want
to
route
those
issues.
D
A
A
F
Is
frankie?
I
just
wanted
to
say
that
I
kind
of
jumped
the
gun
we
still
have.
Oh
yeah
jump
the
gun
with
this
pr
three
one
three
three.
It
follows
the
design
dock
at
three:
zero,
five
zero
and
it's
jumping
the
gun,
because
this
the
design
dock
has
not
yet
been
approved,
and
so
there
are
potentially
a
lot
more
changes.
That
could
happen.
So
that's
why
this
is
even
though
this
code
that
I
submitted
for
the
pull
request
is
complete
and
I
have
all
the
tasks
unit
tests.
F
So
that's
my
update
for
this
week.
Oh-
and
I
guess
I
should
talk
about
what
this
code
is
for,
it's
for
the
the
ability
to
restore
given
multiple
backup
versions
of
api
groups.
So
each
resource
group,
like
you,
know
the
horizontal
plot,
auto
scaler-
can
have
multiple
api
groups
like
v1
or
v2,
beta
1
or
v2
beta
2.,
and
what
raphael
did
was
he
backed
he
enabled
using
a
with
a
feature
flag.
F
G
Yeah,
the
this
is
raphael
me
and
frankie.
We
been
working
this
for
last
couple
weeks,
yeah
if
you
guys
want
to
look
at
the
design
dock,
maybe
the
even
the
code
helps
to
understand
the
design
dock
and
vice
versa,
just
to
give
an
idea
how
to
test
this,
I
had
to
build
six
different
controllers
with
the
same
crd,
but
different
versions
and
the
conversions
back
and
forth.
G
So
we
make
sure
this
code
is
supposed
to
do
what
is
intended
for
so
there's
a
lot
of
details
on
it,
but
we
wanted
to
put
something
for
the
community
to
review
before
the
break.
To
be
doing
you
know,
eggnog
and,
and
so
on.
You
can,
you
know,
check
the
code
and
but
anyways.
I
think
it's
a
lot
of
work
and
I
assume,
in
my
opinion,
needs
to
be
done
and
we've
done
and-
and
I
appreciate
anyone
to
jump
in
and
look
to
see
if
all
this
makes
sense.
F
D
I
haven't
had
a
chance
to
go
back
and
look.
I
know
I
talked
to
frankie
early
last
week
I
hadn't
had
a
chance
to
come
back
to
the
design
doc.
Excuse
me,
I
think
it's
in
really
good
shape.
Personally,
so
I'll
take
a
look
at
this
pr,
this
this
code
as
well,
and
hopefully
we
can
get
it
merged
before
the
break,
at
least
the
design
dock.
B
G
Yes,
that's
a
great
question.
Jonas,
can
you
click
at
that
link
in
the
middle
of
the?
The
page
has
examples
for
project
plural,
so
we
narrowed
down
four
cases
during
migration
and
we
came
up
with
this
crd
called
rock
band
and
they
have
different
versions
so
each
case
on
those
four
cases
we
have
a
source
cluster
and
a
target
cluster.
Each
one
has
a
combination
of
the
version
so
for
each
combination.
We
had
to
build
a
controller
with
the
conversion
logic.
G
So
from
all
those
four
cases
I
got,
I
can
tell
the
case
d.
That's
the
one,
the
trickiest,
because
today,
by
default
we
already
take
care
of
case
a
and
kcc
and
case
b
is
very
unlikely
to
happen.
Unless
someone
really
messed
up
on
your
life
cycle
of
your
crd,
so
case
d.
Is
that
what
we're
going
to
look?
Very?
G
G
G
So
the
conversion
logic
really
in
this
case
gonna,
look
common
versions
that
support
it
and
use
them,
because
today
we
only
use
preferred
so
case
d.
Today,
valero
doesn't
work
right
because
you
know
v1
is
incompatible
with
v2,
but
hey
there's
an
intersection
v2
beta
2
and
v2
beta1.
Why
we
can
not
use
those?
That's
exactly
what
our
code
is
doing,
picking
up
the
common
versions
and
even
giving
ability
to
the
user
to
cherry
pick,
which
version
you
want.
Does
it
make
sense.
G
G
G
B
F
Much
about
computer
yeah
and
to
complicate
matters
we
allowed
for
users
to
to
create
a
config
map
to
to
be
able
to
select
their
own
the
versions
they
prefer.
So
this
is
in
the
in
this
code
as
well
as
well
as
a
design
dock.
It's
described
from
the
design
dog,
the
ability
for
the
users
to
choose
the
version
they
want
to.
G
D
I
think
I
think
this
is
just
to
reiterate
kind
of
what
this
is
the
high
level,
what
this
is
for
it's
for
jumping
between
versions
of
kubernetes,
when
you're
not
going
one
to
one
so
so
you're
not
going
120
to
121
you're,
going
120
to
126,
or
something
like
that,
where
you're
going
between
versions
that
don't
necessarily
have
the
the
web
hooks
built
in
to
do
the
translations
for
you.
So
this
this
expands
valero's
scope
such
that
your
backups
can
can
last
a
little
bit
longer.
They
can
become
more
archival.
D
Yeah,
so
we
had
restore
progress,
and
this
is
actually
kind
of
feeding
into
the
next
point.
So
we
have
this
design
dock
for
the
rest
of
our
restore
progress,
reporting
that
was
merged
and
also
scott's
design,
dock
that
was
merged
for
allowing
our
restore
item
action
plugins
to
wait
on
the
additional
items
to
be
restored,
and
I
wanted
to
check,
I
think,
dylan's
still
on
the
call.
H
H
Probably
we
would
probably
get
started
on
this
the
next
month,
or
so
as
far
as
when
it
would
actually
be
done.
I
can't
really
speak
to
that.
It's
going
to
depend
on
if
we
can
get
more
than
just
a
layover,
but
I
know
he
was
the
one
that
was
looking
into
this.
He
had
to
put
it
on
hold
temporarily,
but
I
can
sync
up
with
him
and
write
you
guys
in
slack.
We
have
a
better
idea.
H
D
So
in
terms
of
a
rough
timeline,
we
were
hoping
december,
given
we
have
two
weeks
left
and
that's
kind
of
a
slow
momentum,
probably
looking
more
like
january
at
this
point,
but
just
to
to
communicate
what
from
our
what
our
team
needs
out
of.
It
is
mostly
the
multiple
credential
handling,
depending
on
how
long
the
timelines
are.
I
I'm
not
going
to
guarantee
we'll
hold
the
release
up
for
certain
things,
but
we
can
certainly
talk
if
it's
a
couple
days,
not
a
big
deal,
but
if
it's
longer
than
that.
H
D
H
D
That's
that
work
was
taking
the
restore
item,
action
plugins
and
allowing
them
to
wait
for
an
additional
item
and
saying:
hey?
Is
that
item
ready
and
if
it's
not,
then
I
will
wait
to
restore
the
current
the
current
item
so
yeah.
If,
if
that
is
a
priority,
I
just
want
to
call
that
out
that
one,
since
it
involves
changes
to
the
plug-ins.
I
know
scott
had
made
some
changes
in
there
to
be
more
careful
around
it,
not
breaking
stuff.
So
I
don't
think
it's
going
to
be
backwards
incompatible.
H
A
G
G
How
can
I
explain
that
it?
I
wish
I
I'll
be
a
little
bit
prepared
to
give
you
the
link,
but
the
fact
that
we
have
to
install
crds
on
the
tar
target
cluster
to
have
our
valero
plugins
to
run
successfully
otherwise
developer
plugins
just
run
on
every
single
item.
Oh,
oh
right!
Okay.
I
know
which
one
you're
talking
about
yeah.
He
fixed
it
a
couple
weeks
ago,
so
I
just
want
to
know
because,
right
now
we
are
installing
a
bunch
of
crds
on
on
target
clusters
just
to
have
our
plugins
to
be
functional.
D
I
can
look
into
that
and
see
if
we
can
get
a
one
five
three
out
in
the
next
week
or
so
for
that
fix
one
1.6
is
probably
going
to
be
january
at
this
point,
but
yeah
for
that
for
bug
fixes.
How
long
has
it
been
since
one
five?
Two,
it's
been
like
two
months:
okay,
almost
two
months:
okay,
yeah.
That
was
kind
of
a
big
big
thing,
because
we
thought
we
had
fixed
it,
but
we
only
fixed
one
side
of
it.
Item.
G
D
G
A
G
G
A
All
right
just
a
reminder:
we
will
not
have
the
community
meeting
on
december
22nd
and
29th
of
of
this
year,
taking
some
time
off
for
the
holidays,
we'll
remind
everyone
next
week
as
well,
but
yeah
cancellations
will
be
coming
if
they
are
still
there
in
your
calendar.
After
those
cancellations
have
gone
out,
make
sure
you
delete
them
because
we
won't
be
here
and
yeah
nolan.
Do
you
want
to
give
a
community
shout.