Bug 58299 - Add interrupt Timer
Summary: Add interrupt Timer
Alias: None
Product: JMeter
Classification: Unclassified
Component: Main (show other bugs)
Version: 2.13
Hardware: All All
: P2 enhancement (vote)
Target Milestone: ---
Assignee: JMeter issues mailing list
Depends on:
Reported: 2015-08-28 14:09 UTC by Sebb
Modified: 2015-12-14 20:41 UTC (History)
0 users

Sample Test case (4.92 KB, application/xml)
2015-08-29 12:59 UTC, Milamber

Note You need to log in before you can comment on or make changes to this bug.
Description Sebb 2015-08-28 14:09:20 UTC
Samplers such as HTTP can take a long time to return all the date.

There are timeout settings, but these apply to each network message, not the overall sample.

Once the sampler has completed, the duration assertion can be used to flag overlong samples, but this could take a very long time.

Rather than update every sampler and its GUI to support an extra timeout option, it is possible to use the Interruptible.interrupt() method in a Timer.

Timers run before the sampler to which they apply, so they can schedule a future task to invoke the interrupt() method on the sampler.
The following is a very basic BeanShell Timer implementation:

==== cut here =====
import java.util.concurrent.Executors;
run=new Runnable() {
    public void run() {
tpool = Executors.newScheduledThreadPool(1);
// schedule the interrupt to occur in 5 seconds from now
tpool.schedule(run, 5L, java.util.concurrent.TimeUnit.SECONDS);
return 0; // don't actually wait
===== cut here =====

It would be useful to have a built-in Timer to perform the same function.
Comment 1 Sebb 2015-08-28 19:09:39 UTC
URL: http://svn.apache.org/r1698393
Add interrupt Timer
Bugzilla Id: 58299

    jmeter/trunk/src/components/org/apache/jmeter/timers/InterruptTimer.java   (with props)
    jmeter/trunk/src/components/org/apache/jmeter/timers/gui/InterruptTimerGui.java   (with props)
    jmeter/trunk/xdocs/images/screenshots/timers/interrupt_timer.png   (with props)
Comment 2 Milamber 2015-08-28 20:03:51 UTC
URL: http://svn.apache.org/r1698403
Add French localization for the new Interrupt timer
(and littel alphabetical ordering in the keys)
Bugzilla Id: 58299

Comment 3 Milamber 2015-08-28 20:06:45 UTC
When I test the new Interrupt timer, I have the message on stdout:

MEvent. CASE!

and this on the sampler response:

java.net.SocketException: Socket closed
	at java.net.SocketInputStream.socketRead0(Native Method)
	at java.net.SocketInputStream.read(SocketInputStream.java:152)
	at java.net.SocketInputStream.read(SocketInputStream.java:122)
	at sun.security.ssl.InputRecord.readFully(InputRecord.java:442)
	at sun.security.ssl.InputRecord.read(InputRecord.java:480)
	at sun.security.ssl.SSLSocketImpl.readRecord(SSLSocketImpl.java:927)
	at sun.security.ssl.SSLSocketImpl.performInitialHandshake(SSLSocketImpl.java:1312)
	at sun.security.ssl.SSLSocketImpl.startHandshake(SSLSocketImpl.java:1339)
	at sun.security.ssl.SSLSocketImpl.startHandshake(SSLSocketImpl.java:1323)
	at org.apache.http.conn.ssl.SSLSocketFactory.connectSocket(SSLSocketFactory.java:436)
	at org.apache.http.impl.conn.DefaultClientConnectionOperator.openConnection(DefaultClientConnectionOperator.java:180)
	at org.apache.http.impl.conn.ManagedClientConnectionImpl.open(ManagedClientConnectionImpl.java:294)
	at org.apache.jmeter.protocol.http.sampler.MeasuringConnectionManager$MeasuredConnection.open(MeasuringConnectionManager.java:107)
	at org.apache.http.impl.client.DefaultRequestDirector.tryConnect(DefaultRequestDirector.java:643)
	at org.apache.http.impl.client.DefaultRequestDirector.execute(DefaultRequestDirector.java:479)
	at org.apache.http.impl.client.AbstractHttpClient.execute(AbstractHttpClient.java:906)
	at org.apache.http.impl.client.AbstractHttpClient.execute(AbstractHttpClient.java:805)
	at org.apache.jmeter.protocol.http.sampler.HTTPHC4Impl.executeRequest(HTTPHC4Impl.java:521)
	at org.apache.jmeter.protocol.http.sampler.HTTPHC4Impl.sample(HTTPHC4Impl.java:329)
	at org.apache.jmeter.protocol.http.sampler.HTTPSamplerProxy.sample(HTTPSamplerProxy.java:74)
	at org.apache.jmeter.protocol.http.sampler.HTTPSamplerBase.followRedirects(HTTPSamplerBase.java:1518)
	at org.apache.jmeter.protocol.http.sampler.HTTPSamplerBase.resultProcessing(HTTPSamplerBase.java:1599)
	at org.apache.jmeter.protocol.http.sampler.HTTPAbstractImpl.resultProcessing(HTTPAbstractImpl.java:463)
	at org.apache.jmeter.protocol.http.sampler.HTTPHC4Impl.sample(HTTPHC4Impl.java:408)
	at org.apache.jmeter.protocol.http.sampler.HTTPSamplerProxy.sample(HTTPSamplerProxy.java:74)
	at org.apache.jmeter.protocol.http.sampler.HTTPSamplerBase.sample(HTTPSamplerBase.java:1147)
	at org.apache.jmeter.protocol.http.sampler.HTTPSamplerBase.sample(HTTPSamplerBase.java:1136)
	at org.apache.jmeter.threads.JMeterThread.process_sampler(JMeterThread.java:434)
	at org.apache.jmeter.threads.JMeterThread.run(JMeterThread.java:261)
	at java.lang.Thread.run(Thread.java:745)

None to think that is the Interrupt timer in action? Normal?
Comment 4 Sebb 2015-08-29 11:49:03 UTC
I do not see the same problem.
However running as non-GUI did show that the executor threads need to be daemon

URL: http://svn.apache.org/r1698462
Add interrupt Timer
Ensure threads are created as daemons
Add some extra debug info
Bugzilla Id: 58299


Can you provide a sample JMX that shows the issue?
Comment 5 Milamber 2015-08-29 12:59:34 UTC
Created attachment 33053 [details]
Sample Test case

Here the sample test
Comment 6 Sebb 2015-08-29 13:50:32 UTC
I see what you mean now; the response does contain the stack trace.
Yes, that is expected when the request is aborted.

The same would happen when stopping the test.

BTW OpenJDK reports "MEvent. CASE!" sometimes:

There is a minor problem with the implementation: if the timeout is set too short, it may fail to cancel the first HTTP sample. That is because the sampler can take a while to initialise itself. The timer can expire before the sample itself is started, so interrupt() does nothing.

A work-round is to add a dummy HTTP request (without timeout) in a setUp group to do the init.

Handling one-off init is probably best addressed in a separate issue, as it would be better if the first sample time were more accurate.
Comment 7 Sebb 2015-12-14 20:41:54 UTC
Work on this was moved to a separate branch.
The code was reworked to use a Pre-Processor instead.
See Bug 58736