|author||Dom G <firstname.lastname@example.org>||Tue Oct 04 16:39:15 2022 -0400|
|committer||GitHub <email@example.com>||Tue Oct 04 21:39:15 2022 +0100|
Enable port search my default for compactor and sserver (#288)
Uno makes it easy for a developer to experiment with Accumulo or Fluo in a realistic environment. Uno is designed for developers who need to frequently upgrade and test their code, and do not care about preserving data. While Uno makes it easy to setup a dev stack running Fluo or Accumulo, it also makes it easy to clear your data and setup your dev stack again. To avoid inadvertent data loss, Uno should not be used in production.
Checkout Muchos for setting up Accumulo or Fluo on multiple machines.
Uno requires the following software to be installed on your machine.
fetchcommand to download tarballs.
fetchcommand builds tarball from local repo.
You should also be able to ssh to localhost without a passphrase. The following instructions can help you setup these requirements in your environment :
The following commands will get you up and running with an Accumulo instance if you have satisfied the requirements mentioned above. Replace
fluo to setup a Fluo instance.
git clone https://github.com/apache/fluo-uno.git cd fluo-uno ./bin/uno fetch accumulo # Fetches binary tarballs of Accumulo and its dependencies ./bin/uno setup accumulo # Sets up Accumulo and its dependencies (Hadoop & ZooKeeper) source <(./bin/uno env) # Bash-specific command that sets up current shell
Note that the Accumulo shell can be accessed in one of two ways. The easiest is method is to use the
You can also access the shell directly. The Accumulo installation is initialized using the username
root and password
secret (set in the
uno.conf file). Therefore, the shell can be accessed directly using:
accumulo shell -u root -p secret
When you're all done testing out Accumulo you can clean up:
For a more complete understanding of Uno, please continue reading.
First, clone the Uno repo on a local disk with enough space to run Hadoop, Accumulo, etc:
git clone https://github.com/apache/fluo-uno.git
uno command uses
conf/uno.conf for its default configuration which should be sufficient for most users.
Optionally, you can customize this configuration by modifying the
uno.conf file for your environment. Inside this script the variable
UNO_HOME defaults to the root of the Uno repository.
If you would like to avoid modifying
uno.conf because it is managed by git, there is a second way to configure uno. If
conf/uno-local.conf exists then it is used instead of
uno.conf. After pulling the latest changes to Uno, a tool like meld can be used to compare
cp conf/uno.conf conf/uno-local.conf vim conf/uno-local.conf
All commands are run using the
uno script in
bin/. Uno has a command that helps you configure your shell so that you can run commands from any directory and easily set common environment variables in your shell for Uno, Hadoop, ZooKeeper, Fluo, and Spark. Run the following command to print this shell configuration. You can also add
--vars to the command below to limit output to PATH or environment variable configuration:
You can either copy and paste this output into your shell or add the following (with a correct path) to your ~/.bashrc automatically configure every new shell.
source <(/path/to/uno/bin/uno env)
uno script set up, you can now use it to download, configure, and run Fluo's dependencies.
uno fetch <component> command fetches the tarballs of a component and its dependencies for later use by the
setup command. By default, the
fetch command downloads tarballs but you can configure it to build Fluo or Accumulo from a local git repo by setting
uno fetch to see a list of possible components.
fetch command is run for the first time, it only needs to run again if you want to upgrade components and need to download/build the latest version.
uno setup command combines
uno install and
uno run into one command. It will install the downloaded tarballs to the directory set by
$INSTALL in your
uno.conf and run you local development cluster. The command can be run in several different ways:
Sets up Apache Accumulo and its dependencies of Hadoop, ZooKeeper. This starts all processes and will wipe Accumulo/Hadoop if this command was run previously.
uno setup accumulo
Sets up Apache Fluo along with Accumulo (and its dependencies). This command will wipe your cluster. While Fluo is set up, it does not start any Fluo applications.
uno setup fluo
For Fluo & Accumulo, you can setup the software again without wiping/setting up their underlying dependencies. You can upgrade Accumulo or Fluo by running
uno fetch before running this command.
uno setup fluo --no-deps uno setup accumulo --no-deps
You can confirm that everything started by checking the monitoring pages below:
If you run some tests and then want a fresh cluster, run the
setup command again which will kill all running processes, clear any data and logs, and restart your cluster.
Uno is focused on running Accumulo & Fluo. Optional features and service can be run using plugins. These plugins can optionally execute after the
run commands. They are configured by setting
These plugins can optionally execute after the
install command for Accumulo and Fluo:
accumulo-encryption- Turns on Accumulo encryption
influx-metrics- Install and run metrics service using InfluxDB & Grafana
These plugins can optionally execute after the
run command for Accumulo and Fluo:
spark- Install Apache Spark and start Spark's History server
accumulo-proxy- Starts an Accumulo Proxy which enables Accumulo clients in other languages.
uno wipe command will kill all running processes for your local development cluster and clear all the data and logs. It does not delete the binary tarballs downloaded by the
fetch command so you can use
setup directly again in the future. If you need to reclaim the space used by the binary tarballs you'll have to manually delete them.
Before running an Apache Fluo application, it is recommended that you configure your shell using
uno env. If this is done, many Fluo example applications (such as Webindex and Phrasecount) can be run by simply cloning their repo and executing their start scripts (which will use environment variables set in your shell by
If you want to create your own Fluo application, you should mimic the scripts of example Fluo applications or follow the instructions starting at the Configure a Fluo application section of the Fluo install instructions. These instructions will guide you through the process of configuring, initializing, and starting your application.