►
From YouTube: GitLab 13.2 Kickoff - Enablement:Geo
Description
Fabian Zimmer (Product Manager, Geo) highlights new features and direction for GitLab Geo in 13.2.
Roadmap: https://gitlab.com/groups/gitlab-org/-/roadmap?scope=all&utf8=%E2%9C%93&state=opened&label_name[]=group%3A%3Ageo
A
Everyone,
this
is
Fabian
the
product
manager
for
Catholic
geo,
and
this
is
the
kidnap
trio
kickoff
video
for
gitlab
13.2.
The
next
upcoming
release,
I'll
walk
you
through
the
geo
roadmap
thematic
overview
and
the
iterations
that
we're
going
to
ship
and
address
in
13.2.
So
let
me
quickly
share
my
screen.
A
So
the
first
thing
is,
if
you're
interested
in
the
product
direction
for
deal
on
a
high
level.
Please
have
a
look
at
the
direction
page.
We
keep
them
updated
so
for
all
the
categories
that
are
part
of
the
GEO
Group,
do
your
application,
dr
and
backup
and
restore,
but
before
I,
go
into
the
specifics
of
13.2.
A
They
are
essentially
two
main
initiatives
that
you've
heard
me
talk
about
a
number
of
times
in
the
past.
The
first
one
is
really
to
add
replication
capabilities
to
geo
for
the
data
that
is
not
currently
replicated
by
geo.
This
is
really
important,
because
if
we
don't
replicate
data
from
a
primary
to
a
secondary,
you
can't
easily
recover
it.
A
We
need
to
go
to
a
backup
so
for
disaster
recovery
replicating
everything
is
really
important
and
there's
a
major
initiative
to
centrally
standardize
a
framework
that
is
scalable
and
allows
other
teams
we
didn't
get
or
the
wide-legged
lab
community
to
add
data
to
get
lab
geo.
So
this
is
one
of
the
main
two
things
that
we're
doing
right
now.
A
Because
if
we're
not
replicating
data
fast,
there
will
be
more
lag
and
making
sure
that
geo
is
scalable.
Performant
even
for
very
large
installations
is
really
important,
and
so
we
found
some
work
in
that
area.
So
these
are
the
the
high-level
themes
here
that
are
happening
and
a
lot
of
the
activities
that
you'll
see
me
discuss
in
more
detail
fit
into
those
there's,
obviously
a
lot
more
that
we
all
aim
to
accomplish
in
the
future,
for
example,
complete
maturity
for
the
AR
and
complete
maturity
for
geo
replication.
A
A
Two,
which
is
geo,
will
add,
capabilities
to
support
three
more
datatypes
terraform
state,
vulnerability,
exports
and
external
write
requests,
all
of
those
are
files
or
blogs,
and
they
will
benefit
from
our
framework,
making
it
much
easier
to
implement
those,
and
we
hope
that
we
can
really
iterate
a
lot
faster.
Now,
that's
sort
of
the
framework
itself
is
becoming
more
stable.
So
that's
number
one
I've
opened
these
issues
here
so.
A
Actually,
I
didn't,
but
that's
fine.
These
are
the
three
main
things
here.
Then
there
are
a
few
bits
that
we
need
to
wrap
up
for
our
scalability
improvements
for
thirteen
point
two.
This
will
actually
result
in
us
removing
a
part
of
deal
by
thirteen
point.
Two
has
got
a
foreign
data
rappin
for
post
rest,
which
will
make
you
a
lot
easier
to
maintain
an
upgrade
and
also
you
know
the
performance
will
be
impacted,
so
quite
excited
to
bring
those
stale
ability
improvements
to
a
close
in
thirteen
point
one.
You
can
see
this
here
as
well.
A
We
work
on
specific
parts
of
the
front
and
capabilities.
So
what
will
be
shown
to
systems
administrators
trying
to
understand
the
state
of
the
replication
and
this
year
is
really
about
some
technical
debt
that
and
making
sure
that
the
framework
itself
also
works
well
with
the
front-end
code,
so
that
will
probably
take
a
little
bit
longer.
There
are
a
few
smaller
items
in
here
that
will
be
shipped
in
thirteen
one
to
another
bit.
This
sort
of
happening
a
little
bit
more
in
the
background
in
small
iterations,
is
to
add
verification
for
package
files.
A
So
this
will
take
a
little
while
longer,
but
we
are
addressing
specific
issues
here
in
13.2
as
well
to
ensure
that
verification
becomes
a
sort
of
included,
out-of-the-box
part
of
the
self-service
framework.
This
is
important
because
we
replicate
data
right
now,
especially
files,
but
we
don't
automatically
verify
that
all
of
the
data
that
was
transferred
was
actually
not
corrupted,
and
so
there
are
some
situations
in
which
that
is
obviously
a
problem.
A
If
you're
trying
to
like
access
data-
and
you
know
maybe
hard
disk
failed
or
the
file
system
had
a
problem,
and
so
we
are
aiming
to
verification
as
a
part
of
the
framework,
which
should
then
also
mean
that,
for
example,
these
items
here,
which
are
firstly
focus
on
replication,
then
later
on,
we'll
come
with
verification
out
of
the
box
with
very
little
effort.
So
these
are
the
parts
and
13.2
would
that
have
to
do
mainly
with
the
self-service
framework
and
some
of
the
scalability
work.
A
There's
a
second
part
of
this,
which
is
focusing
on
the
improvements
for
planned
failures,
and
so
in
13.1.
We
hope
to
finish
a
feature
that
makes
it
easy
to
pause
and
resume
replication,
and
so
hopefully
this
will
close
relatively
soon
and
ship
in
time,
which
makes
it
easier
to
essentially
decouple
a
primary
and
a
secondary
and
we'll
also
have
a
little
bit
of
discovery
left
to
do
on
the
failure
of
processes.
A
But
then-
and
you
heard
me
talk
about
that
in
detail
before
we
are
or
have
been
working
for
a
little
while
on
creating
an
intense
mode
and
actually
put
it
off
a
little
bit
more
in
favor
of
delivering
some
smaller
iterations
in
the
self-service
framework.
But
we've
broken
this
down
a
little
bit.
More
I
can
actually
show
you
this.
A
So
there
are
quite
a
few
very
granular
items
here
for
the
maintenance
mode
that
we
aim
to
ship
in
thirteen
point
one
and
then
there's
a
few
more
items
in
thirteen
point
two,
but
we
enter
to
pay
that
there
is
a
little
bit
more
testing
involved
to
actually
ship
the
entire
functionality.
So
that's
another
focus
for
for
this
release.
A
Another
item
that
sort
of
spun
out
of
interviews
that
we
had
with
our
systems
administrators,
is
that
the
overall
documentation
sometimes
is
a
little
bit
difficult
to
follow,
and
so
we
start
on
simplifying
the
documentation,
maybe
adopting
more
of
a
run
book
style
for
documentation
and
so
forth.
Thirteen
point
one
four,
thirteen
point
two:
that
is
sorry
we
are
going
to
actually
rewrite
some
of
the
documentation
in
more
foreign
book
style.
A
So
that's
probably
going
to
take
us
into
the
release
afterwards
as
well,
and
then.
Finally,
there
are
a
couple
of
items
that
I
also
wanted
to
discuss.
That
are
important,
which
is.
We
are
also
continuing
to
improve
the
user
experience
for
our
systems,
administrators
and
what
this
means
in
13.2
concretely
is.
We
are
going
to
work
on
the
data.
I
need
to
actually
find
it
have
a
look
sorry
on
quite
a
few
small
items.
One
that
is
actually
important
is
forms.
A
So
here
the
forms
that
are
on
the
front
end
for
editing
the
nodes,
and
we
will
make
sure
that
those
really
are
a
little
bit
easier
to
manage
for
assistance,
administrators
and
the
last
thing
which
is
going
to
be
a
pattern
for
the
next
releases.
Is
we
really
need
to
improve
the
promotion
process
itself,
especially
for
larger
installations?
So
we
are
working
towards
sort
of
a
single
command
that
can
be
run
to
promote
an
entire
multi
server
installation
of
a
skate
lab,
and
then
13.2
will
start
with
the
POC.
A
So
this
is
essentially
a
proof
of
concept
to
understand
how
we
could
actually
sort
of
alter
the
configuration
of
nodes
automatically
so
that
if
you
have
a
secondary
site
that
you
want
to
promote
to
a
a
primary
systems,
administrators
can
actually
run
a
single
command.
But
there
is
a
little
bit
of
technical
complexity
involved
and
we
will
start
very
small
to
understand
the
technical
implications
and
then
ship
it
route
is
improvements
along
the
way.
So
that's
another
item
that
is
coming.
A
So
this
is
the
road
map
view,
there's
also
planning
if
you
where
you
can
look
at
the
individual
like
pieces
that
are
important
and
another
thing
that
I
didn't
mention.
Yet
that
is
something
that
will
need
to
happen
is
get
letting
out
ships
with
Patroni,
which
is
a
template
for
the
fault,
tolerant,
post
quiz
and
we're
also
going
to
do
a
little
bit
of
work
to
verify
that
giro
works
well
with
Patroni
and
the
latest
version
of
post
quiz.
A
So
that's
it,
as
you
can
see
this
quite
a
few
moving
parts
now
that
you
have
a
visualization
of
all
of
the
epics,
but
we
are
trying
to
make
these
improvements.
Iteratively
and
I'll
keep
you
posted
as
to
the
progress.
So
thanks
very
much
for
watching
this,
video
and
I
hope
to
see
you
again
next
month.
Bye.