►
Description
How to troubleshoot common issues with Peter Rabbitson, followed by a long Q&A from our community and hosts.
Learn about upcoming events at: https://spacerace.filecoin.io/events
Chapters:
A
So
over
the
next
three
weeks,
we'll
be
hosting
a
series
of
events
where
you
can
hear
from
the
falcon
team
and
from
yourself
the
fall
coin:
mining
community
week,
one
this
week
we're
starting
off
with
a
mining
workshop.
Tomorrow
we
have
the
opening
ceremony
later
in
the
week,
we've
got
the
minor
show-and-tell
and
to
end
off
the
week.
A
We've
got
the
space
race,
roundup,
we'll
be
talking
about
who's
top
of
different
tables,
and
who
is
our
our
community
champion
more
on
that
later,
you
can
find
all
the
details
on
our
upcoming
events
on
space,
race,
dot,
file,
coin
dot,
io,
slash
events
and
you
can
register
to
attend
there.
Also
we're
also
excited
to
hear
from
you.
A
So,
if
you're
interested
in
pretending
and
presenting
at
one
of
our
upcoming
minor
show
and
tells
then
you
can
apply
using
the
link
in
the
chat,
a
reminder
that
by
attending
this
event
and
all
space
race
events
that
you
agree
to
the
falcon
code
of
conduct-
okay,
let's
kick
things
off
with
falcoin
engineer,
peter
rabbitson,
who
you
may
know
from
slack
and
github
as
reba
he'll,
be
taking
us
through
how
to
troubleshoot
common
issues
and
then
follow
with
a
live
cuny.
So
please
get
those
questions
ready
over
to
you.
Reba.
B
Hi
everybody
so
exciting,
exciting
times
we're
we're.
Finally,
here
the
the
the
space
race
is
launched,
sectors
are
ceiling,
people
are
happy,
and
I
guess
for
those
of
you
who
are
just
joining
us.
I
would
like
to
go
through
some
of
the
very
basics
in
this,
like
very
short
presentation,
and
then
we
will
going
to
leave
pretty
much
the
rest
of
this
hour
for
any
questions
that
would
come
up.
So,
as
angie
already
mentioned,
we
are
like
literally
here.
B
There
are
so
many
exciting
things
that
will
happen
going
forward.
So,
first
of
all,
what
are
we
doing?
We're
mining
falcon
for
this
stuff.
B
We
have
a
slack
where
we
talk
about
all
kinds
of
things
and
the
central
channel
where
all
the
fun
is
happening
is
space
race,
and
I
believe
somebody
will
drop
a
link
in
a
moment
of
how
to
actually
join
the
slack
and
we
just
started
a
new
channel
for
extra
announcements
and
things
like
that
like
when
you
should
upgrade
to
a
new
version
of
plato
or
flotus
or
anything
else
that
is
very
important
to
space
race.
B
The
space
the
channel
is
called
space,
race
announcements
and
I
very
much
recommend
everybody
who
is
in
space.
There
is
all
else
or
elsewhere
to
join
it
and
listen
to
any
kind
of
you
know
news
that
we
have
to
share
with
you
in
the
coming
three
weeks
cool.
So,
let's
jump
straight
into
it.
This
is
a
very
small,
concise
list
of
things
you
generally
need
to
do
in
order
to
build
lotus
to
just
you
know
to
just
to
just
join.
B
If
you
have
never
done
any
of
this
before
generally,
you
can
use
any
computer
as
long
as
you
have
128
gigs
of
memory
and
sufficient
hard
drive
space.
I
will
talk
a
little
bit
more
later
in
the
presentation
about
the
different
timings
and
things
like
that.
But
in
order
to
get
to
get
running,
you
need
to
install
some
dependencies,
which
I'm
giving
an
example
here
for
ubuntu
and
debian
based
systems.
But
you
know
on
red
hat
and
others
will
be
something
similar.
B
We
need
to
open
cl,
build
essential
ceiling,
git,
curl,
jq,
regular
stuff.
You
need
to
install
a
recent
goal.
114
go
115
might
work
already
on
the
latest
version
of
lotus.
I'm
not
sure
about
that.
I
believe
it's
already
fixed.
B
You
need
to
install
rust
in
order
to
build
the
really
number
crunchy
part
of
lotus
from
scratch,
because
it
is
really
important
for
the
timings
that
I'm
going
to
have
one
slider
here
and
what
do
we
do
then?
Then
we
clone
lotus.
B
We
check
out
the
currently
latest
version.
0.53
that,
like
will
be
a
number
of
other
versions
as
we
go
forward
and
fix
things
through
our
entire
run
of
the
space
race,
and
then
you
build
it
like
this.
It
is
important
to
supply
these
rust
flags
and
if
I
build
from
swords
this
is
listed
in
several
places
in
our
documentation,
but
when
people
were
asking
questions,
I
have
seen
a
number
of
you
not
do
that.
This
is
important.
This
will
make
a
very
big
difference
in
your
ceiling
performance
going
forward.
B
So
we
will
photos
if
you
already
have
lotus
and
you
want
to
rebuild
it
even
simpler.
You
already
have
everything
checked
out.
You
have
everything
installed.
You
just
need
to
say
git
fetch
to
by
the
way.
There
is
a
lot
of
commands
here
because
git
is,
you
know
excellent,
but
basically
what
you
need
to
do
is
you
need
to
git
fetch
in
order
to
get
all
the
latest
stuff
from
our
server.
You
need
to
check
out
the
tag
that
we
announced
in
the
in
the
channel
that
I
just
talked
about.
B
You
need
to
do
exactly
this
because
it
is
weird
sometimes,
and
then
you
just
make
everything
again
and
you're
good
to
go
cool,
so
you're
wondering
like
why.
Why
do
we
build
all
this?
So
what
do
we
do
once
you
are?
I
lost
a
slide
okay,
so
I
was
going
to
point
out
how
to
how
to
start
your
lotus
note.
B
B
Then
you
create
a
wallet
you
back
it
up
and
then
you
just
start
the
demon,
and
there
are
a
few
more
things
where
you
need
to
go
to
to
a
faucet
grab
some
falcon,
but
many
of
you
already
listening
here
are,
like
you
have
gone
through
this
multiple
times
already
during
our
calibration
and
if
you
are
having
trouble
and
the
documentation
is
not
sufficiently
clear.
B
Please
join
us
on
slag
and
we'll
be
very
happy
to
help
you.
So
let's
get
into
the
middle
of
it
for
those
of
you
who
haven't
done
any
kind
of
mining.
Yet
what
does
mining
actually
involve?
Once
you
have
everything
running
as
a
node
in
the
network,
you
need
to
first
tell
it
like
hey.
B
I
have
some
storage
available
and
you
do
this
by
issuing
the
command
lotus
miner
sectors
pledge,
which
means
that
I
pledge
my
existing
space
to
the
network
and
I'm
ready
to
keep
telling
the
network
that
this
space
is
healthy
and
and
in
good
standing,
and
you
will
be
continuously
doing
this
throughout
the
space
race
in
the
actual
live
network.
A
lot
of
this
will
be
taken
by
actors
on
the
network
which
are
giving
you
data
to
store.
B
So
all
this
allocation
of
sectors
and
not
putting
it
on
the
network
and
so
on
and
so
forth
will
be
happening
automatically
for
you
once
you've
done
this.
You
can
look
how
your
miner
is.
Actually
you
know
how
it
is
doing
by
issuing
lotus
miner
info.
It
will
tell
you
what
your
current
balancing
falcon
is.
It
will
tell
you
how
many
sectors
you
have.
It
will
tell
your
id.
B
It
will
tell
you
if
you
have
any
penalties,
because
you
didn't
communicate
with
the
network
correctly
things
like
that
and
for
every
sector
that
you
are
currently
working
with,
and
I
use
the
works.
The
word
sector,
which
means
32
gigabytes
of
data,
for
this
particular
network.
B
You
can
see
by
calling
lotus
minus
sectors
lists
what
sectors
you
have
and
you
can
see
for
each
individual
sector
what's
happening
with
that
by
issuing
this
command
lotus
miners
sector's
status,
dash
dash
log
and
the
number
the
numbers
start
from
zero
and
go
to
as
many
sectors
as
you
have
cool
so
again,
very
basic
stuff.
I
just
want
to
run
through
it
real
quick
for
those
of
you
who
are
not
sure
you
know
what
what
we're
doing
here.
B
So
what
happens
to
a
sector
while
it
is
going
through
the
through
the
motion,
so
to
speak,
so
the
first
thing
a
sector
either
we
are
filling
it
with
zeros
or
we're
filling
it
with
deals
that
come
from
from
the
internet.
This
takes
about
two
minutes.
Depending
on
on
your
hardware.
B
Then
we
go
through
a
very
very
intensive
process
called
pre-commit
one
which
can
take
about
four
hours,
and
I
apologize
it's
3am
here
and
I
am
making
a
lot
of
mistakes
in
my
slides,
so
this
particular
stage
is
actually
the
one
that
would
greatly
benefit
if
you
have
an
amd
cpu
and
I
will
fix
the
slides
before
I
post
them,
post
them
for
for
general
consumption.
B
So
if
you
have
an
empty
cpu
with
a
recent
instruction
set,
this
part
will
take
four
hours
for
this
process
to
complete,
and
this
is
the
main
portion
that
is
ensuring
that
you
need
to
have
a
relatively
general
purpose:
computer
and
they're
not
building
weird
asics,
like,
for
example,
in
bitcoin,
where
the
entire
point
is
just
to
get
everything
faster
without
without
care
for,
for
you
know,
for
for
the
environment,
for
the
usability
of
hardware
and
so
on
and
so
forth.
B
So
this
is
what
guards
the
network
from
this
kind
of
activity
once
this
takes
place
and
by
the
way,
if
you
don't
have
an
amd
cpu
instead
of
four
hours,
this
very
well
can
take,
I
believe,
11
hours.
So
you
really
want
to
make
sure
that
your
hardware
is
correct
and
this
building
from
source
that
I
was
talking
about
earlier.
You
also
need
to
do
that
in
order
to
hit
this
number
or
better.
B
Then
we
have
a
approach
called
pregnant
two
which
either
will
use
as
many
cores
as
you
have,
or
if
you
have
a
gpu,
it
will
use
that,
so
it
can
take
about
three
hours.
Then,
in
order
to
synchronize
everything
with
the
network,
you
need
to
wait
a
little
bit
without
anything
happening
on
your
machine.
B
This
is
currently
chosen
for
about
one
and
just
a
little
bit
hours
to
make
sure
that
nobody
can
cheat
to
basically
try
to
put
on
the
network,
something
that
you
don't
yet
have,
and
after
this
finishes,
then
we
have
another
two
hours
during
which
we
again
either
work
through
on
all
your
cpus
or
we're
using
the
gpus
that
you
have
on
your
machine
and,
yes,
it
is
connected
in
general.
B
All
of
all
of
this
can
happen
faster.
If
you
have
very
well
tuned
hardware-
and
you
were
really
you
know-
taking
the
time
to
tune
everything
during
the
calibration
period
that
just
finished
it
was
in
the
past
three
weeks,
but
in
general,
if
you
are
like
a
hobbyist
and
you
just
put
a
relatively
good
machine
together,
this
is
kind
of
what
you
are
expecting
to
seal.
B
A
second-
and
you
also
can
always
get
an
up-to-date
requirement
of
various
how
much
memory
is
required
from
which
cpu
is
required
by
running
this
command
lotus
worker
run.
Help
will
tell
you
like,
in
the
current
version
of
photos,
that
you're
running
what
will
be
required
for
the
expensive
ones
of
these
stages
cool.
So
we
covered
that.
B
Let's
move
forward,
we're
almost
there,
so
you
are
mining
and
your
miner
is
like
doing
its
thing
and
so
on
so
forth,
and
you
will
the
moment
you
put
something
on
the
network.
You
will
show
up
on
our
dashboard,
which
is
on
space,
baseball
coin
io,
and
it
kind
of
looks
like
this,
and
if
we
like,
let's
unpack
a
little
bit,
what
do
we
have
here?
So
we
have
the
miners
and
we
have
a
certain
rank.
B
This
is
a
particular
action
that
each
miner
needs
to
perform
during
this
network,
just
to
make
sure
that
this
functionality
will
work
for
everybody.
This
is
how
many
sectors
they
have
sealed
so
far,
and
this
is
measured
in
bytes,
so
each
sector
is
set
to
gigabytes.
So,
for
example,
again
it's
still
in
the
morning.
B
Actually,
I'm
not
sure
how
many
cycles
I
did
terabytes,
but
you
know
you
get
the
idea,
and
then
we
have
an
interesting
thing
here
that
says,
like
storage
of
success
rate
and
the
tv
will
deal
successfully
in
order
to
make
sure
that
not
only
we
just
throw
storage
in
the
network,
mindlessly
we
put
together
a
network
to
simulate
what
the
real
file
coin
mainnet
will
look
like
where
various
other
nodes
will
be
contacting
your
miner
and
trying
to
store
data
on
it
and
then
retrieve
it
back,
and
this
number
here
is
exactly
what
is
happening
for
each
of
these
individual
miners.
B
B
I
will
talk
a
little
bit
about
what
the
exact
timings
are
for
this,
but
if
something
goes
wrong
you
will
get
a
particular
dink
like,
for
example,
this
person
over
here
has
like
six
seven
percent
success
rate,
and
if
this
is
you,
you
can
totally
like
click
open
lock,
and
you
will
see
your
types
of
errors
and
things
like
that
and
we'll
dig
into
this
in
in
a
moment
as
well.
Another
portion
that
you
can
see
here
is
unclaimed.
B
Minor
hold
the
thought
we
will
talk
about
this
at
the
very
at
the
very
end
of
the
presentation
cool.
So
how
do
these
deals
actually
work
every
two
to
six
hours,
depending
on
your
previous
successes,
your
previous
failures
and
so
on
and
so
forth.
The
bots
will
make
a
deal
with
you.
B
They
likely
will
make
several
deals
in
quick
succession
like
two
or
three
of
them.
Again.
It
really
depends
a
lot
on
your
previous
interactions
with
the
pod,
because
the
bot
has
to
be
very
careful
not
to
overload
you
with
deals
that
you
cannot
fulfill
and
then
tell
well.
You
know
this
person
is
not
dealing
with
me,
so
I'm
just
going
to
give
them
0
or
something
the
sizes
of
the
deals
vary
depending
on
how
much
power
you
have
already
on
the
network.
B
How
many
cycles
you
pledged
somebody
else
will
be
as
small
as
128
bytes
and
some
deals
will
be
two
gigabytes
and
that's
normal
after
the
deal
is
taken
by
your
miner
and
after
it's
sealed,
so
this
eight
hours
or
less,
you
will
receive
a
retrieval
deal
from
a
different
node.
B
It's
run
by
the
same
people,
but
it's
going
to
be
actually
different,
address
different
everything
and
in
order
to
make
sure
that,
for
this
particular
competition
that
you
know
who
to
accept
deals
from
if
you
don't
want
to
like
just
open
yourself
to
the
entire
network
and
see
how
this
how
this
goes,
although
there
is
really
no
harm
in
doing
that
by
default,
your
miner
actually
will
exist
from
everybody.
B
You
can
look
in
the
docs
and
get
the
non-bot
addresses
and
put
them
in
your
config,
so
that
the
only
deals
that
you
will
accept
for
the
duration
of
the
space
race
will
be
from
our
both
and,
as
I
was
saying
earlier,
about
the
little
dl
set
lifecycle
thing
here,
where
it
says
n8.
In
order
for
you
to
get
a
thumbs
up
like
some
other
miners.
Already
on
the
live
dashboard,
you
need
to
run
this
command
at
least
once
lotus
finer
sectors.
B
B
I'm
actually
very
happy
to
report
that
most
of
the
issues
that
have
to
do
with
software
itself
have
been
resolved,
and
I
am
not
going
to
talk
about
things
like
insufficient
funds
or
insufficient
gas,
because,
ideally,
you
should
not
see
this
anymore
and
we
have
been
running
for
like
a
little
bit
over
a
day
now
and
fingers
crossed
so
far.
Everything
is
working
just
beautifully.
B
The
the
dev
team
have
done
an
amazing
job
taking
in
the
lessons
from
the
calibration
network
and
making
sure
that
the
not
very
good
situation
that
we
saw
there
will
not
happen
going
forward.
So
if
you
go
and
click
open
log,
you
will
see
a
bunch
of
things
like
that
here,
I'm
giving
you
a
bunch
of
examples
of
errors
that
essentially
mean
that
you
were
somehow
not
reachable.
B
It
is
very
important
to
make
sure
that
your
miner
has
an
exposed
port,
that
this
port
is
static,
that
basically,
if
you
restart
the
miner,
it
will
not
change
and
it
will
not
match
your
firewall
anymore
and
also,
very
importantly,
and
unfortunately
I
I
I
left
off
this
slide.
You
need
to
put
your
address
your
peer
id
and
your
ip
and
port
onto
the
onto
the
blockchain
and
again
journals
on
slide,
blue
and
we'll
make
sure
to
to
again
reiterate
how
to
do
that.
B
If
you
do
all
of
this
and
your
network
is
not
hostile,
you
should
not
see
any
of
those.
If
you
do,
it
might
be
that
there
is
a
temporary
hiccup
on
the
network
and
that's
okay.
We
kind
of
expect
this
and
the
deals
by
design
currently
do
not
retry.
B
Instead,
you
will
receive
another
deal
later
on
and
if
this
one
succeeds
and
once
after
it
succeed,
it's
great
don't
forget
that
this
particular
success
rate
for
you
is
set
at
80,
which
means
that
you
can
stay
offline
for
essentially
four
days
straight
and
you
will
still
be
considered
a
good
participant,
of
course,
in
the
life
network.
B
We
want
to
bring
those
numbers
way
higher
than
that,
because
80
percent
uptime
for
your
storage
is
not
that
great
and
the
space
race
is
a
great
opportunity
for
us
to
really
see
how
many
of
those
we
get,
what
kind
of
network
conditions
we
get
into,
and
so
on,
and
course
so
again.
Thank
you
very
much
for
participating
in
this
in
this,
like
dress
rehearsal
for
our
mainnet
launch
so
and
the
last
slide
here.
B
I
would
like
to
hand
over
to
my
colleague
ian,
who
will
talk
just
a
little
bit
about
the
registration
which
has
to
deal
has
to
do
with
this
unclaimed
minor
label,
and
this
is
what
is
this?
What
ian
is
going
to
talk
to
you
about
ian.
C
Hey
thanks,
reba,
really
appreciate
it,
yeah,
so
registration.
So
thank
you
so
much,
first
of
all
to
all
the
miners
who
have
taken
time
to
register
on
the
dashboard
so
far,
we've
had
at
least
50.,
but
if
you
look
at
the
dashboard
you'll
notice
that
there's
still
tons
and
tons
of
unclaimed
miners,
so
what
we're
doing
we
need
miners
to
do
is
register
their
miners
no
later
than
2200
utc
on
monday
august
31st,
it's
this
coming
monday,
that's
exactly
one
week
after
the
competition
began.
C
C
We've
had
a
lot
of
requests
from
folks
about
how
to
combine
different
minors.
First
of
all,
you
have
two
days
after
the
registration
window
closes
to
undertake
these
operations,
so
it's
2200
utc
on
wednesday
september,
2nd
when
it
comes
to
combining
miners.
What
do
I
mean
so
in
the
live
filecoin
network
once
mainnet
launches?
C
It
won't
matter
very
much
whether
you're
running
you
know
lots
of
small
minors
in
your
operation
or
you
have
a
more
complex,
multi-machine
setup
that
all
sort
of
feeds
into
one
minor
id
it
does
matter
for
the
space
race.
So
what
we're
doing
is
allowing
sort
of
the
administrative
team
on
the
back
end
to
manually
combine
your
multiple
small
minors
into
one
large
minor
just
for
purposes
of
the
competition.
C
Here's
the
steps
to
take
that,
as
you
register
your
miner,
make
sure
that
you
use
the
same
email
address
to
register
each
one,
send
an
email
to
mining
at
filecoin.io
list
out
all
of
your
minor
ids
and
tell
us
what
you
want.
The
display
name
to
be
for
the
final
combined
miner
now
bear
in
mind.
We've
heard
some
miners
talking
about
sort
of
lots
of
miners
getting
together
to
create
a
combination
or
things
like
that.
That
is
not
what
this
tool
is
intended
to
do.
C
So
this
is,
if
you,
as
an
individual
miner
or
an
individual
organization,
have
many
machines
that
you're
running
and
just
want
to
combine
them
administratively
for
purposes
of
the
contest.
The
point
of
this
is
not
for
lots
and
lots
of
different
teams
to
get
together
and
pool
their
resources,
and
if
we
identify
that
you're
doing
that,
then
it's
possible
you'll
be
penalized
or
disqualified.
C
B
Awesome
thank
you
ian,
and
I
have
like
one
last
point
of
general
advice,
which
is
basically
informed
by
the
many
questions
that
we
got
from
miners
that
were
testing
the
network
in
the
run-up
to
to
space-less
race.
We
totally
expect
that
all
these
hundred
percents
will
go
down
throughout
the
space
race,
so
we're
looking
at
99,
98
and
so
on,
because
there
will
be
times
when
the
network
is
congested.
B
There
will
be
times
when
an
emergency
upgrade
needs
to
be
made,
and
during
this
time
deals
will
not
go
through
correctly
and
the
bot
will
say
like
well.
I
couldn't
heal
right
now,
so
we
will
try
again.
So
the
point
of
the
space
race
is
not
to
get
a
hundred
percent
all
across
the
board
and
everything,
although
that's
great
but
the
deal.
Retrieval
and
storage
is
a
pass
fail
scenario,
so
you
will
not
be
rewarded
more
if
you
do
92
percent
over
91..
B
I
really
want
to
stress
that
so,
as
I
say
here
do
not
worry
about
it
and
another
thing
that
is
a
little
bit
jarring
to
pretty
much
everybody
who
hasn't
dealt
with
file
coin.
Before
is
that
your
miner
will
be
operating,
it
will
be
like
doing
its
think
happily
and
then
all
of
a
sudden
something
will
happen
on
the
network
and
your
lotus
miner
info
will
show
that
your
power
is
gone
and
the
dashboard
will
get
you
from
you
know.
B
However
many
terabytes
you
have
down
to
like
half
of
that,
or
even
a
quarter
of
that
this
is
on
by
design
behavior
of
the
network,
that
anytime,
that
one
of
your
sectors
does
not
call
back
home
so
to
speak
at
the
predefined
time.
B
You
will
get
penalized
by
losing
the
power
of
the
sector
for
24
hours.
If
on
the
next
24-hour
cycle,
this
sector
phones
back
like
yes,
I'm
good,
then
your
power
will
be
restored
and
essentially
the
only
penalty
you
pay
is
a
small
number
of
falcon
and
the
time
that
you
spent
without
this
power
within
you
know
within
your
your
miner
and
that's
pretty
much
it
that's
all.
I
have
for
my
short
presentation
again.
B
A
Thanks
so
much
reba
that
was
really
interesting
and
I
think
super
valuable
for
for
a
lot
of
us
here.
If
it's
okay,
we're
going
to
move
over
to
q
a
so
if
you
have
any
questions
for
reba
or
ian
or
the
rest
of
the
team
we're
here
to
help.
Obviously,
we've
just
kicked
off
space
race.
Naturally
there'll
be
things
that
you're
facing
and
there's
some
troubleshooting
to
be
done.
So
these
are
the
people
to
ask
questions
to
this
is
your
moment.
A
So
if
you
look
at
the
bottom
of
the
screen,
you'll
see
a
q
a
panel.
If
you
click
on
that,
then
you
can
ask
whatever
questions
you'd
like
reba.
If
we
start
at
the
top
with
yourself,
neo
has
asked
he
has
got
an
error.
Submitting
sector
proof
failed
exit
equals
seven.
Do
you
know
how
to
fix
it?.
B
Yes
neo,
so,
unfortunately,
this
is
insufficient
information
to
go
by.
We
will
need
a
little
bit
more
of
your
logs
and
in
general,
whenever
you
folks,
ask
us
a
question
on
slack,
I
very
much
recommend
that
you
prepare
a
little
bit
like
larger
package,
so
to
speak
for
us
to
look
through.
This
includes
your
lotus
version.
This
includes
about,
let's
say,
20
lines
of
relevant
log.
B
A
Cool
thanks,
reba
neil.
I
know
that
you're
very
active
on
the
in
in
the
slack.
So,
if
you
can,
you
can
follow
up
there
and
reba
is
in
there
and
is
super
active.
So
if
you
want
to
follow
up
with
your
logs,
we'll
be
happy
to
kind
of
look
into
that
for
you
and
help
you
further
along
okay.
Next
question:
reba
stuart
berman
asks.
Is
there
a
way
to
recover
from
sector
faults?
A
For
instance,
I
rebooted
when
a
completed
sector
was
being
transferred
from
ceiling
path
to
long
storms,
storage
path,
sealed
the
file
was
the
wrong
size,
but
I
was
able
to
manually
copy
the
sealed
sector.
However,
the
logs
show
the
error
and
I
would
like
to
get
it
recovered
and
prove
it
and
stewart
has
provided
some
error
logs.
Can
you
give
us
some
insight
into
what
his
issue
might
be?.
B
Yeah,
so
actually
this
particular
error.
This
is
new.
I
haven't.
I
haven't
seen
it
before.
I
found
this
wrong
side
this.
This
is
really
weird.
My
first
suggestion
would
be
to
definitely
file
a
bug
about
this
and
I'm
saying
file
a
bug
without
looking
at
the
previous
issues,
because
I
haven't
seen
this
issue
before,
but
the
general
workflow
is,
of
course
you
go
to
lotus
issues
on
github.
You
look
for
your
particular
error.
B
Looking
for
something
like
sector
files,
wrong
sites
would
be
excellent
because
it
will
show
up
right
away
if
it's
there
and
if
you
cannot
find
anything,
please
open
an
issue
with
with
exactly
this
type
of
information
in
in
there.
So
as
far
as
your
actual
question,
like
you
were
able
to
manually
copy
it,
and
you
would
like
to
recover
and
prove
it,
I
believe
you
will
not
be
able
to
no
you.
You
should
be
able
to
eventually
prove
it,
but
this
imp.
B
This
requires
that
the
sector
has
not
been
corrupted
in
some
way
and
file.
It's
wrong
size
to
me
implies
that
it
got
truncated
or
something
like
that
so
yeah.
I
unfortunately
don't
have
a
better
answer
here
and
I
think
we
will
have
to
take
this
on
slack
and
debug
a
little
bit
more.
A
Okay,
great
reba
and-
and
that's
another
really
important
point
to
make
like
if
you're,
if
you're,
finding
issues
like
please
report
bugs
and
file
bugs
with
us,
like
that's
a
big
part
of
the
incentivized
test
net
like
we
really
want
to
you
know:
troubleshoot
these
things
we're
lucky,
obviously
to
have
reba
live
in
person.
A
Doing
this
debugging
live
is
not
an
easy
thing,
but
you
know
it:
it's
brilliant
that
you're
bringing
these
things
up
and
we're
really
kind
of
motivated
to
dig
into
them
with
you
and
kind
of
troubleshoot
here,
okay,
next
question
from
stuart
again,
how
do
I
force?
Oh
sorry,
I've
just
moved
slightly.
How
do
I
force
a
weight
deal
sector
to
start
stealing?
Is
this
correct
to
push
weight
deals
in
sector
three
reba.
B
Yes,
so
you,
okay,
let
me
talk
a
little
bit
about
the
the
terminology
here
so
by
default,
when
your
lotus
mind
is
configured,
it
is
configured
to
wait
for
one
hour
from
allocating
the
sector
and
putting
the
first
deal
that
received
into
it
until
it
actually
starts
seeing
the
sector.
The
reason
this
is
in
place
is
that
in
the
real
networking
in
the
main
net
you
for
any
deal
that
you
receive
that
contains
what
we
call
verified
data,
which
is
basically
data.
That
is
real.
B
You
will
get
10
times
the
power
from
the
part
of
the
sector
that
is
occupied
by
this
deal.
So
it's
very
interesting
to
you
to
have
as
much
of
a
32
gigabyte
sector
be
filled
with
deals
of
various
sizes
that
are
actual
real
data.
So
this
is
why
your
miner
will
wait
when
this
is
the
first
deal
in
this
patient
with
more
deals
with
comedians
will
be
able
to
put
them
in
the
same
sector
and
only
then
it
will
go
like
okay.
B
Now
this
sector
is
ready,
I'm
going
to
start
seeing
it
now
right
now
in
the
space
race.
We
we
we're,
obviously
bootstrapping.
So
we
do
not
have
nearly
enough
people
trying
to
make
deals
and,
as
I
actually
said
in
the
presentation
earlier,
auto
miners
choose
to
filter.
Could
they
accept
deals
from
to
only
our
bots
and
that's
okay
for
the
space
race?
B
This
will
not
be
okay
in
the
main
that,
because,
like
the
entire
point
for
the
falcon
networks
for
you
to
store
other
people's
data,
but
so
in
the
in
stewart's
case,
there
is
a
deal
that
came
in
and
is
blocking
his
sector
now
from
starting
to
seal,
because
it
is
waiting
for
the
originally
configured
hour.
So
stewart
you
don't
need
to
update
state
or
anything
like
this,
because
this
is
actually
a
very
destructive
operation.
All
you
need
to
do
is
say,
lotus,
minor
sectors,
cl
and
number.
A
Awesome,
thank
you
so
much
reba
stewart.
I
really
hope
that
helps
you
and
answers
your
question.
Okay.
Next
question
from
emily:
what
causes
a
minor
to
be
slashed.
B
Right,
so
what
one
is
a
minor
to
be
slashed
and
slashing
has
actually
two
two
meanings
in
our
terminology:
we're
still
we're
still
figuring
out,
which
ones
which
ones
which,
but
the
way
this
works.
In
short,
is,
is
this
way
you
prepare
a
sector
either
you
packed
it
with
deals,
or
it
is
just
full
of
zeros
and
you
put
it
on
the
network,
and
then
there
is
an
a
process
that
happens.
Every
24
hours
called
proof
of
space
time.
B
So
the
network
expects
the
network,
as
this
entire
consensus,
ironic
machinery
that
is,
observing
in
a
distributed
fashion,
what
happens
throughout
the
entire
landscape?
It
expects
you,
as
this
particular
miner
id
for
this
particular
sector
number,
and
this
particular
30
minute
window,
to
submit.
B
A
computation
that
you
cannot
compute
in
any
other
way
without
having
the
entire
sector
ready
and
without
expanding
some
some
competing
resources,
be
it
on
a
gpu
or
on
a
cpu,
but
it
takes
like
in
the
best
of
times
it
takes
about.
B
I
believe
I've
seen
people
do
it
in
six
miles
and
if
you're,
just
using
like
a
mid-range
cpu
with
with
the
correct
instructions
and
stuff
it
might
take
up
to
like
22
minutes
or
something
like
that,
and
what
you
do
then,
is
you
submit
this
into
the
network
and
the
letter
goes
like
awesome.
This
person
still
has
the
sector
still
has
the
data
I'm
happy
about
it?
They
keep
their
power.
B
B
So
this
is
not
just
the
32
gigabytes
of
sector
itself,
but
remember
you
could
have
deals
in
it
which
multiplies
power
so
you're
going
to
lose
everything
that
is
given
to
you
by
the
sector
and
this
penalty
will
persist
for
you
for
24
hours
until
the
next
event
for
this
sector
to
be
true
and
rolls
around
and
a
minor
to
be
slashed
entirely
means
that
you
missed
so
many
sectors
and
you
incurred
so
many
penalties
that
your
collateral
that
you
put
together
at
the
time
of
sealing,
which
is
basically
a
small
amount
of
alcohol
that
incentivize
you
to
not
just
seal
the
sector,
mine
some
blocks
and
then
just
throw
the
entire
miner
away,
because
then
you're
going
to
throw
this
far
coin.
B
If
you
get
slashed
for
a
particular
sector,
you
lose
a
little
bit
of
this
collateral
and
if
you,
if,
if
you
do
this
for
enough
sectors
over
enough
time,
you're
basically
going
to
get
in
well
in
that
you
are
going
to
get
a
negative
balance
in
your
on
your
mind,
then
your
miner
is
going
to
be
drawn
out
of
network
entirely
and
it's
as
if
you
never
existed,
and
the
particular
sets
of
collaterals
and
the
entire
timing
and
the
the
frequency
of
retries
and
so
on
and
so
forth,
is
all
tuned
in
such
a
way
that
to
make
it
better
to
make
it
best
economically.
A
Awesome,
that's
a
really
really
great
explanation
reba,
so
I
think
our
bottom
line
is
don't
get
slashed,
make
sure
you're
starting
the
data,
make
sure
that
you're
ready
for
the
bot
to
check
on
it.
Okay,
brilliant.
The
next
question
also
comes
from
emily
she's
asking.
Do
I
need
to
register
to
be
eligible
for
rewards,
so
ian
covered
this
in
the
presentation,
but
I'd
invite
ian
or
or
reba
to
just
reinforce,
and
I
know
that
we've
had
a
couple
of
people
join
us
since
we
started.
C
Yeah
easy
question:
the
answer
is,
yes,
you
have
to
register
your
miner
to
be
eligible
for
rewards
like
we're
very
serious
about
this.
So
if
you
haven't
registered
by
monday
the
31st
of
august
at
2200
utc,
you
might
all
the
work
you've
done
to
go
into
space.
Race
might
be
for
nothing
because
you
won't
be
eligible
for
rewards.
So
please
go
to
spacerace.filecoin.io
click.
The
register
button
in
the
top
right
corner
take
the
five
minutes
that
it
requires
to
register
your
minor
you'll,
be
very
glad
you
did
and
we
will
be
as
well.
A
Thanks
so
much
ian
yeah
and
we
will
post
the
links
here
in
the
chat
so
that
you
can
get
there.
It
just
takes
a
few
minutes
and
obviously
the
the
rewards
are
are
awesome
and
really
worthwhile.
Okay,
brilliant
next
question
from
stuart
reba.
I
am
on
a
single
machine
with
all
tasks.
How
do
I
prevent
unsealed
sectors
from
remaining
in
my
ceiling
directory.
B
B
B
I'm
I'm
starting
to
answer
this
because
I
believe
it's
not
supposed
to
happen,
so
I
think
you
hit
another
another
bug
here
of
of
sorts
and
yeah.
B
I
I
believe
we'll
we'll
need
to
take
this
into
into
strike
again,
because,
quite
honestly,
I'm
not
I'm
not
sure
what
what
desk
is
here.
So
sorry
about
that.
A
No
that's
great
and
like
stewart
like
we
said,
please
like
come
into
slack.
We
can
deep
dive
on
that
on
those
issues,
and
this
is
why
we're
doing
space
race
to
kind
of
like
surface
some
of
these
problems
and
issues
so
excited
to
kind
of
like
learn
more
about
what
you're
facing
and
how
we
can
help
kind
of
and
and
just
bring
some
more
of
the
team
into
that.
Okay.
A
Next
question:
will
the
calibration
network
still
be
running
for
tests
as
the
space
race
is
going
on,
and
that
comes
from
nelson.
B
Yes,
hi
nelson,
so
the
calibration
network
is
actually
done,
so
you
might
be
wondering
if
you're
running
and
and
not
like
well,
it
runs
just
fine
for
me,
and
this
is
the
beauty
of
the
centralized
systems
as
long
as
sufficient
amount
of
people
keep
running
nodes
with
the
calibration
network
setup
and
config
consoles.
Of
course
it
will
continue
running
indefinitely.
B
However,
the
infrastructure
behind
it,
the
dashboards,
the
bootstrapper
nodes,
the
basically
everything
that
makes
this
the
faucet
everything
that
makes
this
network
thick
is
already
gone.
Now
we
will
continue
working
on
what
is
now
test
net,
slash
space
race,
slash.
You
know
our
master
branch
by
further
fine
tuning
lotus
to
make
sure
that
we
are
as
ready
for
main
main
head
launcher
possible.
B
Now,
don't
be
alarmed
we're
not
going
to
like
reset
the
network
anymore
or
do
anything
of
this
of
this
you
know
sort
that
is
really
disruptive,
because
at
this
point
you
know
people
are
really
racing,
but
at
the
same
time
we
need
to
practice
as
much
as
we
can
how
to
change
the
tires
on
the
moving
vehicle,
because
we
will
be
doing
that
with
mainnet
as
well.
Going
forward-
and
this
is
the
time
for
us
to
do-
that-.
A
Awesome,
thank
you
so
much
reba
next
question
is
from
hermann
felker.
Can
we
get
some
revamped
documentation
soon?
It's
a
little
bit
confusing
as
it
stands
now.
B
Yes,
herman,
I
definitely
feel
your
pain.
Actually,
this
is
how
I
got
involved
in
to
falcon
myself.
I
was
originally
working
on
a
different
team,
but
there
were
holes
to
be
filled
and
yeah
here
we
are.
There
is
a
very
con,
concerted
effort
going
on
right
now
to
fix
the
documentation
to
unify
the
documentation,
to
get
into
basically
a
coherent
story.
B
That
is,
you
know,
being
that
that
is
accessible
to
both
existing
users
and
newbies
like,
for
example,
some
of
the
things
I
believe
that
I
had
on
my
slides
today
are
not
reflected
in
parts
of
the
documentation.
B
So
yes,
this
definitely
pinpoint-
and
there
is
a
number
of
people
working
on
this
right
now
as
quickly
as
they
can
and
hopefully
now
that
we
launch
the
network,
there
will
be
a
little
bit
of
you
know,
breathing
room
for
us
to
be
able
to
catch
up
on
on
these
tasks.
But,
yes,
this
is
within
our
top
five
things
that
we
need
to
fix
and
take
care
of.
A
Yeah
absolutely
aruba,
I
think,
from
an
ecosystem
perspective
and
like
I'm
on
that
side
of
filecoin
and
it's
something
that
we're
working
on
really
really
hard
and
it's
great
for
us
to
get
feedback
feedback
like
that
and
and
also
we're
seeing
the
community
sharing
like
tips
and
tricks
and
other
things.
A
So
if
you
have
suggestions,
please,
like
let
us
know
in
github,
if
there
are
things
that
we
can,
you
know
that
would
be
better
represented
through
documentation
or
or
graphs
or
like
just
we're
really
excited
to
hear
from
you.
What
will
work
best
and
and,
like
reba
said
we're
working
really
really
hard
to
unify
documentation.
A
You
know
and
make
sure
that
it's
a
great
experience
for
people
who
are
just
joining
us,
but
also
those
who
have
been
with
us
for
a
while
as
well.
Okay,
next
question
from
jackie
tan
reba:
how
do
you
think
what
do
you
think
about
the
ddos
which
miners
have
suffered
that
will
affect
their
deal
success
rate
and
might
lose
qualification
to
win
the
rewards
ian
or
reba
happy
for
you
to
take
this
question?
Yeah.
B
I
I
actually
would
like
to
take
ian
to
take
this
question
and
he
has
already
attended
his
camera.
Yes,.
C
Yep
perfect
one
quick
thing
for
the
previous
question
by
the
way
on
the
top
of
the
documentation.
There's
this
file
coin
user.
I
think
his
username
is
joe
nada
or
something
like
that
who
made
this
sort
of
like
pirate
guide
to
file
coin
mining,
to
try
to
shore
up
some
of
the
perceived
shortcomings
with
the
doc
so
check
it
out
on
slack,
he
posted
it.
It's
really
good!
C
So
thanks
to
him
for
making
that,
but
like
it's
an
open
source
project,
so
you
know
if,
if
the
documentation
needs
improvement,
please
help
us
help
us
with
suggestions,
help
us
with
things
like
that,
it's
great
okay,
so
the
definition
of
a
mining
operation.
This
is
actually
included
in
the
rules
for
the
space
race,
which
you
can
click
into
on
the
dashboard.
C
But
the
definition
of
a
mining
operation
is
the
physical
location
of
the
storage
and
ceiling
hardware.
So
it
is
acceptable
to
rent
cloud
surfaces
in
europe
or
in
anywhere
else.
That's
fine
might
not
be
cost
effective,
but
it's
within
the
rules.
However,
all
the
ceiling
and
all
the
storage
hardware
has
to
be
within
the
reason.
It's
not
acceptable
to
do
something
like
running
a
proxy
or
something
of
that
nature,
but
as
long
as
the
hardware
you're
using
is
in
that
location,
it's
fine.
C
Yes,
you
can
so
as
long
as
again
as
long
as
the
physical
ceiling
and
storage
hardware
is
based
in
the
location
that
you're
claiming
that's
totally
fine
and
within
the
rules.
A
B
Yeah
are
actually
a
number
of
questions
like
this
in
the
chat
I'll
try
to
answer
like
all
of
them
in
one
go,
so
this
is
getting
into
the
area
of
distributed
networks
where
we
are
kind
of
committing
not
to
make
breaking
change
of
the
protocol.
So
if
you
have
an
older
version,
it
is
supposed
to
kind
of
keep
keep
operating
and
will
eventually
just
drift
off
and
not
be
able
to
either
sink
the
chain
or
not
be
able
to
post
correctly
and
so
on
and
so
forth.
B
Now
this
is
in
theory,
practice.
We
are
still
kind
in
our
late
beta
of
lotus
and
a
number
of
bugs
are
being
discovered
and
fixed
as
we
go.
B
As
I
said,
multiple
that
already
we
are
committing
to
not
break
your
setup
anymore,
so
that
your
power,
your
standing,
your
your
sectors,
will
all
remain
there
for
the
duration
of
the
space
race,
but
we
also
cannot
guarantee
that
your
original
node
will
be
able
to
communicate
with
the
rest
of
the
network.
Without
you
know,
without
without
problems
going
forward.
For
instance,
there
are
a
number
of
fixes
with
the
json
rpc
that
are
going
in
into
into
the
next
version,
actually
zero.
Five.
B
Four
that
I
would
recommend
anyone
to
get
just
because
there
is
a
risk
that
your
minor,
just
by
sitting
on
its
own,
will
disconnect
from
from
the
rest
of
your
notes,
and
you
will
just
start
losing
power
without
without
without
anything
else,
going
wrong.
There
are
some
fixes
for
how
the
selection
of
messages
that
go
to
the
network
is
evaluated,
so
the
risk
is
reduced,
that
some
of
your
post
messages
proof
of
space
time
that
I
was
explaining
about
a
little
bit
earlier
if
they
don't
make
it
on
the
network.
B
Even
if
you
complete
everything
correctly,
you
had
the
data.
You
did
everything
on
time
if
they
don't
actually
make
it
to
the
network
so
that
other
nodes
go
like
aha,
this
person
is
indeed
saying
they
have
what
they,
what
they're
supposed
to
have,
then
you're
still
going
to
get
slashed,
you're
going
to
be
penalized
and
so
on
so
forth.
B
So
the
general
rule
is
this
for
the
space
race,
unless
you
have
very
very
good
reasons
not
to
upgrade
like,
for
example,
you
are
running
a
custom,
build
of
sorts
and
so
on
and
so
forth,
which
I'll
get
into.
B
There
is
another
question
about
this,
but
in
general,
if
you,
unless
you
have
a
very
good
reason,
not
to
upgrade,
I
would
recommend
upgrading
at
least
every
like
two
three
days
and
for
the
remainder
of
the
week
like
check
daily
for,
for
you
know,
for
what
versions
are
available
and
at
the
end
of
the
day,
all
of
this
open
source
like
we're
not
going
to
recompile
binary
or
anything
like
this
and
his
own
github.
You
can
actually
read
the
list
of
commits
and
decide
on
your
own.
Like
do
I
really
want
that?
B
Do
I
want
to
take
that
chance?
Do
I
care
about
this?
Do
I
care
about
that?
Our
team
is
exercise,
excellent,
commit
hygiene.
So
all
the
commitments
just
are
actually
really
nice
to
read.
They
all
passed
their
tests,
so
yeah
be
informed.
That's
all
I
can
say.
A
That's
great
ruba,
so
that's
yeah!
That's
that's.
Awesome.
Update,
update
update
okay
next
question:
from
brian
will,
storage
miners
be
able
to
purchase
the
pre-filled
hard
drives
from
falcon
discovery
store
after
mainnet
is
launched.
B
C
Go
ahead,
I
do
so.
The
first
round
of
filecoin
discover
hard
drive.
Sales
is
over,
there
may
be
more
available
after
launch,
but
it's
really
important
to
note
that,
like
discover
is
going
to
be
only
one.
Very
tiny
source
of
verified
client
deals
after
mainnet
launch,
so
focusing
or
fixating
on
those
drives
is
probably
not
a
great
strategy.
We
don't
have
a
specific
timeline
for
when
there
will
be
more
available
and
like
honestly,
there
are
not
very
many
out
there
in
the
world
even
now.
C
So
I'd
focus
on
you
know
making
your
minor
well
set
up
to
re,
to
receive
deals
from
verified
clients
after
launch,
by
focusing
on
things
like
storage
quality
and
retrieval
quality,
because
that's
gonna
be
much
more
successful
for
you
than
thinking
about
falcon
discovery.
Hard
drives.
B
Yeah
that
that
that
pretty
much
answers
the
question
yes
again
like
the
the
drives,
there
has
been
a
lot
of
conversation
that
I
heard
on
not
only
on
slack
that
the
falcon
discover
drives
are
a
golden
ticket
they're.
Not.
There
are
many
many
ways
to
get
this
data.
A
That's
great,
thank
you
so
much
ian
and
reba.
That's
that's
really
great
to
kind
of
clarify
a
question
from
sanat.
Is
there
a
reason?
Oh
sorry,
my
goodness,
this
is
just
move
for
me.
Give
me
a
second:
is
there
a
reason
why
some
miners
could
not
just
pre-compute
and
store
all
the
sector
window
proofs
when
sealing
a
sector,
so
they
could
send
them
immediately
when
slash
if
asked,
instead
of
having
to
compute
them
on
the
spot.
B
Yes,
thank
you
sanat.
This
is
actually
a
very
a
very
good
question
and
the
answer
to
this
actually
consists
of
multiple
parts,
because
there
are
different
tradeoffs
to
be
made
like
how,
if
you,
if
you
look
back
at
my
presentation
unless
you're
like
super
super
tuned,
it
takes
good
eight
hours
to
see
the
sector
and
the
best
best
times
that
we
have,
since
so
far,
have
been
like
in
the
neighborhood
like
four
and
a
half
hours.
B
So
we
need
to
balance
this
amount
of
complexity
with
exactly
the
attack
that
you
just
described.
So
one
of
the
ways
we
do
this
is
we
rely
on
multiple
sources
of
randomness,
so
it's
the
block
height
itself,
the
type
of
stuff
that
is
already
on
the
chain.
B
We
have
an
entire
separate
project
within
the
protocol,
labs
and
and
and
other
companies
called
called
dirent
run
by
a
group
called
the
league
of
entropy,
which
is
actually
a
distributed
network
just
like
filecoin
or
ethereum,
but
that
only
focuses
on
delivering
randomness
to
parties
in
a
synchronized
fashion.
So,
in
short,
the
reason
you
cannot
pre-compute
and
store
the
window
posts
is
because
the
particular
challenge
that
will
be
issued
for
your.
B
A
Thanks
reba,
I
think
we're
coming
up
to
our
last
couple
of
questions
here.
So
thank
you
so
much
for
staying
with
us
and
and
answering
all
of
these
questions
live.
I
I
think
I
speak
for
everyone
on
here.
It's
really
really
useful
and
and
helpful
to
kind
of
dig
into
these
details,
especially
in
the
early
days
of
space
race.
So
next
question
from,
let
me
see
so
many
questions
to
be
sure.
Thank
you
so
much
to
everyone,
who's
participating
as
well.
A
This
is
really
really
great
and
useful
for
the
team
to
kind
of
dig
into
things.
Next
question
I
think,
bear
with
me
a
second
let's
see
here.
B
Yes,
this
is
an
excellent
question,
something
that
we
unfortunately
have
a
little
bit
of
like
it's
under
documented
present.
So
there
are
two
things
that
we
are
working
on
number
one.
We
are
working
on
a
document
to
describe
when
it
is
safe
to
shut
down
both
a
minor
and
a
node
and
a
worker,
because,
right
now
it's
not
clear
exactly
when
you
can
do
this.
There
are
actual
windows
opportunity
that
once
you
compile
everything
and
you
basically
just
need
a
shutdown
and
turn
back
up.
B
There
are
sufficient
amount
of
windows
where
you
can
do
that,
but
it's
not
very
clear
where
exactly
you
can
do
that.
So
this
one
thing
that
we
are
definitely
going
to
document
in
the
next
week
or
so
I
hope
and
another
thing
that
we're
working
on
is
the
ability
to
actually
post
workers,
so
they
will
kind
of
freeze
doing
what
they're
doing
and
then
fall
off
of
one
after
another
after
they
have
done
their
job.
B
So
essentially,
what
you'll
be
able
to
do
is
you'll
be
able
to
upgrade
your
lotus
node.
First
then,
upgrade
your
miner
by
freezing
all
the
workers
telling
them
like
hold
up
and
the
miner
will
go
up
and
down,
and
then
they'll
reconnect
with
again
and
then
to
the
workers
you
can
tell
like
when
you're
done,
with
whatever
you're
doing
right
now
do
not
do
not
do
anything
else
and
just
shut
yourself
down,
and
then
you
will
have
to
start
a
different
one
in
this
place.
B
We
did
not
get
a
chance
to
fully
test
this,
yet.
This
is
why
it
didn't
ship
with
space
ace
on
day
one,
but
there
is
definitely
work
being
done
to
make
this
possible.
A
Great
thanks,
reba
on
to
the
next
question
from
ian
bakewell:
what's
the
largest
sector
size,
you
can
seal
slash,
assign
if
you
have
a
lot
of
storage,
eight
plus
hdd
in
a
system.
For
example,
could
you
speed
up
bringing
storage
online
by
increasing
sector
size
dramatically.
B
Right
so
for
this
particular
network
and
for
main
net
at
first,
we
really
don't
know
what
next
year
will
bring
for
now.
All
we
can
do
is
32
gigabyte
sectors
and
64
gigabyte
sectors.
B
The
time
to
prove
them
is
identical,
so
you
would
win
more
from
having
64
gigabyte
sectors
as
you
need
to
ongoingly
produce
the
post
messages
to
basically
say
like
yes,
I
have
this
data.
However,
the
actual
ceiling
time
is
double
the
amount
of
time
it
takes
for
the
32
gigabyte
sector,
so
the
onboarding
of
this
data
will
be
about
the
same
amount
of
time.
B
There
are
parts
of
the
64
gigabyte
sector
process
that
are
a
little
bit
faster
per
gigabyte,
but
you
need
more
memory
which
is
difficult
to
scale
like
for
64
gigabytes.
So
I'm
sorry
for
32k
bisectors.
You
already
need
128
gigs
of
memory
for
64
gigabyte
sectors.
You
need
something
in
the
196
range.
B
It
actually
hasn't
been
very,
very
well
tested
right
now,
because
it
just
takes
longer
and
pretty
much
everybody's
kind
of
focusing
on
32
gig
sectors,
but
you're
more
than
welcome
to
play
with
64k
by
sectors,
and
let
us
know
how
it
works.
A
Awesome
we
love
that
a
bit
of
experimentation
goes
a
long
way
in
these
things.
Right.
Okay,
I
think
we're
gonna
take
one
last
question
here,
as
reba
said,
I
think
it's
very
early
in
the
morning
for
him
and
he's
been
a
fantastic
presenter,
so
we'll
let
him
go
to
bed.
The
last
question
comes
from
storage
hive
inc.
We
I've
got
a
mixture
of
different
hardware
and
found
the
default.
Scheduler
logic
doesn't
work
well
for
me,
is
it
okay?
If
I
tweak
scheduler
code
for
my
use
case.
B
Right
so
I'm
going
to
say
something
controversial
here:
nobody
likes
the
default
schedule,
logic
right
now,
so
by
all
means
number
one.
Yes,
you
can
tweak
it
as
much
as
you
want,
because
at
the
end
of
the
day,
schedule
is
literally
some
defense
on
your
machine.
Only
it
doesn't
communicate
with
the
with
the
consensus
to
the
network
in
any
way,
so
you
are
free
to
do
anything
you
want.
Moreover,
I
very
much
encourage
you
to
share
whatever
you
discover,
or
you
know
or
come
up
with.
B
There
is
already
a
group
on
slack
that
is
kind
of
sharing
their
way
of
doing
things.
B
I
believe
one
of
our
earlier
questions
were
from
somebody
that
does
this
stuart
has
been
experimenting
with
running
multiple
workers
within
120
128
gigs
of
memory,
which
is
really
cool,
so
yeah
by
all
means
like
join,
join,
something
on
slack
and
share
anything
you
feel
like
you
could
share,
because
if
your
particular
schedule
logic
is
nice
both
in
terms
of
that
it
is
cleaner,
it
performs
better
or
it
utilizes
the
hardware
better
or
all
of
the
above.
B
The
team
will
be
very,
very
happy
to
both
include
into
lotus
and
maintain
it
for
you
going
forward,
so
you
will
never
have
to
like.
You
know
worry
about.
Oh,
I
have
this
custom
branch
now
I
need
to
upgrade
so
I
need
to
rebase
my
changes,
so
I
need
to
make
sure
it
doesn't
breaks
yeah.
You
know
this
is
how
open
source
works.
A
You
gotta
love,
open
source
and,
and
that's
why
we're
all
here
right
now
we
really
encourage
everyone
to
not
just
participate
in
space
race,
but
you
know
filecoin
is
an
open
source
project.
We're
excited
to
see
the
different
bugs
that
you
discover
like
help
us
improve
the
documentation.
A
Thank
you
so
so
much
to
reba,
who
got
up
at
crazy
o'clock
in
the
morning
to
be
with
us
and
answer
all
of
our
questions
live
and,
and
hopefully
help
a
lot
of
folks
as
they
kind
of
continue
on
their
space
race
journey
a
huge
thank
you
to
our
production
team.
I
think
some
of
them
are
up
in
the
middle
of
the
night
as
well
and
a
huge
thank
you
to
ian
also
on
the
filecoin
team,
who
has
helped
answer
some
questions
tonight,
just
a
couple
of
reminders
from
my
side.
A
So
this
is
just
the
start
of
the
space
race
event
schedule.
We've
got
a
lot
more
coming
your
way.
I
will
just
share
this
now.
I
think
here
we
go
so
upcoming
in
space
race.
If
my
computer
will
be
well
behaved
so
tonight
we
had
the
mining
workshop
tomorrow
at
10
p.m.
Utc
we
have
the
opening
ceremony
with
falcone,
founder,
juan
binay
and
a
number
of
members
of
the
falcon
team,
so
we
hope
you
can
join
us
we're
going
to
be
talking
about
the
space
race
program
generally.
A
On
thursday,
we
have
the
minor
show
intel,
there's
a
link
in
the
chat.
Just
now,
if
you'd
like
to
present
at
that
and
talk
about
your
architecture
about
your
hardware,
about
your
team
about
why
you're
excited
to
mine
on
falcoid,
then
we're
really
excited
for
you
to
kind
of
share
your
journey
and
then,
on
friday,
every
friday
we're
going
to
have
a
space
race
roundup,
so
we're
going
to
be
talking
about
who's
top
of
the
leaderboards
globally
and
in
each
continent.
A
We're
going
to
talk
about
the
folks
who
have
been
really
great
to
the
community
by
helping
folks
in
slack
and
reporting
bugs,
etc
so
like
and
keep
that
awesome
stuff
going
and
then
in
week,
two
we've
got
loads
more
stuff
coming
your
way
as
well.
There
is
a
page
for
this.
It's
just
emily
will
link
it
in
the
chat
and
you
can
register
for
all
of
these
events
there
just
a
huge
huge
thank
you
to
join
us
this
evening.
A
We're
excited
to
spend
more
time
with
you
over
the
coming
weeks
and
that's
it
from
us
this
evening.
Thanks
so
much
again
and
see
you
all
soon
take
care
bye,.