Archives of the TeradataForum

Message Posted: Tue, 12 Feb 2002 @ 14:50:22 GMT

	<Prev	Next>		<<First	<Prev	Next>	Last>>

<-- Anonymously Posted: Tuesday, February 12, 2002 09:37 -->

Cplatt and Raghav, thanks for your comments. I am looking for high level strategies but this is good info. I should give more details so you know the situation better.

We are quite familiar with ARC here and I believe we have taken full advantage of it. We cluster dump our largest tables (> 200GB) individually, use database-level archives for "application" databases at times they are least active, and "database by table" archives for highly active databases (long database-level lock not feasible) or to get the rest of the smaller tables in a database where a big table lives. We have 8 different archive jobs, some running in parallel and some serially based on data volume, tape availability and load schedules. We do not employ journalling although we are looking into the implications of journalling our larger tables.

We share the tape drives with the rest of the enterprise so the constant host traffic eats into our ability to write data when we need to. Because we are an "always open" company, I *never* get a window of inactivity where I can slam data to tape unobstructed. We will not get the funds for a dedicated silo unless I can prove we need it and it has ROI - prior attempts have not impressed management and I haven't heard too many people in the past say they are happy with their NCR-supplied dedicated tape solution.

Hope this gives everyone a better idea of where we are at.

Thanks!

	<Prev	Next>		<<First	<Prev	Next>	Last>>