►
Description
En esta charla, veremos cómo podemos hacer un disaster recovery de nuestro querido cluster de Kubernetes, así como de las diferentes aplicaciones que corran en el. Revisaremos cómo Velero puede ayudarnos en nuestra misión y cómo no, lo veremos en directo con un hands-on donde veremos cómo podemos recuperarnos de un oh shit moment en nuestro cluster de Kubernetes.
Contacta a Ramiro en:
- https://www.linkedin.com/in/ramiro-alvarez-fernandez-3b945b77
- https://twitter.com/kaskol10
A
A
B
C
You
all
very
much
for
attending
this
talk.
I,
hope
you
are
all
enjoying
this
event
of
the
governors
of
Guatemala
and
you
really
like
being
able
to
share
the
bill
with
all
these
speakers
who
are
very,
very
powerful
and
Thank
you
very
much
especially
to
the
organization
for
making
it
possible
and
well,
rulers
is
a
great
choice
to
run
all
our
worlds,,
but
once
we
are
running
rulers
in
production,,
you
start
to
ask
yourself
some
questions
that
go
through
your
head,.
How
do
I
have
backups
to
recover?
the
cluster
of
a
disaster,?
C
How
do
I
deal
with
this
data?
Corruption?
Well,?
The
objective
of
this
talk
is
to
answer
these
questions.
We
are
going
to
see
how
an
open
source
solution
can
help
us
to
make
a
backup,,
restore
and
migration
of
our
cluster.
Cúber,.
You
put
our
volumes
or
there
are
several
of
them
on
today's
agenda,,
because
we
are
going
to
talk
about
sailboat
use
cases
and
we
will
see
the
components
that
gold
has,,
those
data,
that
we
can
weather,
restore
generic
workflow
and
once
the
theory
is
finished,
we
will
see
the
action.
C
We
will
do
a
demo
with
a
couple
of
scenarios.
The
first
one
is
disaster
recovery
where
between
the
dks
clusters
and
the
second
one,
a
data
migration
between
a
ws
cluster
and
a
jica
gtp
cluster
and
at
the
end
some
conclusions
well
commenting
on
a
little
like
how
it
went
if
we
were
able
to
do
the
demo
correctly,
very.
B
C
The
main
use
cases
of
sailboat
would
be
the
disaster
recovery
to
reduce
the
recovery
time
in
case
of
a
catastrophe,
as
well
as
the
integrity
of
the
data
and
will
allow
the
sailboat
allows
us
make
scheduled
backups
for
you
to
protect
our
beloved
workflows
data
from
cuber
nets
and
protect
us
from
data
loss
due
to
hardware
theft,
to
maintain
a
retention
policy
to
comply
with
certain
company
policy.
We
protect
from
data
corruption.
That's
what
data
is
integrity
and
data
migration.
C
A
sailboat
gives
us
that
portability
to
play
between
one
cloud
provider
and
another,
as
you
already
can
from
our
cuber
net
cluster,
that
even
re-index
data
from
our
production
cluster
to
a
smaller
environment,
to
test
for
example,
also
allows
us
and
Consistency
All
this.
That
I
am
commenting
on
sailboat
can
be
done
in
a
declarative
way
and
using
an
approach
and
tops,
as
was
normally
commented
in
the
previous
talk,.
C
The
eros
components
are
worth
it,
since
a
command
line
that
we
will
have
installed
at
our
premises
and
a
sailboat
server
controller
that
we
will
have
installed
in
the
class
of
rulers
and
that
this
displays
a
series
of
this
reale.
In
the
class
of
rulers
that
allows
you
to
interact
with
the
werder
command
line
and
allow
you
to
make
that
backup,
jesse
rector,
those
are
the
two
key
components:
the
berklee
and
the
server
control
value.
C
Yes,
we
are
in
the
case
of
doing
a
data
migration
between
different,
clear
providers,
a
crosstour
to
the
provider,
because
we
need
a
layer
of
agnosticism
at
the
system
level,
and
that
is
what
will
help
us
to
resist
that.
It
is
not
a
velero
component
per
se,
but
for
that
data,
migration
It
is
necessary
for
the
agnosticism
of
the
delphi
system.
Yesterday,,
as
I
commented
very
well,,.
B
C
On,
the
sailboat,
he
communicates
directly
with
the
sepia
and
discovery
ruler
In
this
way,.
You
don't
need
to
enter
the
detective
to
get
all
that.
Information.
Let's.
Remember
that
the
main
cloud
providers,
as
a
scholarship,
are
the
ones
that
fall.
They.
Don't
let
us
enter
Twitter,.
So
there's
no
problem.
C
C
Velero
gives
us
that
flexibility
of
use,
since
it
allows
us
to
launch
back
at
normally
scheduled
when
you
want
each
work,
every
24
when
you
need
it
and
also
filter
the
backup
or
a
multitude
of
fields,
either
in
space
or
a
resource
or
the
carnations
of
our
gubernatis
resources,
and
it
also
offers
us
that
extensibility
extensibility
in
terms
of
plugins,
for
example,
add
a
ws
plugin
that
communicates
with
that
3.
Add
the
gcs
plugin
that
type
of
plugins
that
also
contribute
are
part
of
the
community
and
also
offer
us
our
ability
at
game
level.
C
For
example,
we
make
a
backup
and
after
making
the
backup,
we
want
to
launch
a
self
to
do
whatever
we
can't
do
it
through
a
voucher,
the
sailboat
workflow.
In
the
end,
the
summary
we
have
a
sailboat
to
make
a
backup-
and
we
don't
always-
but
whenever
we
have
a
current
back
to
restore
it
or
try
to
restore
it,
to
make
a
backup
from
our
command
line.
Sailboat,
we
win.
The
ego,
create
baquet,
backup
this
attacks
against
the
complements
and
ceip,
and
there
is
a
list
of
resources
that
we
need
or
that
we
have
filtered.
C
So
that
the
backup
and
it
communicates
with
the
werder
observer
control,
the
veil
of
server
control,
I,
put
the
backup
to
that
cloud.
Diet
stories
once
it
takes
a
while
it
depends
on
the
data
that
we
have
to
weather
and
back
and
after
that,
the
normal
thing
is
that
we
do
a
restore
not
every
day,
but
the
day
the
whole
class
of
rulers
falls.
We
want
to
have
a
guaranteed
rector
so
to
do
the
rest
from
our
command
line.
We
keep
a
restore,
create
genre
from
the
previous
backup.
B
C
C
C
They
selected
a
location
where
we
want
that
backup,.
In
our
case,
it
will
be
a
bouquet
of
w
and
well,
the
necessary
permits,
and
others
are
worth
this
first
example,
well,.
We
were
close
to
that
cluster
1
k
s
class
2,,
it's
worth
what
I
have
in
this
case
and
you
can
have
1,
I
already
have
the
three
clusters
raised,.
Otherwise
it
would
take
us
a
long
time
to
raise
the
three
that
is
better
for
you,
I
have
the
link
space
where
It
is
found,
well,.
C
Several
of
us
have
added
the
drumming
of
ws.
We
have
added
the
default
so
that
this
time
this
back
ds3
comment
here
and
well.
We
put
the
credentials
here.
The
credentials
are
here
because
of
the
crisis-
kiseki,
but
valero
also
accepts
today
from
identity
provider
that
it
would
be
very
good
to
install
it
for
environments
already
in
production
now
for
A
normal
example,.
This
is
fine
because
we
have
our
sailboat
history
and
there
is
a
space
called
ghost
that
has
this
boy:
space
band
and
well,.
C
C
C
And
we
can
create
cluster
1
I
am
going
to
say
that
it
takes
a
while
a
little
bit.
We
include
in
space
ghost,
as
you
saw
in
game
facebook.
It
has
because
it
has
a
service,
it
has
a
set
replica
and
we
forgot
when
commenting,
it
has
those
gdp
and
if
they
are
from
the
three
gp2
classes,
other
gigs
well.
B
C
C
It
looks
good
and
included
the
link
space
ghosts.
S
3
pl
of
720
I,
think
that
it
is
a
week
and
that
he
added,
well,
added
the
control
of
clement's
text.
Replicas
persists
in
bolus.
The
links
persist
with
their
snapshots.
If
we
go
to
the
ws
console
and
we
are
going
to
go
from
2
well,
we
will
also
see
that
we
are
not
going
to
see
if
I
am.
Cheating
is
like
that.
C
He
took
two
snapshots
and
if
we
see
that
there
are,
those
are
the
two
snapshots
that
we
have
very
well
and
we
have
the
backup,
and
since
we
are
in
this
in
the
case
of
ws,
we
are
going
to
go
to
cluster
2.
Well
in
cluster
2,
we
will
return
with
the
px
We,
don't
stop
with
cluster
2
that
we
have
here,
well,.
You
will
see
that
I
don't
have
it,
I
only
have
sailboat
installed,
I
installed
it
the
same
as
in
the
previous
one,
well,.
C
Let's
see
what
we
can
do
with
sailboat,
vent
backups,
there
it
is,.
There
is
our
backup
that
we
want
to
restore
the
backup,.
Then
we
will
backup
it.,
1
and
wake
up.
We
wait
a
bit
this
time
on
arrival,
plus
the
call
as
an
example.
I
didn't
read
much
to
be
honest.
We
are
going
to
do
the
details
of
this
case.
For
example,
6
m
okay,
I
have
completed
go
here.
It
restored
us.
Okay,
confirm
that
now
It
existed,.
C
B
C
Then
the
scenario-
the
first
scenario
would
be
this
first
scenario
where
we
have
made
a
backup
of
kale
cluster
1
and
the
rest
of
cluster
2
is
done
now
we
are
going
to
add
one
more
piece
and
we
are
going
to
see
how
to
do
from
this
cluster
2
in
which
it
is.
We
are
going
to
make
the
backup
and
restore
it
to
a
sheikh.
As
you
can
see,
we
are
missing
the
diamond
set.
Well,
we
would
need
to
do
something,
and
then
it
was
explained
that
the
wall
was
worth.
C
What
I
am
doing
well
I
am
adding
that
it
projects
from
montse
terrestrial
by
default.
In
it
we
call
the
same
official
chart.
I
keep
adding
the
ws
plugin.
We
keep
putting
the
package
of
rulers,
you
see
and
a
couple
of
important
pieces
I'm
going
to
do
a
bit
of
that
today,
just
in
case
a
couple
of
important
pieces,
we
add
the
tourist
table.
C
Football
is
all
it
means,
because
the
volumes
by
default
would
already
be
annotated,
otherwise
the
volumes
would
have
to
be
annotated
and
well
it's
a
tad
good
from
god,
gifts
from
god
and
then
pge
confirms
them.
As
you
kids
know,
and
gp2
gp3,
usually
Being
habitual
of
double
gtp.
We
have
request
and
standard
standard.
It
is
simply
okay,
so
the
option
a
couple
of
options
we
have
is
not
to
create
that
gp2
in
class
and
cai
stores-
or
we
simply
say
today
when
it
is
restored
action.
C
Then
we
give
it
the
change
and
pass
me
on
gp2
That's,
it,
and
well,
again,
well,
we'll
pass
on
the
credentials.
Well,.
How
would
this
look
?
Well,?
Now
we
see
that
we
have
the
stick
there,
and
it's
okay.
It's
fine,.
How
is
the
status
stick,
but
well,
we're
going
to
add
a
dessert,
another
new
post,
since
we're
still
here
in
two
we
are
going
to
add
a
port
in
cluster
2.
C
B
C
A
little
and
the
distraction
no
longer
comes
out.
We
are
going
to
see
the
details
and
in
the
details
well
completed.
Also
s
3,
2
everything
again
and
the
The
difference
is
that
this
time
there
is
no
one
other
than
a
sailboat
except
festival
caps,
and
there
it
is,
okay,.
Once
we
have
that
backup,
we
can
move
on
to
the
3D
cluster.
B
C
B
C
Go
to
the
children
see
and
there
we
have
children
62.
We
go
to
little
things
that
you
have
and
there
it
is
all
and
from
tbc,
and
we
see
that
this
stock
automatically
classes
before
it
was
gp2
I
think
we
had
it.
Oops
I,
don't
know
if
it
had
it
here,
then
the
wolf
taught
it
was
the
gp2,
and
now
it
is
standard.
If
we
see
the
wells
well,
we
will
wait
but
to
see
how
it
is.
C
B
B
B
C
The
x
forward,
we
go
back
to
zero
rifles
and
we
should
go
to
class
two
moon
classes
and
if
we
return
to
our
pleasure
one
from
us
to
form-
and
it
should
go
out
only
there-
you
are
welcome
Guatemala,
so
I
can
go
back
to
our
post
days
and
we
have
seen
that
we
have
perfectly
done
a
recovery
disaster
between
each
s1
and
s2
and
with
a
data
migration
between
that
2
and
a
cluster
of
click
cycles.
Very
good
conclusions.
Well,
as
you
can
see,
the
recovery
disaster
is
very
direct,
very
simple,
very
fast.
C
It
takes
very
little
about
ten
seconds
between
this
recovery
in
a
lower
environment
cluster,
where
I
was
doing
tests
because
it
takes
six
minutes
to
remove
the
entire
cluster
from
between
6
and
15
minutes
integrity
in
the
data,
but
it
can
help
us
above
all
to
have
that
security.
Is
today's
backup
in
sigua
room
volume
due
to
confusion
or
because
the
volume
even
double
is
in
our
cloud
provider.
It
broke
because
nothing
happens.
We
have
the
backup
and
we
restore
and
the
ta
migration,
especially
in
those
cases
where
we
believe
today.
C
I
want
to
try
little
things
from
ws
hcp
I
want
to
see
how
it
works
and
I
want
to
explore
new
things,
because
velero
can
help
us
a
lot
in
it.
Well.
Thank
you
very
much
everyone
for
all.
This
don't
be
shy
and
you
can
consult
me
on
my
website,
twitter
and
well.
We
have
several
open
positions
in
peace
and
dotcom
where
he
worked
in
the
plans
for
engineering
team
and
well,
there
are
several
open
positions
within
the
company,
so
do
not
hesitate
to
apply.
Thank
you
very
much.
A
C
C
You
can
put
more
than
one,,
you
can
put
more
than
one
replica,,
but
the
truth
is
that
we
normally
have
it
as
we
have
it
with
a
replica
and
and
then
They
are
in
fixed
instances,
and
we
are
aware
of
whether,
after
all,.
What
matters
to
you
is
whether
that
backup
was
made
successfully
or
not,
but
hey,.
If
Vélez
has
it,.
C
Velero.
I
think
huerta
n'sue
is
fine
and
I'm
not
sure,
but
surely
it
has
in
its
old,
but
people
art
is
totally
public.
So
surely
you
can
open
an
edition
for
them
and
the
documentation
is
quite
extensive,
very
very
extensive.
They
have
many
use.
Cases
of
gtp
has
been
of
afp
vision
of
ws
there,
so
nothing
to
care
about
the
documentation
and
the
challenges
that
are
public.
C
But
it
also
offers
there
is
quite
a
lot
of
community
dashboard
in
graph
ana.
You
have
enough.
You
can
implement
rawls
commitments
to
have
all
the
ability
to
observe
sailboat
of
those
Well,.
Add
those
metrics
in
the
official
charteau
gene,.
There
is
a
metrix
parameter,,
you
add
the
average
after
the
service
mo
auditor
and
now
automatically,
it
will
come
to
your
loss
if
I
fail,,
let
them.