►
From YouTube: 2020 08 25 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
So
a
couple
items
I
that
caught
my
eye,
while
I
was
reading
through
my
backlog
of
email,
this
first
one
about
application,
load,
balancing
not
resilient
to
network
disconnects
on
replicas
we
were
paying,
but
I'm
not
sure
this
is
something
that
we
need
to
own,
but
we
need
someone
else
to
take
a
look
at
it
is
there
anybody
wants
to
take
a
quick
look
and
see
if
this
is
something
we
should
own
or
hand
off
to
sorry.
B
Yeah
I
took
a
quick
look
in
the
afternoon
and
I
think
it
has
a
quite
straightforward
fix,
so
it
might
just
be
a
configuration
issue.
Okay
jorik
found
that
and
pointed
it
out.
I
like,
I
reproduce
it
and
it
does
fix
the
problem.
B
A
A
A
All
right
and
then
I
was
looking
through
the
audit
events,
epic
and
the
ux
changes
are
not
scheduled,
yet
they
have
a
they
put
a
priority.
One
label
on
it,
so
they're
going
to
go
through
and
break
it
down
to
see
what
further
information
they
need
before
they
implement
it.
But
I
didn't
see
anything
any
issues
in
there
for
follow-up
items
for
us.
A
If
we
need
to
switch
anything
over
to
start
using
that,
once
that
ux
ticket
is
gone
because
I
was
going
to
mark
the
ux
ticket
as
a
blocker
before
we
can
do
what
we
need
to
do
to
actually
start
using
the
other
than
tables.
Did
I
just
miss
it
or
you
know,
do
we
need
to
do?
We
need
to
create
an
issue
for
any
kind
of
switchover
to
start
using
those
tables
or
my
way
off
base
here.
B
A
I
thought
the
plan
was
wait
for
the
ux
changes,
because
those
are
going
to
optimize
based
on
the
partitions
we
created.
I
don't
think
performance
will
be
terrible
based
on
the
usage
patterns
that
we've
seen
of
honda
events,
but
my
understanding
was
the
plan.
Was
we're
always
going
to
wait
on
those
ux
changes.
B
So
basically
we're
cleaning
up
the
migration
preparing
the
table,
so
it's
ready
to
be
switched
and
then
we
wait
for
the
exchanges
once
they're
in
we
can
see.
We
can
do
some
testing
or
we
can
have
an
extra
issue
that
is
actually
about
swapping
tables
and
before
we
do
that
we
test.
If
if
the
partitioning
gets
picked
up
right
and
then
we're
done.
C
Yeah,
I
think
the
swapping
will
be
might
be
a
little
bit
more
complicated
than
we
originally
had
talked
about
due
to
mostly
because
the
in
the
new
table,
the
primary
key
is
multi-column,
which
we
know
rails
doesn't
support
out
of
the
box,
and
I
think,
if
you
swap
the
tables,
while
the
application
is
running
that
could
have
potential,
I
think
initially
we
had
said
you
know
that
if
we
swap
the
tables,
the
definitions
are
the
same.
C
C
But
then,
if,
for
some
reason
like
a
server
were
to
restart-
and
it
would
start
running
that
same
code
but
reload
the
schema,
then
suddenly
everything
is
broken,
because
now
it's
seeing
the
primary
key
which
it
doesn't
recognize
and
like
some
you
know,
simple
things
like
saving
records,
won't
work
anymore
because
it
can't
auto
generate
the
value
and
and
lookups
on
the
id
fail,
because
it
doesn't
know
how
to
handle
that.
So
you
probably
need
to
talk
about
that.
B
Would
it
make
sense
to
after
finishing
up
the
migration
cleanup
that
we
look
we
look
into
how
to
swap
would
work
like,
so
that
we
have
a
plan
and
we
know
how
much
work
that
is
or
have
an
idea
even.
A
Pat,
since
you
did
such
a
good
job
of
describing
the
work
I
assigned
you
an
answer.
A
Thank
you
all
right
see
who
was
here
last
week
was
it.
It
was
just
pat
and
andreas
right.
B
A
All
right,
I
didn't
see
anything
written
down
for
goals
from
last
week,
which
is
fine.
Did
you
have
any
goals
that
you
all
set
and
wanted
to
review,
or
should
we
just
jump
into
what
we're
going
to
do
this
week.
B
I
think
we
kind
of
went
over
the
all
the
milestone
priorities
last
week
and
on
top
of
that
we
have
the
container
registry,
so
part
of
last
week
was
getting
getting
up
to
speed
with
that
and
then
throughout
the
engineer
he's
he's
on
vacation
for
two
weeks
so
that
we
had
a
couple
of,
or
I
had
a
couple
of
like
questions
catching
up
with
him
last
week.
A
To
it
all
right,
let's
just
jump
into
this
week
and.
C
Sorry
one
thing:
I
think
that
we
sort
of
neglected
to
do
next
last
week,
too,
was
go
through
the
mason
and
say
which
issues
would
be
deliverable.
A
C
B
B
A
And
then
I
threw
the
data
dictionary,
epic
into
our
backlog.
It's
cat
would
like
this
has
asked
for
that,
and
it
actually
came
up
in
the
development
channel
too.
Someone
asking
about
columns
and
data
dictionary
might
help
there
that's
more
of
a
long-term
goal.
It's
not
something.
We
need
to
start
this
milestone,
but
something
we
can
talk
about
if
time
allows
all
right.
So
issues
for
this
milestone
can
mark
as
deliverable
here.
B
Probably
some
I
mean
this
is
sort
of
a
one-off
thing
for
for
that
user
right.
But
if
the
pattern
there,
if
the
race-
or
I
think
there
is
a
race
condition
in
there
like
like
pat
described
so
there
is
some
room
that
this
pops
up
more
regularly.
Actually,
if
that's
right,
I
think
we
should
fix
that
or,
or
is
that
pattern
correct
that
we
see.
C
Yeah,
I
think
your
suggestion
of
you
know
if
we
don't
see
the
kind
of
minimum
partition
there
that
we
should
just
skip
it.
I
think
that
makes
sense,
because
that
should
always
be
there
if
the
partition
ran.
Even
if
there
would
be
no
data,
there
would
still
be
at
least
that
one.
A
This
one
assumes
that
we
have
oh
more
of
a
question.
Does
this
one
assume
that
we've
rolled
it
out
in
production?
A
B
Yeah,
that's
kind
of
the
one
we're
working
on
right
now
and
it's
maybe
not
only
about
identifying
the
access
patterns,
sort
of
went
into
like
we're
currently
working
on
an
alternative
proposal
for
a
database,
and
then
we
make
sure
we
want
to
discuss
what
the
benefits
are
and
make
sure
we
understand
the
trade-offs
and
then
finalize
the
database
model
in
this
monster.
B
Yeah,
it
might
just
be
that
we
end
up
doing
what
is
sort
of
currently
out
there,
but
there
is
an
alternative
that
we
can
explore
and
I
think
we
should
spend
some
time
on
that
and
then
discuss
what
the
the
trade-offs
are.
B
Okay,
because
the
the
model
that
is
proposed
is
not
bad,
but
it's
very
flexible
and
it
might
have
some
problems.
A
Right,
like
actual
access
patterns
that
work
with
partitioning
easily
no
yeah,
okay,
giannis
remove
the
get
ignored,
schema
migration
files.
A
B
The
next
one
is,
I
think
the
title
is
a
bit
strangely
phrased,
but
it's
about
like
highlighting
in
the
documentation
that
certain
postgres
requirements
have
changed.
I
think
it's
so
it's
just
a
documents
edition
I
yeah.
I
can
do
that.
A
B
But
it's
so
I
think
it's
at
a
point
where
we
can
also
make
a
call
and
go
with
something
because
we're
kind
of
not
getting
a
lot
of
feedback,
and
I
think
they're
also
quite
thin
at
the
moment-
and
I
have
asked
about
like
do-
do
we
have
plans
to
do
the
repacking,
like
with
a
higher
frequency
like
not
only
once
a
year
but
more
than
that
and
even
automated
and
there's
no
work
schedule
for
that
on
their
side.
So
maybe
that
would
help
get
that
going
or
maybe
help
fix
that
problem.
A
Oh,
if
we
just
did
the
break
task.
B
Yeah
or
it's
not
done
with
the
red
class,
but
if
we
went
in
that
direction
and
offered
them
something
to
use
again.
A
Really
bad
alone
and
this
one's
being
worked
on
right
now,
index
helpers.
B
B
The
issue
is
not
there
is
it's
a
tracking
issue,
basically
tracking
the
problem
but
yeah.
When
we
talk
about
events
partitioning,
then
we
should
also
have
a
solution
for
that.
Yep.
A
Okay,
developer
documentation
for
migration
helpers.
B
A
So
do
we
have
any
specific
goals
for
the
week
that
we
want
to
discuss.
B
So
janice
and
I
we
discussed
sketching
the
container
registry
database
alternative,
let's
say
and
then
getting
together
end
of
the
week
and
discuss
that.
C
Yeah
finish
that
up-
and
I
guess
you
know
investigate
as
you
can
formulate
a
plan
to
swap
the.
C
A
A
A
A
A
A
I'm
gonna
stop
sharing
that
so
thursday
we'll
do
the
team
retro
then
make
sure
that
that's
on
the
agenda.