I am using cassandra 1.* (3GB ram;4Cors) with 1 node (at first). I am saving different sizes (0.01mb – 600mb) of files and indexes in the db and i want to know whats the best size of a value in a column. If the optimal size is smaller then the filesize, i just split it in more columns.
I writed a test to find this size, but i am getting java heap excp. all the time, if i am writing/reading to much values in a row (without pause).
How i can fix this problem? i allready read (http://wiki.apache.org/cassandra/MemtableThresholds) and also the cassandra – FAQ
And i know there is no streaming L
Would help to decrese the flashsize of the memtable? (whould slow the reading right?)
Thanks for help or links:D