tomscut edited a comment on pull request #3839: URL: https://github.com/apache/hadoop/pull/3839#issuecomment-1008483869
> It's good catch here. Would you mind to add new test to cover this case? BTW, what is the root cause about NPE here? Thanks. Thanks @Hexiaoqiao for your review and comments. The main reason is that the `DatanodeDescriptor#storageMap` cannot find the storage that DN reports through `lifeline`. As a result, the `NPE` is thrown in line 460 of the code when the storage is operated. It is similar to the situation encountered in this issue [HDFS-14042](https://issues.apache.org/jira/browse/HDFS-14042).  Here is the scenario we encountered: 1 We add a disk to datanode and reconfig `dfs.datanode.data.dir`. 2 After a while, many values on NN's Web become negative. 3 We found the NPE from the NN log and discovered this issue [HDFS-14042](https://issues.apache.org/jira/browse/HDFS-14042). -- This is an automated message from the Apache Git Service. To respond to the message, please log on to GitHub and use the URL above to go to the specific comment. To unsubscribe, e-mail: [email protected] For queries about this service, please contact Infrastructure at: [email protected] --------------------------------------------------------------------- To unsubscribe, e-mail: [email protected] For additional commands, e-mail: [email protected]
