Several routers on our network are stuck in full / exchange OSPF states and not passing along the routes. The odd part is that the same configuration (automatically generated, BTW) seemed to work OK in older RouterOS versions.
I’ve found it in four links between RB600 with RouterOS 3.10.
Sample network topology where this problem has appeared:
Also, once in these full/exchange states, the adjacency counter never goes above 40 seconds.
If and only if I reboot the “CanVallsET” router does the route exchange take place (and then, both neighbors stay in the full state). Here’s the same link a couple hours after having rebooted “CanVallsET”, no other changes made.
I can easily and consistently replicate the situation after a reboot by disabling the OSPF network in one of the routers and letting 40 seconds pass (the dead-interval). Then, I enable it again -there you go, stuck in full/exchange. The odd thing is that this happens on its own, as if OSPF wasn’t able to converge after a temporary link outage.
We’ve already tried upgrading to RouterOS 3.14 (changelog said something about an OSPF bug), to no avail.
If there’s anything else I can tell you which can be useful to know where this could be coming from, let me know.
Thanks in advance…
I’ve too problems with OSPF after 3.10, tested up to latesd 3.15 with no good results.
Some routes are exchanged, but others with shorter path doesn’t get replicated. May be just the first exchanges until the full state is reached and routes that should be better just don’t get eschanged any more.
In one router I got routes marked with Unreachable, but the gateway was a router conected to same switch and connection between them where working fine.
The default route consistently doesn’t replicate. I nedded to force a 0.0.0.0/1 and 128.0.0.0/1 to have a fallback work around to the default route.
3.19, Routing Package, Just upgraded all my routers from 2.9.38 to 3.19 and have had to add a large amount of static routes as OSPF is not performing as expected.
Costs are not being added properly (all distance marks are 110) Hotspot interface (which is not a gateway, but as it is ethernet (I guess)) is being assigned default route dynamically. Routes are taking higher cost paths (i.e. 4 hops instead of 2). Routes are not distributing between some routers.
Two main routers are not passing routes (this is where I’ve had to add manually).
Some routers not receiving hello packets, or actually any OSPF (as per IP Firewall,Rule 1, Filter, Input - OSPF(89) - accept), Bytes and Packets are staying at 0. Which means the router up the line is not sending in anything, broadcast or point-to-point) however the router is sending HELLO packets and again nothing is being marked down
Any suggestions?
Should I use Routing-test? I would of thought Routing in 3.19 was the same as routing-test in 2.9.x which I had tested and worked fine.
Suggestions appreciated. I’ve put a sticking plaster on it, but right now I have 50+ Hotspots that are not running OSPF correctly and will cause HELL when a gateway dies.
I have same problem with x86 Mtik 3.10 and OSPF. Sometimes router changes gateway state to unreachable even though it’s reachable (it’s done for all ospf routes).
In meantime I tried version 3.22 and on one of our router we can see very strange behaviour. We don’t use user manager or hotspot. We only started ospf and log is now full of these messages. Beside the messages default password for admin is resetted and “Use radius” in AAA for user login is also resetted.
16:32:23 manager,error,info,debug failed to detect payment timeout: unable to op
en database file
16:51:54 manager,error,info,debug failed to cleanup session.acctSessionId: unabl
e to open database file
16:51:54 manager,error,info,debug failed to cleanup log.acctSessionId: unable to
open database file
16:51:54 manager,error,info,debug failed to cleanup incomplete users: unable to
open database file
17:03:35 manager,error,info,debug failed to detect payment timeout: unable to op
en database file
17:32:38 manager,error,info,debug failed to detect payment timeout: unable to op
en database file
18:01:50 manager,error,info,debug failed to detect payment timeout: unable to op
en database file
18:30:03 manager,error,info,debug failed to detect payment timeout: unable to op
en database file