►
From YouTube: Velero Community Meeting/Open Discussion - May 5, 2020
Description
May 5, 2020
Status Updates
steve
updated container base images to ubuntu:focal
fixed a CRD restore issue related to unstructured conversion functions
fixed an issue with unresolvable includes/excludes
updated plugin server code to ignore unrecognized flags, and added special handling for --features flag in Velero
reviews
carlisia
continuing with the new BSL: controller (https://github.com/vmware-tanzu/velero/pull/2490); New: added ValidationFrequency / LastValidationTime, adding validation on BSL create/update events, using status/phase to validate backups
ashish
CSI work:
Deleting CSI objects and cleaningup storage snapshots on backup deletion
Syncing in CSI objects during backup sync-in. Mainly to ensure they are deleted on backup deletion/ expiry.
Testing.
Fixup in persisting backup here
PR reviews
Investigate why we are unable to install ca-certificate package on ARM builds
nrb
Almost done with CRD back up work in https://github.com/vmware-tanzu/velero/pull/2478. Want to grab a few more samples today if I can
specifically want to get prometheus objects as mentioned in https://github.com/vmware-tanzu/velero/issues/2370
Reviewing CSI work, need to finish up the describe output too
On community support this week
Will be handling the v1.4 release
Will be on Crossplane’s TBS YouTube series Thursday with Carlisia
Discussion Topics
[carlisia]
Want to add a Tilt setup for rapid development for 1.5. Here’s an example from the cluster api team: https://cluster-api.sigs.k8s.io/developer/tilt.html
Contributor Shoutouts
@jaygridley - [Azure] add support for static storage account keys with restic (https://github.com/vmware-tanzu/velero/pull/2455)
@jones-paul1525 - [AWS] update AWS SDK to enable some new auth methods (https://github.com/vmware-tanzu/velero-plugin-for-aws/pull/41)
@dunglas - [Helm chart] Add support for secret env vars (https://github.com/vmware-tanzu/helm-charts/pull/99)
@austinbv - [Helm chart] Add documentation for having separate providers (https://github.com/vmware-tanzu/helm-charts/pull/100)
@hasheddan [Velero core] Reviewing #2490 for CRD handling
A
A
We
have
one
discussion
topic
today
and
if
you
want
to
add
in
something
else,
please
do
so
if
you
want
to
just
chime
in
throughout
the
conversation,
please
do
so
as
well
just
unmute
yourself
and
with
that
I'm
gonna
share
my
screen
and
get
started
all
right
so
for
status.
Updates
this
week.
First
off
we
have
Steve.
B
We
were
on
Bionic
before,
but
vocals
out
now
so
I
pitted
those
we
had
a
CRT
restore
issue
related
to
the
unstructured
converters
that
are
part
of
the
probably
API
machinery
where
certain
fields
don't
convert
correctly
and
we've
seen
this
on
the
backup
side
as
well.
But
we
had
another
instance
of
it
in
the
restore
code
so
made
it
fix
there
I
fixed
another
issue
related
to
included
excluded
resources.
Where
there's
a
potential
bug
worth.
B
If
they
were
unable
to
be
resolved
through
discovery,
you
could
potentially
end
up
getting
more
resources,
backed
up
than
you
and
you
hope
to
or
potentially
having
plugins
run
for
more
resources
than
you
intended.
So
I
fixed
this
super
straightforward
fix,
but
but
it's
kind
of
a
big
impact
and
then
I
also
updated
our
plugin
and
the
basically
the
server
code
that
we
intend
plug-in
authors
to
use
to
serve
their
plugins.
B
Previously,
this
code
would
error
if,
if
an
unsupported
flag
was
passed
to
the
plug-in
server-
and
we
changed
this
to
to
basically
ignore
unrecognized
flags,
so
this
lets
us
easily
add
new
flags
in
velaro
core
that
get
passed.
The
plugins
and
plugins
won't
immediately
break
if
they
don't
support
these
flags,
so
we
can,
we
can
add
flags
and
kind
of
a
backwards
compatible
way
and
then
yeah
beyond
that.
Just
just
lots
of
PR
reviews
on
everything
else.
That's
going
on!
So
that's
that's!
That's
pretty
much
been
it
for
me.
For
the
past
few
days,.
A
B
Not
for
me,
I'm
I,
think
most
of
the
stuff
that
I've
been
working
on
is
wrapped
up.
I
do
have
one
PR
that
still
open
the
back
up
progress
PR
that
I
know
everyone
has
has
taken
a
look
at,
but
that
will
need
to
get
finalized
this
week,
but
for
the
most
part,
trying
to
review
other
open,
pr's
and
and
get
our
scope
closed
for
so
that
we
can
ship
the
1.4
beta.
B
C
Yeah
so
I
continued
to
work
on
the
new
backup,
start
location
controller.
Everybody
had
a
chance
to
review
it
and
it
was
very
helpful.
So
basically,
we
have
this
loop
that
validates
if
the
store
is
ready
or
if
it's
valid
so
now
we're
going
to
I'm
going
to
add
the
same
process
to
the
DSL,
creates
and
update
events.
So
we
do
that
validation
there
as
well,
and
keep
updating
the
face
for
the
for
the
location
CRD.
C
C
A
D
We
want
to
clean
up
the
story
snapshots
when
backups
are
deleted
or
like
they
expire,
so
I.
Take
that
and
then
I
have
an
ongoing
PR
to
sync
of
CSI
objects
into
the
cluster
when
Valero
sinks
bass
backups
into
the
cluster.
So
the
idea
is
again
that
if
the
backups
expire
or
anything
or
get
deleted,
we
want
to
make
sure
that
all
the
resources
that
were
created
during
the
backup
get
deleted.
D
There's
one
more
thing
that
I
noticed
when
we,
so
we
also
stored,
like
the
snapshot.
Csi
volume
snapshot
objects
into
as
part
of
the
backup
and
I
noticed
that
we
only
store
volumes
after
contents
that
are
associated
or
bound
to
a
volume
snapshot.
But
we
also
want
to
store,
like
dandling
volumes,
not
shut
contents
that
are
that
may
be
part
of
a
backup
just
so
that
they
get.
They
are
synced
in
and
deleted
whenever
it
is
time
so
I'm
going
to
work
on
a
fix
for
that.
D
D
D
Yeah
there's
this
other
thing
that
we
that
came
up
somebody
was
unable
to
talk
to
the
GCP
from
their
raspberry
pi
from
the
container
running
on
raspberry
5
turns
out.
They
were
missing
the
CSL
packages
on
their
container,
so
somebody
from
the
community
put
up
a
fix
to
install
C
asserts
as
part
of
the
container
builds
in
the
docker
file,
but
unfortunately
we
are
not
able
to
run.
A
D
E
E
Yeah,
the
the
rest
of
the
a
lot
of
CSI
reviewing
I
need
to
finish
up
the
describe
output
that
shouldn't
be
as
involved
as
the
rest.
Most
of
it's
just
making
sure
we
purse
the
stuff
correctly
I
may
have
to
think
about
what
we
want
to
parse
it
off
of
whether
it's
volume
snapshot
or
volumes
natural
contents
and
then
I
don't
have
a
whole
lot
else
to
talk.
E
E
Carly
C
and
I
will
be
on
Cross
Plains
TBS
youtube
series:
that's
a
link
to
their
playlist,
don't
think
we
have
an
episode
link
yet
so
it'll
be
probably
at
the
bottom
of
that,
and
that
will
be
at
noon.
Eastern.
This
Thursday
as
far
as
blockers
I,
don't
think,
can
I
have
any
right.
Now,
the
the
main
things
gonna
be
trying
our
best
to
get
a
get
a
beta
release
out
by
the
end
of
the
week,
4
1,
4
and
testing
across
lots
of
different
versions
of
gruber
Nettie's,
particularly
1:18.
The
CSI
works.
E
A
You're,
so
you
were
on
Bret
Fisher's
YouTube
last
week
and
TBS
podcast
this
week,
khaleesi
air
you'll
be
and
you'll
be
there
as
well.
This
Thursday
lots
of
media
exposure
love.
It
there's
great.
Thank
you
so
much
for
volunteering
and
stepping
up
and
doing
this
really
really
awesome
job
for
the
1.4
release
Nolan
since
you're
leaving
that
do
you
also
want
to
do
the
blog
post,
for
it
sure
I
can
do
that.
Also,
if
you
want
I
can
have
an
issue
and
yeah.
E
Yeah
and
something
that
just
reminded
me
of
like
one
upgrade
step
I
know
people
will
have
to
do
is
with
the
backup
progress
folks
will
have
to
upgrade
their
CR
DS
there.
Since
there's
a
new
field
now
through
fields,
so
I
know,
at
least
one
thing
will
have
to
be
updated
to
update
your
series
in
terms
of
like
mechanics
of
doing
an
upgrade,
but
that's
something
we'll
probably
be
flushing
out
through
the
beta
I.
E
C
Yes,
so
I
have
I,
took
a
look
at
the
tilt
setup
that
the
cluster
API
team
team
did
and
I
thought
was
awesome
and
I
would
love
to
do
if
ever
wanted
to
get
a
check.
If,
if
this
is
something
that
we
would
want,
I
personally
would
love
to
have
that
set
up
for
for
just
myself.
But
there's
you
know
if
I
do
it
for
me,
I
might
as
well
do
it
for
the
project.
C
So
basically
we
would
check
in
that
shield
file
and
the
the
end
result
would
be
that
anyone
could
just
put
a
fill
in
a
JSON
file
with
some
variables
and
that
just
run
tilt
up
and
the
whole
thing
will
put
up
and
start
running
and
the
way
they
did
it.
I
was
playing
around
with
some
time
ago
in
the
way
I
feel
and
I
was
happy.
C
That
I
just
figured
out
how
to
make
a
run,
but
it
was
very
interesting,
but
it
wasn't
very
practical
because
the
way
I
did
it
was
building
an
image
and
then
sending
the
image
to
the
cluster,
which
is
takes
some
time
and
we
do
have
a
way
to
develop
locally.
Just
using
the
binary,
which
is
much
faster,
so
the
way
that
and
the
goats
scene
setup
for
the
this,
this
that
set
up
for
the
class
API
team,
you
you
will
build.
C
The
tilts
will
build
the
binary
based
on
the
fire
that
changed
and
you
specify
what
files
it
should
be
watching
in
the
best
binary
and
sends
it
to
the
pods
and
restarts
the
part
in
that
it
should
be
as
fast
in
angel
side
is
pretty
fast.
It
should
be
as
fast
as
running
the
pilot
in
there
locally,
except
that
you
don't
have
to
like
the
other
binary
start
started
phalaris.
C
It
would
just
keep
running,
and
then
you
also
have
the
pretty
UI
in
green
Reds,
as
as
you
keep
breaking
or
fixing
things
that
would
just
appear
so,
oh
yeah,
so
I
wanted
to
check.
If
anybody
would
not
like
to
add
that,
because
I
think
we
could
also
add
to
our
documentation
and
be
one
additional
way
that
people
could
it,
it
would
be
an
easy,
an
easy
entry
to
Valero
developments
at
least
I.
Think
so.
A
C
If
you
were
developing
for
so
well,
we
will
have
instructions
just
like
that
documentation
that
the
classified
team
has-
and
there
is
a
that-
will
be
a
JSON
file
with
some
specific
variables
you
have
to
fill
in.
For
you
a
use
case,
you
know,
so
you
are
like
what
plug-in
you're
using.
So
you
just
add
the
provider
name
and
we'll
know
what
plug-in
that
corresponds
to,
and
basically
so
you
fill
up
that
basic.
Very
few
things
run
the
command
tilts
up
at
the
root
of
velaro,
and
that's
it.
C
So
much
cycle
so
I
would
say
just
look
watch
over
this
directory
and
as
I'm
developing
I
will
see
in
the
browser
the
the
log
running
and
changing,
and
also
a
visual
cue
of
if,
if
I
broke,
the
server,
for
example,
we
will
turn
red
as
I,
see.
Sarah
logs
will
turn
red
and,
as
you
fix
things
will
turn
green.
C
Accent
I
mean
that
is
really
good
to
be
looking
at
the
log
in
and
seeing
what
what
messages
you
it's,
you
know
you
are
triggering
as
you
make
changes,
but
octane
gets
its
again
through
the
REO,
the
value
of
things,
this
state
of
things
different
different
tools.
Although
you,
you
can
also
look
at
the
log
through
octants
I.
A
C
F
C
F
E
C
B
Yeah
I
can
take
the
top
two
here
and
then
other
folks
can
jump
in
first
one.
We
have
us
from
Jay
Gridley,
who
had
a
previous
contribution
to
the
azure
plugin
to
add,
support
for
static
storage
account
keys,
so
this
is
mostly
relevant
if
you're,
not
using
volume
snapshots,
maybe
you're
using
rustic
or
just
backing
up
the
kubernetes
manifests,
and
you
don't
want
to
use
an
azure
service
principle.
B
So
this
supports
just
using
a
static
storage,
account
key
that
you
pass
in
as
an
environment
variable
through
the
secret,
and
so
this
PR
here
adds
support
for
that
for
using
rustic
with
azure.
In
this
way,
I'm
not
actually
necessitated
some
changes
to
bolero
core
for
reasons
that
we
don't
need
to
get
into
right
now,
but
anyway,
yeah
thanks
Jake
Ridley
for
wrapping
that
work
up.
B
That
should
be
shipping
in
1.4
and
then
the
next
one,
just
a
quick
one
from
Jones
Paul
1525
who
updated
our
AWS
plug-in
to
use
a
much
more
recent
version
of
the
AWS
go
SDK,
which
enabled
some
new
authentication
methods
specifically
for
the
use
case
around
STS,
endpoints
and
I.
Don't
know
all
the
details
but
appreciate
the
update
there.
C
My
turn
the
Arabic,
the
next
two
ones
so
so
to
shouts
for
help
chart
contributions.
Danglers
it's
his
handle.
He
had
a
support
for
adding
secret
environment
variables.
Recent
perfect
credentials
and
I
was
in
the
V.
He
had
a
documentation
for
a
contribution
that
he
had
made
in
the
past,
but
he
was,
it's
really
appreciated
that
he
came
back
and
the
documentation
for
having
separate
providers
and
the
last
one
is
not
mine.
E
Yeah,
that's
for
me,
hashed
dan,
who
I
believe
is
also
the
host
of
the
PBS
show
works
upstream.
A
lot
with
Sierra,
B's
I
believe
there's
a
CRD
repo
might
be
kidnapped
back
home,
so
she
or
he
that
he
owns
so
he
hopefully
came,
came
along
and
provided
some
insight
onto
how
sororities
work
and
also
provided
a
review
on
PR
2498.
So
thank
you
for
that.