Just out of curiosity, why not have two separate databases with links if
needed for the few times that Process needs information from Results?
smb_slb_at_my-deja.com wrote in message <7ihbc5$34q$1_at_nnrp1.deja.com>...
>Hi all,
>
>We have a system composed of a relatively small amount of vital data
>and a huge amount of not-so-vital data (probably like a 1/400 ratio).
>Assuming that I store all the vital tables on one tablespace (called
>PROCESS) and the other data on a second tablespace (called RESULTS), is
>it possible for the system to run if the RESULTS tablespace has been
>trashed? Specifically, if the system crashes, is it okay to restore
>just the PROCESS tablespace but not restore the results tablespace (or
>restore it to a much older state, say when the system first came
>online)?
>
>We're looking at 2-5 GB for PROCESS data and 200GB - 1 TB for RESULTS
>data. The way I figure it, we'll do RAID 5 or 10 on the whole affair,
>but only do tape backups of the PROCESS tablespace. If we have a
>catastrophic failure, we just restore PROCESS plus an empty RESULTS
>tablespace.
>
>The only RI in between the two subsystems is the RESULTS referring to
>some stuff in PROCESS, so losing everything in RESULTS won't leave any
>dangling references.
>
>Will this work?
>
>
>--== Sent via Deja.com http://www.deja.com/ ==--
>---Share what you know. Learn what you don't.---
Received on Wed May 26 1999 - 18:06:04 CDT
