Forum Discussion

Kelemvor's avatar
Kelemvor
Icon for Expert rankExpert
3 days ago

If a collector fails to a down collector, will it then fail to that collectors backup?

Hi,

We have three collectors that can all see the same items.  Normally we've had them setup where 1 and 2 fail to each other, and 3 fails to 1.  However, if we ever lose two at a time, we'd be stuck.

If I set them up so 1 -> 2, 2 -> 3, 3 -> 1, and we lose 1 and 2, what happens?  The devices on 2 will fail to 3 just fine.  What happens with the devices on 1?  Do they try to fail to 2 and then realize that 2's down and then fail to 3?  Or do they just go to 2 and then stop?

Thanks

    • Kelemvor's avatar
      Kelemvor
      Icon for Expert rankExpert

      These are in different locations with a VPN collection.  We don't want them to be checked by other collectors, but in an emergency, it would theoretically work.

      • Mike_Moniz's avatar
        Mike_Moniz
        Icon for Professor rankProfessor

        Ah I understand. Sounds like a good thing to test out in a sandbox instance :) Not sure if anyone has tested that situation (I have now) but let us know if you find the answer.

        Assuming the collectors are all on separate sites. Watch out for the situation where a site with a collector goes down due to ISP/power issue and another collector attempts to take over. Instead of the down collector blocking every device from reporting host down, the working collector will not block it and EVERY device at that site will report down. I've seen a ton of tickets generated from those situations. Enough that we avoid having failover across VPNs since typically something that takes down the collector will also take down all the devices at the site and/or vpn itself. For maintenance, you can do a manual failover (or SDT the site).