PfRv3 priorities purpose inside threshold policies

[damienpicard]

Hello everyone,

I have a small question for which I can't find any clue inside the Cisco documentation;

Can someone explain me the purpose of the Threshold priorties inside a specific policy used for PfRv3 ? For example :

HubMC(config-domain-vrf-mc)# class EXAMPLE sequence 100
HubMC(config-domain-vrf-mc-class)# match dscp default policy custom
HubMC(config-domain-vrf-mc-class-type)# priority 2 loss threshold 5
HubMC(config-domain-vrf-mc-class-type)# priority 1 one-way-delay threshold 50
HubMC(config-domain-vrf-mc-class-type)# priority 3 jitter threshold 200000
HubMC(config-domain-vrf-mc-class-type)# exit
HubMC(config-domain-vrf-mc-class)# path-preference MPLS fallback INET

I can't find any clear information on when the router decide to switch the traffic on the fallback ISP but I assume it happens as soon as a threshold is exceeded. Then what's the purpose of these priorities ? Or am I wrong, and maybe it is not going to switch through the fallback ISP if only a priority 3 threshold is exceeded for example ?

Thanks for your help !

Regards,

Damien.

[WILLIAM STEGMAN]

It's definitely obscure.  After much searching, the only thing I can find is that it looks like it might be used in conjunction with the resolve command when there are conflicting actions to take on traffic classes that might be triggered by multiple thresholds being crossed.

http://www.cisco.com/c/en/us/td/docs/ios-xml/ios/pfr/command/pfr-cr-book/pfr-h1.html#wp2068742995

https://aitaseller.wordpress.com/2012/10/15/pfr-cisco-performance-routing/

[Marek.Bedy]

I agree with you, it is absolutely unclear and not documented, even in 2018.

Anyone can clarify it?