[ http://issues.apache.org/jira/browse/HADOOP-240?page=comments#action_12413006 ]
Hairong Kuang commented on HADOOP-240: -------------------------------------- On a second thought, the right approach to this problem should distiguish different cases when create a directory and then log them differently. When the mkdirs function is called, either the directory is successfully created, or the directory already existed, or the directory is not able to created due to its parent's problem. Currently FSDirectory.addNode returns null when a directory exists or a directory is not able to be created. I am thinking to add a PathNotFound exception. When a directory is not able to be created, addNode throws a PathNotFound exception. When a directory exists, it returns null. Otherwise, returns a new node. When the function mkdirs catches a PathNotFound exception, it logs it as a failure and returns. When it gets a null, simply continues. > namenode should not log failed mkdirs at warning level > ------------------------------------------------------ > > Key: HADOOP-240 > URL: http://issues.apache.org/jira/browse/HADOOP-240 > Project: Hadoop > Type: Bug > Components: dfs > Versions: 0.2.1 > Reporter: Hairong Kuang > Assignee: Hairong Kuang > Priority: Minor > Attachments: mkdir.patch > > when a namenode creates a directory, it also recursively tries to creates its > parent directories. If they already exist, the lastSuccess is false and the > "error" is logged at the warning level. > The right approach is to set the logging leve lower to fine or simply ignore > the the "error". -- This message is automatically generated by JIRA. - If you think it was sent incorrectly contact one of the administrators: http://issues.apache.org/jira/secure/Administrators.jspa - For more information on JIRA, see: http://www.atlassian.com/software/jira
