accumulo-notifications mailing list archives

Site index · List index
Message view « Date » · « Thread »
Top « Date » · « Thread »
From "Josh Elser (JIRA)" <j...@apache.org>
Subject [jira] [Commented] (ACCUMULO-2739) Create an In-Process MiniAccumuloCluster
Date Fri, 30 May 2014 03:26:02 GMT

    [ https://issues.apache.org/jira/browse/ACCUMULO-2739?page=com.atlassian.jira.plugin.system.issuetabpanels:comment-tabpanel&focusedCommentId=14013253#comment-14013253
] 

Josh Elser commented on ACCUMULO-2739:
--------------------------------------

I did a bit more work and got things generally working (at a very high level).

https://github.com/joshelser/accumulo/tree/ACCUMULO-2739-microcluster

Turns out that there might be a bigger issue of stopping tserver and master _cleanly_ because
of {{UtilWaitThread}} eating the InterruptedException (I think). Not a big problem unless
the user wants to continue running the same JVM after stopping the microcluster...

> Create an In-Process MiniAccumuloCluster
> ----------------------------------------
>
>                 Key: ACCUMULO-2739
>                 URL: https://issues.apache.org/jira/browse/ACCUMULO-2739
>             Project: Accumulo
>          Issue Type: New Feature
>          Components: mini
>            Reporter: Josh Elser
>            Assignee: Josh Elser
>             Fix For: 1.7.0
>
>
> Two common griefs about MiniAccumuloCluster is that it's slow to start and it is difficult
to debug things running inside of the Master or TabletServer (because they are in separate
processes).
> Running each Accumulo process inside of the parent process would likely increase startup
speed and would make debugging much easier (as you wouldn't have to remote-attach the debugger
anymore).
> Such a MAC would likely also pave the way for easier migration away from MockAccumulo
as both of the above complaints are some of the primary concerns for removing Mock with no
replacemnt.



--
This message was sent by Atlassian JIRA
(v6.2#6252)

Mime
View raw message