Adam Pocock created HDFS-5127:
---------------------------------
Summary: TestDataNodeVolumeFailureReporting and
TestDataNodeVolumeFailureToleration timeout on ZFS
Key: HDFS-5127
URL: https://issues.apache.org/jira/browse/HDFS-5127
Project: Hadoop HDFS
Issue Type: Bug
Components: test
Affects Versions: 2.0.3-alpha, 2.0.0-alpha
Environment: Solaris 11.1, x86-64
Reporter: Adam Pocock
Priority: Minor
TestDataNodeVolumeFailureReporting and TestDataNodeVolumeFailureToleration
timeout on ZFS, due to an inaccurate method of measuring node capacity.
Filesystems like ZFS which store additional metadata do not return absolutely
consistent capacities after filesystem changes, so the logic which controls
these two tests (in DFSTestUtil.waitForDatanodeStatus) fails as
(expectedCapacity != currTotalCapacity), and the tests timeout. This occurs
even when the ZFS filesystem and pool live on a ramdisk. The tests pass
correctly when using a UFS filesystem on a ramdisk, and so the problem is ZFS
rather than Solaris specific.
--
This message is automatically generated by JIRA.
If you think it was sent incorrectly, please contact your JIRA administrators
For more information on JIRA, see: http://www.atlassian.com/software/jira
