►
From YouTube: Tomas Jelinek: The journey of a virtual machine from VMware to oVirt and how it can be customiz..
Description
virt-v2v is a tool to convert virtual machines from foreign guests to KVM. oVirt is a datacenter virtualization solution managing KVM guests. ManageIQ is a cloud management platform managing both VMware and oVirt (and many others).<br>We have integrated all this to provide an easy migration of VMware guests to oVirt using ManageIQ. <br>In this talk all layers of this integration will be explained giving the audience a firm understanding of how the journey of the virtual machine from VMware to oVirt looks like and how it can be customized.<br>No prior experience from the audience expected.
A
You
can
be
quite
customized,
so
I
will
explain.
Why
would
you
do
that?
And
how
would
you
do
that
and
everything
which
I
will
be
speaking
about
up
until
this
point
has
already
been
done
and
released,
so
you
can
use
it,
but
there
is
some
room
for
improvement,
and
this
is
what
I
will
be
touching
in
the
future
section.
So
first
thing,
first
vertie
to
be
how
many
of
you
know
what
were
three
tweaks?
A
A
A
On
the
right
hand,
side
there
is
some
storage
domain
which
can
be
some
I,
scuzzy
or
NFS
or
whatever
whether
converted
disks
will
go
and
in
the
middle
there
is
some
Linux
books.
I
mean
it
vertically
to
be
installed.
In
this
case,
it
happens
to
be
an
over
hypervisor,
but
it
doesn't
have
has
to
be
so
when
Berkeley
2d
is
asked
to
convert
some
virtual
machine
first
thing
it
does
that
it
asks
the
the
under
API
using
deferred
about
the
metadata
of
the
virtual
machine.
A
These
are
just
pin
overlays,
so
they
need
some
packing,
so
the
backing
is
set
to
the
actual
disks
which
are
still
on
the
VMware
side.
Now
it
needs
to
access
the
guest
filesystem
of
those
images
to
do
some
modifications
inside
them.
For
this
it
boots
up
a
temporal
virtual
machine.
It
is
not
the
virtual
machine.
We
are
converting
Yeti,
just
some
just
some
virtual
machine
machine
to
used
to
access
the
guest
file
system
boot
it
up
attaches
the
and
around
some
modification
inside.
A
So
what
modification
it
runs
depends
heavily
on
what
is
installed
on
that
image.
Actually,
for
example,
if
it
is
Windows,
then
it
copies
the
real
tie
off
drivers
in
that
on
that
image
of
the
transistor
and
modifies
the
windows
in
a
way
that
on
the
first
blue,
it
will
actually
install
those
drivers
when
the
image
is
like
death
is
repaired
and
demise
and
everything
then
it's
time
to
actually
copy
the
discs.
A
A
A
A
This
is
how
it
looks
like
so
it
has
an
engine,
something
we
call
an
engine.
It
is
a
centralized
service
which
contains
all
the
business
logic
and
expose
an
API
and
some
front-end.
It
divides
the
resources
into
data
centers,
which
are
datacenters
more
clusters
of
hypervisors
in
this
hypervisors,
because
some
Linux
boxes
it's
on
services.
It
allows
the
virtual
machines
so
how
this
fits
into
the
picture
of
of
this
vert
v2v.
A
A
This
is
how
so
the
bottom
part
of
the
image
is
simplified.
Version
of
the
image
I
have
shown
before
it's
a
over
hypervisor,
the
sauce
vm
/
and
the
aqua
story,
which
is
managed
by
overt
on
above
that
you
see
the
ovirt-engine
is
that
centralized
logic
and
its
database
where
the
metadata
are
stored
so
overt.
In
order
to
have
a
virtual
machine,
it
can
manage
it.
It
needs
two
things:
it
means
the
metadata.
It
needs
the
disks
of
it.
A
A
After
this
it
needs
the
discs
again
picks
a
random
hypervisor
and
contacts
the
vertically
to
be
on
it
and
cause
it
really
cause
a
command
line
that
convert
the
district
that
Vav
and
REM
and
put
the
disks
over
there
when
it's
done
or
it
has,
the
metadata
has
the
disk.
So
it's
happy
and
it
can
manage
the
virtual
machine.
The
story
ends
for
over
in
this
stage,
so
the
last
component
in
this
in
this
story
is
the
managed
IQ.
A
So
how
many
of
you
guys
know
with
many
Jake
is
down
you
work
with
the
VM
anyway,
so
we
manage
IQ
main
Jack
is
the
manager
of
the
measures.
It
is
a
cloud
management
platform
which
means
that
it
provides
some
centralized
management
logic
for
lots
of
different
services
like
clouds
and
visualizations
and
so
on.
So
what
it
actually
does
is
that
at
the
bottom
of
the
image
you
have,
you
see
some
grey
boxes.
I
will
call
them
here
for
party
services,
because
I
like
that
term,
it
can
be,
for
example,
the
wall
over.
A
It
is
such
a
perk
deserves
on
albey
ember
or
wool
OpenStack
or
opener
shift,
or
whatever
lots
lots
of
different
services.
Now
major
ike,
you
can
talk
to
all
of
these
api's
understand.
All
of
those
api's
takes
all
the
data
from
all
of
those
to
party
services
stores
them
in
his
own
unified
form
in
his
own
database.
So
now
it
has
a
giant
database,
readiness
and
holistic
view
of
lots
of
environments
and
provides
different
services.
A
On
top
of
them,
like
I'm,
not
going
to
go
over
all
week
to
check
them
out
later,
I
will
mention
all
which
are
important
from
our
perspective.
From
the
perspective
of
just
of
this
tall,
one
is
the
UI,
so
you
have
one
unified
user
experience
over
completely
different
third-party
services.
Let's
say,
and
the
second
is
automatic
called
automate
what
it
actually
is.
A
This
is
what
we
use
in
manage
IQ.
This
automate
provides
the
food
for
both
the
front
end
and
the
back
end
for
these
plugins
and
the
back
end
is
organized
as
a
state
machine,
so
you
implement
the
states
of
the
state
machine
and
also
the
state
transitions,
so
how
the
state
machine
is
going
to
work.
So
this
is
what
we
used.
This
is
how
it
ended
up.
So
in
the.
A
A
So
this
is
awkward
engine
I
was
talking
about
above
that,
like
top
part
of
the
picture,
you
have
the
core
component,
some
core
components
of
major
IQ.
The
database
is
the
giant
data
database.
I
was
talking
about
before,
so
it
contains
all
the
data
from
everything,
and
there
is
some
component
for
the
sake
of
simplicity,
I
call
it
here,
refresher.
It
is
something
which
make
sure
that
the
database
is
always
up
to
date
in
between
there
is
the
promise
state
machine.
A
A
Then
a
word
starts
doing
all
of
its
magic
like
taking
the
metadata
contact
inwardly,
to
be
to
start
copying
to
disk,
and
so
on.
This
takes
time
so,
the
next
day
stage,
he
is
just
waiting.
This
state
wakes
up
every
three
minutes
checks
the
database.
It
doesn't
check
the
whole
API,
though
it
checks
its
own
database
checks
its
own
database,
while
the
Refresh
will
make
sure
it's
up
to
date
and
waits
until
about
the
migration.
It's
actually
finished
when
it
finished
the
last
state
in
this
state
machine
is
the
configure
networks.
A
This
is
a
post
modulation
state,
which
means
that
at
this
point,
the
virtual
machine
is
already
in
Albert
successfully.
It
has
the
metadata
it
has
the
disks,
you
can
start
it,
you
can
use
it,
but
it
doesn't
have
networking
set
because
over
takes
care
also
of
the
work
network
of
the
virtual
machine.
So
what
is
needed
here
is
to,
but
what
the
state
does
is.
A
It
looks
at
the
virtual
machine
checks
what
network
interfaces
this
virtual
machine
does
have
and
plot
a
virtual
network
over
twitch
or
network
into
each
of
the
network
interfaces,
and
this
is
80
Madrid.
This
is
what
is
done
by
default,
and
this
what's
it
works
with
the
typical
case.
As
we
all
know,
the
typical
cases
are
very
rare,
so
everyone
needs
to
do
some
customization
something
special
somewhere,
especially
if
you
are
talking
about
pet
virtual
machines,
which
is
over
this
for
pet
virtual
machines,
then
you
need
to
do
something
special
all
the
time.
A
So,
for
example,
just
at
least
I,
don't
know,
send
an
email
that
you
will
have
an
outage
or
around
some
unstable
script
in
the
source
before
it's
migrated
to
let
to
prepare
it
for
being
moved
to
some
other
place
or
I.
Don't
know
after
successful
migration
rename
the
sauce
VM
to
something
like
migrated,
a
good
ever,
and
these
are
all
things
which
we
cannot
really
see
paid,
and
this
is
why
we
have
used
the
automate
to
implement
this
migration
logic,
and
this
is
how
you
can
modify
it.
A
So,
as
I
said,
the
automate
scripts,
not
only
the
UI
and
functionalities
exposed
to
the
user,
but
also
the
implementation.
So
this
is
a
stress
interest
from
the
major
key
UI
weather,
where
you
can
see
how
the
how
this
paid
machine
I
was
just
talking
about,
looks
like
health.
The
three
states
I
was
mentioning
the
launch
rate
for
import
and
configure
networks.
You
can
go
ahead
and
add
your
own
state
in
between
before,
after
wherever
you
want
and
implement
the
content
of
the
state,
as
you
wish,
you
can
implement
whatever
you
want
there.
A
A
So
this
is
what
we
have
done.
This
is
or
released
and
I'm
working.
So
if
you
now
go
ahead
and
install
the
latest
releases
of
all
these
components,
you
can
use
it,
but
there
is
no
room
for
improvement
and
I
have
divided
it
into
two,
the
high
and
the
low
level,
let's
type
in
the
high
level,
which
is
in
this
case
the
manager
IQ.
A
So,
as
you
remember,
in
manage
IQ
I
picked
one
particular
VM
and
told
it
go,
go
ahead
and
baguette
it
over
there.
This
is
good.
If
you
have
a
couple
of
beers,
you
want
to
migrate.
It's
not
so
great.
If
you
have
a
couple
of
thousands
of
VMs,
you
want
to
migrate,
it
will
be
much
better
to
have
some
abstract
entity.
I
call
it
here,
migration
group,
you
can
call
it.
However,
you
like,
which
just
is
some
packet
of
virtual
machines.
A
You
want
to
make
it
in
one
group,
ideally
in
batches,
but
now
you're
migrating
this
bunch
of
beers
together,
you
don't
necessarily
want
to
make
it
all
of
them
to
one
cluster
in
overt.
Well,
you
don't
really
don't
always
want
to
channel
all
the
disks
of
order,
virtual
machines
to
one
storage
of
on
the
over
side.
What
you
want
to
do
instead
is
to
do
some
kind
of
research
resource
snapping.
A
So
you
have
these
two
abstract
entities,
you
join
them
together
and
then
you
have
some
extra
migration
on
how
to
call
it
something
you
can
submit
to
the
manager
infrastructure,
and
it
then
takes
care
of,
for
example,
magnetic
in
batches
in
during
some
aggression
windows,
something
like
that
that
little
to
from
the
low-low
perspective,
one
one
thing
overt
novices
vertie
to
be,
as
is
so.
It
is
really
using
the
command-line
interface
of
the
very
chewy
and
word
feet,
which
is
installed
on
every
hypervisor
of
the
overt.
A
So
overt
is
pretty
much
now,
acting
as
a
mediator
between
major
IQ
and
and
were
creatively.
It
would
be
so
much
better
if
we
or
someone
would
contributing
to
work
b2b
itself
to
some
scripts
around
doesn't
even
matter
the
knowledge
of
the
earth
api,
so
the
overt
suddenly
be
passive
in
this
role,
it
will
be
only
the
receiver
of
the
of
the
incoming
data
and
work
with
to.
We
would
use
the
overts
upload
API
to
upload
the
disk,
use
the
api
over
to
create
the
metadata
and
connect
and
fit
together.
A
This
has
couple
of
advantages,
run
less
work
for
overt
simple
code
and
everything.
Second,
you
don't
necessarily
want
to
have
the
logic
of
migration,
not
just
logic,
but
the
load
of
the
migration
on
a
hypervisor
you
are
using
for
the
core,
virtualization
and
first
possibly
the
most
important.
Then
you
could
leave
the
same
logic
or
very
similar
logic,
to
make
it
to
something
else,
for
example
to
OpenStack.
A
So
this
is
the
one
and
ii
like
an
announcement
on
this
lower
loyalties,
but
currently,
when
the
disks
of
the
virtual
machines
are
copied
over
from
vm
virtu
over
then
HTTP
see
was
a
protocol
to
get
the
BS
from
VMware,
which
is
not
ideal
berkeley
to
be
support,
also
better
ones.
Ssi
saanddk
Mallika
is
a
proprietary
tool
from
from
Denver
which
are
both
much
faster
and
between
each
other.
They
have
advantages
and
disadvantages.
A
We
just
need
to
figure
out
which
one
do
we
want
to
use
and
how
to
integrate
it
in
a
good
way.
So
it
will
be
like
easy
to
use
for
the
user.
So
this
pretty
much
concludes
the
also
the
lower
future
plans.
The
important
point
here
is
that
these
are
not
some
visual
thinking,
so
the
next
slide
is
not
going
to
be
touches.
Welcome,
but
we're
really
working
on
on
these
items
now
and
should
arrive
in
one
of
the
upcoming
releases.
A
The
storages-
and
so
the
question
was
if
there
is
some
work
being
done
or
not,
probably
the
virtual
machines
conversion,
but
also
all
the
infrastructure.
The
VMware
is
providing
so
short
lesson.
No
now
yeah
we
are
expecting
the
user
to
configure
over
to
to
configure
all
the
resources
around
and
then
just
to
migrate.
The
virtual
machines
into
there.