cancel
Showing results for 
Show  only  | Search instead for 
Did you mean: 

FNMS Cloud - Adding beacon #2 to our environment - New beacon recognized but never gets priority

Hey all - 

 

We've been running on just one beacon for over 2 years.  I set up another one.  We use https so we've had a cert created and added to cert.pem on a test machine.

FNMS beacon lists sees the beacon and says both are working.  But in the logs on the test machine, it always gives beacon #1 a  10 and beacon #2 99 even though they are responding at the same rate.  These are peer beacons, not a parent - child relationship.

From the logs:

[Wed 09 Dec 2020 11:47:25 AM PST (N, 2)] {5180} Average round-trip time to beacon1..com is 15 ms
[Wed 09 Dec 2020 11:47:25 AM PST (N, 2)] {5180} Average round-trip time to beacon2.com is 15 ms
[Wed 09 Dec 2020 11:47:25 AM PST (N, 2)] {5180} Assigned server "https://beacon1.com/ManageSoftRL/" a priority of 10
[Wed 09 Dec 2020 11:47:25 AM PST (N, 2)] {5180} Assigned server "https://beacon2l.com/ManageSoftRL/" a priority of 99
[Wed 09 Dec 2020 11:47:25 AM PST (N, 2)] {5180} Applying prioritization algorithm "MgsADSiteMatch" with parameters ""
[Wed 09 Dec 2020 11:47:25 AM PST (N, 2)] {5180} SelectorAlgorithm "MgsADSiteMatch" could not be loaded
[Wed 09 Dec 2020 11:47:25 AM PST (N, 2)] {5180} ERROR: Could not successfully prioritize servers

Any help would be appreciated.

(1) Reply
ChrisG
By Community Manager Community Manager
Community Manager

If the ping (ICMP) response time for both servers is exactly the same (to the closest millisecond) then the default priority will likely end up as you're observing. Here are a few thoughts about this:

If the beacons are in different locations, then I expect the likelihood that the ping response time is exactly the same would typically be pretty low, and the typical operational behavior you would see across many devices is a reasonably breakdown based on which beacon can be reached fastest.

If the beacons are in the same location, then it is more likely that the response time will be the same for two different ping operations - although once you get connections from across different networks that is going to be less likely. If this is a real concern then a couple of options to consider may be:

  • Use a load balancer - configure agents to connect to the load balancer's URL, and something like a round robin algorithm to farm out requests to both beacons.
  • Use a setting like "MgsRandom;MgsPing;MgsADSiteMatch" for the SelectorAlgorithm agent preference - this will randomize the priority order of beacons before doing the ping, such that if the ping response from both beacons is the same then the random order will be preserved.
(Did my reply solve the question? Click "ACCEPT AS SOLUTION" to help others find answers faster. Liked something? Click "KUDO". Anything expressed here is my own view and not necessarily that of my employer, Flexera.)