[ 
https://issues.apache.org/jira/browse/SPARK-17573?page=com.atlassian.jira.plugin.system.issuetabpanels:comment-tabpanel&focusedCommentId=15498996#comment-15498996
 ] 

Jianfei Wang commented on SPARK-17573:
--------------------------------------

fileStream may never be closed when some exceptions occur,such as:
fileStream.getChannel.positition(start) throw IOException.
Oh this problem is not so interesting, I will go to other problems,thank you 
Sean Owen!
this method is in ExternalAppendOnlyMap.scala
{code}
 private def nextBatchStream(): DeserializationStream = {
        // some code ingnore
       
        
        fileStream = new FileInputStream(file)
        fileStream.getChannel.position(start)
        batchIndex += 1

        val end = batchOffsets(batchIndex)

        assert(end >= start, "start = " + start + ", end = " + end +
          ", batchOffsets = " + batchOffsets.mkString("[", ", ", "]"))

        val bufferedStream = new 
BufferedInputStream(ByteStreams.limit(fileStream, end - start))
        val compressedStream = serializerManager.wrapForCompression(blockId, 
bufferedStream)
        ser.deserializeStream(compressedStream)
      } else {
        // No more batches left
        cleanup()
        null
      }
    }
{code}

> The FileInputStream may be uncloseed when some exceptions occurs
> ----------------------------------------------------------------
>
>                 Key: SPARK-17573
>                 URL: https://issues.apache.org/jira/browse/SPARK-17573
>             Project: Spark
>          Issue Type: Improvement
>          Components: SQL
>    Affects Versions: 2.0.0
>            Reporter: Jianfei Wang
>            Priority: Trivial
>              Labels: performance
>
> I think that the InputStream may never be closed when some exceptions occur, 
> we should surround this with try catch.
> {code}
>  private def addFilesToZipStream(parent: String, source: File, output: 
> ZipOutputStream): Unit = {
>     if (source.isDirectory()) {
>       output.putNextEntry(new ZipEntry(parent + source.getName()))
>       for (file <- source.listFiles()) {
>         addFilesToZipStream(parent + source.getName() + File.separator, file, 
> output)
>       }
>     } else {
>       val in = new FileInputStream(source)
>       output.putNextEntry(new ZipEntry(parent + source.getName()))
>       val buf = new Array[Byte](8192)
>       var n = 0
>       while (n != -1) {
>         n = in.read(buf)
>         if (n != -1) {
>           output.write(buf, 0, n)
>         }
>       }
>       output.closeEntry()
>       in.close()
>     }
>   }
> some code in TestUtils.scala
>  val in = new FileInputStream(file)
>       ByteStreams.copy(in, jarStream)
>      in.close()
> some code in IvyTestUtils.scala
>    val in = new FileInputStream(file._2)
>       ByteStreams.copy(in, jarStream)
>       in.close()
> {code} 



--
This message was sent by Atlassian JIRA
(v6.3.4#6332)

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

Reply via email to