►
From YouTube: 20190801 sig arch conformance
Description
No description was provided for this meeting.
If this is YOUR meeting, an easy way to fix this is to add a description to your video, wherever mtngs.io found it (probably YouTube).
A
Hello
today
is
Thursday
August,
1st
2019.
This
is
sick,
arch
conformance
sub-project.
We
are
just
gonna
walk
through
with
some
backlog
grooming
because
we
have
a
big
backlog.
Hayley
have
other
things
you
want
to
prioritize,
so
the
purpose
of
this
call
is
to
basically
shuffle
sort,
which
is
always
fun,
I,
think
for
the
time
being,
I'm
going
to
ignore
pull
requests.
A
B
C
A
C
D
A
A
C
A
C
D
C
C
C
A
A
A
C
B
E
D
C
A
B
B
C
C
This
I
wanted
to
quickly
get
some
vocabulary.
That
may
not
be
clear,
is
within
our
swagger.
We
have
the
operations,
which
are
the
paths
combined
with
HTTP
verbs
that
give
the
operations
each
of
the
end
point.
Flush
operations
has
parameters
and
those
parameters
can
be
required
or
optional.
The
what
you
pass
two
parameters
are:
what
I'm
calling
resources,
sometimes
they're
called
kind,
sometimes
are
called
pipes
that
the
definition
objects
inside
of
swagger.
C
Those
resources
are
the
things
that
kubernetes
uses
types,
earth
kind
and
the
referred
to
as
pipes
elsewhere.
Those
resources
have
another
sub
thing
called
properties
when
they
point
to
other
resources,
so
we
can
have
required
parameters
for
our
operations
and
we
can
also
have
required
properties
for
our
resources
and
so
trying
to
clarify
when
we
look
at
this
query,
we're
looking
at
the
properties
of
resources
that
are
required
for
the
resource,
not
the
required
parameters
for
our
operations,
because
those
are
all
separate-
and
there
may
be.
D
C
For
example,
may
be
that
the
end
point
or
in
the
middle
point
of
a
dag
for
any
particular
API
call,
because
whatever
you're
passing
as
a
parameter
to
an
operation,
might
at
some
point
need
a
pause.
Spec
defined
each
of
their
resources
that
are
called
to
it
at
some
point
hit
a
pods
back,
even
though
pods
back
itself
is
not
directly
referenced
in
any
API
operation.
Right.
B
C
C
C
Trying
to
ask
the
entry
point
so
do
you
want
me
to
focus
on
operations
and
the
parameters
to
other
operations
that
include
with
anyone
within
their
their
graph,
PAH
spec
and
then
within
the
now?
That
sets
the
particular
operations
at
parameters
that
include
prospect
and
then
within
the
properties,
how
far
down
the
tree?
What
are
we
looking
for?
What
are
we
going
to
measure
that
are
we
wanting
to
look
at
all
the
kinds
that
are
referenced
from
any
operation
that
includes
parameters
that
include
some
reference
to
prospect?
I.
A
E
For
example,
like
a
pod
like
this
probe,
pod,
healthiness
probe
and
the
timeouts
related
therein
like
how
would
we
want
to
combinatoric
Li
Express,
this
right,
I
do
want
to
die.
I
did
want
to
dive
down
into
what
kinds
of
probes
can
be
used,
so
I
do
need
to
introspect
that
resource
a
little
bit,
which
is
how
I
discovered
that
we
had
tests
for
HTTP
and
exact,
but
we
didn't
have
tests
for
the
T
sub
D
probe.
We
do
now
what
other
introspection
should
we
be
doing.
C
A
So
measuring
field
coverage
or
resource
property
coverage
seems
like
the
reasonable
thing
to
do,
and
it
doesn't
need
to
transitively
close
across
everything
that
uses
pod
spec
right,
as
well
as
everything
below
that
graph
of
these
reach
of
these
fields.
Basically,
just
what
does
the
raw
field
coverage
that
we
have
ignoring
everything
else
and
then
wait?
A
E
There
there's
a
thing
out
there
that
K
native
uses
right
now,
which
I
keep
poking
hippy
or
somebody
if
they
want
to
try
like
bringing
it
over
here,
it's
a
forest
of
trees
and
how
covered
is
the
forest
and
all
they
check
is
whether
or
not
fields
are
sex.
Don't
check
what
they're
set
to
so
we're
not
covering
over.
Is
it
empty?
Is
it
you
know
the
maximum
length?
Does
it
have,
or
does
it
have
all
of
the
different
genomes?
E
A
And
I
think
that's
that's
a
good
starting
point
and
that's
the
point
you
should
like
to
be
honest.
That's
what
the
point
we
should
have
started
at
like
months
ago
like
this:
it's
a
finite
fixed
amount
of
work
to
verify,
and
it
should
take
no
more
than
a
couple
days
to
get
it
all
sorted
out.
If
this
is
the
primary
function,
if,
like
I,
don't
have
a
couple
days
to
vote
towards
anything
right,
that's
just
not
the
way.
A
A
A
A
A
Think
team
harrisbrown
all
right
so
so
I'll
leave.
These
two
is
at
CD
and
then
networking
and
we
just
closed
a
bunch
of
other
ones
right.
So
that's
pretty
decent
as
far
as
overall
issue
backlog
is
concerned.
So,
let's
start
with
pod
spec.
Let's
get
some
coverage
there,
these
other
things
are
kind
of
in
flight
and
we
should
evaluate
the
PRF
switch
to
PR
view
now
or
do
we
have
other
things
that
we
want
to
talk
about?
So
we
talked
about
pod
spec.
We
said
watches
for
p1,
but
let's
do
0.
First.
B
E
E
D
A
B
A
B
B
A
B
B
D
A
C
A
C
A
A
A
B
A
B
A
A
B
B
Think
this
is,
we
documented
that
they
shouldn't
need
them
for
now,
but
I
thought
that
we
currently
confirms
to
say.
No,
we
shouldn't
need
them,
but
we
do
have
to
deal
with
the
issue
of
you
know.
If
I
don't
know,
if
Red
Hat
wants
to
run
conformance
against
some
multi
cluster
system,
where
they're
not
gonna
happen,
there'd
only
be
able
to
run
in
a
namespace
they
have
in
whole,
or
should
that
be
okay,.
A
B
Can
see
I
know
I'm,
you
know,
obviously,
that's
just
an
example
but
like
if
somebody's
running
a
pure,
a
pure
kubernetes
as
a
service,
that's
some
multi,
tenant
and
back-end
I.
Don't
know
it
right
now.
I
think
the
the
guidance
in
the
document
is
they
must
not
require
privileged
in
order
to
be
a
conformance
test
and
I.
Think
that
this,
whether
it's
this
issue,
another
one
needs
to
be
to
say
how
do
we
handle.
B
Different
user
roles
in
conformance
is
what
does
it
mean?
Conformance
if
confirm
is
a
specifically
just
for
branding
then
maybe
requiring
privileged
is
fine,
because
it's
it's
just
the
providers
that
are
going
to
run
them,
but
if
conformance
is
for
users
to
be
able
to
evaluate
whether
their
workloads
are
whether
the
cluster
is
going
to
allow
them
work
learner
to
be
portable,
then
it's
a
different
question
and
it
has
different
implications
for
different
user
roles.
Well,.
A
Well,
that's
part
of
the
conformance
processes.
They
have
to
submit
the
configuration
that
they
used
so
that
it
can
be
reproducible
to
get
the
conformance
behavior.
And
if
a
person
comes
back
and
says
it's
not
conformant
like
openshift,
for
example,
right.
That's
because
they
are
in
a
multi-tenant
environment
where
they
don't
have
those
capabilities.
The
cluster
itself
is
component,
but
your
view
of
the
world
as
a
user
is
not.
You
cannot
do
all
the
things.
A
A
A
A
A
A
B
B
A
The
problem
is
because
I,
the
inner
workings
of
the
the
goo
goes
through
a
separate
layer
and
the
backend
of
this
of
how
there's
there's
a
separate
area
that
eventually
gets
in
the
same
watcher
code.
So
the
part
of
the
unmarshal
II,
it's
a
totally
separate
code
path
that
is
totally
forked
in
the
API
server
for
config
map
versus
other
things,
for
all
the
resources.
So
every
single
reason
they
all
eventually
back
into
the
storage
layer
but
there's
a
whole
there's
a
whole
middle
tier.
B
A
A
C
A
B
A
C
A
A
C
A
A
A
A
A
We
can
start
down
trying
to
close
out
some
of
these,
like
conformance
tests
for
Esther,
Danny
Act.
Sorry,
not
important,
totally,
not
important
to
me.
I'm
gonna
leave
that
in
backlog,
like
any
good
conformance
test
stock
required,
I
think
that's
open
by
Quinton.
How
old's
this
thing
he's
referring
some
document.
Yeah
we've
already
done
this,
though
we've
already
have
enough
documents.
I
think
we
have
a
lot
of
documents,
yeah.
C
C
D
B
A
B
A
A
A
D
A
B
B
So
what
it
probably
means
is
that,
right,
you
have
the
things
that
are,
incidentally,
access
performance
tests
like
we
just
did
service
validation
and
in
order
to
do
that,
we
use
in
Tennessee
right.
So
if
the
affinity
isn't
covered,
we're,
incidentally,
covering
it
when
we
cover
services,
yeah
I,
edit
long-term
backlog,
yeah.