►
From YouTube: 2021 03 02 Database Team 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
All
right,
this
is:
oh,
it's
actually
march
2nd
all
right
wrong
date,
so
march,
2nd
database
team
weekly
meeting
jump
right
into
it,
get
through
this
infra
dev
issues
with
jose.
I
didn't
see
anything
new
or
unassigned
in
there
jose.
Are
there
any
burning
issues
that
we're
not
looking
at
at
the
moment.
A
Okay,
yeah
and
another
question
I
asked
is
jerry
raised
the
point
that
we
should
go
over
the
saturation
report.
I
believe
that's
what
he's
talking
about
because
he
said
monthly
report.
A
B
A
A
All
right
on
the
billboard,
nothing
in
the
close
column
scrolling
all
the
way
back
to
blocked.
Does
anybody
know
what's
going
on
with
this
one.
D
A
All
right,
let's
see
this
one,
I
haven't
looked
at
this
issue
for
a
while.
Where
are
we
at
with
rolling
out
testing
the
database
maintainers.
D
Right,
we
have
released
it,
so
this
is
a
tracking
issue
to
follow
whether
everyone
so
for
everyone
to
read
the
how
things
are
done
and
just
sign
off
that
they
tried
out
the
solution.
D
Yeah,
so
we
thought.
D
Reverted
it
and
yesterday.
A
D
There
were
some
triggers
that
were
manually
added
in
a
state
in
the
production
and
they
were
added
using
not
using
the
gitlab
user,
so
the
gitlab
user
that
runs
the
migrations,
could
not
drop
the
trigger
into
functions.
So
we
reverted
the
migration
and
we
manually
fixed
the
problem
in
staging.
D
And
I
I
will
update
the
issue
and
call
it
done
okay,
because
this
is
not
relevant
for
instances.
So
we
totally
remove
the
migration
again,
all
right.
C
Yeah
so
I
think
the
primary
key
conversion
stuff,
so
we
had
a
discussions
on
that
last
week
and
I
think
we're
pretty
close,
just
really
settled
on
the
initial
iteration,
what
that's
going
to
be
finishing
up
tests.
I
think
I'll
have
the
test
finished
today
and
then
get
it
into
review.
My
goal
is,
hopefully,
if
we
can
get
it
merged
this
week,
that
would
be
good
and
then
we
can
get
the
migration
rolling
on
events
and
at
least
get
some
real
production
or
staging.
A
Do
we
have
the
estimates
on
how
long
it's
going
to
take
improved,
or
is
that
still
going
to
require
some
subsequent
testing.
C
Yeah,
we
haven't
really,
you
know
we're
guessing,
maybe
that
like
something
like
three
to
five
times
faster
than
what
we
saw
in
in
database
lab,
but
we
really
don't
know,
that's
why
we
just
want
to
get
it
in
production
as
quick
as
we
can
and
then
make
adjustments
and
get
a
feel
for
it.
Then
that
will
you
know
guide
how
long
we
think
the
other
migrations
will
take
too.
Okay.
D
Yeah,
that's
related
to
the
work
done
by
pat,
so
this
is
the
update
that
we
need
in
order
to
be
able
to
also
test
using
our
framework,
the
new
background
migrations,
but
we're
not
going
to
block
what
a
path
is
doing.
D
D
Yeah
and
it's
easy
to
figure
find
new
things
that
we
did
not
know
about
so,
for
example,
in
my
case,
starting
a
monday
morning,
I
found
the
three
issues
I
did
not
know
about,
and
you
know
they
did
not
pop
up,
because
some
of
those
are
not
labeled
as
bad.
They
did
not
pop
up
in
our
three
housing
issue,
so
it's
great
to
have
it
there
and
also
remind
us
that
we
have
to
do
to
do
to
action.
A
A
Okay,
since
it's
just
you
right
now,
pat,
do
we
need
to
talk
about
any
focused
topics
for
tomorrow?
Are
we
good.
B
D
Green
light
from
the
ecosystem
team
today
so
the
moment
we
get
a
green
light,
I'm
going
to
to
start
and
add
the
first
migration,
the
the
first
america
that
we
need
to
start
partitioning,
webco
clones
and
send
it
apart
for
a
review
and
we
can
release
it
by
the
end
of
the
week.
That's
my
planning
list.
A
E
Well,
yeah:
this
is
just
for
for
people's
attention,
so
this
is
a
p1
s1
that
breaks
any
geo
upgrades
to
13.9
and
it
also
breaks
running
back
to
backups
for
any
other
version,
and
this
is
the
consequence
of
a
rename
of
a
column
to
align
with
our
diversity
and
inclusion.
E
I
think
actually,
like
I've
come
to
the
conclusion.
The
reason
for
the
rename
is
not
so
important
for
this.
I
think
my
question
for
for
this
group
is
more
when
people
are
starting
to
rename
stuff,
how
can
we
make
that
easier
or
safer
to
do?
Because
I
think
this
before
we
realized
this
broke
upgrades.
E
It
also,
I
think,
caused
another
production
incident,
and
so
it
seems
like
a
pretty
risky
thing,
but
I
know
too
little
about
that
specific
process
that
has
to
do
with
it,
but
I
just
wanted
to
highlight
it
for
for
this
group,
because
maybe
there's
some
need
for
for
updated
guidelines,
or
you
know
a
section
that
says
hey.
You
know
if
you
were
thinking
about
renaming
something
please
take
into
consideration.
You
know
like
some
of
these
other
things,
so
I
just
wanted
to
like
highlight
it,
because
it
seems
like
a
whole
lot
of
side.
D
Yeah,
so
rename
columns
is
one
of
the
most
tricky
updates.
We
can
do
to
our
schema,
with
renaming
tables
being
the
the
worst
the
most
difficult.
So
I
totally
agree
with
stan
holm.
We
had
the
discussion
where
he
was
saying
that
we
should
not
rename
columns
if
there
is
not
a
really
really
really
good
reason
to
do
so.
That
does
not
mean
that
we
should
not
have
the
migration
helpers
to
allow
you.
D
So
this
is
something
that
we
have
discussed
in
the
past
with
pat,
so
we
have
two
types
of
two
ways
to
rename
columns
right
now
we
have
the
way
that
was
used
in
dcmr,
which
is
called
concurrently,
renamed
columns
called
whatever,
which
is
in
line.
D
So
in
reality,
this
this
runs
and
renames
the
column
in
one
milestone
and
makes
the
switch
in
one
milestone
and
can
cause
all
the
problems
that
we
discussed,
and
then
we
also
have
another
migration
helper
that
we
don't
advertise
that
much
in
in
the
dogs,
it's
called
the
name
using
background
migration,
which
in
theory
that
was
built
back
in
the
day
to
help
rename
columns
for
very
large
stages.
D
D
Please
be
careful
when
you
rename
columns,
that's
one,
and
maybe
we
can
have
a
section
explaining
why
this
is
so
tricky
and
what
the
problems
can
be,
but
from
a
touring
perspective,
my
opinion
is
that
we
should
rebuild
the
renaming
column
using
background
migration
after
we
are
done
with
the
primary
key
conversions
following
what
we
are
doing,
with
primary
key
conversions,
to
use
background
jobs,
to
copy
everything
to
a
new
column
in
one
milestone,
keep
both
columns
there
and
allow
the
and
have
a
guidelines
that
will
tell
developers
to
you
know.
D
Do
it
in
three
steps,
add
a
new
column,
duplicate
the
the
old
one
and
then
remove
the
old
one.
That
should,
I
think,
from
what
I
can
think
of
it
will
remove
problems
with
all
code
new
code
trying
to
access
different
columns
that
are
not
there,
because
during
the
transition
we
will
have
both
columns
there.
E
I
thought
naively,
like
I
was
only
aware
of
the
process
that
you
just
described,
because
that
is
also
the
only
way
to
maintain
zero
downtime
upgrades
right,
because
then
you
can.
Actually,
you
know,
do
this
over
various
versions
right
and
you
don't
need
to
introduce
downtime
to
drop
tables
or
rename
them
or
some
something
like
that
right.
You
can
so
yeah.
That
sounds
good.
E
D
At
some
point
we
have
to
rebuild
this
helper
because
right
now,
this
helper
does
not
work
as
expected.
So
this
helper
is
using
background,
migrations,
background
jobs
to
execute
the
transition,
but
it
also
incurs
the
the
final
swap.
D
So
that
means
that
the
swap
at
some
point
so
it
happens,
the
sub
swap
will
happen
the
same
myosin,
which
could
also
cause
problems.
So
I
think
that
if
we
could
repeat
the
repeat,
some
of
those
helpers
using
what
we
learned
in
the
new
framework
that
patrick
is
building,
we
could
have
way
better,
a
way
better
approach
for
renaming
columns
and
also
for
change
the
the
type
of
a
column
which
is
also
tricky,
not
the
tricky.
D
C
C
I
think
we
need
to
take
an
opportunity
to
re-evaluate
whether
it
makes
sense
to
keep
using
those,
as
is
because,
if
not,
we
should
make
it
clear
that
they're
not
you
know
they
shouldn't
be
using
them,
and
I
agree
also
that
rename
is
it's
risky
and
I
think,
no
matter
what-
and
I
think
we
shouldn't
be
really
doing
it
unless
we
have
a
very
good
reason
to
do
it.
It's
just
kind
of
I
think
some
like
technical
debt
that
might
have
to
just
be
accepted
that
it's
part
of
the
application.
D
Yeah,
to
tell
the
truth:
most
probably,
we
already
have
an
issue
to
update
rename
column
but
yeah.
I
I
will
find
that
and
add
another
one
so
that
that
we
can
update
the
dogs
when
we
get
the
ability
to
breath,
breathe
and
yeah.
I
mean.
E
This
is
not
a
drop
everything
do
that
now,
as
you
explain
right
it's
like,
but
I
think
that
it's
just
something
that
I've
seen
and
the
impact
of
customers
is
not
great
right,
and
so
we
want
to
make
sure
that
in
the
longer
term,
we
don't
run
into
similar
issues.
If
we
can
avoid
it.
D
And
then
another
hidden
issue
in
that
error,
and
if,
if
someone
was
to
check
that
there
from
the
geo
team,
is
that
another
way
to
cause
problems
is
not
to
for
when
an
instance
does
not
follow
the
proper
process
of
running
the
migrations.
D
So
it's
also
when
the
the
instance
tries
to
run
a
no
downtime
update
without
running
first,
the
migrations
then
doing
whatever
we
have
to
do
for
jio
and
then
try
running
the
post
migrations,
and
this
is
one
more
time
we
have
seen
also
other
cases
where
instances
that
they
don't
that.
Don't
follow
the
proper
process
on
how
to
execute
the
post
migrations
that
can
cause
issues,
and
this
is
a
more
general
thing
to
think
about
in
the
future.
E
If
you
follow
the
process
exactly
you're
fine,
if
you
make
a
mistake,
you're,
probably
very
not
fine,
right
and
so
they're,
not
that
many
guardrails
where,
where
people
are,
you
know
like
warrant
that
hey
you've
not
run
these
things
beforehand
right.
It's
like
we
will
not
allow
you
to
do
this
or
some
such
thing.
A
D
Where
yeah
create
view
yeah
that
can
cause
addition.
Other
types
of
problems
because-
and
there
is
also
from
adam-
a
proposal
on
renaming
using
a
variation
of
using
views,
maybe.
D
A
I
just
looked
at
the
migration
testing
tracking
issue.
It
doesn't
look
like
anybody
has
responded
to
that
and
I
haven't
looked
through
the
database
channel
yet
to
see
if
anybody's
talking
about
it.
Do
you
know,
has
anybody
seen
it
actually
being
used
by
other
maintainers.
D
A
A
D
D
With
more
than
10
000
with
around
10,
000
or
2000
commits
inside
them,
let
me
check
he
did
he
has.
D
A
We
are
at
the
end
of
the
agenda.
You
know
the
topics
folks
want
to
talk
about
today.
E
E
A
All
right,
I'm
gonna,
turn
off
recording
here
in
a
second,
let's
go.