►
From YouTube: Kubernetes UG VMware 20220106
Description
December 6, 2022 meeting of the Kubernetes VMware User Group. Recent news and updates related to the vSphere platform (log4j, upcoming deprecation of ESXi boot from SD Card and USB, recent updates to Kubernetes Cloud Provider and CSI storage driver). Followed by free form discussion
A
It
looks
like
we've
got
a
little
bit
light
attendance
today,
just
because
this
is
a
week
that
included
a
holiday
in
much
of
the
world,
but
I've
added
a
few
items
on
the
agenda
and
then,
if
we
get
through
those
and
people
want
to
chat
about
whatever
we
can
go
ahead
with
that
apologies
in
advance.
If
I
get
background
noise,
my
neighbor
is
having
a
roof
replacement,
so
it
seems
quiet
at
the
moment,
but
there
have
been
outbursts
of
hammering
and
sawing
and
things
breaking
out
over
the
past
couple
days.
A
Security
issues
seem
to
have
gone
out
pretty
big
by
my
observation.
This
thing
is
getting
up
there
close
to
the
old
original
heart
bleed
in
terms
of
how
many
things
are
out
there
with
exposures
and
vsphere
is
no
exception.
A
The
company
put
out
a
blog
post
summary
in
mid-december,
so
I've
got
a
link
to
it
there,
where
I'm
not
going
to
read
it
to
you,
but
there's
a
lot
of
information
in
there,
and
the
summary
is
that
vsphere
has
some
exposures
as
well
as
a
lot
of
other
collateral
vmware
products
that
are
often
used
in
conjunction
with
vsphere.
A
There
is
a
q,
a
thread
which
is
being
updated
periodically.
I
looked
at
it
this
morning
and
the
bottom
line
is
that
we're
working
on
patches
but
they're.
At
this
point
in
time,
we
can't
say
that
there
is
a
100,
comprehensive
solution
to
address
all
aspects
of
the
log
for
j
exposure
in
all
products.
So
the
q,
a
thread
like
I
say,
has
the
updates
on
the
latest
data
and
then
for
the
legalese
summary,
the
source
of
truth.
A
A
A
A
Another
thing
that
this
this
isn't
quite
in
the
category
of
the
first
two
items,
but
just
making
people
aware
this
is
something
that
I
took
some
time
off
at
the
holidays
and
updated
my
home
lab
and
discovered
that
vmware
has
put
out
an
advisory
saying
that
we're
going
to
deprecate
booting,
esxi,
hosts
off
of
sd
cards
and
usb
devices,
something
that
in
an
earlier
area
was
popular,
and
I
think
there
are
legacy
situations
where
pieces
of
hardware
are
booting
off
those
devices.
A
If
you're
not
aware,
these
all
forms
of
ssds
tend
to
have
a
phenomena
where
they
can
wear
out,
you
know
too
many
writes
to
the
flash
memory
cause
it
to
eventually
fail.
It
maybe
gets
worse
over
time
like
some
of
the
technologies
that
improve
the
storage
density.
A
A
They
move
to
it
because
you
can
get
much
higher
amounts
of
storage
for
a
given
price,
but
these
sd
cards
and
usdb
devices
tend
to
use
the
cheapest
available
technology.
So
they
are
not.
You
know:
they're
they're
not
like
using
enterprise,
ssds
and
nvme
or
sas
formats.
They
tend
to
be
really
cheap
failure
prone
things
and
bottom
line.
A
For
my
own
personal
story
is,
I
ended
up
doing
my
updates
in
my
home,
lab
rebooting,
rebooting
them
and
finding
that
indeed,
I
had
one
of
these
devices
wear
out
and
the
thing
wouldn't
reboot,
so
there's
a
decent
summary
there
of
what's
involved,
including
a
william
lam
blog
post.
If
you're
on
home
lab
scenarios
like
intel
nooks
or
in
my
case,
I'm
using
kind
of
ancient
dell
servers,
but
you
know
they
they
come
from
that
era
where
some
of
those
were
actually
sold.
A
You
know
from
the
factory
with
esxi
in
pre-installed
on
sd
cards,
so
I
think
if
you're
running
five-year-old
servers,
there
are
definitely
things
out
there
like
that.
So,
anyway,
just
in
case
it's
useful
to
people,
I
threw
that
out
there
there
was
an
update
to
the
vsphere
cloud
provider
that
came
out
in
december.
A
A
If
anybody
is
out
there
still
running
those,
I
think
that
would
commonly
be
a
scenario
of
kind
of
being
on
a
vendor
platform
with
an
older
kubernetes
release
that
you
don't
want
to
take
the
you
know,
you
want
to
kind
of
stick
with
the
same
thread
of
what
you've
got
and
there
there
are
updates
for
those
that
address
the
issues
disclosed
at
that
link,
and
with
that
said,
that's
kind
of
all.
I
brought
to
this
meeting
so
I'll
open
up
the
floor
to
anything.
B
It
was
funny
you
were
talking
about
the
the
nux
running
into
exactly
the
same
problem.
Usb
sticks
failing
because
the
the
vsphere
7
boot
banks
are
very
right.
Intensive.
A
B
A
Yeah,
so
I
I
went
through
the
same
thing
myself
and
it
isn't
just
the
nook
see
like
I
say:
I've
got
like
dell
710s
and
things
and
tried
to
investigate
a
bunch
of
solutions
for
nooks.
A
The
william
lamb
article
is
good
and
for
that
writing
to
logs
frequently
issue
it
once
you've
been
burned
and
it
wore
out
it's
kind
of
a
little
late
to
retroactively
fix
that
problem,
but
you
can
go
down
to
command
line
or
powershell
and
move
where
the
log
file
goes,
so
that
it
really
the
boot
process
itself
only
needs
to
read
that
usb,
which
will
not
wear
it
out.
You
know
you
can
read
those
as
many
times
as
you
want
and
you
can
actually
by
default.
A
The
log
file
is
on
the
same
device
as
the
boot,
but
it
is
possible
to
configure
esx
to
move
the
log
to
another
device,
and
so,
if,
in
your
nook,
you
have
got
some
kind
of
better
storage
than
the
usb,
you
can
fairly
easily
move
the
thing
there.
It
took
me,
maybe
20
minutes
to
go,
find
the
knowledge
base
article
on
how
to
do
it
and
then
make
the
command
line,
calls
to
actually
like
make
it
happen,
and
I
did
have
to
reboot
a
couple
of
times.
A
I
think
to
to
do
this
plus
prove
that
it
actually
was
working,
but
it
wasn't
too
bad.
I
came
to
the
conclusion
that
pretty
much
all
of
the
usb
keys
you
can
buy
these
days
are
kind
of
trashy.
You
know
they've
gone
for
price
points
on
sd
cards,
because
people
use
them
for
professional
cameras
and
photography.
A
Now
this
I
don't
think
nooks
have
sd
cards,
but
maybe
I'm
wrong.
I
don't
actually
have
a
hook
myself,
but
in
the
flash
cards
they
have
ratings
of
of
these
things,
and
there
are
models
that
are
specifically
marked
high
endurance,
so
they
put
excess
capacity
there,
anticipating
that
things
will
wear
out
and
I've
also.
A
I
don't
know
whether
I
saw
blog
articles
with
people
alleging
that
buying
a
higher
capacity
device
than
what
you
need
is
a
viable
solution.
But
I
don't
know
that
I
buy
that,
but
you
know
that
the
thing
would
I
can't
see
how
it
could
steal
from
things
that
it
all
it
could
just
as
easily
still
be
in
use,
but
I
at
least
I
saw
somebody
write,
a
blog
claiming
that
that
was
true.
Maybe
if
you
formatted
it
for
less
than
the
actual
capacity.
C
A
C
So
if
you
have
a
large
ssd
because
it
doesn't
rewrite
like
when
you
delete
and
write,
it
doesn't
write
down
the
same
sector,
so
the
trim
spreads
it
out
over
all
the
sectors.
So
when
you
have
a
large
capacity,
it
spreads
it
out.
So
you
wear
your
nodes
out
over
all
of
them,
so
they
all
wear
at
the
same
level.
So
when
you
do
have
a
failure,
it
kind
of
like
they
all
fail
kind
of
at
the
same
time,.
C
That
you
are
writing
the
same
size
now,
if
you're
writing
more
data,
it's
going
to
wear
faster,
of
course,
but
if
you're
like
doing
one
gig,
the
16
gig
is
gonna
last
longer
than
the
four
gig
drive.
A
Okay,
so
at
least
that
might
be
sort
of
remedial
so
that
you
last
longer
before
you
get
burned
and
I
did
find
on
those
sd
cards,
they
have
a
bunch
of
rating
systems,
there's
three
of
them.
So
you'll
find
old
blogs
that
say.
Class
10
is
the
best,
but
it
turns
out
they
kind
of
technology
reached
the
pinnacle
of
the
top
rating
of
10
and
they
started
a
new
one.
That's
a
little!
U
symbol!
A
So
I
think
a
u3
is
kind
of
the
best
tech
for
these
those
sd
cards
now
and
here's
another
one.
I
think
really
would
work.
But
unfortunately,
at
this
point
in
time,
vsphere
kind
of
has
or
esxi
has
declared
that
all
usb
is
kind
of
in
this
taboo,
but
you
can
actually
buy
little
usb
adapters.
Let's
see,
unfortunately,
I
have
one,
but
I
think
it's
away
from
my
desk
at
the
moment
or
I'd
hold
it
up
to
the
camera,
but
it
actually
is
light.
A
It
looks
almost
like
a
very
fat
usb
flash
device
and
the
inside
it
will
hold
an
nvme
actual
sd
drive,
which
should
not
have
wear
issues
at
all.
You
could
actually
even
buy
an
enterprise
grade,
one
of
those
and
at
least
in
the
dell
710s.
I
believe
that
would
actually
physically
fit
because
the
usb
they
have
a
usb
socket
inside
the
chassis
that
has
plenty
of
room,
and
if
you
were
to
move
to
one
of
those
that,
where
you're
actually
using
a
legit
sd
drive
instead
of
kind
of
these
consumer
flash
drives.
A
There
actually
are.
If
you
have
room
amazon,
actually
has
these
little
usb
adapters
that
go
from
usb
to
a
sata
connector,
and
I
think,
once
again
you
could
put.
If
you
had
the
physical
room,
you
could
connect
those
to
a
sata,
ssd
or
actually
even
a
saturn
rotating
media
drive
would
probably
have
more
durability.
A
You
could
move
your
log
files
off
of
those
sds
if
you
have
any
decent
storage
at
all
inside
those
nooks
and
potentially
also
use
a
usb
gateway
adapter
to
get
it
to
an
actual
ssd
drive.
D
C
In
that
thing
I
was
gonna
say
was
you
could
store
your
logs
to
the
ram
disk?
So
as
long
as
you
don't
reboot,
but
then
you
could
also
set
up
like
a
syslogger
somewhere.
So
yeah
there
you
go.
A
Yeah
that
blog
post
actually
has
another
workaround
of
booting
off
the
network
and
then,
of
course
you
could
use
an
external
syslog
or
something
so
there
might
be
some
solutions,
but
it's
just
in
a
home
lab
if
it's
a
chicken
and
egg
you
know
like
I,
I
have
a
log
inside
log
server,
but
the
thing
is
it's
hosted
on
the
esx
itself.
So
if
the
sx
goes
down
that
isn't
going
to
do
any
good
for
the
log.
D
A
A
C
A
So,
anyway,
opening
up
the
floor
to
anything
else.
People
want
to
talk
about
even
including
nominations
for
meetings
future
meetings
of
this
group.
Another
thing
that
I'll
ask
for
help
on
is
that
the
kubecon
europe
is
coming
up,
and
I
know
at
least
robert
you
are
in
europe
and
scott.
I
don't
know
if,
where
you
are,
is
considered
europe
or
not,
but
close
enough,
it's
probably
closer
to
europe
than
the
other
kubecon
sites
that
will
be
may
14,
and
at
least
for
now
that
is
planned
as
a
physical
event.
A
Only
with
no
remote
attendance
in
valencia,
spain-
and
I
think
this
group
is
likely
to
submit
a
cfp
for
some
topic
to
be
covered.
If
anybody
here
has
suggestions
for
what
you'd
like
to
see
there,
or
even
if
you'd
like
to
volunteer
to
be
part
of
a
presentation
at
that
event,
I
open
I'm
open
open
to
ideas.
D
Yeah,
I
unfortunately
I
already
have
a
cfp
in
and
you're
only
allowed
to
speak
at
one
event
at
cubecon
one
talk,
so
I
can't
I'm
doing
one
on
carvel.
Hopefully
if
that
gets
accepted,
carvel
and
tce
but
yeah-
and
I
think
you
know
I
would
hopefully
I
will
be
there
live
coven
allowing
and
we'll
definitely
attend
the
meeting.
That's
for
sure.
B
Yeah,
I'd
love,
I'd
love
to
be
there,
but
I'm
I'm
not
gonna.
I'm
gonna
bet
on
it.
A
A
Certainly
what
might
happen
but
yeah,
let
me
know-
and
even
if
you
can't
be
there
yourself,
if
you've
got
ideas
for
my
trouble,
is
I've
been
putting
on
these
sessions
year
after
year
and
eventually
you
run
out
of
ideas
and
I'd
rather
have
ideas
generated
by
prospective
audience
members
than
me,
because
it's
kind
of
hard
it's
a
guessing
game,
if
you're
guessing
what
people
are
interested
in
rather
than
having
actual
audience
members
tell
you
flat
out
what
they'd
like
to
learn
about.
B
Yeah,
I
think
the
yeah
I
don't
know
I
haven't
really
been
thinking
about
kubecon,
but
the
I
mean
topics
we've
discussed
on
in
this
in
this
forum
before
are,
which,
which
would
always
be
interesting,
are
we
talking
about
before
like
availability,
you
know,
basically,
the
the
full
stack
of
how
you
build
a
highly
available
kubernetes
solutions
in
this
case
involving
you
know,
the
vmis
flavors
of
kubernetes
distributions,
but
the
basic
principles
are
broader.
B
A
Yeah,
I
I
think
I'd
kind
of
like
to
put
that
together
too
and
there,
but
there's
different
levels
where
you
know.
I
think
the
big
challenge
to
which
unfortunately
they're
yet
to
be
invented,
great
solutions
is,
is
the
whole
thing
of
drilling
down
through
the
layers
of
complexity
there
to
get
to
root
cause
analysis
and
to
have
the
thing
be
a
little
smarter
about
putting
together
things
that
these
different
abstraction
layers
that
you've
you've
got
going
on
a
session
on.
It
could
do
sort
of.
A
I
don't
know,
there's
still
value
in
it,
but
I'd
describe
it
as
superficial
to
go
at
it
like
a
user
panel
where
people
are
more
or
less
summarizing
the
pain
points
but
kind
of
the
more
valuable
one.
If
you
could
pull
it
off,
would
be
something
maybe
getting
into
an
intro
on
the
pain
points
and
actually
coming
up
with
solutions.
B
Yeah,
but
just
I
mean
even
just
creating
awareness
about
about
how
to
how
to
architect
systems
like
that,
because,
as
we've
discussed
before
right
most
most
most
people
only
have
like
a
they've
only
have
a
visibility
on
a
single
slice
of
the
stack
but
yeah.
I
I
mean
yeah
like
what
we
talked
about
with
miles
last
time.
You
know
it'd
be
nice.
If
we
had
like
a
bunch
of
cookbooks
on
how
to
how
to
build.
You
know
a
set
of
standard
solutions,
but.
A
A
A
What
you
do
for
kind
of
a
remote
branch
office
is
a
little
different
from
what
you
do
for
a
full
on-prem
data
center
and
if
you
could
pigeonhole
them
into
those
it
would
help
and
then,
if
you've
got
the
luxury
of
having
something
that
maybe
spans
across
regions,
so
that
part
of
your
disaster
recovery
could
could
entail
standing
up
replacement
resources
in
another
re
location,
in
other
words,
you're
implementing
concepts
of
like
regions.
B
I
I
assume
that
that
tce
will
be
a
a
project
in
the
ecosystem.
That
vmware
is
gonna,
want
to
have
kind
of
gonna
push
well
push
gonna
show
off
at
kubecon
quite
heavily.
It's
a
commit,
such
community
oriented
open
part
of
the
tanza
ecosystem.
B
A
I
think
we
anticipate
doing
that,
but
under
the
current
the
cncf
rules,
it
isn't
yet
accepted
as
a
sandbox
project,
and
you
can't
do
that
when
you,
you
know
when
an
open
source
thing
first
goes
out
it
does.
It
isn't
automatically
eligible
you
have
to
have
outside
contributors.
It
has
to
have
a
certain
critical
mass.
A
They
do
an
analysis
of
what
they
call
the
health
of
the
open
source
project
that
is
based
on
a
bunch
of
criteria
that
you
know
they
don't
want
vendors,
just
putting
a
brand
label
on
something
saying
now
open
source
and
then
it
instantly
becomes
eligible
for
cncf
and
because
it
isn't
there.
Yet
there
are
some
rules
that
prevent
the
scope
of
how
much
presence
it
can
have
in
kind
of
the
official
sessions
and
things
I'm
anticipating.
A
I
don't
even
know
if
it's
been
decided
yet,
but
I'm
guessing
that
vmware
will
have
a
booth
and
at
a
minimum
it
would
be
there
but
yeah.
I
think
we
just
aspire
to
it,
getting
a
little
broader
exposure,
but
in
terms
of
having
actual
official
sessions
on
the
session
agenda,
because
it
isn't
right
now
under
the
cncf.
D
Kind
of
I
think,
no,
I
think
the
one
other
subject
that
could
be
interesting
to
deal
with
is
that
we've
talked
about
here.
Also
a
bit
is
the
challenges
or
the
things
to
note,
with
kubernetes
and
air
gap
scenarios,
and
that's
something
that
is
relatively
specific
to
vsphere,
although
it
is
a
bit
broader,
because
there
are,
for
example,
aws
environments
that
cut
off
access
to
the
internet,
but
much
more
in
the
vsphere
world.
D
A
You're
right,
that
is
a
that.
That
is
a
great
topic.
It's
almost
like
now,
I'm
sorry
that
the
general
cfp
window
is
closed
because
that's
you're
right.
It
is
broader
than
just
vsphere,
but
given
that
vsphere
is
so
popular
when
you
go
on-prem
that
that
is
a
common
scenario
of
just
network
considerations,
standing
up
a
container
image
repository
managing
to
come
up
with
a
scenario
to
still
get
patches
done
and
things.
D
Right,
how
do
you
still
utilize,
the
you
know,
the
velocity
of
the
open
source
community
of
helm,
charts
that
are
being
released
by
bitnami
or
anyone
else
on
a
weekly
basis?
A
new
version
is
out
there,
but
how
do
you
deal
with
the
cycles
of
how
do
you
bring
that
into
your
environment?
How
do
you
deal
with
all
of
this
when
you're
in
a
air
gap
scenario,
is
a
complex
challenge
that
there
are
tools
out
there
open
source
tools
that
out?
D
But
it's
there's
never
been
that
I
have
found
at
least
a
good
session
or
blog
post
or
anything
that
really
goes
into
depth
on
what
the
correct
methodology
is,
and
I
think
that
as
vmware
is
the
most
prevalent
for
these
types
of
cases
as
a
platform
to
run
on
could
be
something
that
would
be
interesting
to
have
in
the
vmware
user
group
session
that
you
may
also
be
able
to
attract
people
that
aren't
necessarily
just
running
on
vsphere
right.
A
You
know
it
isn't
just
air
gap
now
that
I
think
about
it.
There
have
been.
This
has
come
up
in
meetings
of
this
user
group
before
we've
got
a
user
with
a
big
retailer
who
has
the
scenario
where
they
do
have
network
connectivity,
but
it's
bad.
It
isn't
100
reliable,
it
isn't
high
bandwidth
and
they
have
scenarios
where
they
want
to
take
a
power.
Cycling
come
back
and
maybe
when
they
come
back,
the
network
is
down,
but
they
want
that
store
to
come
back
and
be
online,
which
you
know
sort
of
implies.
A
Things
like
any
containers
have
to
have
local
sources
rather
than
fresh
polls
off
the
internet.
I
also
have
heard
users
with
cruise
lines
where
they
have
a
scenario
where
these
cruise
ships
actually
have
a
lot
of
it
equipment
on
them.
A
They,
you
know
even
multiple
racks,
where
they're
running
accounting
and
billing
systems,
entertainment
systems
and
gambling
machine
operations,
and
they
get
connectivity
only
when
the
ship
comes
into
port
and
then
it
can
go
a
week
or
more
out
at
sea
with
no
connectivity
and
it's
similar
to
air
gapped,
but
maybe
not
a
hundred
percent
identical.
But
you
know
there
are
some
real
interesting
challenges
there
for
how
you'd
come
up
with
a
system
that
would
meet
your
requirements
and
so
yeah.
D
Yeah,
no
for
sure-
and
I
I
have
done
I
we're
currently
working
myself
and
a
few
other
people
and
opening
up
another
kind
of
like
user
group
for
specifically
for
on-premise
kubernetes,
for
like
air
gap,
solutions
and
I've
collected
a
bunch
of
you
know,
tools
that
help
with
this
and
a
bunch
of
just
like
data.
D
I
would
be
more
than
if
this
is
a
subject
that
is,
you
know
that,
whatever,
if
you,
if
it's
decided
that
this
is
a
subject,
I
can
definitely
share
some
thoughts,
and
you
know
the
you
know,
whatever
the
notes
that
I
have
in
the
resources
that
I
have
collected,
that
I
have
found
useful
in
this
type
of
use
case.
A
That'd
be
great,
but
yeah.
It's
interesting
that
kind
of
air
gapped
on-prem,
and
then
I
described
the
cruise
ship
as
mobile
and
right.
I've
heard
people
dreaming
of
things
like
kubernetes
in
an
automobile,
not
sure
I
buy
it
yet,
given
the
limited
resource
there,
but
you
know
who
knows:
compute
capacity
tends
to
grow
over
time
and
even
if
it
isn't
viable
on
a
cheap
consumer
automobile,
maybe
it's
viable
on
a
larger
truck
or
other
mobile
vehicle.
That
you
know
might
have
some
interesting
issues
like
that
too.
D
D
The
question
has
been
reliability
and
things
like
that-
and
you
know
there
are
some
complexities,
obviously
of
using
kubernetes
there,
but
I
don't
think
we're
that
far
away
with
these
lightweight
kubernetes
distributions,
like
k3s,
that
run
as
a
single
binary,
I'm
not
sure
we're
that
far
off
from
that
happening.
Actually
yeah.
A
A
So
anybody
else
got
anything
any
topics
you
want
to
chat
about,
throw
it
out
there.
B
Well,
I
was,
I
was
kind
of
kind
of
kind
of
thinking
about
what's
to
happen
this
year
when
it
comes
to
the
the
the
kubernetes
vmware
landscape,
because
I
think
it's
going
to
be
an
interesting
year
for,
for
a
whole
bunch
of
reasons
and
what
I
really
what
I
really
like
right
now
is
you're,
seeing
a
rapid
maturing
of
parts
of
the
tanzu
portfolio
and
and
the
politics
inside
the
map
view
which
is
gonna,
which
is
gonna,
mean
that
we're
gonna
have
there's
gonna,
be
a
more
solid
narrative
around
vmware
and
kubernetes,
and
everything
that
goes
on
top
of
it
this
year
and
I'm
seeing,
I
mean,
remember
I'm
from
a
partner
right,
so
I'm
always
thinking
about
kind
of
our
customers
and
what
we're
seeing
happening-
and
I
mean
there's
a
lot
of
there-
is
a
lot
of
kind
of
noise
right,
there's
a
lot
of
interest
in
in
in
all
of
the
cloud
native
topics.
B
No
one
really
knows
how
to
deal
with
it.
Up
till
you
know
this.
You
know
up
till
last
year,
I've
been
a
bit
hesitant
about
saying
you
know
like
let's
drive
into
the
the
newer
kubernetes
distributions
vmware
is
offering,
because
it's
still
very
much
in
being
built
up
now
this
year.
That
seems
to
be
changing,
we're
reaching
points
where
the
value
add
is
starting
to
outweigh
the
immaturity
and-
and
I
mean
personally,
I've
got
a
bunch
of
pscs
lined
up
with
customers,
which
I'm
really
looking
forward
to.
B
What
what's
what
needs
to
happen-
and
god
knows
we've
covered
this
before-
is
a
bit
of
education
and
a
bit
of
getting
getting
that
that
infra
up
going
when
it
comes
to
these
kinds
of
technologies.
B
This
is
what
I
really
loved
about
tce,
because
it's
because
it
it
it's
an
open
community.
Anyone
can
join
in
right.
Anyone
can
participate
in
that
part
of
the
vmware
ecosystem
around
kubernetes
and
that's
going
to
hopefully
help
incentivize
a
lot
of
interest,
and
it's
it's
a
great
way
of
when
I'm
talking
to
people
saying
look
just
look
at
that
for
now
right,
you
don't
need
to
invest
in
a
you
know,
a
bunch
of
bunch
of
hardware
or
set
it
aside
for
a
pse.
B
B
A
A
I'm
just
gonna
interrupt
you
a
bit
that
this
group,
because
we're
under
this
cncf
charter,
we're
specifically
advised
that
this
is
about
all
vendors
kubernetes,
including
people
who
don't
use
a
commercial
distro
or
who
use
rancher
red
hat.
So
a
lot
of
what
you're
talking
about
is
a
stack,
and
I
believe
you
could
use
carvel
carvel
is
that
is
under
cncf
to
run
on
other
forms
of
kubernetes
that
aren't
even
tonsu,
but
I
appreciate
the
words
that
you,
you
know.
A
A
lot
of
the
complexity
for
users
is
just
that
it
it's
almost
like
at
one
point.
You
know
linux
was
a
thing,
but
it
got
commoditized.
I
think
kubernetes
itself
is
more
or
less
getting
commoditized
or
that's
the
path
it's
on,
but
an
actual
user
who's
going
to
use
this
for
practical
things
needs
to
worry
about
the
applications
on
top
of
it
and
along
with
it,
and
maybe
at
some
point,
kubernetes
disappears
into
invisibility.
A
So
this
concept
that
you
have
a
complete
stack
of
things
that
work
together
is
is
a
key
thing
to
making
this
practical
to
deploy.
B
Yeah
and
and
all
so
many
parts
of
that
tanzania
stack
gonna
break
out
into
the
open
source
community
this
year,
they're
gonna
be
part
of
that.
Tce
stack,
so
it's
it's!
You
know
that
it
just
creates
a
huge
amount
of
opportunity
and
a
way
in.
That's
really,
you
know
the
barrier
to
entry
is
so
much
lower
if
it's
not
immediately
a
commercial
offering
right
so
so
that
you
know
it
makes
it
makes
my
job
easier,
but
it
also.
B
It
also
means
that
user
groups
like
this
all
the
you
know
the
communities
evolving
around
these
open
source
projects
around
where
this
tanza
ecosystem.
You
know
that
they're
all
like
the
perfect
way
to
get
into
this
stuff.
A
Have
the
experience
you
know
my
trouble?
Is
I've
worked
on
kubernetes
so
long,
I'm
like
an
insider
and
almost
know
too
much
and
at
some
point,
you'd
think
that
this
could
be
good
enough.
That
someone
with
zero
background
in
kubernetes
but
is
familiar
with
an
app
development
platform,
should
be
able
to
get
by
with
minimal
understanding
and
investment
and
just
get
things
to
work.
How
close
do
you
think
we.
D
Oh,
I
think
we're
very
close
with
the
you
know
exactly
robert,
why
I'm
saying
that
we're
very
close,
but
I
think
that
in
the
end
with
you
know,
the
currently
beta
when
it
becomes
ga,
becomes
ga
idea
of
tanza
application
platform
and
whatnot,
which
again
that's
a
commercial
offering,
but
I
think
the
general
idea
there
was
a
great
article
that
I
know
robert's
been
sharing
around
also
and
written
by
a
close
friend
actually
from
vmware
here
in
israel,
oded
chopin,
who
used
to
be
a
cloud
foundry
guy.
D
So
the
idea
of
a
pass
right
when
we
looked
at
the
beginning
of
containers
and
all
of
that
it
was
very
difficult.
The
idea
of
writing
applications
at
that
point,
and
then
you
have
the
pcfs
or
the
heroku's
or
whatever
came
out
with
the
idea
of
the
paz
system
and
that's
coming
to
kubernetes
as
well,
whether
that
be
vmware's
offering
or
other
offerings
that
are
out
there,
things
like
cartographer,
which
is
an
open
source
project.
All
of
these
tools.
I
think
that
what
people
are.
D
Finally,
understanding
is
what
kelsey
hightower
posted,
I
think
in
2017,
if
I'm
not
mistaken,
which
is
that
kubernetes
is
not
the
end
game.
It's
a
platform
for
platforms,
it's
just
a
good
place
to
start
from,
and
I
think
that
people
are
starting
to
realize
that
and
that
platforms
are
being
built
above
it
which
allows
what
you're
talking
about
steve
right
like
I'm.
If
we
look
at
like
even
k,
native,
which
ryan
a
huge
open
source
project
on
top
of
kubernetes,
still
required.
D
Writing
a
bunch
of
yaml,
creating
docker
files
doing
all
of
that
they
have
a
sandbox
project
called
funk.
That
allows
you
to
basically
just
give
function
code
and
it
builds
out
a
k,
native
service
and
everything
for
you.
You
don't
need
to
know
about
kubernetes,
so
that's
like
an
open
source
system
for
certain
types
of
applications,
and
then
you
have
argo.
D
Cd
and
you've
got
all
these
different
tools
that
are
kind
of
starting
to
come
together
in
the
community
that
are
really
building
these
platforms
on
top
of
kubernetes
that
are
abstracting
it
all
away,
and
I
think
that
the
key
with
kubernetes
that
I
think
people
are
starting
to
realize
in
the
community
is
that
if
your
developers
need
to
have
raw
access
to
cube
ctl
and
use
it
on
a
daily
basis,
you're
doing
something
wrong.
Your
developers
should
not
need
to
be
with
cube
ctl
on
a
daily
basis.
D
They
should
have
some
level
of
abstraction,
whether
that
be
home
grown
or
a
commercial
offering
or
an
open
source
tool
that
deals
with
the
complexities
for
them,
because
there's
too
many
knobs
in
kubernetes
that
to
teach
a
developer
who
needs
to
know
multiple
programming
languages,
also
yaml
infrastructure,
storage,
networking
everything
we've
already
realized
that
full
stack
development
doesn't
really
work
that
well
and
that
front-end
developers
are
better
at
front
end
than
full
stack
are
and
back
end
developers
are
better
at
back
end
than
full
stack
developers.
We
tried
that
it
moved
back.
D
I
think
the
same
thing
is
going
to
go
with
kubernetes.
It's
got
such
a
wide
spread
that
we
need
these
other
platforms,
and
I
think
that
we
are
getting
there
and
this
year
is
going
to
be
big
with
that.
A
Yeah
that
makes
sense
to
me.
It
almost
seems
like
it
has
a
feel
like
to
use
a
historical
precedent.
I.T
evolved
in,
I
don't
know
the
90s
through
the
early
2000s
into
the
scenario
where
enterprises
had
dbas
and
database
specialists.
They
had
network
specialists,
they
had
storage
specialists
and
they
became
people
who
had
that
deep
knowledge
that
got
called
in
when
you
had
to
go,
make
a
major
architectural
purchase
decision
and
when
you
had
to
troubleshoot
when
things
went
horribly
wrong.
A
But
for
the
most
part
individual
developers
were
not
called
upon
to
do
that,
or
maybe
they
weren't,
even
given
the
admin
credentials
to
touch
the
things
related
to
the
databases,
the
network,
the
storage
and
maybe
eventually
that's
the
way.
Kubernetes
goes
where
enterprises
have
a
handful
of
kubernetes
specialists,
but
it
isn't
expected
to
be
a
an
across-the-board
skill
that
everybody
in
the
it
infrastructure
would
have.
D
Right
you'll
have
a
platform
team
that
is
in
charge
of
the
platform,
be
that
kubernetes
is
the
infrastructure
and
then,
above
that,
whatever
platform
tooling
you're,
using
above
that
to
expose
it
to
your
and
it'll,
be
just
like
databases
as
a
commodity
or
networking
is
a
commodity
that
the
developer
can
say
yeah.
I
need
this
port
open
to
here.
He
doesn't
need
to
know
what
that
means
in
terms
of
what
seven
firewalls
along
the
way
that
needs
to
be
opened
in
and
what
vlans
need
to
be
created,
or
whatever
needs
to
be
done.
D
B
So
that's
this
is
where
I
see
the
challenge.
I
think
we
saw
a
bit
of
a
way
to
go
before
we
can
get
there
there's
I
I
I
still
see
too
much
fragmentation
in
in
in
standards
and
skill
sets
and
tools
around
getting
to
that
point.
Well,.
B
Yeah,
so,
okay,
so
this
is
just
kind
of
literally
what
my
job
is
with.
Customers
is
kind
of
developing
those
kind
of
policies
like
that
and
ideas
like
that,
the
the
but,
but
just
to
kind
of
just
take
a
step
back
to
the
technology.
I
mean
if
we
look
at
if
we
look
at
automation
to
get
stuff
done,
so
I'm
not
talking
about
the
developers
yeah,
I
mean
I.
I
totally
agree
with
scott
that
you
know
you
want
developers
to
just
be
able
to
focus
on
code.
You
don't
want
them.
B
You
want
to
give
them
that
high
level
abstraction,
so
they
don't
need
to
deal
with
the
infrastructure.
But
the
moment
you
get
to
the
what
we're
calling
the
platform
team
things
get.
Things
are
really
messy
out
there
right
now
and-
and
this
worries
me
and
you
know,
I
think
we
have
a
lot
of
work
there.
If
you
think
of
processes
around
automating
say
firewalls
yeah
I
mean
some
companies
are
trying
to
adopt
frameworks
around
automating
this
stuff,
sometimes
with
terraform
it's
from
the
vmware
ecosystems,
often
with
vrealize
automation.
B
They're
we
still,
you
know
it
comes
back
to
the
skill
thing
again.
We
need
we
need
a
lot
of
people
in
for
up
where
they'll
adopt
using
tools.
You
know
high
level
infrastructure,
automation
tools
like
terraform
or
ansible
things
like
that
to
be
able
to
enable
that
stuff.
You
you
can't.
You
can't
successfully
implement
higher
level
platforms
if
you
can't
get
rid
of
the
problem
of
programmatically,
putting
in
a
firewall
rule
somewhere
and-
and
this
may
just
be
the
lens
through
our
custom.
B
You
know
to
the
customers
I
see
in
the
netherlands,
but
many
of
these
companies
aren't
there
yet
they're
still
doing
firewall
rules
manually,
they're
still
setting
you
know,
building
vms
manually
yeah,
I
mean
it's
a
long
tail
right,
so
you're
going
to
have.
You
can
have
companies
that
at
some
point
manage
to
get
to
a
certain
level.
B
I
mean
I
within
within
the
kubernetes
hype
cycle.
We
like
talking
about
the
tip
of
the
spear,
but
it's
not
what
I
see
when
I
look
at
the
market.
B
So
it's
you
know,
the
long
tail
is
getting
broader,
it's
getting
higher
and
as
many
people,
many
companies,
especially
newer
companies,
are
kind
of
being
built
up
with
some
of
this
cloud
native
thinking
already
embedded
in
their
people.
A
lot
of
people
are
still
being
left
behind
and
it's
it's.
You
know
it's
it's
a
it's
still
going
to
be
a
huge
challenge
to
get
to
get
to
the
level
where,
if
you
you
have
to
solve
some
of
those
underlying
problems.
First,
before
you
can
get
to
the
point.
B
Okay,
now
we
have
the
the
the
foundation
to
say:
let's
do
a
higher
level.
You
know
developer
platform
say
like
like
cloud
foundry
or
tap,
or
something
like
that.
The
the
ground
work
is
not
in
place,
culturally
and
and
and
from
a
skeletal
perspective
in
many
companies.
Unfortunately,
so
this
and
the
the
the
the
sound
moves
under
your
feet
all
the
time
when
it
comes
to
the
open
source
community.
B
So
it's
it's
so
like
for
the
average
you
know
kind
of
vsphere
admin
or
person
that
comes
from
infra.
A
lot
of
them
aren't
even
bothering
to
try
and
keep
up
because
it's
moving
too
fast.
So
it's
it's
very
hard.
One
of
the
things
I
try
to
to
recommend
to
kind
of
middle
management
is
makes
you
know
you
guys
have
to
strategically
invest
in
skilling
people
up
and
and
giving
the
the
room
that
people
can
actually
start
investigating
some
of
these
newer
technologies
right.
D
No,
I
mean
I'm
seeing
things
a
bit
differently
here,
but
then
again,
the
israeli
market
is
known
as
very
ahead
of
like
adopting
technology
very
fast
and
very
early
for
good
and
for
bad.
I
will
say
for
bad
for
a
lot
of
things
as
well,
because
they
adopt
technologies
before
they're
richer
enough,
but
that's
just
like.
I
think
it
is
definitely
market
specific,
but
that's
also
where,
when
I
have
seen
customers
that
are
like
that,
that's
where
you
don't
go
in
with
kubernetes,
I
think
at
day
one.
D
I
think
that
kubernetes
or
anything
in
the
modern
application
world
needs
to
come
along
with
cultural
change
and
the
cultural
change
needs
to
come
first.
Otherwise
it
will
be
a
failed
implementation
of
modern
applications,
and
that's
where
I
think
things
like
a
vcf
stack
or
things
like
that,
which
again
just
taking
the
vmware
infrastructure
being
programmatic
having
that
capability.
D
You
then
need
to
build
it
from
the
ground
up.
You
can't
just
unfortunately,
in
today's
world
there's
everyone
goes
according
to
buzzwords,
so
I
mean,
like
you
know,
there's
a
very
famous
meme
of
someone
that
went
to
vmworld
a
cto
who
goes
to
vmworld
and
he
comes
in
saying
cooper.
What
and
he
walks
out
saying
great.
I
need
a
multi-cluster
service
mesh.
You
know
they
start
with.
D
What's
a
container
and
get
all
I
know
everything
needs
to
be
done
in
steps,
and
I
think
that
one
of
the
issues
is
that
people
try
to
rush
because
of
the
buzzword
into
we
need
kubernetes
everywhere,
without
even
understanding
that
they
need
to
break
down
monolithic
applications
into
microservices
the
amount
of
times
I've
seen,
containers
that
are
in
macro
services.
It's
a
monolith,
that's
just
tied
up
into
a
container
and
does
not
know
how
to
scale
and
does
not
know,
does
not
follow.
D
If
I
was
zero
of
the
twelve
factors
of
an
application
that
there's
supposed
to
be,
is
you
know
a
sad
fact
of
the
of
life
and
therefore,
I
think
in
those
cases
kubernetes
is
not
the
right
solution
at
this
point.
It's
a
goal
to
get
to,
but
implementing
kubernetes
at
such
a
place
is
not
the
right
step.
It
needs
to
be
cultural
within
the
legacy
applications
and
then
moving
towards
kubernetes.
B
Yeah
what
I
do
like
about
the
about
the
hype
cycle
is
that
it
does
it
pushes
on
all
levels
right,
so
so
vendors
are
starting
to
release
their
software
in
container
form,
and
you
know
frameworks
that
developer
use
developers
use
you
know.
The
default
at
some
point
becomes
a
helm
chart
instead
of
a
bunch
of
vms
and
and
that's
when
you
start
to
notice
the
pressure
coming
from
the
other
side
yeah.
That's
when
then
you
know
your
kubernetes
platform
that
you
know
that
is
being
minimally
used.
B
Suddenly,
developer
groups
start
approaching,
you
saying:
look,
we
have
this
thing
and
they
gave
us
a
bunch
of
containers
and
we
don't
know
what
to
do
with
this.
Please
help
us.
I
think
I
think,
for
the
next
few
years,
that's
going
to
be
the
level
right,
that's
going
to
be
the
level
that
a
huge
amount
of
people
are
going
to
be
introduced
to
kubernetes.
First
is
at
that
level.
So
we're
not
talking
application
platforms
right,
we're
not
talking
about
high
level
of
automation,
we're
talking
about.
B
I
I
built
this
in
docker,
it's
kind
of
12
factor,
but
not
really.
Please
help
me
run
it
in
kubernetes,
because
I
hear
that's
better
than
docker.
That's
the
level
we're
at
right
now.
I
think,
and
we're
going
to
stay
there
for
a
while
and
then
hopefully,
because
of
all
these
pressures
coming
from
all
these
different
directions,
the
the
the
the
floor
keeps
rising
slowly
right
and
I
I
I'm
kind
of
I'm
the
challenge
right
now.
Is
this
a
lot
of
enterprises
wanna?
They?
B
They
want
a
kubernetes
platform
because
they
see
this
coming
right.
They
they
kind
of
see
like
oh,
you
know.
We
have
a
bunch
of
developer
groups,
they're,
really
enthusiastic
about
this
new
stuff.
We
need
some
kind
of
kubernetes
platform
to
host
these
people,
they're,
not
thinking
about
application
club
thinking
about
you
know,
we
need
a
place
to
run
containers
and
we
don't
want
to
have
a
bunch
of
separate
docker
hosts.
A
I
think
you're
on
to
something
there,
robert
just
because,
having
worked
in
the
past
for
isps,
I
think
it's
true
that
pretty
much
all
of
those
who
are
selling
a
software
product,
even
in
the
old
days
of
vms
they
didn't
want
to.
They
didn't
want
to
deal
with
certifying
it
on
a
zillion,
different
vendors
bare
metal
servers,
so
they
they
ended
up,
insisting
that
it
go
into
these.
B
B
Yeah
and
you're
gonna
I
mean
you're
gonna,
see
the
same
with
low
code
frameworks,
I
mean
I
have
a
bunch
of
customers
that
are
kind
of
in
the
geo
space.
So
you
see
a
lot
of
geo
geo
frameworks
like
air
and
stuff
like
that.
There
are,
I
should
say
I
should
say,
and
the
these
are
very
kind
of
open
source,
driven
low
code
frameworks
and
you're.
B
So
so
so
you
you
do
see
that
pressure
starting
to
come,
but
it's
all
extremely
low
level
right,
it's
not
very
complex
and
and
it's
it's,
the
its
developers
and
analysts
leading
that
push
and
not
you
know,
there's
no
policy
around
it
yet
so
so
it
so
it
is.
I
mean
I
can
see
it's
coming,
it's
just
not
coming
as
fast
as
you.
B
You
know
you
you
might
be
you
might
like,
and
I
I
feel
that
especially
the
vendors
making
kubernetes
distributions
are
they're
over
here
right
there
on
the
other
side
of
that
divide
and
there's
this
huge
gap
in
between
but
yeah
the
another
point
I
can't
remember,
but
it's
my
point,
it's
just
it's
still
it's
going
to
take
longer
than
than
you.
You
think.
If
you
just
attended
kubecon.
A
I
don't
imagine
the
laggards
are
on
to
that
yet.
But
you
know
the
web
assembly
concept
is
that
it's
sort
of
like
a
docker
container,
but
the
code
is
portable
and
runs
everywhere,
and
the
founder
of
docker
is
quoted
as
saying
gee.
If
web
assembly
had
existed
back
when
we
designed
docker,
we
wouldn't
have
needed
to
do
docker
so.
D
D
Technology
hurdles
like
this,
where
there's
the
lack
of
knowledge
or
whatever
is
to
actually
never
try
and
do
it
in
a
brown
field,
always
do
a
green
field
for
kubernetes,
let's
say
or
greenfield
for
the
new
technology,
because
you
get
rid
of
most
of
the
complexities
you
put
it
in
a
new
cluster.
You
put
it
in
a
new
place.
New
networking
stack
whatever.
It
is
in
many
cases,
that's
possible
because
you're
starting
small
and
then
you
say
great,
you
want
to
move
to
this
new
environment
you're
using
new
methodologies.
D
So
you
don't
affect
the
current
platform.
You
don't
try
and
put
in
modern
application
methodologies
into
a
legacy
workload
environment.
You
create
the
new
environment
and
say
here
we
work
programmatically.
Here
we
work
declaratively
and
then
anyone
that
wants
to
use
that
platform
or
manage
that
platform.
You
have
a
small
platform
team,
that's
a
break
out
from
the
current
platform
team
for
the
legacy
applications
that
are
the
people
that
are
that
want
to
learn
and
that
are
eager
to
learn
these
new
technologies
and
they
manage
this
new
environment
that
is
small
at
the
beginning.
D
So
they
can
start
with
a
small
team
and
developers
that
want
to
adopt
the
new
technologies
and
gain
the
brand
new
hardware
that
we
just
acquired
with
awesome,
new
capabilities,
great
use,
our
new
methodologies
or
you
can
continue
using
legacy,
use
the
five-year-old
servers
that
are
here
that
are
running
on
sd
cards.
That
vsphere
is
going
to
break
on
at
some
point
because
it's
going
to
run
out
of
rights-
and
you
know
be
happy
there,
but
I
think
that
separation
of
give
the
choice
you
want
to
work
legacy
work
legacy.
D
D
Happy
new
year,
everyone,
and
by
the
way
in
february,
I
will
be
doing
a
talk
on
air
gapped,
kubernetes
stuff,
with
carvel
for
anyone.
That's
interested
it'll,
be
after
the
next
meeting
and
I'll
have
a
link
by
then
hopefully,
because
that's
relevant
to
this
group,
but
I'll
be
doing
that
on
a
on
tanzu
tuesday
on
some
air-gapped
kubernetes
stuff,
that's
not
relevant
to
only
vmware
distribution.
So
anything
that's
running.
A
B
And
I
too
will
be
on
a
tanzania
tuesdays
with
the
vm
vanguard
group
and
we're
going
to
talk.
What
are
we
talking
about
again
we're
going
to
talk
about
something.