EDB Postgres with Containers

© 2014 EDB All rights reserved. 1
EDB Postgres with Containers

© 2016 EnterpriseDB Corporation. All rights reserved. 2
•  Who Is EnterpriseDB?
•  What Are Containers?
•  How Can I Get Started?
•  How Are Stateful Applications Managed?
•  Can I Run This With My Current Architecture?
What we will learn

About EnterpriseDB

EDB is a Recognized Leader
Magic Quadrant for Operational DBMS, Q4 ‘15
• Gartner Comments
− “EnterpriseDB is responsible for
many features of PostgreSQL,
contributing to JSON, materialized
views and partitioning. Increasing
numbers of users of Gartner’s
inquiry services recognize the
EnterpriseDB brand.”
− “Clients report that the functionality
of EnterpriseDB's Postgres Plus
Oracle compatibility feature is more
than sufficient to run both mission-
critical and non-mission critical
applications.”
− “Reference customers rated
EnterpriseDB highly for the stability
of its DBMS, and were satisfied with
its HAS/DR features.”

POSTGRES
innovation
ENTERPRISE
reliability
24/7
support
Services
& training
Enterprise-class
features & tools
Indemnification
Product
road-map
Control
Thousands
of developers
Fast
development
cycles
Low cost
No vendor
lock-in
Advanced
features
Enabling commercial
adoption of Postgres

EDB Customers
EDB currently has over 3,000 total customers including 63 of the Fortune
500 and 137 of the Forbes Global 2000

Containers Have Changed Everything

Containers
are
kind
of
old
news

Linux
Containers
(LXC)
and
CGroups
and
namespaces
are
extensions
of

concepts
that
started
with
Solaris
and
BSD
Jails

But
projects
like
Docker
and
RKT
are
very
new,
and
very

ac(ve.

●  This
last
week,
Docker
received
107
commits
from
51
authors
on
github

●  More
than
22,000
commits
since
iniPal
in
2013

●  2/3
of
Companies
That
Try
Docker
Adopt
It (according to datadog)
https://www.datadoghq.com/docker-adoption/
Where
did
they
come
from?

Developers
and
administrators
conPnue
to
face

well
known
problems:

How
to
package
applicaPon
code
across

environments

How
to
get
the
best
performance
from

available
physical
resources

Why
the
sudden
popularity?

Packaging
ApplicaPons

Have
you
ever
moved
a
simple
code
ﬁx
into
ProducPon
only

to
have
problems
with
a
library
mismatch?

Which
leads
to
an
easy-‐enough
looking
upgrade?

Which
suddenly
cascades
to
updaPng
mulPple
packages,

tracing
dependencies,
and
hours
of
mess?

. . . Well me neither :-)

Packaging
ApplicaPons

Containers
provides
a
versioned
run-‐Pme
environment.

A
container
is
just
a
tar
file,
or
a
series
of
tar
files.

It
holds:

●  An
OperaPng
System

●  Libraries

●  All
dependencies

Each
step
builds
a
new
layer,
making
the
finished
container

image
instantly
runnable.

• Be_er
performance
than
VMs

• Higher
uPlizaPon
than
bare
metal

• Faster
IOPs
path

RunPme
Architecture

Server
Host OS
Bins / Libs Bins / Libs
Docker
AppA
AppA’
AppA’’
AppB
AppB
AppB

A
container
wraps
one
single
process:

• A
web
server
(nginx,
ligh_pd)

• An
applicaPon
server
(Python,
Java,
C)

• A
middleware
component
(RabbitMQ,
memcached)

• A
stateful
service
(PostgreSQL,
Redis)

• A
one-‐oﬀ
job
(backup
script,
archiving,
ETL)

What’s
in
the
container?

Containers
Change
Architecture

© 2016 EnterpriseDB Corporation. All rights reserved. 15© 2013 EDB All rights reserved 8.1. 15
Getting Started

Containers
are
NOT
Virtual
Machines

They’re
just
tar
ﬁles

That
you
can
run

Keep
one
thing
in
mind

EDB
Postgres
image

Oﬃcial
PostgreSQL
Docker
image
(DockerHub)

Build
your
own

Pulling
images

Containers are cheap to create, easy to dispose
Containers wrap a single process
Containers usually do not alter state
That is to say, most containers don’t write data
Wait. What about logs?
•  manage them with a purpose-driven container
•  send them to stdout
The container mindset

Stateful Applications

The
ﬁle
system
of
a
typical
docker
container
is
not
permanent

But
criPcal
data
can
easily
be
made
permanent
by
using

Volumes

These
map
the
host’s
storage
directly
into
the
container

What
happens
to
the
data?

# docker run -v /my/local/volume:/container/pgdata
--name vol_ppas ppas:9.5

Other
containers
can
share
the
same
volume,
or
mount
the

volume
on
start.

What
happens
to
the
data?

# docker run --name my_ppas
--volumes-from vol_ppas
-p 5432:5432 ppas:9.5 start
# docker run --name ppas_log_mgr
--volumes-from vol_ppas
--volumes-from log_dest
ppas:9.5 log_maint

I
can
haz
container
demo?

Putting Containers Into
Production

Bringing
in
new
technology

Containers
will
run
the
same
way
everywhere

Containers
can
discover
exisPng
infrastructure

Bringing
in
new
technology

Begin
leveraging
containers
in
small
ways.
.
.

Replace
a
single
process
(sogware
build)

Replace
a
single
server
(load
balancer)

Install
a
portable
toolset
(check
out
nsenter!)

Replace
a
long
command
(generate
cerPﬁcates)

Closing Thoughts

Infrastructure needs to scale beyond managing VM’s
Breaking apart applications into microservices helps
View infrastructure as code
Containers are just tar files
Try them in a few small places and work your way up
A
few
general
observaPons.
.
.

Performance
Tuning

ConsulPng

Training
&
CerPﬁcaPon

Remote
DBA
Services

How
can
we
help?

h_ps://enterprisedb.com/ahc

● Send
email
to
sales@enterprisedb.com

● h_ps://www.enterprisedb.com/general-‐inquiry-‐form

● h_ps://www.enterprisedb.com/resources-‐community

EDB Postgres with Containers

More Related Content

What's hot (20)

Similar to EDB Postgres with Containers (20)

More from EDB (20)

Recently uploaded (20)

EDB Postgres with Containers