Tom Lane <tgl@sss.pgh.pa.us> wries:
> >> Does a REINDEX of the table fix it?
>
Yes, positively it does. After reindexing all five indices of the table,
vacuum times have dropped from minutes to seconds. Looking at pg_class, it
is apparent that there was a cut in the number of tuples of about 30% for
all indices, but the number of pages dropped by about 90%!
Other people mentioned that this probably would get fixed in a forthcoming
release. Is there anything we can do in the meantime?
thanks
cl.
_________________________________________________________________
MSN 8 with e-mail virus protection service: 2 months FREE*
http://join.msn.com/?page=features/virus