I’ve been starting seenig strange OSPF errors again. Discarding packet: Locally originated src address = x.x.x.x. The logs are being spammed with this so much, it looks as this is the only thing happening on the router. We have multiple external IP’s. On one of those IP’s we have a VPN endpoint on which foreign offices connect. We have cca. 50 Mikrotik routers in our network all connected with OSPF. As soon as I disconnect the VPN endpoint router from the OSPF routing, the errors stop. Also I must say that everything looks as though it works normally. The foreign offices (FO) are connected via PPTP or SSTP and the OSPF is set to point-to-point on these links, others via broadcast. All other routers seem to communicate between them in a normal manner without ospf errors. It seems my predecessor was aware of these problems since he set out a logging rule, not to show errors from ospf route. This can’t be the only solution, can it? Anyone else had similar problems?
Thank you mrz for your answer. The situation is like this:
The router I found out to be causing the problems is our VPN gateway. Lets say that we have a local address for VPN endpoint and it is 192.168.1.1. VPN clients get IP’s 192.168.1.2, 3, 4, etc. VPN gateway has its own public address, but is also connected through eth2 adapter to our core router (which has its own public IP, but from the same /29 range as VPN endpoint. Public routes are not pushed through OSPF).
So the OSPF network rule is: 192.168.1.0/24 on vpn endpoint site and also on client sides. As soon as I add these routes to VPN router I start getting route ospf error spams. If I disconnect the VPN router from our network I don’t see these errors on any other router but VPN endpoint. If I disconnect all but 1 VPN clients on VPN endpoint router = no errors. Anything else = errors. It seems that this could be something like same IP on multiple interfaces, but even if I change each VPN connection range to it’s own subnet (192.168.1.0/30, 192.168.1.4/30,…) - so no VPN interface is having the same IP, I still get this error.
mrz maybe I am misunderstanding what you are trying to achieve with this, but to me this kind of defeats the purpose of OSPF networks, you are meant to be able to put the /24 in there and have it match any interface that has an IP within that /24.
Having to add every little subnet to the OSPF networks would be tedious.
We are having the same problem. It seemed to start with the more recent versions newer than 5.19. I am joining the thread to follow it for help. We have about 60 routers in our network all on the backbone running OSPF and the core router at the edge which is our BGP gateway is running constant “route ospf error Discarding packet: locally originated” for every IP address that is local to this router. There are no bridge loops. Thanks for the help.
Combo mode. For VPN links i used Point-to-point and for LAN links and subnets I used broadcast. I use NBMA for all now and it’s beautiful. Also tried RIP and also had no problems.
NBMA failed us also. At some (seems) random point I can’t get routes anymore from main to client routers. All links are stable, no loopbacks and each OSPF router has it’s own routerID! Not all OSPF links fall down at the same time. Both routers still see each other as neighbours, one as DR and others as backup. Restarting a client router always helps and I get routes pushed back for again a random time. It really doesn’t matter which RB I use. It can be a true RB or an x86 machine or VM - same problems. I can’t seem to be able to solve this problem so I was asked to move everything back to RIP. RIP also pushes routes immediatelly, where OSPF needs quite some time. We have 15 routers at this location interconnected via copper or fiber. Using RoS versions from 5.24 to 6.4.
