If you do decide to go the exp/imp route, there are a lot of things you can do to speed it up significantly - here are some notes on things you might want to try. Unfortunately, if you do a character set conversion, I'm not sure if you can use direct=y since I believe it requires the exp character set to match that of the database, but maybe you can still do the conversion on the import side. On Export: * DIRECT=Y * RECORDLENGTH=65535 * Consider running export via multiple parallel sessions, e.g. breaking it up by table or schema, and then doing the same for import. * Exp/imp to/from named pipe to avoid disk I/O * Verify OPTIMIZER_MODE != FIRST_ROWS - this can cause very poor performance - see Metalink 181072.1 for more info (included below) On Import: * ALTER DATABASE NOARCHIVELOG; (set back to ARCHIVELOG and take full backup after completion) * BUFFER=10485760 * RECORDLENGTH=65535 o Especially important when going across platforms * Optionally (depending on needs, version, etc.): o COMMIT=N o STATISTICS=NONE o ANALYZE=N o RECALCULATE_STATISTICS=N o Increase size of online logs o Use LMTs o INDEXES=N (followed with INDEXFILE=Y and then parallel, nologging creation of indexes after completion of table imports) o Set _disable_logging=TRUE o Set _wait_for_sync=FALSE ________________________________ From: Robyn [mailto:robyn.sands@xxxxxxxxx] I am slowly accepting that we may have to recreate the databases. if anyone has a specific suggestion on rebuilding a 2 TB database with minimal downtime, please drop me a line. Privileged/Confidential Information may be contained in this message or attachments hereto. Please advise immediately if you or your employer do not consent to Internet email for messages of this kind. Opinions, conclusions and other information in this message that do not relate to the official business of this company shall be understood as neither given nor endorsed by it.