Error while importing schemas using datapump

Question

Hi,
I am trying to import schema from qc to development. after importing i got the following error attached below:
Processing object type SCHEMA_EXPORT/TABLE/GRANT/WITH_GRANT_OPTION/OBJECT_GRANT
Processing object type SCHEMA_EXPORT/TABLE/GRANT/CROSS_SCHEMA/OBJECT_GRANT
Processing object type SCHEMA_EXPORT/TABLE/INDEX/INDEX
ORA-39065: unexpected master process exception in RECEIVE
ORA-39078: unable to dequeue message for agent MCP from queue "KUPC$C_2_20090421161917"
Job "SYS"."uat.210409" stopped due to fatal error at 20:15:13
ORA-39014: One or more workers have prematurely exited.
ORA-39029: worker 2 with process name "DW02" prematurely terminated
ORA-31671: Worker process DW02 had an unhandled exception.
ORA-39078: unable to dequeue message for agent KUPC$A_2_20090421161934 from queue "KUPC$C_2_20090421161917"
ORA-06512: at "SYS.KUPW$WORKER", line 1397
ORA-06512: at line 2
ORA-39029: worker 3 with process name "DW03" prematurely terminated
ORA-31671: Worker process DW03 had an unhandled exception.
ORA-39078: unable to dequeue message for agent KUPC$A_2_20090421162030 from queue "KUPC$C_2_20090421161917"
ORA-06512: at "SYS.KUPW$WORKER", line 1397
ORA-06512: at line 2
ORA-39029: worker 4 with process name "DW04" prematurely terminated
ORA-31671: Worker process DW04 had an unhandled exception.
ORA-39078: unable to dequeue message for agent KUPC$A_2_20090421162031 from queue "KUPC$C_2_20090421161917"
ORA-06512: at "SYS.KUPW$WORKER", line 1397
ORA-06512: at line 2
Is my import completed successfully or not??. please help...

Answer 1

When a datapump job runs, it creates a table called the master table. It has the same name as the job name. This is used to keep track of where all of the information in the dumpfile is located. It is also used when restarting a job. For some reason, this table got dropped. I'm not sure why, but in most cases, datapump jobs are restartable. I don't know why the original message was reported, but I was hoping the job woudl be restartable. You could always just rerun the job. Since the job that failed already created tables and indexes, if you restart the job, all of the objects that are dependent on those objects will not be created by default.
Let's say you have table tab1 with an index ind1 and both table and index are anaylized. Since tab1 is already created, the datapump job will mark all of the objects dependent on tab1 to skip. This includes the index, table_statistics, and index_statistics. To get around this, you could say
table_exists_action=replace
but this will replace all tables that are in the dumpfile. Your other options are:
table_exists_action=
truncate -- to truncate the data in the table and then just reload the data, but not the dependent objects
append -- to append the data from the dumpfile to the existing table, but do not import the dependent objects
skip -- skip the data and dependent objects from the dumpfile.
Hope this helps.
Dean