12-01-2016 05:55 PM - edited 03-01-2019 05:06 AM
Hi,
Was wondering if anyone has come across this? Getting this minor fault on an L3Out with SVIs to 8 diffent links (2 nodes and 4 interfaces each).
The error says MTU but that can't be it as we do have full OSPF adjacencies.
These interfaces go down to ESX hosts in a cluster so lots going on over these trunks.
Thanks in advance for any pointers!
minor
2016-11-30T01:36:29.383+00:00
Raised
topology/pod-1/node-201/local/svc-policyelem-id-0/uni/epp/rtd-[uni/tn-common/out-L3_Edge_RoutedOut/instP-L3_Edge_RoutedOut-EPG]/node-201/stpathatt-[eth1/25]/nwissues
Configuration failed for uni/tn-common/out-L3_Edge_RoutedOut/instP-L3_Edge_RoutedOut-EPG node 201 eth1/25 due to Operational interface parameters cannot be different (e.g. MTU) from different deployments under a context
12-02-2016 06:22 AM
Claudia,
What version APIC are you running? I would go back through your access policies and verify everything is ok from there. Do you still have connectivity? What routing protocol are you using? This could be related to a bug when mutliple l30uts are deployed on the same node using the same context.
Thanks,
Michael G.
12-02-2016 06:59 AM
Hi Michael, Robert,
We are on 2.0(1o) and everything seems to be working well and we do have connectivity. We are using OSPF for this L3Out.
To Roberts point, yes it did look like MTU was an example but the admin for the routers with which we have an adjacency reported that they were seeing MTU error messages but it can't actually be that since we have OSPF adjacency.
Let me review the access policies again.
Thx!
---------- ---------- ------------------------ --------------------
controller 1 ctl01 2.0(1o)
controller 2 ctl02 2.0(1o)
controller 3 ctl03 2.0(1o)
12-02-2016 07:24 AM
In the OSPF Interface Policy are you hard setting the MTU or using "ignore-mtu" ?
Whats configured (MTU) on the other router?
Robert
12-02-2016 07:44 AM
Hey Robert,
MTU is set at 9000. I asked the admin on the other routers (NSX Edge Routers) to set them to 9000 as well. I've asked for confirmation as well the exact error they are seeing.
Access profiles match (10G, CDP off, MCP off, LLDP on, same AEP). They are applied to the right leafs and in a "formed" state.
If there is any ovrlap I can't find it. Interestingly enough, there are 2 different ACI fabrics and two ACI - NSX Edge OSPF adjacencies and they both have the same set of faults (on each interface that goes down to the ESX cluster). Of course, I'm a point of commonality as well :D
12-02-2016 06:31 AM
This is not specifically MTU related. The fault simply states this as an example.
Are you using the same OSPF Interface policy on all Node Interfaces?
Agree with Michael it's most likely an access policy discrepancy.
Robert
Discover and save your favorite ideas. Come back to expert answers, step-by-step guides, recent topics, and more.
New here? Get started with these tips. How to use Community New member guide