►
From YouTube: Velero Community Meeting - September 6, 2022
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
So
yeah
so
I'm
running
a
meeting
again
today,
Orlin
is
still
on
vacation,
he's
the
community
manager.
So
just
a
couple
of
quick
updates.
First,
on
the
on
the
release
timeline,
we
have
a.
We
have
decided
to
do
a
192
release.
We
have
a
PR
here.
Are
there
an
issue
here
that
has
oh
and
just
to
clarify
my
screen?
Sharing
fine
people
see
that.
B
A
So
we
have
a
an
issue
here,
so
it's
not
switching
to
that,
but
the
to
list
the
issues
we
want,
our
PR
as
we
want
to
cherry
pick
to
192..
A
So
if
anybody
has
any
other
merged
issues
on
Main
that
you
think
we
ought
to
consider
for
192,
this
is
the
issue
we
can
discuss
those
and
again
mostly
you
know,
if
they're
bug
fixes
relatively
small
changes
and
no
crd
changes,
we
should
be
able
to
kind
of
get
those
in
if
somebody
has
a
reason
they
need
them
for
192.
for
110
..
Sorry,.
B
I
missed
the
last
meeting,
so
add
any
of
these
like
blockers
or
so
is
there
a
reason
why
we
are
building
192
in
short,
duration,.
A
B
A
Mean
I
think
there
is
this
I
believe
one
of
this
is
a
CDE
was
involved,
and
then
there
were
some
other
issues.
I
know
that
on
the
red
hat
side
and
I
had
identified
that
we
have
customers
that
are
hitting.
A
You
know
that
kind
of
have
a
fairly
significant
impact
on
like
with
one,
for
example,
that
had
a
fairly
significant
impact
on
restore
it
was
causing
restore
errors.
Anytime,
someone
was
using
the
the
data
mover
that
we're
adding
on
the
oedp
side.
So
basically,
these
are,
you
know,
bug
fixes
that
and
then
with
the
CDE.
A
We
wanted
to
get
that
out
anyway,
and
but
but
while
we
have
this
release
identified,
if
there
are
any
other
bug
fixes
that
you
know
are
impacting
people
that
would
make
sense
to
get
into
192..
This
doesn't
affect
the
110
time
frame,
which
is
ongoing.
A
The
the
meeting
two
weeks
ago
was
announced
that
we
were
push
pushing
the
previous
dates
back
by
a
couple
of
weeks
around
110,
so
the
future
fees
for
110
is
now
a
week
from
today,
future
complete
on
the
19th
of
October
I
hope
to
actually
get
the
release
out
release
candidate
starting
early
November
or
ga
mid-november.
B
So
power
protect,
we
are
releasing
power
protect
this
month
and
going
with
191.
So
that's
why
I
was
curious
if.
A
That
is,
if
I'm
trying
to
you
know
you
might
have
a
look
at
those
those
issues
to
see.
If
there's
any
of
those
that
affects
you,
I
I
think
one
of
them
was,
for
the
specific
I
mean
the
cve
that
I
know
of
that
I
think
was
on.
Here
is
relating
to
the
go
Lang
version
where
we're
building
with
one
117
11
and
there's
a
CE
identified,
16
117,
12.
A
I,
don't
remember
if
any
of
the
other
ones
are
cdes
or
not
with
those.
Let's
see
that
okay,
that's
yeah,
there's
an
issue
with
storage
class
name
that
was
causing
a
panic.
Okay.
This
is
just
relating
to
CSI
plugin.
A
This
is
another
one
that
was
causing
a
panic,
so
we
had
a
couple
of
bugs
that
were
causing
the
valeripod
to
crash,
and
then
there
was
actually
a
few
of
those
and
then
again
I,
don't
see
the
CV
listed
here,
but
I
have
seen
the
pr
I
think
it
was
merged
already,
or
that
does
address
that.
So,
if
none
of
those
specific
bugs
are
relevant
to
your
use
cases,
then
you
know
you
don't
need
to
worry
about
that.
I,
don't
know!
A
If
we
have
a
precise
data
identified
for
192,
yet
I
think
the
I
think
when
it
came
up,
although
I
don't
know
if
this
is
firmly
decided,
it
was
the
idea
was
probably
by
the
end
of
the
month.
End.
A
Okay
and
then
on,
there's
a
status
update,
so
I'll
go
ahead
and
go
through
my
own
I've
taken
over
the
plug-in
version.
He
worked
that
Fong
was
working
on.
We
got
his
PR
merged
and
then
I've
got
a
couple
of
other
follow-up
mpr's
relating
to
the
actually
I
said.
That's
the
mistake:
I
should
fix
it.
It's
the
back
of
item
action
in
our
store
pris
that
are
waiting
review.
A
A
That's
part
of
that
to
me
too,
because
part
of
the
way
with
this
works
is
that,
if
you're,
adding
a
new
function,
for
example
to
the
in
your
V2
you're,
also
as
part
of
this
provide
an
adapter
which
will
provide
you
know
default
behavior
for
that
function
as
appropriate,
so
that
in
existing
V1,
plugin
will
still
work.
So
you
don't
have
to
immediately
upgrade
all
your
plugins
as
long
as
the
default
behavior
is
acceptable
for
your
use
case.
A
That's
not
those
aren't
PR's
to
merge.
Those
are
just
to
show
every
kind
of
to
kind
of
test
and
prove
everything
in
the
previous
TRS
do
what
we
need
them
to
do
to
enable
V2
plugins,
but
those
can
also
be
used,
then,
as
a
model,
once
we
identify
what
that
B2,
plug-in
API
looks
like
and
I'll
get
to
something
later
that
relates
to
that
as
well.
A
Next
steps
for
this
would
be
to
do
similar,
refactoring
for
ReStore
item
action,
volume
snapshot
or
another
and
other
plugins
to
kind
of
there's
a
certain
amount
of
work
to
get
everything
ready
kind
of
moved
around,
because
basically,
these
V2
plugins
have
to
exist
in
different
packages,
different
good
packages
and
kind
of
side
by
side.
A
I
also
have
still
the
pr
out
for
the
item.
Action
plug-in
design-
and
this
is
the
design
to
allow
plug-ins
both
for
store,
item
action
and
backup
item
action,
as
well
as
volume
snapshot
just
to
initiate
along
a
long
running
operation,
possibly
run
by
another
controller
as
part
of
the
plug-in
execute
and
then
Valero
will
then
call
back
into
it
later
to
check
on
progress
so
that
it
won't
move
the
plug-in
to
the
backup
or
restore
to
complete
it
until
those
background
tasks
are
done.
A
So
this
is
just
the
design
for
that,
not
the
implementation.
Once
that's
approved
one
of
the
consequences
there
is
then
to
because
that
design
involves
API
changes
to
the
backup
item,
action,
restore
item,
action,
plugins
that
were
and
volume
snapshotter.
So
once
we
have
the
versioning
infrastructure
in
place,
then
we
can
go
ahead
and
Define
those
API
versions
with
the
additional
method,
arguments
and
new
methods
needed
to
start
that
work.
So
that's
that's
the
kind
of
tie
in
there
between
those
two
I
also
have
one
more
PR
of
it's.
A
I
got
feedback
last
week
for
adding
volume,
snapshot,
location
credentials.
This
brings
vsl
API
kind
of
up
to
the
same
level
that
bsls
are
which
allow
you'd
have
different
credentials
for
different
locations
and
the
basically
the
feedback
was.
This
is
good,
but
we
need
to
update
the
doc,
which
is
fine.
So
the
plan
is
to
update
that
Doc
and
comments
to
kind
of
response
to
feedback
this
week.
A
So
I'll
be
working
on
that
and
then
again
ongoing
work
with
the
plug-in
versioning
as
well.
So
that's
it
for
me
in
terms
of
status.
A
A
Okay
and
in
shubham
did
you
you
had
a
topic
about
192.
A
Yeah
yeah
I,
haven't
I,
haven't
seen
any
actual
trade
pprs
yet
for
those
I
know
they're
on
the
list
and
they
were
identified.
I
I
saw
in
slack
that
it
was
kind
of
a
comment
to
say
hey.
We
should
do
this,
but
yeah
I
haven't
seen
those
PRS
Cherry
Picked.
Yet.
B
Okay,
maybe
we
can
see
some
activity
next
week,
like
District
yeah.
A
Right
and
I'm
not
sure
what
the
what
the
code
freezes
for
that
either
something
else
we
should
probably.
Maybe
we
should
Ping
On
slack,
just
kind
of
kind
of
assess
that
offline
to
see
because
yeah,
we
need
to
make
sure
that
basically,
everything
in
that
issue
identified
needs
to
get
into
onto
the
one
nine
Branch
and
again.
If
anybody
else
identifies
anything
else
that
we
need,
we
need
to
make
sure
those
everything
slips
of
the
cracks
there,
because
I
know
for
oedp
we're
relying
on
some
of
those.
B
A
The
ones
where
they
you
know
you
know
they
crashed
the
Valero.
Well,
then,
I
guess
wait.
Two
two
guns
areas:
there's
the
ones
that
crashed
Valero,
there's
two
of
those
really
and
in
there
and
I
think
only
one
of
those
affected
us
directly
and
then
there
was
the
one,
the
the
fix
where
we
actually
are
failing
restorers
then
we
shouldn't
be
failing,
though,.
B
A
Okay,
did
anybody
else
have
any
additional
questions
or
topics
to
bring
up.
C
So
hey
this
is
pradeep,
so
I
was
kind
of
last
week
that
we
will
explore.
I
was
talking
to
some
of
the
customers
who
use
Valero,
and
one
of
the
thing
came
up
in
kind
of
actually
a
couple
of
things
came
up.
One
of
that
like
can
we
support
NFS
as
a
storage,
location
for
backup,
and
this
kind
of
widespread
topic,
I've
heard
multiple
times.
C
A
C
A
Yeah,
that's
good
to
know
and
I
know.
I've
heard
the
same,
I've
seen
the
same
request.
You
know
on
the
right
side
from
customers.
I
know
I
know
our
current
answer,
for
that
is
that
you
can
always.
You
know,
bring
up
an
S3,
compliant
Object
Store
on
your
NFS
infrastructure.
You
know
something
like
you
know,
nuba
or
something
like
that
or
Mineo.
C
The
same
thing,
I
was
mentioning
only
the
question
was
coming:
it's
additional
cost,
yeah,
bringing
it
any
S3
bucket
and
NFS
is
more
of
a
standard
already.
Customers
have
invested
sure,
so
they
look
at
overall
cost
of
ownership
and
that's
what
they
say
maintaining
another
vendor.
Another
kind
of
cost
is
yeah.
A
And
and
I
I
guess
I
guess
the
I
know
there
was
an
issues
like
I
was
I
was
looking
this
up
a
year
or
so
ago,
and
you
know
I
I
know
there
had
been
issues
put
into
GitHub.
You
know
about
this
in
the
past,
I
I,
don't
remember
whether
there
were
any
of
those
were
ever
investigated.
You
know
whether
anyone
looked
into
you
know
how
much
work
it
would
be.
You
know
to
actually
build
in
the
code
to
work
directly
on
a
file
system
rather
than
an
object
store.
B
A
A
Right
exactly
you
know
it,
it
might
be
worth
searching
the
the
closed
issues
to
see.
What
was
how
those
were
you
know
resolved
in
the
past.
I'm
absolutely
didn't
do
anything
for
them,
but
I
mean.
Was
there
a.
C
C
Something
yes
and.
A
And
just
I
guess
the
starting
point
would
be
to
to
search
for
previous
issues
to
say:
okay
did
someone
else,
do
an
investigation.
You
know
two
years
ago
and
we
can
build
off
that
or
do
we
are
we
starting
over
basically
I
think
that
makes
sense,
especially
you
know,
as
we
started
into
the
question
of
111
scope,
you
know.
Is
this
something.
A
Into
is
creating
a
new
GitHub
issue
for
is
kind
of
a
starting
point
of
discussion.
C
Because
I
was
also
looking
at
I
think
we
store
the
backup
as
a
tarp.gz
file.
It's
end
of
the
day
kind
of
for
one
big,
zip
file.
You
should
able
to
kind
of
right
onto
any
file
system
that
files
yeah,
if
you
use
metadata
and
all
information,
maybe
an
object
format,
but
the
final
backup
file
is
going
to
charge
it
right,
which
we
download
and
do
that
so
probably
something
we
can
again.
As
I
said,
maybe
we
can
open
a
new
issue,
but
I
know
there
is
a
existing
issue
which.
C
That
yeah,
it's
still
open.
It's
one
double
two:
nine
I.
A
Guess
if
you
could
just
link
it
into
the
discussion
topics
here
so.
B
A
A
C
Yeah
I'll
just
add
it
to
a
discussion
topic
and
the
second
thing
which
kind
of
question
came
and
then
I
was
not
very
much
sure
that
how
to
do
that
is
all
about
like
if
we
have
a
big
volumes,
how
do
we
kind
of
do
incremental,
snapshotting
and
all
on
those?
So
it
was
around
like
if
you
have
an
actual
database
and
all
and
those
are
the
kind
of
a
flat
file
so
yeah.
C
A
C
However,
what
options
we
can
kind
of
consider
in
future
from
blogspace
or
something
which
is
which
can
make
efficient
incrementals
and
all
especially
the
databases
and
all
kind
of
things.
So
it's
one
consideration,
I
heard
and
another
one
which
kind
of
not
directly,
but
when
I
was
analyzing
like
we
need
a
better
like
for
objects,
backup.
We
have
lot
of
strategies
in
build.
C
For
example,
you
can
override
you
can
skip
this
particular
object,
type
and
all,
but
for
volumes
we
have
a
very
flat
strategy
like
you
do,
annotation
it
will
escape
or
not,
which
is
kind
of
a
more
of
a
someone
has
to
do
a
manual
work,
but
so
strategy
could
be
more
of
like
if
there
are
the
different
kind
of
volumes
in
a
cluster
anything
volume
which
is
a
CSI
compliant.
A
C
Based
yeah,
so
I
have
a
kind
of
a
backup
volume
strategy
base
where
users
say
Okay.
First,
preference
is
my
CSI
volume
snapshot.
Second,
is
my
any
data
copy
option?
I,
don't
want
to
say
rustic
or
copia.
We
should
say
something
more
generic
term
like
copying
data
to
a
backup,
location
or
something
or
right.
A
B
A
C
You
want
to
call
it
something
like
that
and
then
we
Define
user,
to
give
a
strategy
at
Valero
backup
level
so
that
they
don't
need
to
worry
about
annotation
or
or
kind
of
because
today,
if
you
start
using
a
stick
or
in
feature
copier,
that
takes
the
first
preference
and
then
CSI
volume
snapshot
takes
a
second,
but
in
the
future,
since
we
started
support
the
CSI
volume
snapshot.
Ideally
that
should
take
a
first
preference
as
a
more
because
that's
a
more
crash,
consistent
versus
that.
So
we.
A
Should
look
at
those
right
track?
I
should
point
out
in
terms
of
first
preference.
We
actually
have.
We
work
both
ways.
Basically,
there's
there's
a
flag
to
say:
well,
The
Rustic
is
default
or
not
the
default,
and
so
and.
B
C
A
If
CSI
is
the
default
well,
CSI
or
snapshotting
is
the.
A
C
C
C
B
A
That
sounds
good,
yeah,
yeah,
yeah
and
probably
an
issue
for
that
for
that
kind
of
rules
based,
so
we
can
kind
of
discuss,
you
know
pros
and
cons,
and
you
know
what
what
kinds
of
implementation
would
work
there
and
you
know
and
like
you
said
you
know,
identify
you
know.
This
is
something
that
came
from.
You
know
customer
requests,
so
that's
kind
of
so
we
kind
of
know.
This
is
something
that
you
know
there's
a
real
world
use
case
for,
rather
than
just
you
know,
some
idea
hey.
This
might
be
nice
because.
B
A
You
know,
because
we
may
have
you
know:
IDP
users
at
red
hat
that
that
you
know
want
something
similar,
and
so
we
want
to
consider
their
needs
into
this
as
well
and
kind
of
you
know,
even
though
a
single
customer
from
a
single
you
know,
vendor
might
be
the
kind
of
a
reason
to
add
the
issue.
A
Next
week's
meeting
will
be
at
the
the
8
PM
time
kind
of
the
Beijing
Centric
meeting,
and
that's
the
point
where
we'll
see
where
we
are
with
features-
and
you
know
with
the
Feature
Feature
freeze
and
you
know,
kind
of
probably
additional
hey.
This
is
in
or
this
thing
is
not
in
those
kinds
of
discussions
will
be
relevant
there.