can ospf table-map use with metric cost in nexus 9K?

[sanhtethlaing]

 ospf table-map using distance for TE  cisco document;

can i use metric(cost) instead of distance?

please suggest me

thanks

kindly see attached

[Giuseppe Larosa · Accepted solution]

Hello @sanhtethlaing ,

the route-map foo in your example sets the Admin distance for different types of routes or using more specific match conditions.

setting the metric would be an action that interfers with OSPF link state database and SPF algorythm for this reason I think it is not supported.

Edit:

as you can see in the following configuration guide the table map is intended to be used only to modify AD of routes

https://www.cisco.com/c/en/us/td/docs/dcn/nx-os/nexus9000/105x/unicast-routing-configuration/cisco-nexus-9000-series-nx-os-unicast-routing-configuration-guide/m_configuring_ospfv2.html#id_87051

So as I have written above you cannot set the metric using a table map .

Hope to help

Giuseppe

[MHM Cisco World]

I will check this in lab

MHM

[sanhtethlaing]

thanks for your reply

may i know any suggestion, to do traffic engineering on nexus 9k, i currently two 9k as primary and secondary, but routing was going via both switches

so let me know best practice to manipulate route traffic

thanks again

[MHM Cisco World]

the Metric (cost) can not use as @Giuseppe Larosa  mention 

""may i know any suggestion, to do traffic engineering on nexus 9k, i currently two 9k as primary and secondary, but routing was going via both switches""
table-map is inject route into RIB in NSK, you need to prefer path in Peer connect to two NSK not use table-map

MHM

[Giuseppe Larosa]

Hello @sanhtethlaing ,

>> may i know any suggestion, to do traffic engineering on nexus 9k, i currently two 9k as primary and secondary, but routing was going via both switches

by default OSPF supports ECMP = Equal Cost Multi Path and devices perform flow based load balancing.

if  you want to have a primary  / backup the easy move is to increase OSPF cost on both sides of link between Nexus2 (the intended secondary) and the peer device(s).

Take in account also the possible different settings on the autoreference bandwidth that is 40 Gbps in NXOS and 100 Mbps in traditional IOS , IOS XE.

The command in IOS XE is ip ospf cost <value> in interface mode . 

You can check current OSPF costs using the following on all involved devices

show ip ospf interface

You need to use the same command both on peer and on Nexus2.

Other possible options are available depending on the type of OSPF routes and if you are using an OSPF multi area design (as opposed to single area).

If there are not security devices in the path like transparent FWs or IPS you may be fine to stay with ECMP.

Also FWs in routed mode can be in the path and they need to see both directions of each flow to properly work.

Another possible reason for having Primary / Secondary routing are the use of network analyzers.

Hope to help

Giuseppe