Skip to main content
Last updated: 3 Jun 2021

New instances fail to provision

Sometimes, it may happen that a new instance is not provisioned correctly due to unattended reboot rebooting the instance before all the init scripts have time to run.

You may diagnose this issue by looking at the system logs in AWS console (select the failed instance in EC2 console -> Actions -> Monitor and troubleshoot -> Get system log) and seeing whether all init scripts have been run.

This issue is more severe in Staging due to unattended reboot checking every 5 minutes at all hours of the day (see here) rather than between 0-5 for Integration and Production.

You could temporarily disable unattended reboot by adding this to the relevant hiera: govuk_unattended_reboot::enabled: false, deploy the branch of Puppet and create new instances again.