hadoop-common-dev mailing list archives

Site index · List index
Message view « Date » · « Thread »
Top « Date » · « Thread »
From "Steve Loughran (JIRA)" <j...@apache.org>
Subject [jira] [Reopened] (HADOOP-14736) S3AInputStream to implement an efficient skip() call through seeking
Date Tue, 05 Sep 2017 11:38:01 GMT

     [ https://issues.apache.org/jira/browse/HADOOP-14736?page=com.atlassian.jira.plugin.system.issuetabpanels:all-tabpanel
]

Steve Loughran reopened HADOOP-14736:
-------------------------------------

> S3AInputStream to implement an efficient skip() call through seeking
> --------------------------------------------------------------------
>
>                 Key: HADOOP-14736
>                 URL: https://issues.apache.org/jira/browse/HADOOP-14736
>             Project: Hadoop Common
>          Issue Type: Sub-task
>          Components: fs/s3
>            Reporter: Steve Loughran
>            Priority: Minor
>
> {{S3AInputStream}} implements skip() naively through the byte class: Reading and discarding
all data. Efficient on classic "sequential" reads, provided the forward skip is <1MB. For
larger skip values or on random IO, seek() should be used.
> After some range checks/handling past-EOF skips to seek (EOF-1), a seek() should handle
the skip file.
> *there are no FS contract tests for skip semantics*



--
This message was sent by Atlassian JIRA
(v6.4.14#64029)

---------------------------------------------------------------------
To unsubscribe, e-mail: common-dev-unsubscribe@hadoop.apache.org
For additional commands, e-mail: common-dev-help@hadoop.apache.org


Mime
View raw message