fast.html revision fb3fb4f3d76d55b64440afd0af72775dfad3bd1d
<!--
Copyright 2006 Sun Microsystems, Inc. All rights reserved.
Use is subject to license terms.
ident "%Z%%M% %I% %E% SMI"
CDDL HEADER START
The contents of this file are subject to the terms of the
Common Development and Distribution License (the "License").
You may not use this file except in compliance with the License.
You can obtain a copy of the license at usr/src/OPENSOLARIS.LICENSE
See the License for the specific language governing permissions
and limitations under the License.
When distributing Covered Code, include this CDDL HEADER in each
file and include the License file at usr/src/OPENSOLARIS.LICENSE.
If applicable, add the following below this CDDL HEADER, with the
fields enclosed by brackets "[]" replaced with your own identifying
information: Portions Copyright [yyyy] [name of copyright owner]
CDDL HEADER END
-->
<html>
<head>
<title>Quick Start Guide to the Java DTrace API</title>
</head>
<body>
<h1><a name="Quick_Start_Guide_to_the_Java_DTrace_API_"></a>Quick Start
Guide</h1>
<h1><small><small>to the</small> Java DTrace API</small></h1>
<hr style="width: 100%; height: 2px;">
<h2>Contents</h2>
<ul>
<li><a href="#Hello_World">"hello, world" Example</a></li>
<ul>
<li><a href="#Writing_a_Simple_Consumer">Writing a Simple Consumer</a></li>
Script</a></li>
</ul>
<li><a href="#Aggregations">Aggregations</a></li>
<li><a href="#Target_Process">Target Process ID</a></li>
<li><a href="#Closing_Consumers">Closing Consumers</a></li>
<li><a href="#Learning_DTrace">Learning More</a><br>
</li>
</ul>
<h2><a name="Hello_World"></a>"hello, world" Example</h2>
To demonstrate how to use the Java DTrace API, let's write a simple Java
"hello, world" and exits). You will need root permission to use the
Java DTrace API (just as you do to use the <tt>dtrace(1M)</tt> command).
You may want to eliminate this inconvenience by adding the following
<br>
<br>
<tt><i>user-name</i>::::defaultpriv=basic,dtrace_kernel,dtrace_proc</tt>
<br>
<br>
<i>(Substitute your user name.)</i> See the <a
<b>Security</b></a> chapter of the <i>Solaris Dynamic Tracing Guide</i>
for more information.
<br>
<h4><a name="Writing_a_Simple_Consumer"></a>Writing a Simple Consumer</h4>
Creating a DTrace <a
href="/api/org/opensolaris/os/dtrace/Consumer.html">consumer</a>
is easy:
<pre><tt>
Consumer consumer = new LocalConsumer();
</tt></pre>
<br>
Before you can do anything with the consumer, you must first open it.
Then you simply compile and enable one or more D programs and run it:
<pre><tt>
consumer.compile(program);
consumer.go(); // runs in a background thread
</tt></pre>
<br>
To get the data generated by DTrace, you also need to add a <a
href="/api/org/opensolaris/os/dtrace/ConsumerListener.html">listener</a>:
<pre><tt>
consumer.addConsumerListener(new ConsumerAdapter() {
public void dataReceived(DataEvent e) {
}
});
</tt></pre>
<br>
Here is a simple example that runs a given D script:<br>
<br>
<pre><tt><font color=#aaaaaa>
import org.opensolaris.os.dtrace.*;
public class TestAPI {
public static void
main(String[] args)
{
if (args.length < 1) {
System.err.println("Usage: java TestAPI <script> [ macroargs... ]");
System.exit(2);
}
File file = new File(args[0]);
String[] macroArgs = new String[args.length - 1];
Consumer consumer = new LocalConsumer();
consumer.addConsumerListener(new ConsumerAdapter() {
public void dataReceived(DataEvent e) {
}
});
<font color=#aaaaaa>
try {</font>
consumer.compile(file, macroArgs);
} catch (Exception e) {
System.exit(1);
}
}
}</font>
</tt></pre>
<br>
Compile the test program as follows:
<pre><tt>
java -cp dtrace.jar TestAPI.java
</tt></pre>
<br>
Now we need a D script for the program to run. The following is a
simple example that prints "hello, world" and exits:<br>
<pre><tt>
dtrace:::BEGIN
{
trace("hello, world");
exit(0);
}
</tt></pre>
<br>
Run as follows:<br>
On i86:<br>
<pre><tt>
java -cp .:dtrace.jar TestAPI hello.d
</tt></pre>
<br>
On sparc you need to add the <tt>-d64</tt> option to java:
<pre><tt>
</tt></pre>
<br>
You may need to set <tt>LD_LIBRARY_PATH</tt> so that Java can find
this:
<pre><tt>
enabledProbeDescription = dtrace:::BEGIN, flow = null, records =
["hello, world", 0]]
</tt></pre>
<br>
There is one record in the <a
for each action in the D script. The first record is generated by the
<tt>trace()</tt> action. The second is generated by the <tt>exit()</tt>
action. For prettier output, you could change the <tt>ConsumerAdapter <a
href="/api/org/opensolaris/os/dtrace/ConsumerAdapter.html#dataReceived%28org.opensolaris.os.dtrace.DataEvent%29">dataReceived()</a></tt>
implementation as follows:
<pre><tt><font color=#aaaaaa>
consumer.addConsumerListener(new ConsumerAdapter() {
public void dataReceived(DataEvent e) {
ProbeData data = e.getProbeData();
for (Record r : records) {
if (r instanceof ExitRecord) {
} else {
}
}<font color=#aaaaaa>
}
});</font>
</tt></pre>
<br>
<h2><a name="Aggregations"></a>Aggregations</h2>
The example Java program can just as easily run a more complex script,
such as an aggregation:<br>
<pre><tt>
syscall:::entry
/ execname == $$1 /
{
@[probefunc] = count();
}
profile:::tick-1sec
{
printa(@);
clear(@);
}
</tt></pre>
<br>
The above script uses the <tt>$$1</tt> macro variable as a placeholder
for whatever executable you'd like to trace. See the <a
Macro Arguments</b></a> section of the <b>Scripting</b> chapter of the
<i>Solaris Dynamic Tracing Guide</i>. Using two dollar signs (<tt>$$1</tt>)
instead of one (<tt>$1</tt>) forces expansion of the macro variable to
type string.<br>
<br>
To run the example Java program using the above D script, you need to
specify an argument to the <tt>execname</tt> placeholder. To be sure of
getting output, let's use "java":<br>
On i86:
<pre><tt>
java -cp .:dtrace.jar TestAPI syscall.d java
</tt></pre>
<br>
On sparc:
<pre><tt>
java -d64 -cp .:dtrace.jar TestAPI syscall.d java
</tt></pre>
<br>
A data record generated by the <tt>printa()</tt> action is printed to
the console once every second. It contains counts of system calls by
function name made by java. No record is generated by the
<tt>clear()</tt> action.<br>
<br>
If you omit the argument to the <tt>execname</tt> placeholder, the
program fails to compile and the API throws the following exception:
<pre><tt>
syscall.d: line 2: macro argument $$1 is not defined
at org.opensolaris.os.dtrace.LocalConsumer._compileFile(Native Method)
at TestAPI.main(TestAPI.java:26)
</tt></pre>
<br>
A DTrace script may have more than one aggregation. In that case, each
aggregation needs a distinct name:<br>
<pre><tt>
sdt:::interrupt-start
{
self->ts = vtimestamp;
}
sdt:::interrupt-complete
/ self->ts && arg0 /
{
this->devi = (struct dev_info *)arg0;
@counts[stringof(`devnamesp[this->devi->devi_major].dn_name),
this->devi->devi_instance, cpu] = count();
@times[stringof(`devnamesp[this->devi->devi_major].dn_name),
this->devi->devi_instance, cpu] = sum(vtimestamp - self->ts);
self->ts = 0;
}
</tt></pre>
<br>
The <tt>@counts</tt> and <tt>@times</tt> aggregations both accumulate
values for each unique combination of device name, device instance, and
CPU (a three-element tuple). In this example we drop the <tt>tick</tt>
probe to demonstrate a more convenient way to get aggregation data
without the use of the <tt>printa()</tt> action. The <a
href="/api/org/opensolaris/os/dtrace/Consumer.html#getAggregate%28%29">
<tt>getAggregate()</tt></a> method allows us to get all aggregations at
once on a programmatic interval.<br>
<pre><tt><font color=#aaaaaa>
...
try {
consumer.compile(file, macroArgs);
consumer.go();</font>
Aggregate a;
do {
Thread.sleep(1000); // 1 second
a = consumer.getAggregate();
if (!a.asMap().isEmpty()) {
}
} catch (Exception e) {
System.exit(1);
}
...</font>
</tt></pre>
<br>
Compile and run:
<pre><tt>
javac -cp dtrace.jar TestAPI2.java
</tt></pre>
<br>
On i86:
<pre><tt>
java -cp .:dtrace.jar TestAPI2 intrstat.d
</tt></pre>
<br>
On sparc:
<pre><tt>
java -d64 -cp .:dtrace.jar TestAPI2 intrstat.d
</tt></pre>
<br>
and running it again with the above modification (<tt>TestAPI2</tt>).<br>
<br>
By default, the requested aggregate includes every aggregation and
accumulates running totals. To display values per time interval
(instead of running totals), clear the aggregations each time you call
<tt>getAggregate()</tt>. Clearing an aggregation resets all counts to
zero without removing any tuples. The following modification to the
example above clears all aggregations:
<pre><tt><font color=#aaaaaa>
// a = consumer.getAggregate();</font>
a = consumer.getAggregate(null, null); // included, cleared
</tt></pre>
<br>
Each <tt>Set</tt> of aggregation names, <tt>included</tt> and
<tt>cleared</tt>, specifies <i>all</i> aggregations if <tt>null</tt> and
no aggregations if empty. Any subset is possible. However, if an
aggregation has ever been used in the <tt>printa()</tt> action, it is no
longer available to the <tt>getAggregate()</tt> method.<br>
<br>
Be aware that you cannot call <tt>getAggregate()</tt> on an interval
faster that the <tt>aggrate</tt> setting. See the <a
<b>Options and Tunables</b></a> chapter of the <i>Solaris Dynamic
Tracing Guide</i>. See also the <a
<b>Minimizing Drops</b></a> section of the <b>Aggregations</b> chapter
for specific information about the <tt>aggrate</tt> option. The default
<tt>aggrate</tt> is once per second. Here's an example of how you might
double the <tt>aggrate</tt> to minimize drops:
<pre><tt>
</tt></pre>
<br>
Even a single drop terminates the consumer unless you override the <a
href="/api/org/opensolaris/os/dtrace/ConsumerAdapter.html#dataDropped%28org.opensolaris.os.dtrace.DropEvent%29">
<tt>dataDropped()</tt></a> method of <tt>ConsumerAdapter</tt> to handle
drops differently. To avoid drops, it is probably better to increase
the <tt>aggsize</tt> option, since increasing the <tt>aggrate</tt> makes
the consumer work harder. In most cases, the <tt>aggrate</tt> should
only be increased when you need to update a display of aggregation data
more frequently than once per second. Many runtime options, including
<tt>aggrate</tt>, can be changed dynamically while the consumer is
running.<br>
<br>
It's also worth mentioning that a D aggregation may omit square
brackets and aggregate only a single value:
<pre><tt>
@total = count();
</tt></pre>
<br>
The resulting singleton <a
href="/api/org/opensolaris/os/dtrace/Aggregation.html">
<tt>Aggregation</tt></a> instance has one record that may be obtained as
follows:
<pre><tt>
Aggregate a = consumer.getAggregate();
Aggregation total = a.getAggregation("total");
AggregationRecord totalRecord = total.getRecord(Tuple.EMPTY);
</tt></pre>
<br>
<h2><a name="Target_Process"></a>Target Process ID</h2>
aggregation example above), the Java DTrace API also supports the
<tt>$target</tt> macro variable. (See the <a
<b>Target Process ID</b></a> section of the <b>Scripting</b> chapter of
the <i>Solaris Dynamic Tracing Guide</i>.) This allows you to trace a
process from the very beginning of its execution, rather than sometime
after you manually obtain its process ID. The API does this by creating
a process that is initially suspended and allowed to start only after <a
href="/api/org/opensolaris/os/dtrace/Consumer.html#go%28%29">
<tt>go()</tt></a> has initiated tracing. For example, we might want to
aggregate all the system calls from start to finish made by the
<tt>date</tt> command:<br>
<pre><tt>
syscall:::entry
/ pid == $target /
{
@[probefunc] = count();
}
</tt></pre>
<br>
href="/api/org/opensolaris/os/dtrace/Consumer.html#createProcess%28java.lang.String%29">
<tt>createProcess()</tt></a> call to execute the given command but
prevent it from starting until the consumer is running:<br>
<pre><tt><font color=#aaaaaa>
...
consumer.open();</font>
consumer.compile(file);
consumer.go();
...</font>
</tt></pre>
<br>
It also overrides the <a
href="/api/org/opensolaris/os/dtrace/ConsumerAdapter.html#processStateChanged%28org.opensolaris.os.dtrace.ProcessEvent%29">
<tt>processStateChanged()</tt></a> method of the
<tt>ConsumerAdapter</tt> to print a notification when the process has
ended:
<pre><tt><font color=#aaaaaa>
...
consumer.addConsumerListener(new ConsumerAdapter() {
public void dataReceived(DataEvent e) {
}
public void consumerStopped(ConsumerEvent e) {
try {
Aggregate a = consumer.getAggregate();
for (Aggregation agg : a.asMap().values()) {
for (AggregationRecord rec : agg.asMap().values()) {
System.out.println(rec.getTuple() + " " +
rec.getValue());
}
}
} catch (Exception x) {
System.exit(1);
}
}</font>
public void processStateChanged(ProcessEvent e) {
}<font color=#aaaaaa>
});
...</font>
</tt></pre>
<br>
Compile and run:
<pre><tt>
javac -cp dtrace.jar TestTarget.java
</tt></pre>
<br>
On i86:
<pre><tt>
java -cp .:dtrace.jar TestTarget target.d date
</tt></pre>
<br>
On sparc:
<pre><tt>
java -d64 -cp .:dtrace.jar TestTarget target.d date
</tt></pre>
<br>
The consumer exits automatically when the target <tt>date</tt> process
completes.<br>
<h2><a name="Closing_Consumers"></a>Closing Consumers</h2>
An application using the Java DTrace API may run multiple consumers
simultaneously. When a consumer stops running, the programmer is
responsible for closing it in order to release the system resources it
holds. A consumer may stop running for any of the following reasons:
<ul>
<li>It was stopped explicitly by a call to its <a
<tt>stop()</tt></a> method</li>
<li>It encountered the <tt>exit()</tt> action</li>
<li>Its <tt>$target</tt> process or processes (if any) all completed</li>
<li>It encountered an exception</li>
</ul>
By default, an exception prints a stack trace to <tt>stderr</tt> before
notifying listeners that the consumer has stopped. You can define
different behavior by setting an <a
<tt>ExceptionHandler</tt></a>, but the consumer is still stopped.<br>
<br>
The same listener that receives probe data generated by DTrace is also
notified when the consumer stops. This is a good place to close the
consumer:
<pre><tt><font color=#aaaaaa>
consumer.addConsumerListener(new ConsumerAdapter() {
public void dataReceived(DataEvent e) {
}</font>
public void consumerStopped(ConsumerEvent e) {
Consumer consumer = (Consumer)e.getSource();
}
}<font color=#aaaaaa>
});</font>
</tt></pre>
<br>
This releases the resources held by the consumer in all cases, i.e.
after it exits for <i>any</i> of the reasons listed above.<br>
<br>
You can request the last aggregate snapshot made by a stopped consumer,
as long as it has not yet been closed:
<pre><tt>
Aggregate a = consumer.getAggregate();
</tt></pre>
<br>
Note however that any aggregation that has already appeared in a <a
<tt>PrintaRecord</tt></a> as a result of the <tt>printa()</tt> action
action will not be included in the requested aggregate.
<h2><a name="Learning_DTrace"></a>Learning More</h2>
<br>
The <a href="http://www.opensolaris.org/os/community/dtrace/">
OpenSolaris DTrace page</a> has links to resources to help you learn
DTrace. In particular, you should read the <a
href="http://docs.sun.com/db/doc/817-6223"><i>Solaris Dynamic Tracing
Guide</i></a>.<br>
<br>
Try the example Java programs on this page with other D scripts. You
executable script. You may want to remove <tt>profile:::tick*</tt>
clauses if you plan to use the <tt>Consumer</tt> <a
href="/api/org/opensolaris/os/dtrace/Consumer.html#getAggregate%28%29">
<tt>getAggregate()</tt></a> method and control the aggregating interval
programmatically. If the script uses the pre-compiler, you will need to
call the <tt>Consumer</tt> <a
href="/api/org/opensolaris/os/dtrace/Consumer.html#setOption%28java.lang.String%29">
<tt>setOption()</tt></a> method with the <a
href="/api/org/opensolaris/os/dtrace/Option.html#cpp">
<br>
To quickly familiarize yourself with the Java DTrace API, take a look at
<br>
<a href="#Quick_Start_Guide_to_the_Java_DTrace_API_">Back to top</a><br>
<br>
</body>
</html>