apache
fix replicateRate of batchRead in auto-recover is negative
Motivation
Currently the replicateRate of batchRead in auto-recover may exceed Integer.MAX_VALUE and then become negative, which would result in batchRead operation stuck.
Changes
deal with the negative situation
good jobs, I have a question: should maxBytesToReplicate use long type?
rateLimiter.acquire(int) just support int value. Besides, Integer.max_value means 2GB throughput, it is enough for most situation.
good jobs, I have a question: should maxBytesToReplicate use long type?
rateLimiter.acquire(int) just support int value. Besides, Integer.max_value means 2GB throughput, it is enough for most situation because replicate throughput should not be too high.