Hi, Zhong, Thanks for sharing this. We probably should add a sanity check in the broker to make sure that replica.fetch.max.bytes >= message.max.bytes. Could you file a jira for that?
Jun On Thu, Feb 13, 2014 at 8:01 PM, zhong dong <zea...@qq.com> wrote: > We encountered with this problem, too. > > And our problem is that we set the message.max.bytes larger than > replica.fetch.max.bytes. > > After we changed the replica.fetch.max.bytes to a larger number, the > problem solved. > >
