►
From YouTube: Geo HA cluster upgrade walkthrough
Description
Douglas walks through an upgrade of a Geo cluster in high availability configuration from GitLab 12.0 to 12.1
B
Great
first,
thank
you
all
for
joining
this
call.
The
idea
here
is
to
try
to
follow
the
upgrade
process
for
a
joke:
--will
starina,
AJ
environment.
Basically,
what
I
did
I
set
up
a
cluster
primary
foulest,
your
own
United
States
and
a
joke
was
very
own,
the
Netherlands,
so
you
try
to
go
through
the
documentation
and
follow
these
steps.
And,
let's
see
what
happens,
if
you
can,
you
should
upgrade
process
I'm
sure
my
screen.
B
B
This
are
the
machines
that
we
are
going
to
try
to
update
the
post
resistance.
The
right
is
to
get
Lee
in
the
to
application
servers
and
set
up
the
you
have
a
verb
list.
I
followed
these
instructions,
it's
important
to
note
that
they
are
most
alter,
dated
and
missing.
Some
steps.
I
had
some
notes
here.
I
try
to
create
issues,
and
we
did
this
steps
later
today.
B
B
B
B
B
A
B
A
B
And
something
that
we
need
to
start
to
think
about,
for
example,
for
these
demos.
It's
easy
for
me
to
try
to
follow
the
steps
here,
but
as
for
example,
you
have
customers
that
are
going
to
upgrade
their
environments,
but
they
have
some
people
using
the
your
instances.
So
we
have
some
traffic.
So
this
is
important
to
take
an
account.
B
B
A
A
H,
a
H,
a
stands
for
high
availability,
so
in
general
that
just
means
like
within
within
a
single
instance
of
gitlab.
You
want
to
make
sure
that
there's
no
single
point
of
failure,
so
you
break
out
all
the
different
services
that
are
required
to
support
the
get
lab
instance,
and
you
have
multiple
of
each
service.
You
know
running,
and
so
that's
just
within
one
cluster
geo.
B
Thank
you,
okay,
anyways
yeah,
thanks
Mike
I,
find
my
fondest
is
to
consider
further,
which
knowledge
AJ
deployment
so
basically
are.
The
different
here
is
that
we
random
regressions
before
it,
but
they
do
all
the
other
nodes
here
they
suggest
we
did.
The
select
package
in
the
firt
on
the
apply
node
then
update
all
the
other
nodes,
then
around
the
migrations.
B
B
B
We
have
some
rake
tasks,
each
Labs
CTL
that
it
can
check
the
status
of
our
instance.
For
example,
you
know
that
other
services
are
running.
They
have
the
gear
which,
like
rails,
that
you
can
run
rails,
console
for
example,
and
we
also
have
the
lab-
take
command
that
you
can
run
and
rake
tasks
like
database
migrations.
B
B
B
B
Okay,
detect
it
because
I
I'm
gonna
know
what
is
running
each
lab
child
at
1.4
and
our
second
a
closer
running
Jeff
child
at
all
that
farm
and
is
another
warning,
because
we
have
more
tables
now
in
each
of
the
day-to-day
steps.
I
try
to
follow
these
steps
for
the
second.
If
knowledge,
okay,
we
need
you
sure
that
we
have
this
automatic
language
flag
daily
all
sacrament
occasions,
but
the
only
difference
not
to
add
that
you
have
found
a
partner.
B
B
B
A
B
B
A
B
These
migrations,
I,
don't
think
that
we
can
skip
step.
It's
not
necessary.
I
think
that
you
get
right
there
here.
It's
not
such
a
step,
because
the
secondary
node
use
uses
a
capital
database
from
the
planter
node.
So
all
the
migrations
have
been
run
on
the
primary
notes:
Oh
God
applicator
to
the
segment.
B
B
B
B
B
B
A
A
B
B
B
B
B
B
B
Okay,
I
think
that
we
can
have
some
actually
things
from
the
skull
first
I'd
like
to
suggest
for
the
entire
gel
to
try
to
do
more.
Try
to
do
this
demos
more
frequently
like
we
had
during
the
g8
immigration
I,
think
that
it's
good
to
keep
our
documentation
up-to-date
and
make
sure
that
all
the
steps
are
working.
B
A
Think
it'd
be
worth
spending
the
time
with
the
ansible
project
like
letting
us,
you
know,
create
an
H,
a
cluster
with
a
circular
business,
matheran
stigma
yeah.
Do
you
think
it
would
be
worth
spending
the
time
adjusting
the
ansible
project
so
that
we
can
deploy
an
H
a
like
geo
setup?
Or
do
you
think
that
I
mean
because
that
might
be
a
lot
of
work?
I,
don't
know.
Yeah
seven
hours
to
you
know
stand
us
up
because
of
various
issues,
though
it
might
be
worth
it.
Yeah.
B
I
think
so,
I
think
that
the
first
step
we
need
to
make
sure
that
our
documentation
is
working.
It
took
me
seven
hours
because
I
found
a
lot
of
issues
different
stuff.
Then
after
you
have
the
conditions
working,
we
can
try
to
play.
Danceable
Scripture.
She
used
the
HIV
provider.
I'm
not
sure
is
the
name
of
the
two
that
the
infrastructure
King
using
to
use
to
set
up
the
get
shadowed,
conquest,
er,
well,
yeah,
but
I
think
that's.
It's
working
best
take
some
time
to
improve
that
script
as
well.