cancel
Showing results for 
Search instead for 
Did you mean: 
cancel
2400
Views
10
Helpful
11
Replies

Change upstream provider without losing BGP connection

netone.cisco
Level 1
Level 1

Hi everyone,

We are connected to 2 different providers (PROV01 and PROV02) with eBGP full internet tables. PROV01 routes have higher local preference over PROV02.

We are having problems with our provider and would like your expertise. PROV01, higher preference has been having trouble with their router. The BGP remains up but the router stops forwarding traffic. as a resulta we lose connection to the internet but our traffic never goes to PROV02 since PROV01 BGP remains up announcing the 400K routes to us.

Is there a way to test internet connection though PROV01 and as the "internet is DOWN" automatically change traffic to PROV02? Can BGP parameters changes be triggered by IP SLA?

Our router is an ASR1006 RP2.

Thank you

11 Replies 11

Sergey Fer
Level 1
Level 1

You may use OER/PfR for direct BGP manipulation. IP SLA can't directly manipulate BGP, but you may tie some EEM script to IP SLA and do whatever you want inside that script.

Sergey,

Thanks for replying. I will be sincere and say I have never ever heard about what you wrote above.

Of course I can reserch further but can you please give us a brief introduction what these are and links for starting to use these them? OER?PfR?EEM?

Thank you.

OER is for Optimized Edge Routing, PfR is for Perfomance Routing, which are different names of the same technology. It is targeted to control perfomance of a particular routes and change them without changng or resetting all routing. This may be done automatically through BGP attributes (Local Preference) manipulation.

EEM is for Embeeded Event Manager. It is a scripting tool that allows for doing whatever you want (execute IOS command, send syslog or even e-mail, ...) in response to any event. Here your event might be tracking event triggered by IP SLA.

You also may use more tricky tools but they, probably, would be a combination of IP SLA, conditional routing etc.

You may find info about OER/PfR here: http://www.cisco.com/en/US/products/ps6628/products_ios_protocol_option_home.html

and about EEM here: http://www.cisco.com/en/US/products/ps6815/products_ios_protocol_group_home.html

Sergey,

Many thanks. Quick reading seems just the solution. Unfortunatelly I have a loooooooottt of reading to do.

I'll keep you posted on the results and configs (it may take some time because of all the studying and requesting testing windows)

Mario

It's a good chance to study something new OER is a rather complex technology and is not too fast. If you need a solution that reacts to changes in a few seconds, probably better to start from scripting. It is more simple. Don't dig into TCL scripting. This feature will allow you become a Router God but it will be a little later. Start from EEM applets.

Hi,

Apart from OER/PfR, you can also try BFD for the same.

More on BFD can be found on

http://www.cisco.com/en/US/docs/ios/12_0s/feature/guide/fs_bfd.html

HTH,

Smitesh

Smitesh,

Thanks but unfortunately from the "BFP Operation" we read:

BFD provides a low-overhead, short-duration method  of detecting failures in the forwarding path between two adjacent  routers, including the interfaces, data links, and forwarding planes.  BFD is a detection protocol that you enable at the interface and routing  protocol levels. Cisco supports the BFD asynchronous mode, which  depends on the sending of BFD control packets between two systems to  activate and maintain BFD neighbor sessions between routers. Therefore,  in order for a BFD session to be created, you must configure BFD on both  systems (or BFD peers). Once BFD has been enabled on the interfaces and  at the router level for the appropriate routing protocols, a BFD  session is created, BFD timers are negotiated, and the BFD peers will  begin to send BFD control packets to each other at the negotiated  interval.

From the bold parts we see that both devices would have to be neighbours and managed by us. Which is not the case. the next router is from the provider and communication stops 2 hops after that.

A method using ping woul be needed.

Thank you

Tagir Temirgaliyev
Spotlight
Spotlight

Hi

this is very strange

because if the PE router stop pass traffic, and then a BGP session with other neighbors fall.

or it is mistake in provider network, just tell them about it.

I dont think you need here ip sla or something more. just make BGP whithout mistakes.

ttemirgaliye,

It is indeed a provider provider, they have been informed and admit it. Unfortunately no solution has been given until now and we have to find our own solution, avoid being woken up at night and keep service up for our customers.

Its a provider device and unfortunately not much help I can try to offer them on the causes of these.

Regards

MP

Hi

in this case ip sla may be simplest way.

just ping some router inside provider network

if no pings so shutdown interface to provider so BGP will switch all traffic to second provider.

for example

ip sla 10

icmp-echo 172.2.2.198 source-interface GigabitEthernet0/0

frequency 10

ip sla schedule 10 life forever start-time now

track 10 ip sla 10 reachability

delay down 20 up 20

!

event manager applet PE_down

event track 10 state down

action 1.0 syslog msg "PE_down"

action 2.0                                            -- conf t

action 3.0                                            -- int g0/0

action 4.0                                           -- shut

event manager applet PE_up

event track 10 state up

action 1.0 syslog msg "PE_up"

action 2.0

action 3.0

dont forget to rate post if it helps.

ttemirgaliye (whats your name?),

Thank you very helpfull indeed. I'm not familiar with EEM language (did not have time to read it yet) but looks simple to catch up.

One detail: How would this test the PE to be up if the source interface (g0/0 in your example) was shutdown? We would have to  manually go in and "unshut" the interface bringing the problem momentarily to check.

Following your logic I came up with the following (considering we can configure/deconfigure static routes with EEM and the same ip sla as above)

event manager applet PE_down

event track 10 state down

action 1.0 syslog msg "PE_down"

action 2.0                                            -- conf t

action 3.0                                            -- add static route to PE2

event manager applet PE_up

event track 10 state up

action 1.0 syslog msg "PE_up"

action 2.0                                        -- conf t

action 3.0                                        -- remove static route to PE2

Let me know what you all think please

Getting Started

Find answers to your questions by entering keywords or phrases in the Search bar above. New here? Use these resources to familiarize yourself with the community:

Innovations in Cisco Full Stack Observability - A new webinar from Cisco