►
From YouTube: SIG-Testing Biweekly Meeting for 20230919
Description
SIG-Testing Biweekly Meeting for 20230919
A
Cool
all
right,
hey
everybody,
Welcome
to
successing,
meeting
yeah
I.
Think
I
recognize
folks
here
as
having
attended
before.
But
if
anybody
wants
to
unmute
and
introduce
themselves
feel
free
to
otherwise
we
can
get
started
with
first
item.
B
I
do
so
before
we
get
started.
I
think
there's
something
wrong
with
the
meeting
notes
permissions.
Can
you
double
check
that
please.
B
The
meeting
notes
I
can't
edit
them,
even
though
I'm
part
of
the
right
Google
Groups.
A
A
B
Not
now
we'll
do
that
later,
yeah
but
yeah.
So,
as
you
know,
I've
been
working
on
the
getting
cops
already
to
be
used
as
our
primary
Ed
cluster
provisioning
tools.
B
I
think
I've
made
some
rather
well
good
progress,
but
I
have
a
couple
of
PRS
that
are
open.
That
I
need
someone
to
approve,
so
we
can
get
them
merged.
B
I.
Think
for
me,
the
most
important
one
is
our
first
one
there
and
it's
when
you
click
it
and
get
on
the
screen.
It's
kind
of
self-explanatory,
of
what
it's
trying
to
do.
There's
been
a
lot
of
we've
said
a
lot
about
it,
but
it
needs
to
be
fixed.
A
Cool
yeah
anything
else,
besides
just
an
advisory
of
like
asking
for
help
for
approving.
A
B
The
case
authenticated
test
images.
We
have
a
test
that
basically
pulls
images
from
a
private
location
in
Google
cloud
from
from
a
container
registry.
It's
misconfigured
right
now.
I
would
like
to
get
that
changed,
but
It's
Tricky,
so
we're
gonna
recreate
that
in
a
community
and
do
card
project
and
make
it
work
properly
in
there
here.
So
we
just
I
just
need
some
thumbs
up
and
then
you
can
go
and
deploy
that.
C
Program
with
the
pull
request
is
like
we
don't
have.
Ownership
of
the
tests
require
requiring
this
artifact
registry.
Wait.
Bro
I
think
Ben
made
the
comment
about
this
I
think
the
bigger
problem
is
like
for
the
tests
relate
to
private
registry.
We
don't
have
ownership
and
I.
Don't
want
us
to
set
up
a
piece
of
infrastructure.
That's
going
to
be
a
technical
death
in
six
months.
C
C
That
that's
the
intent
it's
like
we
need.
We
need
to
seek
one
thing,
owning
the
test
and
basically
giving
a
plus
one
to
remove
the
test,
because
no
one
is
maintaining
them.
So
what
the
first
one
is
who's
owning
those
tests.
The
second
thing
is
that
do
we
need
them
now,
because
maybe
we
have
a
better
coverage
for
this
test
somewhere
else.
D
Okay,
before
before,
removing
tests
that,
let's
not
passion,
because
the
problem
with
the
moving
and
adding
and
changing
things
quickly,
is
that
in
kubernetes
this
takes
one
or
two
releases.
Until
we
know
the
Fallout,
there
are
a
lot
of
things
here,
like
copying,
shift
that
they're
not
going
to
know
that
you
remove
one
test
until
one
or
two
releases.
Oh,
for
this
specific
case
is
C
gaps
yeah.
We
need
to
go
to
see
that
and
explain
what
is
the
problem
and
and
actually,
you
need
to
provide
a
solution.
D
B
Think
the
problem
is
an
infra
problem.
It's
not
necessarily
with
a
test.
The
test
is
fine,
it
just
doesn't
work
when
you
are
running
it
in
certain
locations.
B
D
And
tag
it
so
it
can
only
run
with
a
special
conditions,
all
right.
So
this
is
the
problem
that
we
are
having.
Now
we
have
this
right
and
that's
the
problem.
We
need
to
carve
it
out,
but
we
cannot
do
extracting
a
painful
strategy.
We
need
to
be
a
strategic
and
and
go
one
by
one.
This
is
going
to
be
no.
B
D
But
within
it,
okay,
but
we
need
to
put
the
project
in
that's.
Why
I
ask
it
to
the
I
mean
this
project
has
to
be
managed.
It's
not
now
we
are
going
to
move
to
everything
and
we
are
going
to
remove
tests
and
do
everything
this.
This
needs
a
communication
with
all
the
state,
but
all
the
stakeholders
are
all
the
kubernetes.
D
C
D
In
the
long
term,
we
eventually
will
get
rid
of
Cuba
right.
So
what
we're
gonna
do
is
now
we
have
coverage
100
and
then
we
move
to
cups
and
we
can
have
over
50
because
they
just
didn't
run
and
we
didn't
do
it.
The
next
thing.
B
B
D
D
B
Problem
with
that
is
at
the
end
of
The
Journey,
when
we
have
both
classes
up
and
running.
The
first
thing
we
want
to
do
is
see
how
the
tests
are
configured,
which
ones
are
passing,
which
was
a
failing.
If
there's
too
many
skips,
these
questions
will
come
off
and
be
like.
Why
are
we
doing
this
yeah
foreign.
B
D
D
D
B
The
other
two
PRS
are
super
simple
right:
we've
got
tests,
they're
not
tagged
properly,
I
added
the
great
tags,
so
I
can
skip
them
properly.
Like.
D
D
So
for
for
any
changes
you
need
to
eat
this.
Usually
you
can
touch
me
or
me,
and
Patrick
Cody
or.
B
D
D
A
Just
remembered
or
sorry,
no
real,
quick
thing:
I'm
gonna
fix
this
after
I
just
remembered
that
my
email
filters
out
KK
PR's
for
some
reason,
so
I
don't
think
I
see
those
I
will
go
fix
that.
D
B
B
Okay
sounds
good.
I'm
gonna
write
that
book
off
this
week.
I
believe
it's
kind
of
helpful
I
am
also
in
a
position
where
I
can
go
that
what
I'm
proposing
has
works,
because
there
are
a
ton
of
bugs
that
had
to
squish
before
I,
even
had
like
a
working
POC
to
bring
in
front
of
someone
wow.
D
B
D
D
History
Founders
and
you
are
going
to
have-
then
they
are
going
to
identify
you.
This
is
the
person
doing
this
so
I'm
going
to
allocate
time
because
right.
A
D
Okay,
okay
and
then
you
have
the
other
one.
One
important
thing
about
this
replacement
and
I
was
talking
with
Justin
Santa
Barbara
today,
right
now,
I
think
that,
because
of
the,
my
suggestion
is,
is
to
make
the
transition
as
similar
as
possible.
Simple
sample
GC
deploys
a
very
simple
password
single
Master
right
three
work,
so
if
you
can
mimic
that
as
much
as
possible
with
it,
because
if
we
start
to
have
multi-master
in
the
CI,
that's
super
thankful
to
divide.
B
Yo,
we're
not
doing
that
Ops
by
default.
Doesn't
even
do
that,
so
it's
just
one
master
and
three
working
out.
That's
it.
D
Okay,
so
perfect,
that's
another
thing:
okay,
you
you
just
no
more
questions,
because
the
moment
that
you
need
to
divide
a
place
with
multi-master
is
like
spending
hours
all
day
and
I
think
a
master.
Everything
is
much
cheaper,
yeah
appreciate
it.
So,
as
I
said
yes,
I
I
want
to
support
you
and
and
I
think
that
you
are
doing
the
ugly
thing.
So,
let's
try
to
to
put
some
structure
on
this
project
and
and
be
able
to
do
your
best
faster.
A
All
right
does
that
sound
good
on
that
topic.
Is
there
anything
else
to
discuss.
A
Oh
hi
I
think
you
have
one
more
just
to
show
off
some
examples.
B
All
right
so
this,
if
let
me
show
you
this
or
the
job,
is
configured
to
do
yeah
so
yeah.
So
this
is
similar
to
the
I
think
the
43
test
that
we
do,
that
checks
everything
except
serial
slow,
and
so
these
features
tests
it's
looking
great
so
far.
These
are
two
failures
that
we
discussed.
This
is
another
thing
that
I'm
planning
on
fixing.
It's
got
nothing
to
do
with
kubernetes.
B
It's
it's
a
test
that
was
written
to
test,
keep
up
Behavior
it
shouldn't
even
be
here,
and
then
this
is
a
DNS
thing
that
I'm
going
to
fix
today.
I
managed
to
get
something.
A
B
Upstream
that
will
make
the
screen
but
yeah.
This
is
what
we're
looking
at.
It's
almost
ready,
I
think
now
I'm
in
a
position
where
I
can
run
more
run
the
functional
equivalence
of
these
tests
that
are
available.
Where
was
it.
B
B
Where
I
can
start
running
these
version
of
these
jobs
that
I've
provisioned
by
cops
and
get
the
same
results
particularly
like
this
one
here,
it's
currently
bugged
but
yeah
I,
think
sometime
next
week,
I
might
be
in
a
position
to
start
running
these
jobs
in
Canary,
I
didn't
see
how
we
look
and
then
by
then
actually
have
that
document
ready
for
the
Journey
of
how
we're
gonna
take
that
to
production.
B
B
B
Yeah
I
think
it's
similar
to
this
one
here,
a
little
bit
more
tests.
Sorry,
not
that
one
sorry
just
give
me
a
second.
B
D
B
So
that's
what's
passing
right
now
we
know
what
the
failures
are.
One
is
DNS
that
would
be.
D
Fixing
okay
and
go
go
to
the
to
the
other
projects.
A
D
D
Right
that
you
are
saving
30
minutes
in
in
50.,
so
yeah.
B
This
is
a
pre-built
version
of
kubernetes,
but
even
then
it
takes
like
30
minutes
on
average
to
run
like
once,
most
of
conformance
on
a
bunch
of
other
extra
stuff,
but
yeah.
It's
I
think
it's
pretty
fast.
If
I
look
at
job
history,
30
minutes-ish
on
average.
B
So
far
this
is
running
I
think
about
eight
times
a
day,
but
yeah
I'm
gonna
get
all
these
greens
soon
and
then
and
write
that
doctor.
B
D
D
B
C
B
B
Yeah,
okay
I'll
write
all
of
that
down
in
that
dark
and
see
where
we
are
next
week.
D
D
A
Awesome
yeah
I
think
with
that
we're
at
the
end
of
the
agenda.
Is
there
any
last
minute
stuff
that
folks
want
to
mention.
A
All
right
sounds
good
thanks
everybody
for
attending
thanks
for
the
detailed
discussion
folks
and
take
care.
Thank.