Kent Yao created SPARK-44632:
--------------------------------

             Summary: DiskBlockManager should check and be able to handle stale 
directories
                 Key: SPARK-44632
                 URL: https://issues.apache.org/jira/browse/SPARK-44632
             Project: Spark
          Issue Type: Improvement
          Components: Spark Core
    Affects Versions: 3.4.1, 3.5.0
            Reporter: Kent Yao


The subDir in the memory cache could be stale, for example, after a damaged 
disk repair or replacement. This dir could be accessed subsequently by others. 
Especially,  `filename` generated by `RDDBlockId` is unchanged between task 
reties, so it probably attempts to access the same subDir repeatedly. 
Therefore, it is necessary to check if the subDir exists. If it is stale and 
the hardware has been recovered without data and directories, we will recreate 
the subDir to prevent FileNotFoundException during writing.



--
This message was sent by Atlassian Jira
(v8.20.10#820010)

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

Reply via email to