[
https://issues.apache.org/jira/browse/HDFS-5289?page=com.atlassian.jira.plugin.system.issuetabpanels:all-tabpanel
]
Aaron T. Myers updated HDFS-5289:
---------------------------------
Attachment: HDFS-5289.patch
Here's a patch which addresses the issue by catching the exception and retrying
the check in this case, instead of assuming that the check will return null. I
tested this by adding a {{Thread.sleep(1000)}} in
{{TestRetryCacheWithHA#testClientRetryWithFailover}} in the thread to execute
the operation. The test fails reliably without this patch, and passes reliably
with it.
> Race condition in TestRetryCacheWithHA#testCreateSymlink causes spurious test
> failure
> -------------------------------------------------------------------------------------
>
> Key: HDFS-5289
> URL: https://issues.apache.org/jira/browse/HDFS-5289
> Project: Hadoop HDFS
> Issue Type: Bug
> Components: test
> Affects Versions: 2.1.1-beta
> Reporter: Aaron T. Myers
> Assignee: Aaron T. Myers
> Attachments: HDFS-5289.patch
>
>
> The code to check if the operation has been completed on the active NN can
> potentially execute before the thread actually doing the operation has run.
> In this case the checking code will retry the check if the result of the
> check is null. However, the test operation does not in fact return null,
> instead throwing an exception if the file doesn't exist yet. We need to catch
> the exception and retry.
--
This message was sent by Atlassian JIRA
(v6.1#6144)
