For example, an HDFS URI of the form hdfs://<namenode-name>[:<port>]/<path>/<file-name> is a static URI. If you deploy the HDFS target service on multiple EDS Nodes, to reduce the load on an HDFS target service, all the instances use the same URI. The instances try to write to the same file. Only one target service succeeds because, at any specified time, only one instance can acquire a lease to write to a file. This scenario results in data loss.

To enable all the instances to write to the HDFS file system, use the #infa.nodename variable in the <file-name> component of the HDFS URI. when configuring the HDFS target service,

The following example shows a URI that uses the #infa.nodename variable:

hdfs://MyNamenode/app_data/logs/#infa.nodename.log

#infa.nodename is the variable that EDS replaces with the EDS Node name.

The name of the file to which a particular instance writes changes with the name of the EDS Node on which the instance runs. Each target service writes to its own HDFS file. For example, the target service in a EDS Node called node1 writes to the file hdfs://MyNamenode/app_data/logs/node1.log and the target service in a EDS Node called node2 writes to the file hdfs://MyNamenode/app_data/logs/node2.log.