> If it helps, I also found quite a few of these in the logs > > org.apache.cassandra.db.UnserializableColumnFamilyException: Couldn't find > cfId=224101 > > However a single cassandra instance locally (OSX, 1.6.0_22, mmap) runs just > perfect for > hours. No exceptions, no OOM.
Given that these exceptions show up on the receiving end of inter-node communication, I think it sounds reasonable that you don't see it on a local instance. Given that the inter-node communication triggers the bug by concurrent use of the ByteBuffer:s - that trigger point would no longer exist on a local node. -- / Peter Schuller
