Archives of the TeradataForum
Message Posted: Thu, 23 Aug 2001 @ 13:12:27 GMT
| Subj: | | Re: Backups & Recoveries |
| |
| From: | | Anomy Anom |
<-- Anonymously Posted: Thursday, August 23, 2001 08:43 ->
If your customer data is loaded at defined intervals and the table(s) are small enough that ARC can copy the entire table set between
load cycles, schedule the archive to start after a load cycle completes. If you are constantly adding to or updating customer data, you
could make use of journalling on the tables and follow the archive strategy NCR has for journalled tables. I'm not too familiar with this
option but I'm sure someone else on the list can jump in to elaborate.
If your tables are larger than ARC could copy between loads or you are constantly loading, consider creating copies of tables at set
times using Insert/Select or "Create Table As With Data" and archiving those while leaving the source tables untouched. This requires a lot
of free perm and uses extra I/Os that may impact performance a bit.
Another suggestion is to consider copying individual tables rather than at a database level. ARC will apply insert/update-blocking read
locks at a table level rather than a database level. While one table in the database is being copied, others in that database can be
loaded. Just remember to add new tables to your ARC script when they are added to the database.
| 
