Sounds like a trigger attached to acs_objects might be causing the problem. You can view the triggers on a table using something like the following (this query excludes RI triggers):
openacs4=# select tgname from pg_trigger where tgrelid = (select oid from pg_class where relname = 'acs_objects') and tgname not like 'RI_Constraint%';
tgname
--------------------------------
acs_objects_context_id_del_tr
acs_objects_context_id_up_tr
acs_objects_context_id_in_tr
acs_objects_update_tr
acs_objects_insert_tr
acs_objects_last_mod_update_tr
acs_objects_mod_ip_insert_tr
(7 rows)
Try dropping the triggers one by one and see if you can get the update speed of object_id = -3 to increase. If at the end of this exercise, you still have a problem, then it's probably one of the RI triggers. You can also try dropping them one by one, but be warned, there's alot of them.
Before you do any of this, make a backup, because your db will be in a bad state afterwards. Or if this a production site, do a dump and restore into a test db, and then do the debugging on the test site.
