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-2842) Allow for use of hflush instead of hsync
Date Tue, 17 Jun 2014 15:25:39 GMT

    [ https://issues.apache.org/jira/browse/ACCUMULO-2842?page=com.atlassian.jira.plugin.system.issuetabpanels:comment-tabpanel&focusedCommentId=14033921#comment-14033921
] 

Josh Elser commented on ACCUMULO-2842:
--------------------------------------

Orthogonal for the most part. I believe this will only save you when you're closing the file
(e.g. a WAL), not the incremental syncs that we expect to happen after we have updated entries
in the metadata table, for example. If you have some power outage without using {{hsync}}
before we close a file (which, remember, happens when we hit that max WAL size), I do not
believe {{dfs.datanode.synconclose}} does anything for you.

I think you could say this would be a middle ground between {{hflush}} everything, {{hflush}}
normally but {{(h)sync}} on close, and {{hsync}} everything, but I would want to re-read some
datanode code before you quote that :)

> Allow for use of hflush instead of hsync
> ----------------------------------------
>
>                 Key: ACCUMULO-2842
>                 URL: https://issues.apache.org/jira/browse/ACCUMULO-2842
>             Project: Accumulo
>          Issue Type: Improvement
>          Components: logger
>            Reporter: Josh Elser
>            Assignee: Josh Elser
>             Fix For: 1.5.2, 1.6.1, 1.7.0
>
>
> We keep getting onto the discussion on hsync versus hflush performance; however, we lack
a way to actually test this.
> Make the invocation of hflush or hsync configurable. It should not change the default,
so users on existing major releases should not be affected unless they choose to change their
configuration.



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

Mime
View raw message