Uploaded image for project: 'IMPALA'
  1. IMPALA
  2. IMPALA-12800

Queries with many nested inline views see performance issues with ExprSubstitutionMap

    XMLWordPrintableJSON

Details

    • Improvement
    • Status: Open
    • Critical
    • Resolution: Unresolved
    • Impala 4.3.0
    • None
    • Frontend
    • None
    • ghx-label-12

    Description

      A user running a query with many layers of inline views saw a large amount of time spent in analysis. 

       

      - Authorization finished (ranger): 7s518ms (13.134ms)
      - Value transfer graph computed: 7s760ms (241.953ms)
      - Single node plan created: 2m47s (2m39s)
      - Distributed plan created: 2m47s (7.430ms)
      - Lineage info computed: 2m47s (39.017ms)
      - Planning finished: 2m47s (672.518ms)

      In reproducing it locally, we found that most of the stacks end up in ExprSubstitutionMap.

       

      Here are the main stacks seen while running jstack every 3 seconds during a 75 second execution:
      Location 1: (ExprSubstitutionMap::compose -> contains -> indexOf -> Expr equals) (4 samples)

         java.lang.Thread.State: RUNNABLE
          at org.apache.impala.analysis.Expr.equals(Expr.java:1008)
          at java.util.ArrayList.indexOf(ArrayList.java:323)
          at java.util.ArrayList.contains(ArrayList.java:306)
          at org.apache.impala.analysis.ExprSubstitutionMap.compose(ExprSubstitutionMap.java:120)

      Location 2:  (ExprSubstitutionMap::compose -> verify -> Expr equals) (9 samples)

         java.lang.Thread.State: RUNNABLE
          at org.apache.impala.analysis.Expr.equals(Expr.java:1008)
          at org.apache.impala.analysis.ExprSubstitutionMap.verify(ExprSubstitutionMap.java:173)
          at org.apache.impala.analysis.ExprSubstitutionMap.compose(ExprSubstitutionMap.java:126)

      Location 3: (ExprSubstitutionMap::combine -> verify -> Expr equals) (5 samples)

         java.lang.Thread.State: RUNNABLE
          at org.apache.impala.analysis.Expr.equals(Expr.java:1008)
          at org.apache.impala.analysis.ExprSubstitutionMap.verify(ExprSubstitutionMap.java:173)
          at org.apache.impala.analysis.ExprSubstitutionMap.combine(ExprSubstitutionMap.java:143)

      Location 4:  (TupleIsNullPredicate.wrapExprs ->  Analyzer.isTrueWithNullSlots -> FeSupport.EvalPredicate -> Thrift serialization) (4 samples)

         java.lang.Thread.State: RUNNABLE
          at java.lang.StringCoding.encode(StringCoding.java:364)
          at java.lang.String.getBytes(String.java:941)
          at org.apache.thrift.protocol.TBinaryProtocol.writeString(TBinaryProtocol.java:227)
          at org.apache.impala.thrift.TClientRequest$TClientRequestStandardScheme.write(TClientRequest.java:532)
          at org.apache.impala.thrift.TClientRequest$TClientRequestStandardScheme.write(TClientRequest.java:467)
          at org.apache.impala.thrift.TClientRequest.write(TClientRequest.java:394)
          at org.apache.impala.thrift.TQueryCtx$TQueryCtxStandardScheme.write(TQueryCtx.java:3034)
          at org.apache.impala.thrift.TQueryCtx$TQueryCtxStandardScheme.write(TQueryCtx.java:2709)
          at org.apache.impala.thrift.TQueryCtx.write(TQueryCtx.java:2400)
          at org.apache.thrift.TSerializer.serialize(TSerializer.java:84)
          at org.apache.impala.service.FeSupport.EvalExprWithoutRowBounded(FeSupport.java:206)
          at org.apache.impala.service.FeSupport.EvalExprWithoutRow(FeSupport.java:194)
          at org.apache.impala.service.FeSupport.EvalPredicate(FeSupport.java:275)
          at org.apache.impala.analysis.Analyzer.isTrueWithNullSlots(Analyzer.java:2888)
          at org.apache.impala.analysis.TupleIsNullPredicate.requiresNullWrapping(TupleIsNullPredicate.java:181)
          at org.apache.impala.analysis.TupleIsNullPredicate.wrapExpr(TupleIsNullPredicate.java:147)
          at org.apache.impala.analysis.TupleIsNullPredicate.wrapExprs(TupleIsNullPredicate.java:136)

      Attachments

        1. long_query_jstacks.tar.gz
          340 kB
          Joe McDonnell
        2. impala12800schema.sql
          0.2 kB
          Joe McDonnell
        3. impala12800repro.sql
          94 kB
          Joe McDonnell

        Activity

          People

            Unassigned Unassigned
            joemcdonnell Joe McDonnell
            Votes:
            0 Vote for this issue
            Watchers:
            3 Start watching this issue

            Dates

              Created:
              Updated: