You are viewing a plain text version of this content. The canonical link for it is here.

Posted to issues@spark.apache.org by "Holden Karau (Jira)" <ji...@apache.org> on 2021/07/28 19:10:00 UTC

[jira] [Created] (SPARK-36330) Use atomic renames for diskstore block writes & do "fall-through" reads

Holden Karau created SPARK-36330:
------------------------------------

             Summary: Use atomic renames for diskstore block writes & do "fall-through" reads
                 Key: SPARK-36330
                 URL: https://issues.apache.org/jira/browse/SPARK-36330
             Project: Spark
          Issue Type: Improvement
          Components: Spark Core
    Affects Versions: 3.3.0
            Reporter: Holden Karau


This is a potential follow-up for SPARK-36058. Since we could end up in the situation where we have a partial write we can't do fall through reads to the diskStore. A solution to this is to use atomic renames on successful write of a block which would make it safe to assume that any on-disk representation of a block represented the full block.



--
This message was sent by Atlassian Jira
(v8.3.4#803005)

---------------------------------------------------------------------
To unsubscribe, e-mail: issues-unsubscribe@spark.apache.org
For additional commands, e-mail: issues-help@spark.apache.org