[infinispan-dev] Classloader leaks?

[Sanne Grinovero]

On 22 February 2017 at 21:20, Dennis Reed <dereed at redhat.com> wrote:
> Are those actually 2 million *unique* descriptors?
>
> I've seen lsof output that listed many duplicates for the same file
> descriptor (one for each thread?), making the list appear much larger
> than it really was.

Good point! You're right, I verified and all instances of e.g. the
jgroups jar were using the same FD, just a different thread id.

This is the full error I'm having, when running the tests from the
"infinispan-compatibility-mode-it" maven module:


java.lang.IllegalStateException: failed to create a child event loop
at io.netty.util.concurrent.MultithreadEventExecutorGroup.<init>(MultithreadEventExecutorGroup.java:88)
at io.netty.util.concurrent.MultithreadEventExecutorGroup.<init>(MultithreadEventExecutorGroup.java:58)
at io.netty.channel.MultithreadEventLoopGroup.<init>(MultithreadEventLoopGroup.java:51)
at io.netty.channel.nio.NioEventLoopGroup.<init>(NioEventLoopGroup.java:87)
at io.netty.channel.nio.NioEventLoopGroup.<init>(NioEventLoopGroup.java:82)
at io.netty.channel.nio.NioEventLoopGroup.<init>(NioEventLoopGroup.java:63)
at io.netty.channel.nio.NioEventLoopGroup.<init>(NioEventLoopGroup.java:51)
at org.jboss.resteasy.plugins.server.netty.NettyJaxrsServer.start(NettyJaxrsServer.java:239)
at org.infinispan.rest.NettyRestServer.start(NettyRestServer.java:81)
at org.infinispan.it.compatibility.CompatibilityCacheFactory.createRestCache(CompatibilityCacheFactory.java:199)
at org.infinispan.it.compatibility.CompatibilityCacheFactory.createRestMemcachedCaches(CompatibilityCacheFactory.java:137)
at org.infinispan.it.compatibility.CompatibilityCacheFactory.setup(CompatibilityCacheFactory.java:123)
at org.infinispan.it.compatibility.ByteArrayKeyReplEmbeddedHotRodTest.setup(ByteArrayKeyReplEmbeddedHotRodTest.java:87)
at sun.reflect.NativeMethodAccessorImpl.invoke0(Native Method)
at sun.reflect.NativeMethodAccessorImpl.invoke(NativeMethodAccessorImpl.java:62)
at sun.reflect.DelegatingMethodAccessorImpl.invoke(DelegatingMethodAccessorImpl.java:43)
at java.lang.reflect.Method.invoke(Method.java:498)
at org.testng.internal.MethodInvocationHelper.invokeMethod(MethodInvocationHelper.java:84)
at org.testng.internal.Invoker.invokeConfigurationMethod(Invoker.java:564)
at org.testng.internal.Invoker.invokeConfigurations(Invoker.java:213)
at org.testng.internal.Invoker.invokeConfigurations(Invoker.java:138)
at org.testng.internal.TestMethodWorker.invokeBeforeClassMethods(TestMethodWorker.java:175)
at org.testng.internal.TestMethodWorker.run(TestMethodWorker.java:107)
at org.testng.TestRunner.privateRun(TestRunner.java:767)
at org.testng.TestRunner.run(TestRunner.java:617)
at org.testng.SuiteRunner.runTest(SuiteRunner.java:348)
at org.testng.SuiteRunner.access$000(SuiteRunner.java:38)
at org.testng.SuiteRunner$SuiteWorker.run(SuiteRunner.java:382)
at org.testng.internal.thread.ThreadUtil$2.call(ThreadUtil.java:64)
at java.util.concurrent.FutureTask.run(FutureTask.java:266)
at java.util.concurrent.ThreadPoolExecutor.runWorker(ThreadPoolExecutor.java:1142)
at java.util.concurrent.ThreadPoolExecutor$Worker.run(ThreadPoolExecutor.java:617)
at java.lang.Thread.run(Thread.java:745)
Caused by: io.netty.channel.ChannelException: failed to open a new selector
at io.netty.channel.nio.NioEventLoop.openSelector(NioEventLoop.java:157)
at io.netty.channel.nio.NioEventLoop.<init>(NioEventLoop.java:148)
at io.netty.channel.nio.NioEventLoopGroup.newChild(NioEventLoopGroup.java:126)
at io.netty.channel.nio.NioEventLoopGroup.newChild(NioEventLoopGroup.java:36)
at io.netty.util.concurrent.MultithreadEventExecutorGroup.<init>(MultithreadEventExecutorGroup.java:84)
... 32 more
Caused by: java.io.IOException: Too many open files
at sun.nio.ch.IOUtil.makePipe(Native Method)
at sun.nio.ch.EPollSelectorImpl.<init>(EPollSelectorImpl.java:65)
at sun.nio.ch.EPollSelectorProvider.openSelector(EPollSelectorProvider.java:36)
at io.netty.channel.nio.NioEventLoop.openSelector(NioEventLoop.java:155)
... 36 more

Now that I know which metrics to look at, I see that before running
this specific module my system is consuming about 12K FDs, and
occasionally it just stays around that same level and the integration
tests will pass w/o any failure.
However most of the times when I run this module specifically, I see
the FD consumption increase during the test run and eventually fail
with the above error. The last sample I could take before failing was
around 15,5K, and not surprising as my limit is set to 16384.. so I
guess it could have attempted to grow further.

Tomorrow I'll try with higher limits, to see if I'm running with
barely enough for this testsuite or if I see it growing more. Still
sounds like a leak though, as the increase is quite significant..

Thanks,
Sanne

>
> -Dennis
>
>
> On 02/22/2017 02:25 PM, Sanne Grinovero wrote:
>> Hi all,
>>
>> our documentation suggest to raise the file limits to about 16K:
>> http://infinispan.org/docs/stable/contributing/contributing.html#running_and_writing_tests
>>
>> I already have this setup since years, yet I've been noticing errors such as:
>>
>> "Caused by: java.io.IOException: Too many open files"
>>
>> Today I decided to finally have a look, and I see that while running
>> the testsuite, my system's consumption of file descriptor raises
>> continuously, up to more than 2 millions.
>> (When not running the suite, I'm consuming 200K - that's including
>> IDEs and other FD hungry systems like Chrome)
>>
>> Trying to get some samples of these file descriptors, it looks like
>> it's really about open files. Jar files to be more precise.
>>
>> What puzzles me is that taking just one jar - jgroups for example - I
>> can count 7852 open instances of it, but distributed among a handful
>> of processes only.
>>
>> My guess is classloaders aren't being closed?
>>
>> Also: why did nobody else notice problems? Do you all have
>> reconfigured your system for unlimited FDs?
>>
>> Thanks,
>> Sanne
>> _______________________________________________
>> infinispan-dev mailing list
>> infinispan-dev at lists.jboss.org
>> https://lists.jboss.org/mailman/listinfo/infinispan-dev
>>
>
> _______________________________________________
> infinispan-dev mailing list
> infinispan-dev at lists.jboss.org
> https://lists.jboss.org/mailman/listinfo/infinispan-dev