Changing the configuration using event policy action

Starting with 12.1, JunOS allows you to change the configuration by the event policy.

In the past this was possible only by using the event policy along with an event script. The first one was detecting that a configuration change had to be done and the second one was commiting the configuration change.

Let’s see this in action.

This is topology used:

(lo0-10.255.21.243/32)R1(ge-0/0/0-1.1.1.1/24) ———- (ge-0/0/0-1.1.1.2/24)R2(lo0-10.255.26.130/32)

R1 has a static route to R2 lo0 interface:

 

[edit]
root@R1# show routing-options 
static {
    route 10.255.26.130/32 next-hop 1.1.1.2;
}
[edit]
root@R1# run ping 10.255.26.130 
PING 10.255.26.130 (10.255.26.130): 56 data bytes
64 bytes from 10.255.26.130: icmp_seq=0 ttl=64 time=2.351 ms
64 bytes from 10.255.26.130: icmp_seq=1 ttl=64 time=2.969 ms
^C
--- 10.255.26.130 ping statistics ---
2 packets transmitted, 2 packets received, 0% packet loss
round-trip min/avg/max/stddev = 2.351/2.660/2.969/0.309 ms

[edit]
root@R1#

Using the below event-options configuration and the RPM configuration, you can have the router to deactivate the static route in case the IP from R2 is not reachable anymore:

[edit]
root@R1# show event-options 
policy PING_NOT_OK {
    events PING_TEST_FAILED;
    attributes-match {
        ping_test_failed.test-owner matches PING-Lo0;
        ping_test_failed.test-name matches to-R2;
    }
    then {
        change-configuration {
            commands {
                "deactivate routing-options static route 10.255.26.130/32";
                "deactivate event-options policy PING_NOT_OK";
                "activate event-options policy PING_OK";
            }
        }
    }
}
policy PING_OK {
    events PING_TEST_COMPLETED;
    attributes-match {
        ping_test_completed.test-owner matches PING-Lo0;
        ping_test_completed.test-name matches to-R2;
    }
    then {
        change-configuration {
            commands {
                "activate routing-options static route 10.255.26.130/32";
                "deactivate event-options policy PING_OK";
                "activate event-options policy PING_NOT_OK";
            }
        }
    }
}

[edit]
root@R1# 

[edit]
root@R1# show services 
rpm {
    probe PING-Lo0 {
        test to-R2 {
            probe-type icmp-ping;
            target address 1.1.1.2;
            probe-interval 1;
            test-interval 5;
        }
    }
}

[edit]
root@R1#

Using this configuration, the router is monitoring only PING_NOT_OK policy as long as the route is active. If the route is inactive, then the router is monitor the policy PING_OK.

As you can see the R2’s Loopback interface is reachable from R1 according from RPM results:

 

[edit]
root@R1# run show services rpm probe-results          
    Owner: PING-Lo0, Test: to-R2
    Target address: 10.255.26.130, Probe type: icmp-ping, Test size: 1 probes
    Probe results:
      Response received, Tue Jul 22 12:08:43 2013, No hardware timestamps
      Rtt: 2965 usec
    Results over current test:
      Probes sent: 1, Probes received: 1, Loss percentage: 0
      Measurement: Round trip time
        Samples: 1, Minimum: 2965 usec, Maximum: 2965 usec, Average: 2965 usec, Peak to peak: 0 usec, Stddev: 0 usec, Sum: 2965 usec
    Results over last test:
      Probes sent: 1, Probes received: 1, Loss percentage: 0
      Test completed on Tue Jul 22 12:08:43 2013
      Measurement: Round trip time
        Samples: 1, Minimum: 2965 usec, Maximum: 2965 usec, Average: 2965 usec, Peak to peak: 0 usec, Stddev: 0 usec, Sum: 2965 usec
    Results over all tests:
      Probes sent: 465, Probes received: 350, Loss percentage: 24
      Measurement: Round trip time
        Samples: 350, Minimum: 1957 usec, Maximum: 40622 usec, Average: 2961 usec, Peak to peak: 38665 usec, Stddev: 2171 usec, Sum: 1036345 usec

[edit]
root@R1# run show services rpm probe-results    
    Owner: PING-Lo0, Test: to-R2
    Target address: 10.255.26.130, Probe type: icmp-ping, Test size: 1 probes
    Probe results:
      Response received, Tue Jul 22 12:08:58 2013, No hardware timestamps
      Rtt: 3073 usec
    Results over current test:
      Probes sent: 1, Probes received: 1, Loss percentage: 0
      Measurement: Round trip time
        Samples: 1, Minimum: 3073 usec, Maximum: 3073 usec, Average: 3073 usec, Peak to peak: 0 usec, Stddev: 0 usec, Sum: 3073 usec
    Results over last test:
      Probes sent: 1, Probes received: 1, Loss percentage: 0
      Test completed on Tue Jul 22 12:08:58 2013
      Measurement: Round trip time
        Samples: 1, Minimum: 3073 usec, Maximum: 3073 usec, Average: 3073 usec, Peak to peak: 0 usec, Stddev: 0 usec, Sum: 3073 usec
    Results over all tests:
      Probes sent: 468, Probes received: 353, Loss percentage: 24
      Measurement: Round trip time
        Samples: 353, Minimum: 1957 usec, Maximum: 40622 usec, Average: 2962 usec, Peak to peak: 38665 usec, Stddev: 2163 usec, Sum: 1045596 usec

[edit]
root@R1#

so this means that PING_OK policy should be deactivated:

 

[edit]
root@R1# show event-options 
policy PING_NOT_OK {
    events PING_TEST_FAILED;
    attributes-match {
        ping_test_failed.test-owner matches PING-Lo0;
        ping_test_failed.test-name matches to-R2;
    }
    then {
        change-configuration {
            commands {
                "deactivate routing-options static route 10.255.26.130/32";
                "deactivate event-options policy PING_NOT_OK";
                "activate event-options policy PING_OK";
            }
        }
    }
}
inactive: policy PING_OK {
    events PING_TEST_COMPLETED;
    attributes-match {
        ping_test_completed.test-owner matches PING-Lo0;
        ping_test_completed.test-name matches to-R2;
    }
    then {
        change-configuration {
            commands {
                "activate routing-options static route 10.255.26.130/32";
                "deactivate event-options policy PING_OK";
                "activate event-options policy PING_NOT_OK";
            }
        }
    }
}

[edit]
root@R1#

Let’s deactivate unit 0 from ge-0/0/0 interface of R2:

 

[edit]
root@R2# deactivate interfaces ge-0/0/0.0 

[edit]
root@R2# commit 
commit complete

[edit]
root@R2#

and let’s try to ping R2 from R1:

 

[edit]
root@R1# run ping 10.255.26.130 
PING 10.255.26.130 (10.255.26.130): 56 data bytes
ping: sendto: No route to host
ping: sendto: No route to host
^C
--- 10.255.26.130 ping statistics ---
2 packets transmitted, 0 packets received, 100% packet loss

[edit]
root@R1#

We should confirm that the configuration of event-options has changed and the static route has been deactivated:

 

[edit]
root@R1# show event-options 
inactive: policy PING_NOT_OK {
    events PING_TEST_FAILED;
    attributes-match {
        ping_test_failed.test-owner matches PING-Lo0;
        ping_test_failed.test-name matches to-R2;
    }
    then {
        change-configuration {
            commands {
                "deactivate routing-options static route 10.255.26.130/32";
                "deactivate event-options policy PING_NOT_OK";
                "activate event-options policy PING_OK";
            }
        }
    }
}
policy PING_OK {
    events PING_TEST_COMPLETED;
    attributes-match {
        ping_test_completed.test-owner matches PING-Lo0;
        ping_test_completed.test-name matches to-R2;
    }
    then {
        change-configuration {
            commands {
                "activate routing-options static route 10.255.26.130/32";
                "deactivate event-options policy PING_OK";
                "activate event-options policy PING_NOT_OK";
            }
        }
    }
}

[edit]
root@R1# show routing-options    
static {
    inactive: route 10.255.26.130/32 next-hop 1.1.1.2;
}

[edit]
root@R1#

When the commit has happened, this was seen in the logs:

 

Jul 22 12:15:27  R1 eventd: EVENTD_CONFIG_CHANGE_SUCCESS: Configuration change successful: while executing policy PING_NOT_OK with user root privileges

The following two tabs change content below.

Paris ARAU

Paris ARAU is a networking professional with strong background on routing and switching technologies. He is a holder of CCIE R&S and dual JNCIE(SP and ENT). The day to day work allows him to dive deeply in networking technologies. Part of the continuously training, he is focusing on Software Defined Network and cloud computing.

Comments

This post currently has 12 responses

  • Nice post! But in my case was not working unless I added to rpm:

    thresholds {
    successive-loss 1;
    }

      • hi,

        is this config supported on Ex switches ? this can be used as a replacement for IP monitoring services which is not supported in EX switches .

        • Yes, this config is supported on EX switches.

          Actually the post is based on a config tested on EX4200.

          Thanks,
          Paris

  • Hi !
    your rpm config tests for the wrong address ( destination of static route instead of next-hop of static route)
    the probe should test the 1.1.1.2 address ( next-hop of the static route)
    it cannot test the destination loopback as there is no route to that address when the interface has been switched off and later switched on.

    • Hi Pawel,

      I haven’t tried this combination, but unless you don’t get an error when you are trying to commit the configuration, it should work.

      Thanks,
      Paris

      • Hi Paris,

        Thanks for your replay.
        Would you have any idea which attributes-match should I use to match the http get request ?

        Pawel

        • Hi Pawel,

          I spent some time going through the attributes-match options(in 15.1R5), but none of them seemed to be the one you are looking for.

          However, it’s possible that the option is there, but not with a self-descriptive name, hence I couldn’t find it.

          Thanks,
          Paris

Leave a Reply to Vlad Cancel reply

Your email address will not be published. Required fields are marked *

Sidebar



%d bloggers like this: