►
Description
KCD Italy 2021* - The video of the talks are now available!
On this talk we'll see how to do a disaster recovery from our dearest Kubernetes cluster and workloads. We'll review how Velero can help us on that and we'll show some hands on recovering a oh shit moment in our Kubernetes cluster.
🤓 SPEAKER: Ramiro Alvarez Fernandez, Platform Engineering Tech Lead @ Empathy.co
👉🏼 Here the slide deck: https://www.slideshare.net/KCDItaly/kubernetes-backup-and-migration-strategies-with-velero-ramiro-alvarez-fernandez
______________________
* KCD Italy 2021 is the first edition of a community-organized non-profit event that will take place on November 17th and 18th and will gather adopters and technologists from open source and cloud native communities for education, collaboration, and networking.
📯 Follow us to be notified about the upcoming MeetUps: https://community.cncf.io/kubernetes-cloud-native-milano/
A
A
A
B
Thank
you.
I
hope
you
are
enjoying
this
this
event
thanks
for
coming
and
thank
you
organization
to
make
it
happen.
Kubernetes
is
a
very
great
place
to
run
all
your
workloads,
but
once
you
start
running
your
nets,
application
in
brazil,
some
questions
come
to
your
mind.
For
instance,
do
we
have
any
kind
of
backup
of
the
cluster?
Can
we
recover
from
oh
moment?
B
B
Okay,
with
today
agenda,
we
will
be
talking
about
the
letter
use
cases.
We
are
going
to
see
the
components
of
balero,
how
backup
restore
the
data
check.
Genetic
workflow
and
the
interesting
part
part
the
showtime,
the
demo.
We
are
going
to
see
a
live
demo
to
proceed
with
how
to
proceed
with
a
disaster
recovery
and
a
data
migration.
We
see
if
we
can
do
it
successfully
or
we
are
going
to
fail
and,
at
the
end,
some
brilliant
conclusions.
B
B
The
data,
integrity
or
data
protection
valeria
allows
schedule
backups
to
protect
our
details,
given
as
workloads
and
data
and
protect
from
data
loss
data
corruption.
So
human
error
are
accomplished
with
compliant
requirements.
These
kinds
of
things-
and
the
last
use
case,
is
the
data
migration.
We
can
validate
portability
because
allow
us
to
manipulate
the
data
across
cloud
providers,
for
instance
for
aws
from
aws
to
gcp
and
allow
us
to
rate
rate
or
a
production
cluster
with
application
data
cluster
to
other
lower
environments.
B
Besides,
if
you
need
a
cross
cloud
provider
across
interaction
in
between
aw
and
gcp
between
clouds,
you
need
an
agnostic
file
system
level.
Therefore,
this
is
where
rustic
entering
action.
You
need
to
deploy
a
rustic
diamond
set
in
your
kubernetes
cluster.
To
has
this
agnosticism
liar
from
a
file
system
level
the
what
velero
can
do
with
the
data?
Okay
interacts
virusly
with
the
kubernetes
api
discovery,
therefore,
that
you
don't
need
to
talk
directly
to
etcd.
B
The
song
key
features
of
balero
are
the
flexibility.
As
I
comment
before,
you
can
schedule
backups
and
you
can
have
fine
access.
Granularity
filtering
by
namespace
by
resource
by
the
levels
on
your
thoughts
whatever
and
extensibility
like
belero,
has
a
multiple
plugins
from
the
community
to
make
it
work
for
with
uploading
with
aws
and
interact
with
s3
and
the
snapshot
api
for
aws
or
gcp.
B
B
Maybe
we
don't
need
to
restore
every
day,
but
when,
in
a
case
of
data
loss
or
disaster
or
whatever,
you
need
to
do
a
restaurant
the
same
way
than
before,
we
are
going
to
hit
balero
typing
our
valero
command
line,
okay,
resto
create
restore,
and
this
is
going
to
interact
with
the
valero
server
controller.
We
are
going
to
pull
the
data
from
the
cloud
object,
storage
and
deploy
all
the
network
resources
in
the
kubernetes
through
the
kubernetes
api.
B
B
We
are
going
to
to
reveal
a
couple
of
use
cases
the
disaster
recovery
and
for
the
disaster
recovery.
We
are
going
to
have
a
couple
of
aws
guys
clusters
because
cluster
one
where
we
are
going
to
proceed
with
hellacap
and
kubernetes
ikea's
kidness
cluster
2,
when
we
are
going
to
simulate
a
restore
for
the
for
the
other
demo,
we
are
going
to
do
the
data
migration
from
cluster
2
in
aws
to
igi.
B
B
B
B
B
B
B
I'm
going
to
hit
it
now,
we'll
we
are
doing
a
valero
backup
created
with
the
name
cluster
one
only
for
the
namespace
host
and
we
are
going
to
wait
until
the
backup
is
completed.
If
we
would
like
to
check
the
details,
we
can
ask
details,
and
this
is
going
to
show
us
how
was
the
backup?
Okay,
the
backup
has
been
completed,
no
errors.
The
nsp
has
included
a
ttl
of
one
month.
B
It
spends
like
two
seconds
and,
as
you
see,
there
are
a
lot
of
things
in
the
kubernetes
api
that
we
back
up
the
config
map,
the
secrets.
The
service
service
account
the
persistent
volume
itself
you
can
check,
but
system
volume.
The
snapshot
has
been
done.
If
we
go
to
the
aws
console,
you
can
check,
we
can
copy
this.
B
B
B
C
B
This
is
up
and
running.
The
database
is
up
and
running.
Oh
wait:
okay,
that's
it!
We
can
do
a
port
forward
again
and
I
hope
we
refresh
okay,
the
image
from
from
cluster
one
is
that,
therefore,
we
complete
this
scenario
where
we
give
our
backup
from
cluster
1
and
with
our
restore
from
cluster
2,
that's
nice,
but
we
would
like.
C
C
B
And
we
are
going
to
publish
it.
Okay,
we
are
going
to
check
if
okay
collector
1
cluster
2,
that's
it.
As
I
said
before,
we
are
going
to
proceed
with
a
data
migration
between
aws
to
gcp.
Therefore,
the
snapshot
api
from
aws
can't
communicate,
can't
interact
or
they
don't
know
how
to
interact
with
the
gcp
snapshot
api.
Therefore,
we
need
a.
B
B
C
B
B
B
B
B
C
B
B
B
B
B
B
B
B
D
B
B
This
is
my
magic
very
easy,
as
the
forward
very
recommended,
any
bad
experience
with
a
stateful
set
well
a
lot
of
because
sometimes
we
have
some
data,
corruption
or,
and
you
lost
one
pvc,
because
whatever
I
think
that
valero
can
help
us
to
be
completely
sure
to
have
this
kind
of
network
of
this
kind
of
security
to
okay.
It
doesn't
it's
not
a
mess.
We
can
restore
our
account
a
previous
backup
and
that's
it.