Manage server packages

If maintenance.patching is not set to manual in a target’s Pillar file, then the unattended-upgrades package is installed and configured.

The example commands below will run on all servers. To run on specific servers, replace '*' with either a glob pattern, like 'cove-*', or with a comma-separated list using the -L flag, like -L kingfisher-main,kingfisher-replica.

As with other deployment tasks, do the setup tasks before (and the cleanup tasks after) the steps below.

1. List upgrades

./run.py '*' pkg.list_upgrades

Consider whether any upgrades are backwards-incompatible or have post-installation steps.

2. Upgrade packages

./run.py '*' pkg.upgrade dist_upgrade=True

Monitor the output for relevant messages.

3. Reboot

1. Find the servers that need to be rebooted:

   ./run.py '*' file.file_exists /var/run/reboot-required
   

2. Reboot the servers that need to be rebooted. For example:

   ./run.py -L server-one,server-two system.reboot
   

   Sometimes, this command hangs, waiting for a response from a server that is already shutting down. Simply wait 30 seconds and stop the command.

3. Check the servers have rebooted without issue:

   ./run.py '*' file.file_exists /var/run/reboot-required
   

   All servers should respond with False.