Details
-
Type:
Bug
-
Status: Closed
-
Priority:
Major
-
Resolution: Fixed
-
Affects Version/s: 1.1.0
-
Fix Version/s: 1.1.0
-
Component/s: tasktracker, test
-
Labels:None
-
Target Version/s:
-
Hadoop Flags:Reviewed
Description
MAPREDUCE-2850 flagged all TT threads that exited in the MiniMRCluster as dead, this breaks a number of the other tests that use MiniMRCluster across restart.
Issue Links
- relates to
-
MAPREDUCE-2850
Add test for TaskTracker disk failure handling (MR-2413)
-
- Closed
-
Activity
- All
- Comments
- Work Log
- History
- Activity
- Transitions
Patch looks good to me.
+1
Hope you ran all tests that were failing earlier.
Show
Ravi Gummadi
added a comment - Patch looks good to me.
+1
Hope you ran all tests that were failing earlier.
Thanks Ravi. Yea, this fixes the other tets. I'll re-run the whole suite on branch-20-security for sanity before committing.
Show
Eli Collins
added a comment - Thanks Ravi. Yea, this fixes the other tets. I'll re-run the whole suite on branch-20-security for sanity before committing.
All tests ran cleanly except for one, which is unrelated, filed HADOOP-7836 for it. I've committed this.
Show
Eli Collins
added a comment - All tests ran cleanly except for one, which is unrelated, filed HADOOP-7836 for it. I've committed this.
Closed upon release of Hadoop-1.1.0.
Show
Matt Foley
added a comment - Closed upon release of Hadoop-1.1.0.
The change to MiniMRCluster to clean up exited TTs threads is actually somewhat invasive so let's just revert the setting of isDead from MR-2850 and add a field that flags what TestDiskFailures would like to check, which is that the TT exited cleanly.