[ https://issues.apache.org/jira/browse/TRAFODION-320?page=com.atlassian.jira.plugin.system.issuetabpanels:comment-tabpanel&focusedCommentId=14696085#comment-14696085 ]
Atanu Mishra commented on TRAFODION-320: ---------------------------------------- taoufik ben abdellatif (taoufik-abdellatif) wrote on 2014-05-29: #1 The issue is specific to the case when one of the key columns is of type varchar. For this case, the internal logic that decodes the boundary values expects the length to be at the beginning of the varchar. The boundary values as read from hbase don't adhere to this condition. Changes are done in NATable.cpp to update the boundary value to include the length so the decoding logic can process it correctly. Changed in trafodion: status: In Progress → Fix Committed Stacey Johnson (sjohnson-w) on 2014-06-10 information type: Proprietary → Public taoufik ben abdellatif (taoufik-abdellatif) on 2014-10-14 Changed in trafodion: status: Fix Committed → Fix Released > LP Bug: 1324299 - compiler crashes in rangePartitioning > ------------------------------------------------------- > > Key: TRAFODION-320 > URL: https://issues.apache.org/jira/browse/TRAFODION-320 > Project: Apache Trafodion > Issue Type: Bug > Components: sql-cmp > Reporter: Apache Trafodion > Assignee: Apache Trafodion > Priority: Blocker > Fix For: 1.0 (pre-incubation) > > > A compiler crash was observed in compiler rangePartitioning after an hbase > partition split. The problem can be easily recreated on a workstation by > forcing a partition split from the hbase master page. > When a crash occurs the following stack trace is generated: > #4 0x00007fcbe54e76e2 in JVM_handle_linux_signal () from > /usr/lib/jvm/java/jre/lib/amd64/server/libjvm.so > #5 <signal handler called> > #6 getValueId (this=0x7fcbe1d2e390, addConvNodes=0, > tf=ExpTupleDesc::SQLARK_EXPLODED_FORMAT, resultBuffer=0x7fcbc7814608 "", > resultBufferLength=360, > length=0x7ffff7779b28, offset=0x7ffff7779b24, diagsArea=0x0) at > ../optimizer/ItemExpr.h:188 > #7 ValueIdList::evalAtCompileTime (this=0x7fcbe1d2e390, addConvNodes=0, > tf=ExpTupleDesc::SQLARK_EXPLODED_FORMAT, resultBuffer=0x7fcbc7814608 "", > resultBufferLength=360, length=0x7ffff7779b28, offset=0x7ffff7779b24, > diagsArea=0x0) at ../optimizer/ValueDesc.cpp:4486 > #8 0x00007fcbde7a3fa6 in getRangePartitionBoundaryValuesFromEncodedKeys > (part_desc_list=<value optimized out>, numberOfPartitions=20, SQLMPKeytag=0, > partColArray= > ..., heap=0x7fcbc786a2a0) at ../optimizer/NATable.cpp:1405 > #9 createRangePartitionBoundaries (part_desc_list=<value optimized out>, > numberOfPartitions=20, SQLMPKeytag=0, partColArray=..., heap=0x7fcbc786a2a0) > at ../optimizer/NATable.cpp:1667 > Assigned to LaunchPad User taoufik ben abdellatif -- This message was sent by Atlassian JIRA (v6.3.4#6332)