Virtualizing Java in Java (jug.ru)

Virtualizing Java in Java

How?
Alexey Ragozin

jug.ru
12 Dec 2013

Virtualizing Java in Java
Single JVM hosting
•
•
•
•
•

multiple “pseudo JVMs”, which
Have independent system properties
Have independent statics
May have different classpath
May be forcefully terminated or
suspended/resumed

You can deploy whole distributed application
topology inside of JUnit test

All started with Oracle Coherence
Native Java distributed peer-to-peer key/value storage
With a lot of extension points
 Distributed data processing
 Listeners for remote events
 Read-through / write-through patterns
 Pluggable serialization
And inability to start two cluster members in one JVM

And we need to test
Behavior of very specific features
Serialization/deserialization defects
Data routing and collocation aspects
Code meant for distributed execution
Threading aspects
Classpath differences between cluster
processes
Cluster configuration tweaks

Multiplexing singletons
Custom classloader
Force to reload classes already loaded by
parent (second copy of class loaded)
Got our first cluster-in-JVM alive
 InheritentThreadLocal to add fancy stuff
– Multiplexing system properties
– Multiplexing multiplex console output

Heated competition
Three open source “cluster virtualizing” effort around
Oracle Coherence

 GridKit (ChTest)
code.google.com/p/gridkit
 Oracle tools
github.com/coherence-community/oracle-tools
 Little
www.littlegrid.net

“Distributing” test case
How to start The Application on “virtual cluster”?

Old school
 Main classes and command line arguments
 But if you need to do verification inside of
vinode?
 A separate main for each test case?

But how all this is relevant to me?
Normal client/server application
 You can use your real main classes instead of mocking server

Hadoop / HBase / Cassandra
 Distributed
 Deployment unfriendly
 Ship with single node – cut down versions

“Distributing” test case
Transparent remotting
client1.exec(new Runnable() {
@Override
public void run() {
NamedCache cache = CacheFactory.getCache(cacheName);
Assert.assertNull(cache.get("A"));
cache.put("A", "aaa");
}
});
@Override
public void run() {
Assert.assertEquals("aaa", cache.get("A"));
}
});

We want to test more!
We are in control class loading
— let’s tweak classpath on flight
 Inject resources
 Remove server classes from client
 Test different codebase versions

Backward compatibility testing

Master JVM, client regression test pack
[trunk version]
Case

Client
[version X]

Server
[trunk version]

Cross version tests

Master JVM, client regression test pack
[trunk version]
Case
Client
[version X]

Client
[version Y]

Server
[version Z]

Managing artifacts
… a bunch of black magic to find local repo
and managing classpath as easy as …
<plugin>
<groupId>org.apache.maven.plugins</groupId>
<artifactId>maven-dependency-plugin</artifactId>
<version>2.8</version>
<executions>
<execution>
<id>viconcurrent-0.7.15</id>
<phase>test-compile</phase>
<goals>
<goal>get</goal>
</goals>
<configuration>
<artifact>org.gridkit.lab:viconcurrent:0.7.15</artifact>
</configuration>
</execution>
</executions>
</plugin>

Managing artifacts
How to get needed artifact on local disk
- Maven will disallow two versions of same artifact
- but we can trick it …
ViNode node;
…
node.x(MAVEN).replace("org.gridkit.lab", "viconcurrent", "0.7.15");

Transitive dependencies are not included, though.

Transparent remotting
Own implementation of RMI
 Standard Java serialization
 Serializing of anonymous Runnable/Callable
 Auto export of Remote interfaces

Writing distributed code by convention!

Bidirectional communications
public interface RemotePut extends Remote {
public void put(Object key, Object value);
}


@SuppressWarnings("unused")
@Test
public void bidirectional_remoting() {
// Present for typical single node cluster
cloud.all().presetFastLocalCluster();


cloud.node("storage.**").localStorage(true);
cloud.node("client.**").localStorage(false);

}

// Simulates DefaultCacheServer based process
cloud.node("storage.**").autoStartServices();
// declaring specific nodes to be created
CohNode storage = cloud.node("storage.1");
CohNode client1 = cloud.node("client.1");
CohNode client2 = cloud.node("client.2");

RemotePut remoteService =
client1.exec(new Callable<RemotePut>() {
@Override
public RemotePut call() {
final NamedCache cache = CacheFactory.getCache(cacheName);

// now we have 3 specific nodes in cloud
// all of then will be initialized in parallel
cloud.all().ensureCluster();
final String cacheName = "distr-a";

return new RemotePut() {
@Override
public void put(Object key, Object value) {
cache.put(key, value);
}
};

@Override
@Override
}
};

}
});
}
});

remoteService.put("A", "aaa");


@Override
public void run() {
Assert.assertEquals("aaa", cache.get("A"));
}
});
}

Bidirectional communications
Extending java.rmi.Remotexec
will mark interface for auto export
}

@Override
@Override
}
};

}
});


Here we got a remote stub, not a
real implementation of interface

Unlike Java RMI, there is no need to
declare RemoteException for
every method
Result of callable will be serialized
and transferred to caller
Objects implementing remote
interfaces are automatically replaced
with remote stub during serialization
Call to a stub, will be converted to
“remote” call to instance we have
created in “virtualized” node few
lines above

Sneak peek: Instrumentation
 System.exit() – is still fatal
 Some cases need “virtual time”
 Tweaking monolithic code
 Fault injection
 Mock injection

PowerMock
 Recompiles everything (Coherence ～ 5000 classes)

AspectJ
 Static interceptors
ByteMan
 Using agent + weird language

ViNode node = ...
ViHookBuilder
.newCallSiteHook()
.onTypes(System.class)
.onMethod("exit")
.doReturn(null)
.apply(node);
node.exec(new Callable<Void>() {
@Override
public Void call() throws Exception {
System.exit(0);
return null;
}
});

LET’S TAKE A BREAK HERE

Questions?

From “virtual” to real cluster

Alexey Ragozin

jug.ru
12 Dec 2013

Virtual stuff is so good
Managing virtual nodes in deterministic way, in
Java, having all luxury of exception handling and
richness of libraries – feeling were so good …
I wished, I could rollout JVMs across real servers

Your network is Big JVM
Same API
3 types of nodes: in-process, local, remote
Transparent remotting
SSH to manage remote server
Automatic classpath replication (with caching)
Zero infrastructure
 Any OS for master host
 SSHd + JVM for slave hosts

New opportunities
Performance testing





deploy system under test
deploy load generators
deploy monitoring agents
gather all result in one place

Deployment (remote execution task for ANT)
Replace your putty with Java IDE
 log scrapping
 parallel execution

As easy as …
@Test
public void remote_hello_world() throws InterruptedException
{
ViManager cloud = CloudFactory.createSimpleSshCloud();
cloud.node("myserver.uk.db.com");
cloud.node("**").exec(new Callable<Void>() {
@Override
public Void call() throws Exception
{
String localHost = InetAddress.getLocalHost().toString();
System.out.println("Hi! I'm running on " + localHost);
return null;
}
});
}

Behind the scene
•
•
•
•

JSCh – SSH client (slightly patched)
Collected classpath artifacts
SCP jars to remote target
Start remote agent (JVM)
 stdOut / stdIn for master – agent communications

• Agent start slave process
• Slave start RMI node and connects (TCP) to master
 agent acts as TCP proxy

Death clock is ticking
Master JVM kills slave processes, unless
 SSH session was interrupted
 someone kill -9 master JVM
 master JVM has crashed (e.g. under debuger)
Death clock is ticking on slave though
 if master is not responding
 slave process will terminate itself

Performance testing
•
•
•
•
•

Master JVM is running on CI
JUnit to start test
Loaded generator farm is deployed on test servers
Monitoring agents deployed on application cluster
Metrics are buffered locally, then send to master and processed

Real test
• Four servers – application
• 50 servers – load farm
• Over 200 – slaves processes

Coding for 200+ processes
Driver - concept
• Driver – Java interface encapsulates test
action
• One way methods
• Friendly for remotting for parallel invokation
+ some utility for parallel execution, workflow
etc
Example:
https://gridkit.googlecode.com/svn/grid-lab/trunk/examples/zk-benchmark-sample

Monitoring
Sigar - http://www.hyperic.com/products/sigar
• collecting common system metrics
Attach API – built-n JVM monitoring
SJK – spin-off CLI tool for JVM diagnostics
• https://github.com/aragozin/jvm-tools

Links
NanoCloud
• https://code.google.com/p/gridkit/wiki/NanoCloudTutorial
• Maven Central: org.gridkit.lab:telecontrol-ssh:0.7.22
• http://blog.ragozin.info/2013/01/remote-code-execution-in-java-made.html

ANT task
• https://github.com/gridkit/gridant

ChTest (Coherence test tool)
• https://code.google.com/p/gridkit/wiki/ChTest
• Maven Central: org.gridkit.coherence-tools:chtest:0.2.4

Thank you
http://blog.ragozin.info
- my articles
http://code.google.com/p/gridkit
http://github.com/gridkit
- my open source code
http://aragozin.timepad.ru
- tech meetups in Moscow

Alexey Ragozin
alexey.ragozin@gmail.com

Virtualizing Java in Java (jug.ru)

More Related Content

What's hot

Viewers also liked

Similar to Virtualizing Java in Java (jug.ru)

More from aragozin

Recently uploaded

Virtualizing Java in Java (jug.ru)