Common Cases Where Routing Loops Occur During Redistribution

  1. Mutual Redistribution Between Two Protocols

  • Scenario: Redistributing OSPF into BGP and BGP back into OSPF.

  • Risk: A route learned from OSPF gets injected into BGP, then redistributed back into OSPF—creating a loop.

  1. Redistributing iBGP into IGP

  • Scenario: iBGP routes redistributed into OSPF or EIGRP.

  • Risk: iBGP routes can be seen as internal by IGP, then re-advertised to other BGP speakers, violating BGP split-horizon rules.

  • Must explicitly deny iBGP in route-maps unless the intension is to redistribute it.

  1. Lack of Route Tagging

  • Scenario: Routes redistributed without tagging.

  • Risk: No way to identify the origin of a route, so it gets re-learned and re-injected endlessly.

  • Best Practice: Use set tag in route-maps and filter with match tag.

  1. Overlapping Administrative Distances

  • Scenario: Redistributed routes have similar ADs across protocols.

  • Risk: RIB may prefer a redistributed route over the native one, causing suboptimal paths or loops.

  • Solution: Adjust ADs using distance command in route-maps.

  1. Improper Filtering

  • Scenario: Redistributing all routes without prefix-lists or route-maps.

  • Risk: Unintended routes get redistributed and re-learned.

  • An empty prefix-list permits all routes—be explicit.

How to Avoid Routing Loops

  1. Use Route Tagging

  1. Apply Prefix Lists and Route Maps Filter only necessary prefixes.

  1. Adjust Administrative Distance

  • Lower AD for trusted sources, raise for redistributed ones.

  1. Avoid Mutual Redistribution

  • Prefer one-way redistribution with clear boundaries.

  • If mutual redistribution is required, tag and filter aggressively.

  1. Use redistribute internal or external selectively

  • IOS XR and NX-OS support filtering based on route type.

Example

  1. Mutual Redistribution Between Two Protocols

  • Scenario: OSPF ↔ BGP redistribution Loop Risk: A route from OSPF gets into BGP, then back into OSPF—creating a loop.

  • Strategy:

    • Tag routes when redistributing from OSPF to BGP (set tag 65001)

    • Deny re-entry of tagged routes into OSPF

  • NX-OS Example:

  1. Redistributing iBGP into IGP

  • Scenario: iBGP routes redistributed into OSPF Loop Risk: iBGP routes re-advertised to other BGP speakers via IGP—violating split-horizon.

  • Strategy:

    • Deny redistribution of iBGP (internal) routes

    • Permit only eBGP (external) routes

  • IOS XE Example:

  1. Lack of Route Tagging

  • Scenario: Redistributed routes not tagged Loop Risk: No way to identify origin—routes re-learned and re-injected

  • Strategy:

    • Use set tag in route-policy

    • Filter tagged routes on re-entry

  • IOS XR Example:

  1. Overlapping Administrative Distances

  • Scenario: Redistributed routes have similar ADs Loop Risk: RIB prefers redistributed route over native one

  • Strategy:

    • Raise AD of redistributed routes to deprioritize them

    • Ensure native IGP routes are preferred

  • NX-OS Example:

  1. Improper Filtering

  • Scenario: Redistributing all routes without control Loop Risk: Unintended routes get redistributed and re-learned

  • Strategy:

  • Use prefix-lists to filter only necessary routes

  • Avoid implicit permit (empty prefix-list)

  • IOS XE Example:

Last updated