Zabbix LinuxHA Monitoring

Tested on Debian.

Install

• Copy the script linuxha.py to the target system (executable by the zabbix user)
• Copy the userparameter to the target system
• Add the sudoers privileges via visudo -f /etc/sudoers.d/zabbix_linuxha
• Restart the Zabbix Agent service
• Import the Template on your Zabbix Server

---

Testing

On the target system:

root@SRV1:~$ su zabbix --login --shell /bin/bash

zabbix@SRV1:~$ python3 /usr/local/bin/zabbix/linuxha.py discover
> {"data": [{"{#LINHA_RES}": "resIP"}, {"{#LINHA_RES}": "resService"}]}

zabbix@SRV1:~$ python3 /usr/local/bin/zabbix/linuxha.py members
> 2

zabbix@SRV1:~$ python3 /usr/local/bin/zabbix/linuxha.py members_active
> 2

zabbix@SRV1:~$ python3 /usr/local/bin/zabbix/linuxha.py votes
> 1

zabbix@SRV1:~$ python3 /usr/local/bin/zabbix/linuxha.py quorum
> 1

zabbix@SRV1:~$ python3 /usr/local/bin/zabbix/linuxha.py resource resIP
> 1

zabbix@SRV1:~$ python3 /usr/local/bin/zabbix/linuxha.py resource_active resIP
> 1

---

Usage

The script expects all resources to have a name that starts with a prefix. By default res. You can change it in the script.

---

Items

• Count of Members in the Cluster
• Count of active Members
• Got all expected quorum votes
• Has quorum
• Resource Discovery
  • Resource is active on any node
  • Resource is active on the current node

Triggers

• Not all Members online (avg)
• No quorum (high)
• Missing quorum votes (avg)
• Resource Discovery
  • Resource offline (inactive on all nodes) (disaster)