[
https://issues.apache.org/jira/browse/ACCUMULO-1378?page=com.atlassian.jira.plugin.system.issuetabpanels:comment-tabpanel&focusedCommentId=13649790#comment-13649790
]
Keith Turner commented on ACCUMULO-1378:
----------------------------------------
One purpose of MAC is to replace MockAccumulo for use in unit testing.
However, its a lot slower than mock even though its behavior is much more
accurate. A possible future goal for MAC would be to speed it up so that unit
test run faster. However making MAC highly configurable (and exposing its
implementation) would give us less room to manuver when trying to speed it up.
I am thinking if we expose as little as possible in the MAC public API it will
give us a lot of flexibiltiy in the future to drastically change its
implementation. Maybe it would be better to create something else for the
purpose of this ticket, that is similar to MAC. Something like
SingleNodeAccumulo.java, that is configurable and the user can make a lot more
assumptions about.
This makes me think we may want to examine the 1.4 and 1.5 APIs of MAC and see
if they can be simplified. But I am still thinking about this. Making MAC
faster in 1.6 would be really nice.
> Allow MiniAccumulo cluster configured directory to be used across different
> runs
> --------------------------------------------------------------------------------
>
> Key: ACCUMULO-1378
> URL: https://issues.apache.org/jira/browse/ACCUMULO-1378
> Project: Accumulo
> Issue Type: Improvement
> Reporter: Corey J. Nolet
> Assignee: Corey J. Nolet
> Priority: Minor
>
> Currently, the MiniAccumuloCluster assumes it should always create the
> directory structure and throws an exception if the passed in directory has
> anything in it.
--
This message is automatically generated by JIRA.
If you think it was sent incorrectly, please contact your JIRA administrators
For more information on JIRA, see: http://www.atlassian.com/software/jira
