This is one of those "works sometimes then stops working" things.

Basically, have 2 border routers (ciscos) talking OSPF with my cores (Mikrotik's).....

The mikrotiks see the OSPF routes to 0.0.0.0 from both borders in LSA at all times, and when you look at the LSA details for each EVERYTHING IS IDENTICAL.

Problem is sometimes they will install a route with both gateways (Correct). Other times they will only have one gateway. I dont get it. It should show both all the time. However, all my cores do the same at the same time making me think its somthing in the network other than the tiks. Is there anything that can make the tik not install a route with both gateways if the LSA's are identical?

Screenshots from webfig attached.

--John

does the cost for both LSAs are equal when one is installed and other is not?

Well, currently its taking the "xxx.xxx.xxx.2" route, and the cost of that is "1". I cant really test dropping that till Thursday when ill be at the site in NY......will let you know then.


--John

You can use "/routing ospf lsa print detail" to look at all LSAs and their costs, even the ones that aren't installed in the routing table.

Can you post that output and point out the two LSAs you're concerned with?

Did the command, but dont see costs in there....


[admin@core3] > /routing ospf lsa print detail

<Deleted all other routes from the output…..>

instance=default area=external type=as-external id=0.0.0.0
originator=xxx.xxx.xxx.1 sequence-number=0x800002B1 age=147 checksum=0x9D4C
options="DC" body=
netmask=0.0.0.0
forwarding-address=0.0.0.0
metric=1
route-tag=0x1
type2


instance=default area=external type=as-external id=0.0.0.0
originator=xxx.xxx.xxx.2 sequence-number=0x800000A9 age=1751
checksum=0xAD45 options="DC" body=
netmask=0.0.0.0
forwarding-address=0.0.0.0
metric=1
route-tag=0x1
type2

The metric is the cost.

Those are identical. They should both be installed as ECMP routes on all receiving routers.

If you only want one gateway installed at any time you should set different costs on the originators. If you want a different gateway to be used by different routers depending on their location in the network make it an external type 1 route so that the cost gets added up across links - type 2 costs are static.

Doesn't change that all routers should be showing both routes, though.

fewi, ok...so im not nuts here. Possible mikrotik bug? Defiantly not seeing both:

[admin@core3] /ip route> print detail
Flags: X - disabled, A - active, D - dynamic,
C - connect, S - static, r - rip, b - bgp, o - ospf, m - mme,
B - blackhole, U - unreachable, P - prohibit
0 ADo dst-address=0.0.0.0/0 gateway=xxx.xxx.xxx.2
gateway-status=xxx.xxx.xxx.2 reachable ether1-Backbone distance=110
scope=20 target-scope=10 ospf-metric=10 ospf-type=external-type-2


Interesting thing here is why is the metric 10 when I look at the installed route in the route list.....is that just the default metric for ospf routes brought into the installed routes list and has nothing to do with the metrics in OSPF routing?

--John

Huh. That's actually very weird. The OSPF metric should be the metric from the LSA, plus whatever your router adds for interface cost. In this case it's a type 2 external route so your router shouldn't add anything and the OSPF metric reported in "/ip route print detail where ospf" should be the same as the metric reported by the OSPF LSA - in this case both should be 1.

If it was a type 1 external route with an LSA metric of 1 (as received by your router) and the interface you're receiving through had a cost of 10 the reported OSPF metric of the LSA would be 1 and the OSPF metric of the IP route should be 10 + 1 = 11. You'd see the same for other routers such as intra-area routes.

What version are you running? There's definitely something wrong with what you're showing, but all of this is working fine for me on 5.6. There have been OSPF fixes in past versions.

Running 5.4 on the ones with the issue....

Didnt look till just now, but the one running 5.5 is actually working properly....so I think you might be right fewi.

Output from the 5.5 mikrotik:

[admin@core1] /ip route> pri detail
Flags: X - disabled, A - active, D - dynamic,
C - connect, S - static, r - rip, b - bgp, o - ospf, m - mme,
B - blackhole, U - unreachable, P - prohibit
0 ADo dst-address=0.0.0.0/0 gateway=xxx.xxx.xx.2,xxx.xxx.xx.1
gateway-status=xxx.xxx.xx.2 reachable ether1-backbone,xxx.xxx.xx.1 reachable ether1-backbone
distance=110 scope=20 target-scope=10 ospf-metric=10
ospf-type=external-type-2


Thank again fewi, wouldnt have thought it was a bug in the last version.... I will upgrade the other 2 boxes thursday and post when done.

--John

Upgaded the 2 1100's to 5.7 (They were at 5.4). So far so good, they are seeing both borders....

We will see if this keeps up -))

You going to mum in Vegas fewi? I definatly owe you a few beers...

--John

Probably. I live in Vegas.

Beers are always welcome...