Let’s first talk about hardware monitoring in general. The Veeam MP monitors both the health and performance of each ESX Host in your environment. It monitors for local fan, power supply, hard drive, and raid controller failures, for instance. It also collects and monitors metrics such as temperature, fan speed, watts, amps, and voltage. We can do this, agentlessly, by tapping into CIM/SMASH on each ESX Host.
With the Refresh Hardware Information events, you are seeing VMware API messaging each time we request the VMware API to refresh hardware information from CIM/SMASH. This is our older method (v5.6 and lower) for gathering hardware information, and is the fallback method for when our new method (available in v5.7 and higher) is not working.
Our new method for gathering hardware information taps directly into CIM/SMASH on each VMware ESX Host via the CIM-XML Service, completely by-passing the VMware APIs. This method requires a connection from each Veeam Collector to each ESX Host. Our Veeam MP Installation Guide defines the following prerequisites to make this happen:
1.) Open TCP port 5989 (two-way) between each Veeam Collector and each ESX Host
2.) Grant the Veeam MP vCenter Connection account CIM Interaction privileges on each vCenter
- a.) Clone the “Read-Only” role and name, “Veeam MP Read-Only”
b.) Edit the new “Veeam MP Read-Only” role and add privilege “HOST\CIM Interaction”
c.) Assign the new “Veeam MP Read-Only” role to the Veeam MP Connection Account
Alternatively, if you would like to get by with the old method for gathering hardware information, you can increase the duration that we refresh the VMware APIs. This can be done from within the Veeam EMS Console.
1.) Log into the Veeam EMS Console
2.) Click on the “Enterprise Manager” tab
3.) Click on the “Collection Settings” button under Administration.
4.) Uncheck “Use defaults”
5.) Increase the Interval Multiplier for Hardware details from 5 (5 min x 5 multiplier = 25 min) to 10 (5 min x 10 multiplier = 50 min) or higher.
The hardware refresh task cycles through every ESX Host within the defined time period set above. So, if you had 10 hosts and had set the Hardware details to 6 (5 min x 6 multiplier = 30 mins) then you would typically see a hardware refresh event in vCenter once every 3 mins (30 min / 10 host = 3 min / 1 host).
In the past, some of our customers had set this multiplier to 96 (8hrs) and even 144 (12hrs) to reduce the number of events generated by the VMware APIs. Ultimately, this is why we replaced this old generation method with the new. Because we did not like the idea of our customers going 8hrs+ without knowing whether they have a faulty hardware condition on a ESX Host.