There are also a number of optional Ceph configuration options.
<property>
<name>fs.ceph.blockSize</name>
<value>67108864</value>
<description>Defaulting to 64MB, this is the size (in bytes) you want Ceph to use in striping data internally and presenting it to Hadoop.</description>
</property>
<property>
<name>fs.ceph.debug</name>
<value>true</value>
<description>If true, the Java-based code will print debugging information to standard error. This is useful if attempting to debug a Ceph issue as it puts both outputs in the same place.</description>
</property>
<property>
<name>fs.ceph.clientDebug</name>
<value>1</value>
<description>If non-zero, the Ceph client will print debugging information to standard error (a higher number=more debugging).</description>
</property>
<property>
<name>fs.ceph.messengerDebug</name>
<value>1</value>
<description>If non-zero, the Ceph messenger will print debugging information to standard error(a higher number=more debugging)</description>
</property>
<property>
<name>fs.ceph.readahead</name>
<value>1</value>
<description>Sets the number of object periods to read ahead in prefetching. This should probably be left at the default of 1.</description>
</property>
<property>
<name>fs.ceph.commandLine</name>
<value>a string</value>
<description>If you prefer, you may enter any of Ceph's command-line configuration here and it will get passed to the C client. Note that any filled-in configuration options will override what you put here.
By default, Ceph performs writes across the network rather than locally. To force local writes, add "set_local_pg" in this property.</description>
</property>