►
From YouTube: November 2018 OpenZFS Leadership Meeting
Description
We discussed status of porting features to all the platforms, and strategies for doing so.
Agenda and meeting notes: https://docs.google.com/document/d/1w2jv2XVYFmBVvG1EGf-9A5HBVsjAYoLIFZAnWHhV-BM/edit?ts=5bb3b66c
A
A
A
A
We're
gonna
after
you
do
the
the
check-ins
on
a
couple
of
key
ones
for
the
last
time
that
we're
going
to
continue
where
we
left
off,
which
is
with
special
devices.
So
first
I
wanted
to
check
in
with
status
updates
on
to
saying
things.
First
is
encryption
of
the
big
one.
So
last
time
we
talked
about
getting
on
to
the
most
and
FreeBSD
and
I
think
we
talked
about
getting
on
the
OS
10,
but
Jurgen
is
already
finished.
That
looks
like
so
great
on
the
loo
most.
C
D
I'm
here
so
well
as
if
they
once
more
you've
found
the
P
R,
there's
a
couple
little
bugs
that
I
found
I
think
we're
just
the
consequences
of
being
from
emerging
so
done
that
there's
been
that
report
from
that
one:
a
user
about
corruption.
When
doing
a
CFS
send/receive
I've
been
trying
to
recreate
that
I've
had
no
luck,
you
or
that
I
was
hitting
these
other
issues.
Not
those
got
resolved
like
I've
been
trying
to
do
that
and
so
I'm.
D
D
The
CFS
folks
haven't
done
a
whole
lot
with
that
particular
part
of
it
in
the
past,
but
that's
kind
of
where
I've
been
standing
who's,
trying
to
figure
out
where
to
go
next
to
try
to
figure
it
out,
didn't
mention
he
has
an
l2
arc
device
which
I
didn't
know
about
before
I.
Don't
know
how
relevant
that
is,
but
that
may
play
into
it,
so
so
that's
kind
of
what
what
I've
been
doing
up
to
date
here.
D
E
If
it's
a
Ross
and
I've
had
some
good
luck
with
having
people
send
me
the
actual,
send
files,
because
there's
nothing
really
sensitive
in
them.
So
I
know
that
some
people
have
been
pretty
comfortable
doing
that
and
allowing
me
to
take
a
look.
Maybe
if
you
can
get
this
in
file
and
try
to
receive
it,
you
can.
D
D
A
Alright
yeah
I
mean
you
might
check
with
them.
I'm
sorry,
I'm
glad
to
hear
that
work
is
continuing
on
this
I'm.
Sorry
to
hear
that
you
haven't
been
able
to
reproduce
it
or
figure
out
what
the
problem
is
that
he's.
Seeing
have
you
I
know
he
was
using
smart
OS.
Have
you
have
you
ported,
the
most
stuff
or
whatever
rebased
John,
just
smart
OS,
and
have
you
got
him
using
your
version
of
it?
So
it
is
or
do
we
think
that's
not
relevant.
C
D
F
G
And-
and
they
don't
I
mean
so
so
you
know
as
soon
as
that's
fixed
I
mean
if
I
mean,
if
I
have
to
if
I
have
to
and
I'm
working
on
the
soft
stuff.
But
if
you
know,
if
I
have
to
do
that,
I
could
do
that,
but
I
mean
I
kind
of
sort
of
lumped
that
in
with
the
original
crypt
at
work.
So
anyway,
that's
so
I.
F
G
A
So
next
I
just
wanted
to
get
a
status
update
on
the
redacted
send/receive
code
review.
That's
out!
This
is
kind
of
swishing
the
agenda
up
a
little
bit
since
this
is
something
that
is
not
getting
any
repo,
but
I'm
kind
of
easy.
I
I
recognize
that
this
is
a
personal
interest
to
myself
and
I'm,
trying
to
elevate
the
visibility
of
it.
A
Like
not
necessarily
you're
saying
not
necessarily
breaking
out
the
review
into
multiple
reviews,
but,
like
saying
like
you
know,
you
Tom
can
review
the
man.
Page
changes
and
and
Brian
will
review
the
colonel
changes
and
Roman
will
review
these
life
changes
yeah
something
online,
does
sort
of
functional
or
like
code
lines,
yeah.
E
Even
something
like
this
cut,
you
know
this
is
the
DNA
changes
that
you
don't
be
supported,
and
these
are
the
art
changes,
and
these
are
the
send/receive
code
changes
and
I
I.
Don't
I
haven't
read
through
the
whole
pen,
so
I
know
those
probably
aren't
equal
segments,
but
I
think
if
we
don't
get
it
divided
up
like
that.
That
would
help
a
lot
with
just
you
know:
cuz
I
kinda
can't
justify
the
time
to
at
least
my
bosses
right
now
like
to
make
that
much
time
to
Brian.
E
A
A
A
A
Usually
you
use
this
like
as
a
an
SS,
you
know,
I
had
one
or
a
few
SSDs
to
your
pool,
that's
mostly
hard
disks
and
used
that
as
SSDs
for
metadata.
Do
you
table
or
other
kinds
of
things
like
that
and
there's
a
bunch
of
a
bunch
of
you
on
this
because
it's
been
presented
in
several
open,
zetas
conferences,
etc.
So
it
seems
like
you
need,
is
full
on
the
other
flat
firms.
I
think
so
we
need
folks
who
are
interested.
B
There
it's
marked
I
think
we're
interested
why
I
think
seasons
maybe
I
mentioned
there
can
say
that
we
started
there.
I
would
probably
prefer
just
trying
to
compare
divergence,
not
like
divergence
and
try
to
just
move
step
by
step.
Rosters
and
merge
disease
specifically
feature
only
but
yeah
I
think
we
are
interested.
A
So
if
I
understand
correctly
on
gender,
sorry,
the
audio
is
a
little
hard
for
me
to
hear
you
would
you
prefer
if,
like
it's
brought
to
a
Lumos
and
then
you
can
port
it
from
Lumos
to
freebsd?
But
if
that
doesn't
happen
then
you
know,
maybe
you
look
resort
to
doing
the
work
yourself.
To
put
it
all
the
way
from
linux,
freebsd.
B
Definitely
be
easier
to
pull
from
illumise,
but
I
think
everybody
understand.
Well,
really
we
should
change
source
vector
to
fortune
from
Linux,
which
means
we
need
to
analyze,
diversions
and
set
up
some
tracker
and
just
try
to
manage
things
more
sequentially
rather
than
trying
to
shape,
because
it
may
be
quite
difficult.
I
see
I
I
made
look
how
exactly
this
specific
change
significantly
problematic
or
not,
but
generally
I
just
actually
began
an
approach
problems
and
creates
more
problems
and
souls.
A
B
That's
I
assume
how
to
find
out
what
states
is
present.
Please
I'd
like
to
mention
that
I
see
significant
problems
in
the
fact
that,
for
example,
when
some
changes
measured
from
one
platform
to
another,
for
example,
from
humans
to
ZFS,
on
Linux,
they
while
formerly
managed
commit
by
commit
the
commits,
are
different
and
for
reasons
completely
unrelated
to
port.
A
in
efforts
sometimes
suggest
some
syntax
changes.
Sometimes
it's
some
minor
call
organization
changes
and
I
would
like
to
bring
people's
attention
that
that's
not
very
nice
as
a
platform.
A
I
I
think
it's
found
the
challenge
on
our
and
with
that
it
that
we
we
don't
really
want
to
have
commits
in
the
repository
that
are
not
themselves
complete
in
terms
of
like
you
know,
if
we're
making
changes
on
import
they're,
probably
you
know
to
make
the
startx
pass
or
possibly
to
make
it
work
at
all.
As
part
of
the
porting
I
mean
I.
B
Saw
examples
when
it
was
purely
syntactical
change
purely
for
margins
like
that
and
some
cases
just
one
function
was
used
instead
of
piece
of
code.
Doing
probably
the
same
well
I
agree
that
it's
one
nice
and
cool
for
a
final
point,
but
how,
after
that
we
can
stay,
is
to
remain
in
sync.
How
can
we
prevent
future
divergence?
How
can
we
leave
together
after
we.
F
A
Yes,
I
think
that
the
issue
that
Alexander
was
bringing
up
sounds
like
like
we're
we're
it's
things
that
are
not
actually
fixing
like
making
it
work,
but
but
stylistic
changes
like
you're
talking
about
I
think
that's,
definitely
something
that
we
can
be
aware
of
and
try
to
not
make
those
you
know
stylistic
or
like
your
code,
cleanliness
changes
while
doing
ports
and
instead
have
those
as
a
follow
on
and
I.
Think
that's.
You
know.
A
We
probably
need
to
make
sure
that
people
who
are
doing
this
points
push
back
on
the
reviewers
that
are
asking
for
stylistic
changes
but
I
definitely
take
kind
of
the
harder
problem
of
you
know
as
like.
After
the
initial
integration
feature,
there's
going
to
be
additional
changes,
maybe
that
fix
things
or
change
this
good
style
with
additional
commits,
and
those
needs
to
also
think
those
should
make
it
to
other
platforms
as
well,
and
that
I
think
that
kind
of
falls
under
the
bike.
A
It
kind
of
brings
attention
to
like
the
real
things
that
we're
missing
out
on
platforms
like
functionality
that
users
care
about
and
if,
if,
if
the
way
munna
addressed
that
is
by
like,
as
you
were,
relying
like,
let's
start
with
you
know
the
next
commit
number
one
and
get
that
onto
a
limo,
some
previous
to
you
and
then
the
next
scoot
number
two
and
then
come
in
number
three
just
kind
of
go
through
it
all
chronologically,
rather
than
looking
at
it
kind
of
feature
by
feature.
I
think!
That's!
A
I
So
literally,
definitely
a
tough
balance
and,
like
I,
do
think
it
will
be
easier
for
us
to
take
the
feature
commit
the
or
the
bodies
of
work
that
represent
observable
features.
If
we
are
also
relatively
up
to
sync
on
all
the
intermediate
stuff
that
maybe
isn't
directly
justified,
I
think
like
it's
important,
not
to
lose
sight
of
the
hidden
cost
of
choosing
not
to
focus
on
the
three
things
that
you're
not
interested
in.
It
comes
time
to
think
up
for
the
one
thing
that
you
are,
that
was.
J
Exactly
oh,
that
was
exactly
my
experience.
I
was
looking
at
the
feature
they
added
instead
of
us
on
Linux
for
I.
Think
it's
Zuko,
io
stat,
where
it
can
tell
if
the
extra
parameters
on
the
end
are
a
pool
name
or
a
V,
dev
name
and
so
on,
and
that
was
going
to
be
useful
for
the
Vita
properties
work.
But
when
looking
at
trying
to
pull
that
over
to
FreeBSD,
it
quickly
became
well.
J
G
Because
in
order
to
get
z-test
working
again
and
fix
it,
some
existing
issues
I
had
to
pull
in
a
bunch
of
other
bits,
and
you
know
the
closure
at
dependencies
means
that
we're
not
just
adding
crypto
support
or
basically
turning
stuff
and,
above
and
beyond
that,
so
limiting
ourselves
to
culture
it
cherry-picking.
The
features
we
want
is
trying
to
doesn't
seem
to
be
really
be
that
viable
solution,
I
mean.
E
K
Also
adds
to
the
fact
that
each
of
the
communities
that
are
involved
here,
like
so
Tom
and
I,
are
you
know
mostly
zo
l
people,
but
like
there's.
Also
people
in
you
know
the
previous
to
the
illumos,
and
you
know
those
are
different
communities
that
have
different
priorities
and
the
cross
community
communication,
like
at
least
from
my
perspective,
is
not
as
if
there.
If
it's
there,
it's
not
as
obvious
as
I,
it
probably
should
be,
or
maybe
it's
not
as
clear
I,
don't
know,
but
like
maybe
some
something
along
those
lines
might
help
improve.
I
Think,
from
our
perspective,
in
illumos,
we
so
like
Ed
again
at
Joint,
we're
not
super
interested
in
multi
input
projection,
for
instance,
and
though
I'm
fascinated
by
the
allocation
classes
stuff,
it's
not
going
to
be
directly
relevant
to
us,
because
we
I
mean
we
were
for
it
to
become
relevant.
We
would
need
other
changes
in
other
parts
of
the
business
that
involve
different
bills
of
materials
and
service.
For
instance,
we
need
to
put
more
SSDs
and
things
like
that,
so
it
is.
I
Not
directly
like
we,
we
see
a
path
to
using
it,
but
I
think
that
we,
if
we
reframe
it
as
making
it
easier
to
get
the
things
that
we
want.
Then
we
can
instead
do
what
we've
been
talking.
We
talked
about
last
meeting
a
little
bit
and
we
haven't
found
time
to
do
it
yet,
but
I
think
the
intent
is
still
to
pick
a
line
in
the
like.
K
I
was
just
about
to
ask
about
that,
but
also
perhaps
I
don't
know.
Maybe
this
may
be.
This
may
or
may
not
make
sense
like
have
some
kind
of
categorization
tag
for
like
clumps
of
things
that,
like
a
specific
feature,
for
example,
when
Tom
contributed
encryption
that
were
follow
up
commits
on
top
of
it
to
deal
with
some
of
the
you
know,
fixes
or
other
aspects
and
like
if
you
wanted
to.
K
A
So
we're
coming
up
on
or
we
have
passed
the
30
minutes,
I
wanted
to
make
a
few
kind
of
remarks
and
before
letting
you
guys
go
on
one
is.
This
is
a
great
discussion.
I,
definitely
appreciate
it
like
that.
We're
having
discussions
about
like
how
should
we
be
doing
this,
who
should
be
involved
in
not
just
like
sticking
to
you
know,
living
by
the
spreadsheet,
and
secondly,
I
also
is
not
lost
to
me
that
this
is
the
second
meeting
that
we
have.
A
Like
an
hour
is
probably
this
funny
as
a
half
an
hour
me,
so
you
got
just
an
hour
still
every
month.
So
if,
if
folks
think
that's
a
good
idea
press
the
green
check,
yes
button.
If
you
want,
if
you
think
something
else,
would
be
preferable,
press
the
no
X
red
button,
that's
that
you
can
see
when
you
click
on
show
participants
all
right.
All.
E
E
I
also
think
that
for
some
of
these,
like
issues
that
were
talking
about
now,
which
are
kind
of
like
really
important,
how
we're
gonna
do
things
in
the
future,
maybe
having
a
few
additional
meetings
just
to
figure
that
stuff
out
and
then
we
can
go
to
one
hour
once
a
month
plus
we
help
them
figure
it
out.
Yeah.
A
I
A
A
From
github
to
the
open
sea,
fest
developer
mailing
list,
we're
gonna
drill,
a
disable
like
last
week
or
something
I'm
gonna
clear
their
responses.
If
you
want,
if
you
want
you
to
re-enable
them,
then
vote
yes,
if
you're
happy
with
the
way
it
early
is
with
not
getting
a
lot
of
spam
on
that
mailing
list,
then
no.
A
L
M
H
M
A
Sure-
and
part
of
that
is,
you
know,
having
multiple
multiple
people
who
are,
you
know
in
charge
to
some
degree
and
part
of
being
in
charge.
It
is
you
know,
career
people,
requests
and
reading
stuff
directing
to
the
gate
so
well,
I'm
happy
to
help
out
in
this
regard
and
I
hope
that
we
any
future
we'll
have
more
more
qualified
people
who
are
all.
A
L
A
All
right
so
I
apologize.
There
were
over
time
again.
We
knew
things
that
we
have
here
well,
I
think
we'll
probably
on
an
agenda
item
to
discuss.
These
pilot
look
concerns
some
more
next
time.
If
you
have
other
agenda
items,
please
add
them
to
the
document.
The
Google
Doc
and
thank
you
all
for
coming.
Thank.