I did some benchmark tests on derby with JDBCBenchmark program.
I set the checkpoint interval as 40MB, and did tests with different
numbers of clients, 1, 10, 20...... I found when the number of clients
is more than 1, sometime derby did not do checkpint on every 40MB
log.
E.g. I tested five times with 10 clients, the maximum uncheckpointed
log sizes I observed were as following (MB):
40, 106.561, 107.118, 107.561, 66.834
The test command I used is:
java -Dderby.storage.pageCacheSize=5000 JDBCBench -driver
org.apache.derby.jdbc.EmbeddedDriver -url jdbc:derby:testdb -tpc 40000
-clients 10 -tps 50
Can anyone tell me how this could happen?
Thanks.
Raymond
_________________________________________________________________
This Easter, Hop Over To Live Maps For Some Cultural Enlightenment.
http://local.live.com/?mkt=en-ca/?v=2&cid=A6D6BDB4586E357F!101
