ORA-39029

平台 P750*2
系统 AIX 6108
数据库 1120310 RAC

导出脚本：

expdp system/oracle DIRECTORY=DBPUMP dumpfile=js_jssi_jsrep_%U.dmp logfile=js_all_exp.log SCHEMAS=JSREP,JSSI,JYCHK,SJQY

导入脚本：

impdp system/oracle DIRECTORY=DBPUMP dumpfile=js_jssi_jsrep_%U.dmp logfile=js_all_imp.log SCHEMAS=JSREP,JSSI,JYCHK,SJQY cluster=N

遇错：

ORA-39014: One or more workers have prematurely exited.
ORA-39029: worker 3 with process name "DW00" prematurely terminated
ORA-31671: Worker process DW00 had an unhandled exception.
ORA-04030: out of process memory when trying to allocate 4194328 bytes (pga heap,buf_kgcbz2state)
ORA-06512: at "SYS.KUPW$WORKER", line 1838
ORA-06512: at line 2

Job "SYSTEM"."SYS_IMPORT_SCHEMA_04" stopped due to fatal error at 10:44:55

实际现场情况是主机资源空闲，不存在资源不够的情况，所以直接就怀疑到Oracle的用户资源分配限制设置上，使用smit user登录检查oracle用户发现在sofe方面都实施RAC的人未做修改，改之。问题消失。

该问题在11.2.0.2之前可能为bug。