Introduction to Docker (and a bit more) at LSPE meetup Sunnyvale

Introduction
to
Docker
(and a bit more...)
#lspe

February 2014—updated for Docker 0.8.1

Who?
●

Jérôme Petazzoni — @jpetazzo
–Not

Maxime Petazzoni (author of Maestro NG)
–Not Thomas Petazzoni (embedded systems & kernel guru)
●

Wrote dotCloud PAAS deployment tools
–EC2,

●

LXC, Puppet, Python, Shell, ØMQ...

Docker contributor (CONTAINERIZE ALL THE THINGS!)
–Docker-in-Docker,

●

VPN-in-Docker, router-in-Docker...

Runs Docker in production
–Psssst...

#lspe

You shouldn't do it, but here's how anyway!

Outline
●

LXC and the container metaphor

●

Docker basics

●

Docker images

●

Docker deployment

#lspe

High level approach:
it's a lightweight VM
●

own process space

●

own network interface

●

can run stuff as root

●

can have its own /sbin/init
(different from the host)

« Machine Container »

#lspe

Low level approach:
it's chroot on steroids
●

can also not have its own /sbin/init

●

container = isolated process(es)

●

share kernel with host

●

no device emulation (neither HVM nor PV)

« Application Container »

#lspe

How does it work?
Isolation with namespaces
●

pid

●

mnt

●

net

●

uts

●

ipc

●

user

#lspe

pid namespace
jpetazzo@tarrasque:~$ ps aux | wc -l
212
jpetazzo@tarrasque:~$ sudo docker run -t -i ubuntu bash
root@ea319b8ac416:/# ps aux
USER
root
root

PID %CPU %MEM
1 0.0 0.0
16 0.0 0.0

(That's 2 processes)

#lspe

VSZ
18044
15276

RSS TTY
1956 ?
1136 ?

STAT START
S
02:54
R+
02:55

TIME COMMAND
0:00 bash
0:00 ps aux

mnt namespace
jpetazzo@tarrasque:~$ wc -l
/proc/mounts
32 /proc/mounts

root@ea319b8ac416:/# wc -l /proc/mounts
10 /proc/mounts

#lspe

net namespace
root@ea319b8ac416:/# ip addr
1: lo: <LOOPBACK,UP,LOWER_UP> mtu 65536 qdisc noqueue state UNKNOWN
link/loopback 00:00:00:00:00:00 brd 00:00:00:00:00:00
inet 127.0.0.1/8 scope host lo
valid_lft forever preferred_lft forever
inet6 ::1/128 scope host
22: eth0: <BROADCAST,MULTICAST,UP,LOWER_UP>
pfifo_fast state UP qlen 1000

mtu 1500 qdisc

link/ether 2a:d1:4b:7e:bf:b5 brd ff:ff:ff:ff:ff:ff
inet 10.1.1.3/24 brd 10.1.1.255 scope global eth0
inet6 fe80::28d1:4bff:fe7e:bfb5/64 scope link

#lspe

uts namespace
jpetazzo@tarrasque:~$ hostname
tarrasque
root@ea319b8ac416:/# hostname
ea319b8ac416

#lspe

ipc namespace
jpetazzo@tarrasque:~$ ipcs
------ Shared Memory Segments -------key
shmid
owner
perms
0x00000000 3178496
jpetazzo
600
0x00000000 557057
jpetazzo
777
0x00000000 3211266
jpetazzo
600

root@ea319b8ac416:/# ipcs
------ Shared Memory Segments -------key
shmid
owner
perms
------ Semaphore Arrays -------key
semid
owner
perms
------ Message Queues -------key
msqid
owner
perms

#lspe

bytes
393216
2778672
393216

nattch
2
0
2

status
dest

bytes

nattch

status

nsems
used-bytes

messages

dest

user namespace
●
●

no « demo » for this one... Yet!
UID 0→1999 in container C1 is mapped to
UID 10000→11999 in host;
UID 0→1999 in container C2 is mapped to
UID 12000→13999 in host; etc.

●

required lots of VFS and FS patches (esp. XFS)

●

what will happen with copy-on-write?
–

double translation at VFS?

–

single root UID on read-only FS?

#lspe

How does it work?
Isolation with cgroups
●

memory

●

cpu

●

blkio

●

devices

#lspe

memory cgroup
●

keeps track pages used by each group:
–

file (read/write/mmap from block devices; swap)

–

anonymous (stack, heap, anonymous mmap)

–

active (recently accessed)

–

inactive (candidate for eviction)

●

each page is « charged » to a group

●

pages can be shared (e.g. if you use any COW FS)

●

Individual (per-cgroup) limits and out-of-memory killer

#lspe

cpu and cpuset cgroups
●

keep track of user/system CPU time

●

set relative weight per group

●

pin groups to specific CPU(s)
–

Can be used to « reserve » CPUs for some apps

–

This is also relevant for big NUMA systems

#lspe

blkio cgroups
●

keep track IOs for each block device
–

read vs write; sync vs async

●

set relative weights

●

set throttle (limits) for each block device
–

read vs write; bytes/sec vs operations/sec

Note: earlier versions (<3.8) didn't account async correctly.
#lspe3.8 is better, but use 3.10 for best results.

devices cgroups
●

controls read/write/mknod permissions

●

typically:
–
–

deny: everything else

–
●

allow: /dev/{tty,zero,random,null}...
maybe: /dev/net/tun, /dev/fuse, /dev/kvm, /dev/dri...

fine-grained control for GPU, virtualization, etc.

#lspe

How does it work?
Copy-on-write storage
●

Create a new machine instantly
(Instead of copying its whole filesystem)

●

Storage keeps track of what has changed

●

Since 0.7, Docker has a storage plugin system

#lspe

Storage:
many options!
Union
Filesystems

Snapshotting
Filesystems

Copy-on-write
block devices

Provisioning

Superfast
Supercheap

Fast
Cheap

Fast
Cheap

Changing
small files
Changing
large files
Diffing

Superfast
Supercheap

Fast
Cheap

Fast
Costly

Slow (first time)
Inefficient (copy-up!)

Fast
Cheap

Fast
Cheap

Superfast

Superfast

Slow

Memory usage

Efficient

Efficient

Inefficient
(at high densities)

Drawbacks

Random quirks
AUFS not mainline
!AUFS more quirks

ZFS not mainline
BTRFS not as nice

Higher disk usage
Great performance
(except diffing)

Bottom line

Ideal for PAAS and
high density things

This is the Future
(probably)

Dodge Ram 3500

#lspe

Compute efficiency:
almost no overhead
●

●

●

●

processes are isolated,
but run straight on the host
CPU performance
= native performance
memory performance
= a few % shaved off for (optional) accounting
network performance
= small overhead; can be reduced to zero

#lspe

Alright, I get this.
Containers = nimble VMs.

#lspe

Problem: shipping goods
?

?

?

?

?

?

?

?

?

?

?

?

?

?

?

?

?

?

?

?

?

?

?

?

?

?

?

?

?

?

#lspe

?

?

?

?

?

?

Solution:
the intermodal shipping container

#lspe

Problem: shipping code
?

?

?

?

?

?

?

?

?

?

?

?

?

?

?

?

?

?

?

?

?

?

?

?

?

?

?

?

?

?

#lspe

?

?

?

?

?

?

Solution:
the Linux container

#lspe

Separation of concerns:
Dave the Developer
●

inside my container:
–

my code

–

my libraries

–

my package manager

–

my app

–

my data

#lspe

Separation of concerns:
Oscar the Ops guy
●

outside the container:
–

logging

–

remote access

–

network configuration

–

monitoring

#lspe

Docker-what?
●

Open Source engine to commoditize LXC

STOP!
HAMMER DEMO TIME.
#lspe

Yes, but...
●

●

●

« I don't need Docker;
I can do all that stuff with LXC tools, rsync,
and some scripts! »
correct on all accounts;
but it's also true for apt, dpkg, rpm, yum, etc.
the whole point is to commoditize,
i.e. make it ridiculously easy to use

#lspe

What this really means…
●

instead of writing « very small shell scripts » to
manage containers, write them to do the rest:
–

continuous deployment/integration/testing

–

orchestration

●

= use Docker as a building block

●

re-use other people images (yay ecosystem!)

#lspe

Docker-what?
The big picture
●

Open Source engine to run containers

●

using copy-on-write for quick provisioning

●

allowing to create and share images

●

standard format for containers
(stack of layers; 1 layer = tarball+metadata)

standard, reproducible way to easily build
trusted images (Dockerfile, Stackbrew...)
#lspe
●

Docker-what?
Under the hood
●

rewrite of dotCloud internal container engine
–
–

●

original version: Python, tied to dotCloud's internal stuff
released version: Go, legacy-free

the Docker daemon runs in the background
–
–

HTTP API (over UNIX or TCP socket)

–
●

manages containers, images, and builds
embedded CLI talking to the API

Open Source (GitHub public repository + issue tracking)

#lspe

Docker-what?
The ecosystem
●

Docker Inc. (formerly dotCloud Inc.)
–
–

●

~30 employees, VC-backed
SAAS and support offering around Docker

Docker, the community
–

more than 300 contributors, 1500 forks on GitHub

–

dozens of projects around/on top of Docker

–

x100k trained developers

#lspe

Authoring images
with run/commit

#lspe

1) docker run ubuntu bash
2) apt-get install this and that
3) docker commit <containerid> <imagename>
4) docker run <imagename> bash
5) git clone git://.../mycode
6) pip install -r requirements.txt
7) docker commit <containerid> <imagename>
8) repeat steps 4-7 as necessary
9) docker tag <imagename> <user/image>
10) docker push <user/image>

#lspe

Authoring images
with run/commit
●

Pros
–
–

●

Convenient, nothing to learn
Can roll back/forward if needed

Cons
–

Manual process

–

Iterative changes stack up

–

Full rebuilds are boring, error-prone

#lspe

Authoring images
with a Dockerfile

#lspe

FROM ubuntu
RUN
RUN
RUN
RUN
RUN

apt-get
apt-get
apt-get
apt-get
apt-get

-y update
install -y
install -y
install -y
install -y

g++
erlang-dev erlang-manpages erlang-base-hipe ...
libmozjs185-dev libicu-dev libtool ...
make wget

RUN wget http://.../apache-couchdb-1.3.1.tar.gz | tar -C /tmp -zxfRUN cd /tmp/apache-couchdb-* && ./configure && make install
RUN printf "[httpd]nport = 8101nbind_address = 0.0.0.0" >
/usr/local/etc/couchdb/local.d/docker.ini

EXPOSE 8101
CMD ["/usr/local/bin/couchdb"]

docker build -t jpetazzo/couchdb .

#lspe

Authoring images
with a Dockerfile
●

Minimal learning curve

●

Rebuilds are easy

●

Caching system makes rebuilds faster

●

Single file to define the whole environment!

#lspe

Do you even
Chef?
Puppet?
Ansible?
Salt?
#lspe

Authoring Images
with Chef/Puppet/Ansible/Salt/...
Plan A: « my other VM is a container »
●

write a Dockerfile to install $YOUR_CM

●

start tons of containers

●

run $YOUR_CM in them

Good if you want a mix of containers/VM/metal
But slower to deploy, and uses more resources

#lspe

Authoring Images
with Chef/Puppet/Ansible/Salt/...
Plan B: « the revolution will be containerized »
●

write a Dockerfile to install $YOUR_CM

●

… and run $YOUR_CM as part of build process

●

deploy fully baked images

Faster to deploy
Easier to rollback

#lspe

Install Docker
●

On your servers (Linux)
–
–

Single binary install (Golang FTW!)

–
●

Packages (Ubuntu, Debian, Fedora, Gentoo, Arch...)
Easy provisioning on Rackspace, Digital Ocean, EC2, GCE...

On your dev env (Linux, OS X, Windows)
–

Vagrantfile

–

boot2docker (25 MB VM image)

–

Natively (if you run Linux)

#lspe

Run containers
SSH+CLI
● REST API
● Maestro NG
● Mesos
● OpenStack
● … and more
●

#lspe

Service discovery
●

you can name your containers

●

you can link your containers

docker run -d -name frontdb mysqlimage
docker run -d -link frontdb:sql nginximage

→ environment vars are injected in web container
→ twelve-factors FTW!
#lspe

Allocating CPU and RAM
●

docker run -c $CPU_SHARES -m $RAM_MB

Docker API will soon expose:
●

total CPU/RAM

●

allocated CPU/RAM

●

used CPU/RAM

WARNING: memory metrics are tricky!

#lspe

Memory metrics
●

free
jpetazzo@tarrasque:~$ free
total
used
Mem:
8076564
7019656
-/+ buffers/cache:
4385120
Swap:
0
0

●

this is useless

#lspe

free
1056908
3691444
0

shared
0

buffers
8

cached
2634528

Memory metrics
●

cat /proc/meminfo

●

GRRRREAT

#lspe

MemTotal:
8076564 kB
MemFree:
1055260 kB
Buffers:
8 kB
Cached:
2634528 kB
SwapCached:
0 kB
Active:
3198088 kB
Inactive:
1407516 kB
Active(anon):
1986692 kB
Inactive(anon):
185252 kB
Active(file):
1211396 kB
Inactive(file): 1222264 kB
Unevictable:
0 kB
Mlocked:
0 kB
SwapTotal:
0 kB
SwapFree:
0 kB
Dirty:
144 kB
Writeback:
0 kB
AnonPages:
1971208 kB
Mapped:
2402276 kB
Shmem:
200876 kB
Slab:
148020 kB
SReclaimable:
90680 kB
SUnreclaim:
57340 kB
KernelStack:
4936 kB
PageTables:
43928 kB
NFS_Unstable:
0 kB
Bounce:
0 kB
WritebackTmp:
0 kB
CommitLimit:
4038280 kB
Committed_AS:
8365088 kB
VmallocTotal:
34359738367 kB
VmallocUsed:
357696 kB
VmallocChunk:
34359349628 kB
HardwareCorrupted:
0 kB
AnonHugePages:
0 kB
HugePages_Total:
0
HugePages_Free:
0
HugePages_Rsvd:
0
HugePages_Surp:
0
Hugepagesize:
2048 kB
DirectMap4k:
210384 kB
DirectMap2M:
8056832 kB

Memory metrics
●

●

cat
/sys/fs/cgroup
/memory
/memory.stat

slightly better...

#lspe

cache 2700967936
rss 1977851904
rss_huge 0
mapped_file 275435520
swap 0
pgpgin 104106567
pgpgout 102964277
pgfault 140459553
pgmajfault 3858
inactive_anon 168316928
active_anon 1993658368
inactive_file 1258921984
active_file 1257791488
unevictable 0
hierarchical_memory_limit 9223372036854775807
hierarchical_memsw_limit 9223372036854775807
total_cache 2700967936
total_rss 1977851904
total_rss_huge 0
total_mapped_file 275435520
total_swap 0
total_pgpgin 104106567
total_pgpgout 102964277
total_pgfault 140459553
total_pgmajfault 3858
total_inactive_anon 168316928
total_active_anon 1993658368
total_inactive_file 1258921984
total_active_file 1257791488
total_unevictable 0

Memory metrics
●

●

cat
/sys/fs/cgroup
/memory
/memory.stat

this looks better

#lspe

cache 2700967936
rss 1977851904
rss_huge 0
swap 0
pgpgin 104106567
pgpgout 102964277
pgfault 140459553
pgmajfault 3858
unevictable 0
total_rss_huge 0
total_swap 0
total_unevictable 0

Memory metrics
●

●

cat
/sys/fs/cgroup
/memory
/memory.stat

but you want this

#lspe

cache 2700967936
rss 1977851904
rss_huge 0
swap 0
pgpgin 104106567
pgpgout 102964277
pgfault 140459553
pgmajfault 3858
unevictable 0
total_rss_huge 0
total_swap 0
total_unevictable 0

Large Doses
of
Network Traffic
#lspe

Networking
●

sometimes, you need to expose range of ports

●

or completely arbitary ports

●

or non-IP protocols

●

or you have special needs:
–

more than 1 Gb/s in containers

–

more than 1,000 connections/s

–

more than 100,000 concurrent connections

#lspe

Get rid of the overhead
●

●

●

●

●

use openvswitch
bridge a container directly with a NIC
(remove iptables out of the equation)
move a (macvlan) NIC to a container
(a bit of overhead; multi-tenant)
move a (physical) NIC to a container
(zero overhead; single-tenant)
move a (virtual function) NIC to a container
(if your hardware supports it)

#lspe

Pipework
●
●

sidekick script for Docker
I replaced the plumber with
a very small small shell script

pipework br1 $APACHE 192.168.1.1/24
pipework br1 $MYSQL 192.168.1.2/24
pipework br1 $CONTAINERID 192.168.4.25/20@192.168.4.1
pipework eth2 $(docker run -d hipache) 50.19.169.157
pipework eth3 $(docker run -d hipache) 107.22.140.5
pipework br0 $(docker run -d zmqworker) dhcp fa:de:b0:99:52:1c

#lspe

Thank you! Questions?
http://docker.io/
http://docker.com/
@docker
@jpetazzo

#lspe

Introduction to Docker (and a bit more) at LSPE meetup Sunnyvale

Recommended

More Related Content

What's hot (20)

Viewers also liked (20)

Similar to Introduction to Docker (and a bit more) at LSPE meetup Sunnyvale (20)

More from Jérôme Petazzoni (20)

Recently uploaded (20)

Introduction to Docker (and a bit more) at LSPE meetup Sunnyvale