The following blog has been updated to reflect changes in version 6.2 to 6.3 as well as repair some other issues.
As you may have found the existing official documentation on vROPS Remote collectors is pretty thin. As I was involved in a project to get this all going in a Enterprise setting I thought I would share some documentation with you.
First of all here is the official doco: http://pubs.vmware.com/vrealizeoperationsmanager-6/index.jsp#com.vmware.vcom.core.doc/GUID-83164C8C-...
And here is a good post about some questions you may have: http://virtsanity.com/2015/05/vrealize-operations-manager-6-remote-collector-information/
If you have not worked with vROPS 6 yet there is a good book I would recommend: https://www.packtpub.com/virtualization-and-cloud/mastering-vcenter-operations-manager
The deployment we are looking at is a vROPS 6.0.2 with vRIN 5.8.4 (vRealize Infrastructure Navigator formally vCenter Infrastructure Navigator, VIN) and with SRM integration of vRIN.
The idea is to have a central vROPS cluster and then use remote collectors to get data from other vCenters that are disbursed throughout the word.
However you can also use this for pure vROPS remote collectors
The main site consists of an vROPS Master, a replica and a Data node. vROPS is connected to the local vCenter (Protected site) as well as to the VRIN that is paired with the same vCenter. vRIN is configured to collect information form the VMs as well as from SRM.
Each remote site has a remote collector that is paired with the remote vCenter (Protected Site) as well as the vRIN instance. vRIN is configured to collect information form the VMs as well as from SRM.
Using vROPS and SRM together is something that needs to be discussed. Some people have the idea that they like to monitor the VMs that fail over from the protected vCenter to the recovery vCenter and that it all then magically works. This is not the case.
Each VM (or actually every object) in vCenter has its unique moref (managed object reference) and even if a VM has the same name in the Protected vCenter as in the Recovery vCenter it’s a different object. When SRM protects a VM it will create a placeholder VM on the recovery site. This placeholder VM is basically only the VMX file and has no VMDKs attached to it. SRM will furnish the VMs with VMDKs at the time of recovery.
So if you are connection a vROPS instance to the Proetced and the Recovery site, VROPS will see two different VMs (each with the same name). One will be active monitored and the other one is powered off. However you just wasted a vROPS VM licence for an essentially dead VM. The placeholder VM on the protected site shouldn’t be on, if it is...you are in a DR scenario.
So what would be the benefit of an vROPS in DR?
The only thing would be the ability to use all the data that is collected from the point on that the placeholder VM is started. You could use the troubleshooting options as well as some of the views etc. But all forecasts will be unusable. Please remember that vROPs needs at least 3 weeks of data collection to make accurate future predictions.
In my personal opinion vROPS in DR is just a waste of licensing and space. I can not see any real benefits. Please feel free to correct me.
For VRIN to be integrated in SRM the user must have permission on the PAIRED SRM instance. Meaning the users needs to have permissions on vCenter as well as on the SRM instance that this vCenter is paired with. As for the role…that’s a bit tricky there isn’t really any great doco about it I successfully used for vCenter the read rights plus Virtual machine | Interaction | Console interaction | Guest operating system management by VIX API. For SRM I haven’t really tested it that much and used the Admin role…properly there is a better solution.
The following figures show all the Network ports that need to be in place for vROPS & vRIN in regards to the above scenario.
The are heaps of posts about how to deploy and configure vROPS and vRIN so I will not cover this.
We will focus on deploying and configuring the remote collectors.
A collector group is a group of collectors. The idea is to have multiple remote collectors in one site and then use them to spread the load of the collection. You don’t have to create a collector group to use a remote collector.
Go to Administration and then to Collection Groups. Here you can create a new Collection group and assign remote collectors to it.
To see if everything is working do the following: