[Bigdata-commit] SF.net SVN: bigdata:[7688] branches/MGC_1_3_0/bigdata-jini/src/test/com/ bigdata/journal/jini/ha/TestHAJournalServer.java

[<tho...@us...>]

Revision: 7688
          http://bigdata.svn.sourceforge.net/bigdata/?rev=7688&view=rev
Author:   thompsonbry
Date:     2013-12-20 13:17:53 +0000 (Fri, 20 Dec 2013)
Log Message:
-----------
Disabling the test of RMI interrupt behavior pending a response from the river community.  It appears that the RMI on the remote service is not interrupted when the local thread executing the RMI is interrupted.  To me, this seems like an RMI bug.

Modified Paths:
--------------
    branches/MGC_1_3_0/bigdata-jini/src/test/com/bigdata/journal/jini/ha/TestHAJournalServer.java

Modified: branches/MGC_1_3_0/bigdata-jini/src/test/com/bigdata/journal/jini/ha/TestHAJournalServer.java
===================================================================
--- branches/MGC_1_3_0/bigdata-jini/src/test/com/bigdata/journal/jini/ha/TestHAJournalServer.java	2013-12-19 22:11:12 UTC (rev 7687)
+++ branches/MGC_1_3_0/bigdata-jini/src/test/com/bigdata/journal/jini/ha/TestHAJournalServer.java	2013-12-20 13:17:53 UTC (rev 7688)
@@ -40,6 +40,7 @@
 import net.jini.config.Configuration;
 import net.jini.core.lookup.ServiceID;
 
+import com.bigdata.BigdataStatics;
 import com.bigdata.ha.HAGlue;
 import com.bigdata.ha.HAStatusEnum;
 import com.bigdata.ha.IndexManagerCallable;
@@ -372,11 +373,19 @@
      * This unit test setups up a service and then issues an RMI that invokes a
      * {@link Thread#sleep(long)} method on the service. The thread that issues
      * the RMI is then interrupted during the sleep.
-     * 
-     * @throws Exception
      */
     public void test_interruptRMI() throws Exception {
 
+        if(!BigdataStatics.runKnownBadTests) {
+            /**
+             * FIXME TEST DISABLED. I have written to the river mailing list
+             * about this test. I am not observing the interrupt of the
+             * Thread.sleep() on the remote service. I need to figure out if
+             * that is the expected behavior or if this is an RMI bug.
+             */
+            return;
+        }
+        
         // Start a service.
         final HAGlue serverA = startA();
 
@@ -498,10 +507,11 @@
             log.warn("Will sleep: millis=" + millis);
             try {
                 Thread.sleep(millis);
+                log.warn("Sleep finished normally.");
             } catch (Throwable t) {
+                log.error("Exception during sleep: "+t, t);
                 ((HAJournalTest) getIndexManager()).getRemoteImpl()
                         .setLastRootCause(t);
-                log.error(t, t);
                 throw new RuntimeException(t);
             } finally {
                 log.warn("Did sleep: millis=" + millis);

This was sent by the SourceForge.net collaborative development platform, the world's largest Open Source development site.