Due to maintenance to the electricity infrastructure in our street, I had to shut down our home server.
It had been up and running for nigh on 192 days, which is a record in our house as far as i remember.
It’s an RPI4, running Raspberry Pi OS, bookworm, serving nextcloud and a VPN, and it’s our ‘NAS’.
My power fails more often than that 
I do remember the days when I would schedule a reboot every 365 days for servers.
That looks a little longer than the 2 weeks my ubuntu server runs before a kernel update 
.
A RPi 4b performs excellently for server use. I have two RPi 4b with 4 GB RAM that run a EnOS base install, so basically Archlinux ARM. They are on a UPS and just do their thing with no problems.
One is a Web server running nginx, noip, and Let’s encrypt and the other is my NAS.
I update them on a monthly basis.
Pudge
Best I ever had was a radio station broadcasting server, with 470-something days uptime. Running Ubuntu 14.04 I believe. I simply forgot to update it, because everything was running so smoothly… 
When you forget about the server, then you know you’ve got a stable system running indeed.
Indeed! the same experience here. Rock solid.
Fact: i started this setup with an RPI3, running the OS on an SSD. Years later i bought the RPI4, made sure it would boot from an external drive, turned off the 3, disconnected the SSD, connected it to the 4 and turned it on.
It booted without problem.
2 weeks does seem a bit short indeed.
Does remind me of an Microsoft knowledge base article which stated that a Windows NT? Server needed a reboot every 76 days (don’t remember the actual number) because running it longer would result in an overflow of some register/counter…
I looked it up [1].
It was W95 and W98 apparently that crashed after 49.7 days.
[1] https://www.quora.com/Why-did-Windows-95-and-98-crash-after-49-7-days-of-uptime
And it did not have a kernel update in 192 days?
Weekly updates with reboot here. If something has an uptime over a week it is usually an oversight or automation failed. Long gone are the days where uptime was a meaningful metric. Uptime = unpatched vulnerabilities.
A valid question, but apparently not.
After an update i always check if ‘/var/run/reboot-required’ exists to check if a reboot is needed, and it didn’t.
My home server is also running debian bookworm. Over the last 192 days, back until 27. November, it received linux-image-6.1.0-28 up to current linux-image-6.1.0-37. That is 9 updates. All of them needed a reboot.
The period for the 9 updates i counted started on 15. january. That is 26 weeks for 9 updates. Approx 3 weeks for a new update
Then I’ll have to have a look at my update and reboot regime. Thanks for the info. 
