This article describes steps to setup split appliance architecture for Opvizor Metrics and Logs.

Purpose

You may consider split appliance setup when monitoring larger VMware vSphere environments which may degrade the performance of the user experience when navigating through the dashboards. In split appliance setup the database layer is separated from processing layer therefore allowing better performance at higher workloads.

Setup

The initial appliance will be referred to as head VM

1. shutdown head VM
2. clone head VM - this clone will be the new db VM
3. startup db VM , change name and IP
4. after reboot stop perfanalyzer services on db VM. Run as root: service perfanalyzer stop
5. startup head VM
6. create ssh key on head VM as root

• sudo su -
• mkdir ~/.ssh/PA6 && ssh-keygen -N "" -f ~/.ssh/PA6/opvsingle

7. add public key to authorized keys on db VM
   • connect as admin
   • add the key to authorized_keys
8. test and accept hosts public key on head VM as root slogin -i .ssh/PA6/opvsingle root@<new_ip>
9. after successful test, exit the shell to return to the head VM
10. edit /data/app/perfanalyzer_hosts and add section [pa6_gsn] with an ip pointing to the new appliance
11. edit /data/app/graphite.inc.sh to point to the clone - e.g.

GRAPHITE_HOST=<name_or_ip_of_clone>
GRAPHITE_PORT=2003
GRAPHITE_SYMLINKPORT=5000
GRAPHITEWEB_HOST=<name_or_ip_of_clone>
GRAPHITEWEB_PORT=8080
START_LOCAL_GRAPHITE=0

12. Converting the clone to a db-vm

• deploy again the patch to the head-vm web ui, so it will be deployed on the db VM

13. if needed, cleanup unused data, compact disks

Patching

Patching is seamless to the database appliance and applied in the same patching process when patch is uploaded to the head appliance.