►
From YouTube: 2020-05-12 :: Crimson SeaStor OSD Weekly Meeting
Description
No description was provided for this meeting.
If this is YOUR meeting, an easy way to fix this is to add a description to your video, wherever mtngs.io found it (probably YouTube).
A
We
have
four
come
here
and,
let's
start
as
weak
area,
what's
deer
walking
on
the
new
achievement
nation
under
and
the
unedited,
the
merge,
support
and
I
think
it's
basically
the
personality-wise
it
its
occasion.
I'm
trying
to
add
more
unit
tests
and
later
I
will
debit
to
based
on
stems
ability,
alienation
to
rewrite
the
mutation
and
the
duplication
and
allocation
and
release
released
them
with
with
they'll,
be
a
tree
based
an
implantation.
That's
my
plan
and
I
said
from
each
other
still
really
bring
peers
and
by
the
way
I.
A
A
A
C
C
D
Back
field
yesterday,
yesterday,
crimson
has
first
at
first
object
to
two
replicas
and
also
switched
to
the
backfield
state.
I
mean
it
went
in
successfully
went
there
active
and
clean
right
now,
I'm
I'm,
making
I'm
cleaning
up
the
branch
implemented.
Some
minor
missed
functionality
like
let's
say
sending
the
information
about
battle
process.
D
I
mean
operations
like
it
was
in
the
chat,
also
I'm
I'm
revising
the
branch
to
to
master
which
is
a
bit
cumbersome
because
because,
because
I'm,
based
on
the
pretty
old
version
of
of
the
recovery
branch
and
since
then
a
lot
of
things
have
moved
I
post
the
project,
I
need
to
catch
up
sit
on
that
I
was
reviewing
a
English
1a
PR
for
for
Adam,
it
looks,
it
looks,
sign
maybe
match.
This
is
very
small
thickness
may
be
that
we
should
the
couple
the
change,
lighted.
D
E
E
C
Yep
I
trust
the
first
round
of
comments
on
this.
The
first
PR
then
pushed
it
and
pushed
a
copy,
I'm
Keef,
who
I
think
you
used
to
wanted
to
look
at
the
cash
right
so
I'm
waiting
on
that
face
no
particular
hurry.
I
have
unit
tests
that
now
successfully
split
notes
and
I'm.
Now
working
on
the
incre
f
and
a
graph
logic
that
involves
removing
papier
mappings.
C
C
A
A
A
C
What
I'm
saying
the
interface
is
present,
but
a
lot
of
it's
not
on
you.
If
you
take
a
look
at
the
current
branch,
the
transaction
manager
has
one
unit
test.
That's
pretty
much
the
only
thing
that
works.
Ok,
if
you
extrapolate
out
to
the
Oba
manager,
you
can
guess
that
you
could
probably
allocate
a
lot
of
things
you
just
can't.
Decrement
ruff-ruff
counts
at
the
moment.
I
expect
to
have
that
this
week,
though,.
G
Yeah
for
the
Ono
tree,
Pro
top
two
things
are
implemented,
the
block
layout
with
index
t2
vacation
and
they
look
up
inside
all
types
of
of
the
nodes
and
they
are
highly
coupled
with
each
other.
So
I
want
to
make
this
stable
first,
and
the
second
part
is:
is
the
recursive
look
up
in
the
tree
down
to
the
leaf?
G
A
A
G
C
G
C
That
says,
omit
the
topple
elements
that
are
common
mm-hmm.
You
already
know
what
they
are
right.
So
if
the
level
n1
is
defined
as
not
having
the
first
three
topple
elements,
then
you
just
remember
those
three.
When
you
begin
reading
that
that
note,
you
know
what
they
are,
because
you
just
send
it
to
that
note
in
the
first
place,
yes,
yeah,
okay,.
C
If
you
just
you
know,
if
you
need
a
functioning
transaction
manager
before
I,
have
one
feel
free
to
create
or
to
interface
out
transaction
manager
and
write
us
and
and
write
a
stupid
in
memory.
Implementation
it'll
be
useful
for
you
to
test
anyway.
C
C
I
mean
this
is
I
tried
to
make
it
as
modular
as
possible
so
that
we
don't
have
to
actually
mount
disks
to
test
these
components.
We
want
to
be
able
to
exhaustively
test
the
v3
stuff
without
starting
up
a
single
sees,
for
instance,
and
then
we'll
sort
of
these,
for
instance
that
tests
it
even
work.
But
that
way,
we'll
see
test
failures
at
like
a
really
particular
component
and
not
just
see
store
crash.
A
F
And
then
we
wait
for
this
should
be
done
this.
This
is
done,
and
this
has
done
in
the
messenger
when
which
is
wait
for
all
the
continuations
or
all
the
event
consumptions
to
finish,
and
then
we
can
clean
up
all
the
things,
and
this
is
this
is
a
basic
basic
idea
and
I
think
the
main
discussion
should
be
around
the
idea
that
to
release
messengers
from
the
job
to
track
all
the
ongoing
event,
consumptions
and
letting
each
component
to
do
that
themselves.
A
Yes,
removing
removing
the
dispatcher
by
itself
when,
when,
when
it's
trying
to
is
notified
to
post,
are
shut
down,
and
then
the
messenger
will
not
expect
more
even
to
to
them
before
it
is
removed
from
from
the
chain
of
dispatcher,
which
is
the
pipe
into
the
messenger
later
on.
Miss
messenger
will
push
start
start
to
tear
down.
F
G
G
It's
event:
consumption
futures,
but
it's
fields
you
need
to
implement
increment
a
gate
to
if
there
are
some
other
internal
continuations
in
the
component,
though,
so
that's
why
I
think
if,
if
the
messenger
is
is,
can
be
the
coupled
with
the
to
track
the
event
consumption
futures,
then
the
entire
shutdown
procedure
can
be
much
much
simpler,
but
it
can
be
in
the
future.
Iterations.