Author: mahadev
Date: Wed Nov 25 19:06:08 2009
New Revision: 884229
URL: http://svn.apache.org/viewvc?rev=884229&view=rev
Log:
ZOOKEEPER-597. ASyncHammerTest is failing intermittently on hudson trunk
(Patrick Hunt via mahadev)
Modified:
hadoop/zookeeper/trunk/CHANGES.txt
hadoop/zookeeper/trunk/src/java/test/org/apache/zookeeper/test/ClientBase.java
Modified: hadoop/zookeeper/trunk/CHANGES.txt
URL:
http://svn.apache.org/viewvc/hadoop/zookeeper/trunk/CHANGES.txt?rev=884229&r1=884228&r2=884229&view=diff
==============================================================================
--- hadoop/zookeeper/trunk/CHANGES.txt (original)
+++ hadoop/zookeeper/trunk/CHANGES.txt Wed Nov 25 19:06:08 2009
@@ -133,6 +133,9 @@
ZOOKEEPER-582. ZooKeeper can revert to old data when a snapshot is created
outside of normal processing (ben reed and mahadev via mahadev)
+ ZOOKEEPER-597. ASyncHammerTest is failing intermittently on hudson trunk
+ (Patrick Hunt via mahadev)
+
IMPROVEMENTS:
ZOOKEEPER-473. cleanup junit tests to eliminate false positives due to
"socket reuse" and failure to close client (phunt via mahadev)
Modified:
hadoop/zookeeper/trunk/src/java/test/org/apache/zookeeper/test/ClientBase.java
URL:
http://svn.apache.org/viewvc/hadoop/zookeeper/trunk/src/java/test/org/apache/zookeeper/test/ClientBase.java?rev=884229&r1=884228&r2=884229&view=diff
==============================================================================
---
hadoop/zookeeper/trunk/src/java/test/org/apache/zookeeper/test/ClientBase.java
(original)
+++
hadoop/zookeeper/trunk/src/java/test/org/apache/zookeeper/test/ClientBase.java
Wed Nov 25 19:06:08 2009
@@ -28,6 +28,8 @@
import java.net.Socket;
import java.util.Arrays;
import java.util.LinkedList;
+import java.util.Map;
+import java.util.Map.Entry;
import java.util.concurrent.CountDownLatch;
import java.util.concurrent.TimeUnit;
import java.util.concurrent.TimeoutException;
@@ -453,6 +455,20 @@
return d.delete();
}
+ private static void logAllStackTraces() {
+ StringBuffer sb = new StringBuffer();
+ sb.append("Starting logAllStackTraces()\n");
+ Map<Thread, StackTraceElement[]> threads = Thread.getAllStackTraces();
+ for (Entry<Thread, StackTraceElement[]> e: threads.entrySet()) {
+ sb.append("Thread " + e.getKey().getName() + "\n");
+ for (StackTraceElement elem: e.getValue()) {
+ sb.append("\tat " + elem + "\n");
+ }
+ }
+ sb.append("Ending logAllStackTraces()\n");
+ LOG.error(sb.toString());
+ }
+
/*
* Verify that all of the servers see the same number of nodes
* at the root
@@ -465,15 +481,24 @@
// run through till the counts no longer change on each server
// max 15 tries, with 2 second sleeps, so approx 30 seconds
int[] counts = new int[parts.length];
+ int failed = 0;
for (int j = 0; j < 100; j++) {
int newcounts[] = new int[parts.length];
int i = 0;
for (String hp : parts) {
- ZooKeeper zk = createClient(hp);
try {
- newcounts[i++] = zk.getChildren("/", false).size();
- } finally {
- zk.close();
+ ZooKeeper zk = createClient(hp);
+
+ try {
+ newcounts[i++] = zk.getChildren("/", false).size();
+ } finally {
+ zk.close();
+ }
+ } catch (Throwable t) {
+ failed++;
+ // if session creation fails dump the thread stack
+ // and try the next server
+ logAllStackTraces();
}
}
@@ -486,6 +511,11 @@
counts = newcounts;
Thread.sleep(10000);
}
+
+ // don't keep this up too long, will assert false below
+ if (failed > 10) {
+ break;
+ }
}
// verify all the servers reporting same number of nodes
