►
From YouTube: Velero Community Meeting - May 16, 2023
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
Okay,
so
this
is
the
May
16th
2023
Villa
Community
meeting
for
the
U.S
Europe
friendly
time
zones
and
I'll
be
leaving
this
today,
so
it
started
with
status.
Shubham.
B
Hey
that's
hard,
so
I've
just
been
doing
some
PR
reviews,
design,
checkups
and
whatever
comes
my
side
regarding
CSI
and
stuff.
That's
it
from
my
side.
Okay,.
A
Yeah
and
I've
also
done
some
design
review
or
sorry
some
PR
reviews
this
week
as
well,
but
that's
pretty
much
all
I've
done
from
the
wellera
of
stream
side
this
week.
So
let's
just
get
into
the
the
discussion
topics.
So
we've
got
a
demo
on
the
AKs
backup
built
a
type
of
Valero.
Let's
go
ahead
and
do
that.
A
C
Okay,
I'll
be
quick
on
the
part
of
going
through
your
demo
video,
it's
a
quick
five
minute
demo,
so
I'll
run
through
you
guys
just
quickly
start
with
what
we
have
built
is
azure.
Backup
in
itself
is
a
platform
that
we
have
internally
within
Azure
to
support
different
different
workloads
and
enabling
backup
capabilities
for
it.
So
it's
already
there
for
VM
for
SQL
Azure
files,
postgres
blobs
a
lot
and
within
that
same
platform
we
brought
in
the
support
for
Azure
kubernetes
service.
C
So
I
have
a
little,
so
basically
they
will.
Basically,
what
we
have
done
is
that
Valero
is
now
integrated
with
Azure
backup
and
the
entire
experience
that
we
have
so
that
it
enables
the
customers
to
backup
the
AKs
cluster
as
they
back
up
any
other
workload
across
Azure,
and
then
they
have
that
single
glass,
pane
of
view
of
our
product
called
backup
center
and
the
newer
version
called
business
continuity.
Center.
C
Just
go
and
have
the
UI
experience
to
manage
their
backups,
their
policies
and
reporting
alerts
and
monitoring
entire
experience,
so
that
becomes
unavailable
as
a
part
of
the
back
side
of
you.
Okay
quickly
highlighted
how
we
brought
in
the
implementation,
so
my
idea
that
we
did
was
that
that
the
Valero
is
now
installed
inside
the
AKs
cluster.
C
As
an
extension,
substantial
extensions
allows
the
Valero
parts
to
be
installed
and
then
we,
the
extension,
connects
with
the
Azure
backup
and
then
the
whole
backup
process
runs
for
the
customer
in
that
context.
So
that's
how
we
have
enabled
a
Valero
inside
a
case
class
transpose
this
demo.
We
can
get
into
it
how
the
enablement
of
the
extension
everything
happened
currently
right
now.
In
this
context,
the
preview
version
that
we
have
launched
our
support
is
available
for
Azure
disks.
C
As
you
are
already
aware,
the
CSI
driver
that
we
have
of
our
Azure
files
doesn't
support
restore
capability,
so
we
are
currently
enabling
only
Azure
disk
based
support
for
backups
in
time
being
once
file
share.
Has
that
capability
enabled
we
will
make
it
available
also
for
our
customers
now
I'll
quickly,
just
jump
into
the
demo,
and
it's
a
bit
vibrant
demo
to
run
through
you
how
the
entire
experience
end-to-end
looks
like
this
involves
installing
the
extensions
which
will
bring
in
the
Valero
pods.
C
C
So
this
is
a
cluster
that
I
have
recreated
with
an
application
deployed
in
it.
So
it's
a
WordPress
that
I
have
deployed.
You
can
see
it's
a
namespace
that
I
have
created.
This
will
be,
I
will
be
backing
up.
This
is
a
deployment
under
that
namespace.
C
This
is
the
underlying
PVC
and
PV,
which
I'm
gonna
go
and
backup
of
it
and
then
there's
a
load
balance
suppose
I've
created
which
allows
me
to
access
the
website,
whereas
extraordinate
another
WordPress
app.
So
you
can
see
this
is
the
website
so
now,
first,
what
I'll
do
is
that
I'll
install
the
backup
extension
Society
cluster,
which
will
enable
it
to
configure
backup
and
performance
too
so
I'll
go
to
the
web
section
of
the
case
portal
and
click
on
the
install
X
Stitch
button.
C
Click
on
next
and
create,
and
with
this
extension
will
be
installed
inside
the
cluster
now
with
extensional
start
cluster
is
ready
to
backup
so
I'm,
going
to
configure
backup
and
I'll
select
the
work,
which
is
the
management
layer
for
backup
backup.
This
all
10
cluster
is
already
as
trusted
access
enabled
in
between
I
click
on
next.
C
And
now
I'll
show
you
the
policy
as
per
which
the
backups
will
happen.
So
it's
a
one
day,
backup
frequency
and
the
backups
will
retake
for
two
days
for
next
and
I'll
specify
the
backup
configuration
so
I
click
on
ADD,
I'll
name,
my
backup
instance.
As
I
said,
we
are
looking
to
backup
only
the
WordPress
app.
C
C
And
now
you
can
see
a
backup
instance
has
been
created
for
this
AKs
cluster.
As
for
the
defined
configuration
now
within
this
backup
instance,
you
can
see
all
the
necessary
details
that
we
have
set
up.
While
configuring,
you
can
see
a
storage
account
and
blob
container
and
the
snapshot
Resource
Group
and
now
click
to
win.
One
demand
backup
this
automatic
backup
will
create
a
recovery
point.
Meanwhile,
you
can
see
we
had
defined
a
backup
instance
notification.
So
you
can
see
the
details
here.
C
C
Meanwhile,
I'll
click
on
refresh
and
now
you
can
see
my
recovery
point
is
recreated.
So
the
recovery
point
is
available
with
this
ready
now,
I
can
perform
a
restore
to
an
alternate
cluster,
so
whatever
I
have
backed
up
like
you
know,
my
WordPress
app
I'll
restore
it
with
alternate
cluster.
C
Now
I'll
go
to
my
alternate
cluster.
This
cluster
already
has
backup
extension
installed
in
it,
but
it
doesn't
have
the
WordPress
app
spaces
section
and
you
can
see.
There's
no
WordPress
app,
which
is
created
so
from
that
backup,
recovery
point
and
the
restore
the
WordPress
application
cluster.
The
original
cluster
go
to
the
backup
section.
You
can
also
do
the
restore
process
from
the
backup
center
open,
Stills,
I'll
click
on
restore.
C
Now
I'll
do
the
resource
configuration
this
namespace
I'm,
just
storing
if
I
need
to
apply
any
filters
enables
API
groups.
Only
those
resources
will
be
stored
if
I
wanted
to
store
my
resources
to
a
different
name.
Space.
The
cluster
or
in
case
of
conflict,
happens
to
what
option
should
I
go
for,
skip
or
pass
the
resources.
C
I
can
say
that
a
store
completed
in
one
minute
and
35
seconds
now
I'll
go
to
the
alternate
cluster,
where
I
perform
the
restore
I'll,
go
to
the
namespaces
and
show
that
the
namespace
has
now
been
created
and
my
WordPress
app
is
deployed.
As
you
can
see,
the
WordPress
Avenue
space
is
created
there.
C
Workload
and
the
deployment
is
visible
as
it's
deployed
this
part
of
the
red
Crystal
process,
I'll
go
to
storage
and
you
can
see
the
pdcrp.
Are
there
I
go
to
the
services,
so
my
load
balance
is
also
deployed
now,
so
it
has
an
alternate
external
IP
and
click
on
it,
and
now
you
can
see.
Website
is
restored
as
this
component
for
the
table.
C
Okay,
next
time,
once
you've
got
to
see
the
demo
for,
in
terms
of
certain
items
that
we
are
working
in
in
the.
C
Is
a
preview
solution
that
is
out
in
public
preview
going
towards
GA,
sometimes
so
as
part
of
it
GA.
We
want
to
bring
out
the
hooks
that
are
there
available
within
Valero.
We
want
to
enable
it
as
part
of
the
ecas
backup
solution
that
we
have
so
that
app
consistency
for
the
customers
can
be
brought
in
other
than
that.
We
have
certain
ideas.
C
The
Azure
backup
platform
in
itself
has
capability
to
enable
ransomware
protection,
so
we
are
currently
working
on
to
integrating
with
that
entire
infrastructure
so
that
the
backups
that
get
created
can
be
predicted
from
ransomware
and
then,
following
by
that,
we
also
are
working
on
to
enable
the
disaster
recovery
part
of
it,
so
that
the
backups
are
available
in
a
different
region
all
together,
so
that
even
if
the
final
region
goes
down,
customer
has
backup
in
the
secondary
and
you
can
enable
it.
C
So
these
are
certain
items
that
we
are
looking
at
currently
within
our
roadmap
and
thinking
about
going
and
building
for
our
solution.
In
terms
of
you
guys
want
to
look,
take
a
look
at
it.
This
is
a
documentation,
link,
I
have
provided
here,
so
you
can
go
and
take
a
look
at
it.
How
it
looks
like
and
can
even
go
and
try
out
a
solution
as
it's
available
in
public
preview.
If
you
have
a
Azure
subscription
to
take
a
look
at
it,
okay,
so.
C
Quick
demo
that
we
wanted
to
bring
out
to
the
community
that
the
solution
that
we
have
built
funny
case
on
top
of
this
is
how
it
works,
and
it
looks
like
just
a
couple
thing
in
the
video
when
you
saw
the
restore
process.
Although
I
started
the
process
within
the
AKs
portal,
the
backups
are
in
no
way
tied
to
the
life
cycle
of
the
AKs
clusters.
C
Even
if
the
cluster
goes
down,
your
backups
remain
as
it
is,
and
you
can
utilize
our
backup
center
Discovery
endpoint
to
go
ahead
and
restore
it
and
down
the
line.
We're
also
introducing
support
for
command
line
interface,
and
also
that
we
won't
even
go
through
the
UI
experience.
You
can
use
the
CLI
or
terraform
and
all
to
perform
backup.
So
that
will
also
be
available
in
time,
see.
A
C
Any
questions
something
please
go
ahead:
yeah
I
had
a
couple
of
questions,
I
think
when
you
are
showing
the
the
demo
you
were
showing
this
from
a
single
cluster
Viewpoint
right
so
from
a
cluster.
You
could
go,
look
at
backups
and
see.
What's
there,
do
you
guys
have
a
multi-cluster
view
where.
B
C
Can
go
to
one
place
and
look
at
all
the
Clusters
and
which
ones
are
protected,
which
ones
are
not
protected,
obviously
trying
to
see
if
there's
a
gap
there
that.
B
A
C
An
entire
so,
as
I
mentioned
this,
this
currently
I
mean
the
in
the
depth,
usually
for
Azure
customers.
The
way
they
deal
with
I
mean
they
work
it
out
as
a
little
keyboard
to
AKs
portal
and
it
becomes
easier
for
them
to
discover
the
backup
experience,
but
the
whole
management
part
of
backup
still
relies
on
backup
center,
where
you
have
the
entire
at
scale
experience,
so
you
can
take
a
look
at
it
and
see
all
the
Clusters
that
are
protected
or
not
protected.
So
you
can
take
a
look
at
that.
C
Okay,
that
view
is
also
available
within
backup
center
and
along
with
it,
any
backups
that
you
have
created
for
any
of
the
cluster
within
your
subscription.
You
can
take
a
look
at
it,
so
that
multi-cluster
management
scenario
is
also
available,
so
you
have
the
view
of
entire
set
of
clusters
that
you
have
within
a
subscription
via
backup
center,
so
that
is
available
and
then
in
business
continuity
center
that
is
currently
in
preview.
We
are
going
to
bring
out
newer
elements
also,
along
with
regular.
C
C
There
it's
part
of
our
backup
center
experience,
okay,
awesome
and
and
the
the
business
continuity
Center
is
the
one
where
you
were
saying
that
you
could
bring
in
third
party
applications
to
provide.
You
know
similar,
better
Services
that
that
we
can
plug
into
as
well
right.
I
think
that's
yeah,
yeah,
so
business
continue.
Center
is
a
fundamental
idea
that
we
are
bringing
is
that
the
customers
can
have
different
different
choices
to
predict
their
workloads
within
Azure.
C
It
could
be
VM,
SQL
databases,
anything
that
could
be
there
potentially
and
the
idea
behind
is
that
that
we
will
provide
a
single
management
layer,
basically
or
an
experience
layer
on
top
of
it,
where
customers
can
go
and
see
protected,
instances,
workloads
and
all
those
details,
and
then
on
the
back
end,
they
can
continue
using
the
in
what
basically
any
third-party
solution
that
they
want
to
use
to
predict
their
clusters,
so
that
capability
will
come
in
via
the
ABC
Center
So.
C
C
C
Do
we
have
to
worry
about
any
conflicts?
If
both
are
you
know,
similar
format
and
available
through
I
think
your
extension
will
just
be
natively
available.
Ours
would
be
available
through
Marketplace,
so
just
trying
to
understand
if
there
are
conflicts
that
we
need
to
work
to
resolve
actually
I
think
in
context
with
that
extension.
It
follows
the
framework
that
test
highlighting
right.
D
So
extensions
we
have
a
internal
team
which
provides
us
an
extension
platform
of
sorts.
We
don't
directly
rely
on
kubernetes,
specific
extension
Frameworks
that
they
have
I
am
not
fully
aware
of
them,
but
as
far
as
I
know,
we
don't
rely
on
them.
As
such,
we
have
our
own
infra
inside
AKs,
which
manages
our
extension
or
the
life
cycle
of
the
extension.
C
D
C
Terms
of
availability,
so
just
to
highlight
that
that
recently
AKs
has
come
out
with
extensions
and
applications
in
a
Marketplace,
so
backup
would
essentially
would
be
available
to
the
customers
via
that
landscape.
Also.
So,
as
you
see
from
The,
Experience
backup
is
available
within
the
AKs
portal,
as
a
first
party
experience
for
it,
but
we
will
also
make
it
available
within
the
marketplace
for
them.
C
So
as
as
part
of
our
Discovery
element
of
it
so
yeah,
we
are
trying
to
cover
all
the
places
that
are
available
to
reach
out
to
our
customers
and
desktop
backup
solution.
Okay
and
you're.
Currently
just
doing
CSI
snapshots
or
non-csi
snapshots
like
for
Azure
files
and
stuff
are
not
covered
at
this
point
correct
and
that
would
be
yeah.
A
C
This
point
we
are
not
moving
towards
us.
Different
I
mean
we're
deliberately
not
moving
towards
a
different
solution
to
support
AFS
days,
persistent
volumes
and
focusing
more
on
displays,
because
CSI
is
the
standard
and
you
want
to
follow
through
with
it.
We
are
currently
working
with
internal
teams
to
get
this
enabled
down
the
line
so
that
we
can
bring
out
support
for
AFS
or
any
future
CSI
driver
that
Azure
comes
out
with,
so
that
it
comes
with
that
inbuilt
snapshotting
and
restoring
capability
so
that
backup
for
it
can
directly
latch
onto
it.
C
C
Right
now
hitting
a
pause
on
bringing
the
support
for
AFS
or
file
share
for
time
being,
as
we
are
working
with
the
internal
teams
to
get
that
particular
functionality
being
enabled
that
that's
the
whole
reason
for
it.
Awesome
and
I
saw
a
fairly
elaborate
roadmap.
Thank
you
for
sharing
that
rajat.
How
much
of
this
is
you
know,
roadmap
on
the
AKs,
backup,
side
and
and
I
saw
things
like
bring
your
own
encryption,
keys
and
and
stuff?
How
much
of
this
will
will
get
upstreamed
for
for
Valero
itself,
and
so.
A
C
Set
of
elements
that
I
brought
out
like
bring
your
keys
and
all
those
are
inbuilt
form
supports
so
Azure,
as
I
mentioned
Azure.
Backup
in
itself
is
a
platform
all
together
on
which
we
enable
our
different
different
Azure
workloads,
so
they
have
their
own
backup
capability
that
we
comes
in,
so
we
utilize
Valero's
extension
to
enable
backup
for
AKs,
but
the
things
like
the
ransomware
protection
and
all
those
are
something
that
are
inbuilt,
Within,
These,
Are
Back
baked
in
within
the
Azure
backup
platform.
So
the
extension
doesn't
provide
those
capabilities.
C
It's
the
backup
platform
provides
that
capability.
So
there
is
a
separation
between
those
two,
at
least
for
time
being,
but
within
us
keep
on
working
on
to
see
what
we
can
bring
into
the
community
also
on
this
parts.
So
we
can
keep
that
consideration.
Also
always
in
our
mind
and
bring
your
own
keys
in
that
context,
is
really
bring
your
own
keys
for
encryption
at
rest.
Not
encryption
in
transit
is
that
yeah.
C
Okay:
okay,
thank
you
that
that
clarifies
the
things
and
I
I
saw
a
bunch
of
questions
on
Aro
from
there
are
multiple
flavors
within
Azure
that
we
will
for
kubernetes
that
we
will
start
tackling
in
time
being
once
we
have
our
AK
solution
out,
so
things
like
Arrow
or
arcbs
kubernetes
clusters,
or
something
that
are
down
the
line
are
Target
to
get
them
enabled
for
our
customers.
D
A
Was
wondering,
are
you
using
like
an
official
Valero
released
images
here,
or
do
you
rebuild
or
Fork
on
your
own
side.
D
We
rebuild
them,
we
don't
use
it
directly.
Yeah.
A
Yeah
and
do
you
have
any
local
changes
or
are
you
rebuilding,
basically
a
full
release,
specific
version?
We.
D
B
D
E
I
have
one
quick
question,
so
is
this
only
snapshot
backup
or
you
can
also
select
file
system?
Backup,
if
you
want
I
mean
I,
I
didn't
see
a
problem.
Yesterday,
okay,
yeah.
C
Well,
so
that's
what
we're
talking
to
Satya
about
this
So
currently
right
now,
our
backup
solution,
for,
if
you
have
an
Azure
disk,
this
EVS,
we
can
take
a
snapshot
of
it.
If
a
customer
comes
with
a
filed
share
based
PV,
we
are
not
correctly
enabling
that
CSI
driver
snapshotting
capability,
because
that
driver
doesn't
have
restore
capability.
Also,
okay,
so
it
can
create
a
snapshot
of
that
afspp
where
it
cannot
create
a
restore
for
it
cannot
restore
backend
AFS
from
that
snapshots
So.
E
C
E
C
C
Different
questions
here,
he
was
just
asking
if
you
enable
restock
or
copia
no.
C
B
B
Hey
rajat
great
demo,.
B
I,
just
had
one
quick
question:
I
think
I'm
sure
answered
it,
but
still
so
you
did
a
cross
cluster
backup
right
for
snapshots.
Css
snapshots.
Yes,
yes,
so
did
both
the
Clusters
have
access
to
the
storage
underlying
storage
or
like.
D
Just
Stitch
all
of
those
flows
on
the
portal,
all
the
permission,
flows
kind
of
happen
like
a
magic.
We
take
care
of
them
automatically
when
the
customer
goes
for
a
research,
because.
D
C
D
We
don't
exactly
stop
it
right
now,
but
I
mean
we
will
plan
to
lock
it
in
the
future,
but
the
expectation
it
has
to
flow
from
our
backend.
It
has
to
come
through
rest
API
portal,
Powershell
CLI,
whatever
they
prefer.
If
they
take
any
backups
from
inside
the
cluster,
it's
it
does
not
reflect
back
to
our
back
end.
So
it's
it's
like
not
a
not
a
use
scenario.
We
support
okay,
at
least
for
now,.
D
B
Inside
so
you
might
have
two
well
arrows
in
the
same
AKs
cluster,
but
you
guys
won't
be
aware
about
the
other
Valero.
D
Instruments
yeah,
we
don't
I
mean
as
long
as
they
share
the
objects
or
it's
up
to
them,
but
I
mean
we
only
care
about
what
is
done
in
context
of
our
backup.
If
a
backup
flows
to
our
API,
we
take
care
of
we
map
that
will
ever
go
back
up
to
our
backup.
We
don't
care
about
anything
else
that
customer
might
be
doing
there.
D
You're
right,
so
we
do
recommend
in
that
new
users
do
not
have
a
existing
wheel,
because
we
don't
really
guarantee
the
existing
installation
to
work.
We
kind
of
call
that
out
there's
nothing
else,
pretty
much.
You
can
do
here.
Yeah.
C
E
Just
a
quick
comment
about
the
craft
cluster
restore
and
this
kind
of
Builds
on
what
shubham
was
saying.
We
also
at
cloudcast
have
been
working
on
similar
class
Cloud
restore
so
I
just
wanted
to
touch
base
on
this
I
guess
what
you're
doing
is
basically
creating
the
backup,
storage
location
to
the
cluster
B,
the
second
cluster
and
let
Valero
think
all
the
backups
and
basically
all
the
recovery
points
and
then
proceed
with
the
restore
right.
That's
what
you're
doing?
E
D
E
D
C
You
yeah,
the
keyboard
just
allows
you
to
run
VMS
inside
kubernetes
and
it's
to
essentially
create
one
API
platform
for
managing
both
kubernetes
clusters,
as
well
as
VMS,
and
there
is
a
plug-in
to
support.
Essentially
each
VM
runs
as
a
pod
and
it
can
have
its
own
disks
again
within
a
cluster
and
and
I.
Think
tiger
is
just
asking.
How
do
we
support
a
configuration
like
that,
where
you
may
have
VMS
running
inside
large
kubernetes
clusters.
D
C
C
We
haven't
brought
in
support
for
this.
Okay,
as
in
case
from
our
customer
standpoint
also
to
highlight
we
haven't
gotten
a
requirement
from
them
that
they
are
looking
to
get
such
level
of
support
within
the
AKs
backup
solution,
but
down
the
line.
C
If
we
see
as
again
potential
to
see
if
we
want
to
enable
such
kind
of
a
backup
capability,
also
where
a
keyword
plugin
is
there
and
VMS
are
created
in
their
clusters,
then
we
will
think
about
it
to
bring
that
support
also
in
time,
but
at
least
right
now,
it's
not
there.
As
as
I
said,
the
customer
demand
is
not
there
for
us
to
bring
that
kind
of
solution,
at
least
for
time.
B
C
Good
we're
on
the
same
boat,
we're
we're
actually
exploring
support
for
keyword
on
on
our
end
and
it
it's
turning
out
to
be
more
work
than
we
thought,
but
we're
we're
on
the
where
it
it's
we're
working
through
it
on
our
end,
so
we
completely
understand
the
complexity
around
supporting
it
in
I.
Think
it's
being
pushed
a
bit
more
by
red
hat
and
Souza
than
than
others
at
this
point,
but
yeah
anyway.
C
That's
that's
all
we
know
about
it
to
be
to
be
fair,
I
didn't
see
any
other
questions
and
I'm
all
shut
up.
A
Okay
and
then
we
had
some
topics
from
Magoo.
Let's
see
here,
oops
sorry,
there
we
go
yeah.
E
Yeah
right,
my
question
is
I
mean
again:
I
was
helping
a
user
on
the
slack,
and
this
came
up
in
that
context
and
and
I
think
somebody
opened
this
PR.
So
this
shouldn't
this
change
also
be
made
in
1.11
and
probably
even
before,
because
if
somebody
goes
to
the
docs
and
let's
say
clicks
on
1.11,
they
wouldn't
see
this
change
right.
Oh.
A
B
A
The
dock
was
added
on
on
Main,
so
we
got
it
in
111.,
so
the
question
was:
should
we
backport
this
to
the
and
because
this
also
affects,
what's
on
the
the
the
doc
site
we
I
I
would
say?
Probably
we
should
make
a
PR
to
extend
this
to
any
specific
versions.
A
A
So
so
I
think
I
think
we
do
need
to
make
sure
the
docs
are
consistent.
There
I
think
that's
a
good
point.
E
All
right,
okay,
the
next
one
yeah.
C
E
E
If
you
scroll
down
you,
you
can
see
a
comment:
oh
yeah
click
on
the
link,
so
we
we
didn't
completely
understand
so
I
wanted
to
ask
you
something
so
in
the
current
implementation,
when
I
set
the
timeout
right-
and
it
is
the
CSA
plugin
that
constantly
checks,
if
the
snapshot
moved
to
ready
to
huge
State
and
that's
the
only
place
right
as
far
as
I
know
the
main
Valero
binary
and
the
application,
it
doesn't
really
check
anything
about
the
CSI
status
right.
Is
that
correct.
B
E
E
Mean
that's
what
this
comment
is
saying
right,
he's
saying:
that's
why
there
are
two
places
that
have
the
polling
logic
somewhere
in
the
middle.
You
see
that
sentence.
There
are
no
two
places
with
the
polling
logic.
As
far
as
I
can
tell.
D
There
are
actually
two
places
where
the
polling
happens:
one
is
in
the
plugin
and
one
is
after
the
backup
is
done
when
they
upload
the
volume
snapshot,
yeah
jsons,
it
happens
in
two
places
and
I
was
all.
We
have
already
been
tracking
this-
that
this
polling
is
kind
of
duplicate
right
now,
right
so
I
think
I
I
will
discuss
with
Jiang
once
more.
It
seems
like
the
second
set
of
polling
right,
which
happens.
D
Post
backup
can
be
removed
pretty
much
because
it
did
not
seem
to
add
any
specific
value
if
you
are
anyways
waiting
for
10
minutes
and
going
out
and
there's
another
set
of
changes
that
we
are
already
bringing
in
in
the
plug-in
where,
in
case
of
failures,
we
clean
the
snap
shot
up
front
like
if
a
failure
happens
in
the
CSI
plugin
will
clean
up
the
snapshot
and
exit,
so
that
polling
should
not
even
ideally
occur
in
the
the
second
time
when
it
happens,
it's
it's
in
the
core.
B
E
Right
so
this
this
issue
is
only
about
they're
ready
to
use
flag,
I
mean
when
the
snapshot
is
ready
right
so
that
second
polling
doesn't
apply
here.
As
far
as
I
know,
right
that,
but.
B
What
sun
is
trying
to
say
is
this
parameter
that
you
are
using
right?
That
is
the
timeout
parameter
that
is
used
in
that
polling.
What's
up
value.
E
B
E
The
as
the
name
suggests
right,
it
says:
CSI
snapshot
timeout
and
that
should
get
used
in
the
plugin
tomorrow.
If
we
decide
that
the
second
polling,
whether
it's
there
or
not,
should
use
a
second
parameter
seems
to
be
to
be
a
separate
issue.
In
any
case,
can
you
guys
look
at
this
comment
and
maybe
you
can
post
and
we
can
discuss
it.
E
D
Yeah
I
think
I
had
a
in-depth
discussion
with
him
around
this,
because
we
were
anyways
tackling
the
performance
issue
in
CSI
around
Bia
and
two
sets
of
pollings.
So
what
I'm
planning
to
propose
to
the
Angus
that
the
second
set
of
polling?
We
try
to
remove
it
completely
if
possible,
and
this
timeouts,
this
timeout
specifically
right
I
feel
that
it
should
be
honored
in
the
plugin,
because
end
of
the
day,
that
polling
will
remain
no
matter.
What
that
is
what.
B
D
D
B
A
B
E
Okay,
okay,
the
other
things
are
just
quick
question
so
again
in
the
context
of
this
issue.
For
example,
I
know
that
David
is
going
to
work
on
this.
Who
can
who
can
assign
this
I'm
sure
some
permissions
are
needed?
How
does
this
thing
work
currently
like.
A
Okay,
so
so
the
issues
time,
it's
not
really
I
mean
that's
been
kind
of.
You
know
a
case
by
case
I
mean
certainly
for
maintainers.
People
tend
to
self-assign
things,
they're
working
on
I.
Think
if
there's
an
issue
that
someone
in
the
community
is
going
to
work
on
I
think,
basically,
you
know,
I've
talked
to
a
maintainer
and-
and
you
know,
get
it
get
them
to
sign
it
to.
You
is
probably
the
best
course
of
action
here.
A
You
know.
If
there's
some
question
about
who
should
work
on
it,
then
we
can
have
the
discussion.
Usually
usually,
if
someone
wants
to
work
on
it.
That's
that's
fine.
You
know
they're
going
to
work
on
it.
Unless
someone
else
is
already,
you
know
planning
on
doing
it
for
something
they
need,
which
is
often
not
the
case.
So
there's
not
really
a
huge
formal
process
around
it.
I
guess
that's
what
I'm
saying.
A
Valero
users
is
more
for
user
questions,
so
I'm
using
Valero
I'm
having
this
error,
you
know
help
me
out
here.
Bolero
Dev
is
the
place
where
we
tend
to
have
those
maintainer
level
discussions
and
Community
discussions
around.
You
know
feature
development
and
Bug
fixes.
E
Sounds
good
and
the
next
question
again
all
came
in
context
of
talking
to
some
people
on
the
little
users
like
so
I
just
want
to
clarify
the
current
semantics
around
this.
So
if
I
set
both
these
snapshot,
volumes
like
to
true
and
default
file
system,
backup
to
True,
which
you
can
certainly
do
right-
these
are
just
completely
independent
Flags.
There's
no
validation.
Now
in
the
code
right
right.
What
is
the
expectation
there
so.
A
So
those
really
control
different
things
snapshot
volumes
says
basically,
if
Valero
would
would
otherwise
and
and
Siobhan
correct
me
if
I'm
getting
any
of
this
wrong
as
I
understand
snapshot,
volume
says:
okay,
if
Valero
was
otherwise
going
to
snapshot
a
volume
setting
it
to
false
tells
Valero.
Don't
do
that
so
that
doesn't
control.
A
You
know,
backup
method
that
just
says:
hey
Skip
the
snapshotting,
because
we're
not
doing
that
for
this
backup
the
default
to
file
system,
backup
that
has
more
to
do
with
the
the
decision,
because,
basically,
anytime
Valero
has
a
volume
to
back
up.
It's
going
to
take
one
of
two
options:
it's
either
going
to
try
to
do
a
snapshot,
or
it's
going
to
try
to
do
a
file
system,
backup
a
file
system
backup
can
either
be
rested
for
copy,
depending
on
how
you've
configured
it
a
snapshot.
A
Backup
is
either
going
to
be
a
CSI
Snapchat
or
a
native
Snapchat
again,
depending
on
how
you've
configured
Valero.
So
if
you
have
a
combination
of
volumes,
say,
for
example,
you're
normally
using
CSI,
but
you
also
have
some
NFS
volumes
that
you
obviously
can't
use
CSI
for,
if
you
have
mostly
CSI
volumes,
you'd
probably
want
to
set
the
default
to
FS
back
up
to
false
and
then
for
those
volumes.
You
can't
use
CSI
snapshots.
A
You
would
set
a
pod
annotation
on
those
to
say,
use
Snapshot
and
use
rest
to
copy
instead
flip
that
around.
If,
if
most
of
your
volumes
are
ones,
you
want
to
use
rustic
for,
then
you
would
set
the
default
to
FS
backup
to
true.
And
then,
if
you
had
a
volume
that
you
wanted
to
skip
rustic
because
you
you
could
use
CSI
for
or
native
snapshots
for,
then
you
would
set
a
pod
annotation
for
that
volume
to
say,
don't
use
rustic
here.
E
A
I,
don't
think
so,
because
the
reason
is
that
the
the
default
to
have
as
backup
doesn't
say
whether
or
not
we're
backing
something
up.
It
doesn't
say
whether
or
not
we're
Snapchatting
it
just
tells
Valero.
This
is
the
default
mode
and
we
change
it
on
a
per
volume
basis
to
the
other
mode.
So
in
other
words,
that's
used
for
that's
used
for
deciding
you
know
which
backup
method
to
use
the
snapshot.
Volumes
selection
is
basically
turning
off
snapshot,
selection,
completely
or
snapshot
backup
completely.
A
A
If
you
don't
want
any
volume
backup,
then
you
then
you
would
probably
set
snapshot
volumes
to
false
and
you
would
not
set
default
volume
step
as
backup,
because
that
way,
no
no
PV!
You
know
data
gets
backed
up
right.
E
A
You
may
want
to
turn
off
snapshotting,
but
still
use
rustic
or
copia
for
selected
volumes.
You
know,
or
things
like
that,
so
so
I
don't
know.
If
there's
any
backup
level
validation.
That
really
makes
sense
to
say
you
know
some
combination
of
these
two
billions
is,
you
know
nonsensical,
because
since
they
mesh
since
they
control
two
different
kinds
of
behavior,
they're,
probably
edge
cases
for
both.
You
know
for
any
combination.
A
Sounds
like
a
bug
to
be
honest,
and
have
you
seen
this
as
far
as
I
understand,
Valero
will
not
and
from
what
I've
seen.
This
is
true,
but
again
I'm
not
saying
this
is
not
a
bug
here,
but
Valero
should
be
only
doing
rustic
if
there's
not
a
snapshot,
backup
for
a
particular
volume.
B
A
E
A
Yeah,
if
you're,
seeing
both
happening,
that
that's
either
a
bug
or
it
may
be
that
the
logs
are
confusing
and
not
describing
properly
what's
happening.
E
E
A
I,
don't
really
have
a
lot
of
background
on
that
particular
issue.
I
I've
noticed
this
myself
and
I
mean
in
terms
of
in
practice,
I'm
not
really
sure
I.
E
A
Yeah
I
I
that
might
be
worth
having
a
opening
issue
and
then
having
a
discussion
around
it
and
I
think
this
is
one
of
these
things
that
you
know
there's
nobody
currently
working
on
Valero.
That
was
on
the
project
when
that
decision
was
made.
So
I
actually
don't
have
any
insight
on
this,
because
this
is.
This
is
behavior
that
predates
my
involvement
in
the
project
and
since
I
think
I'm,
probably
the
longest
longest
standing
person
involved
in
this
project
regularly.
A
At
this
point,
pretty
much
nobody
currently
working
on
Valero
was
was
working
on
Valero.
When
that
decision
was
made,
so
it's
possible
to
revisit
it.
We
can
talk
about
pros
and
cons,
see
what
makes
sense
you
know.
Should
it
be
configurable
should
the
finance
is
but
I
would
say,
open
initiative
discuss
it.
A
And
that's
all
good:
does
anybody
else
have
any
issues
to
bring
up.
A
Yeah,
it
sounds
like
no
we're
almost
we're
running
running
low
on
time
anyway.
So
I
think
it
worked
out
well,
so
that
will
be
it
for
now
next
week
we'll
have
the
meeting
time.
A
8
PM
us
Eastern,
8
AM
Beijing,
or
is
it
so
yeah?
That's
right
and
this
time
in
two
weeks
so
have
a
great
week,
and
you
know
ping
us
on
slack.
If
anything
else
comes
up
or
create
an
issue.