►
Description
en Español from OpenShift Commons Gathering Buenos Aires August 2108
Mariano Vasallo - Sistema GDE en Openshift @ Ministerio de Modernización
A
A
Well,,
we
are
going
to
tell
you
about
the
implementation
of
head
and
a
little
about
our
great
project
of
simplification
and
paper
and
zation
in
the
state,
and
how
was
the
architecture
that
we
used
for
this
first
I
I
present
I?
Am
the
director
of
a
digital
firm
I
have
been
using
Linux
for
20
years
or
so
and
well
I
had
to
collaborate,.
It
is
not
exactly
my
area
but
to
collaborate
on
issues
of
architecture
and
other
definitions
in
gender,,
which
is
a
platform
made
up
of
more
than
20
modules,.
A
Let's
say
at
the
level
of
infrastructure,
the
only
thing
that
perhaps
interests
you
is
that
there
are
different
tomcats
running
different
java
applications
that
are
quite
complex
and
very
complex
from
the
functional
point
of
view
and
in
the
interactions
that
they
have
between
them,.
That
makes
it
quite
difficult
to
manage,
but
functionally
what
it
does
is
implement
everything.
What
is
all
the
rules
that
exist
in
a
government,
all
the
legal
rules,
all
the
processes
that
allow
a
person
to
do,
or
not
an
official
to
do
or
not
certain
things
legally.
A
It
is
not
a
bpm,
but
it
is
a
platform
through
which
all
the
government
processes
of
the
entire
nation
and
of
a
lot
of
municipalities
and
state
entities
can
work,
basically
works
with
documents,
files
procedures,
good
users
and
others,
and
other
entities
are
not
doing
well.
This
formatting
has
gone
out.
A
We
are
not
going
to
talk
about
tests
here
of
concept
or
veins,
but
of
things
that
are
already
implemented,
things
with
which
paper
was
completely
eliminated
in
the
city
systems
that
are
implemented
in
the
23
ministries
and
in
other
organizations,
and
we
already
have
65
million
documents
stored.
All
with
digital
signatures
are
not
digitized,
but
they
were
born.
Digital
all
are
signed,
all
are
the
product
or
have
the
author
clearly
identified
and
have
legal
value
and
in
procedures
we
have
66
million
seven
hundred
thousand
files.
A
In
my
previous
experience,
It
has
happened
to
me
to
have
meetings
and
developments
and
entire
projects
to
carry
out
the
system
of
a
procedure
or
the
system
or
the
entire
project
to
carry
out
the
reengineering
or
computerization
of
a
process..
This
platform
is
much
freer,.
It
has
a
lot
of
internal
rules,
but
it
allowed
us
to
do.
All
of
this
in
less
than
two
years
is
made
up,
as
I
told
you,
of
these
modules
mainly,,
but
there
are
about
20
that
communicate
with
each
other,.
A
Well,
and
what
it
guarantees
is
being
able
to
computerize
the
government
complying
with
all
the
laws,
that
is,
one
cannot
install
a
harp
in
a
government
and
say
that
what
is
worth,
that
is,,
the
government
is
governed
by
files
and
documents
and
things
that
have
to
have
legal
value,.
Then
one
cannot
install
an
erp
and
say
that
a
certain
purchase
was
made
and
that
the
evidence
is
in
the
registry
of
the
database
data.
A
The
total
number
there
must
be
a
document
with
legal
value.
Argentine
law
says
that
the
document
with
legal
value
is
the
document
that
has
a
digital
signature.
So
we
achieved
these
objectives
with
this
platform,
and
what
we
say
is
that
the
most
effective
thing
is
to
speed
up
the
processes
Many
times
it
starts
with
the
part
of
reengineering
the
process,
since
something
is
found
to
be
inefficient
and
a
whole
project
is
proposed
to
do.
It.
A
We
have
an
example
of
these
dates,
or
times
are
real
for
a
customs
procedure
that
took
90
days
before
digitizing
Because.
We
say
that
speeding
up
the
key
more
than
doing
reengineering,
because
in
many
cases
you
do
not
have
the
process
and
you
can
spend
six
months
doing
the
functional
analysis,
we
debureaucratize
and
digitize,,
which
is
what
we
call
simplifying
this
process,.
We
take
60
days
with
which
you
have
left
30
and
the
effectiveness
of
the
effort
that
you
can
put
into
doing
process
reengineering,.
A
You
can
get
a
maximum
of
15
days
out
of
that,,
but
there
is
a
huge,
a
huge
step
forward
just
by
debureaucratizing
and
digitizing
and
debureaucratizing,,
which
is
removing
what
is
not
necessary,,
be
it
the
procedures
or
the
processes
or
steps
that
are
unnecessary,
the
balls
and
the
rest,,
the
digitized
ones,.
We
have
a
perfect
bpm
of
each
process,.
A
They
do
not
continue
to
work
as
now,,
since
it
is
an
open
workflow
in
which
the
processes
exist,,
they
are
in
the
work
mode
of
each
office,
of
each
one
of
them.,
the
23
ministries
and
thousands
of
departments
and
the
rules
that
governed
it
on
paper
continue
to
govern
it
digitally.
But
this
allows
you
to
make
this
type
of
improvement.
I
am
going
a
little
fast
to
try
to
leave
room
for
questions
mainly
here.
We
are
to
explain
why
we
use
docker
and
how
was
how
was
our
implementation
in
everything?
A
That
is
the
part
of
docker.
We
have
a
main
installation
that
is
this,
that
I
told
you.
Those
numbers
are
only
from
the
central
public
administration,
but
we
saw
that
we
had
to
take
it
to
a
lot
of
other
places.
That
is
have
a
lot
of
other
clients
that
in
some
cases,
what
you
would
do
is
a
multi-tenant
installation
that
allows
you
to
have
different
organizations
for
data
governance
and
separation
and
legal
issues.
There
are
also
times
when
you
need
separate
installations
when
we
consider
what
to
do.
A
We
decided
to
go
everywhere
or
for
solutions
everywhere,
because
on
the
one
hand,
we
need
to
optimize
resources,
that
is,
to
make
the
same
installation
that
used
to
occupy
20
or
40
machines.,
virtual
machines
or
20
virtual
machines
without
scalability,
that
is,,
a
single
virtual
machine
per
module
and
in
practice
we
have
eight
virtual
machines
out
of
a
module,
balancing
load,
to
manage
all
of
this,
and
to
manage
all
of
this,.
It
is
necessary
to
optimize
the
resources
of
both
the
administrators
and
both
the
physical
or
datacenter
resource
and
the
number
of
administrators.
A
In
production
running,
which
are
municipalities,
universities,
state
companies,
others
jurisdictions,
and
what
we
saw
is
a
bit
of
an
enormous
degree
of
improvement
in
these
three
steps
that
we
were
doing,,
which
is
to
go
from
doing
an
installation
with
virtual
machines,.
It
took
us
in
the
order
of
months,
that
is,
a
whole
month
to
be
able
to
deliver
the
configured
infrastructure
and
working
at
best
a
month
when
we
went
to
do
care,
which
was
the
first
thing
was
in
the
fear
all
this.
A
Let's
make
this
that
runs
on
a
tom
cut
and
someone
has
to
go
in
and
touch
an
xml
file
and
configure
it
and
so
on.
Let's
make
it
that
became
infrastructure
as
code,,
which
was
the
first
big
step
with
infrastructure.
As
code,,
we
were
able
to
replicate
the
same
thing
in
a
lot
of
places,
that
is,
to
be
able
to
have
control
so
that
the
same
ten
people
can
manage
a
hundred
installations.
A
Just
by
touching
configuration
variables,,
we
exposed
everything
which
was
external
as
configuration
variables,
environment
variables,
and
we
arrived
at
a
solution
with
two
fields
in
which
we
can
set
up
a
new
environment
in
a
matter
of
days.
One
day,
two
days,
install
a
new
environment.
Later
matías
will
tell
you
more,
but
it's
not
just
raising
the
containers.
There
are
also
dns
database
issues
to
register
in
the
external
load
balancer,
but
hey,.
A
And
that
is
let's
say
and
that's
why
we
decided
After
evaluating
several
options
and
due
to
openshift,,
we
want
to
get
to
installing
it
in
a
matter
of
hours
and
also
optimizing
the
datacenter
resources,.
We
are
basing
our
docker
solution
on
raising
the
20
containers
in
a
virtual
machine
and
in
the
case
of
having
to
scaling,
as
we
have
at
the
border
of
the
virtual
machine,,
we
could
not
break
it,.
So
with
these
installations,
based
on
what
we
can
get
to
the
greatest
power
that
we
can
give
to
a
virtual
machine.
A
A
Well,
I
can
tell
you
or
I,
am
not
going
to
tell
you
each
of
the
modules.
There
are
20,
they
do
administrative
issues,
one
is
general
access
to
the
system
from
where
one
chooses
things
another
sends
official
communications
generates
documents
in
another
groups
files
and,
as
you
can
see,
there
are
a
lot
of
other
modules
and
they
are
not
all
here.
A
A
This
is
more
or
less
the
universe
and
a
pan
of
those
who
use
it
is
the
entire
national
government,
anses
pami
and
several
ports.
All
insurers
are
centralizing
or
receiving
from
gd
facilities
of
insurers,
all
the
policies
of
all
the
insurers
in
the
country
in
theory,
that
should
lead
us
to
no
longer
having
to
present
a
policy
and
that
no
one
can
present
a
paper
that
says
it
is
the
insurance
policy
something
and
that
it
is
not
because
everything
will
be
registered.
A
Centrally,
many
municipalities
benefit
from
having
an
electronic
government
system,
an
internal
document
management
system,,
but
they
do
not
have
the
budget
to
do
it,.
They
cannot
face
a
project
of
this
magnitude,.
So
thanks
to
This
solution,
we
can
make
available
to
them
in
the
national
data
center
is
a
cloud
installation
cloud
in
quotes.
Obviously
it
is
infrastructure,
but
they
see
it
as
a
cloud
because
they
are
given
the
credentials
determined
url
they
enter
and
use
it.
We
give
them
the
service,
and
then
we
administer
it
more
precisely.
B
It
well,
we
start
with
the
model
of
the
topics
to
be
discussed.
We
are
going
to
see
the
architecture
more
or
less
of
what
is
in
a
global
way,
the
current
state
that
it
is
what
we
have
at
the
moment,
the
different
models
of
gd
implementation,
because
it
was
born
with
virtual
machines,
and
now
it
has
some
productive
environments
in
openshift
and
where
we
are
going.
B
B
B
B
B
We
currently
have
more
than
100
ecosystems,
stop
operating.
That
includes
the
public
administration
national
with
the
national
public
administration,
also
with
virtual
machines.
We
are
paying
for
anses,
we
are
paying
for
pami
and
there
are
other
little
ones
that
were
born
in
services
that
at
some
point
may
be
entered.
B
B
What?
Where
is
apn
and
before
we
just
had
to
cut
it,
because
the
infrastructure
was
no
longer
enough
to
continue
growing
and
the
production
environment
where
anses
pami
apn
is,
and
there
is
a
need
for
nation
services,
and
we
also
have
several
templates.
What
is
packaged
or
what
is
It
does
in
a
turn,
that
is,
9,
nothing
magical
and
it
is
to
be
traveling
to
the
provinces
that
want
to
implement
it
or
to
autonomous
entities
towards
the
AFIP.
At.
This
moment
it
is
with
that
template
of
that
package,.
B
As
I
mentioned,,
there
are
different
ecosystems
as
soon
as
it
is
complete,,
it
runs
in
its
twenties.
applications
with
all
the
apis.
Very
complete,
pami
anses
run
I
stay
with
core
modules
that
will
be
in
a
total
of
10
applications.
15
with
the
apis.
The
apis
are
dedicated
for
each
ecosystem.
It
is
not
shared
I
stay
here.
There
is
one
for
anses
for
penis
for
whatever,
and
also
there
is
a
thousand
that
is
the
one
that
attends
web
services
also
dedicated
to
each
one
of
the
ecosystems.
B
This
model
has
already
been
delivered
to
neuquén
córdoba,
santa
fe
catamarca
to
the
fib.
There
are
a
lot
of
provinces
that
are
missing
here,
but
not
blood,
because
obviously
there
was
no
place,
and
now
the
possibility
of
interoperating
between
different
ecosystems
is
added,,
whether
they
are
together
or
not,,
as
I
had
already
told
you,
production
ovation
and
test
have
the
pami,
anses
and
apn
that
can
share
documents
interoperating,
but
another
installation
stopped
at
some
point
could
not
do
this
based
on
documents
from
one
entity
to
another.
B
What
disadvantages
we
had
with
virtual
machines?
What
is
it
that
we
experience?
Bone
deployments
are
slow,
although
we
are
using
level
to
deploy
and
make
changes,
In
the
configuration
and
many
other
things
you
always
have
it,.
You
can
put
an
error
that
the
ssh
connection
can
take
time
because
to
give
it
an
idea,
there
are
15
virtual
machines
only
for
gedo,
and
that
takes
time
if
you
deploy
the
guard
one
by
one
and
so
on.
B
B
Do
not
have
auto
scaling,,
so
in
larger
virtual
machines
it
could
be
dumped
because
at
one
point
it
is
bad
for
a
while.
button
was
deployed
incorrectly.
Some
module
did
not
start
and
all
crashed,
that
is,
also
with
the
configuration
of
that
module,.
Perhaps
in
jim's
template,,
something
was
already
wrong
and
you
don't
realize
it
until
someone
reports
it
to
you,
because
there
are
1
or
2.
B
B
B
B
B
B
Advantages:
easy
scale,
independent
ecosystems,
better
implementation
time.
There
is
another
issue
that
of
independent
ecosystems
is
very
important,
because
if
I
deliver
something
to
vicente
lópez
vicente
lópez,
wants
me
not.
Nothing
is
shared
with
another,
that
is,
with
another
with
another
ecosystem,
that
is,,
that
the
different
ones
are
saved
in
a
different
place
in
another
state,,
and
we
improve
the
implementation
and
deployment
times,
that
is,
before
it
was
yes,
when
we
confirmed
for
me
that
it
was
an
appendix
of
what
It
is
apn,
we
were
a
lot
of
time.
B
The
disadvantages
is
a
complexity
in
the
configuration
it
is
a
jam,
the
immense
compost
and
when
the
first
andoian
boys
started
I
could
not
do
it
I
had
it.
What
to
do.
Someone
else
pass
this
knowledge
on
to
someone
who
is
not
used
to
everywhere.
Jamel,
who
did
not
understand
where
to
go,
did
not
understand
that
you
had
to
edit
80%
of
the
birds
are
oversized.
They
have
a
lot
of
disk
many
cpus
a
lot
of
memory
and
they
are
not
using
it
go
one
by
one.
Seeing
what
I
need
that
doesn't
also
get
complicated,.
B
Well,
the
openshift
model,
the
prince's
idea-
was
to
be
able
to
automate
everything
from
the
network
configuration,,
which
is
where
it
is
up,.
We
set
it
up
on
top
of
a
network,
we
wanted
to
up
openstack,
but
due
to
the
times
of
the
state,
we
could
not,
because
you
saw
that
the
cart
is
always
ahead
in
the
state,,
so
you
are
running
behind
the
cart,
accommodating
the
thing
as
power,
so
we
use
network
for
this,
and
for
the
moment
we
have,
we
are
using
nfs
for
the
registration
and
for
persistent
volumes.
B
B
And
from
the
mat
that
it
has
low,
that
is,
the
test
approval
environments,
where
they
are
testing
that
we
are
going
to
change
this,,
in
fact,
we
do
not
have
a
production
course,,
we
are
going
to
put
all
the
same
bag
and
we
are
going
to
work
with
what
we
have,
because
it
also
makes
sense.
What
sense
does
it
make
that
actually,
an
ecosystem
that
we
deliver
is
testing
until
we
put
the
certificate
with
which
it
will
sign
the
document?
B
Officially,
then
there
is
a
flag
that
says:
okay,
this
is
testing
until
what
happens
this,
then
it
does
not
make
sense
to
do
what
we
are
considering.
This
is
what
we
are
going
to
do
is
it
does
not
make
sense
if
we
already
have
an
ecosystem
before
it
is
in
the
test
data
center
to
pass
it
to
production,
because
that
is,
it
is
productive
for
only
one
flan
for
only
one
smallness,
it
was
our
first
approach.
B
Since
we
configured
everything
that
has
to
do
with
the
network,
everything
was
test
rice
error.
Until
the
hit
came
out,
everything
could
be
replicated
again.
That
is
when
it
arrived
and
it
was
running
the
bulletin.
We
did
again,
it
came
out
great
perfect.
That's
what
The
deployment
of
the
bmw
openshift
will
remain,
that
is,
of
all
the
application
nodes,,
the
new
master
node
and
the
infra
node,
everything
is
automated..
We.
B
B
The
model
persists
offensive
system.
Well,
this
means
we
already
have
three
ecosystems
in
the
low
openshift
cluster.
As
I
told
you.
Maybe
we
won't
use
the
low
anymore.
It
is
going
to
start
doing
all
production.
This
is
not
yet
with
the
flash
that
tells
you
that
they
are
productive,
but
they
are
still
being
approved
by
the
modernization
team.
Functional
each
ecosystem
is
a
different
project
without
multi
tinning.
That
is
this,
and
people
told
us
that
it
is
already
100%
productive,.
It
is
already
signed,
documents
are
created,.
B
Files
can
be
interoperated,
that
is,
it
is
complete,,
it
is
ready
to
set
the
flag
and
it
comes
out
in
production,,
as
we
are
doing
the
deployment
of
the
ecosystems,.
We
use
a
giant
conflict,.
The
strategy
is
a
doctor
file
that
It
is
in
an
internal
kit
of
ours
and
when
we
put
the
bills
and
others
we
import
the
configuration
with
that
in
two
hours
we
already
have
an
ecosystem
running,.
It
is
necessary
beforehand
that
we
are
available
in
the
virtual
machine
that
makes
it
ours,.
It
is
seen
in
generally
record
time.