interim-forwarding-loops-in-ospf-or-is.html

---
url: /2017/05/interim-forwarding-loops-in-ospf-or-is/
title: "Interim Forwarding Loops in OSPF or IS-IS Networks"
date: "2017-05-04T08:12:00.000+02:00"
tags: [ IP routing ]
---

<p><a name="_GoBack"></a>One of my readers sent me this question (slightly rephrased):</p>
<blockquote class="cite">Assume you have A,B and C connected in a triangle (with an alternate longer path to C). What happens if C loses its links to A and B? Won’t the traffic to C loop between A and B for a while?</blockquote>
<p>As always, it depends.<!--more--></p>
<p>Here’s the network topology diagram we’ll use (thanks to ASCIIFlow Infinity):</p>
<pre class="code">+-------+       +-------+<br/>|   A   +-------+   B   +-----+<br/>+--+----+       +----+--+     |<br/>   |                 |        |<br/>   |    +-------+    |    +---+---+<br/>   +----+   C   +----+    | Slow  |<br/>        +---+---+         +---+---+<br/>            |                 |<br/>            +-----------------+</pre><p>The actual sequence of events happening on a router obviously depends on particular details of control plane implementation, but it’s reasonable to expect something along these lines:</p>
<ul class="ListParagraph"><li>Forwarding entries using interface X are removed as soon as interface X goes down;</li>
<li>Lacking alternate ECMP entries to the same destination, the router would install pre-computed backup entries (for example, the results of LFA computation) into the forwarding table;</li>
</ul>
<p class="more">We’ll ignore the details of how the backup entries are installed. Ideally the <a href="/2012/01/prefix-independent-convergence-pic/">control-plane software changes the next-hop groups</a> not the actual forwarding entries.</p>
<ul class="ListParagraph"><li>Lacking pre-computed backup entries, the router will recalcuate the main routing table (there might be alternate routes with higher administrative distance) and repopulate the forwarding table if such entries exist.</li>
</ul>
<p>There’s nothing else the routing/forwarding table manipulation software can do at that moment. </p>
<p>In our scenario, A and B have a <a href="/2012/01/loop-free-alternate-ospf-meets-eigrp/">loop-free alternate path</a> to C, and if you configured LFA on them the FIB manipulation software would install the alternate route (A =&gt; B and B =&gt; A) resulting in a temporary forwarding loop… unless you configured <a href="http://www.cisco.com/c/en/us/td/docs/ios-xml/ios/iproute_ospf/configuration/xe-3s/iro-lfa-frr-xe.html#GUID-08D08712-F317-4BAF-A45B-0AC5AFA374AD">LFA Downstream Path</a>.</p>
<p>If you haven’t configured LFA, A and B will have no usable route to C (we’ll ignore summary/default routes for the moment) and the traffic sent to C will be dropped.</p>
<p>At the same time, the routing protocols monitoring interface X kick in and start their work:</p>
<ul class="ListParagraph"><li>New router LSA/LSP is generated (assuming we’re not dealing with P2P links) unless there have been so many recent changes that the LSA generation is throttled;</li>
<li>New LSA/LSP describing local topology change is flooded. These update packets might be delayed based on any OSPF/IS-IS packet pacing configured on the device;</li>
<li>Asynchronously to that, SPF process is eventually run based on how SPF timers are configured, generates new best routes, and sends them to the routing table.</li>
</ul>
<p>Now for the “<em>it depends” </em>part:</p>
<ul class="ListParagraph"><li>If the SPF process is run before the changed LSAs are received from the OSPF/IS-IS neighbor, we’ll get a temporary loop (assuming LFA hasn’t already generated one) that will disappear the second time the SPF process is run;</li>
<li>If the changed LSA is received before the SPF process is run (and LFA was not used), there will be no forwarding loop.</li>
</ul>
<p class="warn">Please don’t read the above paragraphs as <em>LFA is bad</em>. It is not. This blog post is evaluating the consequences of a rare event (multiple link loss) from the <em>understanding RIB, FIB and SPF </em>perspective. If the network loses a single link (A =&gt; C or B =&gt; C) using LFA results in faster convergence.</p>
<p>For even more details, read the <a href="http://packetpushers.net/microloop/">Microloops!</a> blog post by Russ White.</p>