►
From YouTube: Geo Scheduling Call 2021-03-09
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
A
A
D
C
It
was
going
to
be
a
slow
last
week,
just
because
I
was
on
a
pipeline
rotation,
but
I'm
getting
to
the
point
where
so
I've
been
going
through.
I've
been
automating
this
for
get
as
I
go
as
well,
which
is
bigger,
okay,
but
I'm
kind
of
getting
to
a
point
where
it's
difficult
to
know.
If
I
need
to
change
my
automation
or
if
it's
because
the
documentation's
wrong.
D
C
Which
is
the
correct
one
to
do,
because
I
say
we're.
I
know
today
as
well,
like
alex
pointed
out
a
while
back
that
you
know
like
doing
the
hub.
Reload
does
cause
downtime
on
each
of
the
things,
but
on
in
the
geo
docs,
we
just
say
run
it
on
all
the
machines
in
the
non
geo
docs.
It
does
say:
do
it
one
at
a
time,
so
it's
whether
the
correct
process
should
be
documented
as
like,
remove
it
from
the
load
balancer,
do
a
reload
put
it
back
in
the
load.
C
A
A
A
E
E
Right
like
if
you're
monitoring
for
the
top
five
percent
or
whatever
right,
like
of
worst
requests
like
that,
could
still
trigger
an
alert
for
someone
which
could
make
them
be
like.
Well,
it
wasn't
really
zero
downtime,
which
may
I
don't
know
I
don't
know
yeah.
I
I
I
tend
to
agree
with
you,
though.
I
don't
think
that
that's
that
big
of
a
deal,
I
think
a
couple
of
people
can
retry
their
requests.
F
To
actually
be
zero
downtime
because
we
yeah
we
may
want
to.
If
not,
we
may
want
to
try
to
yeah
get
some
clarification
on
that
just
to
yeah.
C
Because
I
can't
I
mean
that's
the
next
step
I
can
do
as
well.
Like
I
mean,
I
think
it
will
be
zero
downtime
if
you
manually,
remove
it
from
the
load
balancer
first
or
set
it
into
maintenance
mode.
So
that
no
requests
go
then
do
the
reload
wait
till
it's
back.
So
if
you
do
it
the
full
every
single
step
way,
it
should
be
zero
downtime.
But
I
suppose
from
an
automation
point,
are
we
happy
with
a
you
can
just?
C
Yes,
if
you
go
to
the
wrong
machine,
you
might
see
an
issue
you
reload
and
you're
on
to
the
you're
on
to
the
next
node
in
the
round
robin
essentially
and
you're
fine,
but
as
well,
it's
more
if
it
was
expected,
I
suppose,
as
well
with
the
hub,
because
I
thought
like
then
this
is
my
naive
approach
of
what's
it
in
between
the
hub
and
just
a
restart
like.
Why
do
we
do
the
hub
if
it
just
doesn't
essentially
the
same
as
what
a
restart
would
from
a
user
perspective.
A
G
Yeah
it
does
it
doesn't.
That
was
one
of
the
sacrifices
we
made
switching
to
puma
is
we
can't
do
like
the
live
reloading
that
unicorn
does
so
unicorn?
We
could
do
the
stair
down
time
where
it
kept.
The
old
workers
around
until
the
new
workers
were
loaded
and
ready,
and
I
started
syncing
requests
to
the
new
workers,
but
kumar
does
not
do
that.
C
A
Okay,
sorry
sorry
to
go
a
little
too
deep
in
here,
but
I
we
do
need
to,
I
think,
defines
your
downtime.
Maybe
like
get
some
other
people
outside
of
the
geo
team
like
involved
in
like
what
are
you
know?
What
are
we
trying
to
accomplish
here,
like
yeah,
all
right,
well,
cool,
that's
good
work.
B
Writing:
okay!
Nick,
do
you
know.
F
Yeah
I
haven't
checked
in
on
this.
I
know
there
was
some
activity.
F
Looks
like
yeah
yeah
gabriel
had
added
some
some
updates
a
few
days
ago.
Rebased
it
oh,
it
looks
like
it
just
got
a
database
approval
from
andreas.
So
I
guess
that
needs
to
needs
to
get
passed
along
again.
So.
A
Still
actively
in
review,
that's.
F
Yeah,
my
my
general
understanding
of
petronius
yeah,
I'm
not
sure
exactly
what's
going
on
with
this
issue
either,
but
I
I
know
that
douglas
and
gabriel
and
matt
and
kind
of
talk
through
some
of
these,
these
different,
like
setup
issues
that
they
were
kind
of
running
into
and
I
think,
had
and
just
kind
of
talking
about
what
what's
needed
for
like
what
else
do
we
want
to
show?
You
know
in
terms
of
like
a
reliable
setup
in
order
to
get
us
to
beta,
and
so
douglas
did
tell
me.
F
There
were
a
couple,
other
follow-up
issues
that
they
were
creating
going
to
create
out
of
that,
but
I
that's
sort
of
my
high
level
understanding
of
what's
going
on
and
I'm
not
too
familiar
with
yeah
progress
on
some
of
these
individual
issues.
F
F
Yeah
yeah,
I
think,
yeah.
I
think,
let's
move
on
from
this
I'll
paying
for
an
update
in
the
epic.
I
I
do
know
that
or
yeah,
let's
yeah,
let's
ping,
for
an
update
in
in
the
yeah
in
the
issue
here,
yeah
and
then
and
then
I'll
also
ping
for
something
in
in
the
epic.
If
we
don't
see
see
an
update.
A
H
Hey
so
if
you
scroll
down
just
a
little
bit
mike,
I
broke
down
an
implementation
checklist.
There's
also
a
retro
item
about
this.
If
you
guys
are
curious
on
the
the
13,
I
think
10
retro
we're
trying
to
do
nick
is
suggesting
trying
to
do
like.
I
don't
know
how
you
call
it
nick,
but
like
rolling
retro
so
like
when
you
have
items
to
maybe
put
them
on
there
before
the
actual
milestones
ended,
but
anyway,
this
one
probably
was
inside
kind
of
a
tldr.
H
This
one
was
sized,
probably
incorrectly
by
myself,
and
I
probably
should
have
made
this
sub
epic,
so
this
is
kind
of
my
non-invasive
approach
to
treating
it
as
if
it's
a
subarabic,
and
maybe
we
could.
I
could
still
do
that,
but
I
think
at
this
point,
creating
a
bunch
of
issues
is
backfill
is
probably
not
very
beneficial
but
yeah.
So
this
is
kind
of
where
I'm
tracking,
the
mrs
you
can
see
we're
on
mr4
of
eight.
H
So
this
moving
along
nicely
a
lot
of
the
mrs
aren't
like
terribly
large,
maybe
minus
one
or
two
but
yeah.
So
this
one's
going
along
nicely.
Just
writing
tests
and
kind
of
getting
the
mrs
out
there,
and
hopefully
I
based
on
the
speed
of
this
I'm
gonna
expect
this
is
probably
going
to
be
done
towards
the
end
of
1311.,
so
definitely
before
14.0
but
yeah.
If
you
have
any
questions,
feel
free
to,
let
me
know
but
yeah.
H
The
ui
is
approved
by
sungjoon,
as
it
has
been
for
a
while,
I'm
just
going
through
the
review
cycles
with
mvc
and
large,
mrs
in
general,.
H
Yes,
it's
if
you
just
click
on
that.
First,
mr
mic,
to
the
ui
utilization.
It's
probably
right,
it'll
be
right
at
the
top
in
the
description
yeah
right
there,
geo
nodes
beta
is
the
name
of
the
of
the
of
the
that's
underhood.
So
that's
the
feature.
A
So
I
was
just
chatting
with
ian
this
morning
about
the
lfs
migration,
so
that
one
there
was
some
weirdness
with,
because
the
old
one
is
kind
of
like
displayed
in
the
ui
via
a
different
code
than
the
new
replicators
right.
So
this
I
don't.
H
Know
we're
gonna
yeah,
so
this
change
is
only
for
the
node
status
bars.
Are
you
talking
about,
like
the
actual
replica
views.
A
The
no
no
no
yeah
the
progress
bars,
I
don't
know
there
was
some
weirdness,
so
I
don't
know
we
might
need
to
make
some
kind
of
front-end
related
change,
just
just
kind
of
a
heads
up
that
that.
H
Might
be
coming
awesome
yeah
this
this
ping
me.
I
can
take
a
look
yeah
because
right
now,
like
there's
an
expectation
that
everything
follows
the
same
pattern
since
our
api
response
is
flat
so
preferably
like
as
long
as
we
could
try
to
stick
to
that
the
model.
That's
there,
then
we
should
be
shouldn't
be
too
hard
if
it's
a
big
outlier,
we'll
probably
have
to
have
some
like
you
know,.
A
Yeah
so
yeah,
the
weirdness
that
we're
looking
at
is
like
the
feature
flag
because
we're
going
to
feature
flag
it
so
that
if
you
enable
the
replicator,
then
the
old
stuff
turns
off,
and
if
you
disable
the
replicator,
then
the
the
old
stuff
turns
on.
And
then
you
know
what
I'm
saying.
So
I
don't
know
what
anyways
but
yeah,
we'll
we'll
definitely
be
thanking
you
for
review
on
that
awesome
sounds
good
cool.
A
Okay,
geo
usage,
ping
alex.
E
Hello
yeah,
so
the
secondary
usage
thing
worker
is
in
review.
E
It
went
through
around
with
ian
and
around
with
douglas,
and
I
just
have
a
few
comments
that
I
am
following
up
on
and
I'm
hoping
that
that'll
get
merged
today
fingers
crossed
and
then
the
next
issue
in
this
is
actually
collecting
the
metrics
and
that
is
in
a
merge
request,
marked
as
work
in
progress
that
went
through
some
product
intelligence
review
and
they
had
some
questions
and
then
it's
got
a
review
and
then
I'm
waiting
for
the
team
mccreedy
to
take
a
look
at
it,
maybe
or
someone,
and
then
I
have
one
other
murder
request
that
I'm
also
having
to
view
I
I
split
up
collecting
the
metrics
and
adding
them
to
the
node
status.
E
E
E
Oh,
I
must
have
missed
a
link
I'll
make
sure
that
I
get
the.
I
must
not
have
the
issue
link
in
the
in
the
merge
request
description,
it's
probably
on
the
it's.
If
you
like
scroll
to
the
bottom,
it
probably
got
like
referenced
in
commit
yeah,
so
it
gets
mentioned
and
commits
because.
A
E
A
A
Okay,
the
rest
are
and
ready
for
development
or
open,
improve
the
self-service
geo
replication
framework.
I
saw
some
emails
come
in
today,
but
I
don't
know
as
valerie
you
wanna
talk
about
this.
D
A
Cool
okay,
yeah,
so,
okay,
cool!
Thank
you.
A
D
Yeah,
it's
also
mine,
so
I
started
to
work
on
it
again
today,
so
I
need
to
like
for
for
now
I'm
just
collecting
data
about
these
queries,
how
heavy
they
are
for
database
so
and
so
on.
A
A
G
Door
mike
helped
me
out
a
little
while
ago
we
talked
about
the
current
status,
so
I'm
going
to
move
forward
with
rolling
this
out
as
a
feature
flag
right
now.
That
should
help
us
kind
of
identify
what's
old,
that
needs
to
go
on
what
doesn't
need
to
go
because
we're
running
into
a
lot
of
different
behaviors
based
on
things,
I'm
still
learning.
So
I'm
going
to
move
forward
with
that
and
hopefully
have
more
results
and
more
definitive
answers
on
things
in
the
next
couple
of
days.
G
A
G
F
So
from
understanding
or
yeah
moving,
all
at
once
doing
it
behind
a
feature
flag
and
then
sort
of
toggling
that
and
staging
to
see
see
what
the
impact
might
might
be
exactly:
okay,
cool
that.
A
E
Bummer,
oh,
I
was
just
saying
it's
a
bummer
that
we
can't
migrate
it
a
little
bit
at
a
time,
but
I'm
glad
that
I'm
good
job
figuring
that
out
sorry.
A
No,
I
think
it's
yeah.
I
think
it's
gonna
be
good,
that
we're
gonna
have
the
old
way
be
able
to
still
be
used
if
something
goes
wrong
with
enabling
the
future
flag,
but
yeah
before
I
forget
zach,
as
you
fade
into
the
darkness
there.
B
A
All
right,
so
the
next
epic
here
is
backup
restore
gitlab
instances,
oh
ready
for
development,
okay,
the
next
epic
is
secondary,
mimicry.
I
I
have
a
poc
open
for
this
issue,
I'm
doing
it
in
a
workhorse,
and
there
was
some
discussion
which
I
haven't
yet
gotten
back
to
this
morning.
A
So
I
need
to
look
at
that
and
I
hope
you
know
to
kind
of
like
get
an
idea
of
weight
and
what
we
can
do
to
move
that
forward,
but
I
think
it's
looking
like
it,
you
know,
might
be
pretty
feasible
to
to
to
do
like
in
a
reasonable
time
frame.
So
why
anyways?
I
don't
know,
I
haven't
seen
the
discussion,
the
recent
discussions
yet,
but
we'll.
F
See
super
exciting
and
and
thanks
for
yeah,
even
putting
the
animated
gif
up
there.
I
think
it's
cool
to
see
that
and
the
implications
so
yeah
nice
work.
A
Thanks
yeah,
it's
fun
to
see
as
from
a
geo
perspective,
but
I
I
just
like
it's
it's
funny
to
think
about
from
from
a
non-geo
perspective,
like
somebody
looking
at
that,
it's
like
why
I
just
why
are
you
just
clicking
around
get
loud
like
it
looks
completely
normal
and
I'm
like?
Yes,
it's
normal
there's,
no
errors
all
right!
A
Next!
Epic
here
is
omnibus
improvements
in
review,
alex.
E
Oh,
I
think
wait.
Did
this
merge.
Did
I
forget
about
this?
Sorry,
oh
it's
not
merged.
Yet
all
right.
I
guess
I
probably
have
to
fix
something
on
this.
Sorry
about
that
yeah.
This
is
just
adding.
E
So
we
we
weren't
honoring
postgres
sql
data,
well
postgres
sql
der
for
a
couple
of
operations
we,
but
we
were
for
everything
else,
specifically
promote
database
and
replicate.
E
F
Awesome
thanks
for
looking
into
that,
and
I
think
did
this
kind
of
I
mean.
I
know
it
came
out
with
the
recent
customer
issue
but
looks
like
we
took
it
on
as
part
of
the
bug
triage
process
or
like
kind
of
prioritized
it
as
part
of
that.
E
E
It
was
like
eric
rediagnosed.
This
issue
was
happening.
It
was
like
it's
weird,
it's
not
recognizing
postcryster
and
then
was
like
oh
yeah.
There's
an
issue
for
it.
I
was
like.
Oh
well
sounds
like
it's
probably
more
important
than
a
than
a
four.
So
we
we
bumped
it
up
kind
of
based
on
that,
and
then
it
was
pretty
simple
to
fix.
So
it
seemed
like
the
right
choice
to
just
fix
it.
F
A
Yeah.
Thank
you.
Thank
you
very
much
alex
okay!
Oh
that's
it
for
the
epics.
We
have
some
things
closed
here.
C
F
Yeah,
I
don't
either,
but
I
will
make
a
note
here
to
to
investigate
and
and
create
an
issue
if
there
is
not.
A
A
A
Was
like
an
urgent
bug
fix?
Did
we
did
we
not
talk
about
this?
I
guess
there
was
a
bunch
of
stuff
in
slackchat
yeah.
I
don't
know
if
anybody
knows
more
about
this,
but
there
there
was
like
an
urgent
bug
fix
needed
to
avoid
avoid
downtime
during
upgrades.
So.
F
Yeah
thanks
douglas.
I
think
this
is.
I
think
this
also
had
a
corresponding
production
ish
like
issue
as
well,
not
geo-related,
but
just
like
due
to
our
mixed
deployment
with
canary
fm.
If
I
remember
that
correctly
so
so
yeah
it
not
only
affected
that
yeah
secondaries,
but
it
was
a
wider,
wider
issue.
A
Okay-
next,
oh
this
was
just
moved
so
I'll
just
skip
that
okay
in
review.
This
thing
is
covering
this.
I
don't
know
why.
A
B
A
Oh
yeah,
so
there
was
a.
This
is
a
customer
issue
where
they're
using
object,
storage,
replication,
which
is
in
beta
for
geo
but
they're
using
it
and
direct
direct
download,
isn't
working
and
I
think
they're
on
azure
or
something
object,
storage
and
there
we
actually
fixed
this
for
like
lfs
files
and
other
things
like
that
a
long
time
ago.
A
But
that
fix
we
didn't
put
it
into
the
ssf
like
framework
for
blobs,
because
I
think
that
fix
kind
of
happened
around
the
same
time
as
we
were
creating
the
ssf
so
anyways.
I
have
a
merger
quest
open
to
fix
that
valerie
reviewed
it.
Thank
you
and
I
need
to
address
some
feedback,
and
I
think
it
shouldn't
be
too
much
longer.
A
F
Yeah
thanks
thanks
for
taking
taking
it
on
yeah,
I
mean
yeah,
I
think,
with
bug
triage.
It's
like
you
know,
might
have
ended
up.
You
know
if
it
was
current
enough
and-
and
someone
had
the
context
they
might
have
taken
on,
but
I
think
it's
nice
to
also
say
like
hey
as
part
of
this,
this
process
there
there
is
somebody
who
can
you
know,
sort
of
make
that
decision
and
and
go
ahead
and
and
run
with
it
with
their
with
the
triage
time
they've
set
aside
so
yeah
thanks
for.
A
Yeah
no
problem,
okay
and
then
we
have
something
in
dev
which
douglas
has.
I
do
not
know
where
this
is.
F
Yeah,
I
think,
there's
some
yeah
some
ongoing
discussion
here
with
ashley
yeah.
When
I
was
talking
to
douglas.
I
think
he's
still
he's
still
kind
of
investigating
it
and
and
working
out,
but
I'm
not
sure
exactly
the
the
details
of
where
he's
at.
A
F
A
All
right,
that's
it
for
in
dev
and
beyond.
I
don't
know
nick.
Do
you
want
to
take
over
the
ready
for
development
and
stuff.
F
Share
my
screen
quickly,
so
yeah
fabian
went
through
earlier
today
and,
and
I
think,
removed
some
items
from
the
board,
so
I
believe
most
of
this
should
be
prioritized,
so
we
have
our
ready
for
development
items
in
in
our
various
epics.
I
won't
go
through
those,
although
I
am
just
noticing,
did
we
skip
over
verifying
all
blob
replicators
before
because
yeah
they're,
both
ready
for
development
and
I
see
akriti-
is
assigned
to
those?
Did
you
have
any
any
thing
you
wanted
to
say
about
those
actualty.
F
Cool
well,
thank
you
for,
for
taking
on
this
work.
Let's
see
we'll
just
go
down.
Yes,
we
kind
of
have
this.
This
pair
down
list
of
of
ready
for
development
items
feature
flag,
I'm
for
yeah
rolling
out
the
feature
flag
for
geo
group.
Wiki
replication!
F
Is
that
something
that
we're
just
kind
of
tracking
here,
but
is
actually
going
to
be
done
by
the
mr
author?
Or
is
that
something
that
we're
going
to
to
manage
ourselves
once
that
mr
is
merged?
Does
anyone
know.
F
Okay,
I'll
yeah
make
a
note
in
here
to
figure
that
out
and
yeah,
and
then
I
think
these
have
been
here
for
a
while
but
and
should
be
yeah.
So
we've
got
the
s2
at
the
top
of
the
list
and
then
a
couple
of
low
weight
severity,
three
issues,
and
then
I
see
these
ones
in
the
open
column.
F
A
F
Yeah
no
problem,
that's
that's
good
news.
I'll!
Just
move
this
over
here!
Yeah!
Is
there
anything
else
you
wanted
to
say
about
it.
A
A
So
I'm
just
attempting
to
kind
of
add
a
starting
point
of
like
the
current
situation
like
back-end
maintainers
are
all
pinged
on
this
thing.
But
if
anybody
has
any
interest
in,
you
know
like
where
we
should
do
permission
checks,
please
feel
free
to
check
out
the
merger
quest
yep.
F
Cool
thanks,
yeah.
I
think
I
think
that
yeah,
that
really
does
it
for
for
ready
for
development,
so
yeah,
if
you're,
if
anyone's
looking
for
for
something
to
work
on
next
outside
of
an
epic.
I
think,
looking
here
at
these
one-off
ready
for
development
items
is
a
good
idea.
We've,
you
know
even
explicitly
stated
it
for
for
1310
that
we
want
to
where
we
can
take
a
look
at
some
of
these.
F
These
one-off
small
bug
fix
quality
of
life,
types
of
issues
so
yeah.
Definitely
if
you're
kind
of
like
between
an
epic
and
looking
for
for
something
feel
free
to
to
take
a
look
at
this
list
before
you
know,
maybe
considering
the
next,
the
next
big
epic-
and
I
guess
I'll,
also
yeah.
This
one
is
under
an
epic,
but
it's
backup
restore
get
lab
instances.
So
it
is
all.
D
F
Its
own,
oh
I'm,
not
sharing
my
screen
anymore,
but
if
you
look
on
the
board,
there's
there's
a
backup
restore
issue.
I
think
it's
in
ready
for
development.
Alex
is,
is
assigned
to
it
now,
but
I
would
also
consider
that
in
the
one-off
issues
category
so
thanks
for
for
assigning
yourself
there
alex
oh.
E
If
someone
else
wanted
to
do
it,
I'm
not
like,
I
don't
have
dibs
I
just
there.
It
seemed
like
something
that
someone
should
be
assigned
to
to
make
because
it's
on
a
customer
issue
and
they
seemed
unhappy
that
no
one
was
working
on
it.
So
I
was
like
oh
well,
I
could
just
be
an
assignee
and
at
least
they'll
feel
like
it's
getting
someone's
attention,
rather
than
no
one's
attention
which
wasn't
getting
no
one's
attention.
Sorry,
I
don't
want
to
put
that.
I
don't
want
it
to
sound
like
it
wasn't
getting
attention.
It
was
just.
F
Yeah
yeah,
no,
no
worries,
yeah.
I
think
it's,
I
think.
That's
that's
a
good
idea,
yeah,
it's
a
a
big
community
community
user
and
I,
I
think,
yeah
I
think,
that's
fine.
I
think
we
did
say
that
it
would
be
scheduled
for
1311.
So
maybe
if
we
we
hit
1311-
and
you
still
haven't-
had
a
chance
to
look
into
it
more,
we
can
unassign
it
and
highlight
it
as
something
for
someone
to
to
take
on.
F
So
let
me
stop
recording
now.