Hello everyone,
I’m posting here to see if anyone else has encountered a similar issue with LDP. I initially opened a bug report a few months ago but haven’t received a response, so I’m hoping the community might have insights.
Issue Summary
We have a ring of 3 CCR2216 routers (B, C, and D) that have been having issues with LDP. From Router A, we’ve been attempting to enable LDP to reach Router D via both B and C. However, once LDP is enabled, traffic forwarding and login access to Routers C and E through F become unreliable or completely broken when the path involves Router C.
This issue first appeared on RouterOS 7.11.2, persisted through 7.16.2, and remains unresolved even after upgrading to 7.18 Stable. We’ve confirmed the issue is reproducible and occurs consistently across all RouterOS v7 versions we’ve tested.
Network Design
- We are running OSPF between sites.
- iBGP is used back to our route reflectors for the default route.
- Traffic from Router A to Router D breaks when both LDP paths (via Router B and C) are active.
- The topology is illustrated in the attached diagram.
MPLS/LDP Configuration:
/mpls interface
add disabled=no interface=all mpls-mtu=9000
/mpls ldp
add afi=ip disabled=no lsr-id=XX.XX.XX.XX transport-addresses=XX.XX.XX.XX
Questions:
Has anyone seen similar ECMP issues with LDP on RouterOS v7?
Are there known bugs or workarounds related to LDP ECMP path handling?
Thanks in advance for any insights or suggestions you might have.
