►
From YouTube: Velero Community Meeting - July 17, 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
A
So
with
that
said
today,
please
I've
pasted
the
link
to
the
hackmd
for
the
community
meeting
notes
into
the
chat.
Please
add
yourself
there
and
add
your
topics
I'm
gonna
share.
A
I
hope
you
can
see
that
so
few
topics
for
today.
Scott,
do
you
want
to
take
off.
B
Yeah
sure
so,
just
a
few
updates,
the
the
most
recent
thing
was
we
had.
We
had
an
issue
with
one
nine
zero,
where
we
changed
the
way
the
validation
was
done
for
bsl's
and
and
one
nine
zero.
There
was
a
situation
where,
basically
being
it
was
being
constantly
validated
over
and
over
again,
several
times
a
second,
rather
than
once
a
minute,
there
was
a
fix
that
was
pushed
to
the
branch
for
that
that
works
in
the
full
upstream
environment,
where
you're
managing
your
bsls
directly.
B
The
validation
for
the
bsl
spec
was
calling
initializing
the
object
store
and
the
object
store
needs
certain
things
added
to
the
configuration
map,
credentials,
file
and
other
things,
and
what
that
code
was
doing
was
actually
modifying
the
in-memory,
bsl
config
and
then
passing
in
that
config
value,
but
not
saving
it.
B
But
the
bsl
validation
was
saving
the
bsl
object
because
it
had
other
things,
changes
that
it
might
need
to
make,
for
example,
as
default
and
the
last
update
time,
and
so
there
was
a
side
effect
of
calling
this
validation
of
inserting
these
unnecessary
fields
in
the
spec,
which
then
our
odp
controller
was,
you
know,
removing
immediately,
and
so
they
were
kind
of
fighting
over
that
config
value,
which
meant
that,
even
though
the
change
was
made
to
the
processing
for
oadp,
we
were
still
doing
the
multivalidation.
B
I
have
a
pr
that
was
approved
on
main
that,
basically,
instead
of
doing
that,
it
makes
a
deep
copy
of
that
config
map
memory
in
memory
and
passes
that
on
to
the
object
store,
because
the
object
store
doesn't
actually
deal
with
backup,
storage
location
objects.
It
just
takes
a
map
of
config
values.
So
there's
no
need
to
modify
that
kubernetes
resource
anyway.
So
with
that
change,
we
shouldn't
see
this
kind
of
flip-flopping
of
config
in
environments
where
there's
an
external
controller,
managing
backup
storage
locations.
So
that's
merge
to
main.
B
We
just
need
to
also
get
it
onto
the
release19
branch.
I
think
that
was
because
we
were
hoping
to
get
191
to
fully
fix
that
validation
bug
once
that
comes
out,
so
we
just
need
one
more
act
for
that
one.
If
we
can
get
that
today,
that
would
be
great,
but
you
know
which,
whenever
it
can
be
so
I
have
a
couple
of
other
pros
out-
item
action,
progress,
monitoring,
design.
We
talked
about
this
in
a
little
more
detail.
Last
week
I
made
some
updates
in
response
to
some
feedback.
B
I
still
need
to
look
at
that
open
invitation
pr
from
dave
from
a
few
months
back
to
see
if
that
pr
introduces
any
new
concepts
that
are
missing
in
the
in
the
design.
So
I
will
look
at
that
this
week
as
well
and
if
anyone
else
has
further
commentary
on
that,
that
would
be
great
and
then
I
have
a
implementation
pr
for
vsl
credentials
submitted
before
1
9,
but
we
decided
it
was
too
risky
to
get
that
in
before
the
1
9
deadline.
So
we're
hoping
to
get
that
in
110..
B
I
did
see
there
was
a
comment.
I
need
to
go
back
and
look
to
see
if
I
need
to
make
a
change
or
not
to
that.
So
that's
pretty
much
it
for
me.
B
Yeah
and
if
you,
if
you
have
a
question
now,
we
can
talk
about
it,
if
not
comment
on
the
pros
and
we
can
take
it
that
way
on
github.
A
B
I
have
not
heard
anything.
I
don't
really
know
what
his
level
of
participation
is.
B
At
this
point,
the
one
big
question
I
kind
of
had
for
dave,
and
you
know
in
the
absence
of
any
feedback,
it
may
be
that,
because,
basically
one
of
the
one
of
the
issues
with
and
this
this
was
brought
up-
I
believe
one
of
the
reviewers
was
together
so
we're
on
the
item,
which
one
is
that
yeah,
the
the
item,
action,
progress,
monitoring
and
one
of
the
things
that
today's
proposal
it
had
was
that
when
the
backup
was
com
was
otherwise
completed
and
we're
just
waiting
for
uploads,
it
would
go
into
this
uploading
state
and
we
we've
renamed
it
in
my
pr
to
be
a
more
general
and
waiting
for
an
action,
completion
or
something.
B
But
the
point
was
that
in
his
design
we
weren't
writing
the
backup
to
the
object
store
until
we
all
that
waiting
was
done,
and
then
we
moved
on
to
completed,
and
the
question
was
on
the
post
on
the
pr.
You
know
why
don't
we
just
write
it
now
because
we're
not,
even
though
we're
waiting
for
action
to
finish
we're
waiting
for
external
processes.
To
finish,
there's
nothing!
That's
going
to
happen
in
that
upload,
that's
going
to
affect
the
content
of
the
valeria
backup.
B
Why
don't
we
just
write
it
now
that
we
have
it
instead
of
holding
it
in
memory,
while
the
next
backup
goes
and-
and
I
actually
agree
with
that
change-
I
think
this
also
gives
us
if
the
valero
server
were
to
crash,
say
during
that
next
backup,
the
backup
that
was
otherwise
completed
and
just
waiting
well
that
backup,
it's
already
written.
So
we
have
it.
You
know,
whereas
we'd
lose
it
in
the
case
where
we're
waiting
to
upload
it.
So
I
actually
think
we
should
probably
do
that.
B
I
was
kind
of
curious
whether
dave
had
a
reason
for
the
wait
waiting
or
it
was
one
of
these
things
where
he
just
kind
of
thought,
since
we're
currently
not
writing
until
we're
done.
Let's
just
wait
so
that
might
be
a
change
we
could
make,
without
any
detriment
to
functionality
and
with
some
obvious
kind
of
improvements.
So
my
inclination
at
this
point
just
to
go
with
the
idea
that
once
we
are
at
the
point
where
we
update
the
backup
state
to
hey
we're
just
waiting
for
you
know
uploads
to
complete.
B
That's
the
point
where
we
need
to
write
the
back
of
the
object
store,
but
there's
so
there's
some
question
about.
Like
logs,
do
we
write
logs?
We
may
have
to
refactor
the
logs
so
that
we
can
then
add
additional
logs,
because
obviously
this
there
will
be
some
logs
associated
with
this
follow-up.
You
know
checking,
but
that
would
just
be
one
more
set
of
message
and
then
we
may
not
even
need
that,
because,
basically,
at
this
point
the
backup's
done.
B
The
only
thing
left
for
this
backup
is
for
valeria
to
wait
until
we're.
You
know
if
there
are
any
errors
from
those
processes
that
would
obviously
need
to
go
into
the
log.
So
that
may
be
the
complication
of
needing
to
you
know,
update
or
add
additional
log.
I
think
that's
fine,
I
think
we
even
may
have.
I
thought
there
might
have
been
a
pr
a
while
back
where
we
had
some
notion
of
adding
another
log
after
the
first
log,
so
we'll
have
to
think
about
how
this
affects
logs.
B
I
think
that's,
probably
the
main
complication,
but
my
inclination
is
to
do
that
to
just
to
write
the
backup
as
soon
as
valero
is
done
with
it,
because
anything
else
is
going
to
reduce
reliability
over
what
we
do
now.
B
Any
I
don't
know
if
there's
any
other
comments
or
feedback
from
on
that.
But
again
you
know
feel
free
to
respond
now
or
respond
on
the
pr.
B
If
you,
if
you
have,
if
you
can
reach
out
to
him,
just
to
see
if
he
has
any
comments
at
a
big
and
mainly
just
because
I'm
taking
a
proposal
he
made
and-
and
you
know
deliberately
changing
it
and
combining
another
use
case,
so
it'd
be
great
to
hear
his
feedback
on
whether
I
missed
anything
important
or
whether
I'm
going
in
the
wrong
direction.
You
know
we
can
go
for
without
it
if
he's
totally
unavailable.
But
if
he's
got
the
time
to
look
at
it,
that
would
obviously
be
welcome.
B
And
and
if
if
he
says,
everything's
fine-
or
he
doesn't
have
time
to
look
at
it-
that's
fine
too
just,
but
if,
if
if
he
wants
to
provide
some
feedback,
you
know
you
know
as
to
suggest
the
changes
or
you
know
you
know
pointing
out
why
certain
things
were
done
the
way,
and
maybe
we
need
to
go
back
to
that.
That's
fine,
too,.
A
Yeah,
okay,
cool
anyone
else
anything
to
discuss
on
this
one
or,
if
not,
anyone
planning
any
vacations
for
the
upcoming
months,
because
it's
like
for
at
least
for
our
region,
it's
like
middle
of
the
summer
and
I'm
I'll
be
on
vacation,
the
fu
august
and
half
of
september.
A
So
I
wanna
plan
yeah.
B
Not
much
for
me,
I'm
gonna
be
gone
for
about
a
week
kind
of
end
of
july,
beginning
of
august
kind
of
half
of
the
for
half
of
the
end
of
july
and
half
of
the
first
week
of
august,
so
yeah
so
so,
but
about
a
week
total
kind
of
july,
27th
until
august
2nd
something
in
that
range.
A
B
Yeah
yeah,
that's
fine,
so
so
that
would
be
the
meetings,
the
the
the
the
noon
meetings
twice
a
month
during
august
and
the
first
fourth
of
july.
If
you
just
send
me
a
message
offline.
Just
to
you,
know
slap.
A
Yeah,
I
have
to
add
you
as
a
host,
so
I'm
not
sure
how
I
can
do
that
because
you're
external
to
to
that
zoom
account
yeah.
It
shouldn't
be
a
problem.
I
think
yeah,
okay
and
they're,
all
in
the
calendar
already,
so
you
have
it
yeah.
B
Just
right,
right
and
and
yeah
I
might
suggest
to
if
you,
if
you,
if
you
have
to
go
to
that
trouble
and
you
can
get
it
to
work
if
you
can
add
shoebom
as
well,
because
that
way,
if
there's
a
week
that
I'm
gone,
he
might
be
available.
You
know
just
because,
for
example,
I
know
I'm
gonna
miss
one
tuesday
and
the
for
the
first
tuesday
in
august.
I
don't
know,
I
don't
that's
a,
I
think.
B
That's
the
us
timezone
meeting
mm-hmm
yeah,
so
so
I
know
the
first
tuesday
in
august
I'm
not
gonna
be
available
and
there
may
be
others,
but
that's
kind
of
you
know
it
it.
Basically,
if
shubham
can
be
a
backup-
and
I
haven't
talked
to
him
yet
so
I
don't
mean
to
volunteer
him,
but
you
know
I.
A
Okay,
let's
okay
I'll
take
that
offline,
get
all
the
dates
and-
and
we
can
discuss
like
in
common
chat
in
slack
or
something
can
take,
what
okay
and
the
next
one.
Thank
you
very
much
about
that
and
the
next
one
updates
on
the
meeting
time
change.
I
think
it's
still
an
ongoing
discussion.
Unfortunately,
there
is
no
like
much
of
attention
right
now,
so.
B
And
obviously,
with
your
upcoming
outages,
let's
not
change
it
until
you
get
back
probably,
but
I
I
just
wanted
to
mention
briefly,
if
you
don't
mind
kind
of
what
I
what
I
commented
there
in
case,
there's
feedback.
I
know
you'd
initially
hoped
to
have
a
single
time
that
worked
for
everybody
and
I
think
the
feedback
from
that
I've
seen-
and
I
think
I
agree
with
that-
is
that
there's
really
not
one
time
that
works
for
everybody.
C
B
I
think
we
probably-
and
I
think
the
beijing
time
is
fine-
that
works
for
us
east
coast
that
works
for
us
west
coast.
It
doesn't
work
for
you,
unfortunately,
which
is
but
again.
This
goes
back
to
the
problem
of
the
number
of
time
zones,
but
the
u.s
time,
though
the
current
meeting,
I'm
thinking
that
if
we
move
this
meeting
two
hours
earlier,
that
makes
it
better
for
beijing.
So
it's
still
it's
still
not
ideal
for
them.
B
But
it's
close
enough
so
that
if
there's
an
important
issue
discussion,
they
can
come
in
because
it's
going
to
be,
you
know
10
o'clock,
not
midnight,
so,
in
other
words,
my
thinking
is,
if
we
can
make,
the
best
schedule
is
one
that,
for
most
people
is
convenient
twice
a
month
and
slightly
inconvenient,
but
still
possible
twice
a
month
we're
trying
to
avoid.
You
know
having
a
time
where
it's
totally
impossible
for
a
maintainer
to
come
half
the
time.
So
I
would
recommend
moving
the
u.s
centric
meeting,
say
two
hours
earlier.
B
That
does
make
it
less
convenient
for
the
west
coast
for
california
because,
instead
of
it
nine
it's
seven
but
again
that
goes
back
into
the
it's
slightly
inconvenient,
but
still
possible
if
it's
an
important
issue,
but
otherwise,
I
believe
the
beijing
centric
time
is
actually
much
more
convenient
for
the
west
coast,
because
you
know
it's
four
o'clock
and
five
o'clock.
I
guess
there.
So
I
don't
know.
I
that's
just
my
my
you
know
kind
of
my
opinion
just
as
a
kind
of
a
factor
there
and
what
I
would
support.
C
I'm
okay
with
that
moving
the
9
a.m;
pacific
time
to
7
a.m!
Pacific
time,
also,
I'm
okay
with
that,
because
that
also
helped
my
teammate
jeon
chen
in
boston.
So
he
can
attend
that
time
zone,
and
that
sounds
okay
too.
So
I
I'm
I'm
okay
with
that.
If
we
want
to
move
right
if
we
agree
to
move
because
I
also
know
that
you
know
people
might
not,
may
not
work
that
early
in
california
right.
B
Yeah
yeah,
my
thinking,
is
that
moving
in
two
hours
earlier
makes
it
a
little
bit
harder
for
you
in
california,
but
a
lot
easier
for
beijing.
That's
kind
of
my
thinking
there,
the
beijing
time,
I
think,
is
fine,
as
is
I
think
any
earlier,
is
bad
for
them
and
any
later
is
kind
of
bad.
B
Yeah,
that's
my
thinking.
Is
it
it's
still
kind
of
late
for
them?
So
that's
not
ideal,
but
it's
better
than
midnight,
because
you
know
right
now.
It's
midnight
in
beijing
and
it'll
be
10
p.m,
which
is
you
know
it's
late
for
a
meeting,
but
if
there's
an
important
issue
that
you
want
to
discuss,
it's
still,
probably
when
you're
awake.
So
it's
it's
an
improvement,
even
though
it's
not
ideal-
and
you
know
any
earlier
than
that-
I
mean
I,
you
know
even
even
say-
9
a.m
at
10
a.m.
B
My
time
would
be
fine
for
me,
but
that's
even
harder
for
california.
So
that's
why?
I'm
that's
why
I'm
saying
make
it
two
hours
earlier
instead
of
three
hours
earlier,
but
you
know
from
for
the
east
coast.
Three
hours
earlier
would
also
work
and
would
be
even
better
for
beijing,
but
again
that's
worse
for
california.
So.
B
And
the
other
question
is,
I
don't
know
who's
going
to
be.
I
know
in
terms
of
people
in
california
other
than
fong.
If
dave
is
going
to
be
participating
regularly,
then
you
know
his.
His
availability
is
also
relevant
here.
If
he's
actually
not
going
to
be
in
the
meeting
instead
of
he
hasn't
been
in
the
last
few
weeks,
then
we
don't
really
need
to
accommodate
that.
B
I
don't
know
what
to
say
so,
if
you're,
if
you're
reaching
out
to
him
on
this
other
issue,
it
might
be
worth
bringing
that
up
as
well
just
to
see
where
he
is
on
it.
If
he's
not
planning
on
being
on
the
meetings,
then
we
cannot
worry
about
that
aspect
of
the
scheduling,
but
if
he
is,
then
you
know
knowing
what
is
available
is
helpful.
A
B
Yeah,
as
far
as
I
know,
right
now,
everybody
on
the
red
hat
side-
that's
contributing
is
in
the
us
east
coast,
so
any
of
the
times
we're
discussing,
I
think,
it'd
be
fine.
You
know
anything
anything
from
9am
later,
so
you
know
the
only
area
where
I
think
from
the
red
hat
point
of
view
would
be
not
ideal
would
be
if
we
decided
to
make
the
beijing
time
later
than
it
currently
is.
But
if,
as
long
as
it's
the
current
time,
I
think
we're
good
on
that.
One.
A
A
That
change
should
have
how
how
to
put
that
should
take
a
bit
longer
to
be
discussed.
Oh.
B
Yeah
and
and
and
I
I
don't
think
it
makes
sense
to
change
it
right
before
you
leave,
for
you
know,
you
know
vacation
and
other
people
going
to
be
gone
anyway.
So
so
I
I
think
discussing
it
now
is
good,
but
I
think,
let's
plan
for
making
a
change
back
when
vacations
are
resolved
and
we're
back
on.
You
know
kind
of
more
predictable
schedules.
I.
A
Think,
like
beginning
of
september
mid
september,
we
can
yeah
sure
start
doing
it:
okay,
great
okay.
So
anyone
else
any
any
other
topics
to
discuss.
A
Nope
we're
all
still
waiting
for
the
kipcon
announcements.
I
think
it
will
be
like
mid
august
or
beginning
of
the
first
week
of
august.
B
A
Yeah
did
they
announce
the
the
final
cut
for
the
for
the
talks?
So
let's
see
what's
accepted
and
what's
not.
C
B
Oh
fun
was
there
anything
the
issue
with
the
protobuf
stuff,
and
I
saw
a
comment
that
the
the
suggested
change.
I
that
the
work
for
me
wasn't
working
in
europe.
Were
you
able
to
work
that
out?
I
saw
another
comment
from
someone
else.
C
Update,
I
I
was
able
to
update
it
to
your
version,
the
portal
3.6
or
something
whatever
the
version
you
hold
on.
I
was
able
to
make
it
to
run
update
com
completely
without
error.
However,
I
I
got
stuck
on
other
issues,
so
I
haven't
okay.
C
B
I
do
know
that
that,
from
the
discussions
when
I
was
trying
to
follow
through
and
figure
this
out,
this
was,
I
believe,
you
know
a
more
recent
functionality
update.
So
if
you
had
an
old
version,
it
might
not
work,
but
I
know
it
worked
for
me
and
I
think
I
saw
another
comment
from
another
maintainer
that
the
suggested
I
made
suggestions.
B
I
had
also
worked
for
him,
so
hopefully
with
the
same
version,
and
hopefully
the
version-
that's
in
you
know
that
the
other
question
we
have
to
figure
out
is
once
it
works
for
us
in
our
local
environments.
You
know
in
terms
of
the
builds
and
the
container
builds
and
the
versions
that
valero
uses.
We
need
to
make
sure
that's
working.
Obviously
we
may
need
to
update
that
dependency
if
it
needs
something
newer,
but
that's
something
we
can
handle
once
it's
all
working
locally
and
everything
looks
right.
You
know
that
way.
B
Right,
there's
a
pr
for
it.
I
just
don't
remember
the
name
of
it
all
right,
so
yeah,
if
you,
if
you,
can
look
up
the
pr
and
send
and
link
it
to
slack,
so
everyone
can
see
it,
but
it's
basically
it's
basically
the
pr
where
we're
implementing
the
plug-in
versioning
for
or
backups.
I
think
you
started
with
backup.
Service
location
was
the
first
one.
Sorry
backup
item
action.
Yes,
backup
item
action
view
one,
I
believe
is
in
the
title
of
the
pr
or
v2.
No,
that
could
be
one
sorry.
C
A
B
B
Yeah
yeah,
because
because
because
this
involved
moving
some
of
our
of
the
the
protobuf
stuff
into
multiple
directories
with
different
packages,
and
so
we
needed
to
change
the
way
we
were
building
to
get
those
files
to
show
up
in
the
right
place,
the
generated
go
files
so.