RouterOS v7 CCR: OSPF Routes Visible but No Ping After Link Down

RouterOS Forwarding Protocols
1

Good afternoon everyone,
I’d like to ask if anyone here is currently using or has experience running CCR with RouterOS v7 and operating both BGP and OSPF simultaneously.
I’m currently facing an issue related to OSPF behavior on my setup.
Here’s my topology overview — I have four routers running OSPF and all are peering successfully with each other:

Huawei-A <-> CCR-A
Huawei-A <-> Huawei-B
Huawei-B <-> CCR-B
CCR-A <-> CCR-B

So basically, it’s a square-like topology with 2 Huawei routers and 2 CCRs.
All routers are fully connected and OSPF adjacencies are established correctly.
However, the issue happens when one of the Huawei–CCR links goes down.
For example:

When the Huawei-A <-> CCR-A link is down, CCR-A cannot ping any IPs on Huawei-A or Huawei-B, but CCR-B can.

Similarly, when the Huawei-B <-> CCR-B link is down, CCR-B cannot ping any IPs on Huawei-B or Huawei-A, but CCR-A can.

I’ve checked the routing tables on both CCRs — all routes from Huawei-A and Huawei-B are still present and learned via OSPF when one link is down.
However, for some reason, the ping only replies from one CCR, not both.
Has anyone experienced similar behavior before?
Any insights or suggestions would be greatly appreciated.
Thanks in advance!

2

Can anyone help with this?
My initial suspicion is that there seems to be a problem between the ccr-a and ccr-b routing, but there's also the possibility that there's a bug in the ccr router.

3

Could you please share your configuration?
Do routes change somehow when links go down?
Do gateways change?
What are distances in the routing table? Do they change?

4

Likely return-path asymmetry when the link drops: routes stay visible but replies take the other CCR.
Test with a src-address ping and check RP-filter/firewall and OSPF costs to ensure a symmetric return path.

Indeed BartoszP is right, share more informations :slight_smile: