Published 11th July 2019
A standard VxRail deployment provides the ability to deploy a VMware vRealize Log Insight appliance in Standalone Mode. In order to provide high availability for this Log Insight instance, the single vRLI appliance should be expanded to a 3-node vRLI cluster.
The VxRail Manager first run will allow the user to deploy a single node vRLI instance which will run in Standalone Mode.
The vRLI license included with VxRail is for standalone single cluster deployments only. With this, that single VxRail cluster can be expanded up to 64 (dual CPU) nodes and still retain the Log Insight licensing.
Also within scope of this licensing is the ability to expand vRLI from a single node to a single cluster (3 nodes).
Please note that a 2-node vRLI cluster is not supported by VMware.
A couple of reasons why a customer would want to expand from the single node vRLI to a Single vRLI Cluster:
- To provide high availability for vRLI
- To help satisfy vRLI compute requirements due to increased VxRail cluster size and workload
Compute Requirements
The initial vRLI appliance is deployed with the following compute resources:
- Memory: 8 GB
- CPU: 4 CPUs
- Disk: 530 GB
So in a Single vRLI Cluster (3-nodes), we will require in total:
- Memory: 24 GB
- CPU: 12 CPUs
- Disk: 1590 GB
Expand vRLI Single Node to a Single vRLI Cluster
First a quick summary of what we will do:
- Configure vRLI Virtual IP
- Create DNS entries for additional vRLI nodes
- Download vRLI OVF
- Deploy OVF template for 2 new vRLI appliances/nodes
- Add each new vRLI node to existing vRLI single node deployment
OK then … on we crack … Configure the vRLI Virtual IP address, as this is not configured during the VxRail first run.
Create the necessary DNS records for the Virtual IP address and additional vRLI nodes, then in the vRLI Admin UI, select Cluster > +New Virtual IP Address
Enter the IP and FQDN details, click Save, give it a few seconds, and you should then have something similar to the following:
The vRLI cluster can be created simply by deploying two more vRLI nodes and adding them to this first vRLI deployment.
Go to the VMware Product Download site and download the required vRLI appliance (same version as the existing standalone node).
Alternatively, if you cannot access the VMware Product Download page, or the required version is no longer available to download, the vRLI ovf files can be accessed from the local datastore of any of the VxRail nodes (Datastores > Files > Images > log-insight-original), so you can deploy the exact same vRLI version as required.
Deploy OVF Template for the 2nd and 3rd additional Log Insight nodes, selecting the Small Configuration option for each (ensure their respective DNS records are in place).
When deployment is complete, for each new node, power on the new appliance, and when ready, browse to https://<FQDN or IP>
When prompted, select JOIN EXISTING DEPLOYMENT
Enter the FQDN or IP address of the master node, and select GO
… shortly followed by ….
At this stage a request is made to the Master vRLI node, from where the user must browse-to and select ALLOW
Once both new vRLI nodes have been added, the vRLI Cluster view should look like the following:
At this stage the expansion is complete. What was a single node Standalone Mode deployment of vRLI, is not a fully HA 3-node vRLI cluster.
Note that this expansion is completely separate from any expansion of VxRail nodes. This was purely at the vRLI level.
No other reconfig of the vRLI is required. Nothing has changed on the VxRail side of things, so no need to go rediscover any of the infrastructure etc.
As part of the VxRail First Run automated installation process, vRLI is configured to discover and monitor the ESXi Hosts under vCenter, as shown above. The screenshot below shows an example of the out of the box vRLI General Overview Dashboard:
And that’s that, your vRealize Log Insight with VxRail is now fully HA. Hope that helps!
Steve
Scamallach 