So, you've configured two Loadbalancer.org appliances as a clustered pair for high availability, the pair has been working fine but then one day you receive an email alert that one of the pair has failed for some reason, for example a disk failure. The remaining unit continues to provide load balanced services without issue, but now you want to repair the failed unit and re-create the clustered pair without any disruption to running services.
From v8.2 this now possible in 3 easy steps.
STEP 1 - Repair or replace the failed unit & power it up
After power up you'll see the following screen on the console:
STEP 2 - Run through the setup wizard
As mentioned in the first screen, login as user 'setup' with password 'setup'
Specify the IP address, default gateway & DNS server(s) of the failed appliance:
Once configured, a summary of these settings is displayed:
You'll then be asked if you're recovering from node failure:
**** At this point Type 'Y' and press ENTER ****
You'll now be prompted to enter information about the remaining appliance:
Once the details are entered, the recovery process continues:
Once complete, the following message is displayed:
STEP 3 - Reboot the new appliance
Finally, as mentioned on the screen reboot the new appliance to complete the recovery process.
You'll now have a fully re-synchronized HA pair.
NOTE: the same procedure applies whether it's the master that's failed or the slave.