[
https://issues.apache.org/jira/browse/HDFS-2103?page=com.atlassian.jira.plugin.system.issuetabpanels:all-tabpanel
]
Bharath Mundlapudi resolved HDFS-2103.
--------------------------------------
Resolution: Not A Problem
Didn't notice the finally block, where read lock is released. I am closing this
Jira.
> Read lock must be released before acquiring a write lock
> --------------------------------------------------------
>
> Key: HDFS-2103
> URL: https://issues.apache.org/jira/browse/HDFS-2103
> Project: Hadoop HDFS
> Issue Type: Bug
> Affects Versions: 0.23.0
> Reporter: Bharath Mundlapudi
> Assignee: Bharath Mundlapudi
> Fix For: 0.23.0
>
>
> In FSNamesystem.getBlockLocationsUpdateTimes function, we have the following
> code:
> {code}
> for (int attempt = 0; attempt < 2; attempt++) {
> if (attempt == 0) { // first attempt is with readlock
> readLock();
> } else { // second attempt is with write lock
> writeLock(); // writelock is needed to set accesstime
> }
> ...
> if (attempt == 0) {
> continue;
> }
> {code}
> In the above code, readLock is acquired in attempt 0 and if the execution
> enters in the continue block, then it tries to acquire writeLock before
> releasing the readLock.
>
--
This message is automatically generated by JIRA.
For more information on JIRA, see: http://www.atlassian.com/software/jira
