| |
Virtual injected into Hadoop unprecedented vigor, production management from the IT point of view, the following points:
- Hadoop and other resource consumption of different types of applications deployed along a shared data center can improve overall resource utilization;
Flexible virtual machine operation so that the user can dynamically created based on data center resources, expand their Hadoop cluster, you can narrow the current cluster, freeing resources to support other applications, if desired;
- Through virtualization architecture provides HA, FT integration, to avoid the traditional Hadoop clusters of single point of failure, coupled with the reliability of the data itself Hadoop, big data for enterprise applications to provide a reliable guarantee.
For these reasons, vSphere Big Data Extensions (BDE) to provide users with flexible deployment and management of Hadoop clusters provide effective support in a virtualized environment. Apart from these advantages, whether virtualization will hurt the performance of Hadoop to run it? To this end, we made performance comparison virtualization deployments and physical deployment of Hadoop clusters and optimization on the same scale, experiments show that virtualization Hadoop clusters can well support the production environment.
Performance comparison virtualized environment and physical environment
Figure 1 shows the deployment style tuning test, deploy a virtual machine on a single physical server only, Tasktracker and Datanode run together on the same node. Because each virtual node can use all server resources, facilitate Hadoop virtualization and the deployment of traditional physical environment do performance comparison and analysis. The test results are shown in Figure 2, the virtualization Hadoop relative performance comparison of the physical environment is almost flat.
At the same time, we put these experiments experience is built into vSphere BDE deployed Hadoop cluster configurations, the shielding performance optimization complexity. Although different data center and cluster configuration settings may result in different manifestations, where according to create, configure, extend Hadoop cluster order to share some common experiences with you |
|
|
|