PG uses MVCC to manage concurrency. A downside of this is that to verify the exact number of rows in a table you have to visit them all.
There's plenty on this in the archives, and probably the FAQ too.
What are you using the count() for?
select logid, agentid, logbody from log where logid=3000000;
this query also returns after about 120 seconds. The table log has about 7 million records, and logid is the primary key of log table. What about that? Why is it too slow?
-sezai
---------------------------(end of broadcast)--------------------------- TIP 9: the planner will ignore your desire to choose an index scan if your joining column's datatypes do not match
