Details
-
Sub-task
-
Status: Resolved
-
Major
-
Resolution: Fixed
-
None
-
None
Description
The test cases there fail because tests are leaking resources into others. For this particular case I noticed that the test cases are checking if all threads are being shut down and they are finding threads created by other tests in the same jvm instance.
To verify this, I added a couple of log messages to print the thread id when the thread starts and when the thread is interrupted. I also interrupt the offending thread in the test case to force it to print its id. Here is one such output:
$ grep 1572 build/test-results/*.xml build/test-results/TEST-kafka.controller.ControllerFailoverTest.xml:[2015-09-16 15:03:42,398] INFO [Controller-0-to-broker-1-send-thread], Starting Controller-0-to-broker-1-send-thread 1572 (kafka.controller.RequestSendThread:68) build/test-results/TEST-kafka.server.ServerShutdownTest.xml:[2015-09-16 15:06:33,334] INFO [Controller-0-to-broker-1-send-thread], Got interrupted: Controller-0-to-broker-1-send-thread 1572 (kafka.controller.RequestSendThread:68)
The number 1572 is the thread id.
Attachments
Issue Links
- links to