I experienced some VPS issues and I hope by sharing I can prevent some other sys admin from going through the same frustrations as I have recently. Here’s a graph of my recent server uptime……its not pretty….
My server was initially configured with Ubuntu to run an nginx web server serving basic blog pages via Kirby. Having created the server and then seen close to 2 months steady uptime, I was surprised to start seeing random reboots occur every few days. No changes had been made to the server in this time except an occasional blog upload so nothing immediately sprang to mind as the potential source of the problems. Nothing in the various system logs shed any further light on the matter and everything looked fine until out of nowhere a series of boot up records were logged.
Having pulled out what little hair I had remaining and cursed my lack of ‘nix experience, paranoia started to take hold and I began to suspect I had been hacked and some April fools virus had been injected into my server. It was quite a relief when I received an email from Ramnode acknowledging they had seen issues on my host hardware and confirmation that to avoid any further problems they would swap the hardware out. Sadly my initial hope and relief faded as the random reboots continued to occur.
Another week or so of experimentation, various software upgrades and researching into Ubuntu’s underlying systems to hopefully enable me to identify how this situation had occurred and therefore how to resolve and prevent a reoccurrence was unfruitful. In desperation I restored my last known good backup but this didn’t resolve the matter either. Just as I was about to give up and reconfigure the server from scratch a mail from Ramnode acknowledged they had tracked the issue to another VPS running on my node that was likely the cause of these reboots. That VPS has now been removed from their systems so fingers crossed this is an end to the matter.
For me it was a useful reminder that a VPS isn’t a dedicated physical server, its a VM running on a shared box somewhere and that other VMs can affect its stability and/or performance. For anyone else suffering random reboots with no sign of reason in the logs, it may be worth mentioning it to your host just in case. I’d like to thank Ramnode for handling a difficult situation as well as could be expected.