Corbin, There are broadly 2 ways you can get into this situation -
1. Your consumer's fetch size is smaller than the largest message it is trying to consume. Try increasing your fetch.size to make it larger than the max.message.size on your producers 2. The log on the server is corrupted. To verify this, try running the DumpLogSegments tool on the log segments for that topic/partition and see if it reports a problem. HTH, Neha On Tue, Dec 11, 2012 at 8:43 PM, Corbin Hoenes <[email protected]> wrote: > possible causes (1) a sin
