[
https://issues.apache.org/jira/browse/CASSANDRA-19101?page=com.atlassian.jira.plugin.system.issuetabpanels:comment-tabpanel&focusedCommentId=17936954#comment-17936954
]
Dmitry Konstantinov edited comment on CASSANDRA-19101 at 3/19/25 9:37 PM:
--------------------------------------------------------------------------
I have adjusted the test to drain + added a check for unused segments and it is
failing :)
{code:java}
final Mutation m = new RowUpdateBuilder(cfs1.metadata.get(), 0, "k")
.clustering("bytes")
.add("val", ByteBuffer.wrap(entropy))
.build();
// force creating several commitlog files
for (int i = 0; i < 10; i++)
{
m.apply();
}
StorageService.instance.drain();
for (CommitLogSegment segment :
CommitLog.instance.segmentManager.getActiveSegments())
{
Assert.assertFalse("An unused segment is left after drain: " +
segment.getName()
+ ", dirty tables: " + segment.dirtyString()
+ ", total segments: " +
CommitLog.instance.segmentManager.getActiveSegments().size()
+ ", commit log files: " + Arrays.toString(new
File(DatabaseDescriptor.getCommitLogLocation()).listFiles()),
segment.isUnused());
}
}
}
{code}
{code}
java.lang.AssertionError: An unused segment is left after drain:
CommitLog-7-1742418190647.log, dirty tables: , total segments: 1, commit log
files: [build/test/cassandra/commitlog/CommitLog-7-1742418190647.log]
at org.junit.Assert.fail(Assert.java:88)
at org.junit.Assert.assertTrue(Assert.java:41)
at org.junit.Assert.assertFalse(Assert.java:64)
at
org.apache.cassandra.db.commitlog.CommitlogShutdownTest.testShutdownWithPendingTasks(CommitlogShutdownTest.java:100)
{code}
We have 1 remaining unused segment which is remaining after drain (in a queue
and on disk).
drain() invokes
org.apache.cassandra.db.commitlog.CommitLog#forceRecycleAllSegments()
forceRecycleAllSegments method iterates over all active segments to check if
they are unused:
{code}
for (CommitLogSegment segment : activeSegments)
{
if (segment.isUnused())
archiveAndDiscard(segment);
}
{code}
but for the last segment: isStillAllocating() = allocatePosition.get() <
endOfBuffer returns true when segment.isUnused() is invoked in this loop, so it
is skipped here.
Later during
org.apache.cassandra.db.commitlog.AbstractCommitLogSegmentManager#awaitTermination
we invoke CommitLogSegment#close for the last remaining segment which as a part
of discardUnusedTail() logic and it changes allocatePosition, so
isStillAllocating() now is false. As a result when we check segment.isUnused()
after drain - it now returns true.
So, it looks like we need to skip at least the last segment in the test check
(if we consider keeping the last segment as an expected behaviour).
was (Author: dnk):
I have adjusted the test to drain + added a check for unused segments and it is
failing :)
{code:java}
final Mutation m = new RowUpdateBuilder(cfs1.metadata.get(), 0, "k")
.clustering("bytes")
.add("val", ByteBuffer.wrap(entropy))
.build();
// force creating several commitlog files
for (int i = 0; i < 10; i++)
{
m.apply();
}
StorageService.instance.drain();
for (CommitLogSegment segment :
CommitLog.instance.segmentManager.getActiveSegments())
{
Assert.assertFalse("An unused segment is left after drain: " +
segment.getName()
+ ", dirty tables: " + segment.dirtyString()
+ ", total segments: " +
CommitLog.instance.segmentManager.getActiveSegments().size()
+ ", commit log files: " + Arrays.toString(new
File(DatabaseDescriptor.getCommitLogLocation()).listFiles()),
segment.isUnused());
}
}
}
{code}
{code}
java.lang.AssertionError: An unused segment is left after drain:
CommitLog-7-1742418190647.log, dirty tables: , total segments: 1, commit log
files: [build/test/cassandra/commitlog/CommitLog-7-1742418190647.log]
at org.junit.Assert.fail(Assert.java:88)
at org.junit.Assert.assertTrue(Assert.java:41)
at org.junit.Assert.assertFalse(Assert.java:64)
at
org.apache.cassandra.db.commitlog.CommitlogShutdownTest.testShutdownWithPendingTasks(CommitlogShutdownTest.java:100)
{code}
We have 1 remaining unused segment which is remaining after drain (in a queue
and on disk).
drain() invokes
org.apache.cassandra.db.commitlog.CommitLog#forceRecycleAllSegments()
forceRecycleAllSegments method iterates over all active segments to check if
they are unused:
{code}
for (CommitLogSegment segment : activeSegments)
{
if (segment.isUnused())
archiveAndDiscard(segment);
}
{code}
but for the last segment: isStillAllocating() = allocatePosition.get() <
endOfBuffer returns true when segment.isUnused() is invoked in this loop, so it
is skipped here.
Later during
org.apache.cassandra.db.commitlog.AbstractCommitLogSegmentManager#awaitTermination
we invoke CommitLogSegment#close for the last remaining segment which as a part
of discardUnusedTail() logic and it changes allocatePosition, so
isStillAllocating() now is false. As a result when we check segment.isUnused()
after drain - it now returns true.
So, it looks like we need to skip at least the last segment in the test check...
> Test Failure: org.apache.cassandra.db.commitlog.CommitlogShutdownTest failed
> on trunk
> -------------------------------------------------------------------------------------
>
> Key: CASSANDRA-19101
> URL: https://issues.apache.org/jira/browse/CASSANDRA-19101
> Project: Apache Cassandra
> Issue Type: Bug
> Components: Test/unit
> Reporter: Jacek Lewandowski
> Assignee: Dmitry Konstantinov
> Priority: Normal
> Fix For: 4.0.x, 4.1.x, 5.x
>
> Attachments: Cassandra-trunk_2057_CommitlogShutdownTest.log,
> diff_discard_logic.png
>
>
> {noformat}
> java.lang.AssertionError
> at org.junit.Assert.fail(Assert.java:86)
> at org.junit.Assert.assertTrue(Assert.java:41)
> at org.junit.Assert.assertTrue(Assert.java:52)
> at
> org.apache.cassandra.db.commitlog.CommitlogShutdownTest.testShutdownWithPendingTasks(CommitlogShutdownTest.java:96)
> at
> java.base/jdk.internal.reflect.NativeMethodAccessorImpl.invoke0(Native Method)
> at
> java.base/jdk.internal.reflect.NativeMethodAccessorImpl.invoke(NativeMethodAccessorImpl.java:62)
> at
> java.base/jdk.internal.reflect.DelegatingMethodAccessorImpl.invoke(DelegatingMethodAccessorImpl.java:43)
> at java.base/java.lang.reflect.Method.invoke(Method.java:566)
> at
> org.junit.runners.model.FrameworkMethod$1.runReflectiveCall(FrameworkMethod.java:50)
> at
> org.junit.internal.runners.model.ReflectiveCallable.run(ReflectiveCallable.java:12)
> at
> org.junit.runners.model.FrameworkMethod.invokeExplosively(FrameworkMethod.java:47)
> at
> org.junit.internal.runners.statements.InvokeMethod.evaluate(InvokeMethod.java:17)
> at
> org.jboss.byteman.contrib.bmunit.BMUnitRunner$10.evaluate(BMUnitRunner.java:393)
> at
> org.jboss.byteman.contrib.bmunit.BMUnitRunner$6.evaluate(BMUnitRunner.java:263)
> at org.junit.runners.ParentRunner.runLeaf(ParentRunner.java:325)
> at
> org.junit.runners.BlockJUnit4ClassRunner.runChild(BlockJUnit4ClassRunner.java:78)
> at
> org.junit.runners.BlockJUnit4ClassRunner.runChild(BlockJUnit4ClassRunner.java:57)
> at org.junit.runners.ParentRunner$3.run(ParentRunner.java:290)
> at org.junit.runners.ParentRunner$1.schedule(ParentRunner.java:71)
> at org.junit.runners.ParentRunner.runChildren(ParentRunner.java:288)
> at org.junit.runners.ParentRunner.access$000(ParentRunner.java:58)
> at org.junit.runners.ParentRunner$2.evaluate(ParentRunner.java:268)
> at
> org.jboss.byteman.contrib.bmunit.BMUnitRunner$1.evaluate(BMUnitRunner.java:97)
> at org.junit.runners.ParentRunner.run(ParentRunner.java:363)
> at org.junit.runner.JUnitCore.run(JUnitCore.java:137)
> at
> com.intellij.junit4.JUnit4IdeaTestRunner.startRunnerWithArgs(JUnit4IdeaTestRunner.java:69)
> at
> com.intellij.rt.junit.IdeaTestRunner$Repeater$1.execute(IdeaTestRunner.java:38)
> at
> com.intellij.rt.execution.junit.TestsRepeater.repeat(TestsRepeater.java:11)
> at
> com.intellij.rt.junit.IdeaTestRunner$Repeater.startRunnerWithArgs(IdeaTestRunner.java:35)
> at
> com.intellij.rt.junit.JUnitStarter.prepareStreamsAndStart(JUnitStarter.java:232)
> at com.intellij.rt.junit.JUnitStarter.main(JUnitStarter.java:55)
> {noformat}
> Manual testing to confirm issues found by CircleCI when testing
> CASSANDRA-18464. Run with Java 11 / IntelliJ
--
This message was sent by Atlassian Jira
(v8.20.10#820010)
---------------------------------------------------------------------
To unsubscribe, e-mail: commits-unsubscr...@cassandra.apache.org
For additional commands, e-mail: commits-h...@cassandra.apache.org
