AnsweredAssumed Answered

Defining Service Setup in a Small Cluster

Question asked by dataanalyst on Aug 7, 2016
Latest reply on Aug 14, 2016 by MichaelSegel

Hi Folks, I am doing a POC on MapR Hadoop on a 3 node cluster using Converged Community Edition license. Cluster has heterogeneous hardware and have got few queries w.r.t. optimal service layout, volume management etc. Below are node configurations:

 

 

1st Node: 24 Cores, 528 GB RAM, SATA HDDs, OS – Ubuntu 14.04

2nd Node: 24 Cores, 528 GB RAM, SATA HDDs, SSDs, SAS, OS – Ubuntu 14.04

3rd Node: 24 Cores, 528 GB RAM, SATA HDDs, SSDs, SAS, OS – Ubuntu 14.04

 

Below service layout is planned for implementation in cluster with all nodes running in YARN mode. How performance will be impacted if CLDB and ZooKeeper are on same node? Also, please share your comments on whether this is optimal service layout for a small cluster?

 

Many Thanks in Advance

 

Untitled11.png

Outcomes