You are viewing a plain text version of this content. The canonical link for it is here.
Posted to common-issues@hadoop.apache.org by "Steve Loughran (Jira)" <ji...@apache.org> on 2020/01/03 14:53:00 UTC
[jira] [Commented] (HADOOP-16786) SequenceFile.Writer#hsync does
not sync to disk when checksums are used
[ https://issues.apache.org/jira/browse/HADOOP-16786?page=com.atlassian.jira.plugin.system.issuetabpanels:comment-tabpanel&focusedCommentId=17007538#comment-17007538 ]
Steve Loughran commented on HADOOP-16786:
-----------------------------------------
yeah, it's a known issue. I should make sure that the docs in HADOOP-13327 cover this "detail"
FWIW, calling hsync all the time has adverse consequences on HDFS, object stores etc. Only use it when you absolutely need to
> SequenceFile.Writer#hsync does not sync to disk when checksums are used
> -----------------------------------------------------------------------
>
> Key: HADOOP-16786
> URL: https://issues.apache.org/jira/browse/HADOOP-16786
> Project: Hadoop Common
> Issue Type: Bug
> Reporter: Mathias Herberts
> Priority: Major
>
> Recently I created a SequenceFile Writer and was surprised to see that the file size was not changing despite hsync being called after each append.
> After some fiddling around, I noticed that if write checksums are disabled then the hsync is effective.
> I checked the documentation and could not find any mention of the fact that hsync could be ineffective when write checksums are on.
> The Writer was using the LocalFilesystem.
--
This message was sent by Atlassian Jira
(v8.3.4#803005)
---------------------------------------------------------------------
To unsubscribe, e-mail: common-issues-unsubscribe@hadoop.apache.org
For additional commands, e-mail: common-issues-help@hadoop.apache.org