Details
-
Improvement
-
Status: Open
-
Major
-
Resolution: Unresolved
-
None
-
None
-
None
-
None
Description
Crashing AMs has been a real pain for users since the beginning. And there are already a few tickets floating around, filing this to consolidate them.
Attachments
Issue Links
- depends upon
-
YARN-2560 Diagnostics is delayed to passed to ApplicationReport
- Open
-
YARN-2688 Better diagnostics on Container Launch failures
- Open
-
YARN-4065 container-executor error should include effective user id
- Open
- is related to
-
YARN-1441 The diagnostic shows empty information in the console when Application failed to complete successfully in Distributed Shell
- Open
-
MAPREDUCE-3820 Improve logging when containers run beyond memory limits
- Open
-
YARN-1551 Allow user-specified reason for killApplication
- Open
- relates to
-
YARN-649 Make container logs available over HTTP in plain text
- Closed
-
MAPREDUCE-3688 Need better Error message if AM is killed/throws exception
- Open
-
MAPREDUCE-4362 If possible, we should get back the feature of propagating task logs back to JobClient
- Open
-
YARN-560 If AM fails due to overrunning resource limits, error not visible through UI sometimes
- Open
-
YARN-1438 When a container fails, the text of the exception isn't included in the diagnostics
- Resolved
-
YARN-564 Job history logs do not log anything when JVM fails to start
- Resolved
-
MAPREDUCE-4937 MR AM handles an oversized split metainfo file poorly
- Closed
-
HADOOP-10312 Shell.ExitCodeException to have more useful toString
- Closed
-
YARN-257 NM should gracefully handle a full local disk
- Open
-
YARN-2934 Improve handling of container's stderr
- Resolved