►
Description
March 16, 2021
Status Updates - items working towards roadmap
carlisia
Theme: unblocking items for the v1.6 release.
Description Status
#3489 [velero] Use Credential from BSL for restic commands (release-blocker) reviewed PR ✅
#3559 [velero] Add E2E test for multiple credentials (release-blocker) reviewed PR ✅
#3527 [velero] Prefer conditional waiting over magic sleep (release-blocker) reviewed PR ✅
#3583 [velero] Restore progress reporting bug fix (release-blocker) reviewed PR ✅
#3568 [velero] Update upgrade docs (release-blocker) PR merged 🆗
#3584 [velero] Improve GH Action PR assign + labeling PR open for review 📬
#3591 [velero] Improve plugin release instructions PR open for review 📬
AWS / GCP / Azure [plugins] Update changelogs PRs merged 🆗
Issue 3493 Initial scoping out of work for packaging with Carvel Meeting ☎️
-Next:
- push tags and run tests against v1.2.0 of the 3 Velero maintainted plugins
- plan the rest of the week and start planning/organizing for v1.7
- PR reviews
bridget
Focusing on 1.6 release blockers
Multiple credentials work is complete
dave
E2E tests with volume snapshots
Release candidate when E2E tests running/passing on AWS & vSphere
RC tests with latest/earliest K8S supported, Azure, AWS, Kind, vSphere, GCP
Manual tests pass
Discussion topics
carlisia
Welcome Eleanor, who officially starts this week! (PM for Velero)
e2e tests for Velero upgrades?
Contributor Shoutouts PRs
@codegold79: Restore progress reporting bug fix
Contributor Shoutouts Velero Q&A
None this week. We do have unanswered questions.
A
A
A
I
there
was
a
small
pr
to
update
the
upgrades
documentation
for
1.6.
That's
coming
up.
I
have
a
couple
small
prs
open.
I
mean
one
pr
open,
but
that's
not
release
blocker,
so
it's
gonna
sit
there
for
a
while
once
we
finish
this
release
at
least
the
release
candidates,
the
pr's
for
getting
the
plugins
ready
to
for
release
they
are
merged.
A
A
C
B
Everything
I
did
that's
a
nice
idea,
carly's
gonna,
like
it
I
yeah.
So
in
the
last
week
I
have
been
mostly
focusing
on
finishing
up
any
work
that
we
need
to
do
in
order
to
get
a
1.6
release
candidate,
so
as
part
of
that
all
the
multiple
credentials
works
finally
merged.
Thank
you
very
much
everyone
for
the
reviews
and
feedback.
There.
That's
been
really
helpful.
Yes
with
that
that
workup
merged
with
et
tests
and
then
yeah
I've
been
working
on
some
of
the
1.6
release
blockers.
D
Okay,
yeah,
so
we're
all
concentrating
on
getting
the
1
6
release
out.
So
I've
got
a
pr
in
where
we
have
ed
test
with
volume
snapshots,
and
that
is
that
will
give
us
some
better
testing
for
right.
Now
we're
doing
rustic,
we'll
test
the
volume
snapshots
as
well
that'll.
Let
us
exercise
all
the
plugins
and
we'll
need
to
set
up
the
test
so
that
we
run
with
the
multiple
credentials
as
well,
so
bridget
added
stuff
in
so
that
we
can
testimon
credentials.
So
we
can.
D
If
we
find
any
bugs
or
stuff
we'll
make
a
decision,
then
whether
or
not
those
are
show
stoppers
and
we
may
or
may
not
go
back
and
cut
another
release
candidate
and
we're
going
to
test
with
our
latest
and
earliest
kubernetes
supported
versions.
Don't
ask
me
what
those
are.
I
have
to
look
those
up
and
we'll
do
that
against
azure
aws,
the
local
kind
clusters,
vsphere
and
google
cloud
and,
as
bridget
mentioned
she's
going
through
the
manual
checklist
of
things
we'll
perform
against
those
as
well.
A
A
E
A
D
No,
and
that
should
be
relatively
easy
to
add,
since
we
can
specify
which
version
of
bolero
we
use
when,
when
the
tests
install,
so
we
could
have
a
upgrade
version
and
we
could
have
an
upgrade
test
but
yeah.
If
we
don't
have
a
pr
or
we
don't
have
a
issue
for
that.
Would
you
put
one
in
because
that's
a
good
thing
to
have.
D
D
F
A
A
He
had
he
had
to
go
through
a
lot
of
codes
to
pick
out
what
needed
to
be
put
back.
Basically,
there
was
a
pr
that
removed
some
codes
that
frankie
put
in,
and
she
should
put
it
back
basically
and,
and
it
had
to
do
with
the
api
group
versioning,
and
when
that
code
was
removed,
it
started
breaking
tests
yay
for
tests.
G
A
D
G
A
C
I
want
to
talk
a
little
bit
about
the
the
versioning
for
plugin,
the
plugin
right,
so
my
one
of
my
design
was
stuck
because
of
that
because,
basically,
when
we
have
the
valero
plug-in,
we
have
multiple
version
of
this
and
it
have
multiple
api.
We
need
a
way
to
control.
Listening
on
it,
and-
and
I
I
saw
there
is
a
somebody-
somebody
else
also
interested
in
that
topic
of
creative
versioning.
I
think
nolan
have
start
the
conversation
about
that.
C
D
C
H
Just
to
jump
in
on
that
as
well,
I
don't
have
a
lot
of
feedback
necessarily
on
the
version
design
itself,
but
I
will
be
involved
because
I
have
a
feature
as
well.
That's
kind
of
waiting
on
that,
because
the
the
wait
for
additional
items
to
be
ready
is
another
change
to
the
plugin
api.
So
we
really
can't
get
that
upstream
until
we
have
the
versioning.
D
H
Yeah,
okay
and
then,
by
the
way
I
I
do
have
an
open
pr.
I
had
a
design
for
this
that
was
approved
that
didn't
at
that
point
did
not
involve
plug-in
changes,
but
I
realized-
and
I
talked
to
nolan
about
this-
that
we
actually
do
need
to
change
the
plug-in.
So
I
have
a
pro
open
on
the
design
update,
so
I
don't
have
so.
D
A
All
right,
so
just
cleanup
kind
of
thing
I
am.
I
was
looking
at
the
v7
project
boards
and
I
saw
some
prs
in
there
and
pr
should
not
be
in
the
boards.
I
don't
know
if
it's
the
automated
system,
that's
adding
them.
So
I'm
gonna
clean
that
up
just
remove
the
prs
and
make
sure
the
corresponding
issues
are
in
there
if
they
are
for
1.7.
A
So
if
you
are
looking
at
that
and
you
see
a
pr
disappear,
just
it's
just
me
cleaning
things
up,
because
if
we
have
the
pr
and
the
ish
I
mean
we
definitely
need
the
issue.
So
we
don't
want
the
pr,
because
then
we
have
duplicates
of
everything
in
this.
It
gets
messy,
so
yeah.
We
just
need
to
make
sure
that
everything
we
are
working
on
has
an
issue,
so
we
can
track
on
the
board
at
least
some
of
them.
A
D
D
B
So
I'm
confused
about
which
so
there's
the
project
board,
but
then
there's
also
the
miles
the
milestone
as
well.
Sorry,
I'm
just
looking
at
the
current
line
project
v1,
I'm
looking
at
the
1
6
board
that
we
currently
have
and
we
do
have
columns
for
like
review
and
progress
and
reviewer
approved
like
it.
B
A
A
So
basically,
if
that's
what
it
is,
we
have
to
re
like
go
to
clean
up
the
like
the
way
I
was
talking
about.
You
go
on
the
pr
remove
the
the
issue.
That
is
a
board
thing
on
the
github
page.
I
don't
know
how
to
describe
it.
You
know
what
I'm
talking
about
so
that
that
lists
the
the
there's
a
board's
fields,
so
you
just
clear
the
the
milestone
from
the
board
field
for
that
pr.
A
B
A
A
D
D
A
A
A
D
A
I
Hi
dave,
so
I'm
just
wondering
if
time
permits
can
we
discuss
about
the
challenges
associated
with
running
multiple
snapshot
operations
in
general,
I
we
did
have
a
brief
chat
about
it
earlier.
I
remember
and
we
discuss
about
some
challenges
associated
with,
perhaps
twice
in
the
same
applications.
At
the
same
time,
if
we
run
concurrent
snapshot
was
that
is
that
the
only
challenge-
or
there
are
any
anything
additional
to
that?
Well,.
D
We're
gonna
have
to
go
back,
so
the
plugins
have
basically
been
written
to
work
in
the
environment.
They're
working
and
the
current
environment
is
essentially
single
threaded,
so
we'd
have
to
go
through
and
do
quite
a
bit
of
testing
against
the
existing
plug-ins,
all
those
to
make
sure
that
it
works
or
have
them
export
something
like
a
flag
that
says
that
they
support
concurrency
and
we
have
we
have
to.
We
have
to
understand
what
the
limits
of
concurrency
are
on
the
various
storage
systems.
I
I
I
D
D
We
want
to
start
bringing
things
up
the
tree
where
we
can
start
snapshotting
applications
and
have
the
application
respond
to
a
snapshot
request
if
it
supports
that.
So
the
the
reason
for
having
the
plugins
run
in
a
different
process
is
it's
really
it's
go.
So
if
we
were
doing
like
java
java
lets
you,
you
know
import
jars
fairly
easily
and
there's
mechanisms
there
for
handling
the
different
well
or
even
just
doing
it
at
runtime
and
go
doesn't
do
that
very
well.
D
D
Those
are
going
to
be
things
that,
for
example,
change
a
tag.
Add
an
annotation,
you
know,
manipulate
things
on
restore
to
make
things,
make
the
resources
fit
and
then
there's
going
to
be
things
that
are
more
like
snapshotty
things,
so
things
like
snapshotting,
a
a
volume
or
snapshotting
a
you
know:
kubernetes
cluster
eventually
and
we're
gonna
leave
those
external
and
that
will
also
we're
to
start
opening
up
the
api.
D
So
it's
not
just
callable
from
within
valero,
but
this
actually
becomes
a
piece
of
infrastructure
that
we
can
then
share
with
our
partners.
So
that's
on
the
roadmap.
We're
trying
to
figure
all
that
out
yeah.
So
the
those
are
the
goals
and
certainly
concurrency
in
snapshot
operations
are
important.
We
probably
have
to
do
things
like.
I
guess,
right
now,
the
quiesce
hooks
are
on
a
per
pod
basis.
D
So
if
there's
a
if
there's
an
execution
hook
involved,
we'd
have
to
do
like
we,
we
shouldn't
snapshot
items
from
a
pod
where
the
execution
hook
hasn't
been
executed.
Yet
if
there's
an
execution
hook
for
the
pod,
otherwise
we
could
kind
of
do
them
in
parallel.
So
I
think
we
need
to
look
at
it
because
it's
it's
relatively
there's
a
lot
of
room
for
complexity,
though.
I
G
D
D
No,
no,
no!
No!
It's
not
going
to
we're
going
to
be
able
to
tell
the
application
snapshot
yourself,
and
it
may
do
something
completely
different
than
snapshotting
volumes.
It
may,
for
example,
dump
out
like,
like
we
have
a
poc
for
postgres,
where
it
dumps
out
it
does
a
pg
dump
right.
So,
if
you
look
at
what
people
have
done
in
other
systems
or
like
cassandra,
you
know
has
its
own
backup
utilities.
We
want
to
start
triggering
those
rather
than
just
doing
volume
snapshots,
because
the
volume
snapshots.
I
D
So,
for
example,
say
we
have
like
a
cassandra
running
with
multi-way
replication,
so
each
of
its
disks
is
already
replicated
three
waves.
Now
you
snap
all
of
them
and
you
back
all
of
them
up
you're,
not
actually
gaining
yeah
right.
So
we
would
like
to
have
the
application.
Cassandra
actually
has
a
backup
and
restore
application.
That
is,
you
know,
applica,
that
is,
that
is
consistent
at
the
application
level,
and
we
don't
need
to
quiet
it.
We
can
say:
hey,
go
snapshot
yourself
and
it
may
be
able
to
do
that
without
stopping
the
world.
I
D
I
D
Yeah
so
there's
the
the
kubernetes
app
because
it's
there's
like
an
app
crd
and
we'll
take
advantage
of
that,
but
I
don't
know
that
everybody's,
so
we
can
start
to
build
things
around
that
the
I
want
to
take
this
problem
in
so
we
can
build
something
around
the
app
crd
we
will,
but
I
expect
to
see
applications
actually
exporting
their
own
snapshot
apis
because
even
building
something
around
the
app
crd,
you're
kind
of
guessing
at
what
the
application
does
right.
You
start
looking
at
you
go
okay!