►
From YouTube: KubeVirt Community Meeting 2022-08-31
Description
Meeting Notes: https://docs.google.com/document/d/1kyhpWlEPzZtQJSjJlAqhPcn3t0Mt_o0amhpuNPGs1Ls/
A
A
B
Might
be
helpful
to
other
members
of
the
community,
let's
see
Andre,
it
looks
like
you
said,
or
Angie
you're
having
access
trouble
to
the
doc.
If
you
can.
B
All
right
looks
like
everybody's
logging
attendance,
so
thank
you
for
that
definitely
add
anything
to
the
agenda:
an
open
floor
that
comes
to
your
mind,
either
now
or
during
the
call,
and
we
will
Circle
back
to
that.
B
A
Hi
I'm
new-
this
is
Bob
Adair
I'm
joining
from
from
cataloging
software
I'm,
just
here
to
kind
of
listen
into.
What's
going
on
we're
doing
some
work
on
we'll
get
catalogica
provides
backup
and
data
protection
software
and
we're
looking
into
how
we
may
add
support
for
Cuba
to
our
our
data
protection
software.
So.
A
B
It
looks
like
today's
agenda
might
be
a
bit
late,
usually
is
towards
the
end
of
the
summer
when
everyone's
getting
ready
to
head
back
into
their
normal
routines.
But.
B
So
we
have
a
talking
point
on
aligning
Cooper
with
kubernetes
release
guidance
I'm,
going
to
pull
that
link
up
I,
don't
see
any
initials
on
who
submitted
that
one.
Does
anyone
want
to
talk
to
this.
D
Sure,
hey
it
says:
Ryan
I
I
had
this
and
I
started
to
throw
out
a
bad
one
on
the
mailing
list,
and
so
my
my
pitch
to
you
know
the
community
and
what
I
wanted
to
discuss
here
and
then
this
pull
request
is
so
the
you
know.
The
current
state
of
the
release
process
in
Cuba
is
like
everyone
knows,
it's
a
monthly
process,
and
basically,
what
I'm
proposing
is
is
changing
it
and
changing
it
to
match
what
kubernetes
does
and
then
a
number
of
reasons
for
why.
D
This
is
a
good
idea
that
I
have
been
listed
in
the
pr,
so
I'll
just
go
over
those
real
quickly
and
I
would
say
some
of
the
big
ones
are
like
when
and
look
at
the
current
say
the
project.
It's
it's
very
mature.
It's
production
ready.
You
know
many
vendors
use
it
and
in
this
its
current
state,
the
release,
Cadence
monthly
release
can
be
difficult
for
for
vendors
and
and
other
end
users
to
keep
up
with,
and
so
one
of
the
challenges
that
always
comes
about
with
this
very
very
quick
release.
D
Cycles
sort
of
the
question
of
like
when
you
know
when,
when
are
you
gonna
upgrade
to
the
new
version?
It's
something
when
things
are
monthly.
This
becomes
a
very
challenging
question
to
answer,
because
there's
always
a
new
version.
That's
got
something
that
you
maybe
want
every
month
and
well
that's
great,
there's
always
there's
always
other
there's
downsides
to
this
and
things
like
supportability
stability
and
and
things
like
that,
like
when
I
I
guess.
A
good
example
is
like,
if
I
wanted
to
back
Port,
something
that's
important
to
me.
Usually
wouldn't
end
up
happening.
D
D
Things
only
get
packed
for
in
a
few
releases
because
it
becomes
too
challenging,
and
so
what
I'm
proposing
is
to
slow
things
down
a
little
bit
to
make
this
process
a
little
bit
easier
for
for
a
whole
number
of
those
reasons:
portability
and
making
those
back
ports
easier,
making
it
easier
for
vendors
to
to
catch
up
and
consume
these
things,
and
and
so
the
process
that
kubernetes
has,
which
is
releasing
three
times
a
year.
Every
15
weeks
seems
pretty
sensible
and
they've
had
some
success
with
it.
D
So
my
thought
was
that
what
we
could
do
is
we
could
follow
the
same
cycle
and
we
could
Trail
kubernetes
with
their
releases.
So
the
idea
is
they
go
through
a
15-week
cycle
release
and
then
at
some
point
in
time,
maybe
a
few
weeks
later,
Cooper
would
follow
up
with
their
release.
And
then
you
know
we
have
supported
releases.
You
know
within
the
same
framework
we
do
now,
which
is
like
three
releases
with
kubernetes,
and
so
we
keep
the
same
promises.
D
B
Okay,
thanks
for
bringing
that
up,
do
we
have
feedback
from
anyone
on
the
call
right
now,
thoughts
that
might
contribute
to
scoping
some
of
the
call
or
further
the
conclusions
that
we
should
be
headed
towards.
E
Okay,
this
the
first
issue,
the
progress
one
is:
is
a
plan
from
Daniel.
E
How
we
can
add
cluster
and
I
have
done
the.
E
But
I
want
someone
can
review
this
issue.
F
E
Okay,
cool
cool
and
I
also
have
a
question
for
this.
We
have
another,
and
currently
we
have
one
server
for
the
kubernetes
community
and
it's
have
been
deployed
kubernetes
by
I.
Don't
know
how
to
pronounce
it
pronounce.
It
think
Lindsay
yeah,
but
I
I
saw
he
was
in
the
retired,
reviewer.
E
So
I
suppose
he
not
he's
not
your
contribution
to
cubie
World
Community.
Now
and
now
we
have
another
server
for
the
community.
So.
E
So
how
who
I
can
ask
to
to
like
Deploy
kubernetes
on
each
or
should
I
do
it.
F
I
I
think
that
you
are
talking
all
right,
you're
you're,
you
were
mentioning
Federico
Jimenez.
Is
that
correct?
Yes,
yes,
yes,
yeah
and
I
think
he
he
deployed
kubernetes
on
that
cluster.
To
be
honest,
I'm,
not
exactly
sure,
but
what
you,
what
you,
what
you
could
always
do,
would
probably
a
brown
eye
on
the
call.
A
F
Doing
hey
thanks
thanks
for
chiming
in
I
was
just
thinking.
Probably
if
Howard
was
asking
us
for
deployment
of
kubernetes
on
the
on
the
cluster
and
do
you
have
information
on
how
Federico
did
it
last
time?
Do
you
know.
A
I
was
taking
a
quick
look
at
this
a
couple
days
ago
and
it
looks
like
it
was
deployed
using
ansible
that
are
using
Cube
spray
similar
to
other
clusters,
so
we
do
have
some
ansible
defined
in
the
projecting
for
repo.
So
I
could
look
into
that
further.
A
That's
as
much
as
I
know.
Right
now,
I
was
meant
to
review
Howard's
ticket
as
well,
but
I
didn't
get
around
to
it.
Unfortunately,.
E
B
B
Last
time,
I'm
on
open
floor
right
now,
if
you
have
anything
else
to
add,
feel
free
to
drop
it.
While
we're
covering
this
one
ability
to
specify
exact
image,
names
for
cube,
vert
operator
and
just
like
CDI
is
doing.
C
Foreign
I
just
want
to
say
that
we
are
trying
to
implement
as
model
to
our
kubernetes
TV
desktop,
which
is
called
deck
house,
but
we
have
very
specific
Logic
for
building
and
pushing
images.
C
Yeah,
please
open
first
link
on
the
second
one
and
we
would
like
to
have
an
ability
to
specify
specific
image
names
like
for
all
the
components,
because
currently
it
is
not
possible
Cubit,
always
expecting
that
all
these
suffixes
will
be
existing
and
they
hard
code
it
and
there's
no
opportunity
to
change
them
and
actually
I
wrote
some
initial
implementation
with
just
changing,
takes
the
image
names
from
the
environment
variables,
but
I
think
it
might
be
nice
to
contribute
it
to
upstream
and
I
just
want
to
know.
C
G
Hey
I
had
a
question
regarding
this,
so
I'm
still
new
to
the
project.
I'm
curious.
There
is
a
so
the
operator
has
an
image
and
then
operator
reconciles
a
cubert
CR.
G
G
C
Actually
in
operator
there
is
already
a
support
of
environment
variables,
but
it
is
done
just
for
specifying
the
digest
for
the
images,
so
you
can
specify
specific
digest
for
every
image
and
you
can
also
specify
Cubit
version
there,
because
it's
calculating
the
version
number
out
of
the
image
tag,
its
own
image
tag,
so
I
just
changed
this
logic
to
specify
the
exact
image
names.
Of
course,
most
of
the
operators
are
doing
that
by
the
specifying
it
in
a
custom
resource
in
custom
resource
of
cubers.
C
Also,
there
is
an
opportunity
to
override
the
repository
name,
image,
prefix
and
tag,
but
the
problem
is
it
always
expecting
images
under
a
specific
path,
for
example,
virt
API,
weird
operator,
weird,
Handler
and
stuff,
like
that
I
need
actually
an
opportunity
to
overwrite
these
images
fully
for
every
component.
G
Right
yeah,
so
what
I
was
wondering
is
like?
Can
we
put
that
in
Indus
pack,
instead
of
the
environment
variable
that
way?
Only
the
operator
image
goes
in
the
environment
variable
and
the
spec,
which
can
change
on
the
Fly,
can
populate
different.
G
C
Yeah
and
I
like
this
idea
to
putting
it
into
shirai,
because
there
is
2cr
to
custom
resource
sorry,
because
there
is
a
tool
with
vertical,
which
is
also
downloading
images
according
to
the
configuration
from
Custom
resource,
for
example,
if
you
use
a
command
vertical
guests,.
C
Fish,
it
will
download
image
using
standard
Cube,
vert
repository
name
tag
and
with
the
hard-coded
suffix
guest
tools,
so
I
think
I
think
it
would
be
nice
to
have
this
such
opportunity.
What
do
you
think
folks,
and
if
you
are
into
that
I,
will
prepare
PR
or
should
I
make
some
proposal
before
what
them
actually
should
I
do
with
that.
B
Sounds
like
going
direct
to
PR
is
probably
a
good
idea.
The
switching
it
from
environment
variables
to
spec
probably
makes
sense,
and
if
I
am
speaking
out
of
turn,
anyone
please
go
ahead.
Jump
in
and
correct
me
and.
H
Can
you
hear
me
yes,
yep,
oh
hi,
it's
evil
I
wanted
to
address
the
question
of
a
live
Hotel
regarding
whether,
if
it's
possible
to
modify
the
components
that
are
being
controlled
by
weird
operator,
so
it
is
possible.
You
can
see
the
customized
components
field
in
which
you
can
specify
in
adjacent
patch
format
what
you
would
like
to
change
in
a
particular
deployment
or
a
diamond
set.
Currently
we
can
patch
only
deployments
and
Diamond
sets
only
we
gonna.
There
is
no
support
for
patching
other
kinds
of
resources.
For
example,
Prometheus
rules.
H
We
don't
have
support,
for
example,
patient
this
kind
of
custom
resource,
so
yeah
you
can
use
customized
components
and,
regarding
your
suggestion,
are
doing
yes.
I'm
do
some
make
sense
if
it
is
required
in
your
case,
to
have
this
kind
of
control
over
the
full
as
the
full
name
and
the
full
parts
of
the
of
the
containers
that
are
provided
by
Cuban
so
yeah.
C
C
B
B
B
F
I
I
can
charm
a
little
bit
into
that.
Probably
we
have
this
popping
up
every
now
and
then
I
think,
and
we
were
looking
into
probably
adding
some
container
scanning
Tools
in
to
our
regular
build
process.
F
Unfortunately,
we
didn't
get
to
that.
For
now.
We
are
just
currently
just
evaluating
because
the
cncf
provides
a
Snick
and
that's
something
that
we
were
looking
into,
which
then
probably
should
yeah.
How
should
I
say
it
should
show
all
the
issues
that
they
are
currently
creating
manually
so
yeah
for
now,
I
think
it's
always
good
to
to
answer
all
those
issues,
but
I'm
not
exactly
sure
if
we
should
contact
that
to
a
night
Milestone
or
something
like
that.
What
what
do
the
rest
of
the
people
on
the
call
think.
H
Yeah
I
I,
totally
agree
I
think
we
should
have
some
automated
way
or
some
policy,
so
we
can
address
such
such
events
and
do
layout
to
respond
on
them,
I
think
as
soon
as
possible,
because
we
don't
want
to
be
vulnerable.
F
Another
way,
probably
what
I
would
be,
what
would
be
some
some
kind
of
solution
at
least,
would
be
something
to
try
to
automate
the
golang
update,
but
I'm
not
sure
how
how
valid
of
a
path
that
is
so
I
I,
don't
know.
H
I
think
for
the
short
term
at
least
we
have
to
bump
it
manually,
I
think
so
too,
on
the
shoulder
for
the
long
yeah
for
the
long
term,
we
should
think
about
some
kind
of
automation
or
that
will
connect
the
scanning
tool.
The
review
of
the
CVS
related
to
the
tool
chain
and
creation,
a
job
that
will
create
an
automatic
buff.
F
H
I
remember
it
tomorrow,
I
was
documenting
it.
We
do
have
something
you
just
need
to
look
at
the
dogs.
F
Oh
there,
it
is
I'm
just
going
to
drop
the
link
into
the
into
the
chat.
A
F
F
I
think
that
could
probably
still
be
a
good
place.
I
think
it's
a
valid
stuff
to
to
just
add
and
do
an
issue
but
yeah
in
general.
Normally,
of
course,
you
would
get
more
direct
reaction,
probably
when
you
would
would
try
the
slack
Channel
but
yeah,
but
for
for,
when
you
have
more
stuff
inside
that
you
want
to
want
to
share
like,
for
example,
logs
and
so
on.
I
think
that
an
issue
would
be
the
perfect
place
for
that.
A
B
B
Do
we
have
something
like
audit
logging
on
VMI
events?
I,
don't
remember
what
actually
lives
in
the
operator
logs.
B
B
F
Think
this
looks
like
a
follow-up
I
think
Lee
yarwood
is
the
one
who
created
this
one
and
I
think
he's
actually
the
one
who
is
actually
working
on
the
deprecation.
If
I
understand
correctly.
Okay,.
B
B
Okay,
if
someone
feels
inspired
and
can
jump
on
that
and
provide
any
useful
feedback,
that
would
be
fantastic,
I'm
sure
their
Community
would
appreciate
you
for
it.
F
And
other
sorry,
the
the
question
was
whether
what
on
What
kubernetes
versions,
We
Run,
our.
B
Tests
is
that
correct,
no
I'm
wondering
what
the
VMI
spec
and
image
used
in
CIS.
F
F
A
A
A
B
Surprising,
but
actually
I
think
that
originated
from
a
fedora
update
issue,
so
I
I
can't
jump
in
and
recreate
this
in
a
timely
fashion.
Right
now,
I
don't
have
time
to
fix
my
hypervisor,
even
even
though
I
need
it
right
now.
B
B
F
B
All
right
and
with
that
that
concludes
the
last
community
meeting
of
August.
So
thank
you
all
for
joining
and
participating.
Welcome,
Andre
and
Bob
I
will
see
you
all
same
time
same
place
next
week.