Hi,
Question 1: If I changed server.properties and added or modified the following
line:
log.flush.intervals.ms.per.topic=topic1:1000,
topic2:3000
Do I have to restart the server to pick up the change?
Question 2: I have three machines with one broker on each of them. 100
producers are
running on the same three machines. The producers will
produce 1M
20k-byte messages altogether. The batch size is fixed at
100. For 5 topics
each with five partitions, the performance is extremely
good. For each topic,
30k messages were produced per second . For 10 topics with
10 partitions,
the performance is really bad. For each topic, 400
messages were produced
per second. The gap is so huge. Any insight on what may
cause the difference?
Regards,
Libo