[HADOOP-15571] Multiple FileContexts created with the same configuration object should be allowed to have different umask

• Type: Bug
• Status: Resolved
• Priority: Critical
• Resolution: Fixed
• Fix Version/s: 3.2.0, 3.1.1, 3.0.4
• Target Version/s: 3.2.0, 3.1.1, 3.0.4
• Hadoop Flags: Reviewed
• Assignee: Vinod Kumar Vavilapalli
• Reporter: Vinod Kumar Vavilapalli
• Created: 28/Jun/18 23:10
• Updated: 07/Jul/18 04:19
• Resolved: 06/Jul/18 18:59

Description

Ran into a super hard-to-debug issue due to this. [Edit: Turns out the same issue as YARN-5749 that Tao Yang ran into]

Issue

Configuration conf = new Configuration();
fc1 = FileContext.getFileContext(uri1, conf);
fc2 = FileContext.getFileContext(uri2, conf);
fc.setUMask(umask_for_fc1); // Screws up umask for fc2 also!

This was not the case before HADOOP-13440.

Symptoms:

Scenario I ran into

When trying to localize a HDFS directory (hdfs:///my/dir/1.txt), NodeManager tries to replicate the directory structure on the local file-system ($yarn-local-dirs/filecache/my/dir/1.txt).

Now depending on whether NM has ever done a log-aggregation (completely unrelated code that sets umask to be 137 for its own files on HDFS), the directories /my and /my/dir on local-fs may have different permissions. In the specific case where NM did log-aggregation, /my/dir was created with 137 umask and so localization of 1.txt completely failed due to absent directory executable permissions!

Previous scenarios:

We ran into this before in test-cases and instead of fixing the root-cause, we just fixed the test-cases: YARN-5679 / YARN-5749

Issue Links

• is caused by: HADOOP-13440 FileContext does not react on changing umask via configuration
• is related to: HADOOP-13073 RawLocalFileSystem does not react on changing umask
• relates to: YARN-5679 TestAHSWebServices is failing
• relates to: YARN-5749 Fail to localize resources after health status for local dirs changed