►
From YouTube: New User Training: 11 Databases
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
Okay,
so
this
is
just
a
short
introduction
to
the
databases
we
provided.
So
not
everyone
wants
to
store
their
data
and
files
on
on
their.
You
know
parallel
file
system.
In
some
cases
you
know
it's
better
fit
with
the
databases
and
we
do
provide
some
database
services
and
ask
so
basically,
we
provide
both
relational
databases,
SQL
databases,
MySQL
and
Postgres,
and
also
MongoDB
for
those
who
prefer
a
schema
less
database
and
there's
that
what
the
only
important
thing
in
this
talk
is
there's
a
form
in.
A
If
you,
if
you
want
one,
you
should
go
to
that
site,
two
to
request
it
now.
These
services
are
only
meant
to
accompany
computational
work
done
at
nurse
they're,
not
like.
The
only
thing
that
you
do
at
nursing
is
to
host
data
in
these
databases,
and
so,
for
example,
unless
this
is
a
specific
good
reason
why
you
need
to
be
able
to
access
these
from
outside
nurse
they're
they're,
generally
only
accessible
from
within
the
nurse
network
and
they're.
A
Also
they're
just
hosted
on
single
servers,
they're
shared
with
many
other
users,
so
they're
not
designed
for
high
performance
really,
but
in
the
future
and
Tony
will
talk
about
this.
We
have
will
have
these
services
called
spin
and
we
do
to
some
extent,
already
have
that
that
allows
people
to
spin
up
their
own
user
configured
databases,
but
at
the
moment
you
can,
you
can
just
request
these
and
for
certain
workloads.
These
will
still
continue
to
exist
and
and
be
a
better
option.
A
Okay,
so
just
a
quick
introduction
to
the
SQL
databases.
So
you
know
some
general
sort
of
tips.
These
are
pretty
standard
options
where
you
know
in
advance
the
schema
that
you
that
you
want
to
store
in
here
they're
relational
databases.
So
this
means
you
know,
tables
of
rows
and
columns
that
some
connected
together
with
with
keys
that
you
can
use
to
join
together
a
table
these
tables,
the
the
kind
of
setup
we
have
here
is
really
only
meant
for
sort
of
mid-sized
databases,
so
a
few
gigabytes
or
whatever.
A
But
you
know
these
are
very
stable
services,
so
they're
pretty
good
for,
for
example,
transactional
operations
or
things
where
you
want
to
know
that
your
database
is
consistent
as
I
mentioned
again.
These
are
single
databases
on
single
servers,
and
so
this
is
the
main
point
of
reiterating
here
is
to
say
that
if
you
access
it
from
10000
compute
nodes
on
Cori
you
they
will,
you
won't
get
10,000
times
the
performance,
because
it
will
be
serialized
at
this
server.
A
You
know
people's
choice
between
Postgres
and
MySQL
is
normally
because
of
other
reasons
like
they
already
are
familiar
with
what
other
of
them.
But
some
example.
You
know
advantages
of
Postgres
are
particularly
for
certain
communities
like
the
astro
community.
There's
various
extensions
that
have
already
been
written,
various
functions
that
can
just
be
used
off-the-shelf,
and
it
also
has
a
good
adherence
to
the
SQL
standards,
whereas
MySQL
is
probably
one
of
them.
You
know
they're,
probably
the
most
popular
option
for
this
kind
of
thing
and
a
good
place
to
start.
A
You
know
this,
isn't
this
very
basic
SQL
isn't
going
to
get
you
anywhere,
but
the
first
thing
you'll
want
to
do
is
change
the
password
that
I
give
you
the
default
password
that
I
provide
you,
and
then
you
know
the
you
can
create
tables
and
so
forth,
and
this
is
a
good
resource.
If
you
didn't
know
about
it,
for
learning
SQL.
A
It
also
how
queries
can
be
very
fast,
and
this
is
one
of
the
people.
The
reasons
people
like
this
is
somewhat
mitigated
in
our
case,
because
we
have
this
shared
service.
So
you
know
different.
Other
users
require
what
you
just
because
you're
frequently
accessing
it
doesn't
mean
it
will
necessarily
be
on
the
cache
because
of
the
users
data
may
be
in
there
sort
of
related
to
that,
even
though
it
can
be.
You
know
a
sharded
service,
that's
extremely
high
performance.
A
A
Okay
and
here
are
similarly
very
basic
commands.
We
actually
have
a
module
on
compute
services
here
that
you
have
to
load,
and
then
you
can
just
access
it
via
this
shell
command
and
then
you
can
write
these
sort
of
JavaScript
expressions
and
put
documents
as
they're
called
in
into
the
database.
But
most
people
like
to
interact
as
with
most
things
these
days.
Five,
and
so
you
can
use
the
PI
library
and
connect
to
that
and
do
everything
in
Python.
A
Ok,
so
just
a
couple
of
comments
about
where
databases
are
actually
used
here.
So,
as
I
mentioned,
you
can
access
by
the
command
line
or
in
your
scripted
code.
But
a
lot
of
the
use
comes
from
these
science
gateway
applications
which
are
like
Web
Services,
where
people
want
to
query,
derive
data
that
they've
put
in
the
databases
here.
A
Just
I
just
put
some
some
observations
about
where
people
tend
to
use
files
or
databases
in
science
which
might
be
different
in
industry
or
whatever
you,
but
generally
speaking,
as
Gillian
indicated
that
large
massively
parallel
HPC
programs
normally
write
their
data
out
in
files.
This
has
advantages
for
large,
shared
collaborations,
for
example,
that
it's
easier
to
share
the
data
as
well.
You
can
just
copy
and
we'll
hear
about
data
transfer
later
you
can
just
copy
the
files
elsewhere.