[
https://issues.apache.org/jira/browse/HDFS-5211?page=com.atlassian.jira.plugin.system.issuetabpanels:all-tabpanel
]
Chris Nauroth updated HDFS-5211:
--------------------------------
Attachment: HDFS-5211.1.patch
Here is a patch that switches the order of operations in
{{DistributedFileSystem#close}}. It calls the superclass to remove the cache
entry first before closing the contained {{DFSClient}}.
I did a full test run, and the only failure was {{TestBalancerWithNodeGroup}}.
This is unrelated, because the test is known to fail intermittently.
Here is the output from test-patch. The Findbugs warnings are unrelated. This
patch only touches {{DistributedFileSystem#close}}, and none of the warnings
were reported in that method. BTW, I reported a bug in "ant test-patch" in
HADOOP-9965, and the workaround is to revert HADOOP-9573 locally.
[exec] -1 overall.
[exec]
[exec] +1 @author. The patch does not contain any @author tags.
[exec]
[exec] -1 tests included. The patch doesn't appear to include any new
or modified tests.
[exec] Please justify why no tests are needed for
this patch.
[exec]
[exec] +1 javadoc. The javadoc tool did not generate any warning
messages.
[exec]
[exec] +1 javac. The applied patch does not increase the total number
of javac compiler warnings.
[exec]
[exec] -1 findbugs. The patch appears to introduce 22 new Findbugs
(version 1.3.9) warnings.
> Race condition between DistributedFileSystem#close and FileSystem#close can
> cause return of a closed DistributedFileSystem instance from the FileSystem
> cache.
> --------------------------------------------------------------------------------------------------------------------------------------------------------------
>
> Key: HDFS-5211
> URL: https://issues.apache.org/jira/browse/HDFS-5211
> Project: Hadoop HDFS
> Issue Type: Bug
> Components: hdfs-client
> Affects Versions: 1.2.1
> Reporter: Chris Nauroth
> Assignee: Chris Nauroth
> Attachments: HDFS-5211.1.patch
>
>
> {{DistributedFileSystem#close}} closes the underlying {{DFSClient}} before
> the superclass call to {{FileSystem#close}}, which removes the instance from
> the cache. This leaves a small window during which another thread can call
> {{FileSystem#get}} and receive the closed instance from the cache.
--
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
