hadoop-hdfs-issues mailing list archives

Site index · List index
Message view « Date » · « Thread »
Top « Date » · « Thread »
From "Andy Isaacson (JIRA)" <j...@apache.org>
Subject [jira] [Commented] (HDFS-3640) Don't use Util#now or System#currentTimeMillis for calculating intervals
Date Tue, 17 Jul 2012 18:35:34 GMT

    [ https://issues.apache.org/jira/browse/HDFS-3640?page=com.atlassian.jira.plugin.system.issuetabpanels:comment-tabpanel&focusedCommentId=13416432#comment-13416432
] 

Andy Isaacson commented on HDFS-3640:
-------------------------------------

There is a subtle distinction here, between cases where *timeofday* should be used and cases
where *monotonic* time should be used.  Both have their place in a distributed system.

If a timer is used to compute a delay period as in HDFS-3485, then it's important to use *monotonic*
time.

If a clock is used to generate a time, or a unique identifier, or an opaque token that is
written to a persistent store or sent over the wire to another node (even if in an "opaque"
object), then it's important to use *timeofday*.
                
> Don't use Util#now or System#currentTimeMillis for calculating intervals
> ------------------------------------------------------------------------
>
>                 Key: HDFS-3640
>                 URL: https://issues.apache.org/jira/browse/HDFS-3640
>             Project: Hadoop HDFS
>          Issue Type: Bug
>    Affects Versions: 2.0.0-alpha
>            Reporter: Eli Collins
>            Assignee: Eli Collins
>
> Per HDFS-3485 we shouldn't use Util#now or System#currentTimeMillis to calculate intervals
as they can be affected by system clock changes.

--
This message is automatically generated by JIRA.
If you think it was sent incorrectly, please contact your JIRA administrators: https://issues.apache.org/jira/secure/ContactAdministrators!default.jspa
For more information on JIRA, see: http://www.atlassian.com/software/jira

        

Mime
View raw message