Hi all,
I am running a data import in a Postgres 8.0.1 data base. The target
table is one with ~ 100 million rows. The imported data is around 40
million rows. The import is inserting 10000 rows per transaction. The
table has a few indexes on it, a few foreign constraints, and one insert
trigger which normally inserts a row in another table for each insert on
this one, but for the duration of the data transfer it was modified to
not insert anything for the data which is transfered, but it does an
extra lookup on the index of a ~ 10000 big table to get the exclusion
condition.
Given this context, at transfer start I've had a transfer rate at ~ 1,3
million rows per hour, which dropped progressively to ~ 300000 now after
transferring ~ 24 million rows.
My question is what could cause this progressive slow-down ? I would
have expected a more or less constant transfer rate, as the table was
not empty in the first place to say that the indexes slow down as they
grow... so after a growth of ~ 20% I have a slow-down for inserts of
more than 4 times. What should I suspect here ? I can't just drop the
indexes/foreign keys/triggers, the db is in production use.
TIA for any ideas,
Csaba.
---------------------------(end of broadcast)---------------------------
TIP 9: In versions below 8.0, the planner will ignore your desire to
choose an index scan if your joining column's datatypes do not
match
