Big table with UNION ALL or partitioning with Tsearch2

Hi,

I have a really big Tsearch2 table (100s GB) that takes a while to perform
queries and takes days to index. Is there any way to fix these issues
using UNIONs or partitioning? I was thinking that I could partition the
data by date but since I am always performing queries on the Tsearch2
field I do not know if this will help performance. I think paritioning
will help the indexing problem since I can incrementally re-index the data
but again I figured it would be better to ask.

Any suggestions will be greatly appreciated. Thanks in advance.

System I am running on:

-Raid 5 with 16x drives
-Quad core XEON
16 GB of memory (Any suggestion on the postgresql.conf setup would also be
great! Currently I am just setting shared mem to 8192MB)
-x86_64 but Redhat 5 Ent

Benjamin