6.2 Clustering for High Availability
vRealize Operations Manager supports application level high availability (HA). Enabling HA has a penalty in the maximum number of objects and metrics that can be stored. With HA enabled, each piece of data contained on the master node is stored twice, halving the maximum number of supported objects and metrics and doubling the IO and disk space requirements.
Table 4. Impact of HA on Scalability – vRealize Operations Manager 6.0.2
| Maximum Data Nodes | Maximum Supported Objects (8 node cluster) | Maximum Supported Metrics (8 node cluster) |
Without HA | 8 | 75 K | 20 m |
With HA | 8 | 37.5 K | 10 m |
Design Considerations
• To enable HA, at least one more data node must be deployed in addition to the master node. If possible when HA is enabled, deploy the replica node within the same cluster and put DRS rules in place to keep the replica separate from the master so there is physical redundancy.
• The HA feature of vRealize Operations Manager operates independently of, but complementary to VMware vSphere High Availability.