Gridmix3 - Emulating Production Workload for Apache Hadoop (Hadoop and Distributed Computing at Yahoo!)

At Yahoo!, grids running Hadoop have attracted a wide range of applications from a diverse set of functional groups. The workload submitted by each group is distinguished not only from others in the cluster, but its profile also changes as users gain experience with Hadoop. Users tune their jobs to consume available resources; some may circumvent the assumptions and fairness control mechanisms of

[yuzuru_k]

hadoopベンチマークツール

[shiumachi]

Gridmix3では、本番環境のジョブのデータに基づいてシミュレーションジョブを生成してくれる、ということが書いてある。