docker-library
/
docs
mirror of https://github.com/docker-library/docs.git
1
0
Fork 0
Code Issues Packages Projects Releases Wiki Activity

Better documentation

This commit is contained in:
Elisey Zanko 2016-05-02 17:56:02 +05:00
parent a9be8c8086
commit a65f571914
1 changed files with 36 additions and 10 deletions
Show Stats Download Patch File Download Diff File Expand all files Collapse all files

46
storm/content.md
View File

@ -8,22 +8,48 @@ Apache Storm is a distributed computation framework written predominantly in the
# How to use this image
## Setting up a Storm Cluster
This image entry point is a [Storm command line client](http://storm.apache.org/releases/1.0.0/Command-line-client.html). See couple of usage examples below.
1. Start Zookeeper
## Running topologies in local mode
$ docker run -d --name zookeeper jplock/zookeeper:3.4.8
Assuming you have `storm-starter-topologies-1.0.0.jar` in the current directory.
2. Start Nimbus
$ docker run -it -v $(pwd)/storm-starter-topologies-1.0.0.jar:/topology.jar 31z4/storm:1.0.0 jar /topology.jar org.apache.storm.starter.ExclamationTopology
$ docker run -d --name nimbus --net container:zookeeper 31z4/storm:1.0.0 nimbus
## Setting up a minimal Storm cluster
3. Start Supervisor
1. [Apache Zookeeper](https://zookeeper.apache.org/) is a must for running a Storm cluster. Start it first. Since the Zookeeper "fails fast" it's better to always restart it.
$ docker run -d --name supervisor --net container:nimbus 31z4/storm:1.0.0 supervisor
$ docker run -d --restart always --name zookeeper jplock/zookeeper:3.4.8
## Running Topologies
2. The Nimbus daemon has to be connected with the Zookeeper. It's also a "fail fast" system.
Assuming you set up a Storm cluster as written above and have `storm-starter-topologies-1.0.0.jar` in your current directory.
$ docker run -d --restart always --name nimbus --net container:zookeeper 31z4/storm:1.0.0 nimbus
$ docker run -it --net container:nimbus -v $(pwd)/storm-starter-topologies-1.0.0.jar:/topology.jar storm:1.0.0 jar /topology.jar org.apache.storm.starter.WordCountTopology topology
3. Finally start a single Supervisor node. It will talk to the Nimbus and Zookeeper.
$ docker run -d --restart always --name supervisor --net container:nimbus --net container:zookeeper 31z4/storm:1.0.0 supervisor
4. Now we can submit a topology to our cluster.
$ docker run -it --net container:nimbus -v $(pwd)/storm-starter-topologies-1.0.0.jar:/topology.jar 31z4/storm:1.0.0 jar /topology.jar org.apache.storm.starter.WordCountTopology topology
## Configuration
This image uses [default configuration](https://github.com/apache/storm/blob/v1.0.0/conf/defaults.yaml) of the Apache Storm. There are two main ways to change it.
1. Using command line arguments.
$ docker run -d --restart always --name nimbus 31z4/storm:1.0.0 nimbus -c storm.log.dir="/logs" -c storm.zookeeper.servers="[\"zookeeper\"]"
2. Assuming you have `storm.yaml` in the current directory you can mount it as a volume.
$ docker run -it -v $(pwd)/storm.yaml:/apache-storm-1.0.0/conf/storm.yaml 31z4/storm:1.0.0 nimbus
## Data persistence
No data are persisted by default. For convenience there are `/data` and `/logs` directories in the image owned by `storm` user. Use them accordingly to persist data and logs using volumes.
$ docker run -it -v /logs -v /data 31z4/storm:1.0.0 nimbus -c storm.log.dir="/logs" -c storm.local.dir="/data"
*Please be noticed that using paths other than those predefined is likely to cause permission denied errors. It's because for [security reasons](https://docs.docker.com/engine/userguide/eng-image/dockerfile_best-practices/#user) the Storm is running under the non-root `storm` user.*