I have a table with ~ 3 million records. I'm indexing a field holding
names, no more than 200 bytes each. Indexing the resulting tsvector
takes forever. It's been running now for more than 40 hours on a Linux
with PG 8.01, a single Xeon & 4GB RAM. My  work_mem postgresql.conf
parameter is at 240960 and maintenance_work_mem at 96384, although the
index task is using at most 12MB. Task is 99% cpu bound. Is there any
way I may speed up the indexing?


IDS de Costa Rica S.A.

---------------------------(end of broadcast)---------------------------
TIP 8: explain analyze is your friend

Reply via email to