►
From YouTube: Velero Community Meeting/Open Discussion - June 29, 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
Hello,
everyone
and
welcome
to
the
valerie
community
meeting
a
slash,
open
discussion
today
is
june.
29Th
2021.
Please
add
yourself
to
the
attended
list.
With
your
affiliation
listed
as
well
we're
going
to
dive
into
some
status
updates.
Then
we
have
discussion
topics
and
some
shout
outs.
So,
first
off
we're
gonna
roll
get
started
with
dave.
B
Yeah,
it
looks
like
we
have
a
lot
of
a
lot
of
status
this
week,
so
yeah,
I'm
still
working
on
the
upload
progress
beijing
team
bring
up,
I've
been
working
on
some
poc
stuff
on
the
astrolabe
side,
moving
discs
and
things
and
I'll
get
back
to
doing
your
code
reviews
this
week.
A
Awesome
any
any
help
needed
with
any
of
the
the
code
reviews.
B
No
we'll
set
things
up
I
think
carlesia
is
gonna,
is
still
around
to
help
me
on
some
of
those
things
as
well.
So
let's
roll
them
forward.
A
All
right
that
was
that
was
super
quick.
So,
let's
dive
into
some
discussion
topics,
dylan.
C
Hey
so
we
have
been
kind
of
digging
into
the
blair
winduin
tests
and
trying
to
figure
out.
Well,
we
have
a
set
of
requirements
for
testing
components
on
our
end
that
we
would
like
to
kind
of
re-leverage
the
valero
end-to-end
testing
framework,
since
valero
is
kind
of
part
of
our
stack
and
digging
into
how
the
current
valero
unwind
tests
are
written.
C
There's
an
open
pr
up
from
carlesia
that
is
refactors
them
a
bit
and
one
of
the
changes
that
she's
making
in
that
pr
is
to
remove
the
ability
to
even
specify
a
single
namespace
to
install
valeroin,
which
is
fine,
but
it
actually
kind
of
isn't
even
the
core
issue
of
what
we
need
changed.
We
were
trying
to
figure
out
like
why
can't
we
supply
to
the
endun
test,
an
instance
of
valero,
to
run
the
test
suite
against
right.
You
can't.
C
Having
the
end-to-end
test
actually
run
through
the
install
you
can,
we
would
like
to
just
be
able
to
kind
of.
B
B
If
so,
we
haven't
merged
felicia's
change,
it
hasn't
gone
through
yet
so
I
don't
know
that
the
I'm
not
so.
I
need
to
go
through
it
with
her
so
like
there's,
no
reason
to
have
a
different
name
space
for
each
for
valero
to
get
installed
in
you
know
and
have
that
change
every
time
she's.
I
think
she's
fixing
a
problem
that
doesn't
really
exist.
C
B
So
there's
a
flag:
minus
minus
install
valero
equals
false,
okay,
which
you
can
pass
to
the
ginkgo
test,
and
that
says,
don't
install
and
don't
uninstall
yeah,
and
that's
specifically
there
so
that
you
don't
you
know
so
that
you
can
work
against
running
valero
and
that's
a
key
part
of
the
design
for
the
test,
because
you
know,
if
you
have
your
own
setup,
that
you
want
to
run
the
test
against,
you,
don't
want
it
like
over
overriding
it
or
whacking
it.
Something
else
right,
right:
okay,.
E
B
C
It
should
be
yeah
yeah,
so
that
helps
because
we
kind
of
did
feel
like
that
was
a
bad
design
decision,
mainly
because
you
know
in
a
test.
We
shouldn't
necessarily
rely
on
the
install
bits.
E
C
So,
okay,
that's
good
to
know
so
so,
if
we're
on
the
same
page,
there
we
can
depend
upon
the
install
valero
flag
still
being
there
and
we
can
set
it
to
false
and
kind
of
what
we're
thinking
is
it
we
were
looking
at
when
I
was
digging
through
the
endowment
test,
a
lot
of
them
kind
of
seem
provider
specific,
which
is
okay,
because
there's
obviously
a
lot
of
provider,
specific
kind
of
education
that
you
want
to
test
with
calero
parts.
C
Tests
so
right
right,
right,
right,
yeah,
I
didn't
mean
to
imply
it's
like
thousands,
so
what
we
were
hoping
to
do
is
because
you
know
there's
a
lot
of
tests
that
are
basically,
I
guess
what
you
call
conformance
testing.
So
we
would
like
to
add
testing
to
say,
like
hey
here,
here's
what
an
a
resource
looks
like
pre-backup.
We
run
the
package
run.
The
disaster
stay
out,
run
the
restore
yeah
just
does
the
resource
match
the
expected
restore?
C
Yes,
that
sounds
really
great
okay,
so
we
have
a
lot
of
that
that
you
know
belongs
upstream,
because
it
would
be
testing
native
kubernetes
resources
and
then
we
also
have
a
set
of
tests
that
would
need
to
test
kind
of
the
openshift
specific
edge
cases
here.
That
probably
doesn't
necessarily
belong
in
the
valero
repo.
But
if
you,
I
guess,
I'm
just
kind
of
bringing
salt
to
say
that
we're
going
to
start
demoning
some
resources
to
writing
those
tests.
C
So
if
we're
on
the
same
page
that
we
can
keep
this
installed
valero
flag
for
the
e2e
test,
then
we
will
probably
have
a
few
resources
devoted
to
doing
that
and
contributing
that
back
up
stream.
So
that's
good
to
know.
I
just
want
to
make
sure
that
we're
on
the
same
page
in
terms
of
that
install
flag.
So
I
can
add
a
comment
on
that
pr
as
well.
C
Just
mentioning
you
know,
we
we
need
that
that
flag
to
still
be
there,
and
so
that
would
kind
of
be
a
breaker
for
us
right
and
then,
if
we're
on
the
same
page,
there
then
I
think
we're
in
good
shape
to
kind
of
move
forward.
In
our
end.
B
Yeah,
I
think
that
that
would
be
excellent.
Thank
you
very
much.
I
mean
we
would
really
appreciate
that
and
welcome
that
yeah.
As
far
as
the
cloud
specific
nature,
I
guess
there's
two
pieces
to
it
I
mean
one:
is
you
have
to
have
an
object,
store
plug-in
to
do
a
backup,
and
so
you
know
we
have
to
set
which
which
object
store.
B
It
is
you
know,
aws
is
kind
of
the
default,
for
you
know,
handling
minio
or
what
have
you
and
then
the
the
kabishi
tests
allocate
persistent
volumes,
and
so
they
need
a
storage
class
and
default.
Didn't
always
cut
it.
So,
but
if
you
don't
run
those,
then
there's
no
reason
to
specify.
C
C
It
is,
I
mean
it's
not
going
to
work
right
right
right
and
I
understand
that,
but
I
mean
I
think,
that's
kind
of
gets
into
the
testing
thing
where,
like
it
really
shouldn't
crash.
In
that
case
right
I
mean
it
should
be
supplying
some
status
info
back
to
the
user
to
say,
like
hey,
we're
missing
some
set
of
configuration,
so
that
would
be
much
nicer.
Yes,
yeah
yeah,
so
I'm
kind
of
hoping
as
we
go
through
this
process.
C
We
can
harden
things
like
that,
because
that's
anything
that
would
break
the
the
pattern
of
an
endowment
test
suite
that
can
run
against
some.
You
know
some
single
instance
of
valero.
That's
kind
of
what
I
noticed
is
that
it
seems
like
the
whole.
Install
and
install
pattern
is
to
prevent
misconfiguration
from
crashing
the
server
right.
We
would
kind
of
want
to
like
tease
what
all
those
things
are
and
then
start
fixing
it
right.
B
Well,
it's
it's
one
is
the
to
make
sure
that
it's
in
configured
correctly,
which
you
know
it's
a
it's
a
pain
to
like
you,
know,
go
through
and
do
all
the
the
configuration
by
hand
every
time.
So
that's
one
of
the
reasons.
The
other
is
so
that
we
don't
have
any
state
left
over
that
carries
between
tests.
Sure
and
you
know
one
it
might
break
the
next
test
or
worse
tests
start
becoming
dependent
on
each
other
and
because
test
a
did
something
test
b
passes.
B
So
we
want
to
avoid
that
situation,
but
you
know
there's
certainly
room
to
be
flexible.
Here
I
mean
it's,
you
know
it's.
It's
pretty
early,
really
pretty
early
stages
for
this
stuff,
so
we
can
think
about
different
ways
and
definitely
having
valero
come
up
and
report
status.
Like
hey,
I'm.
B
You
know
I
started,
but
this
configuration
is
missing.
Therefore
it
won't
work,
that's
a
much
better
thing
than
simply
crashing,
but
you
know
crashing
says
that,
well,
you
got
to
fix
it
out,
go
figure
it
out,
so
it's
better
than
leaving
it
there
non-functional
and
no
status.
B
So
we've
already
got
that
issue
like
you
know
like
even
like
the
the
plug-ins
come
up
and
you
give
them
bad
credentials,
they
don't
check
the
credentials
as
part
of
startup,
so
you
don't
find
out
that
the
credentials
are
bad
until
you
actually
run
your
first
backup
there's
a
bunch
of
places
we
can
improve.
Oh.
C
C
Okay,
all
right
well,
great
thanks.
I
just
wanted
to
make
sure
that
we're
on
the
same
page
with
this
install
flag
and
then
I
will
probably
have
I'll
leave
a
comment
on
carly's
pr.
Just
mentioning
that
we
kind
of
want
to
keep
that
in
and.
B
Yeah,
let's
coordinate
on
test
writing
because,
as
the
beijing
team
spins
up
they're
going
to
be
concentrating
initially
on
things
like
automating
the
test
test
execution
and
also
bringing
up
some
tests.
So
if
we
can
collaborate
a
bit
make
sure
you
know,
maybe
we
get
like
issues
put
in
for
tests
that
you're
planning
to
write
and
they
can
put
in
tests.
And
you
know
we've
got
the
community
meeting,
that's
more
asia-friendly.
If
we
could
have
somebody
pop
in
there
or
we
can
do
it
by
slack
or
whatever.
B
But
let's
let's,
let's,
let's
see
if
we
can
get
kind
of
a
coherent
set
of
tests
that
are.
C
Being
written
yeah
that
sounds
great
I'll,
make
sure
that
whatever
we
do
end
up
starting
on
we'll
we'll
make
sure
there's
issues
open
around
it,
and
then
we
can
get
someone
to
join
the
the
asia
friendly
meeting
and
you
know,
make
sure
we're
keeping
you
guys
updated
on
that.
C
Before
I
kind
of
wrap
up
anyone
from
our
team,
that's
on
the
call
did
I
miss
anything
that
we
needed
out
of
this.
It
sounded
like
that
was
kind
of
the
big
concern
was
the
install
flag.
So
if
we
can
keep
that,
I
think
we
have
enough
to
move
forward
and
start
kind
of
making
sure
we
get
the
requirements
in
our
end
and
yeah.
E
I
mean,
I
think,
the
only
thing
we
talked
about,
which
I
think
is
less
important
now
that
we
can
we're
preserving
the
ability
to
run
against
an
existing
install
is
that
it
would
be
nice
to
be
able
to
do
the
in-end
test
where
the
test
installs
and
have
it
succeed.
If
the
cluster
happens
to
be
an
open
shift
cluster-
and
I
know
there's
the
rustic
permission
issue
that
we
may
need
to
fix
around
that,
but
that
seems
less
important
than
the
other.
I
think
at
this
point.
E
Yeah
and
and
the
issue
we
hit
with
using
an
openshift
cluster,
there
is
just
that
restick
has
to
you,
know
their
permissions
issues
with
the
rustic
damon
set
in
an
openshift
environment,
where
we
have
to
set
an
sec
around
that
and
be
nice
to
be
able
to
get
the
install
to
be
able
to
do
that.
If
the
cluster
happens
to
be
an
open
shift
cluster,
but.
B
Yeah
I
mean:
would
that
be
something
that
makes
sense
to
just
have
it
automatically?
Can
we
automatically
detect
an
openshift
cluster
and
have
install
always
do
this,
because
that
would
seems
reasonable
to
me.
E
That
that
might
be
the
way
to
go
there,
because
that
would
that
then
we
wouldn't
have
to
add
this.
You
know
extra
cluster
type
or
whatever
I
mean
an
open
shift.
Cluster
in
one
sense
is
kind
of
like
a
kind
cluster
for
install
where
you'd
have
to
tell
it.
You
know,
use
aws
for
storage,
it's
not
an
aws
cluster,
but
you
know
you
need
to
use
aws
for
your
storage
or.
B
So
anything
we
can
do
that
makes
that
better.
If
we
can
come
up
with
you
know,
even
if
it's
even
if
it's
like
some
special
casing
there
in
the
installer,
I'm
not
really
against
that,
we
probably
want
to
make
sure
the
helm
chart
does
the
same
thing
too,
but
yeah
I
mean,
let's
figure
out
a
way
to
make
this
work,
because
I
think
it
makes
a
lot
more
sense
to
just
have
it.
Valero
install
should
be
figuring
out
more
and
more
stuff
we
should
have.
B
D
B
Yeah
and
we've
run
into
some
issues.
You
know
where
it's
like.
You
know
the
oh
actually
like
rustic
pads,
you
know
so
where,
where
host
path
volumes,
I
think
it's
like
where
they
get
mounted.
That's
actually
changed
between
different
kubernetes
installs,
so,
like
our
tkgi
version
of
kubernetes,
puts
the
host
path
somewhere
different
and
that
has
to
get
fixed
or
valero
needs
to
get
set
properly
to
handle
that
so.
A
Just
a
a
quick
question
on
conformance
because
I
I
heard
you
talk
about
conformance
here
as
well-
have
has
anyone
looked
into
sonaboy
for
conformance
testing?
Is
that
relevant
here
at
all.
B
We
haven't
looked
at
it,
yet
it
might
be
worthwhile.
I
think
bridget's
kind
of
an
expert
on
sonoboy,
so
we.
B
Her
on
that,
when
she's
back,
I
got
a
ping
from
somebody.
I
can't
remember
exactly
what
it
was
the
other
day,
but
they
were
asking
for
integration,
I
think
between
valero
and
santa
boy
I
wasn't
quite
sure
what
they
were.
Thinking
of.
I
have
to
follow
up
with
them,
but
maybe
that's
like
part
of
the
restore
process
is
making
sure
the
cluster
is
healthy.
A
Yeah,
so
they
they
have
these
plugins,
where
you
can
add
in
more
things
that
sonaboy
would
check
for
so
they
they
check
for
the
standard
like
4
000,
kubernetes
end-to-end
tests,
but
then
you
can
add
other
things
that
they
could
check.
For
so
an
easy
way
would
be
like
to
add
a
valero
plug-in
here,
so
we
can
run
a
sony
boy
check
and
say
yes,
this
valero
setup
is
conformant.
A
All
right
that
conversation
wrapped
up
any
other
questions
comments
around
this.
A
All
right,
then
we,
if
we
don't
have
any
other
discussion
topics.
We
only
have
one
thing
left
to
do
today,
and
that
is
the
shout
outs.
So
the
shout
outs
today
go
to
genting
for
bumping
the
version
of
the
helm
chart
up
to
version
161.
So
thank
you
for
that
and
that's
it
for
today.
Thank
you.
Everyone
for
joining
thanks
for
the
discussions
have
a
fantastic
rest
of
the
week.
Everyone
and
I'll
tell
you
soon.