►
From YouTube: Velero Community Meeting - Nov 2, 2021
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
A
We're
going
to
go
through
some
status
updates
and
then,
if
you
have
any
discussion,
topics
that
you
want
to
bring
up,
please
raise
them
here.
I
know
we
had
some
discussions
yesterday
that
we
might
want
to
bring
up
here
or
we
might
want
to
bring
up
next
week
as
well,
but
yeah
first
off
some
status
updates
from
me.
I
want
to
welcome
abigail
mccarthy
back
from
parental
leave,
super
happy
to
have
you
back
abby.
A
Super
happy
to
have
you
back
yeah,
I
I
I
I
heard
a
baby
awesome
yeah
super
happy
to
have
you
back
abby,
it's
going
to
be
great
to
have
you
here
and
help
out
with
all
the
technical
documentation
again.
So
thank
you
for
for
being
here.
A
The
next
item
is
something
that
was
brought
up
yesterday.
We
had
an
internal
call
and
apparently,
when
I
set
up
the
community
meeting
for
the
beijing
time
zone
community
meeting
or
the
beijing
timezone
focused
community
meeting,
I
set
it
up
in
eastern
u.s
time
and
just
added
it
as
the
as
the
us
as
the
beijing
time
zone.
So
I
will
change
out
that
community
meeting
because
it
should
follow
the
beijing
time
zone,
so
it
shouldn't
be
affected
by
our
daylight
savings
times
or
anything
like
that.
C
Yeah
I
clicked
the
button
before
I
clicked
the
button.
So
last
week
I
was
on
community
support,
so
I
went
through
a
few
of
those.
I
don't
think,
there's
anything
terribly
exciting
there
we're
going
to
be
updating
the
support
rotation
with
ming
and
bruce
being
added
to
that,
I'm
going
to
drop
out
of
support
for
a
while
and
then
also
just
continuing
on
with
the
upload
progress,
the
project
that
never
ends
and
working
towards
testing
out
making
sure
everything
works
there
and
we're
working
on
getting
the
item.
D
Hi
everyone
so
last
week
I
felt
like
I
spent
quite
a
lot
of
it
catching
up
after
pto,
but
in
the
middle
of
all
that
I've
been
working
on
the
astrolabe
crd
design
with
dave,
so
we
know
we're
going
to
be
eventually
kind
of
integrating
that
into
valero
more
as
part
of
the
daily
movement,
and
so
this
is
kind
of
some
of
the
groundwork.
For
for
that,
and
I
also
had
a
a
thinking
phone
last
week
to
chat
about
plug-in
versioning.
D
So
we've
got
some
kind
of
follow-up
steps
and
things
that
we're
going
to
be
doing
as
part
of
that.
So
I
so
we're
going
to
be
working
on
primarily
doing
like
reviews
and
getting
testing
in
place
so
unit
testing
integration,
testing
and
then
we'll
try
and
get
the
branch
without
work
set
up
for
to
run
on
some
of
the
internal
kind
of
testing
pipelines
that
we
have
to
make
sure
that
we're
not
breaking
any
existing
functionality.
That's
something
we
definitely
want
to
avoid.
D
Then,
like
dave
mentioned,
we
need
to
like
his
upload
progress
prs
and
things
are
outstanding.
So
there's
there's
that
and
then
quite
a
few
other
pr's
that
design
docs
and
so
on
that
I
really
want
to
try
and
get
reviewed
this
week.
So
that's
going
to
be
a
focus
of
mine
as
well
to
try
and
get
those
things
closed
and
get
them
in.
A
D
No,
it's
not
it's
still
kind
of
early
days
for
that,
but
once
it's
in
place
it
will
it'll,
be
obviously
put
for
review
and
so
on.
A
E
Sorry
about
that
more
used
to
the
google
me
keyboard
shortcuts
than
the
zoom.
I
am
from
cloudbees,
I'm
which
I'm
here,
because
we
have
customers
who
are
very
interested
in
cross
region
restorers,
particularly
on
aws,
which
is
something
I
know
it's
not
currently
supported.
E
I've
been
playing
with
this
some
and
I
put
together
kind
of
a
proof
of
concept,
pull
request
that
is
contained
in
the
aws
plugin
specific
to
bbs,
so
it
doesn't
apply
to
other
kinds
of
volumes,
but
just
for
ebs,
then
that
seems
to
work,
but
that
basically
does
a
snapshot
replication
between
regions
and
then
it
builds
on
another
open,
pull
request
in
that
plug-in
that
that
did
metadata
rewrites
of
the
volumes
to
annotate
them
with
the
correct
zone.
E
So
I
just
extended
that
to
annotate
it
with
the
right
region
as
well
and
there's
some
other
miscellaneous
changes
in
there
to
do
thing
too,
to
fix
some
other
problems.
They
noticed
at
scale
on
aws
for
things
like
api,
retry
requests,
and
things
like
that
so-
and
this
also
is
only
practical
with
a
lot
of
volumes.
E
E
E
Those
pr's
are
currently
closed
and
there
was
a
comment
that
this
is
something
that's
on
the
backlog
or
being
planned
for
some
future
version
of
lero.
But
I
couldn't
get
information
about
what
what
specifically
was
planned
here
or
what
form
that
would
take.
So
I'm
wondering
if
people
here
have
more
information
about
that
or
know
what
I
could
help
contribute
to
yeah.
So
currently,
this.
C
C
We
are
going
to
have
both
cross-region
movement
and
actual
data
movement,
so
we
can
do
cross-cloud,
so
I
want
to
get
all
that
stuff
into
those
apis
and
down
there.
Rather
than
doing
like
one-off
solutions
for
the
different
cloud
providers,
we
could
consider
doing
something
specifically
for
ebs,
but
then,
as
you
said,
it
doesn't
work
really
well
until
we
go
into
running
all
the
snapshotters
in
parallel
and
running.
Those
things
in
parallel
is
not
something
we're
ready
to
do.
Yet.
It's
really
something
that
I
mean.
C
There's
there's
ordering
issues
in
terms
of
you
know:
when
do
the
pvs
become
available,
or
when
do
we
snapshot
things
on
the
on
the
other
side,
so
that
you
know
we
have
to
do
things
in
the
right
order?
Also,
there's
going
to
be
issues
in
some
systems
with
the
number
of
outstanding
requests.
E
So
with
the
parallelism,
you
can
complete
all
of
the
backups
within
three
or
four
minutes
and
it
works
fine.
You
do
get
you
do
get
some
api
rejections
from
aws
saying
you
have
too
many
concurrent
snapshot
requests
in
flight,
which
is
why
I
also
had
to
amend
it
to
to,
as
I
said,
do
api
retries
under
certain
conditions.
If
you
do
the
retry,
then
it
does
work.
C
The
region,
each
each
sword
system
has
got
its
own
set
of
funkiness
there
that
we
need
to
take
care
of
there's
other
things
like
we
want
to
be
able
to
do
the
ordering.
In
fact,
it's
been
reported
that
the
pre
and
post
restore
hooks
are
not
pre
and
post
backup
hooks
on
the
pods
are
not
bookending
the
snapshotting
and
backups
properly
of
the
volumes.
C
C
E
I
thought
about
having
it
also
wait
for
the
replication
to
complete
you.
Could
it
would
add
a
little
bit
of
time
to
the
backup
I
think
that
depends
on
it
doesn't
seem
to
be
necessary
because
the
backup
actually
becomes
usable
immediately
because
of
the
way
aws
implements
the
the
volume
restore
from
snapshot.
It
does
a
lazy
pull
of
blocks,
so
in
that
case
it
seems
to
work
as
is,
but
yeah,
certainly
for
different
cloud
providers.
It
would
be
very
different,
so.
C
Yeah,
so
that's
kind
of
what
we
want
to
do
is
going
to
get
the
general
model
in
there
for
this,
so
the
parallelism,
I
hear
you
it's
an
issue,
but
it's
not
an
easy
fix
and
there's
really
there's
like
a
hierarchy
of
things
that
need
to
happen.
So
that's
that's.
Why
we're
kind
of
like
putting
it
off
until
we
understand
it
better
and
get
a
better
overall
design
together.
C
Going
to
happen
in
asteroids,
some
stuff
will
happen
in
valero
and
the
the
next
move
is
to
get
things
like
data
movement
into
valero,
which
would
then
allow
us
to
back
up
and
restore
snapshots
across
regions.
So
that'll,
probably
wind
up
being
sometime
this
summer
for
lucky
to
get
all
that
together.
E
Okay,
all
right
are
there
tracking
issues
somewhere
in
astrolabe
that
cover
this.
This
sort
of
thing
is
that
already
filed
or
discoverable.
C
Most
of
the
things
are
in
the
valero
project,
I'm
not
sure
if
we've
got
an
issue
open
on
parallelism.
Let's
take
a
look.
C
Yeah
so
there's
28.88
in.
C
I'll
put
in
the
chat
but
yeah,
so
we
had
some
some
things
there
figuring
out,
like
ordered
resources,
concurrent
resources.
C
Yeah
I
mean
there's,
there's
definitely
issues
here,
but
there's
also
reasons
why
we're
not
just
jumping
right
into
it,
because
it
it's
getting
worse
and
worse,
like
the
ordering
of
restore
resources
on
restore
as
we
get
into
the
the
applications
like
operator
driven
things.
C
C
Yeah
and
we
could
definitely
use
some
help
working
through
the
parallelism
issues
so,
but
I
think
we
need
to
work
through
them.
You
know
kind
of
from
a
top-down
way
in
in
terms
of
like
how
do
we
get
things
running
in
parallel,
we're
going
to
be
doing
snapshots
for
things
besides
simply
volumes,
so
that
we'll
be
able
to
start
snapchatting
applications
and
we're
currently
working
on
a
project
with
cast
and
to
implement
their
canister
backup
description
language.
C
I
guess
and
use
that
from
astrolabe
and
that's
you
know
that's
going
to,
for
example
that
may
call
out
and
snapshot
volumes,
and
so
we
need
for
that
to
be
able
to
issue
things
in
parallel.
But
then
we
need
to
be
able
to
block
at
the
right
points
so
that,
for
example,
like
a
canister
blueprint
may
snapshot
six
volumes.
It
can
issue
all
those
in
parallel,
but
then
it
would
like,
for
example,
be
doing
a
quies
on
class
and
it
should
be
doing
a
class.
It
can
issue
all
the
six
snapshots
in
parallel.
C
C
So
rustic,
you
know
we'll
we'll
and
we
may
be
combined
well,
actually
we're
going
to
try
to
combine
rustic
backups
with
say
snapshots
so
that
we
can
take
a
snapshot
then,
on
systems
that
don't
have
data
pads
like
a
generic
csi
system,
you
need
to
clone
and
then
attach
the
volume
and
then
extract
the
data.
But
that
can
be
done.
You
know
in
the
background,
because
once
the
snapshot's
taken,
we
have
a
consistent
point
in
time.
C
F
Yeah,
so
so
thank
you
jesse
also
yesterday,
he
bring
our
very
good
point.
Is
we
don't
have
a
formal
list
of
what's
going
into
the
plugin
interface
version?
2.
F
right
now
I
only
implemented
you
know,
adding
a
contact
into
the
function,
but
what
else
should
those
should
go
with
that
version?
Two?
I
do
not
have
any
formal
list
yet
so
should
we
start
you
know
creating
that
list
and
we
start
adding
it
up,
adding
it
in
the
implementation,
because
now
we
have
the
feature
branch,
and
this
is
the
time
to
add
in
if
we
need
to
add
anything-
and
this
is
the
time
so.
C
D
Yeah
yeah,
so
there
was
them.
There
was
another
design
proposal
I
know
from
from
scott.
Oh
he's,
it's
got
something
to
call
him,
so
I
know
that
he
was.
I
can't
remember
what
it
was,
but
I
know
that
he
needed
to
add
another
function
to
one
of
the
item.
Action.
B
Yeah
that
was
basically
to
allow
items
on
restore
to
wait
until
so
so
we
have
the
the
additional
items
that
are
returned
and
then
there's
a
another
call
that
was
going
to
be
added
to
the
stratum
action.
To
call
for
an
item
to
say:
is
this
thing
ready,
so
that
would
be
kind
of
a
new
method
added
to
the
stratum,
action.
D
Yeah,
I
think
I
think,
to
your
to
your
point,
but
listing
them
all
and
getting
them
in
this
branch.
I
think
that
we
can.
D
I
hope,
with
with
the
with
the
branch
that
we're
currently
working
on,
is
that
getting
the
plugins
ready
to
be
versioned
would
be
separate
from
any
specific
changes
in
the
plugin
apis,
which
would
require
versioning
yeah.
B
I
think
it
would
be
confusing
if
we,
if
we
started
adding
new
things
to
the
you
know,
we
want
a
version
of
all
of
these
plug-ins
exactly
as
they
exist
now,
but
as
a
version
one
that
is
versioned.
You
know
so
so
that
the
first
version
supportable
iteration
of
this
changes
nothing
and
then
after
that's
done,
then
we
can
put
the
version.
To
I
mean
we
don't
need
a
release
with
that.
B
If
we're
calling
the
second
one
version
two,
you
know,
but
I
mean
what
it
would
so
that
you
know
you
have
basically
the
first
change
that
adds.
Versioning
doesn't
change
any
of
these
specific
interfaces,
it's
all
there
and
then
once
that's
committed
to
maine
or
the
branch
that
we're
working
on
whatever.
At
that
point
you
know,
then
you
can
add
the
version
two
to
specific
plugins
types.
You
know
we're
sweating
action,
whatever
that
will
add
whatever
we
think
we
need
to
add
to
the
support.
You
know
the
first
round
of
new
features.
D
Yes,
I
think
yeah
we
we
can
go
through
the
branch
that
we
currently
have
and
try
and
see
like
what.
Specifically
from
that,
we
can,
because
I
know
that
you
mentioned
you
started
adding
some
of
the
context.
But
maybe
I
think
that
I
think
that's
like
really
important
to
do
from,
like
a
perspective
of
like
making
sure
that
the
approach
works
and
and
so
on,
but
I
think
yeah,
perhaps
as
part
of
that
we
can
try
and
get
the
the
work.
D
That
is
specifically
just
rearranging
the
the
code
into
the
new
version
format
and
get
that
in
first
and
then
we
can
build
upon
the
the
work
of
like
adding
the
contacts
and
so
on
into
each
of
the
apis
and
then
actually
versioning
them,
because
I
think,
as
long
as
it's
all
like.
So
I
think
whenever,
whenever
we
all
like
initially
discussed
this,
I
think
it
was
outside
that
we
could
have
multiple
changes
go
into
a
particular
versioned
api
as
long
as
it
was
all
within
like
one
release
cycle.
D
So,
for
example,
we
could
get
the
context
work
in,
but
we
don't
necessarily
need
to
get
the
changes
that
scott
wanted.
In
with,
like
the
v2
committed
for
store
item
action,
so
say
we
had
a
change
for
adding
context.
Then
we
could
have
another
change
within,
for
example,
the
one
eight
time
frame
and
I
could
all
go
in
under
v2,
and
then
it
was
only
like
at
one
one
line
when
we
would
need
to
version
it
again.
F
How
about
this?
Let's
call
a
separate
meeting
when
we
walk
through
all
the
chain
that
I
have
made
so
far
into
that
feature,
future
branch,
and
then
we
will,
you
know
if
we
need
to
split
out
the
chain
into
two
separate
chain.
We
will
do
that
during
that
meeting.
Let's
take
offline,
maybe
we
can
discuss
continue
on
slack
and
then
we
I
I
recommend
that
we
have
at
least
one
or
two
hour.
You
know
code
walkthrough,
each
of
the
team
there's
so
many.
F
So
yeah,
so
we
don't
have
to
check
out
the
meeting
time
with
the
team
here.
B
Yeah
yeah.
That
would
be
good.
F
Okay,
we
will,
we
will
try
to
schedule
a
meeting
between
three
of
us
to
walk
through
all
the
change,
and
you
can
also
have
comments,
and
you
know
feedback
on
that.
Okay,.
A
A
No
takers,
all
right,
then,
let's
dive
into
some
contributor
shout
outs
got
a
few
here.
It's
been
a
while,
since
we
did
contributor
shout
outs,
I've
been
traveling
and
forgot
to
do
them,
and
then
we
also,
I
was
also
on
ptl
last
week.
So
now
we
have
a
few
here.
So
this
one.
Does
anyone
want
to
help
me
out
with
these
ones
we're
good.
D
Yes,
I
can,
I
can
help
I
I've
seen
some
of
these.
I
may
not
know
the
context
and
I
can't
involve
them,
but
I
can
help
where
I
can
here.
So
this
is
a
change
from
from
frankie.
So
there
was
a
a
bug
discovered
with
some
recent
changes,
we're
making
to
use
cube
builder,
where
the
valero
controllers
were
operating
on
flair
resources
created
outside
of
the
blair
namespace.
So
this
is
a
change
to
fix
high
we're
using
cube
builder
to
only
operate
on
the
blair
namespace.
A
Yeah,
I
remember
I
remember
the
discussion
about
this.
It
was
a
little
a
little
strange
so
good
to
have
that
fixed.
Thank
you.
D
So
this
is
a
change
from
danfung
to
to
update
the
one
eight
roadmap,
so
she's
gonna
be
focusing
on
some
technical
health
work
within
valero.
So
I
think
that
was
just
yeah
just
adding
the
the
issue.
That's
could
be
tracking
that
so
thank
you,
john
funk,
for
taking
that
work
on.
D
This
is
from
eleanor
updating
the
road
map
and
what
we're
planning
to
do
for
for
one
eight.
So
that's
the
focus
for
the
blair
team
until
kind
of
next
year.
Early
next
wednesday.
D
D
This
is
from
urgi
kame
of
fixing
the
indentation
within
the
helm.
Chart.
I
didn't
review
this
one,
but
I'm
guessing.
There
was
a
bug
that
was.
D
D
C
A
D
Maybe
there
was
something
maybe
it
was.
There
was
something
that
was
maybe
missed
out
in
one
of
the
pr's.
I
haven't
reviewed
these,
but
maybe
it
was
a
follow-up
or
that
was
required.
So
I
missed.
F
D
Awesome,
thank
you
very
much
both
for
for
adding
that
and
then
making
sure
that
it
works
correctly.
So
this
is
from
matthew.
I
think
this
was
changing
high.
I
can't
remember
which
parts
of
it,
but
I
think
it's
some
of
the
the
service
ports
or
something
it's.
C
The
prometheus
ports
and
how
they're
named-
and
he
added
in
the
http
dash
in
there
and
that's
actually
something
that
I
wanted
to
bring
back
here-
is
because
this
only
fixed
the
fixes
the
helm
chart.
So
we
don't
set,
we
still
use,
we
use
a
completely
different
name
when
regular
valero
install
runs.
So
there
was
the
question
of
you
know.
Should
this
be
something
that
we
you
know?
Is
this
right
fix
and
should
this
be
fixer,
but
it
got
merged
so
yeah.
D
So
simply
different
and
like
the
the
service
that
gets
generated,
if
you
do
like
valero
install
so.