I am giving this a try now
On the subject of maintaining servers, and I can’t remember if this has been suggested before, would it make any sense for Trellis to add/keep an additional playbook that runs updates and reboots?
./bin/update.sh production example.com?
That would make it easy to incorporate into a regular WordPress maintenance schedule, which I already do approximately monthly for my clients.
It’s not uncommon for servers to be online for months or possibly a year+ without a reboot. However in today’s idea of servers being so expendable and replaceable, I think that requirement is rather moot (unless you get to work on a legacy app like I do). I would think that unless there is a security patch that’s required, or something goes wrong, these servers that Trellis provisions are more or less set and forget, and if something happens, provision another one.
I guess I just always see my servers complaining of pending updates, or “needing to be rebooted” when I ssh to them.
I haven’t actually had a problem though.
I’ve had to set myself a task to routinely check-in on my servers. Tempted to roll all my simpler Roots installs into one server, when I get the time.
Separate servers for each site saves you from the typical “bad neighbor” issues shared hosts get. Email blacklists, etc.
Yeah I’m still weighing up the pros/cons.
For emails in particular I only use SendGrid anyway so my Trellis servers never send anything.