►
From YouTube: Sig-Testing Weekly Meeting for 20230418
Description
Sig-Testing Weekly Meeting for 20230418
A
Okay,
cool
yeah
welcome.
Folks,
today
might
be
a
short
one,
I
think
it's
mentioned,
but
we
can
go
ahead
and
go
through
just
to
start.
If
there's
anybody
who
is
new
here
and
or
wants
to
give
a
quick
introduction
for
themselves,
please
feel
free
to
unmute
and
do
so.
B
Hello
I
am
new,
wish
I
used
to
participate
way
back
when
but
kind
of
stopped.
So
my
name
is
Vladimir
Vivian
I'm,
with
with
VMware
about
a
couple
years
ago,
I
started
the
the
e2e
framework
project
as
a
way
to
help
folks
write
kubernetes
tests
without
depending
on
the
intrinsic
built-in
kubernetes
test
framework.
So
this
one
is
completely
independent,
so
yeah
so
kind
of
came
back
to
we'll
we'll
be
trying
to
participate
more
and
more
in
the
in
the
Sig
testing
meetings
as
often
as
possible.
C
A
A
All
right,
I
think
that's
all
right.
That.
C
Said
Benjamin
just
showed
up
I
wouldn't
be
surprised
if
he's
got
something
to
say
right.
D
I,
don't
have
any
agenda
items
I
just
wanted
to
check
in
certainly
no.
A
Worries
yo
all
right,
I
think
with
that
said,
Vladimir
I
think
you've
got
the
agenda
in
him,
so
go
for
it.
B
Yeah
absolutely
so,
as
I
mentioned,
I
I'm
the
lead
for
the
e2e
framework,
which
is
a
Sig
testing,
sponsored
sponsored
project,
and
it's
been
around
for
a
couple
years
now,
I
decided
to
kind
of
jump
back
in
and
and
try
to
participate
as
much
as
possible,
because
the
project
is
actually
gaining
momentum
and
getting
adopters.
B
So
this
is
the
one
place
that
I
know
where
we
can
have
continuous
face
to
face
with
with
the
community,
so
definitely
want
to
jump
back
in
and
and
put
my
face
in
the
meeting
and
participate
so
that,
if
folks
have
any
questions
and
about
the
project,
the
last
thing
that
just
happened
is
I.
I
had
a
chance
to
finally
do
a
release.
B
Last
week
or
last
weekend,
I
don't
remember
version
0.2
and
again
the
the
project
is,
you
know
when
it
first
started,
it
was
I,
I
wrote
a
lot
of
code,
but
for
the
last
couple
years
it's
been
Community
Driven
completely,
which
is
great
and
folks
are
adding
things
that
they
want
to
see
and
and
the
project
is
healthy
and,
like
I
said
it's
getting
adopted,
I,
don't
know
if
there's
any
question,
but
I
can
just
talk
about
the
project.
If
there's
no
question.
B
So
if
you
don't
know
about
the
e2e
framework,
like
I
said,
like
I
mentioned,
it
is
a
Standalone
project,
not
Standalone,
in
a
sense
that
it's
it
helps.
You
write
tests
against
a
running
cluster.
B
It
has
zero
dependencies.
As
far
as
external
framework,
you
basically
use
gold
test
tooling
to
run
your
tests.
Other
than
that.
It's
you
writing
your
your
tests
using
go.
B
The
project
comes
with
several
support
tools
to
help
you
write
your
tests,
one
being
it
has
built-in
support
for
starting
a
kind
cluster
as
you're
as
you
run
your
test,
so
you
can,
with
literally
one
line
of
code,
tell
it
to
start
kind
as
part
of
your
setup
sequence
for
the
test,
and
once
that's
done,
you'll
have
a
handle
to
the
to
the
client,
that's
already
configured
and
pointing
to
your
current
cluster
and
then
from
that
from
there.
B
You
can
write
your
tests
and
then
you
also
have
handles
to
kind
of
hook
into
the
shutdown
sequence
of
the
tests,
so
you
can
take
control
there
as
well,
and
obviously
the
tests
themselves
they're
written
in
pure
gold
test
functions.
B
The
only
thing
is:
there's
a
kind
of
a
build,
a
pattern
that
you
get
with
with
the
project
that
allows
you
to
do
some
further
setup
of
your
tests,
even
in
inside
the
test
function.
So
you
can
do
things
like
labeling
your
tests,
give
it
a
name.
So
you
can
see
in
the
report
which
test
was
running
and
you
know
the
main
thing.
B
Is
you
get
a
the
access
to
the
to
a
kubernetes
client
that
allows
you
to
run
your
tests
and
have
access
to
the
to
to
the
to
the
API
server?
There
are
other
there
are
other
supporting,
tooling
or
apis
that
you
get
such
as
there's
a
weight
API,
which
is
a
a
wrapper
of
I,
think
it's
API
machine
weight,
but
it's
it's
a
little
bit
simpler
to
use
where
you
can,
during
a
test,
set
up
weight
conditions
to
wait
for
a
specific
condition
to
become
true.
Before
the
test
can
proceed.
B
B
Instead
of
just
having
to
read
the
example.
So
now
you
can
go
and
see
some
documented
stuff,
and
you
can
read
about
some
of
the
supporting
features.
B
Let
me
see
so
you
know
there
are
things
like
you
can
app.
When
you
run
your
test,
you
can
specify
how
you
wanna
the
behavior
of
the
test,
meaning
you
can
provide
skip
labels
or
inclusion
labels
of
of
the
tests
that
you
wanna.
You
want
to
see
executed.
B
A
news
this
morning
too
I'm
like
am
I
like
no
okay,
all
right
good
to
know
so,
like
I
said
there's
now
you
have
docs,
you
can
read
and
and
jump
directly
to
different
examples
in
the
tests.
B
So,
like
I
said
you
could
do
filtering
you.
Can
somebody
added
examples
on
how
to
add
your
own
CLI
Flags?
If
you
want
to
even
more
custom
customization,
you
can
test
custom.
Somebody
provided
ways
to
do
to
test
crds
custom
resources,
I'm
just
going
down
the
list
of
examples.
B
Parallel,
you
can
do
you
can
execute
tests
in
parallel
and,
like
I
said,
you
can
do
well.
Resource
watch
is
another
thing
you
can
do,
but
you
also
have
conditions
where
you
can
set
up
conditions,
as
I
had
mentioned
earlier.
B
B
C
Assume
the
question
in
the
Channel's
been
answered
by
Ben,
maybe
for
folks
not
looking
in
the
chat
check
the
chat
for
a
question.
My
question
is
just
to
be
very
specific
and
clear
about
but
you're
talking
about
you're
talking
about
the
test
framework
used
in
the
kubernetes
test
directory
for
e2e
and
stuff
of
that
sort.
B
No,
that
is
that
was
one
of
them
driving
motivation
for
this,
because
we
didn't
want
to
because
the
the
test
that's
in
kubernetes
kubernetes
has
its
own
framework.
It
uses
I,
think
it
uses
Ginkgo
and
it's
very
closely
coupled
with
kubernetes
itself
and
if
I'm
creating
a
project,
and
at
least
two
years
ago,
when
this
started.
That's
how
that
was
the
state
of
things.
I,
don't
know.
B
If
things
have
changed
since
then,
and
it
was,
you
know
a
bit
difficult
to
to
kind
of
vendor
that
so
this
project,
the
one
of
the
motivation
of
this
project,
is
hey.
You
don't
need
to
you,
don't
need
trying
to.
You,
know,
reuse.
What's
in
kubernetes,
because
again,
it's
very
tightly
coupled
and
organically
grown
with
kubernetes.
C
D
Was
actually
a
cap
about
staging
the
existing
ewe
framework
thing,
the
old
one
and
like
this
was
kind
of
like
building?
This
was
kind
of
suggested
as
an
alternative.
D
For
Vladimir,
taking
that
over
I
think
there's
a
lot
of
Tech
that
organic
growth
is
a
nice
way
of
putting
it
in
the
existing
UWE
framework
and
we've
made
work
to
improve
it,
but
that's
kind
of
at
its
own
pace
and
it's
a
lot
more
challenging
with
the
thousands
of
existing
tests
cases
absolutely
there's
a
lot
of
Oddball
kubernetes
type,
half-baked
Concepts
in
there
around
things
like
providers
that
are
going
to
take
a
lot
of
effort
to
extricate,
and
you
could
build
something.
D
A
lot
lighter
if
you
could
start
from
scratch
today,
like
this
project,
has
because
the
rest
of
the
ecosystem
has
evolved
like
when
kubernetes
adopted
Ginkgo
early
on
go
didn't
have
a
way
to
do
control
flow
for
like
cleanup
right.
They
had
added
test
main,
but
that
was
it
is
test
made
and
test,
and
you
had
no
way
to
do
things
like
cleanup.
You
had
no
way
to
get
structured
output
out
of
go
tests,
so
I
think
even
Ginkgo
didn't
have
that.
D
How
kubernetes
wanted
and
there's
a
custom
junit
reporter
until
Ginkgo
V2
right,
there's
been
some
really
major
changes
to
the
entry
one
to
switch
to
like
Ginkgo,
V2
and
clean
up
a
bunch
of
this
stuff,
but
it
still
has
a
ton
of
tech,
debt
and
I
would
say
for
most
projects
unless
you
like
this
particular
style
and
want
to
use
Ginkgo,
Inc,
Omega
and
things
I.
D
Think
it's
a
little
bit
more
go
like
to
use
this
e2e
framework
that
Vladimir
has
been
building
that
you
know
using
some
of
the
more
modern
go
Concepts
that
doesn't
depend
on
all
these
really
heavy
dependencies
right.
B
Yeah,
thank
you
for
for
adding
that
and
and
that's
exactly
what
it
is.
It's
it's.
It
was
kind
of
a
reset
of
of
what
was
and
a
lot
of
the
concepts
that
are
in
here.
You
know,
I
took
when,
when
we
were
first
starting
took
a
good
few
days
to
look
at.
What's
what
folks
have
been
doing
in
inside
kubernetes
kubernetes
and
there's
a
lot
of
good
stuff
in
there,
you
know
stuff
like
the
ability
to
wait
for
a
condition.
We
wanted,
that
the
ability
to
construct
objects
with
fragments.
We
wanted
that.
B
So
some
of
this
stuff
is
in
there
the
ability
to
you
know
to
to
test
crds
to
make
sure
that
you
know
you
can
actually
not
just
workload
objects,
but
also
any
kind
of
custom
objects.
Yes,.
C
B
B
B
So
originally,
the
intention
was
not
to
basically
leave
kubernetes
kubernetes,
as
as
it
is
because,
as
Ben
mentioned,
it
has
its
own
Cadence
its
own
pace.
This
is
more
for
folks
who
are
building
tools
in
one
an
easy
way
to
test.
You
know
if
you're
building
on
top
of
kubernetes,
you
want
to
test
whatever
you're
building.
This
would
be
the
tool
rather
than
you
going
and
recreating
your
test,
your
own
test
framework
and
end
up
at
a
place
very
close
to
this.
B
Why
not
just
use
this
to
as
a
starting
point,
so
that's
the
intended
audience,
but
the
the
stuff
that
maintained
testing
for
kubernetes
kubernetes,
completely
separate
but
I
mean
in
the
you
know
some
sort
of
future.
If
somebody
is,
you
know
brave
enough
and
have
a
lot
of
time,
and
they
want
to
explore
using
some
of
this
for
kubernetes
kubernetes
they're
welcome
to,
but
for
now
the
two
projects
are
intended
to
be.
You
know
to
be
separate.
D
We
have
an
enormous
mountain
of
tests
in
kubernetes
kubernetes
that
so
it
that
stuff
is
mostly
in
maintenance
mode
and
we've
had
some
iterative
improvements,
but
it's
not
something
where
we
can
go
back
and
say:
oh,
let's
build
something
for
everyone
to
use
that
that's
this
project,
so
that
is
a
bit
different.
D
The
test,
the
the
test
framework
in
kubernetes
kubernetes,
is,
is
used
by
the
tests
in
kubernetes
kubernetes
for
testing
kubernetes
itself,
with
the
exception
of
the
conformance
program,
those
tests
aren't
super
useful
if
you're
doing
other
things
other
than
developing
kubernetes
itself.
The
the
conformance
tests
on
a
handful
of
the
other
tests
may
be
useful
if
you're
developing
like
a
kubernetes,
distro
or
something,
but
if
you're
developing
kubernetes
applications,
then
you
you
would
want
to
be
looking
at
Vladimir's
project
instead
of
instead
of
the
the
the
things
we
have
in
kubernetes
kubernetes.
D
There's
a
couple
of
good
ideas
in
in
those
tests.
There's
a
bunch
of
bad
ideas
as
well,
and
a
lot
of
TechNet
and
porting
it
to
another
framework
is
not
a
task
that
I
think
we
have
stacked
in
the
foreseeable
future.
D
There's
just
so
many
tests,
I
think
the
biggest
thing
that
we're
doing
with
that
is
just
trying
to
keep
them
rolling
forward
with
the
ecosystem
as
much
as
we
can
and
trying
to
decouple
them
a
bit
from
even
some
of
the
tests
and
kubernetes
kubernetes
are
like
coupled
to
the
the
cluster
tools
like
Cube
up.sh,
and
would
don't
we
didn't
want
that
concept
coming
back
as
well,
but
it's
going
to
be
very
hard
to
remove
that.
There's.
D
A
really
large
context,
object
that
gets
passed
around
in
the
entry
framework
that
just
has
a
snarl
of
fields
in
it,
but
trimming
those
out
is
a
lot
of
work
and
it
you
know
it
doesn't
have
super
obvious
return
on
investment
for
people
versus
you
know
just
working
on
a
clean
slate
for
developing
applications
and
then
add-ons
and
so
on.
B
B
Yeah
I
see
I
was
going
back
to
chat
and
I,
see
that
Brady
mentioned
that
he
is
using
the
existing
utility
test
framework.
Thanks
for
that,
and
the
thing
that's
very
interesting
that
he
mentioned
is
that
he
is
using
it
with
ginco.
B
So
one
of
the
plus
sign,
plus
side
of
taking
the
simplified
pure
go
test
function
is
that
you
can
integrate
that
test
in
any
existing
tool
that
you
already
have,
because
again
it's
it's
just
regular
pure
go
function,
pure
gold
test
functions
so
that
you
can
you
know
if
you're
already,
if
you
like,
Ginkgo,
you
can
use
it
with
Genco,
you
can
actually
you
could
possibly
not
even
use.
B
So
if
I
don't
know,
if
anybody
has
the
project
pulled
up,
but
there's
a
specific
package
name
client
with
a
K
and
I
intentionally
pulled
that
out
with
the
intent,
maybe
at
some
later
Point
down
the
line.
Maybe
pull
that
to
be
its
own
thing,
its
own
simplified
way
of
managing
kubernetes
client
goal
client,
but
you
could
use
that
with
your
existing
test
framework
and
you
get
the
same.
B
You
know
you
get
the
same
benefits,
even
if
you
didn't
want
to
use
E2
the
this
project
need
to
be
test
framework.
So
that's
that
that
that's
good
to
hear
that
somebody
is
using
it
with
Ginkgo
and
the
other
thing
I
I
didn't
mention,
but
past
past
releases.
One
of
the
thing
I
worked
on
in
the
previous
release
was
the
the
integration
with
Cube
test2.
B
So
you
can,
if
you
didn't,
want,
if
you're
already
using
Cube,
Test
2
or
you,
you
know,
you're
looking
if
you're
already
using
Cube
tester,
you
could
still
use
E2
test
framework,
you
let
keep
test2
set
up
your
and
tear
up
down
your
environment
and
and
then
E2
test
framework
will
actually
run
your
tests.
Yeah
go
ahead.
C
Yeah
Brady
that
Brady
Vladimir,
it's
you're
you're
mentioned
client
with
a
K
that
sounds
cool
and
it
sounds
like
you're
drawing
attention
to
it.
I'd
recommend
you
put
a
section
in
the
readme,
you
know
talk
it
up
a
little
bit
because
yeah.
B
B
Yeah,
that's
a
good
idea.
I
probably
should
add
it
in
the
dock
that
the
intent
of
that
package
is
to
show
that
you
can.
There's
I
wanted
to
get
to
a
point
where
we
have
a
another
layer
of
abstraction,
on
top
of,
what's
available
with
a
client
go
and
that's
what
client
with
a
K
does,
is
it's
kind
of
it
high?
Actually
what
it's
one
of
the
things
it's
doing,
it's
reusing
God!
B
What's
the
it's
an
it's
reusing,
another
project
and
the
name
controller
runtime,
so
the
client
there
I
like
what
they
were
doing
so
I
was
like.
Let
me
wrap
some
of
that
and
hide
some
of
the
complexity
so,
whether
or
not
you're
using
a
type
type
client
or
generic
client,
it
doesn't
matter
it
hides
all
that
it
makes
that
decision
for
you
and
hides
it
for
a
and
gives
you
something
you
can
just
use
and
what
else
is
in
there.
B
The
all
the
like
some
folks,
some
one
of
the
a
contribution
I
think
it
was
last
year
that
somebody
did
was
the
ability
to
encode,
basically
any
kind
of
arbitrary
yaml
chunk.
You
can
pass
it
to
the
client
and
it'll
give
you
the
action.
It'll
realize
the
objects
for
you
so
which
is
very
useful
in
tests,
especially
if
you
have
something
that's
very
structured
and
you
don't
want
to
take
time
to
kind
of,
or
you
don't
even
you're,
not
even
sure
what
the
structure
is.
You
can
pass
it.
B
The
the
yaml,
Chuck
and
it'll
give
you
a
realized
object.
So
there's
a
lot
of
good
things
in
there
that
I'm
like
well
I
would
like
to
see
this
in
a
generalized
API,
not
just
for
testing.
So
but
it's
a
good
fertile
ground
to
kind
of
play
around
with
these
ideas-
and
you
know,
hopefully
time
permits,
push
them
further
up
the
stack
as
far
as
maybe
you
know,
other
cigs
that
may
be
interested.
B
Yeah,
so
if
there's
any
not
anymore
questions,
let
me
see
look
at
the
yeah,
it
doesn't
look
like
there's
any
more
questions.
I.
B
B
A
All
right
yeah
last
minute,
if
anybody
has
anything
else,
that
they
want
to
talk
about,
feel
free,
otherwise,
I
think
we
can
call
it
good
for
today.
D
I
just
want
to
add
if
folks
are
interested
in
the
topic.
This
applies
more
to
the
entry
framework,
but
we
actually
just
had
some
Publications
that
had
a
kubecon
about
like
how
to
write
good
ewe
tests
and
best
testing
practices.
I
think
there'll
be
some
overlap
there
pretty
interesting
from
one
of
our
recent
leads
Patrick
and.
B
D
D
If
you,
if
you
can
catch
some
time
with
them,
I
would
recommend
connecting
up
with
Antonio
and
Patrick,
who
are
our
most
recently
promoted.
Tech
leads
in
the
Sig.
They
both
were
I
nominated,
both
of
them
because
of
all
their
work
on
the
entry
framework,
which
suddenly
got
a
ton
of
work
past
couple
years,
I
still
wouldn't
recommend
that
people
like
reuse
that
thing
but
I
think
there's
been
some
good
ideas
brewed
in
there.
D
Okay,
like
I,
was
I
was
just
poking
around
in
here
now,
and
one
thing
that
comes
to
mind
that
I
would
suggest
you
might
want
to
poke
some
folks
about
and
and
see
about
kind
of
copying.
The
idea
into
here
is
the
the
host
exec
pod
trick.
I
didn't
see
anything
for
that
poking
around
the
current
state
of
the
the
kubernetes
six
ew
framework,
and
that's
been
a
really
useful
thing
that
we've
moved
towards.
B
D
B
Of
the
contribution
it
was
about
pod,
exec
and
I'm
wondering
if
whoever.
D
I
I
need
to
I
need
to
find
the
the
exact
pod
stuff
again
dig
that
up.
D
It's
been
a
little
while,
since
I
looked
at
that
yeah
the
basically,
the
idea
is
that
you
can
use
a
privileged
pod
with
the
right
configuration
combined
with
NS
enter
to
to
using
only
the
kubernetes
API,
no
SSH
access
to
to
execute
things
on
the
on
the
host
and
we've
been
migrating
towards
that
as
one
of
the
ways
we're
trying
to
get
rid
of
all
the
broken
provider,
contexts,
okay
and
the
core
framework-
and
this
has
been
working
really
well
for
us-
you
basically
just
create
a
pod
that
is
crafted
to
allow
you
to
get
away
with
exacting
things
as
if
you
were
just
executing
directly
to
the
underlying
node.
D
That
is
a
really
useful,
pretty
portable
trick.
You
know
it
won't
work
on
all
clusters,
they're
going
to
be
clusters
that,
like
block
this
sort
of
behavior,
but
if
you
have
sufficient
privileges
just
using
the
kubernetes
API,
you
can
do
that
so
almost
everywhere
that
we
were
using
SSH
to
do
things
like
checkmates
nodes,
away,
move
this
I
think
for
like
application
devs.
This
is
less
relevant,
but
for
people
that
are
building
like
critical
add-ons
for
things
like
storage
or
networking.
D
This
is
a
really
useful
trick
that
we've
been
adopting
in
the
core
framework.
B
Okay,
yeah
that
that's
that
would
be
interesting,
I'll
I'll
dig
is
that
I'm
not.
D
Sure,
yeah,
that
that
link
in
the
that
Brady
posted
is
okay.
D
Is,
oh
sorry
that
just
shorthand
for
being
in
the
in
the
main,
kubernetes
repo?
We
talk
a
lot
about
projects
being
entry
or
or
out
of
tree.
D
Six
repo
would
be
out
of
tree.
We
try
to
encourage
a
lot
more
things
to
be
out
of
tree
now
right,
so
they
can
develop
at
their
own
pace.
So.
B
Someone
contributed
in
this
release
something
called
exec
and
pod,
so
I'm
wondering
if
they're
you
reusing
I'll,
look
at
the
code,
but
I'll
dig
to
see
if
it's
the
same
thing,
because
it
may
already
be
in
in
have
been
contributed
to
to
I.
D
That's
been
crafted
to
allow
you
to
escape
to
the
host
you're
like
intentionally
from
a
pod
to
the
host,
so
that
you
can
so
you
can
do
things
like
you
can
expect
mounts
or
the
network
state
or
whatever,
but
without,
but
in
a
poor,
in
a
relatively
portable
way
that
that's
only
using
the
kubernetes
API
and
isn't,
depending
on
out
of
band
things
like
SSH
access
and.
B
It
is,
is
Patrick
discussing
that
at
all
in
the
blog
post.
D
D
The
other
controllers
tend
to
not
need
to
do
this,
but
you
can
imagine
needing
to
do
this
for
developing
things.
Like
storage,
add-ons,
yeah.
B
Okay,
I'll
dig
through
that
length
that
under
hosts
exec
util
to
see
what
yeah,
because
one
other
thing
I
wanted
to
do
and
I'm
pushing
forward
with
this
project,
too,
is
to
kind
of
integrate.
I
guess,
integrate
kind
of
a
strong
word,
but
make
sure
that
we're
we're
leveraging
other
tools
like
somebody
is
already
looking
into
using
the
new
the
quack
project
as
a.
C
B
Instead
of
running
kind,
which
I
think
quack
kind
of
started,
something
like
kind
in
the
background,
but
quack
we're
looking
into
integrating
with
Quark
we've
done
it
with
Cube
test
two,
so
anything
that
helps.
You
know
that
that
stays
and
and
reach
out
in
in
the
ecosystem.
It
makes
it
easy
I'm.
Definitely
looking
into
that.
Yeah.
D
I
think
we
have
a
couple
of
just
a
few
of
these
actually
good
ideas
that
are
in
the
that
are
in
the
entry
framework
that
may
make
sense
to
clone
over
here
that
the
host
exec
is
one
I,
think
you
have
a
lot
of
the
other
things
like
that.
How
like
how
to
Wade
on
things
properly.
Most
of
what
Patrick's
talking
about
here
is
just
like
the
approach
to
doing
things
reliably.
Where
is
this
more
of
like?
B
Yeah
I
do
remember
way
back
when
he
did
wrote.
He
wrote
a
previous
blog
that
was
kind
of
in
the
same
path,
but
it
seems
like
now
it's
it's
even
more
taking
it
further.
B
Yeah
I
see
he's
talking
about
polling
and
timeouts
long
waiting
times.
D
Yeah
I
think,
if
I
think
we
connect
you
with
Antonio
and
and
Patrick
Ayo,
Jaya
and
Pioli,
they
they
both
have
done
a
lot
of
work
on
the
entry
one
recently
and
there's
probably
a
few
other
ideas
like
this.
That
I'm,
not
thinking
about
thought,
my
head,
that
are,
are
worth
looking
at
and
seeing
if
they
can
be
of
use
for
the
yeah.
B
Absolutely
framework
Absolutely
I'll,
definitely
reach
out
to
Patrick
when
I
was
doing
storage
way
back
when
we
crossed
paths
so
I
definitely
reach
out
to
him
yeah
and
I'm.
Looking
at
the
blog,
it
looks
like
some
of
the
stuff
like
it
looks
like
they're
directly
built
in
can't
go
now
like
the
waiting
and
the
polling
and
stuff
like
that.
Yeah.
D
We
actually
had
the
the
author,
the
core
maintainer
came
and
interacted
with
suggesting
for
a
while
over
the
past
year,
leading
into
Ginkgo
V2
to
like
get
some
feedback
and
took
a
lot
of
suggestions,
and
you
know,
is
it
fun,
because
a
lot
of
us
have
had
some
frustration
with
it
over
the
years,
but
it
was
so
good
back
and
forth
and
and
the
Corp
the
framework
in
the
kubernetes
repo
has
adopted
some
of
these
Ginkgo
V2
improvements
now
and
I
think
you
know,
there's
good
ideas
in
in
the
midst
of
the
mess.
B
No,
absolutely
absolutely
I
mean
I
can
see
a
little
bit
of
overlap
with,
what's
already
in
in
the
e2e
firmware,
but
definitely
stuff
like
exact,
no
pot,
exact
and
running
stuff
directly
on
the
Node.
We
definitely
need
to
look
into
that
and
I'm
always
looking
for
it
for
ideas
anyway.
So
that's.
D
The
trickiest
one
off
the
top
of
my
head,
that
was
I've,
been
like.
Oh,
we
should,
like
you
know,
document
this
better.
This
is
a
good
idea.
I
mean
it's
a
good
bad
idea.
It's
a
little
confusing
yeah
I'll
be
abusing
some
functionality,
but
I
would
guess
that
between
the
two
of
them,
there's
a
couple
more
of
these
that
I'm,
not
thinking
of
since
I,
haven't
spent
as
much
time
on
it
recently
cool.
B
All
right
awesome,
thank
you
for
the
for
the
heads.
D
Up,
thank
you
for
working
on
this
I'm,
really
glad
that
there's
a
place
for
people
to
collaborate
on
the
auto
tree
one.
My
focus
has
usually
been
on
kubernetes
itself,
but,
like
you
know,
we
want
to
be
able
to
provide
good
practices,
but
every
time
someone
would
come
and
talk
to
us
we're
like
don't
use
the
thing
where.
B
A
All
right,
I
think
we're
good
to
call
it
there
thanks
everyone
thanks
for
the
Lively
discussion
and
have
a
good
rest
of
the
week,
take
care.