Best Practice For Setting Up nWorks Collection for Failover and Disaster Recovery

Additional information given about how collectors work within nWorks and guidelines on how to set up collectors for disaster situations.

The Best Practice for Setting Up nWorks Collection for Failover and Disaster Recovery
To get started with monitoring through the nWorks Enterprise Manager, you will have to add hosts individually or add a VC to the connected server’s page.  Once all of your hosts are added in and check boxed which hosts you want to monitor, you are ready for setting up your collectors for data collection.  

By Default, you will have the Enterprise Manager tab and a “Default Failover Group” tab (v5.6 and earlier) or a Monitoring Group Tab (5.7 and newer) added in to your Enterprise manager.  These 2 options are important because at the ‘Failover/Monitoring Group’ level gives you one set of load balancing between collectors for all collectors under the ‘Failover/Monitoring Group’ and the Enterprise Manager will allow you to load balance across ALL failover groups.  This is important to know because load balancing happens automatically under Failover groups but not automatically under the Enterprise Manager group, but can still be load balanced.  This is important when setting up your Disaster Recovery scenario’s for downed collectors.  

Load Balancing happening between collectors under ‘Failover/Monitoring Group’ happen automatically.  You can create different ‘Failover/Monitoring Groups’ under the “Enterprise Manager” tab --> Create ‘Failover/Monitoring Group’ Button.  Once done, you can left click and drag any collector to the different ‘Failover/Monitoring Groups’.  You can also click on a collector to find the Monitoring Load per collector.  As long as you have enough Monitoring Load % free per collector for another collector to take over and remain under 100% Monitoring Load, you will have enough space to run smoothly and collect correctly.  Good Example:  Collector A is at 25% load, Collector B is at 30% load and Collector C is at 50% load.  This is an acceptable collector balance as any one of them can fail and still be under a 100%.  If you need/plan to have more than one collector go down, it is encouraged to have additional collectors to balance to (active or standby).  Bad Example:  Collector A is at 50% load, Collector B is at 70% load and Collector C is at 80% load.  This is a bad example of collector failover because even at best scenario (A failing) and failing over to B, you will be over 100% collection load (120%).  Once collection goes above 100%, we may see some undesired behavior such as events coming in delayed or not at all.  

As for nWorks Enterprise Manager Failover, at this point (v5.7) there isn’t a way to do this automatically.  The manual way to do this for failover purposes is to either A) restore the nWorks Enterprise Manager VM from a backup (or other like device/software) or B) have a 2nd nWorks Enterprise Manager ready to go, and uninstall and reinstall all of the collectors (nWorksVIC.exe) and point it to the 2nd nWorks Enterprise Manager server during the install.  Keep in mind, if done the 2nd way (B Method) you will have to recreate your Failover Group settings (or have them mirrored) and move the collectors to the desired locations.  

Any other details as for how to set up and install nWorks or how many collectors you will need for your environment, please check out our nWorks documentation online.  
http://www.veeam.com/vmware-microsoft-esx-monitoring/resources.html
http://www.veeam.com/vmware-esx-monitoring-hp-operations/resources.html