►
From YouTube: Magento MSI Open Demo. June 1, 2018
Description
Today’s agenda:
-) Regression result
-) Travis builds update (https://github.com/magento-engcom/msi/issues/1054)
-) Consistency between Catalog and Inventory
-) Synchronization between Catalog and Inventory
A
I'm,
switching
on
her
Gordon,
so
I
believe
that
you
were
that
we
were
working
on
a
regression
cycle
for
MSI
release
preparation
during
this
week
and
I
believe
we'll
start
from
the
sharing
the
status
of
our
regression,
because
we've
just
finished
their
aggression
and
wishes
Lau
will
describe.
You
know,
share
the
status
of
MSI
regression
and
the
next
steps.
We
will
go
on
a
way
to
release
Amazon
swallow
so
free
to
show
your
screen
and
just
show
you
first.
B
B
B
Finally,
we
got
proper
installation
for
b2b
resume
aside
and
we
already
had
issue
what
is
closed
and
I.
Think
Monday.
We
will
do
small
regression
for
b2b
for
sheriff,
catalog
and
other.
Since
what
is
new,
we
have
on
JIRA
small
cycle
what
already
chosen,
and
hopefully
we
will
check
that
the
maasai
working
properly
will
be
to
be
caught,
then
I
think
the
soul
from
me.
Do
you
have
any
questions?
Okay,.
A
So
I
believe
the
next
one
gonna
be
Valeri
and
let
me
share
my
screen,
so
we
we
had
pretty
interesting
situation
because
I
believe
those
of
you
who
made
who
requests
to
MSI
all
who
review
the
MSI
code
on
on
our
Reaper
notice
that
we
had
red,
builds
on
Travis
and
actually
the
main
reason
for
the
red
build.
Was
it
like
we
start
to
exceed
as
a
maximum
time
for
most
of
the
jobs.
A
We're
on
on
on
an
integration
test
and
all
of
the
agents
was
not
able
to
proceed
was
not
able
to
proceed
and
run
all
the
tests
you'd
we
have
on
MSI.
We
were
discovering
this
issue
for
pretty
long
time,
probably
during
the
whole
week,
and
we
found
out
that
the
main
root
cause
of
the
issue
actually
beer
it
in
the
scope
was
there
in
the
scope
of
one
of
the
bug
fixing
where
the
particular
integration
test
was
fixed,
adding
their
generate
sequence
and
story
need.
Actually
that
happened
in
the
website
website
with
this
source.
A
Each
HP
fixture,
and
actually
this
operation
generate
sequence
under
the
hood,
generates
up
to
40
different
tables
and
this
the
separation
supposed
to
be
run
to
clean
the
clean
and
prepare
environment
for
the
full
integration
issued
to
run.
But
we
reuse
this
very
uses
feature
for
quite
many
tests,
and
actually
this
means
for
us
that
we
generate
to
many
additional
tables
during
the
cool
run
of
all
the
tests
use
of
integration
tests
and
that
take
us
too
much
time
to
proceed
that
we
don't
have
enough
room
of
time
to
run
all
the
integration
tests
use.
A
We
currently
have
it
on
the
side,
so
we
stuck
with
this
issue
and
actually
this,
if
you
want
a
root
cause
for
an
ability
to
process
on
Travis,
your
integration
test,
which
you
actually
our
integration
test,
which
we're
on
for
all
the
all
of
the
cars
deliver
it
to
em.
Aside
and
now,
Valerie
will
share
the
status,
how
that
was
fixed
and.
C
D
A
This
actually,
this
issue
show
very
well
that
that's
how
it's
important
to
keep
the
environment
and
the
keep
their
state
clean
after
yourself
after
your
test
when
the
test
is
executed
and
the
test
should
not
pollute
their
state,
the
global
state
of
the
application
framework
with
additional
data,
and
because
this
data
would
will
definitely
harm
the
other
test,
and
this
can
affect
the
pool
test
you
so
now.
Let
me
share
my
screen
again
and
I
will
show
you
update
regarding
the
regarding
the
synchronization
between
catalog
and
inventory.
A
So
this
short
introduction,
like
you,
probably
know-
and
we
described
in
several
times,
that
we
are
trying
to
build
MSI
and
actually
the
cool
new
inventory
system
as
much
decoupled
from
catalog
as
possible
and
the
we
are
moving
towards
a
micro
service
architecture.
That's
part
that
actually
means
that
we
communicate
in
an
MSI
to
do
the
magenta
for
just
we
service
layer.
So
it
would
be
potentially
possible
later
to
deploy
MSI
on
a
dedicated
Hardware
environment
and
make
it
communicate.
A
It
was
discussed
several
times
that
probably
we
no
need
to
make
it
synchronization
between
the
catalog
data
and
the
inventory
data
and
what
it
actually
means
that,
for
example,
it
is.
It
should
be
possible
to
have
asked
in
catalog,
which
does
not
exist
in
inventory
and
vice
versa,
so
it
is
potentially
possible
to
have
an
inventory
for
the
SQ
which
does
not
exist
in
the
catalog.
What
we
gain
with
this
with
this,
let's
say,
not
consistency,
we
current
because
we
wind-
we
don't
have
this.
A
We
are
not
keeping
this
consistency,
actually
that
make
us
possible
to
run,
for
example,
indexation
for
catalog
and
for
inventory
independently.
So
we
know
more
like
we,
our
inventory
indexation
and
the
inventory
import/export
is
not
dependent
on
catalog
import/export,
so
we
can
firstly
to
import
all
the
inventory
data
and
just
afterwards
we
can
import
the
catalog
data
so
and,
moreover,
we
can
do
it
simultaneously.
A
Improving
the
scalability
of
the
whole
system.
Another
thing-
and
this
thing
I
would
like
to
demonstrate
you
that,
for
example,
if
for
some
reasons,
our
product
from
catalog
would
be
removed
or
deleted,
we
still
can
purchase
order
and
we
can
make
shipment
invoice
or
order
cancellation
depending
on
what
merchants
wants
to
do
with
them
on
this
particular
sq.
So
we
don't
have
a
limitation
here
that
the
products
supposed
to
supposed
to
exist
in
catalog,
so
I
will
I
will
demonstrate
this
behavior
to
you.
So
what
the
precondition
we
have.
We
have
a
stock.
A
A
A
A
A
A
A
A
Here
we
get
to
the
page
with
the
source
selection,
algorithm
and
the
source
selection
algorithm
proposed,
so
the
only
source
selection
algorithm
we
have
is
a
priority
based,
so
the
priority
of
the
custom
source
is
higher
than
the
probe
or
equal
to
cubes
or
that's
why
algorithm
proposes
to
make
the
deduction
from
custom
source
in
the
quantity
of
100?
And
after
that,
when
we
run
out
of
all
products
on
a
custom
source,
we
get
a
left
over
to
move
through
the
water
from
the
Q
source.
So
let's
make
deduction,
for
example,
from
custom
stores.
A
Here
you
see
that
the
partial
shipment
was
created
and
we
sent
50
our
50
pieces
from
the
from
the
custom
source.
So
we
can
get
back
to
shipment
page
again
and
see
that
we
have
150
pieces
need
to
be
delivered.
Oh
so
sorry
we
have
just
100
pieces
need
to
be
delivered
and
the
availability
is
150.
So
taking
into
account
that
we
made
the
church
shipment
from
the
custom
source,
we
have
just
50
items
left
on
a
custom
shop,
so
everything
works
properly.
But
now,
let's,
for
example,
make
an
invoice.
A
A
A
So
you
see
that
we
can
make
all
the
operation-
probably
it's
only
appraised,
I
did
not
show
you
the
order
cancellation.
This
also
work,
but
they
don't
have
order
to
show
you
so
like.
You
can
see
that
all
the
operation
like
placing
placing
shipment
will
voice
in
order
cancellation
and
the
credit
lemma
works
properly
for
product
which
does
not
exist
anymore.
In.
D
Igor
it's
marking,
as
we
talked
about
yesterday,
he
didn't
show
this
in
the
demo,
but
those
same
things
do
not
work
on
core
magenta
without
MSI.
So
if
we
try
to
do
that
now,
if
we
wouldn't
it
just
wouldn't
work,
we
get
an
error.
There
I
believed.
So
it's
great
to
see
this
as
a
feature
improvement
for
merchants.
A
That's
why?
Actually,
we
can
get
back
product
to
to
stock
returning
the
product,
so
yeah
compare
comparing
with
comparing
with
poor
implementation.
I'm
aside,
my
provides
provides
a
variety
of
different
business.
Cases
actually
provides
all
variety
of
business
cases
working
with
with
order
processing.
A
And
probably
another
case,
I
would
like
to
show
you
regarding
regarding
the
keys
we
we
discussed
with
you
Markey
yesterday,
when
we
don't
have
enough
products
to
fulfill
the
order.
So
let
me
share
the
existing
behavior.
We
modified
it
a
little
bit,
so,
let's
create
a
product.
So
just
let
me
let
me
describe
a
little
bit
what
I'm
going
to
show
taking
into
account
that
our
system,
our
inventory
inventory
system,
is
the
system
with
the
eventual
consistency
we
currently,
we
have
two
points.
A
A
point
where
we
make
a
check
is
happens
after
some
time
when
the,
when
the
merchant
comes
to
the
admin
page
and
the
processes
order
and
between
this
between
this
checkpoint.
Between
these
points,
it
could
be
potentially
situation
when
we
sync
up
with
with
external
system,
with
your
P
system,
and
the
European
is
a
source
of
truth
and
the
ERP
system
overrides
the
data
in
magenta.
D
A
Magenta
to
green
disc,
their
inventory
data
and
recalculate,
they
so
saleability
quantity.
So,
at
the
time
when
merchant
comes
to
the
admin
panel
and
decide
to
invoice
and
proceed
with
the
shipment,
it's
not
possible
anymore
to
fulfill
the
order
in
the
whole
scope,
because
potentially
we
can,
we
can
have
a
walk
of
products
to
fulfill
it,
but
the
one
with
that
we
still
need
to
make
a
partial
shipment,
and
this.
This
is
especially
a
business
case
for
b2b,
because
if
we
have
too
many
products
in
a
as
a
part
of
the
order,
we
can
have.
A
For
example,
some
of
the
products
already
ready
to
be
shipped
and
other
products
are
not
ready
to
be
shipped,
and
maybe
some
warehouse
is
already
prepared
to
make
a
delivery,
and
some
of
them
are
not
so
we
can.
We
should
have
an
ability
to
make
a
partial
shipment
and
I
will
show
you
how
to
implement
it
right
now.
Let's
create,
let's
create
new
product.
A
A
A
But
to
do
this,
for
example,
let's
update
the
quantity
of
product
on
some
of
some
of
the
sources
where
this
product,
where
this
particular
product
is
stored.
For
example,
let's
update
the
quantity
on
keep
source
and
let's
specify
it
like
25
and
save
the
products.
So
currently
we
have
a
order
which
is
placed
for
150
items,
but
after
this
save
which
which
run
or
indexation
we
have
actually
product
saleability
applies
the
reservation
which
we
already
place
and
if
the
reservation
is
for
150
items.
A
A
A
A
A
A
C
A
D
D
So
you
design,
custom
and
Kiev
source
we
have
for
both
of
those
sources
and
notify
quantity,
and
then
the
ability
to
have
notify
quantity
is
the
default
when
we
assign
sources
for
the
first
time,
those
are
both
empty
and
then,
when
you
save
the
product,
the
notify
quantity
becomes
zero.
So
the
default
state
ideally
should
be
that
the
notify
quantity
use
default
box
is
checked
the
first,
the
first
time
we
could.