[CASSANDRA-8410] Select with many IN values on clustering columns can result in a StackOverflowError

• Type: Bug
• Status: Resolved
• Priority: Normal
• Resolution: Fixed
• Fix Version/s: 2.0.12, 2.1.3
• Severity: Normal
• Assignee: Tom Hobbs
• Reporter: Tom Hobbs

Authors

Tom Hobbs

Reviewers

Benjamin Lerer

• Created: 02/Dec/14 23:53
• Updated: 16/Apr/19 09:31
• Resolved: 12/Dec/14 17:45

Description

When executing a SELECT statement with an IN clause on the clustering columns, a StackOverflowError can occur if the memtable doesn't contain any of the requested slices. In 2.0, this happens with the following stack trace:

ERROR [ReadStage:23] 2014-12-02 14:53:11,077 CassandraDaemon.java (line 199) Exception in thread Thread[ReadStage:23,5,main]
java.lang.StackOverflowError
at org.apache.cassandra.db.marshal.Int32Type.compare(Int32Type.java:52)
at org.apache.cassandra.db.marshal.Int32Type.compare(Int32Type.java:28)
at org.apache.cassandra.db.marshal.AbstractType.compareCollectionMembers(AbstractType.java:279)
at org.apache.cassandra.db.marshal.AbstractCompositeType.compare(AbstractCompositeType.java:64)
at org.apache.cassandra.db.marshal.AbstractCompositeType.compare(AbstractCompositeType.java:36)
at edu.stanford.ppl.concurrent.SnapTreeMap$1.compareTo(SnapTreeMap.java:538)
at edu.stanford.ppl.concurrent.SnapTreeMap.boundedMax(SnapTreeMap.java:905)
at edu.stanford.ppl.concurrent.SnapTreeMap.boundedExtreme(SnapTreeMap.java:833)
at edu.stanford.ppl.concurrent.SnapTreeMap.access$1000(SnapTreeMap.java:90)
at edu.stanford.ppl.concurrent.SnapTreeMap$AbstractIter.<init>(SnapTreeMap.java:2028)
at edu.stanford.ppl.concurrent.SnapTreeMap$EntryIter.<init>(SnapTreeMap.java:1951)
at edu.stanford.ppl.concurrent.SnapTreeMap$EntryIter.<init>(SnapTreeMap.java:1940)
at edu.stanford.ppl.concurrent.SnapTreeMap$SubMap$EntrySubSet.iterator(SnapTreeMap.java:2462)
at java.util.AbstractMap$2$1.<init>(AbstractMap.java:378)
at java.util.AbstractMap$2.iterator(AbstractMap.java:377)
at org.apache.cassandra.db.filter.ColumnSlice$NavigableMapIterator.computeNext(ColumnSlice.java:154)
at org.apache.cassandra.db.filter.ColumnSlice$NavigableMapIterator.computeNext(ColumnSlice.java:162)
at org.apache.cassandra.db.filter.ColumnSlice$NavigableMapIterator.computeNext(ColumnSlice.java:162)

In 2.1, there's a similar error, but it occurs in AtomicBTreeColumns.