Description

Northstar (or any PCE controller) gets overwhelmed by path computation requests sent by the PCC nodes, for delegated LSPs that are in a down state.

This has the effect of filling the logs / DB on the PCE controller.

This is especially of concern when there are LSPs that are down permanently (or for a prolonged time).

 

This can be seen in the PCEP traceoptions on the PCC.

This can also be seen in the output of "show mpls lsp name <name> extensive" output on the PCC

Such as the following:

(output truncated for clarity)

 

user@router> show mpls lsp name <name> extensive | no-more 

Ingress LSP: 6 sessions

 

11.0.0.104

 From: 11.0.0.101, State: Dn, ActiveRoute: 0, LSPname: <name>, LSPid: 5

...

  69428 Apr 9 01:53:26.664 11.104.107.1: No Route toward dest

  69427 Apr 9 01:53:26.651 EXTCTRL LSP: Sent Path computation request and LSP status

  69426 Apr 9 01:53:26.651 EXTCTRL_LSP: Computation request/lsp status contains: signalled bw 0 req BW 0 admin group(exclude 0 include any 0 include all 0) priority setup 7 hold 0 hops: 11.101.105.2 11.105.107.2 11.104.107.1

  69425 Apr 9 01:53:26.650 LSP-ID: 3230 created

  69424 Apr 9 01:53:26.650 Originate Call

  69423 Apr 9 01:53:26.649 Clear Call

...

  69417 Apr 9 01:52:57.181 11.104.107.1: No Route toward dest

  69416 Apr 9 01:52:57.154 EXTCTRL LSP: Sent Path computation request and LSP status

  69415 Apr 9 01:52:57.154 EXTCTRL_LSP: Computation request/lsp status contains: signalled bw 0 req BW 0 admin group(exclude 0 include any 0 include all 0) priority setup 7 hold 0 hops: 11.101.105.2 11.105.107.2 11.104.107.1

  69414 Apr 9 01:52:57.153 LSP-ID: 3229 created

  69413 Apr 9 01:52:57.153 Originate Call

  69412 Apr 9 01:52:57.152 Clear Call

...

  69406 Apr 9 01:52:28.593 11.104.107.1: No Route toward dest

  69405 Apr 9 01:52:28.578 EXTCTRL LSP: Sent Path computation request and LSP status

  69404 Apr 9 01:52:28.578 EXTCTRL_LSP: Computation request/lsp status contains: signalled bw 0 req BW 0 admin group(exclude 0 include any 0 include all 0) priority setup 7 hold 0 hops: 11.101.105.2 11.105.107.2 11.104.107.1

  69403 Apr 9 01:52:28.577 LSP-ID: 3228 created

  69402 Apr 9 01:52:28.577 Originate Call

  69401 Apr 9 01:52:28.576 Clear Call

...

  69395 Apr 9 01:51:59.948 11.104.107.1: No Route toward dest

  69394 Apr 9 01:51:59.931 EXTCTRL LSP: Sent Path computation request and LSP status

  69393 Apr 9 01:51:59.930 EXTCTRL_LSP: Computation request/lsp status contains: signalled bw 0 req BW 0 admin group(exclude 0 include any 0 include all 0) priority setup 7 hold 0 hops: 11.101.105.2 11.105.107.2 11.104.107.1

  69392 Apr 9 01:51:59.930 LSP-ID: 3227 created

  69391 Apr 9 01:51:59.930 Originate Call

...

 Created: Thu Mar 28 14:37:51 2024

Total 1 displayed, Up 0, Down 1

...

 

user@router> 

 

Solution

From the side of the PCC node that is a Juniper router, this interval can be controlled by this configuration:
set protocols mpls label-switched-path <LSP-name> retry-timer <value in seconds>
The default value of this setting is 30 seconds.
What occurs here is that, after an LSP goes down, the node will keep trying to bring it up, and every time, a report gets sent to the controller.

Modification History

2024-04-09 : Article Created