►
From YouTube: Kubernetes UG VMware 20230302
Description
March 2, 2023 meeting of the Kubernetes VMware User Group. Discussed migrating from legacy in tree storage and cloud provider to current out of tree storage provider (CSI). Observation was that resources are out there but not easy to find and a number of these were offered.
A
Hi
welcome
to
the
March
2nd
meeting
of
the
kubernetes
VMware
User
Group,
where
this
group
is
concerned
with
running
all
forms
of
kubernetes,
whether
a
VMware
distro
or
some
other
distro
on
the
vsphere
infrastructure.
We
don't
have
any
action
items
any
items
on
the
agenda
for
today,
as
at
least
as
the
last
time,
I
looked.
Let
me
I'm
just
looking
in
yeah,
nobody
adding
anything,
but
when
that
happens,
we
just
engage
in
freeform
birds
of
a
feather
discussion
on
anything.
Anybody
wants
to
talk
about
before
the
meeting
officially
started.
A
Bryson
and
I
started
chatting
about
the
subject
of
what
documentation
might
be
available
for
migrations
of
Legacy
installations,
to
kind
of
the
modern
thing
and
what
comes
into
play.
There
is
both
the
CSI
storage
plug-in
as
well
as
the
cloud
provider.
That's
the
shim
layer
that
adapts
kubernetes
to
run
on
vsphere
they're
kind
of
orthogonal,
where
you
could
potentially
use
one
or
both
so
Bryson's
point
was
and
I'll.
Let
you
step
in,
but
I'll
try
to
summarize
that
he
wasn't
able
to
easily
find
docs
on
how
you'd
go
about
doing
a
migration.
A
A
It's
not
coming
across
blog
posts,
but
these
are
now
at
this
point,
one
or
two
years
old,
because
CSI
kind
of
moved
independently
of
the
cloud
provider.
So
there
was
an
initiative
that
the
container
storage
interface
that
started
Jay
at
this
point
even
five
years
ago,
was
when
the
planning
started
to
get
the
storage
providers
out
of
tree.
A
Primary
reason
was
that
if
you
even
had
a
security
patch
related
to
a
storage
driver,
you
had
to
wait
for
a
whole
kubernetes
release
to
get
out
the
door
in
order
to
get
a
vehicle
to
put
it
out
there,
and
it
meant
that
you
couldn't
be
very
responsive
and
those
got
split
off,
but
operated
for
a
good
couple
of
years
where
you
could
have
used
the
entry
or
the
out
of
tree
and
on
vsphere.
If
you're
stuck
on
really
old
Hardware.
A
You
know,
and
by
really
old
I
mean
something
that
couldn't
run
these
sphere
6.7
you,
you
too
I
think
even
to
this
day,
you're
stuck
it's
no
longer
supported,
but
there
are
people
who
are
in
that
situation.
I
realize
that,
but
you
couldn't
move
to
the
newer
stuff
and
you
still
can't
so
anyway.
Bryce
said:
do
you
want
to
jump
in
here
and.
B
B
C
B
A
That
was
very
tonzu-specific,
but
then,
on
the
other
hand,
from
the
perspective
of
us
being
a
vendor.
The
people
who
are
on
tanzu
kind
of
expect
easy
to
consume
documentation
where
it
gets
them
down.
A
path
that
is
very
you
know,
is
specific
to
the
distro
they
are
using
and,
in
some
cases,
tonzu
automates
away
things
that
you
might
be
concerned
with.
A
You
know
if
that
didn't
happen,
and
it's
kind
of
a
one
of
these
scenarios,
where
how
you
do
the
docs
there's
no
great
answer
that
is
going
to
satisfy
everybody
and
it's
kind
of
flip-flopped
over
the
years
to
where
we've
done
it.
A
couple
you
know
we,
we
started
out
doing
it
with
our
distros
flipped
it
over
to
Pure
open
source,
then
had
complaints
from
people
who
are
on
the
commercial
products
that
they
couldn't
find
stuff
and
had
to
go
to
a
second
place.
So
it
it
I.
A
That
it's
not
super
straightforward.
Probably
if
I
were
on
tanzu
I'd
go
there
first,
but
if
you're
not,
then
it
is
the
GitHub
and
secondary
sources
blogs
by
Cormac
Hogan
miles
gray
are
great.
You
know,
who
else
does
a
good
job?
In
my
opinion,
documenting
this
is
act.
You
know
my
advantage
might
might
shoot
me
for
bringing
this
up,
but
red
hat
actually
has
a
little
section
on
their
website.
It's
intended
for
openshift
users,
but
much
of
the
stuff
that
you
would
find
there
could
applies
more
or
less
generically
and.
B
C
A
Yeah
I
mean
when
you
get
to
it.
This
is
a
big
picture
issue
where
it's
not
just
the
docs,
but
even
support.
Where
we've
got
people
asking
questions
both
in
this
user
group
slack
Channel,
then
there
is
a
cloud
provider,
slack
Channel
where
they
come
in
as
well
a
lot
of
times,
people
duplicate,
post
them
and
then
they
come
in
through
the
standard.
Vmware
support
Channel
as
well,
and
if
you
are
a
vsphere
user
on
support,
you're
free
to
put
them
in
there.
C
A
One
because
some
of
these
things,
like
the
kubernetes
slack
at
least
in
theory,
give
users
an
opportunity
to
share
with
each
other,
rather
than
being
restricted
to
go
through
the
vendor.
Channel.
Only-
and
you
know
it's
up
to
users
and
customers
to
what
extent
they
want
this
public
that
as
a
vendor,
you
know
our
standard
commercial
support
Channel
by
Design,
because
some
people
are
very
concerned
with
privacy.
They
might
not
even
want
competitors
to
know
what
their
vendor
choices
were,
so
that
we
don't
open
that
up
for
sharing
but
kind
of
by
Def
yeah.
A
C
A
Versa,
VMware,
with
our
multiple
kubernetes
distros,
you
know
having
the
old
PKS,
which
is
still
out
there
and
supported
ends
up
with
forks
in
documentation
and
support
and
things
where
some
solution
that
applies,
if
you're
on
one
form
of
kubernetes,
maybe
isn't
applicable.
If
you
run
a
different
one,.
B
A
A
C
A
They,
the
cloud
provider,
migration,
the
peak
of
activity
in
getting
a
word
on
that
happened
about
a
year
ago,
yeah
and
I
still
remember
there
was
there
was
a
whole
talk
on
the
migration
from
out
of
tree
entry,
Cloud
providered
out
of
tree.
That
happened
at
kubecon
Europe
in
May
of
last
year,
and
they
gave
a
forum
kind
of
for
every
cloud
provider.
You
know
the
public
clouds
have
these
same
issues
too.
You
know
whether
you
were
on
if
you
were
on
Amazon
Google
cloud
or
on-prem
with
vsphere
every
one
of
these.
A
Every
one
of
these
platforms
had
that
scenario
of
going
from
the
entry
to
the
out
of
tree
and
for
right
or
wrong.
There
was
a
feeling
that
kind
of
about
a
year
ago
was
sort
of
that
Peak
transition
where
most
of
the
people
had
had
initiated
the
move,
and
there
were
there
were
small
numbers
of
people
left
behind
now,
if
you're
an
organization
that
got
on
kubernetes
early,
you're
kind
of
in
a
different
category
where
you're
you
know
the
earlier,
you
got
on
top
of
kubernetes.
B
A
B
A
A
B
A
B
A
A
C
B
A
B
A
A
That's
the
Upstream
kubernetes
those
vendors
of
distros
that
the
kubernetes
project
says
that
if
a
vendor
wants
to
take
responsibility
for
this,
they
can
continue
to
support
these
old
versions.
But
the
project
itself
doesn't
have
the
resource
to
be
doing
that,
and
you
know
it
presents
a
dilemma
for
somebody
trying
to
cling
to
these
old
things
and
at
some
point,
you're
you're
going
to
be
sort
of
forced
to
move
I
mean
the
story
of
of
Enterprise.
I.T
really
is
that
same
pattern
has
repeated
that
you
can't
I.
C
C
B
B
A
A
Promise
it'll
be
the
next
one,
because
the
next
one
is
coming
pretty
close
to
when
travel
for
kubecon
Europe
starts
and
I
was
planning
on
attending
there.
So
I
that
might
be
pretty
disruptive
on
my
ability
to
get
that
put
together
by
the
the
first
Thursday
in
April.
Maybe
it
will
be
May
instead,
but
we'll
we'll
try
to
get
that
queued
up.
C
C
To
look
at
this
Bryson
and
Steve,
but
you
know
one's
a
Rancher
issue,
but
they're
talking
about
the
migration
from
the
CPI
so
like,
and
then
someone
gives
a
link
to
like
a
Helm
chart
for
the
old
versions
and
you
know
a
bunch
of
stuff.
But
then
someone
also
gives
a
link
for
like
a
migration
which
looks
like
it's
like
buried
in
the
GitHub
Docs.
B
Yeah,
so
that's
I
mean
that's
one
of
the
things
besides,
just
like.
Maybe
a
presentation
would
be
like
I,
really
just
making
sure
if
we
have
that
stuff
that
it's
it's
like
at
the
bottom
of
the
readme.
On
the
the
first
page,
like
hey,
here's,
how
you
do
the
new
one,
because
it
does
say
here's
how
you
do
the
old
one.
But
then
it
doesn't
say:
here's
how
you
migrate
from
the
old
one
to
the
new
one
right.
C
A
Yeah
and
I'll
have
to
see
I
think
that
we've
maybe
not
put
it
in
the
official
docs,
but
there
have
been
enough
people
who've
been
through
these,
these
migrations
that
they
may
be
in
blogs
or
in
other
forms,
even
videos
for
all
I
know,
but
they're
they
are
out
there.
This
is
you're
far
from
being
the
first
one
to
ever.
Do
this
you're
you're,
maybe
in
the
second
half
of
the
migration
here,
but
I-
understand
that
that
happens
by.
B
A
Yeah
I
have
a
feeling
I
I,
recollect
coming
across
Maybe
10
different
sources
but
like
I,
say
I'm
gonna
have
to
I'm
going
to
have
to
go,
look
around
and
find
all
my
bookmarks
and
things
after
the
meeting,
but
I
think
there
are
quite
a
few
of
them
out
there
and
from
many
categories.
Blog
stocks,
yeah
actual
actual
meeting
presentations
online
in
that
physical
conferences,
so
I
I,
I'll
I'll
try
to
do
a
research
assembly
of
those
sorts
of
things
and
get
back
to
you
on
that.
A
C
A
Yeah
there's
kind
of
things
that
can
happen
even
if
you're
running,
like
persistent
applications,
where
this
is
unrelated
to
cloud
provider
or
really
like
a
version
of
CSI,
but
some
people
occasionally
even
have
a
diff
move
from
one
backing
store
to
another,
and
you
know
that
sort
of
migration
isn't
something
that
just
magically
happens
by
itself
and
I.
Don't
know,
I,
guess
that!
That's
why?
C
Yeah,
interestingly
enough,
to
like
the
stuff
you
talk
about
like
the
back
end
storage
migration,
like
that,
there's
like
a
lot
of
issues
that
can
happen
with
be
serious.
Csi
like
if
you
move
the
first
class
discs
and
stuff
so
like
I,
could
see.
I
like
that,
could
be
pretty
large
and
complex
problem
with
like
an
infrastructure
like
vsphere
and
vcenter,
managing
all
the
this
creation
and
then
like
yeah.
B
A
C
Yeah,
we
definitely
have
like
like
something
like
what
Bryson
was
saying
like
where
someone
will
be
using
tanzu,
but
they
haven't
looked
through
like
the
CSI
docs,
which
are
now
on
VMware
docs,
but
they
used
to
be
only
on
GitHub
and
like
there's
like
feature
components
that
you
wouldn't
know
you
didn't
have
available.
Until
you
know
you
did
something
like
storage,
CRS
or
something
like
that
or
like
historically
motion,
where
it's
like
as
a
v
stream
admin.
C
B
C
Using
first
class
disk
so
like
if
the
v-series
PS5
driver
doesn't
support
storage
view
motion
of
first
class
disks
like
on
the
back
end
so
like.
If
you
were
to
move
it,
there's
a
component
in
vcenter
that
monitors
as
like
a
database
to
monitors
where
the
first
class
disks
are
located
and
then
CSI
interacts
with
that.
So,
if
you
move
that
out
from
under
where
that
CNS
component
monitors
like
you,
can
have
CSI
not
be
able
to
locate
disks.
C
B
C
Usually
like
if
storage
GRS
is
on
or
if
the
storage
has
been
moved
by
like
either
a
user
or
backup
software
that
for
the
first
class,
just
it's
not
something
that
they
manage,
or
they
have
a
feature
that
does
that
yet
I
know
that
they're
working
on
it
and
I
know
that
they
have
some
tools.
That
will
be
upcoming
in
like
newer
versions
of
vsphere
and
also
newer
versions
of
the
CSI,
but
like
currently
like
with
some
of
the
Tennessee
Products,
like
that.
Those
are
the
big
hitters
and
then
multi-attached
issues.
C
But
those
are
more
Upstream
kubernetes.
But
those
seem
to
be
fixed,
hopefully
in
the
next
couple
releases
and
then
redirect
many
volumes.
I
guess,
because
the
only
way
you
can
have
them
with
VCR
CSI
is
with
vsan
file
services
so
like.
If
you
wanted
to
read,
write
menu
volume
and
you
tried
to
provision
it
with
the
vsphere
CSI
driver
you
it
would
just
fail,
and
then
some
people
just
don't
know
that
the
requirement
is
to
have
like
an
NFS
type
storage.
C
Yeah
I
mean
the
ones
I
was
discussing
weren't,
specifically
the
migration,
just
more
like
functionality
of
the
driver,
but
I
know
that
a
lot
of
the
migration
issues
are
oh,
like
open
issues
in
in
GitHub,
like
I've
seen
them
before,
like
even
when
tkgi
first
did
the
migration
firm
entry
to
out
of
tree
like
they
had
a
lot
of
issues
and
a
lot
of
them
were
ending
up
in,
like
a
usually
when
I
get
someone
who's
doing
that
like
if
they're
on
a
legacy
version.
Most
of
them
I'll
find
the
answer
for
in
GitHub.
C
A
I'm
just
kind
of
curious
Bryson
is
this
migration
kind
of
the
biggest
problem
or
opportunity
you're
having
right
now
for
getting
stuff
done,
I'm
just
like
to
have
the
opportunity
to
discuss
with
an
actual
user.
What
the
you
know,
what
the
pain
points
are,
or
for
or
the
opportunities
to
improve
the
kubernetes
itself
as
well
as
no.
B
B
B
C
C
B
C
B
Yeah,
that's
one
of
the
main
things
why
we
started
this
group
was
to
not
be
so
just
stuck
in
the
VMware
side
of
things,
because
there's
so
many
different,
you
know
I,
guess
flavors
of
how
to
install
it
like
stay
more
to
that
base
of
like,
what's
in
common
to
everyone
like
all
the
open
source
stuff.
Oh.
A
C
C
A
C
A
Just
on
top
of
vsphere,
but
you
know,
cluster
API
is
just
sort
of
a
generic
extraction
layer
for
deploying
kubernetes
clusters
to
platform,
and
now
there
there
aren't
kind
of
implementations
of
it
for
every
public
Cloud
out
there,
but
certainly
for
AWS.
There
is
I
think
for
Google
I
think
there
is
for
Azure
as
well,
but
maybe
Kyle.
You
know
more
than
I
do
yeah.
C
A
And
it
isn't
just
VMware
using
cluster
API,
so
there
are
other
people
using
it
as
well.
Like
AWS
I
know,
does
I
personally
just
got
a
curiosity.
Did
an
eks
anywhere
install
in
my
vsphere
home
lab
I
was
actually
surprised
how
similar
it
is
to
what
tonzu
does
I
mean.
If,
if
you
have
used,
one
you'd
swear
that
you
know
the
same
person
did
them
both
almost
from
the
experience
I
had,
but
in
terms
of
cluster
API
I.
A
Think
it's
an
actual
cncf
monitored
project,
so
everything's
out
there
open
source,
there's
a
slack
channel
for
it.
It's
the
same
sort
of
thing
with
any
open
source
where
you
know
it's
Community
Support,
potentially,
unless
you
consume
it
through
some
kind
of
a
vendor
supported
commercial
product,
but
cluster
API
is
out
there
for
standing
up
your
kubernetes
clusters
and
I.
B
C
Yeah
I
mean
does
even
like
get
like.
If
you
look
I
guess
you
wouldn't
have
very
like
unless
you
like,
winch
their
profile
and
they
had
their
company
name,
but
like
a
bug
that
I
would
open
for
feaster
like
tanzu,
but
have
a
pro
like
I
would
be
able
to
tell
the
engineers
for
tanzu
that
it's
a
problem,
they'll
identify
as
like
a
cluster
API
issue
and
they'll
go
upstream
and
fix
in
cluster
API.
C
So
you
almost
get
like
the
benefit
of
that,
if
you're
using
it
as
open
source
because,
like
there's
full-time
employees
working
on
some
of
those
components
to
make
sure
they
work
Downstream
for
any
tanzu
product.
And
that's
you
know,
tkgm
uses
AWS,
Azure
and
vsphere
so,
like
all
of
those
different
controllers
or
Cappy,
workflows
are
like
updated
not
only
by
just
VMware
employees
but
like
Azure.
They
have
their
their
stuff
and
then
AWS.
They
have
their
stuff
and
they
have
all
their
employees
contributing
as
well.
A
A
There's
a
slack
channel
on
kubernetes
slack,
that
is,
cluster
hyphen
API
hyphen
vsphere,
there's
717
members
in
there.
So
that's
pretty
huge
as
these
things
go
and
it
seems
pretty
active,
there's,
obviously
a
GitHub
associated
with
it
too.
But
you
know
the
slack
channel
would
give
you
kind
of
an
initial
read
of
you
know
with
anything
open
source.
You
don't
want
to
attach
your
organization
to
some
open
source
project
that
isn't
healthy
and
one
of
the
aspects
of
health
is
just
how
many
people
are
active
on
this
and
I'd
say.
C
B
C
A
A
Aren't
often
the
end
users
so
much
as
other
disco,
vendors,
so
I
would
say
when
it
comes
to
Cluster
API,
the
most
vigorously
input
interested
users.
The
people
with
skin
in
the
game
are
actually
people
writing
their
own
kubernetes
distros,
because
they
don't
want
to
write
that
disco
over
and
over
again
for
every
cloud
out
there.
A
That's
that's
exactly
why
VMware
got
interested
in
this
as
a
concept
that
we
took
it
as
a
given
that
we're
going
to
have
multi-cloud
and
people
using
multiple
public
clouds
as
well
as
on-prem,
and
we
don't
want
to
write
a
kubernetes
distro.
That's
a
snowflake
for
everyone
out
there
and
the
cluster
API
was
in
an
attempt.
Just
like
cloud
provider
is
it's
to
abstract
whole
clouds,
it
for
purposes
of
deploying
kubernetes
clusters,
so
yeah.
C
A
Not
a
trivial
decision
to
be
made,
but
I'm
just
saying
you
know,
the
cluster
API
has
been
the
choice
of
a
number
of
people.
If
your
target
is
multiple
backing
clouds.
Now,
if
you're,
if
you
weren't
going
after
multiple
variations
of
cloud
that
you
know
you,
you
decided
that
the
only
thing
we're
going
to
do
is
on-prem
period.
Maybe.
C
A
B
C
Yeah
I
mean
they
give
you
a
bunch
of
options
too,
like
and
even
still
not
even
just
vsphere,
like
the
AWS
and
Azure
ones
like
like.
We
were
saying,
like
Steve,
said
the
like
more
of
the
time
it's
another
vendor
asking
for
something
for
like,
like
cluster
API,
to
be
fixed,
so
like
cap
Z,
which
is
the
Azure
one
like
you,
you'll,
see
like
VMware
employees
in
there
trying
to
fix
it
for
tanzu
like
so.
You
get
a
lot
of
different
interactions
with
like
a
different
stakeholders
which
helps
make
the
project
a
lot
better.
A
C
A
C
A
A
Well,
thanks
for
thanks
for
coming,
it's
good
to
chat
and,
like
I,
say
I'll
I'm
gonna
go
away
and
try
to
gather
up
resources
on
migration.
You
know
it
might
be
a
day
or
two,
but
I'll
see
what
I
can
gather
up
and
I'll
I'll
drop
it
in
slack
and
I'll
in
the
longer
term,
I
think
you're
right
that
getting
it
updated
on
GitHub.
So
these
things
are
shared
with
the
whole.
Community
is
the
longer
term
goal
here.