I have a cron job which restarts my computer every day (Linux Mint Debian edition). I like waking up to a "fresh" computer, and since I know that it will restart every day - this is a "forcing function" to (1) be diligent about saving things that are important, (2) treating browser tabs, random notes, etc as ephemeral, and bookmarking the important stuff.
I used to work at an office where we pair-programmed with clients all day (Pivotal Labs), and most of our computers had some sort of "automatically restart / restore from a known-good image". I liked this, as it resulted in less cruft over time, and some intentionality about what getting a computer into a productive state means. It also got me thinking of using automatic routines to accomplish goals, and not being so attached to my open tabs, etc. Let it gooo....
To be more specific about this - for those wanting to get into blogging/publishing, this could mean auto-opening the website project folder using VSCodium upon user login, so its ready to go for the morning coffee. More half the time I just close it - but as a "default", it makes it easy for me to do the thing I want to do.
sudo crontab -e -u root
#-----------------------------
# RESTART COMPUTER DAILY
#-----------------------------
00 04 * * * /usr/sbin/shutdown -r +5 "Rebooting in 5 minutes. Run 'shutdown -c' to cancel"
If your computer has any kind of secure boot / boot chain attestation stuff going on, like all phones do now, rebooting clears out pretty much all malware. Software bugs can put the system in to a bad state which can be abused by malware, but none of it can persist through a reboot since everything is checked and measured from scratch during the boot.
> any kind of secure boot / boot chain attestation [...] rebooting clears out pretty much all malware
IMHO "pretty much" understates the risk.
Malware can easily install itself as a system service, timer unit, XDG autostart or your shell profile among other places. I'll be the first to admit I never check all these places regularly.
The only thing that should be putting minds at ease is regular OS installs from fresh images.
Resist the temptation to do an "in place" upgrade and go with a clean ISO each time your distro comes out with a new major version.
It's a pain but thanks to configuration management or even shell scripts it's manageable for me now.
Admittedly six months is probably too long as well but at least it stops something lurking on a server or my desktop for years.
If you go with an immutable OS like Fedora Silverblue or Bazzite, every single update is a clean ISO install. And then in your user only install things through Flatpak while checking there aren't dangerous permissions enabled you will be much better off than a normal linux setup were mutations can be placed all over the place in a way that is impossible to audit or secure.
This feels rather overstated. Your firmware check isn't going to tell you if a repo package installed/updated by your package manager got compromised, which is apparently increasingly common.
It's more relevant on ios/Android where you can't through normal means install malware at the system level. Linux is moving in this direction with immutable OSs and better secure boot support but there isn't an easy way to have a secure linux desktop setup currently.
Exploits break out of the sandbox for higher system access. Rebooting flushes them out because the exploits have abused some memory bug to load custom code in the system, but this can not persist a reboot since everything gets cyrptographically verified on boot.
This is why things like the Nintendo Switch hacks and iphone jailbreaks have to be reloaded every time the system reboots, while they could modify the system files to try to persist the malicious code, it would simply leave the device unbootable since the boot chain attestation process would refuse to boot when the checksum doesn't match. The malware can only come back when the initial bug has been triggered again, through receiving a malicious text message attachment for example.
> Software bugs can put the system in to a bad state which can be abused by malware, but none of it can persist through a reboot since everything is checked and measured from scratch during the boot.
This does not account for supply chain attacks nor trojan attacks which seek consensual installation. A reboot will not eradicate these threat vectors.
It relies on your software to be properly sandboxed and permissioned.
What we have seen is state funded hacking groups can often exploit bugs which break out of sandboxes and embed themselves in the system. But the malicious code only exists in memory, it can’t persist between reboots because the next boot would fail the checksum and signature checks.
Which is why security focused OSs will periodically reboot themselves to flush out possible malware and why jailbreaks and hacks for phones and game consoles usually have to be reinitiated after every boot.
MacOS has a few measures including SIP and boot chain security to implement this, though it’s currently hard to set up a properly secure Linux desktop.
> It relies on your software to be properly sandboxed and permissioned.
And organizations rely on their employees not falling for phishing email. How well has that worked?
> What we have seen is state funded hacking groups can often exploit bugs which break out of sandboxes and embed themselves in the system. But the malicious code only exists in memory, it can’t persist between reboots because the next boot would fail the checksum and signature checks.
"Malicious code existing only in memory" is patently false by not addressing supply chain attacks previously mentioned. To wit, the npm left-pad attack[0]. See also the thousands of PHP CVEs.
This is true, which is why it's extremely hard to create a secure linux desktop today. But things are moving in the right direction, if you use Wayland + Flatpak on a immutable OS with secureboot set up you are in a much better position than a standard distro.
Most methods of persisting state have absolutely nothing whatosoever with kernel or modules which are a powerful way to hide malicious activity but far from the only way to persist state including malicious state. The only sane response to being infected with malicious software is to reinstall everything starting with the OS and scan user files AFTER reinstalling.
It is possible even plausible that you can disinfect many sorts of malware from inside a compromised machine but it is inherently less trustworthy and installing from scratch barely takes any more time.
> about what getting a computer into a productive state means
Why would you do that? Is your OS not running stable? Do you not have tight control over the software that is running? Why are you unable to keep it in a productive state?
Greetings from a Computer that only restarts for updates every few days, next to an airgapped machine that has been running for years.
One of our customers has a daily automated restart for their servers. This is because the IT people have long lost track of what the system actually does, and feel they have no time to figure it out.
It is the bane of my existence when it comes to the predictability of uptime, not to mention long-term processes getting interrupted over the day boundary.
I remember when Linux users were practically obsessive about uptime and restarting felt like a sign of failure. This was at a time when Windows seemingly needed to restart once or twice a day, at least.
These days I like to turn my work Mac off at the end of the week just so I feel a literal sense of closure. It's not really the applications minimizing and running in the background; it's ME.
Yes, I remember feeling pride in the stability of my systems when I saw a large uptime. I had a server that had 1000 days of uptime, once. Now when I see a large uptime, I'm terrified of what security patches the kernel may be missing!
I remember installing some new computers for a small shop around y2k, and their NT4 server was acting up a bit when I was adding the new users to the domain controller.
I opened Task Manager to see if any processes was running wild. Imagine my surprise when I saw it had well over 1100 days of uptime!
I still remember the days of servers as pets, rather than cattle, and I was harping about server uptime. A wizened server admin piped in and said he rebooted his servers once a week. Said, if you do it any less frequently, then the odds of catching an error causing change while the person who made said change (possibly himself) is still around and can remember what they did go down precipitously. So, to avoid headaches and potential downtime when it mattered, he would just take servers out of rotation and reboot them, and make sure they came back online.
So true. We have one older, rather large machine in a data center that's been up for.. (checks uptime): 963 days. It has IPMI but at some point something stopped working and now we have to physically go to the data center to restart it. And since we use it every day we can't really afford to lose access to it.
Live Kernel Patching has been around for about 20 years[-1] now.
Red Hat Enterprise Linux[1] and Oracle (Enterprise Linux) Unbreakable Linux[2] both use it as a selling point.
This feature is still a bit ad hoc because, in most setups, rebooting a system isn't a huge burden and is much simpler than using boutique commands to live-patch it.
Live patching exists but it tends to cap out at around a year of updates for any particular kernel version. It's not getting you anywhere near 1000 days.
I worked in a place with a lot of Solaris servers with years-long uptime. It would be my job to patch them. Having no idea what config changes that may have happened over the last 3 years which would take effect on boot was always terrifying.
There's also `systemctl soft-reboot` which initiates a userspace-only reboot, which quickly restarts the system without going through the full hardware and kernel initialization process.
I'm glad to see this. Almost 18 years ago I implemented a similar kexec device+memory preservation for a storage vendor. It was done on a Linux kernel of that day, and it had had a memory reservation and handoff protocol between the two kernels to keep some specific PCI device alive, allowing for state restoration at the application side. I'm proud of the fact that the kernel replacement was just under 1 second in execution (after init process optimization) and the whole kernel+app was less than 10 seconds.
Currently serving: vm's, dns, email, mx-relay, and multiple shoutcast radio relays 24/7 and some other miscellaneous stuff. Colocation is fun, do I win?
5years; I'm 37 now, I was 32. Life seemed easier then.
I had a box set up as NAT (running amazon linux) when we moved from a local datacenter to AWS in 2012. Shut it down last year. It had not been rebooted. Should have grabbed a screen capture of the uptime. Part of me wanted to leave it to reach 5000 days....
I remember when I was asked to replace a core router with a more powerful model. The uptime of the Cisco router was ten years - and it was ten years after the datacenter went into service.
I do still enjoy the odd >30 day uptime on my PC. Usually only reboot when a new kernel version is cut.
I used to reboot into every kernel patch but often I leave .0 running for a very long time now. They seem stable and the kernel moves fast enough nowadays there's often another .0 right around the corner. There might be exploits but they're not a valid threat model for my little desktop.
If something smaller like Mesa updates, I can reload everything simply by logging out/back in, no need for a full reboot/LUKS unlock.
In the mid-2000s I ran a. Fleet of RedHat servers that hosted millions of domains. I had boxes in that fleet that were up for over a year. Netcraft confirmed it!
Microsoft literally bought these 6 or 7 servers to migrate to IIS so they could “beat” Apache. It took more than double the servers, but after I did the initial work it was moved to a different team and I don’t know how the uptime compared.
My machine was rebooted this week due to a power outage. I don't recall the last time prior to that. It generally goes weeks if not months without a reboot.
Just close them. You're never going to read them. If you really think there's something you need, export the browser state to an archive file, then delete in 10 years after you've never consulted it once.
(Disclaimer: I'm aware that there may be valid reasons for this workflow, but in most cases it's just digital hoarding and the above advice is sorely needed. If you really need 1800 tabs, you know who you are and you can safely ignore me.)
Brave has a function to bookmark all open tabs. I have used that from time to time. Or just made a temp file and written all the URLs to it, in case I want to open a few/dozen of them after rebooting.
I think all browsers have that now. I use it 3-4 times a year with Safari on both macOS and iOS. That way I feel comfortable closing them all without worrying about “losing” anything.
I’ve been doing this, or something similar, for at least 15 years now. Dozens of mass-bookmark “folders”. I’ve never once looked at the bookmarks I made, not a single time. I even have old bookmark files archived here and there, from machines and browsers for which the bookmarks weren’t auto-backed-up to “cloud” storage like my Safari stuff is. As soon as I can get a local llm running that’s up to the task, I’ll probably have it build some kind of table with categorization out of all of them, then have it edit out entire categories of crap until I have something I might actually scan over to recover a few interesting tidbits. Finally make some use of them, now that tools can take enough of the pain out to make that a less-daunting prospect.
The irony is that now, in the LLM age, it wouldn't be that hard to feed a few thousand browser tabs and/or bookmarks into a vector database or whatever and then if some fuzzy memory tickles your brain about something you saw a year ago, you can just query it.
It's the same logic as keeping RAW digital photos. Lightroom is already gaining "find the keepers" AI features. Sooner or later it'll be possible to feed a bunch of burst shots into an AI that just weren't worth the trouble to manually sort through. The AI can do the drudge-work of digging up any gems in the rough...
how do you see how many tabs you have open, an extension? i have tab session manager and it shows i have 80 tabs. about 60 of those are ephemeral, and the other 20 i'd have open on a new browser anyhow (email x3, goog cal, caldav cal, nextcloud files, router, local and remote proxmox, navidrome, the documentation server, etc) everything else is superfluous. although i'd probably be a bit sad if i lost all my tabs right now; hence tab session manager.
We had a vp of engineering at a retail place who came from windows shop and wanted to restart the servers every night of the holiday season. It took some doing but we finally convinced him “this is Linux we don’t have to that!”
At Javasoft there were Solaris test machines that had been up for 2+ years, and we had to reboot the windows test machines multiple times a day. It felt really good to leave a large queue of work at the end of the day on the Solaris/Sparc machines, knowing it would be done the next morning.
> I remember when Linux users were practically obsessive about uptime and restarting felt like a sign of failure
I remember it too, like it was yesterday. Wait - it was yesterday.
I see people rebooting Linux boxes to cargo cult trying to fix all kinds of issues and I’m like - rebooting is not a solution. This is not Microsoft Windows.
It was a differentiator when distro updates where sparser, and in start comparison with Windows at the time which couldn't stay up for more than a couple of days without crashing (particularly the XP era).
As a rule, if you don't reboot your servers while you are near them watching for problems, they will reboot by themselves at 3am in the first day you get sick or are traveling.
I don't feel like keeping my tower pc running more than necessary. When I don't use it for hours, I usually turn it off. At least when I go to bed I turn it off. Next day I turn it on again. Like a normal person. I don't see a need to save 1 minute or so, by keeping it running.
I'm astonished I had to wade through so many other replies to find this one. I had never considered I might be doing things abnormally when I turned my machine off at the end of the day.
I've been keeping my desktop machines running 24/7 for ~25 years. I only reboot when necessary.
I remember having around 280 days of uptime on Windows 7 when it went end of life. Having a UPS helps a lot to protect against short power outages or blips.
Nowadays I run Arch Linux, it's been 12 days since a reboot. Not trying to break records, I reboot to apply kernel updates when it's convenient. Since I use tmux and have terminal heavy workflows it takes 1 command and a few seconds to resurrect all of my sessions to get back to where I was at before.
I'm with the UNIX old timers on this one; cleanliness is overrated. Life is messy, creation is messy. More uptime means more mess. Uptime is longevity; there are few things more valuable in life than longevity.
Agreed. I put stuff in ~/tmp only to create ~/tmp2 for even more temporary stuff that I don't delete, and so on. I clean up Desktop though as it is in my face.
You reminded me of this story. Windows, not *nix, though.
Years ago my cube-mate was called to support a desktop user who has having problems with their machine that couldn't be fixed remotely. He realized the user's HDD was full, so the first thing he did was clear the user's trash bin. Mysteriously the machine came up with a ton of free space.
The user seemed happy their machine was up. Nor long after, the help-desk called my cube-mate asking what they did to the user's machine; The user couldn't find any of their files that they had organized in the trash bin.
This is actually useful for smartphones. Sometimes smartphone malware is capable of infecting a device but not persisting, so reboots clean it back up.
I remember something where on maybe mainstream Androids you'd have to fully turn off in order to turn it into BFU, but maybe it does just by restarting now.
I shut down my desktop every night. No reason to keep it on. I can save my work, all my tabs persist, it’s fine.
My main complaint is that my PC seems to have crazy clock drift? And windows just doesn’t seem to care?
It doesn’t actually cause problems but I will check and see that it’s off by >5 seconds and that just shouldn’t ever happen. My phone is always accurate to <0.05s and same in Linux. Idk why windows isn’t.
Tangent about clock drift: my phone (Pixel 9) drifts by about 5-10 minutes a day when in airplane mode. This doesn't sound like it would be a huge deal, but I do a lot of wilderness backpacking and very long flights where I'm keeping it in airplane mode for 36+ hours to save battery, and then at the end I need an accurate time to make the next flight or meet a ride.
Not really sure what to do about it because there is no official support for Pixels in this country.
I do typically, but I haven't always been the best at managing watch charge. It's a coros fitness watch and I've killed it on some hikes by using the GPS too much.
Not OP, but the sleep and hibernate states on my desktop are pretty finicky and can't truly be trusted. I'd rather know everything shut down properly than risk it, especially since the thing boots in seconds anyway, and I have to input the encryption password anyway.
doesn't clear up your ram. One big security reason for regular reboots is that you simply get rid of any potential non-persistent crap on your machine. Also performance obviously, with a full shutdown you get back to a known, clean state.
Worth noting on Windows the restart function only does that if you hold Shift or have Fast Startup disabled.
I don't have any data to prove it but I think Mac users don't bother "cleaning up" after they are done with their computers.
I think windows and Linux users usually shut down their laptops when they are done.
I believe this is because of how Mac is designed, nothing really closes. You close an app and it's just "minimized". Same behavior as with the lid, you close the lid and it suspends.
If I recall correctly, at some point, this also affected the iPhone, you were not able to "fully close" apps and they decided to add a screen so you could swipe and "close" the app (some run in the background, same as android)
I think your model of open/closed is incomplete and thus misleading. There are more states to a process than "active" and "inactive," and it is not optimal for the system to simply move processes between those two gross states. The obvious example is non-foreground apps during multitasking. A less obvious example is an app during a background refresh.
"Fully closing" a process is not necessarily cleaner than letting the system allocate intelligently, despite what one's puritanical upbringing might make them believe. (Consider how artists often need a messy space to optimally hold their processes.)
I think my point is that minimizing the process doesn't free the ram it's using. Closing it, even if it stays running in the background, should free up resources
I still remember reaching into my backpack to retrieve my Macbook in sleep mode only to find it hot to the touch, having woken up for some goddamn stupid reason. We had "proper" sleep decades ago: we called it "hibernate". Modern "smart sleep" is a technological regression, and Apple is as big an offender as any.
So I don't have to spend 10-15 minutes saving all my open files (including the scratchpad ones I haven't named) and then later reopening all my applications, recreating all of my tmux windows and panes, setting up my vim splits and tabs again, and starting all of my stopped Docker containers?
I have a feeling someone is going to jump in with a "solution" to each of these forgetting that lots of work is ephemeral/transient and undeserving of even the slightest bit of automation or pre-configuration.
It takes ages to cold boot. My desktop is ok-ish, but it doens't matter since I only use it occasionally.
My old HP laptop had a slow-ass BIOS that I was convinced had some kind of bug. I replaced it with a brand spanking new thinkpad 2-3 months ago. Guess what? The freaking BIOS is EVEN SLOWER somehow!
They all wake up instantly from sleep.
I therefore only shut them down when I know they'll be unplugged for a while, because for some reason the HP eats through the battery even when off. If suspended, the battery will be out of juice in like two days. Haven't tried any of this with the Lenovo yet.
Suspend used to work great, but since MS figured they should copy Apple half-assedly, suspend is borken. And I have really no idea what we've gained in exchange.
Now that I have an nvme SSD in all my computers and boot times are roughly 10-15 seconds or so unless something has gone wrong, the advantage of sleeping is somewhat mitigated.
Back in the spinning rust era, though, a good unsuspend could be something like 50 times faster to get to a running computer. Possibly more, depending on what your OS needed to start up.
It is still more convenient to have my previous environment most of the time, and still faster to unsuspend than boot, but it isn't as much of an advantage as it used to be, no.
On Windows and Linux it is better to just turn it off because sleep is broken on most hardware. But on Mac there is next to no advantage to doing that. MacOS sleep is also more of a hybrid between sleep and hibernate where it will after enough time offload the ram to the SSD and fully power down so the machine can be in "sleep" without any power drain.
I think a large part is also how long it takes to restart a Mac. Every so often a coworker has to restart and I could probably restart my Linux (or even Windows) laptop 3 times before they're back on.
Kind of reminds me of how slow Windows computers used to boot back in the Vista and 7 era.
I'd argue the opposite. My mac wakes up from sleep when I open the lid, and is functional in seconds with the fingerprint. Meanwhile sleep on windows is a complete dumpster truck and can result in any of "works fine", "a bunch of apps have got stuck" or "your battery drained in your backpack".
Also, my Win11 desktop is "fast" to get from POST (which takes > 2 minutes to do RAM check on every boot with 192GB RAM) to the login screen, but it's a good few minutes from log in before windows has started all the background stuff and it's actually functional.
I'm glad it works for you, but I have witnessed several coworkers restart their macbooks (some M1, some M2, possibly M3) and I don't think I've ever seen a reboot shorter than about two minutes.
At one instance, I rolled over to a coworker who has just rebooted theirs and had a whole 5+ minute conversation.
I don't know what current Windows 11 does, but my typical Linux Mint is single-digit seconds, and my t490 is 7 years old already. Windows 10 is a bit slower but still single second territory. With more current hardware we are talking about 2 seconds plus whatever BIOS needs, see sibling comments. Maybe you have corporate AV on the Windows PC, or funny drivers?
On linux a lot of the time you kinda have to shut down because it either won't go to sleep ("failed to freeze" some process or another) or it will hang or lock up during/after wake - especially if it has nvidia with its binary blobs, but i've had problems with (in kernel, open source) network card drivers too. Since I first had problems like this with linux 2 decades ago, it seems like they will never be able to figure it out.
Some of this seems to be getting worse with the move to wayland. There is no design concept like windows D3D device reset or lost device, or display timeout/recovery, so drivers are apparently charged with perfectly remembering all the state that compositors create, through all the sleep states, with predictably bad results.
I have only ever had one sleep over multiple Linux laptops over nearly 20 years, my own and other people's, including some tricky hardware (an MS surface). I only ever use sleep/suspend to RAM though. I avoid Nvidia though.
That's a perfectly acceptable practice for apps that only ever have one window. For document-based apps or apps that have multiple windows, usually the convention on Mac is for closing the last window to not also quit the app. But of course there are exceptions in both directions, it's not a hard and fast rule.
Linux users do not clean up AFAIK, and the article is by a Mac user who does clean up. Why would Linux and Windows users not use sleep? AFAIK everyone does.
I avoid sleep on my desktop because it often wakes up in a weird broken state, particularly with audio glitching. I avoid sleep on my Steamdeck because the battery slowly drains in sleep.
On my Macbook I always use sleep since there are no downsides.
Not having to stop sessions of all kinds of things and restart them for no good reason.
It doesn't require it to stay up, and if things were better at retaining state across restarts I would care less, but it's a nuisance to have to log back into things, and get things back exactly how I left them.
I often have half a dozen projects up on different virtual desktops, and leaving them how they were when I worked on it last makes it easier to get back up to speed.
EDIT: I used to leave screen sessions running on servers instead, as the workaround to having to reboot my local machine. But it's nice not to need to.
Different reasons. Mine is on the table and I use it more like a desktop. It will just idle when I'm not around, because I come and go often. My current uptime shows on Debian 30 days, 49 min.
Although... 30 days is maybe a bit misleading, because I ran some heavy shaders without thinking that triggered the GPU watchdog and forced me out of my session. I think killing all user processes is almost like a reboot, although not according to uptime.
Nothing requires it. But if I turn off my computer, it takes 15s or so to boot up. If I leave it on, it's available for me instantly. The latter is a more pleasant experience, so I gravitate towards it.
I think what you're talking about is how on Mac, an application can stay running without having any windows. Separate from closing all the windows, you usually have to "Quit" the application if you really want it gone. This is useful for a couple reasons:
- You can "Quit" the application without closing all the windows, and then the next time you start the application, your windows can come back.
- You can close all the windows without "Quit"ting the application, and you don't have to wait for the application to load again in order to open a window later.
Additionally, since application lifecycle is managed separately from the open windows, apps can do cool things like saving and restoring the set of open windows through a system restart. Which Windows and Linux still haven't managed. (Maybe Windows can try to restart the processes... I think I saw that becoming an option more recently)
I've never rebooted often in general, even when I daily-drove Windows. Then, it was because it was annoying to get my preferred workspace back after a Windows restart. Now, I daily-drive macOS and I don't often reboot until the machine gets slow/janky because the machine doesn't really need a reboot until then. And I don't hate reboots as much as I would for Windows because macOS is a lot better at session restoration
I used to be in this camp, maybe not 150+ days but with month+ uptimes, but now with Docker I have to restart regularly as I frequently get notes about 'no more disk space' and the only way to reclaim is is to reboot.
My Lenovo Legion gaming laptop running Linux Mint Cinnamon takes ~10s to boot. I hardly ever leave it on, I mean why??
It is silly to use a cron job to do something that only takes 2s for a human to do.
My homelab Debian Netinst servers tho, run non-stop for months or even years without reboot.
Half century ago Linux had a bug where servers had to be rebooted after X hours, that is no longer an issue. Windows however requires a reboot because its pagination memory gets full and the longer you run it, the more the performance goes to shit.
Months to years without a reboot sounds like a pretty big gamble as to whether the hardware still boots, or not. More frequent reboots tend to catch "the hardware has soured" (and how would you know?) issues sooner, like maybe the BIOS decides to come up with the clock set four years into the future and then time sync can fail and then everything fails. The prior example was strictly hypothetical, of course no real system would ever do that.
I reboot one of my machines all the time. It's a Mac Mini running 'Little Snitch', which informs me about all of the network connects my apps make, especially browsers. One cool feature of that program is that one can temporarily grant access to a connection, for instance until the request quits, or one logs out, or one restarts a machine. I'm slowly developing permanent allow / ban lists, but I also use this feature to temporarily utilize some website, aand then reboot to avoid cross-contamination.
The downside is that I often get shunted off into additional authentication workflows, since the prolonged delay caused by my manual approvals triggers some alerts. One of the entertainment ticket buying services is really convinced I'm committing some kind of fraud.
So, in general, I reboot everytime I start using that machine, at least once per day, sometimes more frequently.
Fun fact: in a former life, I worked for a retailer with 1000s of remotely deployed machines and no field-based tech support. One of the OSes we used back in the day had a bug that caused their license authorization service to fail after a certain amount of uptime. We had hundreds of machines that reached that uptime, all on the same day. Suffice it to say, that was not fun.
I boot my computer every time I will use it and shut it down when I am done. No matter whether on Linux, Windows, or Mac. It's a mindset thing: I will sit at the computer, do what I have to do and then shut it down. Just sharing my way of using it.
I remember a big debate "back in the day" about the effects of shutting down your PC at night and the effect on the health of the CPU. The theory at the time was that by shutting down your PC every night, the temperature fluctuations would put more stress on the CPU, shorting its life span. However, the other side said that leaving on your PC all day and night would spend more time running and reduce its life span, so only having it running when you needed it was better. I think a website did a test and found it makes no difference, but I remember it being a big topic around the 2000's (it might have been related to AMD chips at the time running extremely hot)
AMD's first Athlon chips did not run particularily hot. They needed a fan, just like Williamette chips, but actually used less power than the competition. The difference was they had no internal, on-chip diode. If the heatsink and fan fell off, then the chip could cook itself. The boards had a thermal sensor to initiate shutdown if the fan stopped working. But it's debateable if it would detect it quickly enough if the entire cooler fell off.
In that scenario, shutting it off seems better. But it might roast itself on start up anyway.
Is it weird that we don't have a way to update the system without rebooting by this day and age?
I think there are things in Linux like live kernel patching (kexec, ksplice), but why by 2026 is this sort of architecture of live system updates not a common or included feature yet?
People hate updates and having to reboot and have downtime right? Security updates to core systems are more important than ever now too. So why hasn't this problem been tackled I wonder?
Wouldn't it be great if our systems could update without us having to reboot and interrupt our workloads?
Many modern Linux distros encourage restarting too. Atomic distros make restarting to boot into a new image the default, though some have flags to do a "live" update too.
You still have to reboot for things like kernel updates. As far as I'm aware, live patching the kernel is just for security issues.
Not only that, my Arch laptop is not even able to continue functioning after full system update without reboot. It unloads nvidia kernel modules crashing newly opened graphical programs.
Windows does more of this than people notice actually. I was on a system the other day, Server 2016, it had downloaded 2 updates, both system file related, one had installed already and the other one was nagging for a reboot.
Stuff like Windows Defender doesnt even really ask if you want to install the update anymore, it just installs the new virus definitions immediately.
Anyone who is using full disk encryption will be turning off their computer when they're not around. Hibernation is an option if you want to keep your state.
True but in practice it's still worth something if a screen lock is used. The police won't be foiled but a casual thief is likely to unplug or reboot and at that point your data is protected.
Sadly, my computer has apparently rebooted multiple times this week. I didn't do it, but Microsoft decided it was for the best. I remember when a restart was something you were asked to consent to, and before that you had to affirmatively decide to perform an update.
> Maybe I’m a bit weird, but I like seeing all these warnings: “You have unsaved work! Are you sure?”
One of the reasons I don't like to reboot: Windows taking a few seconds to show the warning, so I have to either babysit the computer until it shuts down or come back the next day with an unrebooted computer.
Now I just use shutdown /f to force shutdown/reboot and forget about it.
I shut my desktop down every night and bring it back up in the morning. It only takes all of 2 seconds these days with NVME drives, even with Fast Boot disabled.
It's kind of a power hog and generates a lot of heat, so I try not to run it if I won't be around.
I also shut my PC down every evening and boot it up in the morning. Suspend to RAM draws too much power for my tastes and suspend to disk does too much writing to the NVME and I don't want to have to buy a new one before the AI death spiral ends in 2028.
A couple of years ago I noticed that my mac starts collecting weird little bugs if I don’t reboot for a really long time. The cursor starts misbehaving (it won’t reliably change over links, or in graphic editors), switching between apps might take a few seconds, and once I had my keyboard input latency increased by ~500-700ms for every keystroke. These issues go away on reboot. I’m trying rebooting once a week or so now.
Can confirm. I have similar problems with very long uptimes on my MBA M1. By long, I mean two months or more. The most common recurring problem is the Mac forgetting it's connected to a monitor. But the problems are often very random. I've had most of the ones you described.
Though I shouldn't use the word "common" as the occurrences are rare. My guess is 4-6 times a year over the past five years. Would love to know what causes it but the randomness of the symptoms would make investigating a bit difficult.
I can confirm this happened, at least in the past, with audio issues for me. If I didn't reboot recently, the audio would start stuttering under high CPU load, with lots of pops and crackles and junk.
This lasted many years, across multiple apple laptops, across multiple different OS versions, across multiple different Corporate management solutions!
Mac users, like Linux users, refuse to acknowledge that it can be possible for there to be a problem that they have not personally experienced. They never extend this doubt to problems on Windows.
This is why everyone believes Apple hardware is always flawless despite nearly every machine having serious design issues that hamper or even break the machine, like a display cable getting changed to be too short in a design update such that after a while your display goes really wonky in a way that is really difficult to google, such that your IT department wont figure out that it's a known and unfixed design fault, and spend weeks trying to reinstall the OS and other troubleshooting steps.
No extra hardware, stock macbooks and an imac. No fancy apps either, RayCast is probably the only one that overrides default system behavior. It used to be better, a few years ago macOS would run stable for months, but even then I had to resort to ‘killall Dock’ for occasional glitches.
I used to leave my computer running 24/7, but then I got another computer to make a server, and it really improved my routine to have a "Shutdown and I'm done with computers for today"
I’ve been on macOS since 2012 and only have to restart during Os updates. What is this person on about? Slow news day? Clickbait? I was clearly baited.
No. And in fact it annoys me greatly that our work IT department forces a reboot once a month.
My desktop is like an actual desktop. I have my tools laid out exactly how I want them for the task I'm working on. I want to come back to them in exactly the same place when I pause for the night or weekend.
I have to clean it up occasionally, to close out things that I'm no longer working on or forgot about. But I'd rather do that incrementally than all at once. Forcing a reboot is like tipping over a physical workbench with a bunch of tools and in-progress projects on it. Just awful, and often takes hours to set back to its original position (at least browser tabs restore reliably these days).
This whole time I thought IT keeps sending out those emails about rebooting for updates because people just couldn't be bothered to remember otherwise. Turns out some remember very well, they just don't care!
sounds like some software projects to work on, being able to reboot back to like exactly where you left off (I've been trying to figure this out more for my own computer)
Ever since I got my MacBook, I've only restarted it about twice a month. With Windows, I used to restart my pc/laptop almost every time I finished using it.
my department manages a fleet of ~10k Windows PCs and it's pulling teeth to get the users to allow an automated reboot once a month for updates. they invent all kinds of silly ways to avoid it. your situation seems atypical.
Windows is leading the pack here as usual. They've introduced a new "all your RAM are belong to us" feature where all 48GB of my RAM slowly drips into a black hole over time, ne'er to be released again until reboot.
I shut down every night and power on every morning. The servers with ECC are an exception, I usually keep those on a typical 1 month cycle for all the updates to apply. It's so damn fast these days I active the "Start Day" home automation scene while rolling out of bed and by the time I've reached the computer it's already sitting at the desktop.
Is it really different from that short time ago? It feels like this has been the normal for much longer than 5 years. I do remember there being fewer restarts to update, but I feel like that was way back in the snow leopard days.
As to why, I assumed it’s because security fixes are in basically every update these days. And it’s easier to change core systems with a restart than live
My understanding for Win10 was that Restart was "more complete" than rebooting by a manual power-off/power-on.
Found it: "By default, Windows computers use a "Fast Startup" feature when you click Shut Down." It actually performs a sort of hibernate, saving state data of the system (but not necessarily of all running programs - that's another setting). Restart clears those state registers and begins a new, fresh Windows session.
So, ideally: Perform a Restart, not Shut Down, at the end of your day.
I restarted my Mac the other day and my free disk space went from 4GB to 40GB. Things are also noticeably zippier soon after a restart. Annoyingly, I can't reboot often because I'd lose all my incognito tabs. It's possible that quitting all applications would have the same effect (except the swap space, which seems not to fully release until reboot), but it's just as much hassle to do that as to fully restart the machine.
FreeBSD is my daily driver and I only reboot for major upgrades (which is required). I never power off cause I work on it, off and on, throughout the day and night (for my own company).
Lightning is electricity that goes through hundreds of meters or even kilometers of air (supposedly a good electrical insulator) to reach the ground, it's not above travelling through electrical lines the opposite way they're intended to be used.
No. This isn't some theoretical Star Trek neutrino emissions scenario. Lightning punches-out wherever it hits, including ground, which is directly connected to neutral.
i think this isn't common knowledge; as it's only tied at one panel, and separate in the other (in the US); so there's two busses in the main breaker panel, one for white and one for green/bare. It's only the utility/ingress panel that has the neutral and ground bonded. It's a fluke of the way we deliver power in the US, split phase, maybe.
High voltage electric arcs can be surprisingly long. The tiny cap in strip is not enough to protect in nearby hits. For that fully disconnecting or proper separation is need.
i thought half of lightning came from the ground, like half the potential. Slow mo lightning strikes have them meeting in the center in the air, rather than striking from cloud to ground. I could be wrong.
Lightning hit an antenna that was disconnected but near a radio and blew up the radio, the PC it was connected to, and then everything connected to the switch that was connected to the PC via cat5, and just for good measure, everything connected to an outlet on that side of the house within 15' of the computer outlet. once it gets in, it doesn't matter if stuff is off or on or whatever.
Yes, i shutdown my pc every day. Maybe it's a habit from old times when i was living with my family and we really had to count every cent.
The only thing that usually runs all the time is my mini-pc, which i use as a server but due to the ongoing heat (and me not using it much at the moment), i've shut that down as well.
I guess I'm old-school then too. I `shutdown now` once I'm done for the day. It takes all of a few seconds to start back up. I see literally no value in keeping it on overnight.
Then again, I primarily use a desktop, so that probably factors in.
I have a home office with a desktop workstation. I shut down at EOD. When I worked in an office we were instructed by IT to leave the machines running overnight and they would run their scans and nonsense outside working hours, which was neat.
I do the same. I am not a fan of laptops preferring instead to sit down at a desktop to work. When I'm done I exit all my programs and shut down. I don't understand people who drag their computer state around everywhere.
I have a terrible work / non-work balance, and one trivial habit I've established is turning off my (Mac mini & MacBook Air) computers & screens when my work time is done.
I don't want it to be trivially easy for me to just do one more thing…there be dragons.
My Saturday mornings are more often markers for running Onyx[1] for maintenance.
I use NixOS and haven't restarted my machines in years. Always on. There are power outages once in a while but everything is set to restart when power is available so it comes back up.
I have a Lenovo Legion and a Macbook Pro - I've had to restart the mac a couple of times due to VPN issues with work, but the Lenovo has probably been running for a few months.
I helped manage 1500 desktops and thousands of VMs over twelve years in a call center and I preached rebooting at the end of the day/shift. There is no doubt that this reduced ticket volume compared to other sites. This included individual and shared desktops.
Sometimes I shut down my computer at the end of the day to symbolically end my week.
That being said, I hibernate at the end of my day. For some reason, merely closing my Dell laptop just isn't as smooth on reopen as my Mac. The startup is almost as long as a full reboot.
My UnRAID server has been up for more than a month, and would be much longer were it not for a system update there, too. The uptime of the VMs on the server are also affected by this.
Rebooting desktops or laptops is ok, rebooting your servers is an anxiety induced task, rebooting your archos linux, well, be prepared to spend an hour troubleshooting afterwards.
I not only rebooted but upgraded my server from 22.04 to 24.04 and every single package on it this weekend with AI help - took a few hours of talking to the thing, even though the server had accumulated some ugly hacks (like installed KDE on top of Pop OS, had a newer version of glibc than supported) and runs a ton of containers. The AI figured out all the apt and dependency hell bullshit that I would never have gotten through. Its a new world!
I turn off my work Mac on the weekends because I cannot for the life of me figure out how to make it shut the fuck up. If I close the lid I still hear email, calendar and slack notifications. If I explicitly put it to sleep I still hear notifications. The only way I can walk away from that cursed computer knowing I won't hear it nag me from the next room is to shut it down.
Actually, it's more or less the opposite. In some OS's (Windows, at least) a reboot does additional purging and loss of state that a shutdown will not do. That's by design, so a normal shutdown/startup is fast, but if you're having a problem a reboot will be more thorough.
Every couple of months typically I do an arch linux update and reboot. But that is about it.
I do hibernate sometimes though, and that is pretty much the same final state power-wise as doing a shutdown (more so for my laptop as it does not keep keyboard/mouse powered in S4 and its the same with the hall effect sensor for the lid).
Once I began the “start from zero” routine at work, I could
not stop. It meant I had to know what to do to get back to my ideal dev env every morning. It meant I had the choice to automate that, or to leave some
bits (or all bits) manual. It surprised me how much I preferred manual, and taught me to recognize when things were actually tedious vs premature shortcuts. Clarity trumps convenience, but, you know, not always.
‘“Microsoft Edge is preve…” Bam! Force quit! Kill kill kill!’
Wait, what? Why is OP using Edge on a Mac? To each their own, it just caught me as odd.
And, as Betteridge’s Corollary or whatever demands, the answer to the headline is “no”. Is this like ancient wisdom about batteries, you’ve got to run them to zero once in a while or they’ll get a “memory”? (Which, of course, hasn’t been true for, like, twenty years.)
Nope. When I want to know when was the last time I came back from vacation, I type this at the CLI:
uptime
I turn off my desktop when I go on vacation for more than a few days. If I just leave for the week-end I don't turn it off.
Very rarely there's a published kernel fix leading to an exploit that could potentially affect my setup that requires rebooting, but that is exceedingly rare.
FWIW my desktop regularly reaches six months of uptime and I had a server at OVH which I kept just because I could that reached something silly like 3400 days of uptime (it just didn't reach 10 years). At some point (after maybe three years) the uptime was so cool I decided to just keep it and see how long it'd stay up (and, no, that one wasn't secure at all: kids, don't try this at home). When the fire at OVH took entire bays off, I wasn't affected so the thing kept cranking.
If we leave security concerns aside, OSes are really that stable now (unless we're talking about Microsoft products of course).
> Have You Restarted Your Computer This Week?
Now of course I've got something like 12 computers at home so it really depends which computer you're talking about. For example I've got a server with ECC memory that runs VMs and containers but... I only need it when I'm awake. So that one I typically turn off at night (for the energy consumption). I know, I know: desktop up and server down at night I must be doing something wrong right? But then it's my setup and I do what I want.
I have a cron job which restarts my computer every day (Linux Mint Debian edition). I like waking up to a "fresh" computer, and since I know that it will restart every day - this is a "forcing function" to (1) be diligent about saving things that are important, (2) treating browser tabs, random notes, etc as ephemeral, and bookmarking the important stuff.
I used to work at an office where we pair-programmed with clients all day (Pivotal Labs), and most of our computers had some sort of "automatically restart / restore from a known-good image". I liked this, as it resulted in less cruft over time, and some intentionality about what getting a computer into a productive state means. It also got me thinking of using automatic routines to accomplish goals, and not being so attached to my open tabs, etc. Let it gooo....
To be more specific about this - for those wanting to get into blogging/publishing, this could mean auto-opening the website project folder using VSCodium upon user login, so its ready to go for the morning coffee. More half the time I just close it - but as a "default", it makes it easy for me to do the thing I want to do.
A slightly more fun version:
Some would say "Chaotic evil". Some would say "Always be prepared!!"
Chaotic evil is my favorite article from The Daily WTF:
Some would say, "Oh, Windows Update again"
I’m glad _someone_ is working on bringing the windows experience to Linux!
Regarding browser tabs you can always just reopen Chrome and press control shift T to get your previous tabs even after a restart
> Pivotal Labs
RIP Pivotal Tracker, and thank you for your involvement! It was a nice piece of software.
If your computer has any kind of secure boot / boot chain attestation stuff going on, like all phones do now, rebooting clears out pretty much all malware. Software bugs can put the system in to a bad state which can be abused by malware, but none of it can persist through a reboot since everything is checked and measured from scratch during the boot.
> any kind of secure boot / boot chain attestation [...] rebooting clears out pretty much all malware
IMHO "pretty much" understates the risk.
Malware can easily install itself as a system service, timer unit, XDG autostart or your shell profile among other places. I'll be the first to admit I never check all these places regularly.
The only thing that should be putting minds at ease is regular OS installs from fresh images.
Resist the temptation to do an "in place" upgrade and go with a clean ISO each time your distro comes out with a new major version.
It's a pain but thanks to configuration management or even shell scripts it's manageable for me now.
Admittedly six months is probably too long as well but at least it stops something lurking on a server or my desktop for years.
If you go with an immutable OS like Fedora Silverblue or Bazzite, every single update is a clean ISO install. And then in your user only install things through Flatpak while checking there aren't dangerous permissions enabled you will be much better off than a normal linux setup were mutations can be placed all over the place in a way that is impossible to audit or secure.
This feels rather overstated. Your firmware check isn't going to tell you if a repo package installed/updated by your package manager got compromised, which is apparently increasingly common.
It's more relevant on ios/Android where you can't through normal means install malware at the system level. Linux is moving in this direction with immutable OSs and better secure boot support but there isn't an easy way to have a secure linux desktop setup currently.
On phones malware is more restrained by the sandboxing. But now the reboot isn't necessary in the first place.
Exploits break out of the sandbox for higher system access. Rebooting flushes them out because the exploits have abused some memory bug to load custom code in the system, but this can not persist a reboot since everything gets cyrptographically verified on boot.
This is why things like the Nintendo Switch hacks and iphone jailbreaks have to be reloaded every time the system reboots, while they could modify the system files to try to persist the malicious code, it would simply leave the device unbootable since the boot chain attestation process would refuse to boot when the checksum doesn't match. The malware can only come back when the initial bug has been triggered again, through receiving a malicious text message attachment for example.
> Software bugs can put the system in to a bad state which can be abused by malware, but none of it can persist through a reboot since everything is checked and measured from scratch during the boot.
This does not account for supply chain attacks nor trojan attacks which seek consensual installation. A reboot will not eradicate these threat vectors.
It relies on your software to be properly sandboxed and permissioned.
What we have seen is state funded hacking groups can often exploit bugs which break out of sandboxes and embed themselves in the system. But the malicious code only exists in memory, it can’t persist between reboots because the next boot would fail the checksum and signature checks.
Which is why security focused OSs will periodically reboot themselves to flush out possible malware and why jailbreaks and hacks for phones and game consoles usually have to be reinitiated after every boot.
MacOS has a few measures including SIP and boot chain security to implement this, though it’s currently hard to set up a properly secure Linux desktop.
> It relies on your software to be properly sandboxed and permissioned.
And organizations rely on their employees not falling for phishing email. How well has that worked?
> What we have seen is state funded hacking groups can often exploit bugs which break out of sandboxes and embed themselves in the system. But the malicious code only exists in memory, it can’t persist between reboots because the next boot would fail the checksum and signature checks.
"Malicious code existing only in memory" is patently false by not addressing supply chain attacks previously mentioned. To wit, the npm left-pad attack[0]. See also the thousands of PHP CVEs.
0 - https://en.wikipedia.org/wiki/Npm_left-pad_incident
This is true, which is why it's extremely hard to create a secure linux desktop today. But things are moving in the right direction, if you use Wayland + Flatpak on a immutable OS with secureboot set up you are in a much better position than a standard distro.
Most methods of persisting state have absolutely nothing whatosoever with kernel or modules which are a powerful way to hide malicious activity but far from the only way to persist state including malicious state. The only sane response to being infected with malicious software is to reinstall everything starting with the OS and scan user files AFTER reinstalling.
It is possible even plausible that you can disinfect many sorts of malware from inside a compromised machine but it is inherently less trustworthy and installing from scratch barely takes any more time.
> about what getting a computer into a productive state means
Why would you do that? Is your OS not running stable? Do you not have tight control over the software that is running? Why are you unable to keep it in a productive state?
Greetings from a Computer that only restarts for updates every few days, next to an airgapped machine that has been running for years.
One of our customers has a daily automated restart for their servers. This is because the IT people have long lost track of what the system actually does, and feel they have no time to figure it out.
It is the bane of my existence when it comes to the predictability of uptime, not to mention long-term processes getting interrupted over the day boundary.
I remember when Linux users were practically obsessive about uptime and restarting felt like a sign of failure. This was at a time when Windows seemingly needed to restart once or twice a day, at least.
These days I like to turn my work Mac off at the end of the week just so I feel a literal sense of closure. It's not really the applications minimizing and running in the background; it's ME.
Yes, I remember feeling pride in the stability of my systems when I saw a large uptime. I had a server that had 1000 days of uptime, once. Now when I see a large uptime, I'm terrified of what security patches the kernel may be missing!
I remember installing some new computers for a small shop around y2k, and their NT4 server was acting up a bit when I was adding the new users to the domain controller.
I opened Task Manager to see if any processes was running wild. Imagine my surprise when I saw it had well over 1100 days of uptime!
I still remember the days of servers as pets, rather than cattle, and I was harping about server uptime. A wizened server admin piped in and said he rebooted his servers once a week. Said, if you do it any less frequently, then the odds of catching an error causing change while the person who made said change (possibly himself) is still around and can remember what they did go down precipitously. So, to avoid headaches and potential downtime when it mattered, he would just take servers out of rotation and reboot them, and make sure they came back online.
So true. We have one older, rather large machine in a data center that's been up for.. (checks uptime): 963 days. It has IPMI but at some point something stopped working and now we have to physically go to the data center to restart it. And since we use it every day we can't really afford to lose access to it.
This is kind of like making sure your backups actually work.
You need to test when servers go down, and people who use them should know and understand what happens when the are off.
Live Kernel Patching has been around for about 20 years[-1] now.
Red Hat Enterprise Linux[1] and Oracle (Enterprise Linux) Unbreakable Linux[2] both use it as a selling point.
This feature is still a bit ad hoc because, in most setups, rebooting a system isn't a huge burden and is much simpler than using boutique commands to live-patch it.
[-1] https://en.wikipedia.org/wiki/Ksplice
[0] https://www.ksplice.com/
[1] https://www.redhat.com/en/topics/linux/what-is-linux-kernel-...
[2] https://docs.oracle.com/en/learn/ol-ksplice/
Live patching exists but it tends to cap out at around a year of updates for any particular kernel version. It's not getting you anywhere near 1000 days.
I worked in a place with a lot of Solaris servers with years-long uptime. It would be my job to patch them. Having no idea what config changes that may have happened over the last 3 years which would take effect on boot was always terrifying.
Thankfully there's livepatching (e.g. https://ubuntu.com/security/livepatch )
6.19 added a new Live Update Orchestrator, which allows significantly more of the system to be retained while doing a kexec / Kernel Handover like transisiton to a new kernel too. https://www.phoronix.com/news/Linux-6.19-Live-Update-LUO https://lwn.net/Articles/1033364/
Systemd added support in recent 2.61. Theres also now ways to have user stores, that survive across switches. https://www.phoronix.com/news/systemd-261
There's also `systemctl soft-reboot` which initiates a userspace-only reboot, which quickly restarts the system without going through the full hardware and kernel initialization process.
But that doesn't give you kernel updates, being userspace only?
Real men use: killall -9
I'm glad to see this. Almost 18 years ago I implemented a similar kexec device+memory preservation for a storage vendor. It was done on a Linux kernel of that day, and it had had a memory reservation and handoff protocol between the two kernels to keep some specific PCI device alive, allowing for state restoration at the application side. I'm proud of the fact that the kernel replacement was just under 1 second in execution (after init process optimization) and the whole kernel+app was less than 10 seconds.
5years; I'm 37 now, I was 32. Life seemed easier then.
I had a box set up as NAT (running amazon linux) when we moved from a local datacenter to AWS in 2012. Shut it down last year. It had not been rebooted. Should have grabbed a screen capture of the uptime. Part of me wanted to leave it to reach 5000 days....
I dunno how much easier life was back in May 2021... a year into the pandemonium, heh.
Easier for some, harder for others. The economy was more unfucked because the government printed tons of free money (mostly to business owners).
Ksplice came out of MIT in 2008, which updates your kernel while it's running. No need to reboot! Supports Ubuntu.
I remember when I was asked to replace a core router with a more powerful model. The uptime of the Cisco router was ten years - and it was ten years after the datacenter went into service.
I do still enjoy the odd >30 day uptime on my PC. Usually only reboot when a new kernel version is cut.
I used to reboot into every kernel patch but often I leave .0 running for a very long time now. They seem stable and the kernel moves fast enough nowadays there's often another .0 right around the corner. There might be exploits but they're not a valid threat model for my little desktop.
If something smaller like Mesa updates, I can reload everything simply by logging out/back in, no need for a full reboot/LUKS unlock.
In the mid-2000s I ran a. Fleet of RedHat servers that hosted millions of domains. I had boxes in that fleet that were up for over a year. Netcraft confirmed it!
Microsoft literally bought these 6 or 7 servers to migrate to IIS so they could “beat” Apache. It took more than double the servers, but after I did the initial work it was moved to a different team and I don’t know how the uptime compared.
My machine was rebooted this week due to a power outage. I don't recall the last time prior to that. It generally goes weeks if not months without a reboot.
> These days I like to turn my work Mac off at the end of the week just so I feel a literal sense of closure
It's also just nice to start Monday with a fresh boot.
If nothing else, it keeps me from getting to the point of 200 tabs open that I'm totally definitely going to need again "soon"
200 tabs? The children of summer are still among us, it seems (he says, glancing at the current tab count of slightly over 1800).
Just close them. You're never going to read them. If you really think there's something you need, export the browser state to an archive file, then delete in 10 years after you've never consulted it once.
(Disclaimer: I'm aware that there may be valid reasons for this workflow, but in most cases it's just digital hoarding and the above advice is sorely needed. If you really need 1800 tabs, you know who you are and you can safely ignore me.)
Brave has a function to bookmark all open tabs. I have used that from time to time. Or just made a temp file and written all the URLs to it, in case I want to open a few/dozen of them after rebooting.
I think all browsers have that now. I use it 3-4 times a year with Safari on both macOS and iOS. That way I feel comfortable closing them all without worrying about “losing” anything.
I’ve been doing this, or something similar, for at least 15 years now. Dozens of mass-bookmark “folders”. I’ve never once looked at the bookmarks I made, not a single time. I even have old bookmark files archived here and there, from machines and browsers for which the bookmarks weren’t auto-backed-up to “cloud” storage like my Safari stuff is. As soon as I can get a local llm running that’s up to the task, I’ll probably have it build some kind of table with categorization out of all of them, then have it edit out entire categories of crap until I have something I might actually scan over to recover a few interesting tidbits. Finally make some use of them, now that tools can take enough of the pain out to make that a less-daunting prospect.
The irony is that now, in the LLM age, it wouldn't be that hard to feed a few thousand browser tabs and/or bookmarks into a vector database or whatever and then if some fuzzy memory tickles your brain about something you saw a year ago, you can just query it.
It's the same logic as keeping RAW digital photos. Lightroom is already gaining "find the keepers" AI features. Sooner or later it'll be possible to feed a bunch of burst shots into an AI that just weren't worth the trouble to manually sort through. The AI can do the drudge-work of digging up any gems in the rough...
how do you see how many tabs you have open, an extension? i have tab session manager and it shows i have 80 tabs. about 60 of those are ephemeral, and the other 20 i'd have open on a new browser anyhow (email x3, goog cal, caldav cal, nextcloud files, router, local and remote proxmox, navidrome, the documentation server, etc) everything else is superfluous. although i'd probably be a bit sad if i lost all my tabs right now; hence tab session manager.
Sideberry
I think I have found my soulmate :)
Every crash cuts deep if it doesn't resume correctly.
2616. On an iPad pro. I am not updating the os because I am pretty sure that the current behavior is a bug. The hard limit has always been 500.
… Why…?
Think of it as a memory aide combined with a to-do list.
I can't. My mind explodes just having 10 tabs open. I need a clean browser, done with whatever, close tab. Reopen if I need that site again.
The implication here is that “done with whatever” happens, rather than “start another thing” while the previous thing is still in progress.
And an indictment of the favorites/bookmarks UX
We had a vp of engineering at a retail place who came from windows shop and wanted to restart the servers every night of the holiday season. It took some doing but we finally convinced him “this is Linux we don’t have to that!”
At Javasoft there were Solaris test machines that had been up for 2+ years, and we had to reboot the windows test machines multiple times a day. It felt really good to leave a large queue of work at the end of the day on the Solaris/Sparc machines, knowing it would be done the next morning.
> I remember when Linux users were practically obsessive about uptime and restarting felt like a sign of failure
I remember it too, like it was yesterday. Wait - it was yesterday.
I see people rebooting Linux boxes to cargo cult trying to fix all kinds of issues and I’m like - rebooting is not a solution. This is not Microsoft Windows.
I've followed the same routine each Friday for at least the past 10 years.
- Install all updates
- Save tabs off to Obsidian (or Raindrop now)
- Reboot
Feels good coming in on Monday to a fresh session.
It was a differentiator when distro updates where sparser, and in start comparison with Windows at the time which couldn't stay up for more than a couple of days without crashing (particularly the XP era).
Feels like a distinction between server and workstation. I'd be ashamed to see only 7 days uptime on my servers D:
Restarting windows twice a day meant a productive day, back in the days of windows 98 (which by the way lived well past windows 2000 and windows xp)
Ahem... yeah... "were"...
I do actually reboot occasionally these days, because the world is so serious now.
As a rule, if you don't reboot your servers while you are near them watching for problems, they will reboot by themselves at 3am in the first day you get sick or are traveling.
I don't feel like keeping my tower pc running more than necessary. When I don't use it for hours, I usually turn it off. At least when I go to bed I turn it off. Next day I turn it on again. Like a normal person. I don't see a need to save 1 minute or so, by keeping it running.
Same here. But I hibernate it, which defeats the purpose.
I'm astonished I had to wade through so many other replies to find this one. I had never considered I might be doing things abnormally when I turned my machine off at the end of the day.
If nothing else, think of the power usage!
I've been keeping my desktop machines running 24/7 for ~25 years. I only reboot when necessary.
I remember having around 280 days of uptime on Windows 7 when it went end of life. Having a UPS helps a lot to protect against short power outages or blips.
Nowadays I run Arch Linux, it's been 12 days since a reboot. Not trying to break records, I reboot to apply kernel updates when it's convenient. Since I use tmux and have terminal heavy workflows it takes 1 command and a few seconds to resurrect all of my sessions to get back to where I was at before.
I'm with the UNIX old timers on this one; cleanliness is overrated. Life is messy, creation is messy. More uptime means more mess. Uptime is longevity; there are few things more valuable in life than longevity.
Agreed. I put stuff in ~/tmp only to create ~/tmp2 for even more temporary stuff that I don't delete, and so on. I clean up Desktop though as it is in my face.
You reminded me of this story. Windows, not *nix, though.
Years ago my cube-mate was called to support a desktop user who has having problems with their machine that couldn't be fixed remotely. He realized the user's HDD was full, so the first thing he did was clear the user's trash bin. Mysteriously the machine came up with a ton of free space.
The user seemed happy their machine was up. Nor long after, the help-desk called my cube-mate asking what they did to the user's machine; The user couldn't find any of their files that they had organized in the trash bin.
Mine grow like this: /tmp, /tmp/t, /tmp/tt, /tmp/ttt, [...]
Every now and then I might take the time to sweep through and nuke them all, but that might take months or years.
I agree with the synthesis of what you said and the OP. Let mess grow then prune. Workspace as a garden.
This is actually useful for smartphones. Sometimes smartphone malware is capable of infecting a device but not persisting, so reboots clean it back up.
At least if you trust the NSA's advice: https://s3.documentcloud.org/documents/21018353/nsa-mobile-d...
On GrapheneOS, restarting also brings the device to a BFU state, which makes it loads harder for an adversary with physical access to unlock it.
yep, and this is the case for both major mobile operating systems
fun thing is that graphene had the feature to automatically reboot on a timer after inactivity for a while, but now ios also ships it :D
I remember something where on maybe mainstream Androids you'd have to fully turn off in order to turn it into BFU, but maybe it does just by restarting now.
The same reason I barely restarted my iPhone for a few years. The `checkra1n` jailbreak had to be reinstalled each time (though it was easy):
https://checkra.in/
This is one of the features I appreciate with OneUI, I can set it to automatically restart it in the mornings.
The people that I know, in the security field, restart their phones, multiple times per day.
I shut down my desktop every night. No reason to keep it on. I can save my work, all my tabs persist, it’s fine.
My main complaint is that my PC seems to have crazy clock drift? And windows just doesn’t seem to care? It doesn’t actually cause problems but I will check and see that it’s off by >5 seconds and that just shouldn’t ever happen. My phone is always accurate to <0.05s and same in Linux. Idk why windows isn’t.
Tangent about clock drift: my phone (Pixel 9) drifts by about 5-10 minutes a day when in airplane mode. This doesn't sound like it would be a huge deal, but I do a lot of wilderness backpacking and very long flights where I'm keeping it in airplane mode for 36+ hours to save battery, and then at the end I need an accurate time to make the next flight or meet a ride.
Not really sure what to do about it because there is no official support for Pixels in this country.
Not judging, but I'm curious why you don't wear a watch?
I do typically, but I haven't always been the best at managing watch charge. It's a coros fitness watch and I've killed it on some hikes by using the GPS too much.
Overall not a huge problem, but it is annoying.
There's many cheap watches with effectively infinite battery. But no gps
>No reason to keep it on.
any reason not to simply let it sleep?
Not OP, but the sleep and hibernate states on my desktop are pretty finicky and can't truly be trusted. I'd rather know everything shut down properly than risk it, especially since the thing boots in seconds anyway, and I have to input the encryption password anyway.
doesn't clear up your ram. One big security reason for regular reboots is that you simply get rid of any potential non-persistent crap on your machine. Also performance obviously, with a full shutdown you get back to a known, clean state.
Worth noting on Windows the restart function only does that if you hold Shift or have Fast Startup disabled.
Replace the CMOS battery yet?
I don't have any data to prove it but I think Mac users don't bother "cleaning up" after they are done with their computers.
I think windows and Linux users usually shut down their laptops when they are done.
I believe this is because of how Mac is designed, nothing really closes. You close an app and it's just "minimized". Same behavior as with the lid, you close the lid and it suspends.
If I recall correctly, at some point, this also affected the iPhone, you were not able to "fully close" apps and they decided to add a screen so you could swipe and "close" the app (some run in the background, same as android)
I think your model of open/closed is incomplete and thus misleading. There are more states to a process than "active" and "inactive," and it is not optimal for the system to simply move processes between those two gross states. The obvious example is non-foreground apps during multitasking. A less obvious example is an app during a background refresh.
"Fully closing" a process is not necessarily cleaner than letting the system allocate intelligently, despite what one's puritanical upbringing might make them believe. (Consider how artists often need a messy space to optimally hold their processes.)
I think my point is that minimizing the process doesn't free the ram it's using. Closing it, even if it stays running in the background, should free up resources
> I believe this is because of how Mac is designed
Yes, that they actually got sleep working properly.
I still remember reaching into my backpack to retrieve my Macbook in sleep mode only to find it hot to the touch, having woken up for some goddamn stupid reason. We had "proper" sleep decades ago: we called it "hibernate". Modern "smart sleep" is a technological regression, and Apple is as big an offender as any.
Sure. Do you shut off your phone every time you put it in your pocket?
What's the point of sleeping your computer? If you are not using it then it's better to just turn it off
So I don't have to spend 10-15 minutes saving all my open files (including the scratchpad ones I haven't named) and then later reopening all my applications, recreating all of my tmux windows and panes, setting up my vim splits and tabs again, and starting all of my stopped Docker containers?
I have a feeling someone is going to jump in with a "solution" to each of these forgetting that lots of work is ephemeral/transient and undeserving of even the slightest bit of automation or pre-configuration.
funny, decent text editors like VS Code save even the unnamed unsaved files
Vim~ #emacs#
It takes ages to cold boot. My desktop is ok-ish, but it doens't matter since I only use it occasionally.
My old HP laptop had a slow-ass BIOS that I was convinced had some kind of bug. I replaced it with a brand spanking new thinkpad 2-3 months ago. Guess what? The freaking BIOS is EVEN SLOWER somehow!
They all wake up instantly from sleep.
I therefore only shut them down when I know they'll be unplugged for a while, because for some reason the HP eats through the battery even when off. If suspended, the battery will be out of juice in like two days. Haven't tried any of this with the Lenovo yet.
Suspend used to work great, but since MS figured they should copy Apple half-assedly, suspend is borken. And I have really no idea what we've gained in exchange.
Now that I have an nvme SSD in all my computers and boot times are roughly 10-15 seconds or so unless something has gone wrong, the advantage of sleeping is somewhat mitigated.
Back in the spinning rust era, though, a good unsuspend could be something like 50 times faster to get to a running computer. Possibly more, depending on what your OS needed to start up.
It is still more convenient to have my previous environment most of the time, and still faster to unsuspend than boot, but it isn't as much of an advantage as it used to be, no.
On Windows and Linux it is better to just turn it off because sleep is broken on most hardware. But on Mac there is next to no advantage to doing that. MacOS sleep is also more of a hybrid between sleep and hibernate where it will after enough time offload the ram to the SSD and fully power down so the machine can be in "sleep" without any power drain.
If you turn it off, it might attempt a 45 minute update when you start it again (yes this has happened to me with macos).
The same reason that you don't do a full shutdown every time you walk away from your computer. I don't really believe that you don't understand this.
I think a large part is also how long it takes to restart a Mac. Every so often a coworker has to restart and I could probably restart my Linux (or even Windows) laptop 3 times before they're back on.
Kind of reminds me of how slow Windows computers used to boot back in the Vista and 7 era.
I'd argue the opposite. My mac wakes up from sleep when I open the lid, and is functional in seconds with the fingerprint. Meanwhile sleep on windows is a complete dumpster truck and can result in any of "works fine", "a bunch of apps have got stuck" or "your battery drained in your backpack".
Also, my Win11 desktop is "fast" to get from POST (which takes > 2 minutes to do RAM check on every boot with 192GB RAM) to the login screen, but it's a good few minutes from log in before windows has started all the background stuff and it's actually functional.
What’s crazy is that boot time was a headline feature of Snow Leopard.
Wat? The apple silicon one on my desk restarts in under 30 seconds. Markedly faster than the Windows PC next to it
I'm glad it works for you, but I have witnessed several coworkers restart their macbooks (some M1, some M2, possibly M3) and I don't think I've ever seen a reboot shorter than about two minutes.
At one instance, I rolled over to a coworker who has just rebooted theirs and had a whole 5+ minute conversation.
FWIW, I experienced the same, and worse, with corporate managed Macs.
I don't know what current Windows 11 does, but my typical Linux Mint is single-digit seconds, and my t490 is 7 years old already. Windows 10 is a bit slower but still single second territory. With more current hardware we are talking about 2 seconds plus whatever BIOS needs, see sibling comments. Maybe you have corporate AV on the Windows PC, or funny drivers?
On linux a lot of the time you kinda have to shut down because it either won't go to sleep ("failed to freeze" some process or another) or it will hang or lock up during/after wake - especially if it has nvidia with its binary blobs, but i've had problems with (in kernel, open source) network card drivers too. Since I first had problems like this with linux 2 decades ago, it seems like they will never be able to figure it out.
Some of this seems to be getting worse with the move to wayland. There is no design concept like windows D3D device reset or lost device, or display timeout/recovery, so drivers are apparently charged with perfectly remembering all the state that compositors create, through all the sleep states, with predictably bad results.
I have only ever had one sleep over multiple Linux laptops over nearly 20 years, my own and other people's, including some tricky hardware (an MS surface). I only ever use sleep/suspend to RAM though. I avoid Nvidia though.
That is not true, when you close apps on macOS the process(es) really are finished, just like on Windows or Linux.
I think they think of "closing an app" as closing its windows, which on Mac does not always "Quit" the app.
Right, I see. That is the traditional behaviour but there’s lots of apps that don’t do that, and closing the last window does also quit the app.
That's a perfectly acceptable practice for apps that only ever have one window. For document-based apps or apps that have multiple windows, usually the convention on Mac is for closing the last window to not also quit the app. But of course there are exceptions in both directions, it's not a hard and fast rule.
For sure, I agree.
Linux users do not clean up AFAIK, and the article is by a Mac user who does clean up. Why would Linux and Windows users not use sleep? AFAIK everyone does.
I avoid sleep on my desktop because it often wakes up in a weird broken state, particularly with audio glitching. I avoid sleep on my Steamdeck because the battery slowly drains in sleep.
On my Macbook I always use sleep since there are no downsides.
As a Linux user, it pisses me off whenever I am forced to reboot, and I'm very disappointed if my laptop uptime is measured in less than months.
Needing to shut down to me indicates something is broken.
What can possibly require a laptop to be up for months?
I have a Linux server that can run for years without needing a reboot. But my laptop I just shut it down after my work is done
What can possibly require a laptop to be shut down instead of put to sleep?
Not having to stop sessions of all kinds of things and restart them for no good reason.
It doesn't require it to stay up, and if things were better at retaining state across restarts I would care less, but it's a nuisance to have to log back into things, and get things back exactly how I left them.
I often have half a dozen projects up on different virtual desktops, and leaving them how they were when I worked on it last makes it easier to get back up to speed.
EDIT: I used to leave screen sessions running on servers instead, as the workaround to having to reboot my local machine. But it's nice not to need to.
Different reasons. Mine is on the table and I use it more like a desktop. It will just idle when I'm not around, because I come and go often. My current uptime shows on Debian 30 days, 49 min.
Although... 30 days is maybe a bit misleading, because I ran some heavy shaders without thinking that triggered the GPU watchdog and forced me out of my session. I think killing all user processes is almost like a reboot, although not according to uptime.
Nothing requires it. But if I turn off my computer, it takes 15s or so to boot up. If I leave it on, it's available for me instantly. The latter is a more pleasant experience, so I gravitate towards it.
Keeping a PC running when you aren't doing anything with it sounds like a waste of electricity to me.
Electricity also wears down electronic components, so I think it also shortens the lifespan of your PC parts.
I can't remember the last time I got a new PC because the old one broke, vs. because I wanted one that was better.
As for being wasteful, sure, but it adds up to an insignificant rounding error of my total energy use so it's not something I lose any sleep over.
How much power is consumed when suspended?
I think what you're talking about is how on Mac, an application can stay running without having any windows. Separate from closing all the windows, you usually have to "Quit" the application if you really want it gone. This is useful for a couple reasons:
- You can "Quit" the application without closing all the windows, and then the next time you start the application, your windows can come back.
- You can close all the windows without "Quit"ting the application, and you don't have to wait for the application to load again in order to open a window later.
Additionally, since application lifecycle is managed separately from the open windows, apps can do cool things like saving and restoring the set of open windows through a system restart. Which Windows and Linux still haven't managed. (Maybe Windows can try to restart the processes... I think I saw that becoming an option more recently)
I've never rebooted often in general, even when I daily-drove Windows. Then, it was because it was annoying to get my preferred workspace back after a Windows restart. Now, I daily-drive macOS and I don't often reboot until the machine gets slow/janky because the machine doesn't really need a reboot until then. And I don't hate reboots as much as I would for Windows because macOS is a lot better at session restoration
my every day use macbook I expect 150+ days uptime before something goes wonky that forces a reboot
I used to be in this camp, maybe not 150+ days but with month+ uptimes, but now with Docker I have to restart regularly as I frequently get notes about 'no more disk space' and the only way to reclaim is is to reboot.
I never shut off my Windows laptop. Just close it. Once a month or so when there's a major update it gets rebooted.
My 15 year old kid annoys me because he will do a full shut down just to go eat his dinner. I'm like "just let it sleep" and he's like "No!"
My Lenovo Legion gaming laptop running Linux Mint Cinnamon takes ~10s to boot. I hardly ever leave it on, I mean why??
It is silly to use a cron job to do something that only takes 2s for a human to do.
My homelab Debian Netinst servers tho, run non-stop for months or even years without reboot.
Half century ago Linux had a bug where servers had to be rebooted after X hours, that is no longer an issue. Windows however requires a reboot because its pagination memory gets full and the longer you run it, the more the performance goes to shit.
Don't you reboot servers to get kernel security updates? I usually have unattended upgrades with the option to reboot during the night if necessary
Months to years without a reboot sounds like a pretty big gamble as to whether the hardware still boots, or not. More frequent reboots tend to catch "the hardware has soured" (and how would you know?) issues sooner, like maybe the BIOS decides to come up with the clock set four years into the future and then time sync can fail and then everything fails. The prior example was strictly hypothetical, of course no real system would ever do that.
I reboot one of my machines all the time. It's a Mac Mini running 'Little Snitch', which informs me about all of the network connects my apps make, especially browsers. One cool feature of that program is that one can temporarily grant access to a connection, for instance until the request quits, or one logs out, or one restarts a machine. I'm slowly developing permanent allow / ban lists, but I also use this feature to temporarily utilize some website, aand then reboot to avoid cross-contamination.
The downside is that I often get shunted off into additional authentication workflows, since the prolonged delay caused by my manual approvals triggers some alerts. One of the entertainment ticket buying services is really convinced I'm committing some kind of fraud.
So, in general, I reboot everytime I start using that machine, at least once per day, sometimes more frequently.
Fun fact: in a former life, I worked for a retailer with 1000s of remotely deployed machines and no field-based tech support. One of the OSes we used back in the day had a bug that caused their license authorization service to fail after a certain amount of uptime. We had hundreds of machines that reached that uptime, all on the same day. Suffice it to say, that was not fun.
I boot my computer every time I will use it and shut it down when I am done. No matter whether on Linux, Windows, or Mac. It's a mindset thing: I will sit at the computer, do what I have to do and then shut it down. Just sharing my way of using it.
I remember a big debate "back in the day" about the effects of shutting down your PC at night and the effect on the health of the CPU. The theory at the time was that by shutting down your PC every night, the temperature fluctuations would put more stress on the CPU, shorting its life span. However, the other side said that leaving on your PC all day and night would spend more time running and reduce its life span, so only having it running when you needed it was better. I think a website did a test and found it makes no difference, but I remember it being a big topic around the 2000's (it might have been related to AMD chips at the time running extremely hot)
AMD's first Athlon chips did not run particularily hot. They needed a fan, just like Williamette chips, but actually used less power than the competition. The difference was they had no internal, on-chip diode. If the heatsink and fan fell off, then the chip could cook itself. The boards had a thermal sensor to initiate shutdown if the fan stopped working. But it's debateable if it would detect it quickly enough if the entire cooler fell off.
In that scenario, shutting it off seems better. But it might roast itself on start up anyway.
In my recollection, the tale was not temperature changes on startup, but the various electrical issues of getting things moving from standstill.
Either way, I just always used sleep-mode.
Is it weird that we don't have a way to update the system without rebooting by this day and age?
I think there are things in Linux like live kernel patching (kexec, ksplice), but why by 2026 is this sort of architecture of live system updates not a common or included feature yet?
People hate updates and having to reboot and have downtime right? Security updates to core systems are more important than ever now too. So why hasn't this problem been tackled I wonder?
Wouldn't it be great if our systems could update without us having to reboot and interrupt our workloads?
Many modern Linux distros encourage restarting too. Atomic distros make restarting to boot into a new image the default, though some have flags to do a "live" update too.
You still have to reboot for things like kernel updates. As far as I'm aware, live patching the kernel is just for security issues.
Not only that, my Arch laptop is not even able to continue functioning after full system update without reboot. It unloads nvidia kernel modules crashing newly opened graphical programs.
Windows does more of this than people notice actually. I was on a system the other day, Server 2016, it had downloaded 2 updates, both system file related, one had installed already and the other one was nagging for a reboot.
Stuff like Windows Defender doesnt even really ask if you want to install the update anymore, it just installs the new virus definitions immediately.
Windows has been gradually "undocking" more and more stuff from the main OS install into separately updatable/restartable/etc components for a while.
Yeah, several times. Thanks windows.
Anyone who is using full disk encryption will be turning off their computer when they're not around. Hibernation is an option if you want to keep your state.
I was about to comment that exactly. If you computer is never switched off, your encryption is permanently bypassed.
True but in practice it's still worth something if a screen lock is used. The police won't be foiled but a casual thief is likely to unplug or reboot and at that point your data is protected.
Unless you use iOS which properly handles clearing encryption keys when the device is locked.
My employer frequently reminds us to shut down when we're done for the day. Otherwise a thief will have a laptop with the BitLocker already bypassed.
Sadly, my computer has apparently rebooted multiple times this week. I didn't do it, but Microsoft decided it was for the best. I remember when a restart was something you were asked to consent to, and before that you had to affirmatively decide to perform an update.
Yes the good 'ol days when the user was considered the owner rather than a security liability. I do miss that philosophy immensely.
(env-godsart) kernel_2026 git:(main) uptime 0:44 up 445 days, 10:50, 18 users, load averages: 5.03 4.35 3.73
> Maybe I’m a bit weird, but I like seeing all these warnings: “You have unsaved work! Are you sure?”
One of the reasons I don't like to reboot: Windows taking a few seconds to show the warning, so I have to either babysit the computer until it shuts down or come back the next day with an unrebooted computer.
Now I just use shutdown /f to force shutdown/reboot and forget about it.
I shut my desktop down every night and bring it back up in the morning. It only takes all of 2 seconds these days with NVME drives, even with Fast Boot disabled.
It's kind of a power hog and generates a lot of heat, so I try not to run it if I won't be around.
Shut down != reboot, but you get the idea.
I also shut my PC down every evening and boot it up in the morning. Suspend to RAM draws too much power for my tastes and suspend to disk does too much writing to the NVME and I don't want to have to buy a new one before the AI death spiral ends in 2028.
A couple of years ago I noticed that my mac starts collecting weird little bugs if I don’t reboot for a really long time. The cursor starts misbehaving (it won’t reliably change over links, or in graphic editors), switching between apps might take a few seconds, and once I had my keyboard input latency increased by ~500-700ms for every keystroke. These issues go away on reboot. I’m trying rebooting once a week or so now.
Can confirm. I have similar problems with very long uptimes on my MBA M1. By long, I mean two months or more. The most common recurring problem is the Mac forgetting it's connected to a monitor. But the problems are often very random. I've had most of the ones you described.
Though I shouldn't use the word "common" as the occurrences are rare. My guess is 4-6 times a year over the past five years. Would love to know what causes it but the randomness of the symptoms would make investigating a bit difficult.
I can confirm this happened, at least in the past, with audio issues for me. If I didn't reboot recently, the audio would start stuttering under high CPU load, with lots of pops and crackles and junk.
This lasted many years, across multiple apple laptops, across multiple different OS versions, across multiple different Corporate management solutions!
Mac users, like Linux users, refuse to acknowledge that it can be possible for there to be a problem that they have not personally experienced. They never extend this doubt to problems on Windows.
This is why everyone believes Apple hardware is always flawless despite nearly every machine having serious design issues that hamper or even break the machine, like a display cable getting changed to be too short in a design update such that after a while your display goes really wonky in a way that is really difficult to google, such that your IT department wont figure out that it's a known and unfixed design fault, and spend weeks trying to reinstall the OS and other troubleshooting steps.
Uhm, do you have a virus or other spyware?
hope not : ) I observed these across three computers on different versions of macOS.
I've been using MacOS for 20 years and I haven't encountered it. I rarely reboot.
Do you have some unusual hardware that requires a driver, or otherwise install something that requires a driver?
No extra hardware, stock macbooks and an imac. No fancy apps either, RayCast is probably the only one that overrides default system behavior. It used to be better, a few years ago macOS would run stable for months, but even then I had to resort to ‘killall Dock’ for occasional glitches.
does logging out fix it?
I used to leave my computer running 24/7, but then I got another computer to make a server, and it really improved my routine to have a "Shutdown and I'm done with computers for today"
I’ve been on macOS since 2012 and only have to restart during Os updates. What is this person on about? Slow news day? Clickbait? I was clearly baited.
No. And in fact it annoys me greatly that our work IT department forces a reboot once a month.
My desktop is like an actual desktop. I have my tools laid out exactly how I want them for the task I'm working on. I want to come back to them in exactly the same place when I pause for the night or weekend.
I have to clean it up occasionally, to close out things that I'm no longer working on or forgot about. But I'd rather do that incrementally than all at once. Forcing a reboot is like tipping over a physical workbench with a bunch of tools and in-progress projects on it. Just awful, and often takes hours to set back to its original position (at least browser tabs restore reliably these days).
This whole time I thought IT keeps sending out those emails about rebooting for updates because people just couldn't be bothered to remember otherwise. Turns out some remember very well, they just don't care!
sounds like some software projects to work on, being able to reboot back to like exactly where you left off (I've been trying to figure this out more for my own computer)
At the end of every workday I shut all the windows and shut down my laptop. Unless I'm in a rush. It's soothing!
Desktop for gaming and Macbook get shut down when I'm not using them, the Mac Minis stay on.
Ever since I got my MacBook, I've only restarted it about twice a month. With Windows, I used to restart my pc/laptop almost every time I finished using it.
my department manages a fleet of ~10k Windows PCs and it's pulling teeth to get the users to allow an automated reboot once a month for updates. they invent all kinds of silly ways to avoid it. your situation seems atypical.
Windows is leading the pack here as usual. They've introduced a new "all your RAM are belong to us" feature where all 48GB of my RAM slowly drips into a black hole over time, ne'er to be released again until reboot.
I shut down every night and power on every morning. The servers with ECC are an exception, I usually keep those on a typical 1 month cycle for all the updates to apply. It's so damn fast these days I active the "Start Day" home automation scene while rolling out of bed and by the time I've reached the computer it's already sitting at the desktop.
Anecdotally was discussing with friend circle recently how our Macs have "you must restart now" updates far more often than 1/2/5 years ago.
Not sure how much is that quality of dev went down vs quantity of threats went up.
Is it really different from that short time ago? It feels like this has been the normal for much longer than 5 years. I do remember there being fewer restarts to update, but I feel like that was way back in the snow leopard days.
As to why, I assumed it’s because security fixes are in basically every update these days. And it’s easier to change core systems with a restart than live
No, restarting is an occasional unfortunate workaround for subsystems that don't properly update in place (e.g. OS kernel).
I'm running QubesOS, my various qubes reboot about a dozen times a day.
I too like running the same kernel for months on end. Just kidding, I restart periodically, especially when there's a kernel patch.
To those who worship uptime, how many kmods do you have loaded? Must be nice not subscribing to openwall/NVD/CISA/cve.org
I update the kernel frequently. I reboot when I'm sufficiently annoyed by not being able to use USB devices.
Yes.
And I'll likely do it again.
Not always by choice. I can crash the system by playing a certain game (they still treat the Apple platform like crap).
I can also put it into limbo, with Xcode, one of the most bountiful bug farms on Earth.
Just got back from vacation, so actually yes. But my uptimes usually hit about 150 days before I get around to running OS updates forcing a restart.
No I power it off first and then I start it up again
For Windows, this may not be enough: https://m.majorgeeks.com/content/page/the_truth_about_window...
Indeed, power off and power on sometimes doesn't even reset uptime.
My understanding for Win10 was that Restart was "more complete" than rebooting by a manual power-off/power-on.
Found it: "By default, Windows computers use a "Fast Startup" feature when you click Shut Down." It actually performs a sort of hibernate, saving state data of the system (but not necessarily of all running programs - that's another setting). Restart clears those state registers and begins a new, fresh Windows session.
So, ideally: Perform a Restart, not Shut Down, at the end of your day.
I restarted my Mac the other day and my free disk space went from 4GB to 40GB. Things are also noticeably zippier soon after a restart. Annoyingly, I can't reboot often because I'd lose all my incognito tabs. It's possible that quitting all applications would have the same effect (except the swap space, which seems not to fully release until reboot), but it's just as much hassle to do that as to fully restart the machine.
If you need "incognito" tabs that last longer, it might be worth looking into something like the Temporary Containers addon for Firefox.
Appreciate it, will check this out!
FreeBSD is my daily driver and I only reboot for major upgrades (which is required). I never power off cause I work on it, off and on, throughout the day and night (for my own company).
BSD daily driver too.
Just recently bought a new NVme so I had to reformat, but went 80+ days without a reboot.
Corporate windows laptop; need to restart every 2-3 days at best.
Thankfully I’ve got admin access on my corporate laptop so disabled this nonsense.
I delayed upgraded to Tahoe, now I need to restart about every week or so, or I will get random lag and freezes.
I also need to restart my iPhone or my airpods will refuse to connect.
I preferred the days where I would restart after 6 months, just because it felt right.
Only when the OS security update requires it. Maybe once every couple of weeks on Ubuntu.
23:58:27 up 716 days, 16:44, 1 user, load average: 9.06, 6.93, 6.40
I guess not. However, this is sort of dishonest since I do sometimes do execute kill -1 or reload individual kernel modules.
aren't there other old-school people like me that shut down everything at end of day and restart the next?
can't be hacked if it's completely off
can't get struck by lightning or surges if the surge-strip is flipped off
fans and spinning drives have lifetime on motors
>can't get struck by lightning or surges if the surge-strip is flipped off
That's not how electricity works. Hot may be open but your ground and return is not.
that's an interesting thought but would violate physics for lightning to come in through the ground line?
Lightning is electricity that goes through hundreds of meters or even kilometers of air (supposedly a good electrical insulator) to reach the ground, it's not above travelling through electrical lines the opposite way they're intended to be used.
No. This isn't some theoretical Star Trek neutrino emissions scenario. Lightning punches-out wherever it hits, including ground, which is directly connected to neutral.
i think this isn't common knowledge; as it's only tied at one panel, and separate in the other (in the US); so there's two busses in the main breaker panel, one for white and one for green/bare. It's only the utility/ingress panel that has the neutral and ground bonded. It's a fluke of the way we deliver power in the US, split phase, maybe.
High voltage electric arcs can be surprisingly long. The tiny cap in strip is not enough to protect in nearby hits. For that fully disconnecting or proper separation is need.
i thought half of lightning came from the ground, like half the potential. Slow mo lightning strikes have them meeting in the center in the air, rather than striking from cloud to ground. I could be wrong.
Lightning hit an antenna that was disconnected but near a radio and blew up the radio, the PC it was connected to, and then everything connected to the switch that was connected to the PC via cat5, and just for good measure, everything connected to an outlet on that side of the house within 15' of the computer outlet. once it gets in, it doesn't matter if stuff is off or on or whatever.
“Ground is 0v everywhere” is a rule of thumb with a surprisingly limited scope of validity, not physics.
it can jump from wherever it is to your highly attractive ground line
Yes, i shutdown my pc every day. Maybe it's a habit from old times when i was living with my family and we really had to count every cent.
The only thing that usually runs all the time is my mini-pc, which i use as a server but due to the ongoing heat (and me not using it much at the moment), i've shut that down as well.
I guess I'm old-school then too. I `shutdown now` once I'm done for the day. It takes all of a few seconds to start back up. I see literally no value in keeping it on overnight.
Then again, I primarily use a desktop, so that probably factors in.
I have a home office with a desktop workstation. I shut down at EOD. When I worked in an office we were instructed by IT to leave the machines running overnight and they would run their scans and nonsense outside working hours, which was neat.
I do the same. I am not a fan of laptops preferring instead to sit down at a desktop to work. When I'm done I exit all my programs and shut down. I don't understand people who drag their computer state around everywhere.
Lightning is a special case because it will jump through the air, regardless of the surge switch status.
My Mac and Linux machine get a reboot every now and then. On the other hand, my work Windows machine, welp…
There are enough crappy win32 applications that you probably should restart Windows PCs nightly.
The usual "Windoze suxxor amirite?" reply.
I have a terrible work / non-work balance, and one trivial habit I've established is turning off my (Mac mini & MacBook Air) computers & screens when my work time is done. I don't want it to be trivially easy for me to just do one more thing…there be dragons. My Saturday mornings are more often markers for running Onyx[1] for maintenance.
[1] https://www.titanium-software.fr/en/onyx.html
Yes, a few times. Thank you for asking, I hope you restart your computer some as well. :)
I use NixOS and haven't restarted my machines in years. Always on. There are power outages once in a while but everything is set to restart when power is available so it comes back up.
IMO the power of NixOS is at the other extreme. Set it up with Impermanence (https://wiki.nixos.org/wiki/Impermanence) and reboot often.
Also worth flagging if you are only ever switching in new configs Kernel patches, firmware, or updates to systemd won’t apply until next boot.
uptime: `up 35 days`
it feels bad in some sense but I don't like my environment being interrupted!
I have a Lenovo Legion and a Macbook Pro - I've had to restart the mac a couple of times due to VPN issues with work, but the Lenovo has probably been running for a few months.
I helped manage 1500 desktops and thousands of VMs over twelve years in a call center and I preached rebooting at the end of the day/shift. There is no doubt that this reduced ticket volume compared to other sites. This included individual and shared desktops.
08:15:42 up 14 days, 14:27
Thats actually a long time for me... I'm using Fedora and they ship a lot of updates fairly frequently... and I tend to get twitchy about updates...
Sometimes I shut down my computer at the end of the day to symbolically end my week.
That being said, I hibernate at the end of my day. For some reason, merely closing my Dell laptop just isn't as smooth on reopen as my Mac. The startup is almost as long as a full reboot.
Yes, but only because of a MacOS update.
My UnRAID server has been up for more than a month, and would be much longer were it not for a system update there, too. The uptime of the VMs on the server are also affected by this.
0:43 up 204 days
I need to reboot
Rebooting desktops or laptops is ok, rebooting your servers is an anxiety induced task, rebooting your archos linux, well, be prepared to spend an hour troubleshooting afterwards.
If the server can't be rebooted without anxiety, something has been configured very poorly.
I not only rebooted but upgraded my server from 22.04 to 24.04 and every single package on it this weekend with AI help - took a few hours of talking to the thing, even though the server had accumulated some ugly hacks (like installed KDE on top of Pop OS, had a newer version of glibc than supported) and runs a ton of containers. The AI figured out all the apt and dependency hell bullshit that I would never have gotten through. Its a new world!
I run an XMPP server (Prosody) on Arch Linux since 2020. I upgrade and restart once every week and so far catastrophe has only struck once.
I turn off my work Mac on the weekends because I cannot for the life of me figure out how to make it shut the fuck up. If I close the lid I still hear email, calendar and slack notifications. If I explicitly put it to sleep I still hear notifications. The only way I can walk away from that cursed computer knowing I won't hear it nag me from the next room is to shut it down.
notice that for some hw parts restart ≠ shutdown & reboot. if you really want to start fresh, shut your machine down once a week.
Actually, it's more or less the opposite. In some OS's (Windows, at least) a reboot does additional purging and loss of state that a shutdown will not do. That's by design, so a normal shutdown/startup is fast, but if you're having a problem a reboot will be more thorough.
Every couple of months typically I do an arch linux update and reboot. But that is about it.
I do hibernate sometimes though, and that is pretty much the same final state power-wise as doing a shutdown (more so for my laptop as it does not keep keyboard/mouse powered in S4 and its the same with the hall effect sensor for the lid).
yes, I always do.
But today I had to power it off, I accidentally created a fork bomb changing a couple of scripts on OpenBSD.
It did not freeze the system but I could not create any more processes. shutdown(8) could not even run, so a hard power off :)
Once I began the “start from zero” routine at work, I could not stop. It meant I had to know what to do to get back to my ideal dev env every morning. It meant I had the choice to automate that, or to leave some bits (or all bits) manual. It surprised me how much I preferred manual, and taught me to recognize when things were actually tedious vs premature shortcuts. Clarity trumps convenience, but, you know, not always.
‘“Microsoft Edge is preve…” Bam! Force quit! Kill kill kill!’
Wait, what? Why is OP using Edge on a Mac? To each their own, it just caught me as odd.
And, as Betteridge’s Corollary or whatever demands, the answer to the headline is “no”. Is this like ancient wisdom about batteries, you’ve got to run them to zero once in a while or they’ll get a “memory”? (Which, of course, hasn’t been true for, like, twenty years.)
I've seen a lot of discussion on this recently but I'm not a mac user so my opinion isn't very relevant. Here's a recent recap though <https://windowsforum.com/threads/microsoft-edge-on-mac-wins-...>
Nope. When I want to know when was the last time I came back from vacation, I type this at the CLI:
I turn off my desktop when I go on vacation for more than a few days. If I just leave for the week-end I don't turn it off.Very rarely there's a published kernel fix leading to an exploit that could potentially affect my setup that requires rebooting, but that is exceedingly rare.
FWIW my desktop regularly reaches six months of uptime and I had a server at OVH which I kept just because I could that reached something silly like 3400 days of uptime (it just didn't reach 10 years). At some point (after maybe three years) the uptime was so cool I decided to just keep it and see how long it'd stay up (and, no, that one wasn't secure at all: kids, don't try this at home). When the fire at OVH took entire bays off, I wasn't affected so the thing kept cranking.
If we leave security concerns aside, OSes are really that stable now (unless we're talking about Microsoft products of course).
> Have You Restarted Your Computer This Week?
Now of course I've got something like 12 computers at home so it really depends which computer you're talking about. For example I've got a server with ECC memory that runs VMs and containers but... I only need it when I'm awake. So that one I typically turn off at night (for the energy consumption). I know, I know: desktop up and server down at night I must be doing something wrong right? But then it's my setup and I do what I want.