i need a consistent export of an ONLINE, EBS R12.1.3 database ( 30 Terra bytes in Size ) based on a flashback_scn. ( took current_scn before export)
There are 240 schemas and 35239 tables
Export log shows following errors for a big number of tables:partitions
ORA-31693: Table data object "XXFAH"."XLA_AE_LINES_H":"XXCTF"."XXCTF_P20152603" failed to load/unload and is being skipped due to error:
ORA-02354: error in exporting/importing data
ORA-01555: snapshot too old: rollback segment number 215 with name "_SYSSMU215_2978801081$" too small
. . exported "XXFAH"."XLA_AE_HEADERS_H":"XXCHN"."XXCHN_P20141908" 17.53 KB 466 rows
. . exported "XXFAH"."XLA_AE_HEADERS_H":"XXCHN"."XXCHN_P20151302" 18.67 KB 507 rows
I checked when this partition was last modified.
SYS-ebsprd>select min(timestamp), max(timestamp) from dba_tab_modifications where SUBPARTITION_NAME='XXCTF_P20152603';
Some are modified same day of export, but still skipped.
We have sufficient size of UNDO and undo_retention parameter to manage 3 days transactions.
On querying dba_tab_modifications , i discovered that above tables have last transactions in past i.e in one month to 24 months back.
Why export skipped them to keep my dump consistent?.
i think following two options can help me, need advise from experienced guys, with reference if possible
Option 1. export separtely above skipped schema.table:partitions let say skipped.dmp,
then import this dump skipped.dmp
then import the first full dump with table_exist_action=APPEND
(my doubt here is that will it skip only the partition,subpartition or full table ? )
import the first full dump ( that lacks the tables )
then import second dump ( skipped.dmp) with table_exist_action=APPEND
Will appreciate help on this issue.