Fix hbase client's write bug in debug mode.

Hbase client has a bug to write all values as empty byte arrays in debug
mode. It is caused by an illegal use of ByteIterator. Values returned
from toString() or toArray() should be cached if they are needed later.

Use the clear buffer on fail option. Not using it is now deprecated.

Synchronize Measurements.getSummary() method to prevent ConcurrentModificationExceptions

Add an HBase 1.0 client.

Remove mapkeeper so that build passes

* Bumps `hbase.version` to 1.0.0
* Adds a new database, hbase-10, which targets the HBase 1.0 API.

Update README.md

Fix typo in redis/README.md

Add Accumulo as a backing data store

* Increment Accumulo Version to 1.6.0
* Fix accumulo dependency versions
  - Harmonize guava version for Accumulo.
  - Automatically manage slf4j version.
* Clean up depr. code
* Clean up return codes and error logging

Make the batchwriter's number of threads configurable.

clean up dead code in client.

* set .gitignore to standard eclipse
* update versions of Accumulo to 1.4.4
* Pin slf4j binding to 1.5.8

Accumulo, Thrift, and YCSB all use different versions of slf4j, causing
runtime issues because everything is bundled together. Pin the version
to something in the middle that seems to work, even if it's not the
latest and greatest version.

Based on YCSB++

http://www.pdl.cmu.edu/ycsb++/

[core] Add timestamp in ycsb qps and latency log 

[gemfire] Fix NullPointException when execute "ycsb run gemfire ..."

[redis] add README explaining how to run redid load tests

provides TravisCI configuration files 

ignore all target folder contents

[mongo] Update the MongoDB Java Driver version to 2.11.2.

JDBC scans did not instruct the driver to use a specific fetchsize,
resulting in most cases in an OOM; add support for setting fetchsize.

Also add support for disabling autocommit on the JDBC connections,
mostly because Postgres for example doesn't support setting fetchsize
without disabling autocommit.