Dawid Weiss created LUCENE-5573:
-----------------------------------
Summary: Deadlock during class loading/ initialization
Key: LUCENE-5573
URL: https://issues.apache.org/jira/browse/LUCENE-5573
Project: Lucene - Core
Issue Type: Bug
Reporter: Dawid Weiss
Assignee: Robert Muir
Fix For: 4.8, 5.0
Attachments: A.java, C.java, Main.java, X.java
It's always worth looking into those randomized failures.
http://builds.flonkings.com/job/Lucene-trunk-Linux-Java7-64-test-only/81259/console
Log quote:
{code}
[junit4] 2> ==== jstack at approximately timeout time ====
[junit4] 2> "Lucene Merge Thread #0" ID=25 RUNNABLE
[junit4] 2> at
org.apache.lucene.codecs.lucene45.Lucene45DocValuesProducer.getSortedSet(Lucene45DocValuesProducer.java:541)
[junit4] 2> at
org.apache.lucene.codecs.perfield.PerFieldDocValuesFormat$FieldsReader.getSortedSet(PerFieldDocValuesFormat.java:285)
[junit4] 2> at
org.apache.lucene.index.SegmentReader.getSortedSetDocValues(SegmentReader.java:500)
[junit4] 2> at
org.apache.lucene.index.SegmentMerger.mergeDocValues(SegmentMerger.java:204)
[junit4] 2> at
org.apache.lucene.index.SegmentMerger.merge(SegmentMerger.java:119)
[junit4] 2> at
org.apache.lucene.index.IndexWriter.mergeMiddle(IndexWriter.java:4068)
[junit4] 2> at
org.apache.lucene.index.IndexWriter.merge(IndexWriter.java:3664)
[junit4] 2> at
org.apache.lucene.index.ConcurrentMergeScheduler.doMerge(ConcurrentMergeScheduler.java:405)
[junit4] 2> at
org.apache.lucene.index.ConcurrentMergeScheduler$MergeThread.run(ConcurrentMergeScheduler.java:482)
[junit4] 2>
[junit4] 2>
"TEST-TestLucene45DocValuesFormat.testSortedSetVariableLengthVsUninvertedField-seed#[7362FE36DE729D42]"
ID=23 RUNNABLE
[junit4] 2> at
org.apache.lucene.index.SortedSetDocValues.<clinit>(SortedSetDocValues.java:72)
[junit4] 2> at
org.apache.lucene.index.DocTermOrds.iterator(DocTermOrds.java:767)
[junit4] 2> at
org.apache.lucene.search.FieldCacheImpl.getDocTermOrds(FieldCacheImpl.java:1214)
[junit4] 2> at
org.apache.lucene.index.BaseDocValuesFormatTestCase.doTestSortedSetVsUninvertedField(BaseDocValuesFormatTestCase.java:2342)
[junit4] 2> at
org.apache.lucene.index.BaseDocValuesFormatTestCase.testSortedSetVariableLengthVsUninvertedField(BaseDocValuesFormatTestCase.java:2375)
[junit4] 2> at sun.reflect.NativeMethodAccessorImpl.invoke0(Native
Method)
[junit4] 2> at
sun.reflect.NativeMethodAccessorImpl.invoke(NativeMethodAccessorImpl.java:57)
[junit4] 2> at
sun.reflect.DelegatingMethodAccessorImpl.invoke(DelegatingMethodAccessorImpl.java:43)
[junit4] 2> at java.lang.reflect.Method.invoke(Method.java:606)
[junit4] 2> at
com.carrotsearch.randomizedtesting.RandomizedRunner.invoke(RandomizedRunner.java:1617)
[junit4] 2> at
com.carrotsearch.randomizedtesting.RandomizedRunner$6.evaluate(RandomizedRunner.java:826)
[junit4] 2> at
com.carrotsearch.randomizedtesting.RandomizedRunner$7.evaluate(RandomizedRunner.java:862)
[junit4] 2> at
com.carrotsearch.randomizedtesting.RandomizedRunner$8.evaluate(RandomizedRunner.java:876)
[junit4] 2> at
org.apache.lucene.util.TestRuleSetupTeardownChained$1.evaluate(TestRuleSetupTeardownChained.java:50)
[junit4] 2> at
org.apache.lucene.util.TestRuleFieldCacheSanity$1.evaluate(TestRuleFieldCacheSanity.java:51)
[junit4] 2> at
org.apache.lucene.util.AbstractBeforeAfterRule$1.evaluate(AbstractBeforeAfterRule.java:46)
[junit4] 2> at
com.carrotsearch.randomizedtesting.rules.SystemPropertiesInvariantRule$1.evaluate(SystemPropertiesInvariantRule.java:55)
[junit4] 2> at
org.apache.lucene.util.TestRuleThreadAndTestName$1.evaluate(TestRuleThreadAndTestName.java:49)
[junit4] 2> at
org.apache.lucene.util.TestRuleIgnoreAfterMaxFailures$1.evaluate(TestRuleIgnoreAfterMaxFailures.java:70)
[junit4] 2> at
org.apache.lucene.util.TestRuleMarkFailure$1.evaluate(TestRuleMarkFailure.java:48)
[junit4] 2> at
com.carrotsearch.randomizedtesting.rules.StatementAdapter.evaluate(StatementAdapter.java:36)
[junit4] 2> at
com.carrotsearch.randomizedtesting.ThreadLeakControl$StatementRunner.run(ThreadLeakControl.java:359)
[junit4] 2> at
com.carrotsearch.randomizedtesting.ThreadLeakControl.forkTimeoutingTask(ThreadLeakControl.java:783)
[junit4] 2> at
com.carrotsearch.randomizedtesting.ThreadLeakControl$3.evaluate(ThreadLeakControl.java:443)
[junit4] 2> at
com.carrotsearch.randomizedtesting.RandomizedRunner.runSingleTest(RandomizedRunner.java:835)
[junit4] 2> at
com.carrotsearch.randomizedtesting.RandomizedRunner$3.evaluate(RandomizedRunner.java:737)
[junit4] 2> at
com.carrotsearch.randomizedtesting.RandomizedRunner$4.evaluate(RandomizedRunner.java:771)
[junit4] 2> at
com.carrotsearch.randomizedtesting.RandomizedRunner$5.evaluate(RandomizedRunner.java:782)
[junit4] 2> at
org.apache.lucene.util.AbstractBeforeAfterRule$1.evaluate(AbstractBeforeAfterRule.java:46)
[junit4] 2> at
org.apache.lucene.util.TestRuleStoreClassName$1.evaluate(TestRuleStoreClassName.java:42)
[junit4] 2> at
com.carrotsearch.randomizedtesting.rules.SystemPropertiesInvariantRule$1.evaluate(SystemPropertiesInvariantRule.java:55)
[junit4] 2> at
com.carrotsearch.randomizedtesting.rules.NoShadowingOrOverridesOnMethodsRule$1.evaluate(NoShadowingOrOverridesOnMethodsRule.java:39)
[junit4] 2> at
com.carrotsearch.randomizedtesting.rules.NoShadowingOrOverridesOnMethodsRule$1.evaluate(NoShadowingOrOverridesOnMethodsRule.java:39)
[junit4] 2> at
com.carrotsearch.randomizedtesting.rules.StatementAdapter.evaluate(StatementAdapter.java:36)
[junit4] 2> at
org.apache.lucene.util.TestRuleAssertionsRequired$1.evaluate(TestRuleAssertionsRequired.java:43)
[junit4] 2> at
org.apache.lucene.util.TestRuleMarkFailure$1.evaluate(TestRuleMarkFailure.java:48)
[junit4] 2> at
org.apache.lucene.util.TestRuleIgnoreAfterMaxFailures$1.evaluate(TestRuleIgnoreAfterMaxFailures.java:70)
[junit4] 2> at
org.apache.lucene.util.TestRuleIgnoreTestSuites$1.evaluate(TestRuleIgnoreTestSuites.java:55)
[junit4] 2> at
com.carrotsearch.randomizedtesting.rules.StatementAdapter.evaluate(StatementAdapter.java:36)
[junit4] 2> at
com.carrotsearch.randomizedtesting.ThreadLeakControl$StatementRunner.run(ThreadLeakControl.java:359)
{code}
Robert looked into it and was wtf'd, quote:
{code}
what happened here...
{code}
I looked into it and was wtf'd, quote (from my head):
{code}
wtf?!
{code}
I looked deeper at the code and it's a beautiful and classic class loading
deadlock. I don't think I've seen an example of this in real life *ever*,
except for this one case.
Problem description.
1. Thread A attempts to return a new instance of RandomAccessOrds:
{code}
return new RandomAccessOrds() {
{code}
RandomAccessOrds extends SortedSetDocValues and has a final static field which
in turn loads RandomAccessOrds (circular reference).
2. Thread B attempts to create:
{code}
private class Iterator extends SortedSetDocValues {
{code}
3. If thread B starts loading SortedSetDocValues, it blocks other threads from
doing so. If, at the same time, thread A starts loading RandomAccessOrds then
thread A will eventually attempt to initialize SortedSetDocValues and both will
wait for each other indefinitely.
I attach a simpler example that does the same as a POC.
--
This message was sent by Atlassian JIRA
(v6.2#6252)
---------------------------------------------------------------------
To unsubscribe, e-mail: [email protected]
For additional commands, e-mail: [email protected]