What is the maximum number of files allowed in a HDFS directory?

Question

What is the maximum number of files and directories allowed in a HDFS (hadoop) directory?

Answer 1

In modern Apache Hadoop versions, various HDFS limits are controlled by configuration properties with fs-limits in the name, all which have reasonable default values. This question specifically asked about number of children in a directory. That's defined by dfs.namenode.fs-limits.max-directory-items, and its default value is 1048576.

Refer to the Apache Hadoop documentation in hdfs-default.xml for the full list of fs-limits configuration properties and their default values. Copy-pasting here for convenience:

<property>
  <name>dfs.namenode.fs-limits.max-component-length</name>
  <value>255</value>
  <description>Defines the maximum number of bytes in UTF-8 encoding in each
      component of a path.  A value of 0 will disable the check.</description>
</property>

<property>
  <name>dfs.namenode.fs-limits.max-directory-items</name>
  <value>1048576</value>
  <description>Defines the maximum number of items that a directory may
      contain. Cannot set the property to a value less than 1 or more than
      6400000.</description>
</property>

<property>
  <name>dfs.namenode.fs-limits.min-block-size</name>
  <value>1048576</value>
  <description>Minimum block size in bytes, enforced by the Namenode at create
      time. This prevents the accidental creation of files with tiny block
      sizes (and thus many blocks), which can degrade
      performance.</description>
</property>

<property>
    <name>dfs.namenode.fs-limits.max-blocks-per-file</name>
    <value>1048576</value>
    <description>Maximum number of blocks per file, enforced by the Namenode on
        write. This prevents the creation of extremely large files which can
        degrade performance.</description>
</property>

<property>
  <name>dfs.namenode.fs-limits.max-xattrs-per-inode</name>
  <value>32</value>
  <description>
    Maximum number of extended attributes per inode.
  </description>
</property>

<property>
  <name>dfs.namenode.fs-limits.max-xattr-size</name>
  <value>16384</value>
  <description>
    The maximum combined size of the name and value of an extended attribute
    in bytes. It should be larger than 0, and less than or equal to maximum
    size hard limit which is 32768.
  </description>
</property>

All of these settings use reasonable default values as decided upon by the Apache Hadoop community. It is generally recommended that users do not tune these values except in very unusual circumstances.

Answer 2

From http://blog.cloudera.com/blog/2009/02/the-small-files-problem/:

Every file, directory and block in HDFS is represented as an object in the namenode’s memory, each of which occupies 150 bytes, as a rule of thumb. So 10 million files, each using a block, would use about 3 gigabytes of memory. Scaling up much beyond this level is a problem with current hardware. Certainly a billion files is not feasible.