We have recently decreased the dfs.replication value on our cluster from 3 to 2 and see behavior similar to that described in issue HADOOP-2270 (https://issues.apache.org/jira/browse/HADOOP-2270?page=com.atlassian.jira.plugin.system.issuetabpanels%3Aall-tabpanel). Even though the parameter dfs.replication has been made "final" at the cluster level, we are still seeing files being created from jobs & "put" operations with replication 3.
Based on the issue information (HADOOP-2270), it seems that this bug has not yet been fixed. We are currently running Cloudera Hadoop 0.20.1+169.68... is there any workaround for this issue? We plan to change the client configuration but are hoping for a way to enforce a non-default replication value. Thanks in advance, -Bobby
