(imported topic written by SystemAdmin)
Is there a way to exempt the main TEM server from being selected from endpoints other than my relay infrastructure? I currently have over 800 endpoints relying on the main server, and I want to lessen that.
(imported comment written by SystemAdmin)
I should note other details.
Environment is all set to automatic selection. I also use fallback relays selection for my DMZ, for devices outside of the domain.
(imported comment written by PeterLoer)
Do you already have an accessible failover relay set for all machines? (i.e., using _BESClient_RelaySelect_FailoverRelay). That is the normal mechanism for keeping endpoints that can’t find a relay via normal means from ending up on the root server.
There isn’t a setting to completely remove the root server from the relay selection list; if an agent tries and fails to contact the Failover relay, it will still try to go to the root server as a final step. The thinking here is that it is a lot better to have 800 endpoints reporting directly to your root server than to have 800 agents that you have no management control of at all…
cheers,
peter
(imported comment written by SystemAdmin)
I do have two failover relays set, but it is reserved for those that are not currently in the domain (i.e. Laptops that are out of the office). Im there are other relays that these endpoints should be selecting in the domain, before they would select the main MES server, but for some reason they are just registering tot he BES server. I would prefer nothing did (other than relays of course).
I do agree that it is better than no control.
(imported comment written by MBARTOSH)
I would also like to control how many relays attach to the BES Server. I have two head relays that I would like all of the relays to communication with, but I still have 223 relays attached to the BES server, 332 to one head relay and 316 to the other head relay. I tried setting the __BESRELAY_SELECTION_RELAYPRIORITY = 10 and that seemed to make it worse. It seems like it should have worked according to this.
_BESRelay_Selection_RelayPriority
Component: BESRelay
Registry Hive:
Type: Numeric
Version:
Platform: All
MinNumeric:
MaxNumeric:
Default: 0
Requires Restart: No
Description: All relays of priority 0 are selected before all relays of priority 1, etc.
Another setting is:
_BESRelay_Selection_RelayWeight
Component: BESRelay
Registry Hive:
Type: Boolean
Version:
Platform: All
Default: 100
Requires Restart: No
Description: The likelihood of a relay being selected is equal to its weight divided by the cumulative weight of all
the relays that are the same distance from the selecting client (NOTE: a relay with a weight of a million will still be
selected after a relay of weight two, if the clients think the latter relay is a hop closer)
Let me know if you find the combination. It is way too complex in my opinion.
(imported comment written by r.kelly)
Consider bringing up a relay with a CNAME of the root server. All clients should will resolve this DNS name. Within you BigFix topology, modify the host file on the other relays, sql, et al to reflect the true root server.
Clients that are selecting the root server will now go to the “false root” instead.
As an aside, I would first recommend placing a relay for the endpoints that is closer in hops that the root server.