►
From YouTube: Kubernetes SIG Scheduling meeting 2018-04-19
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).
B
All
right,
hi
everyone
for
those
who
don't
know
me.
My
name
is
Paris.
I
am
a
chair
of
the
special
interest
group
for
contributor
experience.
I
am
going
to
jam
a
lot
of
information
down
right
now.
What
we're
doing
is
going
around
to
all
of
the
SIG's
and
a
lot
of
the
working
groups
to
which
is
40
at
this
point
and
really
just
trying
to
establish
some
grounds
between
the
things
and
kind
of
the
experience
and
hopefully
engage
you
in
other
ways
that
we
have
into
it.
So
first
things.
B
B
What
a
lot
of
people
are
contributing
to
multiple
repos,
which
could
be
problematic
if
people
are
using
different,
for
instance,
different
commands
for
different
things
or
different,
different
meanings
or
workflows
are
different,
etc.
So
what
we're
doing
is
trying
to
figure
out
to
how
we
can
better
define
these
labels
and
really
keep
things
consistent.
So
in
the
agenda
that
I
have
today,
there
are
several
different
links
in
there.
B
I
don't
necessarily
need
any
feedback
right
now
by
any
means,
but
just
feel
free
to
give
us
feedback
at
any
time
about
this
stuff,
especially
again
how
you're,
using
certain
workflows
and
things
like
BOTS
and
how
you're
doing
issue
triage
one
of
the
things.
So
that
leads
me
into
issue
triage,
which
is
something
we've
also
been
working
on.
As
far
as
guidelines
are
concerned,
we
have
taken
a
stab
at
what
we
think
is
the
best
guidelines,
and
we
wanted
to
see
what
the
thoughts
were
on
applying
these
guidelines
across
all
of
the
communities
github
repositories.
B
So,
first
we're
going
to
be
lazy,
we're
going
to
participate
in
lazy
consensus
with
a
time
box
and
at
least
72
hours
to
the
following
mailing
lists,
with
a
github
issue:
link
label
with
the
subject
of
notice
and
we're
going
to
be
that's
the
contributor
experience
mailing
list.
That's
the
sake
leads
mailing
list
and
that's
also
kubernetes
dev.
B
Community
meeting
speaking
of
charters,
we
have
completed
ours
I'm,
not
too
sure
what
the
status
is
for
scheduling,
but
if
you
need
any
help
with
that,
we're
happy
to
provide
any
help
or
assistance
or
feedback
or
anything
like
that
before
you
have
the
steering
committee
look
at
it
and
then
we've
also
started
to
list
out
what
projects
we're
working
on
as
well.
We
know
this
isn't
a
great
way
for
us
to
project
manage
necessarily,
but
we
just
wanted
to
kind
of
start
getting
the
stuff
in
our
heads
down
on
paper
in
a
more
professional
way.
B
So
we
do
have
a
list
of
our
projects
that
we're
working
on
and
we'd
love
to
hear
feedback
as
to
its.
You
know
what
you're
excited
about,
or
maybe
other
projects
that
you
think
contribute.
Your
experience
should
work
on.
So
we
also
have
a
new
version
of
a
contributor
guide,
so
I'd
love
for
you
to
poke
holes
in
that
and
take
a
look
at
it
does
it
differ
from
the
way
that
people
can
contribute
to
sig
scheduling
and
please
help
us
out
also
with
the
developer
guide.
That
portion
is
coming.
B
That's
going
to
be
a
part
of
the
contributor
guide
and
the
issue
in
the
agenda
is
linked
in
there
feel
free
to
add
your
comments
to
the
issue
and
also
the
contributor
experience
deep
died
of
a
cube.
Conning
Copenhagen
is
actually
going
to
start
attacking
this
issue
and
we're
going
to
do
sort
of
a
live
process
of
feedback
and
try
to
get
a
large
part
of
the
developer
guide
done
during
that
deep
dive,
so
next
section
and
then
we're
actually
almost
done,
which
is
mentoring.
B
We
have
a
ton
of
programs
that
we
spun
out
to
address
many
many
different
use
cases
and
all
of
the
things
that
we've
developed
take
your
time
into
consideration
as
well
as
the
time
of
the
mentor.
Our
excuse
me
mentee,
that's
very
important
to
us
because
we
know
times
the
number
one
barrier
here:
I
mean
if
we
all
had
the
time
we'd
mentor.
You
know
hundreds
of
people
right,
but
no.
We
know
that
we
can't
so
I.
Take
me
taking
care
of
some
of
those
some
of
those
roadblocks.
B
So,
for
instance,
now
we
do
a
monthly
web
series
for
mentors
on
demand.
It's
a
meet.
Our
contributors,
people
ask
questions
from
things
like.
How
did
you
get
involved
in
Fornetti
is
what
sig
do
I
join
all
the
way
to?
Why
is
my
test?
Failing
how
do
I
know
when
my
tests
failing,
etc?
This
is
a
great
thing
for
scheduling
to
be
a
part
of,
and
you
can
advertise
for
your
group
there
etc
and
then
I'm
gonna
jump
down
to
group
mentoring.
B
This
is
a
concept
where
I
think
we
could
really
scale
with,
and
this
is
for
all
contributor
ladders.
We
ran
a
test
cohort
of
current
review
or,
if
she's
me,
current
members
of
the
org
to
get
them
to
your
viewership
and
owner's
files,
and
it
worked
out
really
really
well
and
it's
based
off
of
peer
mentoring
and
we
set
up
different
communication
platforms
and
we
do
a
for
mentees
to
one
mentor
and
senses
a
group.
B
We
now
have
a
name
for
it,
it
didn't
mean
is
the
one
on
one
our.
This
is
the
closest
thing
to
traditional
mentoring
that
we
have
a
lot
of
people
come
to
us
and
say:
can
you
just
pair
me
with
someone
immediately,
I
really
need
some
help,
but
that
is
really
really
time-consuming.
Traditional
mentoring,
you
have
to
set
up
goals
with
people
you
meet
every.
So
often
this
is
taking
a
lot
of
that
out.
This
is
a
one-time,
one-hour
commitment
with
either
someone
that
is
your
peer.
B
B
So
I'm
actually
not
going
to
talk
about
them
right
now,
but
feel
free
to
click
any
of
the
links
in
the
doc
and
then
almost
last
but
not
least,
is
slack
maintenance.
We
are
onboarding
a
ton
of
people
thousands
by
the
month
at
this
point,
so
it's
just
a
quick
note
to
maybe
pin
some
docs
to
your
slack
channels,
like
your
charter,
slack
guidelines,
meeting
notes
anything
that
would
help
people
on
board
and
find
out
a
little
bit
more
about
your
sake
easily.
That's
just
a
tip,
that's
just
a
tip!
That's
not!
B
You
know
a
mandate,
that's
just
from
what
I
see
is
a
slack
admin
that
really
helps
out
with
that
and
then,
of
course,
we
do
have
user
office
hours
as
well,
and
we're
always
looking
for
folks
to
help
out
with
that,
and
just
answer
calls
not
calls.
Oh,
my
gosh,
that's
so
old-school
answer,
questions
that
come
through
from
Stack
Overflow
and
black
and
places
like
that
and
Jorge
Castro
with
hefty
overruns
dad
and
that's
really
it
feed
but,
like
I,
said
feedback
for
us.
How
can
we
help
you?
A
A
D
D
Here
is
the
details
of
the
clutter
and
there
are
32
CPUs
in
the
class,
and
I
will
create
powder
first
to
occupy
all
the
crustal
resources,
and
today,
I
will
create
two
crops.
Now
we
can
see
that
there
are
first
Qdoba
has
has
to
post
and
each
photo
request,
service,
abused
and
the
second
queue
Java
contained
report
and
each
corner
also
request
service
abuse
so
totally
not
a
class.
The
cluster
could
support.
D
D
D
Now,
after
terminating
and
pada,
pada
larger
capacity
will
schedule
the
powder
from
the
two
key
rocks,
and
now
we
can
see
that
two
powder
from
the
first
cure
are
running
and
the
street
job
from
secondary
table
or
pendings
there
in
Pina
Staters.
Even
if
the
cluster
left
some
resource
could
run
some
code.
D
A
A
F
I'm,
sorry,
you
might
have
covered
it
earlier,
but
I
had
a
question.
What
would
happen
if
one
of
the
pods
would
fail
say
well
gang
scheduling
if
one
of
the
pods
failed
to
come
up?
Will
all
the
rest
of
the
pods
will
go
back
to
pending
state
or
God,
but
in
first
case
the
scheduling
itself
won't
happen.
D
C
Yes,
some
more
point:
some
more
input
here
is
relied
to
full
torque
should
draw
with.
There
are
two
kids,
no,
the
first
one
is
something
like
MPI
drop.
If
one
of
task
is
terminated
as
a
whole,
the
whole
job
which
should
be
should
be
Q
and
a
recreate
and
the
other,
the
other
can
over
draw
with
something
like
spark
for
them
or
if
one
office
part
is
killed.
Her
is
the
terminated.
It's
doesn't
need
its.
We
don't
have
to
rerun
the
whole
job.
C
So
by
before
we
will,
we
will
handle
the
first
or
one,
but
in
our
before
the
implementation,
but
we
also
provide
the
interface
to
for
the
user
to
build
their
customized,
your
job
controller,
so
they
can
decide
whether
the
whole
job
ho
travel
is
cured.
Oh
it
just
a
oh,
they
can
just
be
wrong.
One
of
the
one
rerun
to
fail.
The
task.
D
D
A
C
G
G
A
C
For
for
the
starvation
part,
I
think
there
may
be
a
there's,
the
rule
case
and
a
similar
way
to
handle
die
I
listed
some
case
in
the
counseling
design
document.
But
if
this
is
not
finished
yet
yes
for
the
thoughtful
karna
stager,
if
there
we
will
try
to
tread
to
miss
the
high
priority
he'll
drop
yesterday,
at
least
this
is
the
one
case
for
this
one,
but
you
know
there
Fred
resource
of
fragment
is
the
is
the
is
a
really
complex,
the
usable
for
such
kind
of
a
backdrop
so
I
think
we
don't.