►
From YouTube: Kubernetes UG VMware 20230504
Description
May 4, 2023 meeting of the Kubernetes VMware User Group. Covered recent updated to vSphere as it relates to hosting Kubernetes including the CSI storage plugin and the Cloud Provider. Also announced the plan "twilight" this group and perhaps reformulate another user group with a broader on-prem coverage mission.
A
Hi
welcome
to
the
May
4th
meeting
of
the
kubernetes
VMware
User
Group
on
the
agenda.
Today
there
wasn't
much
unless
somebody
just
added
something
to
the
agenda.
Node
stock
I
do
have
a
few
announcements
of
updates
to
the
CSI
plug-in
and
the
vsphere
cloud
provider,
but
I,
don't
think
those
will
take
more
than
a
few
minutes
and
as
usual,
if
we
don't
have
enough
agenda,
anybody
here
is
welcome
to
do
late
nominations
verbally
on
things.
They
want
to
talk
about,
discuss
or
just
share
interesting
observations
with
others.
A
Before
we
get
into
this
I
think
before
I
hit
record,
somebody
was
mentioning
that
they
believe
that
Google
Groups
are
being
end
of
life.
I,
don't
really
know.
I
do
know
that
within
the
kubernetes
sigs
working
groups,
Etc
they've
had
an
aspirational
goal
to
move
to
something
other
than
Google
Groups
for
years
literally,
but
nothing's
ever
happened.
A
One
of
the
issues
that
makes
the
Google
Groups
attractive
is
that
the
kubernetes
project,
like
many
other
open
source
projects,
shares
documents
by
using
the
Google
Docs
platform
and
Google
Docs
conveniently
allows
you
to
give
permissions
to
membership
lists
for
Google
Groups.
A
So
if
you
were
to
move
to
something
that
maintained
membership
list
kind
of
out
of
the
Google
Universe
I
think
that
you
might
be
signing
up
for
quite
a
bit
of
work,
because
you
might
have
to
do
something
like
maintain
a
list
of
emails
and
apply
permissions
to
those
individually
and
even
when
you
get
into
these
things
like
lists
of
emails,
parts
of
the
world,
like
the
EU
in
California,
have
strict
privacy
standards
on
those
and
personally,
as
a
group,
maintainer
I
would
really
not
like
to
go.
A
There
I'd
rather
kick
that
off
on
someone
else
where
you
have
things
that
potentially
could
go
on
like
filing
forms
to
be
forgotten,
and
things
like
that.
So
I'm,
not
really
sure
I
haven't
been
informed
that
in
the
near
term,
any
plans
are
underway,
but
it
could
be
I
shared
the
agenda
notes
document.
If
anybody
doesn't
have
it
speak
up
and
I'll
post
it
again
in
the
chat.
A
If
you
join
late,
the
chat
rolls
off
and
you
don't
get
an
opportunity
to
see
it
again,
but
what
I
have
in
there
is
I
cut
in
pasted
updates
to
the
CSI
plug-in
there
were.
There
was
an
update
to
CSI,
just
in
the
last
week
that
fixed
a
bug
related
to
sync
caused
by
deleting
a
persistent
volume
claim
and
says
that
they've
resolved
a
panic
observed
in
the
Sinker
container
during
the
sync
process.
So
those
sound
like
pretty
significant
bugs
that
got
fixed
a
week
ago.
A
There
was
also
about
a
month
ago,
a
271
release
that
included
new
sidecar
versions.
I
believe
these
sidecars
are
actually
inherited
from
work
and
storage
Sig.
A
So
when
new
versions
are
released,
they
they
get
picked
up
by
individual
plugins.
Potentially
the
links
to
the
actual
full
descriptions
of
those
are
in
the
agenda
notes
document
and
moving
on
to
the
vsphere
cloud
provider.
A
version
1.26.1
was
released
just
yesterday,
the
it
updated
the
version
of
Debian
to
fix
a
cve.
It
fixed
the
CV
in
golang
and
updated
a
TLS
cert
for
end-to-end
test.
A
There
were
also
similar
releases
if
you're
on
older
versions
of
kubernetes
a
release
1.25.2
for
kubernetes
1.25
and
similarly
one
for
1.24
went
out
yesterday.
A
It
was
a
Helm
if
you
install
lap
by
health
chart
that
was
updated.
Four
days
ago,
on
the
vsphere
front,
there
was
an
update
to
the
VCR
8
U1
docs
just
yesterday,
and
there
were
updated,
install
binaries
or
patch
binaries
put
out
on
April
18.
once
again
for
vsphere
Aid
update
one.
There
have
been
no
recent
updates
to
vsphere
7
update
3
in
the
past
month.
A
So
that's
kind
of
my
recap
of
what's
been
going
on
on
the
summary
front
and
I
see
in
the
chat
bands.
Thanks
for
posting
that
update
on
the
demise
of
Google
Groups.
A
Now
we
had
one
person
who
just
joined
the
group
for
the
first
time
because
because
he
had
a
conflict,
but
I
wanted
to
point
out
to
people
and
I
discussed
this
last
month
with
Bryson
that
the
kubernetes
steering
committee
has
asked
that
this
group
be
end
of
life.
They
originally
had
a
plan
that
these
user
groups
would
be
commissioned
for
all
of
the
cloud
providers,
VMware
vsphere.
A
Being
one
of
them,
but
other
examples
would
be
the
Amazon
AWS
running
on
Google's
GCS
or
Microsoft's
azure,
but
as
it
turns
out,
nobody
with
these
other
vendors
or
other
Cloud
providers
ever
put
in
place
any
structure
to
start
one
of
these
groups,
so
they
never
took
place,
and
this
has
been
the
only
group
operating
it
also
being
tied
to
the
cloud
provider,
has
kind
of
left
the
impression
that
some
things
have
been
shut
out.
You
know
nominally
this
group
isn't
geared
around
the
VMware
distribution
of
kubernetes.
A
It
could
support
any
distribution
of
kubernetes,
whether
it
be
Rancher,
red
hat,
openshift
or
whatever
so
long
as
it
runs
on
the
vsphere
platform.
But
there
is
an
interest
in
the
kubernetes
and
the
cncf
to
morph
this
group
into
something
more
like
an
on-prem
user
group,
so
that
maybe
we'd
get
more
vigorous
participation
with
from
people
like
red
hat,
who
might
support
on-prem
kubernetes
and,
as
it
turns
out,
a
lot
of
the
things
involved
with
running
on
VMware
might
apply
generically
to
any
on-prem
use
of
kubernetes.
A
You
features
like
a
load,
balancer,
perhaps
features
and
related
to
Ingress
when
you
do
kubernetes
on
top
of
VMware
vsphere
or
bare
metal,
on-prem
you're,
taking
a
lot
of
those
collateral
tools
on
yourself
and
historically,
when
we've
had
group
meetings
at
least
half
the
meanings
have
related
to
sidebar
things,
that
kind
of
are
used
in
conjunction
with
kubernetes
on
vsphere,
but
might
also
apply
generically
to
running
kubernetes
just
in
any
on-prem
environment,
perhaps
on
a
different
hypervisor
or
no
hypervisor
at
all.
A
So
we've
kind
of
held
discussions
on
opening
up
an
on-prem
User
Group
over
under
the
cncf,
rather
than
the
the
kubernetes
project,
and
the
reason
they
would
prefer
to
have
that
under
the
cncf
rather
than
under
the
kubernetes
project,
which
is
where
this
group
is
operating
now.
Is
that
a
lot
of
these
collateral
tools
like
take
load
balancers,
where
you
might
have
a
load
balancer
such
as
cube
VIP?
A
Maybe
we
could
aspire
to
getting
that
running
in
the
fall,
but
starting
up
these
kinds
of
things
in
the
summer
is
kind
of
rough
too,
because
having
watched
these
groups,
much
of
the
world
has
heavy
usage
of
vacation
during
the
summer
months
and
attendance
at
anything
from
you
know,
online
Zoom
meetings
to
physical
conferences
tends
to
drop
pretty
dramatically
during
the
summer
period,
so
I'm
thinking
that
there
might
be
a
replacement
group
popping
up
sometime
in
the
fall,
but
anyway
that's
the
summary
that
I
wanted
to
point
out
to
people
so
they're
not
shocked.
A
When
should
should
you
see
this
group
stop
having
meetings?
That
would
be
the
explanation.
I'm
told
we'll
keep
the
slack
Channel
around,
and
the
Google
group
does
have
a
mailing
list.
So
if
some
replacement
group
pops
up
under
the
cncf
I
would
expect
that
I'd
be
able
to
send
notices
indicating
to
people
where
this
now
lives
so
I'll
give
up
the
I'll
get
off
the
soapbox
and
open
it
up
for
just
group
discussion.
B
B
Just
you
know
introduce
myself
a
minute
ago,
but
just
to
add
to
that
we
run
openshift
at
Vanderbilt
and
the
the
engineers
that
set
up
the
environment
have
left
the
company
and
we
don't
have
too
much
deployed
there.
But
the
rest
of
us
are
kind
of
dipping
our
toes
in
the
water
and
kind
of
learning
the
environment,
and
you
know
getting
a
feel
for
what
our
service
offerings
you
know
should
look
like
and
how
best
to
support
it.
B
So
you
know
we
we
do
run
everything
on
VMware
and
you
know
openshift
kind
of
has
their
own
forums
and
things
as
well
too,
but
it
is.
It
is
nice
to
have
other
resources
for
for
on-prem,
whether
it's
through
you
know
be
aware
whether
it's
through
Docker
vmug
might
even
be
a
good
option.
If
there's
not
already
a
vmug
community
set
up
under
the
Community
section
of
vmug
for
for
kubernetes,
that
would
probably
be
a
potential
place
to
Pivot
to.
A
C
A
B
B
They're,
mostly
Regional,
at
least
on
the
vmug
site
and
and
I
oftentimes
I,
don't
even
think
to
go
back
and
you
know
visit
these
regularly.
But
there's
like
a
community
section
under.
B
Vmug
in
my
communities-
and
you
know
some
of
those
are
like
you
know-
there's
one
for
for
home
Labs,
you
know,
there's
like
I,
don't
see
it
now
under
my
roots,
but
it
seems
like
there
used
to
be
a
health
care
related
one,
and
so
these,
if
nothing
else,
it's
kind
of
a
place
to
migrate
user
discussion.
Yeah
potentially,
and
you
kind
of
have
support
from
you-
know
the
mug
leadership
as
well
just
kind
of
something
to
throw
out
as
a
as
a
possibility.
B
D
B
Well,
I
mean
slack
is
great,
for
you
know
a
chat
room
communication
anyway,.
A
Yeah
I
would
say,
the
criticisms
of
Slack
are
within
a
specialized
area
like
take
VMware
I,
think
I
have
been
somewhat
policing
it
or
trying
to
do
the
best.
I
can
and
you
can
get
response,
but
you
know
you,
you
normally
wouldn't
expect
to
get
it
within
an
hour
or
something
at
least
in
the
VMware
sector,
because
it
kind
of
would
work
better
I
think
if
there
was
more
activity,
but
if
there
aren't
that
many
inquiries,
what
happens?
Is
the
engineers
don't
look
at
it
that
often
and
it
results
in
high
latencies
you've.
A
Though,
like
a
there's,
a
generic
one
called
kubernetes
users
that
has
the
opposite
problem,
where
you've
got
users,
putting
questions
about
everything
under
the
sun
to
the
point
where
you
might
get
20
messages
in
an
hour,
sometimes
at
Peak
periods,
and
nobody
can
afford
to
pay
attention
to
that.
Or
you
know
I
think
most
people
have
regular
day
jobs
and
and
that
job
isn't
policing
a
slack
channel.
A
So
even
if
you
think
you'll
go
in
there
and
try
to
help
some
users
out,
you're,
not
you're,
maybe
going
to
do
that
once
a
day
or
once
a
week,
but
then
there's
hundreds
of
messages
in
there
and
you're
not
going
to
dig
back
through
them,
or
at
least
that's
my
own
perception.
So
it
results
in
a
lot
of
inquiries
in
that
kubernetes
user
thing
being
more
or
less
seems
to
get
ignored
because
they
kind
of
roll
off
beyond
the
history
range.
A
Where
anybody
who
might
answer
them
is
willing
to
scroll
back
that
far
so
kind
of
not
enough
categorization
I
think
is,
if
you
don't
categorize,
you
can
get
a
lot
of
volume.
But
if
the
topic
is
too
General,
you
can't
get
anybody
willing
to
sift
through
them
for
the
things
that
they
might
care
to
comment
on.
B
Yeah
I
think
discussion,
forums
and
chat
rooms,
kind
of
meet
different
needs.
You
know.
For
that
reason
you
know,
and
particularly
if
you're
I
mean
oftentimes
when
I'm
having
an
issue
I'll
go
to
a
forum
and
you
know
do
a
search
to
see
if
somebody's
already
posted
the
same
topic
and
answered
it,
and
it
makes
it
easier
to
kind
of
archive
things
or
look,
look
up
keywords
and
but
forms
aren't
as
good
for
just
socializing
and.
B
Community,
you
know
getting
to
know
people
or
giving
those
immediate
responses,
so
I
think
they
they
both
serve
a
place
and
for
the
time
being
your
your
Google
group
is,
you
know,
serving
The
Forum
part
of
the
need,
but
there's
I've.
A
Got
a
recommendation
for
how
you
might
make
it
work
that
I
I
haven't
heard
this
frequently
recommended,
but
you
know
if
I
were
in
the
shoes
of
a
user
and
of
course,
I
worked
for
VMware,
so
it's
a
vendor,
but
I
would
go
to
one
of
these
slack
channels
and
search
for
something
in
the
same
category
that
I
have
a
question
about,
or
an
inquiry
about
and
find
the
names
involved.
Now
it
might
not
even
be
your
exact
question.
A
It
might
be
something
that's
a
year
old,
but
get
the
name
and
I
can
tell
you
that
at
least,
if
they're,
a
VMware
employee
you
could,
then
you
know
put
their
at
sign
in
their
handle
in
one
of
the
channels
and
say:
hey
I
I
have
this
question
about
this
and
I
I
saw
that
you
commented
on
something
similar
in
the
past
by
any
chance.
A
A
You
put
your
name
out
there
in
public
in
a
slack
like
the
kubernetes
slack
you're
kind
of
signing
up.
For
that
to
be
something
that
might
happen.
You
know
it's
not
like
and
I
think
the
odds
are
fairly
high.
That
people
would
pay
attention
as
opposed
to
I.
Don't
know,
there's
some
forums
that
are
almost
always
going
to
be
spam,
like
I'm
I'm
thinking
LinkedIn
from
my
perspective
is
something
that
if
people
try
to
ping
me
on
LinkedIn,
my
general
default
is
I'm
likely
to
ignore
it.
D
I
I,
that's
where
I've
got
I
mean
that
might
have
been
how
we
got
in
contact
initially,
it
is
just
like
what
you're
doing
exactly
what
you're
saying
you
know:
finding
the
person
that's
working
on
whatever
project
or
whatever
thing,
you're
working
on
and
paying
them
asking
for
help,
but
you
know
find
them
even
on
the
the
GitHub
right
and
then
go
see
if
they're
in
and
slack
yeah
in
the
kubernetes.
A
Slack
I
can
tell
you,
as
on
the
kubernetes
project,
with
regard
to
GitHub.
There
are
so
many
notifications,
like
you
might
touch
a
piece
of
code
and
then
it
it
sends.
You
emails
about
a
change
to
that
code.
Three
years
later
and
I
think
most
devs
I
know
have
their
GitHub
notifications
at
least
muted,
if
not
completely
turned
off
or
going
to
some
folder.
They
never
look
at
because
the
signal
to
noise
ratio,
yeah.
D
Looking
at
it
to
find
who
it
is
and
then
try
to
find
them
on,
kubernetes
seems
to
be
a
successful
on
a
kubernetes
needs
to
be
more
successful,
but
a
point
I
was
going
to
make.
Is
you
know
what
that
actually
is?
One
of
the
benefits
that
we've
had
with
this
meeting
is
being
able
to
get
when
people
do
have
questions
that
may
not
be
nice,
maybe
a
broader
question,
I
guess.
But
when
we've
had
questions
about,
say,
hey
how's,
the
CSI
driver
work
for
this
how's,
the
migration
different
things.
D
A
A
The
early
days
you
had
a
lot
of
first-time
adopters
and
that
results
in
a
different
sort
of
question
that
is
broader
in
nature
and
more
applicable
to
everybody,
and
one
of
the
things
I
think
kubernetes
is
reaching
maturity.
Level
kind
of
like
Linux,
did
10
years
ahead
of
it
to
where
it's
dropping
down
below
the
water
line
and
people
are
even
moving
up
into
higher
level.
A
Stacks
that
run
on
top
of
kubernetes
or
tool
chains
that
run
on
top
of
kubernetes,
so
that
the
average
I.T
person
doesn't
necessarily
have
to
get
into
the
internal
Plumbing
understanding
of
kubernetes
that
much
anymore.
It's
also
more
stable
and
there's
less
feature
churn
going
on.
D
C
D
We
move
on
to
that,
so
there
there's
still
questions
I,
think
out
there
to
be
had
by
I'm
sure,
there's
others
that
are
in
a
similar
situation
that
maybe
not
part
of
this,
but.
D
Yeah
I
think
that's
the
been
the
biggest
benefit
of
this.
These
meetings,
this
group.
A
A
Worthwhile
too,
is
that
the
Ingress
mechanism
within
kubernetes
is
on
a
a
path
to
being
deprecated
and
they've
moved
to
something
new
called
API
Gateway
and
the
reason
they
moved
is
that
there
was
a
recognition
that,
when
they
specced
out
Ingress,
they
were
a
little
too
loose
on
the
spec,
so
that
people's
actual
experience
with
using
Ingress
is
that
you
could
use
something
like
nginx
or
Contour.
Or
you
know,
there
must
have
been
10
different
ingresses
and
they
all
work
differently
enough.
That
stuff
wasn't
actually
portable,
which
is
you
know.
A
The
claim
of
being
Cloud
native
is
that
you
could
pick
up
your
workloads
migrate
them
to
a
different
platform
that
was
kubernetes
compliant
and
it
should
just
work.
But
the
fact
is,
with
Ingress,
any
implementation
of
Ingress
tended
to
be
different
enough
from
any
other
one
that
if
you
move
from
one
to
another,
it
didn't
just
work
and
the
API
Gateway
is
a
mechanism,
that's
kind
of
locking
things
down
to
a
tighter
degree
under
the
covers
it.
A
A
But
if
you're,
using
Ingress
at
all,
I
would
say
that
API
Gateway
is
probably
something
you
want
to
take
a
look
at
now
and
maybe
make
yourself
a
one
or
two
year
plan
to
you
know,
look
at
look
and
see
if
it
looks
like
you
might
get
any
benefit
out
of
it
and
if
so,
make
a
plan
to
do
a
migration
see.
Vince
has
to
move,
but
a
great
chatting
to
events.
A
And
you
know
the
other
thing:
I
I
got
back
from
kubecon
Europe,
the
other
one
that
was
interesting.
If
you
use
service
mesh
was
that
it
looks
like
the
upcoming
release
of
istio
has
some
very
significant
improvements.
They're
they're,
getting
rid
of
the
sidecars
and
it
appears
I
have
not
used
it
myself
yet,
but
it
appears
to
be
a
quite
significant
re-architecture.
A
A
They
do
by
the
way
I
was
talking
to
the
people
at
the
istio
booth
it
they
were
inviting
me
to
help
kind
of
document
it
or
maybe
give
a
future
presentation
on
it,
but
it's
anticipated
that
some
of
these
features
will
indeed
still
work
to
do
migrations
of
non-kubernetes
resources
participating
in
a
service
mesh.
That's
managed
within
kubernetes.
So
by
that
I
mean
you
can
create
a
mesh
in
kubernetes,
but
then
join
services
or
workloads
that
are
not
in
kubernetes.
A
I
personally
think
that
that
has
great
attraction,
I've
I've,
sometimes
been
skeptical
of
efforts
to
run
stateful
services
on
kubernetes,
as
opposed
to
maybe
hosting
them
in
VMS.
You
know
take
something
like
a
stateful
database
server
or
something
that
has
had
Decades
of
experience
running
it
reliably.
You
know
in
in
a
on
something
like
the
VMware
hypervisor
and
it
treats
it
in
a
different
way.
You
know
the
the
the
standard
summary
of
kubernetes
versus
VMware
VMS,
where
you'd
have
things
like
high
availability.
A
Are
that
when
it
runs
under
VM,
the
infrastructure
is
prepared
to
treat
it
as
a
pet
where
you,
if
it
gets
unhealthy,
you
keep
it
alive.
You
nurse
it
back
to
life,
and
the
kubernetes
philosophy
is
more
like
you
run
a
bunch
of
them
in
parallel
with
horizontal
load
balancing
and
if
something
gets
sick
you
just
shoot
it
in
the
head
and
grow
yourself,
another
one,
and
it
sounds
good
in
theory,
and
there
are
some
stateful
services.
A
I,
don't
know,
Cassandra
mongodb,
that
from
the
get-go
were
architected
for
that
kind
of
philosophy,
but
there
are
others
that
really
grew
up
in
an
earlier
era.
That
Maybe
word
architected
with
that
kubernetes-like
philosophy
and
if
you
could
join
them,
hosted
in
Legacy
Technologies,
but
still
take
advantage
of
service
mesh.
That
might
be
rather
interesting
and
result
in
kind
of
more
desirable
behavior
for
certain
organizations
it.
A
You
know
everybody
is
different
and
I
understand
that
there
are
those
who
believe
that
you
want
to
move
everything
under
kubernetes,
but
I
think
there's
room
for
other
options,
particularly
when
you're,
using
some
of
these
things
in
conjunction
with
Legacy
workloads
that
have
not
made
the
move
to
kubernetes
but
need
to.
You
know
where
you
need
to
share
access
with
these
kinds
of
tools
with
your
kubernetes
workloads.
A
Anyway,
I
think,
if
that
subject
is
of
interest
to
you,
there
were
some
talks
on
that
at
kubecon.
Europe
I,
don't
think
that
they've
posted
those
for
Public
Access,
yet
they
kind
of
embargo
them
for
a
little
while
to
encourage
people
to
pay
for
tickets.
But
historically
the
the
cncf
has
published
those
within
a
month
or
two
I
did
notice
that
some
of
the
the
so-called
co-located
events,
the
ones
they
had
on
the
first
day
of
kubecon.
Those
already
have
been
published
up
on
YouTube
and
the
videos
seem
to
be
all
up
there.
D
So
I
I
have
a
question.
This
isn't
specifically
I,
guess
kubernetes,
but
could
be
like
any
of
VMS
on
VMware,
but
we've
seen
some
stuff
with
maybe
the
network
or
storage,
or
maybe
the
network
causing
vsan
to
have
some
issues
that
would
cause
the
VMS
to
go
into
a
read-only
file
system.
State
I'm
I
mean
up
to
this
point.
We
are
just
we're
monitoring
for
that
and
then
rebooting
the
nodes.
D
Sometimes
they
reboot
fine
and
start
to
go,
read,
read,
write
again,
and
sometimes
they
need
like
an
fsck
to
get
going
again
and
so
then,
that
right
now
is
currently
requiring
manual
intervention
I'm
just
curious
in
general.
Is
there
a
what's
the
best
way
to
handle
that
situation?
Well,.
A
I
think
that
if
you're
on
vsan,
presumably
that
you've
got
full
support,
so
if,
if
you
do,
I
would
just
go
through
the
actual
support
organization,
rather
than
try
to
go
through
the
you
know,
the
open
source
kubernetes
slack
to
get
to
somebody.
If
you
really
wanted
to
go
to
the
you
know-
and
it
isn't
really
a
kubernetes
issue
either
if
it's
vsan
itself
going
out
and
it
affects
VMS,
whether
they're
running
kubernetes
or
not,
and.
D
I'm,
not
even
I'm,
not
even
sure,
if
it's
it's
not
necessarily
A
VMware
issue
like
it
could
be
an
underlying
networking
issue.
That's
that's
impacting
vsan.
C
A
Yeah
in
terms
of
correlating
it,
you
know
the
best
tool
I've
used
for
once
again.
You'd
have
this.
This
is
not
open
source
and
it's
licensed
technology,
but
the
V
realize
Automation
and
kind
of
the
observability
tools
that
VMware
has
we'll
look
at
all
the
logs
of
the
different
layers.
You
know
all
the
way
down
from
the
hypervisor,
the
networking,
the
storage
and
kind
of
correlate
things
to
attempt
to
deduce
a
root
cause.
Not
everybody
has
those
though,
but
if
you
do,
those
I
think
you
get
lucky
over
half
the
time.
A
I
actually
believe
that
if
you
ran
something
like
the
pure
open
source,
elk
stack
or
something
like
that,
that
has
those
sorts
of
capabilities
as
well
and
for
all
I
know,
I've
never
run
them,
but
you
know
kind
of
alternate
commercial,
observability,
stacks
datadog
and
whatever
I
I
kind
of
believe.
They
all
have
at
least
attempted
to
put
tools
like
that
in
there.
That
would
try
to
have
the
tool
help
you
out
as
opposed
to
having
humans
who
know
all
these
different
levels.
A
If,
for
some
reason
you
know
pursuing
support,
isn't
working
for
you,
there
are
the
people
on
kubernetes
slack
who
are
into
storage
like
Shang
Zhang,
who
I
think
has
talked
to
this
group
before
and
if
you
look
at
the
people,
who've
responded
to
inquiries
about
the
CSI
plug-in.
You
could
ping
one
of
them
just
because
you
know
to
inquire
they.
A
They
aren't
likely
to
have
the
networking
knowledge,
but
they
maybe
could
peer
up
or
find
somebody
who
does
are
you
when
you
do
this
I
mean
I,
think
the
recommendation
has
always
been
that
you
dedicate
some
storage
bandwidth
to
the
storage,
as
opposed
to
your
workloads
too.
A
Not
everybody
does
that,
but
I
think
that
that
is
the
suggestion
in
the
docs
and
if
you're
using
virtualized
networking
you
know
through
the
VMware
platform,
there
are
ways
where
you
can
dedicate
bandwidth,
although
in
a
lot
of
on-prem
data
centers
they
actually
even
dedicate
physical
infrastructure
to
the
storage
Network.
You
know
so.
They'll
have
multiple
Nicks
inside
each
server,
and
some
of
these
are
really
only
serving
the
storage
backplane
or
maybe
storage,
plus
control,
plane
and
the
actual
VM
workloads
are
on
their
own
independent,
dedicated
mix.
A
C
C
A
Yeah,
so
you
know
generically
all
I
could
suggest.
Is
you
know
going
down
into
the
logs
looking
for
warnings
and
errors
related
to
networking
and
seeing
if
any
of
those
are
occurring
at
the
same
time,
you're
seeing
these
storage
issues,
maybe
even
not
only
looking
for
errors
and
warnings,
but
maybe
looking
at
utilization
graphs
and
seeing
if
maybe
these
anomalies
correspond
to
things
going
up
to
100
percent.
A
Because
most
of
these
things,
you
know,
if
you
get
a
lot
of
Network
retries
and
things
that
pretty
much
always
is
going
to
end
up
in
a
log
somewhere.
A
Same
thing
with
connection
dropouts.
D
D
I
mean
there's,
there's
some
redundancy
there,
but
it
we
have
smaller
clusters,
there's
yeah,
so
I
I
mean
we
know
that
there's
going
to
be
some
of
that
impact
at
the
infrastructure
level.
So
the
question
I
I,
guess
my
question
really
is
not
getting
back
to
rooting
that
out,
but
the
best
way
to
handle
when
those
file
systems
go
into
read
only
are.
C
D
We
put
in
like
a
a
monitor
and
memory
that
will
see.
Hey
I
went
read
only
it's
been
read
only
too
long,
so
we're
going
to
reboot
the
node
and
that
that'll
that'll
come
back
up
and
we'll
save
quite
a
bit
at
the
time.
But
you
may
get
multiple
VMS
on
the
same
cluster
that
come
up
fine
and
then
one
that
doesn't
and
says
it
needs
enough.
It
needs
the
file
system,
checked.
D
A
D
D
A
Yeah,
sorry-
and
you
know
any
anything
else-
we've
got
15
minutes
left.
If
anything,
anybody
else
wants
to
bring
anything
up.
Todd
I
I
see
you're
still
on,
but
I
haven't
heard
from
you.
A
And
either
you
if
we
do
go
through
with
this
idea
of
getting
an
on-prem
User
Group,
you
know
the
the
cncf
would
like
to
have
some
users
sign
up
to
be
kind
of
management
of
the
group
if
you
will,
or
at
least
sponsors,
because
I
think
they're
always
cautious
about
having
them
turn
into
kind
of
promotional
vendor
pitch
forums
and
I
I.
Think
that
that's
a
good
philosophy.
A
So,
if
either
of
you
are
interested
in
being
one
of
those
user
sponsors
should
we
be
able
to
form
a
group,
an
on-prem,
User
Group
under
the
cncf,
let
me
know
because
I
can
I
should
be
able
to
get
you
listed
in
that
role.
There
might
be
some
benefits
of
that
role.
If
you
ever
go
to
kubecon
they'll
at
least
recognize
you
or
you
know,
maybe
there
might
be
circumstances
where
you
could
even
get
a
free
pass
or
something.
If
you
give
a
talk.
A
A
A
That
said,
I
plan
to
hold
another
meeting
first
Thursday
in
June,
thanks
for
coming
and
we'll
see
you
then
bye.
Okay,.