accumulo-dev mailing list archives

Site index · List index
Message view « Date » · « Thread »
Top « Date » · « Thread »
From Josh Elser <josh.el...@gmail.com>
Subject Notes from a local co-working day
Date Fri, 04 Mar 2016 00:07:42 GMT
Hi all,

A group of us in central MD met today at a local business (which was 
gracious enough to donate a conference room today) and enjoy each 
others' presence while hacking on some code.

In the spirit of no private discussions, I made sure to keep a record of 
the relevant Accumulo talking points we had. We had myself, Billie, 
Keith, Mike Walch, Mike Wall, Christopher, and Ed "Avocado" Coleman (he 
asked for that one) in attendance.

# Where does 1.8 stand?

I made a quick pass over the 1.8 JIRA issues trying to write down what 
we've already done and what is outstanding. The highlights mentioned:

* Tablet locator public API
* Atomic add/remove authorizations API for Accumulo users
* Automated Accumulo Iterator test harness
* Upgrade to Thrift 0.9.3
* Ports for services are moved out of the Linux ephemeral port range 
(the monitor 50095 -> 9995)
* Internal WAL state-tracking changes (Eric's uber change about how we 
record WALs in the metadata table).
* Introduction of a daemon script for Accumulo services and pidfile 
support (!)
* Scanner context-classpath feature from Dave (not sure details)
* Per-table Sampling from Keith (needs a test according to him)

There were a couple of things mentioned that people want to try to also 
get into a 1.8, identified by "owner"

* Christopher: no-chop merges
* Keith: A public API for reading/writing RFiles
* Josh: Finish monitor improvements (not sure about this one, it really 
fell off my radar)
* Mike Wall: Split out examples from "core" repository.

# RFile read/write public API

Just before I stepped out, Keith was giving a brief explanation of a new 
API that he's been working on to read and write RFiles directly.

https://github.com/keith-turner/accumulo/tree/rfile-api

He mentioned that we should expect some JIRA issue/pull-request from him 
soon.

# Incubator discussion

We also had a general discussion about how the incubator process at the 
ASF works. Billie had some great information to pass along, and I tried 
to chime in now and again with things I've learned :)

- Josh

Mime
View raw message