Ce diaporama a bien été signalé.
Nous utilisons votre profil LinkedIn et vos données d’activité pour vous proposer des publicités personnalisées et pertinentes. Vous pouvez changer vos préférences de publicités à tout moment.
Separation
of Concerns
with Docker
and Containers
1 / 60
Who am I?
Jérôme Petazzoni (@jpetazzo)
French software engineer living in California
Joined Docker (dotCloud) more than 4 ...
Outline
Intro / pop quizz about Docker and containers
VMs and containers: technical differences
VMs and containers: functi...
Recap
about Docker
and containers
4 / 60
Build, ship, and run
any app, anywhere
5 / 60
Take any Linux program, and put it in a container
6 / 60
Take any Linux program, and put it in a container
Web apps and services, workers
(Go, Java, Node, PHP, Python, Ruby...)
7 ...
Take any Linux program, and put it in a container
Web apps and services, workers
(Go, Java, Node, PHP, Python, Ruby...)
Da...
Take any Linux program, and put it in a container
Web apps and services, workers
(Go, Java, Node, PHP, Python, Ruby...)
Da...
Take any Linux program, and put it in a container
Web apps and services, workers
(Go, Java, Node, PHP, Python, Ruby...)
Da...
Take any Linux program, and put it in a container
Web apps and services, workers
(Go, Java, Node, PHP, Python, Ruby...)
Da...
What about non-Linux programs?
12 / 60
What about non-Linux programs?
Desktop apps with WINE
(e.g.: Spotify client)
13 / 60
What about non-Linux programs?
Desktop apps with WINE
(e.g.: Spotify client)
Coming soon: Docker for Windows
(run Windows ...
What about non-Linux programs?
Desktop apps with WINE
(e.g.: Spotify client)
Coming soon: Docker for Windows
(run Windows ...
What about non-Linux programs?
Desktop apps with WINE
(e.g.: Spotify client)
Coming soon: Docker for Windows
(run Windows ...
Ship that container easily and efficiently
Docker comes with an image distribution protocol
Distribution server can be hos...
Run those containers anywhere
Containers can run in VMs or in physical machines
Docker is available on all modern Linux va...
VMs and
containers:
technical
differences
19 / 60
Portability
Containers can run on top of public cloud
(run the same container image everywhere)
Nested hypervisors (VMs in...
Format & environment
VM
executes machine code
environment = something that looks like a computer
JVM
executes JVM bytecode...
Containers have low overhead
Normal* process(es) running on top of normal kernel
No device emulation (no extra code path i...
VMs have stronger isolation
Inter-VM communication must happen over the network
(Some hypervisors have custom paths, but n...
Analogy: brick walls vs. room dividers
Brick walls
sturdy
slow to build
messy to move
Room dividers
fragile
deployed in se...
Blurring lines
Intel Clear Containers; Clever Cloud
(stripped down VMs, boot super fast, tiny footprint)
Joyent Triton
(So...
VMs and
containers:
functional
differences
26 / 60
Inside
VMs need a full OS and associated tools
(Backups, logging, periodic job execution, remote access...)
Containers can...
VM lifecycle
Option 1: long lifecycle
(provisioning→update→update→…→update→disposal)
easily leads to configuration drift
(...
Container lifecycle
Containers are created from an image
Image creation is easy
Image upgrade is fast
Immutable infrastruc...
Development process (VMs)
Best practice in production = 1 VM per component
Not realistic to have 1 VM per component in dev...
Development process (containers)
Run tons of containers on dev machines
Build the same container for dev & prod
How do we ...
Bloated containers
Containers have all the software required for production
In dev mode, only essential processes are star...
Lean containers
33 / 60
Principle
"Do one thing, do it well"
One container for the component itself
One container for logging
One container for mo...
Implementation (general principles)
Containers can share almost anything, selectively
files
(logs, data at rest, audit)
ne...
Let's dive
into the details
36 / 60
Logging (option 1: Docker logging drivers)
Containers write to standard output
Docker has different logging drivers:
write...
Logging (option 2: shared log directory)
Containers write regular files to a directory
That directory is shared with anoth...
Logging takeaways
Application can be "dumb" about logging
Log collection and shipping happens in Docker,
or in separate(s)...
"Yes, but..."
"What about performance overhead?"
no performance overhead
both containers access files directly
(just like ...
Backups (file-based)
Store mutable data on Docker volumes
(same mechanism as for logs)
Share volumes with special-purpose ...
Backups (network-based)
Run the backup job (pg_dump, mysqldump, etc.)
from a separate container
Advantages (vs. running in...
Network analysis
Packet capture (tcpdump, ngrep, ntop, etc.)
Low-level metrics (netstat, ss, etc.)
Install required tools ...
Service discovery
Docker can do linking and generic DNS injection
Your code connects to e.g. redis
(pretending that redisr...
Service discovery in practice
When service A needs to talk to service B...
1. Start container B on a Docker host
45 / 60
Service discovery in practice
When service A needs to talk to service B...
1. Start container B on a Docker host
2. Retrie...
Service discovery in practice
When service A needs to talk to service B...
1. Start container B on a Docker host
2. Retrie...
Service discovery in practice
When service A needs to talk to service B...
1. Start container B on a Docker host
2. Retrie...
Service discovery in practice
When service A needs to talk to service B...
1. Start container B on a Docker host
2. Retrie...
General pattern
Your code runs in the same container in dev and prod
Add "sidekick*" containers for additional tasks
Devel...
Composing
stacks of
containers
51 / 60
Docker Compose
52 / 60
docker-compose.yml
rng:
build:rng
hasher:
build:hasher
webui:
build:webui
links:
-redis
ports:
-"80:80"
volumes:
-"webui/f...
Docker Compose
Start whole stack with docker-composeup
Start individual containers (and their dependencies)
with docker-co...
Docker Compose
Start whole stack with docker-composeup
Start individual containers (and their dependencies)
with docker-co...
docker-compose.yml, reloaded
hasher:
build:hasher
worker:
build:worker
links:
-rng
-hasherproxy:hasher
-redis
hasherproxy:...
Heads up!
Docker networking is evolving quickly
Docker 1.7 (to be released Real Soon Now) will support:
"networks" as firs...
Conclusions
58 / 60
Conclusions
Containers can share more context than VMs
We can use this to decouple complexity
(think "microservices" but f...
Thanks!
Questions?
@jpetazzo
@docker
60 / 60
Prochain SlideShare
Chargement dans…5
×

Implementing Separation of Concerns with Docker and Containers

11 379 vues

Publié le

One of the promises of containers is to offer "separation of concerns" between Devs and Ops. Devs are supposedly able to put their code in containers, run those containers locally, and ship them to the Ops team. When deployed on other platforms (from test to staging to QA to production), those containers will behave exactly the same way as they did in development.

But what should we do when our production platform uses a totally different logging system? How should we implement backups or metrics collection in our production containers, without bloating our development containers? In development, it's easy to use a mechanism like Docker Compose to bring up multiple containers together, but how does that work in production stacks?

Jérôme will present many techniques to answer those questions (and a few more), and show how our application containers can remain simple and lean, yet have all the indispensable bells and whistles that are required to run in stable production environments.

Publié dans : Technologie
  • Soyez le premier à commenter

Implementing Separation of Concerns with Docker and Containers

  1. 1. Separation of Concerns with Docker and Containers 1 / 60
  2. 2. Who am I? Jérôme Petazzoni (@jpetazzo) French software engineer living in California Joined Docker (dotCloud) more than 4 years ago (I was at Docker before it was cool!) I have built and scaled the dotCloud PaaS I learned a few things about running containers (in production) 2 / 60
  3. 3. Outline Intro / pop quizz about Docker and containers VMs and containers: technical differences VMs and containers: functional differences Our goal: lean containers Composing stacks of containers 3 / 60
  4. 4. Recap about Docker and containers 4 / 60
  5. 5. Build, ship, and run any app, anywhere 5 / 60
  6. 6. Take any Linux program, and put it in a container 6 / 60
  7. 7. Take any Linux program, and put it in a container Web apps and services, workers (Go, Java, Node, PHP, Python, Ruby...) 7 / 60
  8. 8. Take any Linux program, and put it in a container Web apps and services, workers (Go, Java, Node, PHP, Python, Ruby...) Data stores: SQL, NoSQL, big data (Cassandra, ElasticSearch, Hadoop, Mongo, MySQL, PostgreSQL, Redis...) 8 / 60
  9. 9. Take any Linux program, and put it in a container Web apps and services, workers (Go, Java, Node, PHP, Python, Ruby...) Data stores: SQL, NoSQL, big data (Cassandra, ElasticSearch, Hadoop, Mongo, MySQL, PostgreSQL, Redis...) Other server-y things (Consul, Etcd, Mesos, RabbitMQ, Zookeeper...) 9 / 60
  10. 10. Take any Linux program, and put it in a container Web apps and services, workers (Go, Java, Node, PHP, Python, Ruby...) Data stores: SQL, NoSQL, big data (Cassandra, ElasticSearch, Hadoop, Mongo, MySQL, PostgreSQL, Redis...) Other server-y things (Consul, Etcd, Mesos, RabbitMQ, Zookeeper...) Command-line tools (AWS CLI, Ffmpeg...) 10 / 60
  11. 11. Take any Linux program, and put it in a container Web apps and services, workers (Go, Java, Node, PHP, Python, Ruby...) Data stores: SQL, NoSQL, big data (Cassandra, ElasticSearch, Hadoop, Mongo, MySQL, PostgreSQL, Redis...) Other server-y things (Consul, Etcd, Mesos, RabbitMQ, Zookeeper...) Command-line tools (AWS CLI, Ffmpeg...) Desktop apps (Chrome, LibreOffice, Skype, Steam...) 11 / 60
  12. 12. What about non-Linux programs? 12 / 60
  13. 13. What about non-Linux programs? Desktop apps with WINE (e.g.: Spotify client) 13 / 60
  14. 14. What about non-Linux programs? Desktop apps with WINE (e.g.: Spotify client) Coming soon: Docker for Windows (run Windows apps on Windows machines) 14 / 60
  15. 15. What about non-Linux programs? Desktop apps with WINE (e.g.: Spotify client) Coming soon: Docker for Windows (run Windows apps on Windows machines) Coming soon: Docker for FreeBSD (port in progress) 15 / 60
  16. 16. What about non-Linux programs? Desktop apps with WINE (e.g.: Spotify client) Coming soon: Docker for Windows (run Windows apps on Windows machines) Coming soon: Docker for FreeBSD (port in progress) Coming eventually: Docker for OS X (technically possible; but is this useful?) 16 / 60
  17. 17. Ship that container easily and efficiently Docker comes with an image distribution protocol Distribution server can be hosted by Docker Inc. (free for public images) Distribution protocol is public Open source reference implementation (used by Docker Inc. for the public registry) Container images are broken down into layers When updating and distributing an image, only ship relevant layers 17 / 60
  18. 18. Run those containers anywhere Containers can run in VMs or in physical machines Docker is available on all modern Linux variants Many IAAS providers have server images with Docker On OS X and Windows dev machines: boot2docker There are distros dedicated to run Docker containers (Atomic, CoreOS, RancherOS, Snappy Core...) Other Docker implementations exist (e.g. Joyent Triton) Docker-as-a-Service providers are blooming 18 / 60
  19. 19. VMs and containers: technical differences 19 / 60
  20. 20. Portability Containers can run on top of public cloud (run the same container image everywhere) Nested hypervisors (VMs in VMs) exist, but still rare Containers are easy to move (thanks to layers, distribution protocol, registry...) VM images have to be converted and transferred (both are slow operations) 20 / 60
  21. 21. Format & environment VM executes machine code environment = something that looks like a computer JVM executes JVM bytecode environment = Java APIs Container executes machine code environment = Linux kernel system calls interface 21 / 60
  22. 22. Containers have low overhead Normal* process(es) running on top of normal kernel No device emulation (no extra code path involved in I/O) Context switch between containers = context switch between processes Benchmarks show no difference at all between containers and bare metal (after adequate tuning and options have been selected) Containers have higher density * There are extra "labels" denoting membership to given namespaces and control groups. Similar to regular UID. 22 / 60
  23. 23. VMs have stronger isolation Inter-VM communication must happen over the network (Some hypervisors have custom paths, but non-standard) VMs can run as non-privileged processes on the host (Breaking out of a VM will have ~zero security impact) Containers run on top of a single kernel (Kernel vulnerability can lead to full scale compromise) Containers can share files, sockets, FIFOs, memory areas... (They can communicate with standard UNIX mechanisms) 23 / 60
  24. 24. Analogy: brick walls vs. room dividers Brick walls sturdy slow to build messy to move Room dividers fragile deployed in seconds moved easily 24 / 60
  25. 25. Blurring lines Intel Clear Containers; Clever Cloud (stripped down VMs, boot super fast, tiny footprint) Joyent Triton (Solaris "branded zones," running Linux binaries securely, exposing the Docker API) Ongoing efforts to harden containers (GRSEC, SELinux, AppArmor) 25 / 60
  26. 26. VMs and containers: functional differences 26 / 60
  27. 27. Inside VMs need a full OS and associated tools (Backups, logging, periodic job execution, remote access...) Containers can go both ways: machine container (runs init, cron, ssh, syslog ... and the app) application container (runs the app and nothing else; relies on external mechanisms) 27 / 60
  28. 28. VM lifecycle Option 1: long lifecycle (provisioning→update→update→…→update→disposal) easily leads to configuration drift (subtle differences that add up over time) requires tight configuration management Option 2: golden images (phoenix servers, immutable infrastructure ...) create new image for each modification deploy by replacing old servers with new servers nice and clean, but heavy and complex to setup 28 / 60
  29. 29. Container lifecycle Containers are created from an image Image creation is easy Image upgrade is fast Immutable infrastructure is easy to implement Why? Because container snapshots are extremely fast and cheap. 29 / 60
  30. 30. Development process (VMs) Best practice in production = 1 VM per component Not realistic to have 1 VM per component in dev Also: prod has additional/different components (e.g.: logging, monitoring, service discovery...) Result: very different environment for dev & prod 30 / 60
  31. 31. Development process (containers) Run tons of containers on dev machines Build the same container for dev & prod How do we provide container variants? 31 / 60
  32. 32. Bloated containers Containers have all the software required for production In dev mode, only essential processes are started In prod mode, additional processes run as well Problems: bigger containers behavior can differ (because of extra processes) extra processes duplicated between containers hard to test those extra processes in isolation 32 / 60
  33. 33. Lean containers 33 / 60
  34. 34. Principle "Do one thing, do it well" One container for the component itself One container for logging One container for monitoring One container for backups One container for debugging (when needed) etc. 34 / 60
  35. 35. Implementation (general principles) Containers can share almost anything, selectively files (logs, data at rest, audit) network stack (traffic routing and analysis, monitoring) process space, memory (process tracing and debugging) 35 / 60
  36. 36. Let's dive into the details 36 / 60
  37. 37. Logging (option 1: Docker logging drivers) Containers write to standard output Docker has different logging drivers: writes to local JSON files by default can send to syslog Imperfect solution for now, but will be improved. Preferred in the long run. 37 / 60
  38. 38. Logging (option 2: shared log directory) Containers write regular files to a directory That directory is shared with another container dockerrun-d--namemyapp1-v/var/logmyapp:v1.0 In development setup: dockerrun--volumes-frommyapp1ubuntu sh-c'tail-F/var/log/*' In production: dockerrun-d--volumes-frommyapp1logcollector 38 / 60
  39. 39. Logging takeaways Application can be "dumb" about logging Log collection and shipping happens in Docker, or in separate(s) container(s) Run custom log analyzer without changing app container (e.g. apachetop) Migrate logging system without changing app container 39 / 60
  40. 40. "Yes, but..." "What about performance overhead?" no performance overhead both containers access files directly (just like processes running on the same machine) "What about synchronization issues?" same as previous answer! 40 / 60
  41. 41. Backups (file-based) Store mutable data on Docker volumes (same mechanism as for logs) Share volumes with special-purpose backup containers Put backup tools in the backup container (boto, rsync, s3cmd, unison...) dockerrun--volumes-frommydb1ubuntu rsync-av/var/lib/backup@remotehost:mydb1/ The whole setup doesn't touch the app (or DB) container 41 / 60
  42. 42. Backups (network-based) Run the backup job (pg_dump, mysqldump, etc.) from a separate container Advantages (vs. running in the same container): nothing to install in the app (or DB) container if the backup job runs amok, it remains contained (!) another team can maintain backup jobs (and be responsible for them) 42 / 60
  43. 43. Network analysis Packet capture (tcpdump, ngrep, ntop, etc.) Low-level metrics (netstat, ss, etc.) Install required tools in a separate container image Run a container in the same network namespace dockerrun-d--nameweb1nginx dockerrun-ti--netcontainer:web1tcpdump-pnieth0 dockerrun-ti--netcontainer:web1ubuntuss-n--tcp 43 / 60
  44. 44. Service discovery Docker can do linking and generic DNS injection Your code connects to e.g. redis (pretending that redisresolves to something) Docker adds a DNS alias* so that redisresolves to the right container, or to some external service In dev, Docker Compose manages service dependencies In prod, you abstract service discovery from the container * Really, an entry in the container's /etc/hosts. 44 / 60
  45. 45. Service discovery in practice When service A needs to talk to service B... 1. Start container B on a Docker host 45 / 60
  46. 46. Service discovery in practice When service A needs to talk to service B... 1. Start container B on a Docker host 2. Retrieve host+port allocated for B 46 / 60
  47. 47. Service discovery in practice When service A needs to talk to service B... 1. Start container B on a Docker host 2. Retrieve host+port allocated for B 3. Start ambassador (relaying to this host+port) 47 / 60
  48. 48. Service discovery in practice When service A needs to talk to service B... 1. Start container B on a Docker host 2. Retrieve host+port allocated for B 3. Start ambassador (relaying to this host+port) 4. Start container A linked to ambassador 48 / 60
  49. 49. Service discovery in practice When service A needs to talk to service B... 1. Start container B on a Docker host 2. Retrieve host+port allocated for B 3. Start ambassador (relaying to this host+port) 4. Start container A linked to ambassador 5. Profit! 49 / 60
  50. 50. General pattern Your code runs in the same container in dev and prod Add "sidekick*" containers for additional tasks Developers don't have to be bothered about ops Ops can do their job without messing with devs' code * Kubernetes sometimes calls them "sidecars." 50 / 60
  51. 51. Composing stacks of containers 51 / 60
  52. 52. Docker Compose 52 / 60
  53. 53. docker-compose.yml rng: build:rng hasher: build:hasher webui: build:webui links: -redis ports: -"80:80" volumes: -"webui/files/:/files/" redis: image:redis worker: build:worker links: -rng -hasher -redis 53 / 60
  54. 54. Docker Compose Start whole stack with docker-composeup Start individual containers (and their dependencies) with docker-composeupxyz Takes care of container lifecycle (creation, update, data persistence, scaling up/down...) Doesn't automatically solve networking and discovery (yet) 54 / 60
  55. 55. Docker Compose Start whole stack with docker-composeup Start individual containers (and their dependencies) with docker-composeupxyz Takes care of container lifecycle (creation, update, data persistence, scaling up/down...) Doesn't automatically solve networking and discovery (yet) ... However ... 55 / 60
  56. 56. docker-compose.yml, reloaded hasher: build:hasher worker: build:worker links: -rng -hasherproxy:hasher -redis hasherproxy: image:jpetazzo/hamba links: -hasher command:80hasher80 (This was automatically generated by a tiny Python script.) 56 / 60
  57. 57. Heads up! Docker networking is evolving quickly Docker 1.7 (to be released Real Soon Now) will support: "networks" as first class objects multiple networks overlay driver allowing to span networks across multiple hosts networking plugins from ecosystem partners 57 / 60
  58. 58. Conclusions 58 / 60
  59. 59. Conclusions Containers can share more context than VMs We can use this to decouple complexity (think "microservices" but for ops/devs separation) All tasks typically requiring VM access can be done in separate containers As a result, deployments are broken down in smaller, simpler pieces Complex stacks are expressed with simple YAML files Docker isn't a "silver bullet" to solve all problems, but it gives us tools that make our jobs easier 59 / 60
  60. 60. Thanks! Questions? @jpetazzo @docker 60 / 60

×