I want to less overload cassandra nodes , in order so I am assuming spark.cassandra.input.readsPerSec will solve this problem . I am reading cassandra table from spark . In order to overload less cassandra node , is this property enough to solve the problem .
Can someone please illustrate this with example and values ?
The readsPerSec property puts a limit on the maximum number of read requests a Spark app ca perform per second.
For example, if you set it to 10:
--conf spark.cassandra.input.readsPerSec=10
then only 10 read requests will run every second. By default read requests are not limited so Spark will send as many requests as it needs to unless you throttle it. Cheers!
6 People are following this question.
