I would like to modify, creat code based on the examples in accumulo-1.4.3 accumulo QuickStart.
I understand I can import using maven pom.xml for the examples.
Is there any documentation on how I can deploy the code as jars, ect that I build?
Is it possible to real-time debugging with breakpoints etc?
thnaks for your assistance
If you use the binary + source distribution (which is what you'll get when you do an installation from tarballs according to our release documentation), then when you invoke maven with the package goal you should get updated jars in the Accumulo lib direcotry. Those jars should then be used after Accumulo restarts.
In general, when creating a client that interacts with Accumulo through either MapReduce or the rest of the client API you do not need to do special deployment of jars for either Accumulo or Hadoop. Instead, you'd need to make sure your application had the correct classpath and run from wherever you choose to deploy the application jar.
Accumulo comes with a script to ease setting correct classpaths (both for standalone apps and when submitting MapReduce jobs). You'll find it at $ACCUMULO_HOME/bin/tool.sh
[accumulo@localhost ~]$ $ACCUMULO_HOME/bin/tool.sh Usage: tool.sh path/to/myTool.jar my.tool.class.Name [-libjars my1.jar,my2.jar] [accumulo@localhost ~]$
The only time this isn't sufficient is if you need to deploy a custom iterator that will be used server side. Due to the instability such iterators can introduce, we generally don't recommend customers make use of them unless there's no other way to implement a use case.
If you want to do debugging with breakpoints, etc, especially in an IDE, you'll have to make use of the MiniAccumuloCluster (or MAC). The MAC is not available in our initial release of Accumulo 1.4.3-cdh4.3.0. It is available starting in our more resent release of Accumulo 1.4.4-cdh4.5.0. Using it will only be useful with code you write outside of the main Accumulo codebase, and it is covered very briefly in section 5.2 of the user manual.