[HADOOP-18216] Document "io.file.buffer.size" must be greater than zero

• Type: Bug
• Status: Resolved
• Priority: Minor
• Resolution: Fixed
• Affects Version/s: 3.4.0
• Fix Version/s: 3.4.0
• Component/s: io
• Labels: pull-request-available
• Target Version/s: 3.4.0
• Hadoop Flags: Reviewed
• Flags: Patch
• Assignee: ECFuzz
• Reporter: ECFuzz
• Created: 22/Apr/22 10:39
• Updated: 28/Jan/24 08:26
• Resolved: 26/Apr/22 18:54

Description

when the configuration file in the "io.file.buffer.size" field is set to a value less than or equal to zero, hdfs can start normally, but read and write data will have problems.

When the value is less than zero, the shell will throw the following exception:

hadoop@ljq1:~/hadoop-3.1.3-work/bin$ ./hdfs dfs -cat mapred
-cat: Fatal internal error
java.lang.NegativeArraySizeException: -4096
        at org.apache.hadoop.io.IOUtils.copyBytes(IOUtils.java:93)
        at org.apache.hadoop.io.IOUtils.copyBytes(IOUtils.java:68)
        at org.apache.hadoop.io.IOUtils.copyBytes(IOUtils.java:129)
        at org.apache.hadoop.fs.shell.Display$Cat.printToStdout(Display.java:101)
        at org.apache.hadoop.fs.shell.Display$Cat.processPath(Display.java:96)
        at org.apache.hadoop.fs.shell.Command.processPaths(Command.java:331)
        at org.apache.hadoop.fs.shell.Command.processPathArgument(Command.java:303)
        at org.apache.hadoop.fs.shell.Command.processArgument(Command.java:285)
        at org.apache.hadoop.fs.shell.Command.processArguments(Command.java:269)
        at org.apache.hadoop.fs.shell.FsCommand.processRawArguments(FsCommand.java:120)
        at org.apache.hadoop.fs.shell.Command.run(Command.java:176)
        at org.apache.hadoop.fs.FsShell.run(FsShell.java:328)
        at org.apache.hadoop.util.ToolRunner.run(ToolRunner.java:76)
        at org.apache.hadoop.util.ToolRunner.run(ToolRunner.java:90)
        at org.apache.hadoop.fs.FsShell.main(FsShell.java:391)

When the value is equal to zero, the shell command will always block

hadoop@ljq1:~/hadoop-3.1.3-work/bin$ ./hdfs dfs -cat mapred
^Z
[2]+  Stopped                 ./hdfs dfs -cat mapred

The description of the configuration file is not clear enough, it may make people think that set to 0 to enter the non-blocking mode.
 

<property>   
    <name>io.file.buffer.size</name>   
    <value>4096</value>   
    <description>The size of buffer for use in sequence files.   
    The size of this buffer should probably be a multiple of hardware   
    page size (4096 on Intel x86), and it determines how much data is   
    buffered during read and write operations.</description> 
</property>

 
Considering that this value is uesd by hdfs and mapreduce frequently, we should make this value must be a number greater than zero.

 

Issue Links

• links to: GitHub Pull Request #4220