I am new to Cassandra and I am trying to optimize my reads (write times are ok, about 1000 lines per 0.1 sec) and I read key caching and Row.
Using Java with the Hector library to insert data seems to be going well, but reading is very slow (about 2-3 times slower than the SQL Server instance I'm comparing with). I started looking at the statistics cfstats show:
Key cache capacity: 200000 Key cache size: 0 Key cache hit rate: NaN Row cache capacity: 40000 Row cache size: 0
Shouldn't there be something in the Key or Row cache after insertion? I think thatβs why it works so slowly when I try to read.
Any help would be appreciated ...
source share