I do not think the mmap OOM is connected with the row size. 

Add some info to the other thread if you think it's relevant. 


Aaron Morton
Freelance Developer

On 7/02/2012, at 10:33 AM, Ajeet Grewal wrote:

It's also a good idea to partition time series data so that the rows do not
grow too big. You can have 2 billion columns in a row, but big rows have
operational down sides.

What are the down sides here? Unfortunately I have an existing system
which I modeled with large rows (because I use the sorted nature of
columns to get column ranges). After the amount of data grows, I get
"mmap failed" exceptions (See my other thread "Cassandra OOM"). I
wonder if there is a connection.