HealthBot and NorthStar on JCL

Owned by Jean-luc KRIKER
Last updated: Nov 30, 2020
3 min read




blueprint:

NorthStar (v5) and HealthBot (v2.0.2)

https://portal.cloudlabs.juniper.net/RM/Diagram/Index/fb6c1af8-a503-47f1-b283-e8fb3dbd3309?diagramType=Topology



video
Log into the GUI
NorthStar or NS login

https://<ip@>:<port>         (   admin // Juniper!1    )

HealthBot

SSH or https://                      ( jcluser // Juniper!1    )

Add Allowed network prefixCommands menu >> Add Allowed network prefix >>> <enter ip address of client>
CLI on Healthbot

Start healthbot_listener.py

( Webhook + trigger NS to put the link in maintenance)

Log into ssh jcluser@healthbot     ( jcluser // Juniper!1  )

It's a webhook, listen for HB messages 

Usage:

jcluser@ubuntu:~$ ls
healthbot-2.0.2-1.deb self_healing


jcluser@ubuntu:~$ cd self_healing/

jcluser@ubuntu:~/self_healing$ python healthbot_listener.py

Create two Tunnels: PCEP and NETCONF

Tunnel 1 = jlk_4_to_1_PCEP  ( Control type: PCEInitiated )     Path Computation Element Protocol

Tunnel 2 = jlk_4_to_1_Netconf  ( Control type: Device Controlled )

and

Application Menu >> Path Optimization >>  path optimization >> Enable + 1 minute




Set the Path Optimization every minutes



Display the Delay Tab




Connect on the CentOS ( to create delay )
add delay 

Log into the CentOS  ( jcluser / Juniper!1 )

su -   ( Juniper!1 )

./add_delay.sh


[root@CentOS ~]# more add_delay.sh
tc qdisc add dev eth1 root netem delay 300ms
echo "300 ms delay added between vMX-5 and vMX-6"
[root@CentOS ~]#



SSH to HealthBot
healthbot_listener.py output

Monitor Output 

received delay alert
HIGH DELAY DETECTED for vMX-6 7.105.106.2
PERFORMING EXHUASTIVE LINK FAILURE SIMULATION for vMX-6 7.105.106.2

SIMULATION RESULT true
HIGH DELAY DETECTED PUT LINK UNDER MAINTENANCE::

On NorthStar


1- Link goes on maintenance mode

2- Tunnel re-router to avoid the link

Go to the maintenance tab