Multicast RP time-out

[bryanrobh]

This is the scenario. I have multiple routers running in pim spare-dense mode. When the RP goes down they will retry to find the RP for 30 seconds and then 2 more times until it finally times out after a minute and a half then rolls over to dense mode. 

I want ant to have those timers much faster but I also don’t want to flood the link with hello packets. Any ideas

[Hector Gustavo Serrano Gutierrez]

Perhaps you may find deploying Anycast-RP technique in two or more Routes with the use of MSDP in Cisco IOS based Routers a better approach to accomplish fast failover in case of a failure of the RP in PIM sparse mode.

See Anycast RP

PIM Dense mode (or Sparse-Dense mode) is not recommended and not even supported by some other platforms (like the Cisco NX-OS family of Layer 3 Switches) due to Dense's mode inefficient method to deliver Multicast to the intended receivers.

Back to your original question, you are probably advertising the RP via BSR or Auto-RP methods. You may need to tune these timers to remove the RP from the Routers in the network.

HTH.

[Georg Pauwen]

Hello,

try and set the 'ip pim query-interval' to the lowest configurable value (1, default is 30):

interface GigabitEthernet0/0

ip pim query-interval 1

[paul driver]

Hello

---

@bryanrobh wrote:

This is the scenario. I have multiple routers running in pim spare-dense mode. When the RP goes down they will retry to find the RP for 30 seconds and then 2 more times until it finally times out after a minute and a half then rolls over to dense mode. 

---

@Hector Gustavo Serrano Gutierrez  mentions a valid point the best options for RP resiliency would be BSR/anycast

Depending on what RP resiliency you are using, (auto-rp, bsr) you can configure the RP announcement interval when you specify the RP.
If i remember correctly the expiration value is 3x the specified announcement interval

Auto-rp  - ip pim send-rp-announce <interface> <scope> interval xx
BSR - ip pim rp-candidate  <interface> interval xx 

[Hector Gustavo Serrano Gutierrez]

I agree with @paul driver 

The deployment can take advantage of the Anycast-RP technique for redundancy and fast failover in case one of the RP Routers is lost.

Anycast-RP and BSR can happily work together:

- Anycast-RP is accomplished thanks to MSDP.

- The RP information can still be advertised using BSR.

Worst case scenario, IP PIM Sparse-Dense mode can be kept in place in case of both (or any number of configured RPs) are all lost.

About the timers:

RP-A

RP-A(config)#ip pim rp-candidate lo0 interval ?
  <1-16383>  number of seconds

Default is 60 seconds.

ip pim rp-candidate Loopback0
ip pim bsr-candidate Loopback0

Routers with PIM Sparse-Dense fall back to Sparse mode only when loosing the RP information. Hence, only by tunning the BSR timers is how fall back could happen faster.

In my own experience, leveraging to Anycast-RP is a better approach than relying on PIM Dense mode.

DISCLAIMER:

The configurations discussed in this post are merely templates and not final configurations that you can copy & paste to your network devices in production. You need to review, evaluate and modify the configurations at your best convenience to make sure the results are the ones you intended and introduce those changes to you network only during maintenance windows. The author of this post is not responsible of unintended consequences by failing to follow this disclaimer note.