A new update each day?

I used openSUSE Tumbleweed on a webserver for years and got used to daily rebooting every OS (if no updates that day then it gets rebooted just to clear up any potential odd firmware stuff, memory leaks, or rogue longstanding connections); on a low-end VPS it'd be down maybe 15 secs (my laptop with GNOME boots faster so better server specs could probably have <10s reboots)
As scottro pointed out, sometimes rebooting is not an option. I have administrated SuSe enterprise systems that emulated mainframes and you can't really tell someone that payroll is unexpectedly going down for a reboot unless the COBOL code was made to allow that feature initially. I have also, administrated simple web servers that were rebooted all the time. This is a edge case in the market that FreeBSD has traditionally been in.
 
My personal workflow. I wait until things stabelise, then i build everything from source & install. When done i reboot.
Between times sometimes 3 months. Eg this is personal. I take quarterly ports,not latest. But i wait two months.
Why ? sometimes a very small openssh patch takes 1000 packages to renew with it. poudriere dont takes risks.
[Until now never been hacked, i dont run a server with a public IP, thats a very different workflow]
PS: suppose i would run a server, how do i fix CVE's without down time ...
------------> Rolling Updates
 
I don't understand why some of you think you must reboot for FreeBSD updates. I never reboot just because it asks me to.
Sorry for my ignorance. Do you ignore it or you manually update that part of the system via the patch and rebuild/install and run the new code, if possible?
 
Sorry for my ignorance. Do you ignore it or you manually update that part of the system via the patch and rebuild/install and run the new code, if possible?

When building from source no part even makes a suggestion about rebooting.

On releases I usually run auto-updates using freebsd-update from crontab. I reboot when I know there was a security bug in the kernel, which I learn from the mailing list announcements. Very low tech.

In both cases I need to reboot when somebody fiddled with the kernel ABI, but that is rare. Most "enforced" updates are to userland to make newest pkg binaries work. The kernel is actually pretty flexible wrt running a different userland. But userland and packages are not. A reboot is not required for the latter kind of update.
 
Well, I didn't read all the detail of this kernel fix. When I see that the running kernel is not the one that is installed, I reboot. And I reboot not only to be protected (yes, this time it was useless), but also to test my server. And it's somewhat long: VMs, jails, pf, other services...

In theory, no reboot needed for userland changes, but that's theory. In practice, you can have some surprises because what are running in RAM aren't updated. So, I reboot anyway and test again the system to know if this update has affected something important for me.
 
freebsd-version -kru not all showing the same value bothers some people. We've had lots of threads when an update was only userland and "how come my kernel isn't updated?" "Because this was only a userland update"
Some people have greater tolerance for a mismatch between the "k" (installed kernel for next reboot) and "r" (currently running kernel).
I don't advocate for a specific plan, but yes it does get annoying when updates need reboots and you do them daily.
 
I just upgraded my stations and my server to p7. And then, new patches come. And they come in a way I'm obliged to reboot. Never seen that in 6 years at least.
Sorry
This isn't exactly on topic, but it's close enough.
We May Be Living Through the Most Consequential Hundred Days in Cyber History, and Almost Nobody Has Noticed
The first four months of 2026 have produced a sequence of cyber incidents that, if any one of them had landed in 2014 or 2017, would have dominated a news cycle for a week.
 
In theory, no reboot needed for userland changes, but that's theory. In practice, you can have some surprises because what are running in RAM aren't updated. So, I reboot anyway and test again the system to know if this update has affected something important for me.

There isn't that much in the way of daemons running from the base system. Most are ports/packages. You have to recognize by hand when some have to be restarted.

This is, BTW, a big advantage over most Linuxens. You cannot restart PID 1. In FreeBSD PID 1 is init, a very minimal program that I don't even recall when it had the last security problem. In most Linuxens PID 1 is systemd, a huge thing with constant security issues. Any of those always demand a reboot.
 
I have to admit that the main difference between AIs and humans is that these latter complain.
I just proved I'm a human being. :)

Nevertheless, don't overlook that the world is speeding up in a way that most humans will be useless soon. I don't know how we will stand that.
 
I have to admit that the main difference between AIs and humans is that these latter complain.
I just proved I'm a human being. :)

Nevertheless, don't overlook that the world is speeding up in a way that most humans will be useless soon. I don't know how we will stand that.
Electricity will have to be cut off for a long time at the global level Disconnect UPS batteries In isolation, then start the servers hosting the AI, eradicate the code Erase all computers in the world Reinstall OS And connect the internet in txt mail, html basic mode
Impossible
Too late
AI is a huge multi ver
 
Due to my FreeBSD systems all being virtual machines, I like to either put my host machine in S3 suspend while I'm asleep at night. I opt do shutdown all of my VMs, including those that aren't FreeBSD, so that my Windows 11 Pro machine that they're hosted on either gets suspended or completely shutdown when i go to bed at night. I don't have to worry about 99.999 percent uptime or the need to deal with patching a live system.
 
Back
Top