It would be nice to have a mechanism to keep code formatting in RadarGun uniform, as proper formatting = happy reviewer. Investigate ways to achieve this.

Some ideas:

• Code formatting templates for major IDEs
• Look into https://maven.apache.org/plugins/maven-checkstyle-plugin/ (thanks @rvansa for his idea)

RG tries to read original server config file and fails with NPE while doing so:

07:58:02,532 ERROR [org.radargun.service.ServerConfigurationProvider] (sc-main) Error while reading configuration file (/standalone/configuration/st-indexed-sync.xml)
java.lang.NullPointerException
        at org.radargun.utils.Utils.readAsBytes(Utils.java:189)
        at org.radargun.service.AbstractConfigurationProvider.loadConfigFile(AbstractConfigurationProvider.java:39)
        at org.radargun.service.AbstractConfigurationProvider.getOriginalConfigs(AbstractConfigurationProvider.java:28)
        at org.radargun.stages.lifecycle.ServiceStartStage.executeOnSlave(ServiceStartStage.java:82)
        at org.radargun.SlaveBase.scenarioLoop(SlaveBase.java:86)
        at org.radargun.SlaveBase$ScenarioRunner.run(SlaveBase.java:150)

Create a wiki page describing all the options in #{ ... } and ${ ... } syntax, functions, behaviour with lists etc. Also, document all the built-in variables (group.size, cluster.size, group.foo.size, env.ENV_VAR, random.int....)

Add a hudson plugin so that CBF can be started from HUDSON.

Hi @rvansa @mcimbora @alanfx, this is just a dump of the necessary steps I got from Radim on running the dist.sh script locally. Could be added to the wiki, either to the 5 minute tutorial, or as a separate page. (I don't have the rights to edit the wiki, and Github doesn't support PR's for wiki.)

Title: Using the dist.sh script on a single machine (Fedora 22):

1. Create virtual interfaces (eth0 is some physical interface)
   sudo ifconfig eth0:1 192.168.11.101 netmask 255.255.255.0
   sudo ifconfig eth0:2 192.168.11.102 netmask 255.255.255.0
   sudo ifconfig eth0:3 192.168.11.103 netmask 255.255.255.0
   sudo ifconfig eth0:4 192.168.11.104 netmask 255.255.255.0
   sudo ifconfig eth0:5 192.168.11.105 netmask 255.255.255.0
   sudo ifconfig eth0:6 192.168.11.106 netmask 255.255.255.0
   sudo ifconfig eth0:7 192.168.11.107 netmask 255.255.255.0
   sudo ifconfig eth0:8 192.168.11.108 netmask 255.255.255.0
   sudo ifconfig eth0:9 192.168.11.109 netmask 255.255.255.0
2. Make sshd listen on the new interfaces - add the following to /etc/ssh/sshd_config
   ListenAddress 192.168.11.101
   ListenAddress 192.168.11.102
   ListenAddress 192.168.11.103
   ListenAddress 192.168.11.104
   ListenAddress 192.168.11.105
   ListenAddress 192.168.11.106
   ListenAddress 192.168.11.107
   ListenAddress 192.168.11.108
   ListenAddress 192.168.11.109
3. Create aliases - add the following to /etc/hosts
   192.168.11.101 test1
   192.168.11.102 test2
   192.168.11.103 test3
   192.168.11.104 test4
   192.168.11.105 test5
   192.168.11.106 test6
   192.168.11.107 test7
   192.168.11.108 test8
   192.168.11.109 test9
4. Restart everything
   sudo systemctl restart network
   sudo systemctl restart sshd
5. If you don't have a passwordless ssh key, you will need to create one. Make sure you set an output file other than ~/.ssh/id_rsa, so as not to override your old key. Other than that, just hit enter.
   ssh-keygen -t rsa
6. Test it
   ssh -i --path to the passwordless key-- test1 echo "Interfaces set up"
7. Finally, run Radargun with the dist.sh script
   sh bin/dist.sh -t -r 'ssh -q -o "StrictHostKeyChecking false" -i ~/.ssh/<passwordless_key>' test1 test2 test3

It's been a long while since there was the 1.0.0 release, and we should probably do another one. There has been a ton of improvements and no plan, so, I'd like to ask you what's your opinion: is there something you want to do before this thick line of release? There's no hurry, but we should start looking into final feature list of 1.1.0, and possibly, some ideas for 1.2.0.
@alanfx @mlinhard @andyuk1986

Refactor Query and ContinuousQuery interface so that the builder of the QueryBuilder is separate interface and ContinuousQueryStage is not dependent on Queryable trait.

Running client-server scenarios using RG master branch fails with

Caused by: java.lang.NoClassDefFoundError: com/sun/tools/attach/AttachNotSupportedException
        at org.radargun.service.JavaProcessService.createConnectionProvider(JavaProcessService.java:37)
        ... 7 more
Caused by: java.lang.ClassNotFoundException: com.sun.tools.attach.AttachNotSupportedException

2.X branch doesn't seems to be affected by this issue and it runs without any problem. I used Oracle 1.8. Updating com.sun.tool dependency to 1.8.0 in plugins/process/pom.xml doesn't solve the issue.

right now we have a check to verify that memory stays the same between runs. Do the same for the number of threads. This will give us the additional guarantee that we do a proper cleanup.

2014-05-09 17:30:36,442 [main] ERROR [org.radargun.Master] Error in reporter html
java.lang.NullPointerException
at sun.awt.X11FontManager.getDefaultPlatformFont(X11FontManager.java:779)
at sun.font.SunFontManager$2.run(SunFontManager.java:432)
at java.security.AccessController.doPrivileged(Native Method)
at sun.font.SunFontManager.(SunFontManager.java:375)
at sun.awt.X11FontManager.(X11FontManager.java:32)
at sun.reflect.NativeConstructorAccessorImpl.newInstance0(Native Method)
at sun.reflect.NativeConstructorAccessorImpl.newInstance(NativeConstructorAccessorImpl.java:57)
at sun.reflect.DelegatingConstructorAccessorImpl.newInstance(DelegatingConstructorAccessorImpl.java:45)
at java.lang.reflect.Constructor.newInstance(Constructor.java:525)
at java.lang.Class.newInstance0(Class.java:372)
at java.lang.Class.newInstance(Class.java:325)
at sun.font.FontManagerFactory$1.run(FontManagerFactory.java:83)
at java.security.AccessController.doPrivileged(Native Method)
at sun.font.FontManagerFactory.getInstance(FontManagerFactory.java:74)
at java.awt.Font.getFont2D(Font.java:490)
at java.awt.Font.defaultLineMetrics(Font.java:2143)
at java.awt.Font.getLineMetrics(Font.java:2213)
at org.jfree.chart.axis.DateAxis.estimateMaximumTickLabelWidth(DateAxis.java:1437)
at org.jfree.chart.axis.DateAxis.selectHorizontalAutoTickUnit(DateAxis.java:1350)
at org.jfree.chart.axis.DateAxis.selectAutoTickUnit(DateAxis.java:1325)
at org.jfree.chart.axis.DateAxis.refreshTicksHorizontal(DateAxis.java:1597)
at org.jfree.chart.axis.DateAxis.refreshTicks(DateAxis.java:1537)
at org.jfree.chart.axis.DateAxis.draw(DateAxis.java:1836)
at org.jfree.chart.plot.XYPlot.drawAxes(XYPlot.java:3688)
at org.jfree.chart.plot.XYPlot.draw(XYPlot.java:3186)
at org.jfree.chart.JFreeChart.draw(JFreeChart.java:1242)
at org.jfree.chart.JFreeChart.createBufferedImage(JFreeChart.java:1412)
at org.jfree.chart.JFreeChart.createBufferedImage(JFreeChart.java:1392)
at org.radargun.reporting.html.TimelineChart.saveChart(TimelineChart.java:130)
at org.radargun.reporting.html.TimelineDocument.writeTimelines(TimelineDocument.java:119)
at org.radargun.reporting.html.HtmlReporter.run(HtmlReporter.java:65)
at org.radargun.Master.run(Master.java:94)
at org.radargun.LaunchMaster.main(LaunchMaster.java:32)
2014-05-09 17:30:36,443 [main] INFO [org.radargun.Master] All reporters have been executed, exiting.
2014-05-09 17:30:36,446 [Thread-0] INFO [org.radargun.ShutDownHook] Master process is being shutdown

In client-server mode RG fails to start as it waits for more servers in the cluster than actually configured. It seems it counts also client into cluster size, e.g. for this configuration

<cluster name="1client">
  <group name="server" size="2" />
  <group name="client" size="1" />
</cluster>

if expects cluster of size 3:

[org.radargun.stages.lifecycle.LifecycleHelper] (sc-main) Number of members=[slave1/clustered(local=true, coord=true), slave0/clustered(local=false, coord=false)] is not the one expected: 3

Radargun 3.0 is Java 8+, therefore we can use lambdas from now on. Review where Projections is used and replace the uses with lambdas where appropriate.

@rmacor I cannot assign that to you since you don't have write permissions in Radargun, but this could be a nice introduction into lambdas for you, if you're interested.

I found this benching Infinispan under 4 nodes using a TCP stack. Hangs for hours until I kill the processes.
Stack trace on master:
"main" prio=10 tid=0x08059000 nid=0x7055 runnable [0xf745a000]
java.lang.Thread.State: RUNNABLE
at sun.nio.ch.EPollArrayWrapper.epollWait(Native Method)
at sun.nio.ch.EPollArrayWrapper.poll(EPollArrayWrapper.java:210)
at sun.nio.ch.EPollSelectorImpl.doSelect(EPollSelectorImpl.java:65)
at sun.nio.ch.SelectorImpl.lockAndDoSelect(SelectorImpl.java:69)
- locked <0xdf0d5490> (a sun.nio.ch.Util$1)
- locked <0xdf0d5480> (a java.util.Collections$UnmodifiableSet)
- locked <0xdf0d50e0> (a sun.nio.ch.EPollSelectorImpl)
at sun.nio.ch.SelectorImpl.select(SelectorImpl.java:80)
at sun.nio.ch.SelectorImpl.select(SelectorImpl.java:84)
at org.cachebench.Master.startCommunicationWithSlaves(Master.java:154)
at org.cachebench.Master.start(Master.java:69)
at org.cachebench.LaunchMaster.main(LaunchMaster.java:28)
cluster01:
"main" prio=10 tid=0x08059000 nid=0x319c waiting on condition [0xf745b000]
java.lang.Thread.State: TIMED_WAITING (parking)
at sun.misc.Unsafe.park(Native Method)
- parking to wait for <0xdee93940> (a java.util.concurrent.FutureTask$Sync)
at java.util.concurrent.locks.LockSupport.parkNanos(LockSupport.java:198)
at java.util.concurrent.locks.AbstractQueuedSynchronizer.doAcquireSharedNanos(AbstractQueuedSynchronizer.java:947)
at java.util.concurrent.locks.AbstractQueuedSynchronizer.tryAcquireSharedNanos(AbstractQueuedSynchronizer.java:1239)
at java.util.concurrent.FutureTask$Sync.innerGet(FutureTask.java:227)
at java.util.concurrent.FutureTask.get(FutureTask.java:91)
at org.cachebench.Slave.startCommunicationWithMaster(Slave.java:126)
at org.cachebench.Slave.start(Slave.java:57)
at org.cachebench.Slave.main(Slave.java:212)
cluster02 .. cluster04
"main" prio=10 tid=0x08059000 nid=0xa50 runnable [0xf745b000]
java.lang.Thread.State: RUNNABLE
at sun.nio.ch.EPollArrayWrapper.epollWait(Native Method)
at sun.nio.ch.EPollArrayWrapper.poll(EPollArrayWrapper.java:210)
at sun.nio.ch.EPollSelectorImpl.doSelect(EPollSelectorImpl.java:65)
at sun.nio.ch.SelectorImpl.lockAndDoSelect(SelectorImpl.java:69)
- locked <0xb42c04c0> (a sun.nio.ch.Util$1)
- locked <0xb42c1120> (a java.util.Collections$UnmodifiableSet)
- locked <0xb42c0480> (a sun.nio.ch.EPollSelectorImpl)
at sun.nio.ch.SelectorImpl.select(SelectorImpl.java:80)
at sun.nio.ch.SelectorImpl.select(SelectorImpl.java:84)
at org.cachebench.Slave.startCommunicationWithMaster(Slave.java:66)
at org.cachebench.Slave.start(Slave.java:57)
at org.cachebench.Slave.main(Slave.java:212)

Lately I've been thinking about using freemarker templating engine to generate HTML reports. IMO current approach is a bit clumsy - generating HTML pages in java code makes reporting quite complex. If we could separate view&model, it would make it easier to perform further changes.

See POC at:
mcimbora@0a771fe

Creating index page was pretty straightforward. When writing page with test results, after a while I realized that I'm just moving logic from backend to frontend, which is not really what we want. After some refactoring, I think it will be feasible just to plug in pojos with data to be displayed. Just wanted to share it here to get your feedback. Do you see current approach as sufficient or do you find using templates appealing (I hope you do!)?

@alanfx @rvansa ?

This simple web based console (deployed as a WAR file) could be used to attach to the master process and manage slave nodes.
Additional commands would be needed for the master to be able to communicate with slaves to retrieve information, which the master would then use to update the GUI.
GUI functionality would include:
Slave status
Details of test being run (including fill view of XML config file)
Details of current run being performed
If running from the GUI, we probably want to be able to launch the master in the same VM as the GUI so the WAR file would need to include the CBF jars and start the master process (perhaps a button to launch, re-launch the master?)
In future we could also include the ability to kick-start slaves by making use of some Java-based SSH tool. This would then replace the SSH scripts. :)

In RadarGun 2.0, cluster size often should be replaced by group size. Check the use and replace.

It would be nice to have a historical perf graph: given 10 tests, run over 2 months for example, plot a graph showing performance over the 10 test runs.

org.radargun.reporting.html.TimelineDocument Failed to generate on of the charts:
java.util.concurrent.ExecutionException: java.lang.ArrayIndexOutOfBoundsException: 1000
at java.util.concurrent.FutureTask.report(FutureTask.java:122)
at java.util.concurrent.FutureTask.get(FutureTask.java:188)
at org.radargun.reporting.html.TimelineDocument.writeTimelines(TimelineDocument.java:186)
at org.radargun.reporting.html.HtmlReporter.writeTimelineDocuments(HtmlReporter.java:169)
at org.radargun.reporting.html.HtmlReporter.run(HtmlReporter.java:50)
at org.radargun.Master.run(Master.java:117)
at org.radargun.LaunchMaster.main(LaunchMaster.java:32)
Caused by: java.lang.ArrayIndexOutOfBoundsException: 1000
at org.radargun.reporting.html.TimelineChart.setEvents(TimelineChart.java:101)
at org.radargun.reporting.html.TimelineDocument$1.call(TimelineDocument.java:131)
at org.radargun.reporting.html.TimelineDocument$1.call(TimelineDocument.java:124)
at java.util.concurrent.FutureTask.run(FutureTask.java:262)
at java.util.concurrent.ThreadPoolExecutor.runWorker(ThreadPoolExecutor.java:1145)
at java.util.concurrent.ThreadPoolExecutor$Worker.run(ThreadPoolExecutor.java:615)
at java.lang.Thread.run(Thread.java:745)

org.radargun.Master Error in reporter org.radargun.reporting.html.HtmlReporter@5e838dcd
java.lang.NullPointerException
at org.radargun.reporting.html.ReportDocument.writeResult(ReportDocument.java:146)
at org.radargun.reporting.html.TestReportDocument.writeTest(TestReportDocument.java:51)
at org.radargun.reporting.html.HtmlReporter.writeTestReportDocuments(HtmlReporter.java:152)
at org.radargun.reporting.html.HtmlReporter.run(HtmlReporter.java:51)
at org.radargun.Master.run(Master.java:117)

The existing clustered test is a session emulation test. This emulates the HTTP session access pattern which assumes session affinity, and this means threads and nodes do not compete for the same keys.
We should also create a ConcurrentClusteredSessionTest?, which will force a certain degree of overlap of key accesses, from different threads as well as nodes.

Let's unify coding style through automation tools rather than complaining during review.

https://github.com/michalrus/git-hooks-code-autoformat could give at least inspiration how to do that.

New plugin to test JGroups 3.6.6

2014-05-09 17:13:47,151 [main] ERROR [org.radargun.Master] Error in reporter html
org.jfree.data.general.SeriesException: You are attempting to add an observation for the time period Wed Dec 31 18:00:11 CST 1969 but the series already contains an observation for that time period. Duplicates are not permitted. Try using the addOrUpdate() method.
at org.jfree.data.time.TimeSeries.add(TimeSeries.java:612)
at org.jfree.data.time.TimeSeries.add(TimeSeries.java:684)
at org.jfree.data.time.TimeSeries.add(TimeSeries.java:670)
at org.radargun.reporting.html.TimelineChart.setEvents(TimelineChart.java:98)
at org.radargun.reporting.html.TimelineDocument.writeTimelines(TimelineDocument.java:114)
at org.radargun.reporting.html.HtmlReporter.run(HtmlReporter.java:65)
at org.radargun.Master.run(Master.java:94)
at org.radargun.LaunchMaster.main(LaunchMaster.java:32)
2014-05-09 17:13:47,151 [main] INFO [org.radargun.Master] All reporters have been executed, exiting.
2014-05-09 17:13:47,155 [Thread-0] INFO [org.radargun.ShutDownHook] Master process is being shutdown

NPE in isRunning() as ch might be null

Looks like 95th & 99th percentiles are missing in csv reporter output, let's add those. Additionally check whether other data is correct (e.g. 'threads' doesn't seem to be correct when thread count increases throughout the test).
I'd be nice to add some unit tests to check the output.

An additional metric could optionally be recorded: memory usage on each node. Need to think about the most efficient way to do this (so we don't affect performance too much). Perhaps btrace?
Also need to think of a way to graphically represent this.

Continuous query stage should support adding multiple continuous queries, for reasoning see @rvansa 's comment

add a script to only generate a report, not run the benchmark.
Workaround: for now just comment out, in the benchmark.xml, all stages, update the report xml element the way you want and then run benchmark.sh

such as

• CPU usage
• memory usage
• thread context switches (e.g. using vmstat)
• dropped packets
  etc
  This should be configurable, and might not be usable on all platforms (e.g. vmstat doesn't exist on Windows).

There is an infinite cycle in DistStageAck.setRemoteException() method. I will look into it and create tests when I have a moment.

Merge functionality in pom-links.xml to pom.xml (-Dlinks=true will make the zip with links)

Hibernate's 2nd level cache is a major use case for many distributed caches. We should create a benchmark for this, possibly incorporating the Hibernate libraries and Hibernate's 2nd level cache interfaces to be as realistic as possible - but not exercising Hibernate itself, or any actual database.

Thanks @alanfx for proposal. Currently repeat loop exposes its state only via repeat.counter variable. It would be nice if repeat loop enabled us to loop over pre-defined list of custom values & use their reference inside the loop.

e.g.
<repeat vars="size:[1, 2, 5, 111]"> <stage1 value-size="${repeat.size}"/> </repeat>

Multiple properties could be defined if we could enforce property lists to be of the same sizes for all defined properties.

<repeat vars="size:[1, 2, 5, 111];count:[3, 10, 50, 68]"> <stage1 value-size="${repeat.size}" count="${repeat.count}"/> </repeat>

It looks like only one pattern could be used at the same time (i.e. from-to/times vs var).

@alanfx @rvansa Wdyt?

As shall scripts should be used for operating RadarGun, it should propagate RadarGun exit code or have an option for it.

Check that pooling keys (default) provides any performance difference compared to generating the keys before each request, and possibly remove the pooling.

Update POMs to use Maven's Release Plugin (the same way infinispan-archetypes does)
Release plugin to generate version data for Version.java

Zip distribution of RadarGun doesn't contains some */bin directories, e.g. plugins/process/bin/, while from parts of RG depends on the scripts in this dir (e.g. ProcessLifecycle.java#L138)

Perhaps in a summary box or a label somewhere, we should show some extra info on the test: num loops, num threads, read/write ratio, payload size (bytes), payload type (Strings, byte array, etc).

... except using different thread counts, put/get ratios and num loops. This will ensure that the correct code paths are getting warmed up and optimized but the JIT compiler.

Currently, all stages are inserted into single XSD file, that means all of them are in single XML namespace. With move towards supporting multiple types of plugins (not just Caches, but also JPA implementors and possibly others). We should allow stages to declare their namespaces to prevent naming conflicts.

Currently the variables in scenarios are expanded only when passed as java properties or when starts with env. prefix. It would be nice to expand env. variables directly, without needs to add env. prefix.

Investigate: https://github.com/brianfrankcooper/YCSB/wiki/
also look at faban

The link referred to on the initial GitHub repo page (https://github.com/radargun/radargun) is a broken link. Need to setup GitHub pages to redirect to the wiki?

Running test with RegisterListenersStage fails with

08:26:19,050 ERROR [org.radargun.Master] (main) Processing acks on master failed
java.lang.ClassCastException: org.radargun.DistStageAck cannot be cast to org.radargun.stages.cache.listeners.cluster.RegisterListenersStage$ListenersAck
        at org.radargun.stages.cache.listeners.cluster.RegisterListenersStage.processAckOnMaster(RegisterListenersStage.java:84)
        at org.radargun.Master.executeDistStage(Master.java:250)
        at org.radargun.Master.executeStage(Master.java:144)
        at org.radargun.Master.run(Master.java:89)
        at org.radargun.LaunchMaster.main(LaunchMaster.java:32)

@rmacor Could you look into this?

Thrown when running benchmark-cs.xml

12:32:36,732 ERROR [freemarker.runtime] (main) Error executing FreeMarker template
FreeMarker template error:
For ".." right-hand operand: Expected a number, but this has evaluated to a sequence (Collections$UnmodifiableRandomAccessList wrapped into f.t.DefaultListAdapter):
==> groups  [in template "timelineReport.ftl" at line 91, column 20]

----
FTL stack trace ("~" means nesting-related):
        - Failed at: #list 0..groups as groupID  [in template "timelineReport.ftl" at line 91, column 10]
----

Java stack trace (for programmers):
----
freemarker.core.NonNumericalException: [... Exception message was already printed; see it above ...]
        at freemarker.core.Expression.modelToNumber(Expression.java:105)
        at freemarker.core.Expression.evalToNumber(Expression.java:98)
        at freemarker.core.Range._eval(Range.java:50)
        at freemarker.core.Expression.eval(Expression.java:78)
        at freemarker.core.IteratorBlock.acceptWithResult(IteratorBlock.java:69)
        at freemarker.core.IteratorBlock.accept(IteratorBlock.java:64)
        at freemarker.core.Environment.visitByHiddingParent(Environment.java:345)
        at freemarker.core.ConditionalBlock.accept(ConditionalBlock.java:48)
        at freemarker.core.Environment.visit(Environment.java:324)
        at freemarker.core.MixedContent.accept(MixedContent.java:54)
        at freemarker.core.Environment.visit(Environment.java:324)
        at freemarker.core.Environment.process(Environment.java:302)
        at freemarker.template.Template.process(Template.java:325)
        at org.radargun.reporting.html.HtmlReporter.processTemplate(HtmlReporter.java:240)
        at org.radargun.reporting.html.HtmlReporter.writeTimelineDocuments(HtmlReporter.java:189)
        at org.radargun.reporting.html.HtmlReporter.run(HtmlReporter.java:62)
        at org.radargun.Master.run(Master.java:135)
        at org.radargun.LaunchMaster.main(LaunchMaster.java:32)

In order to get more accurate values for a benchmarks, an approach would be to run it several times and merge the results in an 'average' like report.

I am hitting this exception

07:37:22 java.lang.IllegalStateException: Something not sent to slaves yet: {java.nio.channels.SocketChannel[connected local=/10.0.0.74:2103 remote=/10.0.0.74:56234]=java.nio.HeapByteBuffer[pos=0 lim=59 cap=8192]}
07:37:22 at org.radargun.RemoteSlaveConnection.clearBuffer(RemoteSlaveConnection.java:131)
07:37:22 at org.radargun.RemoteSlaveConnection.runStage(RemoteSlaveConnection.java:158)
07:37:22 at org.radargun.Master.executeDistStage(Master.java:236)
07:37:22 at org.radargun.Master.executeStage(Master.java:144)
07:37:22 at org.radargun.Master.run(Master.java:98)
07:37:22 at org.radargun.LaunchMaster.main(LaunchMaster.java:32)
07:37:22 04:37:22,259 WARN org.radargun.RemoteSlaveConnection Failed to send termination to slaves.
07:37:22 java.lang.IllegalStateException: Something not sent to slaves yet: {java.nio.channels.SocketChannel[connected local=/10.0.0.74:2103 remote=/10.0.0.74:56234]=java.nio.HeapByteBuffer[pos=0 lim=59 cap=8192]}
07:37:22 at org.radargun.RemoteSlaveConnection.clearBuffer(RemoteSlaveConnection.java:131)
07:37:22 at org.radargun.RemoteSlaveConnection.mcastObject(RemoteSlaveConnection.java:144)
07:37:22 at org.radargun.RemoteSlaveConnection.release(RemoteSlaveConnection.java:252)
07:37:22 at org.radargun.Master.run(Master.java:131)
07:37:22 at org.radargun.LaunchMaster.main(LaunchMaster.java:32)
07:37:22 04:37:22,260 INFO org.radargun.ShutDownHook Master process is being shutdown
07:37:22 Master 3383 finished with value 127

Currently Radargun doesn't work if one follows the Quickstart (https://github.com/radargun/radargun/wiki/Five-Minute-Tutorial). I had to make the following changes to get it working:

• comment out lines relating to infinispan53 in pom.xml (it tries to download Infinispan 5.3.0-SNAPSHOT from the repos which doesn't exist - perhaps change it so that it uses 5.3.0rc2?)
• updated benchmark.xml so that it tests infinispan52 rather than a very old version (4)
• copied all the missing files over from plugins/infinispan51 to plugins/infinispan52
• updated dist-sync.xml so that namedCache contains "testCache" rather than "x"
• also, the quickstart mentions cachebench.log which is now called radargun.log

Data loading (using load-data stage) includes multiple steps (getting next key id, key/value generation). Investigate performance impact of the steps & provide optimization if possible.

Hi Team,
I am very much interested in running Benchmak with different products like Coherence and Terracotta. I didnt got much documentation todo that. Looking for some help in runing those, if possible send me detailed email with the steps to [email protected]. . Appreciate your help.
Thanks a lot!
Regards
Prasad Kasthuri