> Are there any details on how much of an issue fragmentation is (with > Cassandra ) ? > With all the merging and deletes that happen with during compactions, > how does the disk fragmentation look like over time ? Any thumb-rules > on how frequently and how to defrag ?
Cassandra is about as friendly as it can be with respect to fragmentation. All data is written sequentially and in bulk. Assuming a modern file system, and assuming you don't run at 99% disk capacity (which would explode in your face anyway), fragmentation should be a non-issue. The only realistic circumstances where I would expect fragmentation is if you run Cassandra on a file system that also does something else that fragments the hell out of it. -- / Peter Schuller