[
https://issues.apache.org/jira/browse/DIRAPI-94?page=com.atlassian.jira.plugin.system.issuetabpanels:comment-tabpanel&focusedCommentId=13475672#comment-13475672
]
Stefan Seelmann commented on DIRAPI-94:
---------------------------------------
Thanks Emmanuel for the pointers.
When adding a Thread.sleep(1000) just after the connection.search() call the
test fails always. So in that case it seems the server already sent all
results. This proves your theory.
When increasing numEntires from 100 to 1000 the test passes (with 100000
repetitions). However such a change is still not reliable.
I followed your advise with the interceptor and an entry filter. Such an
interceptor already existed as inner class of SearchLimitsIT. I extracted that
interceptor and moved it to test-framework module so it can be reused in
ClientAbandonRequestTest.
> Random failure of ClientAbandonRequestTest
> ------------------------------------------
>
> Key: DIRAPI-94
> URL: https://issues.apache.org/jira/browse/DIRAPI-94
> Project: Directory Client API
> Issue Type: Bug
> Affects Versions: 1.0.0-M13
> Reporter: Stefan Seelmann
> Fix For: 1.0.0-M14
>
>
> ClientAbandonRequestTest.testAbandonSearch() fails occasionally on Jenkins.
> To demonstrate it I added the MulitThreadedMultiInvoker, please run the test
> with system properties "-Dmtmi.invocations=1000 -Dmtmi.trace=true" which runs
> each test method multiple times. The problem is that sometimes the abandon
> request has no effect but the client receives all search result. I'm not sure
> if it is a problem is located in the client API or on the server side.
--
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
