►
From YouTube: GMT 2018-01-25 Containerization WG
Description
Agenda and notes:
https://docs.google.com/document/d/1z55a7tLZFoRWVuUxz1FZwgxkHeugtc2nHR89skFXSpU/edit?usp=sharing
A
B
B
A
A
B
A
A
B
Oh
sorry,
hi,
hey
guys,
I
I
I'm,
not
aware
that
I
was
muted,
circular
yeah,
alright,
so
yeah
I
think
let's
get
started
so
so
I
think
the
first
thing
I
want
to
do
is
since
we
just
cut
1/5,
and
we
want
to
do
a
planning
for
1/6
I.
Think
let's
go
back
to
this
spreadsheet
first
and
then
go
over
everything
we
have
so
far
and
make
sure
the
status
is
up
to
date.
First
I
think
I
think
the
last
time
we
reviewed
this
one
is
probably
like
the
early
of
1
Phi
release
cycle.
B
B
Milled
it
as
well-
yes,
yes,
ok!
So
what
an
ownership
cleanup
I
think
we
have
a
kind
of
tentative
kind
of
thought,
dump
and
design
doc,
but
we
haven't
thought
to
the
point
where
we
can't
actually
on
create
the
tickets
and
start
implementation
so
status.
Is
your
design
I?
Think
that's
accurate!
Mrs.
can
terrorize
the
documentation.
I
think
made
a
lot
of
progress
on
that,
so
I
linked
the
current
documentation
for
maces
container
riser,
which
points
to
a
bunch
of
documentation
for
different
Isolators,
which
is
much
better
than
before.
B
So
so
this
one
I
don't
know:
let's
go
back
to
the
target
release
for
the
island
that
we
don't
know
and
then
we're
pretty
to
do.
A
prioritization
depends
on
the
priority
and
then
figure
out
the
target
release.
Is
that
sound
good,
Greg
yeah
for
sure,
ok,
CPU
affinity?
We
do
have
a
design
somewhere,
but
I
think
we
haven't
gone
through
that
design.
We
haven't
approved
that
design
down
edge
and
we're
still
in
today
on
what
we
plan
to
do
so.
That's
accurate,
io,
isolation.
A
D
B
B
The
rest
has
not
started
to
just
go
through
all
of
them.
Just
make
sure
we
support
registry
Dhaka,
credential
helper,
you
know,
I
started
pre
post
on
pre
post
finish
hook,
yeah,
so
that's
not
started.
We
have
a
discussion.
I
remember,
but
we
haven't
started.
Artifacts,
cashing
no
limit
the
room
size
of
a
container
cheetah
wants
to
work
on,
Deb
I
think
haven't,
started
yeah
I,
don't
know
we
have
a
design
decision,
I.
Think
I
marked
as
with
design.
B
Pick
one
as
user
defined
command
yeah.
This
one
is
highly
related
to
this
one
which
we
to
have
a
discussion
of
this.
So
James
I
put
the
the
containerized
supervise
I'm
kind
of
related
to
this
cool
thanks,
yeah
I
think
we
got
to
a
point
that
we
have
a
reasonable
design.
We
haven't
designed
talk
here,
but
I
think
I
just
mark
it
as
a
design.
D
B
A
B
So
yeah
so
okay,
James
I,
remember
so
basically
like
they
don't
want
to
use
a
continuum
edge
they,
but
they
want
to
provide
isolation
between
different
containers
and
the
way
they
suggest
was
like
having
a
writable
layer.
On
top
of
that,
so
that
anything
you
write
into
the
host
file
system
will
not
be
visible
to
other
container.
Okay
that
make
sense.
B
It's
not
started
I'm,
not
sure
if
it
still
related
to
the
supervising
I,
just
move
those
things
together,
I
might
be
related
support
for
our
statistic
is
in
CNI,
not
started
soap
or
nailing
sucker
for
entering,
contain
our
namespace
I.
Don't
know
what
is
I
think
they
just
bring,
but
a
mirage,
yeah
I
think
it's
we
haven't
started
yet.
Oh,
who
did
that?
B
B
D
D
B
B
It
out
I
need
to
find
it
out.
Okay,
so
I
mark
it
as
in
progress,
yeah
librarian,
which
are
not
doing
anything
that
their
support
cor.
I
know
started
now
with
bandwidth
as
a
resource
not
started
to
do
I
support
enhancement,
yeah
I
haven't
done
that
so
I
also
add
this
CRI
for
like
task
exactly
attached
in
a
socio
I
I,
just
ask
them
some
some
entry
there
yesterday.
So
anything
else
that
you
guys
wants
to
add.
Before
we
doing
a
sorting,
yeah
I
have.
B
D
Have
some
vm
requirements
coming
down?
The
pipe
I
think
unlikely
to
look
at
cgroups
device
device
cgroups
for
those
kinds
of
things,
because
I
just
think
that
generalized
VM
support
is
way
too
big
a
task
mm-hmm.
It's
not
gonna
happen
in
a
useful
and
we're
never
met
it.
Do
it
in
at
least
four
times
right
for
us?
What
exactly
do
you
mean
by
senior
device
support?
D
D
B
E
B
B
B
Yeah
so
that
would
be
great.
Yes
sounds
like
it
sounds
good,
so
we
should
should
clean
this
up
and
JIRA
for
everything
and
if
not,
if
there's
no
exists,
that's
Korean,
epic,
gerrae,
okay,
so
great,
let's
do
some
prioritization,
then,
let's
just
let
me
just
move
all
these
done
ticket
to
the
top
to
the
top
is
done.
B
B
B
A
Yeah
at
least
I
can
see
for
for
those
big
company
many
big
user
day.
They
might
still
stick
with
the
dopa
image
for
a
long
time,
not
only
the
OCR
yet
because,
and
especially
they
can
correct
to
the
requried
of
the
doctor
demon.
They
can
run
the
other
image
in
cloud
continuation.
They
can
do
their
own
view,
so
they
can
tell
they
like
run.
Everything
is
that
okay.
B
Especially
yes,
yeah
the
reason.
I
say
that
is
because
I
don't
see
short
trend
that
people
it's
gonna,
ditch
in
darker
workflow
and
published
OCI
image
to
their
Ostia
image,
store
I,
think
it's
a
long
way
for
them
to
achieve
that
and
people
are
still
using
darker
tuning
and
docker
registry.
So
I
don't
see
like
strong
motivation
to
do
that
right
now
immediately
for
this,
but
we
should
keep
that
in
the
backlog
to
make
sure
that
we're
not
like
far
behind
this
and
also
I,
think
one
thing
is
I.
B
A
B
But
yeah
coo
nest
caress
does
not
yeah,
so
I
think
they
think
that
they
delegate
that
to
the
CRI
Renta
and
the
Sun
see
our
intensive
port
status
and
some
doesn't
yeah
well
I
guess
from
users
experience
like
I
haven't
seen
any
customer
debt
and
that
explicitly
asking
for
using
that
it's
hard
to
justify.
Why
I
want
to
do
that,
but
I
think
it's
a
good
opportunity
for
us
to
think
about
a
unified
artifact
store
because
it's
gonna
introduce
a
new
one,
more
store,
that's
better!
Just
not
do
that
and
have
a
unified
store.
B
So
I
move
this
up
a
little
bit
and
let's
do
the
relative
soaring
later
so
user
name.
Space
James
is
that
before
that
so
I'm
containerize
it
documentation
at
the
end,
seen
progress
in
seeing
a
pretty
good
shape
right
now,
but
we're
still
doing
true
I
still
think
that's
a
top
priority
documentation.
So
I
put
that
at
the
top.
Do
you
guys
agree?
Yep,
yeah,
okay,
great
use,
earn
a
space
James.
Do
you
guys
have
a
target
release
white
when
you
want
to
upstream
those
patches,
1.6
well,.
D
D
B
A
B
B
D
B
B
I
think
replace
this
one
is
m2.
Let's
oh
yeah
I
think
target
this
four
one.
Six,
we
I
don't
think
we
can
complete,
remove
command
and
securing
one
six,
but
I
think
one
thing
we
can
do
is
like
let
comment
executors
lucky
for
I
secure
support,
single
task
command.
So
that's!
Obviously
we
can
definitely
do
I'm,
not
sure
if
you
guys
can
yeah.
So,
let's
just
pull
one
six
there
on
definite
one
to
get
that
done.
Yeah
my
pet
shorty
does
that
okay,
so.
D
B
So
so
I
think
Greg
I
think
the
the
the
fix
for
that
issue
is
simple.
I
think
the
majority
work
should
be
testing
because
we
don't
really
come
I
mean
just
like
a
confidence
issue,
because
we
don't
really
sure
that
that,
if
that's
gonna
cause
more
issue
because
I
think
you
change
the
container
like
hierarchy
structure
a
little
bit,
yeah.
C
B
D
So
we're
taking
our
one
thing,
I
realized,
was
that
we
don't
really
have
any
test
coverage
for
their
custom
executive
case,
so
the
command
executor
works
completely
differently
to
a
custom
executor
and
the
default.
Is
that
key?
No
difference
who
are
customers?
I
can
show
you
say:
yeah.
The
people
have
no
outer
fighting
custom
executives.
We
don't
have
any
specific
containerization
files.
B
Right
yeah,
I
agree,
so
yeah
I
think
the
test
will
custom
Messick
here
might
yeah
I
mean
while
we
are
doing
this,
we
probably
want
to
cover
the
custom,
execute
I
think
the
problem
is
custom
and
security
I,
don't
know
why
we
don't
have
too
many
tests
for
that.
Just
heart,
I
guess:
I
have
to
write
a
mock
aiesec
here
somehow
in
the
repo
yeah.
C
B
B
Okay,
see
your
namespace.
Also,
yes,
is
that
okay,
so
small?
What's
the
release,
do
we
want
to
shoot
for
like
adding
some
minimum
test
for
custom
a
secure
in
the
next
release?
Who
has
cycle
for
this
so
Jamie?
Something
I
have
a
lot
of
stuff
to
do.
I,
don't
think
you
have
I,
don't
think
I
have
cycles
yeah.
B
D
A
B
B
D
B
B
It's
gonna
buy
us
eventually
pretty
soon
I
would
say,
given
the
CSI
stuff
I
think
we
did
some
design
meetings,
but
I
think
everyone
just
don't
have
time
to
time,
but
I
think
we
should
go
back
and
trying
to
fix
that
I
think
we
don't
have
to
we'll
probably
figure
out
solution
for
Linux,
I,
think
figure
out
in
solution
that
works
for
both
Linux
and
the
windows.
A
common
solution
is
gonna
be
hard.
Let's
just
do
separate
solution
for
Windows
and
Linux,
based
on
the
discussion
that
we
have
last
time.
B
B
C
B
B
A
With
me
like
yesterday
on
this
one-
and
he
also
checked
out-
actually
he
checked
with
me
on
the
limiting
the
request
sighs
a
couple
times
before,
and
this
is
gonna
be
the
highest
priority
for
uber
okay.
They
hope
this
can
learn
for
one
six,
so
yeah
they
have
some
other
interesting
parts,
but
this
one
cannot
be
pretty
sure
it
ain't
gonna
be
the
highest
rated
for
there.
Okay,
I'll
move
up.
A
B
So
in
so
Demetri
is
now
here,
I
feel
that
this
still
need
a
lot
more
discussion
and
we
don't
have
a
clear
target
premise
with
target
release,
for
that
way
sounds
like
the
progress
is
being
haunted.
It's
not
really
an
important
things
for
us
at
the
moment.
Let
me
know
if
this
becomes
an
issue
and
that
we
can
reprioritize
this.
If
not,
we
just
keep
that
below
those
things
that
we
we
say
that
we're
gonna
do
set
comp
James.
B
A
Yeah
so
for
second
we
had,
we
did
have
the
decider
and
some
patches
from
J.
We
have
said
we
had
a
son
discussion
on
the
decide
and
then
he
sent
out
some
check
which
we
just
don't
have
tried
to.
Then
they
yet
depends
on
the
priority,
how
we
could
send
it
for
once
expert
three
one,
how
important
it
is
well.
B
I'll
just
put
that
there
on
under,
like
the
uber,
limine,
FS,
stuff
and
I,
tend
to
report.
I
won
six.
We
can
reduce
that
later.
Was
the
party
changes?
Okay,
so
we're
halfway
at
the
very
least
okay.
So
these
are
not
started
tickets
on,
let's
see
which
one
you
guys
really
want
to
see
in
the
next
release
or
in
the
next
few
releases
that
should
be
bumped
higher,
so
change
dimension,
this
right,
general
device
access
support
this
one.
B
B
Us
so
I
think
can
include
steel,
ravine
coach,
I.
Think
I'm
on
from
mrs.
fear,
was
working
on
that
I.
Think
the
skeleton
like
the
framework
of
OCR
has
been
committed.
We
just
haven't
add
those
command,
yet
a
whole
host
whole
pipeline
is
being
committed
and
just
the
sub
command
needs
to
be
added
and
and
I
think
for
CRI
pro
need
to
have
a
like
user
experience
review,
but
anyway,
so
I
think
it's
not
that
big
of
a
task
given
the
stuff
has
been
committed
already.
B
I'll
just
need
your
alwa
sub
command,
for
that
sounds
like
it
needs
someone
to
ship
it
up
in
the
project
manager,
yeah,
yeah,
I,
just
just
just
I
just
move
this
up.
I
need
to
I'll
talk
to
someone
to
find
a
cycle
for
this.
One
I
think
I
have
seen
I
see
a
main
list
question
yesterday
on
this
I
think
from
Yelp
guys,
I'm,
not
sure
if
they
have
psyche
on
here
reach
out
to
them
see
if
they
are
willing
to
work
on
this.
B
B
B
B
If
we
do
or
if
you
launch
and
continue
with
GTI
in
a
boat,
if
you
type
TTY
command
inside
a
container,
it
will
show
another
TTY,
but
application
that
depends
on
TTY
still
works
like
them
like
still
works
as
expected
as
if
it
was
it
has
a
TTY.
So
I
think
there
is
some
issue
is
like
you
have
to
mount
them
like
DV
counsel,
at
a
given
location.
So
the
tty
program
can
see
that,
oh
just
about
them,
yeah
just
a
bug.
So
it's
not
really
like
a
big
issue.
B
I
think
a
TTY
application
still
works
properly.
So
it's
not
a
hard,
hardly
stuff,
but
it's
it's
not.
It
would
be
nice
to
have
that
fixed
I.
Think
that's
what
doctor
did
I
think
that
could
just
buy
my
own,
like
the
UV
counsel,
just
by
mounded,
the
actual
pqi
to
the
to
DV
console,
and
then
you,
if
you
type
TTY
inside
the
aqua
container,
should
slash
the
V
/
console
yeah.
It's
not
a
big
deal.
Just
put
it
there.
B
Okay
sounds
like
a
pretty
good
plan.
We
still
need
to
follow
up
on
some
nylons,
especially
for
the
default
I.
Sakura
connect
secure,
it's
a
big
chance.
I
mean
it's
not
it's
not
a
big
change,
but
it's
a
big
impact
change
and
to
make
sure
that
we
have
a
proper
repair
need
to
have
some
follow-up
on
this
yeah
I'll.
Take
a
notes
on
this.
B
B
Let's
go
back
okay,
so
that's
the
planning
sheet
on
the
other
thing
is
I
think
we
have
a
lot
of
bugs.
So
so
that's
the
filter
that
Gilbert
sent
me
yesterday.
So
just
do
a
search,
unresolved,
containerization,
I'll
container,
all
that
container
bugs
we
have
like
how
many
like
93
and
resolved
bugs,
which
is
not
great
so
I,
guess
we
don't
have
enough
time
today.
B
I
encourage
you
guys,
can
take
a
look
at
those,
especially
for
those.
The
bug
that
you
report
make
sure
they
are
up
to
date
and
also
on
we're
gonna.
Do
a
grooming
and
then
to
a
proper
prioritizing
prioritization
see
which,
but
we
want
to
fix
in
the
next
release
and
then
properly
mark
the
target
version
as
one
six
I.
Don't
think
we
have
enough
time
right
now,
how's
that
sure
I
guess
we
can
do
that
offline,
I!
Think
it's
too
many
tickets,
like
93,
there's
no
way
we
can
cover
them.
B
D
B
D
B
Yeah,
okay,
so
for
those
think
that
you
do
have
a
plan
and
you
do
plan
to
land
at
u16
mark.
The
target
version
has
one
six
so
that
I
can
track
and
then
we're
gonna
do
the
same.
I
think
we're
gonna
go
through
everything:
go
those
bug
ticket
in
the
property
market
target
version
so
that
we
know
what
we
want
to
work
out.
The
next
release,
Oh
read-only
most.
B
Yeah,
so
that's
the
reason,
I
think
I
think
I
am
I'm
pretty
sure
like
doing
some
refactor
I
fixed
that
by
having
closed
the
ticket
yeah.
That's
really
pretty
to
do.
Recruit
me
on
this.
You
know
I'll
double-check.
We
need
to
do
a
grooming
I'll,
send
you
the
yeah
I
mean
the
the
future
is
actually
here
in
the
talk.
So
this
is
bugs.