Select & Prepare Platform

Commission server cluster

The target server cluster must have adequate resources for deploying the chosen PNDA flavor in terms of numbers of nodes, CPU, memory and storage capacity.

Building and configuring a suitable server cluster is outside the scope of this guide. Quite frequently technologies such as Cobbler are used to automate this process.

Required resources

In addition to the existing PNDA flavors, described elsewhere in this guide, we have defined a flavor specifically intended for creating PNDA on physical server clusters.

Production

Production flavor is intended for a reasonably sized physical cluster. It runs the core services in high-availability mode and provides reasonable storage space and compute resource.

Role Number required Cores Memory Storage
tools 1 16 64 GB 1TB
edge 1 20 256 GB 1TB
mgr1 1 16 128 GB 1TB
mgr2 1 16 128 GB 1TB
mgr3 1 16 128 GB 1TB
mgr4 1 16 128 GB 1TB
datanode 3 24 128 GB As required
opentsdb 2 16 64 GB 1TB
hadoop-manager 1 16 64 GB 1TB
kafka 3 20 128 GB As required
zookeeper 3 16 64 GB 1TB

We recommended dividing the storage on every node into at least a root volume (100GB is sufficient) and a data volume. We also recommend JBOD for Hadoop datanodes and RAID10 for Kafka. However, all these aspects are configurable.

Next

Home Prepare Mirror Build Stage Configure Create

results matching ""

    No results matching ""