Apache
/
hbase
1
0
Fork 0
Code Issues Pull Requests Packages Projects Releases Wiki Activity

HBASE-5029 TestDistributedLogSplitting fails on occasion 

git-svn-id: https://svn.apache.org/repos/asf/hbase/trunk@1215370 13f79535-47bb-0310-9956-ffa450edef68
This commit is contained in:
Michael Stack 2011-12-17 00:40:53 +00:00
parent 324eec9a44
commit 3cad85bc77
1 changed files with 23 additions and 11 deletions
Show Stats Download Patch File Download Diff File Expand all files Collapse all files

32
src/test/java/org/apache/hadoop/hbase/master/TestDistributedLogSplitting.java
View File

@ -19,12 +19,10 @@
*/ */
package org.apache.hadoop.hbase.master; package org.apache.hadoop.hbase.master;
import static org.apache.hadoop.hbase.zookeeper.ZKSplitLog.Counters.tot_wkr_final_transistion_failed; import static org.apache.hadoop.hbase.zookeeper.ZKSplitLog.Counters.*;
import static org.apache.hadoop.hbase.zookeeper.ZKSplitLog.Counters.tot_wkr_task_acquired;
import static org.apache.hadoop.hbase.zookeeper.ZKSplitLog.Counters.tot_wkr_task_err;
import static org.apache.hadoop.hbase.zookeeper.ZKSplitLog.Counters.tot_wkr_task_resigned;
import static org.junit.Assert.assertEquals; import static org.junit.Assert.assertEquals;
import static org.junit.Assert.assertTrue; import static org.junit.Assert.assertTrue;
import static org.junit.Assert.fail;
import java.io.IOException; import java.io.IOException;
import java.util.Iterator; import java.util.Iterator;
@ -246,6 +244,15 @@ public class TestDistributedLogSplitting {
assertEquals(NUM_LOG_LINES, count); assertEquals(NUM_LOG_LINES, count);
} }
/**
* The original intention of this test was to force an abort of a region
* server and to make sure that the failure path in the region servers is
* properly evaluated. But it is difficult to ensure that the region server
* doesn't finish the log splitting before it aborts. Also now, there is
* this code path where the master will preempt the region server when master
* detects that the region server has aborted.
* @throws Exception
*/
@Test (timeout=300000) @Test (timeout=300000)
public void testWorkerAbort() throws Exception { public void testWorkerAbort() throws Exception {
LOG.info("testWorkerAbort"); LOG.info("testWorkerAbort");
@ -279,21 +286,26 @@ public class TestDistributedLogSplitting {
slm.enqueueSplitTask(logfiles[0].getPath().toString(), batch); slm.enqueueSplitTask(logfiles[0].getPath().toString(), batch);
//waitForCounter but for one of the 2 counters //waitForCounter but for one of the 2 counters
long curt = System.currentTimeMillis(); long curt = System.currentTimeMillis();
long endt = curt + 30000; long waitTime = 30000;
long endt = curt + waitTime;
while (curt < endt) { while (curt < endt) {
if ((tot_wkr_task_resigned.get() + tot_wkr_task_err.get() + if ((tot_wkr_task_resigned.get() + tot_wkr_task_err.get() +
tot_wkr_final_transistion_failed.get()) == 0) { tot_wkr_final_transistion_failed.get() + tot_wkr_task_done.get() +
tot_wkr_preempt_task.get()) == 0) {
Thread.yield(); Thread.yield();
curt = System.currentTimeMillis(); curt = System.currentTimeMillis();
} else { } else {
assertEquals(1, (tot_wkr_task_resigned.get() + tot_wkr_task_err.get() + assertEquals(1, (tot_wkr_task_resigned.get() + tot_wkr_task_err.get() +
tot_wkr_final_transistion_failed.get())); tot_wkr_final_transistion_failed.get() + tot_wkr_task_done.get() +
tot_wkr_preempt_task.get()));
return; return;
} }
} }
assertEquals(1, batch.done); fail("none of the following counters went up in " + waitTime +
// fail("region server completed the split before aborting"); " milliseconds - " +
return; "tot_wkr_task_resigned, tot_wkr_task_err, " +
"tot_wkr_final_transistion_failed, tot_wkr_task_done, " +
"tot_wkr_preempt_task");
} }
HTable installTable(ZooKeeperWatcher zkw, String tname, String fname, HTable installTable(ZooKeeperWatcher zkw, String tname, String fname,