hadoop - multiple Datanode directories write operation -
iam using hadoop apache 2.7.1 , have configured data node directory have multiple directories
<property> <name>dfs.data.dir</name> <value>/opt/hadoop/data_dir,file:///hdd/data_dir/</value> <final>true</final> </property>
according configuration writing file data should happens on both directories /opt/hadoop/data_dir , file:///hdd/data_dir/ same blocks names , on same sub directories names
but in cluster behavior not happening times writes data blocks local directory /opt/hadoop/data_dir , times writes data blocks
external hard directory file:///hdd/data_dir
what possible reasons , how control behavior
your interpretation of dfs.data.dir
wrong. stated:
with same blocks names , on same sub directories names
files might replicated happens on different datanodes, not on same datanode different directories. property dfs.data.dir
deprecated , should dfs.datanode.data.dir
has been there since hadoop 2.
from hdfs-default.xml
dfs.datanode.data.dir
determines on local filesystem dfs data node should store blocks. if comma-delimited list of directories, data stored in named directories, typically on different devices.
Comments
Post a Comment