[
https://issues.apache.org/jira/browse/HDFS-9607?page=com.atlassian.jira.plugin.system.issuetabpanels:comment-tabpanel&focusedCommentId=15083066#comment-15083066
]
Steve Loughran commented on HDFS-9607:
--------------------------------------
> I believe currently HDFS is not POSIX compliant.
neither was NFSv1 due to its relaxed consistency semantics if local caching was
enabled: doesn't mean there's a need to abandon the API calls. In particular,
if the team were ever to implement full seek-past-end-of-file + write, that
write() call would still be the one to use. Having a new method purely for some
operations would just become a maintenance and test problem.
so: +1 to not trying to implement POSIX; -1 for proposing to implement a subset
of it with new operations. Just use the existing API and have it fall when
appropriate —because that failure point can be changed over time.
Note also that new additions to filesystem APIs should go into
AbstractFileSystem and its delegation-based APIs; the question of when/how to
backport stuff to class {{FileSystem}} subclasses is always a point of
contention.
> Advance Hadoop Architecture (AHA) - HDFS
> ----------------------------------------
>
> Key: HDFS-9607
> URL: https://issues.apache.org/jira/browse/HDFS-9607
> Project: Hadoop HDFS
> Issue Type: New Feature
> Reporter: Dinesh S. Atreya
>
> Link to Umbrella JIRA
> https://issues.apache.org/jira/browse/HADOOP-12620
> Provide capability to carry out in-place writes/updates. Only writes in-place
> are supported where the existing length does not change.
> For example, "Hello World" can be replaced by "Hello HDFS!"
> See
> https://issues.apache.org/jira/browse/HADOOP-12620?focusedCommentId=15046300&page=com.atlassian.jira.plugin.system.issuetabpanels:comment-tabpanel#comment-15046300
> for more details.
--
This message was sent by Atlassian JIRA
(v6.3.4#6332)
