Any word on Cassandra storage backend? - Google Groups

Any word on Cassandra storage backend? - Google Groups

Debugging the PIO Framework is different than the Templates. If you are primarily working on a Cassandra version of the backend you may not need a template until you want to test reading and there are ways to see what is in the EventServer (read) without a template.

I have done template dev and never have to build from the CLI except when the config engine.json changes. IDEA does the incremental compile when code changes.

That said, using IDEA on the Framework is new to me but I'm planning some PRs to do with the storage API soon and will require getting IDEA setup. 

1) the PIO sbt is really only required for the install script for the first time, thereafter you can use your own installation of sbt.

2) let's try it and solve problems as they some up. Seems like most PIO folks don't use a debugger as a matter of preference. Judging from template debugging, it may take some work but all things are possible :-) To try out a new backing store it may be you only need to mess with the EventServer by building and restarting. The CLI for starting the EventServer can be mimicked with a debug config in IDEA. Once set up the incremental compile will be triggered automatically. We should get some guidance from others here.

3) Speaking for myself unit tests are great if they can be done at compile time and don't require a running system, which is where scripted integration tests come in.

Using Hbase as a starting point is ok if you are alright with requiring Elasticsearch for PIO "metadata". The massively scalable architecture for PIO uses HBase + ES. Other templates run fine with a single store for metadata and event data. If you want a single backing store for metadata and events you may want to look at the JDBC backend since it does both. Cassandra should be able handle both and scale indefinitely.

Read full article from Any word on Cassandra storage backend? - Google Groups