►
From YouTube: CNCF SIG Storage 2020-12-09 (2 of 2)
Description
CNCF SIG Storage 2020-12-09 (2 of 2)
A
B
A
C
A
That's
the
old
there,
I'm
I'm
not
hearing
you
very
well,
but
that's
the
way
of
thinking
right
way
as
soon
as
you
say,
restore
as
in
I
was
backing
up,
and
then
I
have
to
restore
to
somewhere
else.
That's
exactly
the
old
way
of
thinking
we
now
can
can
do
constant
sync
between
our
failure
domains
so
that
each
have
a
copy
of
the
state
or
enough
of
the
state
so
that
they
don't
have
to.
A
They
can
continue
work
if,
if
you
lose
one
of
the
failure
domain-
and
that's
really
the
message
that
we
that
that
we
have
to
give
in
this
paper
to
say,
look,
there
is
a
new
way
of
thinking
now
and
new
options
provided
by
the
technology,
and
you
can
really
set
up
your
failure,
domains.
Okay
in
such
a
way
that
if
you
lose
one,
you
don't
have
to
stop
your
you.
Your
service
is
still
up
and
can
you
can
continue
to
serve.
D
B
Oh,
I
was
just
confirming
what
he
said.
I
was
using
a
bad
mic
and
what
he
clarified
was
kind
of
what
I
meant
about
geo-replication
kind
of
taking
over
disaster
recovery.
A
E
I
would
like
to
see
a
discussion
around
that,
so
we
can
get
to
the
same
nouns
and
verbs
and
who's
do
who's,
taking
on
what
role
or
what
entity
is
taking
on
what
action.
D
C
D
Of
components.
C
D
Know
a
group
of
components
or
or
you
know,
and
and
often
that's
tended
to
me
in
a
data
center,
but
I
think
honestly,
we
should
kind
of
look
at
this
in
terms
of
a
failure
domain
or
or
or
an
availability
zone,
because
the
the
the
same
exact
concepts
that
apply
to
two
data
centers
also
apply
to
two
racks,
or
you
know,
two
availability
zones
in
a
cloud
provider
or
whatever
right
the.
I.
D
I
think,
what
we're
what
we're
talking
about
is
is
sort
of
losing
the
entirety
of
of
of
a
of
a
physical
of
a
physical
location,
and
that
can
be
a
small
or
a
large
component,
and
I
think
yeah,
just
just
following
on
from
what
tom
said
there
as
well.
You
know
how
the
topologies
affect
the
orchestration.
D
Where
you're
you
know
where,
where
you're
discussing
where
we
can,
you
know
potentially
discuss
how
different
systems
replicate
or
or
or
provide
you
know,
data
protection
across
sites
or
across
failure.
Domains
versus,
for
example,
you
know,
what's
what's
also
quite
common
is-
is
to
have
distributed
topologies,
which
are
actually
stretched
across
multiple
locations
or
across
multiple.
You
know,
sites
or
whatever.
As
well
and
middleware,
that's
required
to
do
that,
so
so
yeah
there's
there's
there's
a
lot
of
interesting
topics
there.
D
I
I
really
do
think
that
this
this
could
be
one
of
those
things
that
you
know
it
could
be
a
100-page
document.
If,
if
we're
not
careful
and-
and
I
would-
I
would
strongly
suggest
based
on
you
know-
what's
happened
in
the
past
right,
you
know
we
want
to
make
sure
that
we
actually
produce
something
so
scoping
out
at
the
beginning
and
figuring
out
which
sections
we
want
to
to
deal
with
is
is
is
important
because
I
would
it
would
be
lovely
to
do
this
incrementally
I
you
know
we
we
may.
C
D
F
So
I
have
seen
and
hurt
people
using
disaster
recovery
for
making
sure
that
your
workforce
still
has
an
office
to
work
or
your
workforce
still
has
internet
connectivity
or
that
you
have
a
contract
with
home
depot
to
bring
in
power
generators.
Things
like
that,
so
the
term
is
really
also
used
in
different
with
a
different
meaning.
D
The
term
I
I
I
suggest
that
we,
you
know
that
that
can
be
something
we
can.
We
can
talk
about
for
sure.
I'm
not
you're
right,
I'm
not
entirely
sure
what
what
the
what
the
final
terminology
should
be,
but
I
I
do
strongly
believe
that,
once
we
kind
of
highlight
the
scope
we'll
be
able
to
we'll
be
able
to
to
make
that
something
useful.
D
D
So,
just
just
in
terms
of
next
steps:
rafaela,
if
you,
if
you
have
the
time
to
to
put
together
some
some
bullet
points
or
or
topics
or
areas
that
you
think
we
we
should.
We
should
cover
in
in
a
document
that
that
would
be
a
good
start.
And
then
maybe
you
can
share
it,
share
the
google
doc
with
the
group
and
I'm
I'm
sure,
we'll
get
sort
of
lots
of
comments
and
lots
of
feedback.
A
Okay,
I
can
do
that.
I
also
would
like
to
do
what
quinton
was
suggesting.
Maybe
I
can
bring
this
team
up
to
speed
with
what
I
have
already
created
sure.
C
A
And
just
give
you
a
brief,
you
know
maybe
I'll
create
a
deck
that
summarizes
all
the
content
that
I
have
to
see.
If
we
can,
you.
D
That
would
be
absolutely
fantastic
yeah.
Do
you
think
that
this
would
be
that
this
would
be
something
you
could
do
for
the
beginning
of
january?
Maybe.
D
That
sounds
great
and
thank
you
so
much
rafael.
This
is
this
is
a
brilliant,
a
brilliant
topic.
Does
anybody
else
have
any
other
questions
or
any
other
items
we
want
to
raise,
because
otherwise,
I
think
maybe
we
can
finish
this
goal.
A
few
minutes
early.
F
Yes,
phil,
I
just
maybe
pretty
unrelated,
I
I
tried
to
file
periods
for
cncf
sandbox
and
filled
out
the
form
and
I'm
a
little
concerned
because
nothing
came
back,
not
even
an
email.
D
Yeah
I
I
can
yeah
I
can.
I
can
send
you
that
thing
as
well.
The
best
person,
if,
if,
if
you're
on
the
cncs
slack,
the
the
best
person
to
to
ping
is,
is
amy
scovard,
perrin,
who's,
the
the
cncf
project
manager
and
she
does
the
she
organizes
the
scheduling
and
the
project
voting
for
the
sandbox.
B
B
D
G
Just
just
one
kind
of
public
service
announcement:
we
we
probably
have
more
deep
technical
work.
Then
we
have
capacity
to
do
it
next
year.
We're
gonna
have
to
do
a
lot
of
project,
check-ins
and
evaluations
of
various
things.
G
So
short
summary,
I
think
if
anyone
foresees
themselves
having
enough
time
in
the
new
year
to
to
serve
in
the
capacity
as
a
sort
of
senior
technical
person
on
on
this
sig,
I
think
it
would
be
a
good
time
to
give
it
some
thought
over
the
over
the
holiday
period,
and
then
we
can.
We
can
sort
of
start
talking
about
that
more
in
earnest
in
the
new
year.
D
Indeed,
if
just
echoing
what
quinton
said
you
know,
we're
we're
looking
for
tech
leads,
who
can
help
join
the
stick
and
can
help
with
project
reviews,
but
also
you
know,
working
on
documents
like
like
we're
discussing
today.