[jira] [Created] (FLINK-12097) Flink SQL hangs in StreamTableEnvironment.sqlUpdate, keeps executing and seems never stop, finally lead to java.lang.OutOfMemoryError: GC overhead limit exceeded

["xu (JIRA)" <ji...@apache.org>]

xu created FLINK-12097:
--------------------------

             Summary: Flink SQL hangs in StreamTableEnvironment.sqlUpdate, keeps executing and seems never stop, finally lead to java.lang.OutOfMemoryError: GC overhead limit exceeded
                 Key: FLINK-12097
                 URL: https://issues.apache.org/jira/browse/FLINK-12097
             Project: Flink
          Issue Type: Bug
          Components: Table SQL / Planner
    Affects Versions: 1.7.2
            Reporter: xu
         Attachments: DSL.txt

Hi Experts,
There is a Flink application(Version 1.7.2) which is written in Flink SQL, and the SQL in the application is quite long, consists of about 10 tables, 1500 lines in total. When executing, I found it is hanged in StreamTableEnvironment.sqlUpdate, keep executing some code about calcite and the memory usage keeps grown up, after several minutes java.lang.OutOfMemoryError: GC overhead limit exceeded is got.
 
I get some thread dumps:
        at org.apache.calcite.plan.volcano.RuleQueue.popMatch(RuleQueue.java:475)
        at org.apache.calcite.plan.volcano.VolcanoPlanner.findBestExp(VolcanoPlanner.java:640)
        at org.apache.calcite.tools.Programs$RuleSetProgram.run(Programs.java:339)
        at org.apache.flink.table.api.TableEnvironment.runVolcanoPlanner(TableEnvironment.scala:373)
        at org.apache.flink.table.api.TableEnvironment.optimizeLogicalPlan(TableEnvironment.scala:292)
        at org.apache.flink.table.api.StreamTableEnvironment.optimize(StreamTableEnvironment.scala:812)
        at org.apache.flink.table.api.StreamTableEnvironment.translate(StreamTableEnvironment.scala:860)
        at org.apache.flink.table.api.StreamTableEnvironment.writeToSink(StreamTableEnvironment.scala:344)
        at org.apache.flink.table.api.TableEnvironment.insertInto(TableEnvironment.scala:879)
        at org.apache.flink.table.api.TableEnvironment.sqlUpdate(TableEnvironment.scala:817)
        at org.apache.flink.table.api.TableEnvironment.sqlUpdate(TableEnvironment.scala:777)
 
 
        at java.io.PrintWriter.write(PrintWriter.java:473)
        at org.apache.calcite.rel.AbstractRelNode$1.explain_(AbstractRelNode.java:415)
        at org.apache.calcite.rel.externalize.RelWriterImpl.done(RelWriterImpl.java:156)
        at org.apache.calcite.rel.AbstractRelNode.explain(AbstractRelNode.java:312)
        at org.apache.calcite.rel.AbstractRelNode.computeDigest(AbstractRelNode.java:420)
        at org.apache.calcite.rel.AbstractRelNode.recomputeDigest(AbstractRelNode.java:356)
        at org.apache.calcite.rel.AbstractRelNode.onRegister(AbstractRelNode.java:350)
        at org.apache.calcite.plan.volcano.VolcanoPlanner.registerImpl(VolcanoPlanner.java:1484)
        at org.apache.calcite.plan.volcano.VolcanoPlanner.register(VolcanoPlanner.java:859)
        at org.apache.calcite.plan.volcano.VolcanoPlanner.ensureRegistered(VolcanoPlanner.java:879)
        at org.apache.calcite.plan.volcano.VolcanoPlanner.ensureRegistered(VolcanoPlanner.java:1755)
        at org.apache.calcite.plan.volcano.VolcanoRuleCall.transformTo(VolcanoRuleCall.java:135)
        at org.apache.calcite.plan.RelOptRuleCall.transformTo(RelOptRuleCall.java:234)
        at org.apache.calcite.rel.convert.ConverterRule.onMatch(ConverterRule.java:141)
        at org.apache.calcite.plan.volcano.VolcanoRuleCall.onMatch(VolcanoRuleCall.java:212)
        at org.apache.calcite.plan.volcano.VolcanoPlanner.findBestExp(VolcanoPlanner.java:646)
        at org.apache.calcite.tools.Programs$RuleSetProgram.run(Programs.java:339)
        at org.apache.flink.table.api.TableEnvironment.runVolcanoPlanner(TableEnvironment.scala:373)
        at org.apache.flink.table.api.TableEnvironment.optimizeLogicalPlan(TableEnvironment.scala:292)
        at org.apache.flink.table.api.StreamTableEnvironment.optimize(StreamTableEnvironment.scala:812)
        at org.apache.flink.table.api.StreamTableEnvironment.translate(StreamTableEnvironment.scala:860)
        at org.apache.flink.table.api.StreamTableEnvironment.writeToSink(StreamTableEnvironment.scala:344)
        at org.apache.flink.table.api.TableEnvironment.insertInto(TableEnvironment.scala:879)
        at org.apache.flink.table.api.TableEnvironment.sqlUpdate(TableEnvironment.scala:817)
        at org.apache.flink.table.api.TableEnvironment.sqlUpdate(TableEnvironment.scala:777)
 *Both point to some code about calcite.*
 
And also I get the heap dump, found that there are *5703373 RexCall instances, and 5909525 String instances, 5909692 char[] instances ,**size is 6.8G*. I wonder why there are so many RexCall instances here, why it keeps on executing some calcite code and seems never stop. 
|char[]| |5,909,692 (16.4%)|6,873,430,938 (84.3%)|
|java.lang.String| |5,909,525 (16.4%)|165,466,700 (2%)|

|org.apache.calcite.rex.RexLocalRef| |5,901,313 (16.4%)|259,657,772 (3.2%)|
|org.apache.flink.calcite.shaded.com.google.common.collect.RegularImmutableList| |5,739,479 (15.9%)|229,579,160 (2.8%)|
|java.lang.Object[]| |5,732,702 (15.9%)|279,902,336 (3.4%)|
|org.apache.calcite.rex.RexCall| |5,703,373 (15.8%)|273,761,904 (3.4%)|
 
 Because the SQL is a bit long, I put the SQL in the attachment. Also the SQL is written in DSL I invented. But it is translated to Flink SQL later.
 
 
Look forward to your reply.
Thanks a lot.
 
 
Best
Henry



--
This message was sent by Atlassian JIRA
(v7.6.3#76005)