Please enable JavaScript.
Coggle requires JavaScript to display documents.
cassandra data modelling (goals (Spread data evenly around the cluster,…
cassandra data modelling
goals
Spread data evenly around the cluster
Minimize the number of partitions read
cassandra data write
log data in commit log
write to memtable
flush memtable
write to sstable in disk
includes
partition index on disk to map token to location on disk
commitlog_total_space_in_mb
when exceeds, memtable put to flush queue
queue config param
memtable_heap_space_in_mb
data write is blocked if data to be flushed exceeds memtable_cleanup_threshold
is purged once data flushed to sstable in disk
SSTable
immutable
not modified after memtable is flushed
partition
stored across multiple
compaction
merge multiple sstables into one
improves read performance