Hello,

Currently we're running a MapReduce over live htables to do table merge
(introduced at https://hbase.apache.org/0.94/book/mapreduce.example.html);
At the samtime these tables are still serving user scan requests; As this
is a full table scan which may take much server resources, we want to
control the impact to users during the MapReduce, avoding remarkable
performance downgrade during the MR; I see there are two parameters might
be related: caching and cacheBlocks, like :


scan.setCaching(500);

scan.setCacheBlocks(false);  // don't set to true for MR jobs


But still want to double check with the experts here, is there other ways
to control this? Thanks!

Shaofeng Shi
Apache Kylin (incubation)

Reply via email to