I can SSH into my Oracle free VPS but after a while I find that the console is no longer working, and when I try to reconnect it doesn’t respond. I can still ping it, but even logging in to the website and trying to “Launch Cloud Shell Connection” doesn’t work, so it seems to be locking up completely after I SSH in, and I have to reboot it.
I’m only trying to use it as a reverse proxy with caddy, but it obviously needs to be reliable for that. There’s a few errors in the boot log, which I’ve pasted here, but I’m not sure if they’re important.
Oracle boot log - Pastebin.com
I’ve enabled Cockpit and the Services tab under that shows that kdump and SETroubleshoot failed to start. Should I just delete the instance and try creating a new one?
I followed this guide to setup the required rules for the VPS instance, and to create the iptables rules in the VPS, and to create the IPV4 and IPV6 A and AAA records on Cloudflare.
How to Host a Website in Oracle Cloud Free Tier (yoursunny.com)
then I followed Steps 2 and 3 in this more recent guide to create the caddy files
Install & Configure Caddy Web Server with PHP on Oracle Linux 8 (atlantic.net)
but even before the VPS stops working the caddy server doesn’t serve anything when I try to access it in my browser.
This is not a question for /r/selfhosted. This is a configuration issue — that I am inclined to ascribe to installing too much unnecessary software — and would suggest directing to a more generalist tech support forum.
Ohh. You’re on that free server. I thought you had meant the big boy - 24gb memory, 4core ARM cpu free server. Yea I’ve never had a problem getting the little one deployed, probably could ignore my suggestion in this case.
Just as fyi however, tutorial may have Debian in the title, but it’s on using NetBoot.xyz which can be used to install any one of dozens of distro that they maintain and offer for net installation.
Do not delete the instance, being that it’s the free tier, you could be waiting weeks or months for another one to be made available for you. If you do end up needing to reinstall the OS, do it “in-place” on your current instance.
There are a handful of ways you could accomplish this, this one being my favorite and the one I always go to.
Take a look at RAM and CPU usage. Those VMs are extremely useful but have limits: swapspace is almost mandatory, lightweight custom images like debian recommendable, and CPU load should be kept low.
On the last point consider that you have just 1/8 of oCPU guaranteed, indeed the VM has two CPUs but these are for short time bursts. If CPU time is abused the hypervisor will severely throttle the instance possibly locking it (keyword: steal time).
With proper configuration there shouldn’t be any problems with running reverse proxies. I have caddy running on arch and SWAG running on debian on those VMs alongside with many other containers without issues.
I had the same issue, create a swap file and use it
2G should be enough
