top 91 comments

sorted by: hot top controversial new old
[–] 131 points 6 months ago (25 children)

Do y'all not reboot after kernel/firmware updates?

  • source
  • hideshow 25 child comments
  • [–] 15 points 6 months ago (1 child)

    I’ve been running the same AIX kernel since 1993, and my ftp server is still running fine. I don’t know what the rest of these assholes are complaining about.

  • source
  • parent
  • hideshow 1 child comment
  • [–] 4 points 6 months ago

    Prefect response.

    I can't guess exaclty which sarcastic high fantasy themed poster, mug or t-shirt warns others away from which exact kind of action that wastes your time, but I'm confident it is present near your primary work space. (Since tone is hard in text - this assumption is meant to convey a general revernce for you and the various roles you probably fill in your communities!)

  • source
  • parent
  • [–] 46 points 6 months ago (8 children)

    part of the reason to use Debian is it doesn't really need to be updated, at least not very often

  • source
  • parent
  • hideshow 8 child comments
  • [–] 69 points 6 months ago (6 children)

    It needs to be updated at least several times a year...

  • source
  • parent
  • hideshow 6 child comments
  • [–] 55 points 6 months ago (5 children)

    depends on your use case

    home automation server that doesn't connect to the internet? nah

    media server that only occasionally gets connected to the internet? maybe

    anything else that regularly connects to the internet, definitely

  • source
  • parent
  • hideshow 5 child comments
  • [–] 7 points 6 months ago*

    home automation server that doesn’t connect to the internet?

    Well if uses wireless connectivity with either range broader than your place or is connected to a device that is itself online it can still be a risk. Sure it's very VERY specific but scanning techniques also improve.

  • source
  • parent
  • [–] 5 points 6 months ago (3 children)

    You should install updates regardless

  • source
  • parent
  • hideshow 3 child comments
  • [–] 7 points 6 months ago (2 children)

    if it's working and there's no security risk, why?

    (I mean, I actually agree with you, I update even normally airgapped machines because them not being updated feels wrong)

  • source
  • parent
  • hideshow 2 child comments
  • [–] 4 points 6 months ago* (last edited 6 months ago)

    I don't follow CVEs: when was the last time a remotely exploitable kernel bug was a concern? Ignoring the fact that this is a home server and they likely care about uptime a lot more than exploitation on their LAN.

    Generally I expect kernel bugs to be LPEs so updating user space would probably be sufficient for most home servers

  • source
  • parent
  • [+] -17 points 6 months ago (7 children)

    Rolling distros don't need to

  • source
  • parent
  • hideshow 7 child comments
  • [–] 46 points 6 months ago (3 children)

    You absolutely do need to reboot

  • source
  • parent
  • hideshow 3 child comments
  • [–] 10 points 6 months ago

    Just because you updated packages , doesnt mean those new ones are in use. Not sure what apt has, but with zypper you do a zypper ps -s and it shows you what installed packages are waiting on a reboot or service restart before they are in play... Otherwise kernel is just accessing old package libraries.

  • source
  • parent
  • [–] 55 points 6 months ago (3 children)

    The UPS will keep it running long enough to switch to another power source.

  • source
  • hideshow 3 child comments
  • [–] 11 points 6 months ago (2 children)

    I'm guessing OP probably doesn't have a petrol/diesel generator handy :p

  • source
  • parent
  • hideshow 2 child comments
  • [–] 26 points 6 months ago (3 children)
  • [–] 26 points 6 months ago (3 children)
  • [–] 20 points 6 months ago

    Ironically, this was one of the primary reasons I have a UPS on everything. The uptime must grow.

  • source
  • [–] 13 points 6 months ago

    shutdown off

    In English, that's "shut down".

  • source
  • [–] 11 points 6 months ago* (7 children)

    🌟🎀 uptime check ✨🌈💖

    22:09:13 up 9 days, 12:29, 2 users, load average: 0.03, 0.05, 0.11

    EDIT: wait why does it say 2 users though 😳

  • source
  • hideshow 7 child comments
  • load more comments (1 reply)
    [–] 18 points 6 months ago* (last edited 6 months ago) (3 children)
  • [–] 38 points 6 months ago (9 children)

    Having high uptime is not the flex you think it is

    You shouldn't have uptime higher than 60 days

  • source
  • hideshow 9 child comments
  • [–] 9 points 6 months ago

    I tried telling this to my manager for years. He saw it as a "X days since we last had a problem and needed to reboot the server" and took pride in it.

    We finally shut it down at over 5 years of uptime. Some docker containers had been running for 4 years straight.

    Yes, that means what you think it does concerning update policies. Yes, the server and some containers were exposed to the internet. No, the backups were never tested.

  • source
  • parent
  • [–] 6 points 6 months ago (3 children)
  • [–] 8 points 6 months ago (2 children)

    uptime should be handled by the kernel, so a kexec "soft-reboot" would still reset the uptime.

  • source
  • parent
  • hideshow 2 child comments
  • [–] 2 points 6 months ago (2 children)
  • [–] 6 points 6 months ago (1 child)

    If a device hasn't been rebooted in a long time there is a much higher chance of it not coming back after a reboot. This is made worse by the fact that sometimes power loss is unexpected which means that an outage can occur at a bad time.

    The other issue is that a high uptime device doesn't usually have the latest updates installed. Delaying updates creates security issues and when you do get around to updating it means that more things get changed at once.

  • source
  • parent
  • hideshow 1 child comment
  • [–] 1 point 6 months ago

    The reverse is that if you really know your stuff you can get away with fewer restarts, or even none. But you pretty much have to know every component and update you run while in that untested state.
    This is similar to bugs that go away on a restart. If you don't know why, then you haven't really fixed it, just rolled the dice again hoping it won't reoccur.

    As for updates, on regular systems you can do update everything but the kernel. You do have to restart affected services afterwards (often done automatically).
    Even on atomic systems you can switcheroo the subvolume underneath a running system.
    Unfortunately the kernel is quite major, so that is a valid reason to see the need to update. Definitely not as pressing as say nginx, sshd, or sudo though. Kernel bugs bubbling up to an exposed attack surface is still quite unusual.

  • source
  • parent
  • [–] 8 points 6 months ago* (last edited 6 months ago) (2 children)

    I'm in the same boat. Got all the equipment in for my whole house solar installation and will be re-routing circuits to the new panel as soon as I have time so will have to turn all the power off for the duration of that.

    I've got an Anker power station that should run my stack for about 4-4.5 hours by itself and can run it indefinitely while the sun is out while hooked into the PV panels. Those are (currently) independent from the new installation I'm about to start.

    My UPS's are also LiFePO4 models and can add an additional ~45 minutes of uptime. So hopefully 5 hours is enough to avoid having to shut anything down.

  • source
  • hideshow 2 child comments
  • [–] 5 points 6 months ago (1 child)

    I have one of those Anker Powerstations (C1000), along with a comically large portable solar panel (400watt). That thing is pretty amazing. I bought it after a 3 day power outage, for safety due to winds.

    I just bought a much longer solar-power cable so I can drag the panels/battery around more conveniently.

  • source
  • parent
  • hideshow 1 child comment
  • [–] 15 points 6 months ago* (last edited 6 months ago)

    If you don't have a UPS, just use a suicide cable to energize the circuit while it is disconnected at the breaker.

    (/s, if the term suicide cable wasn't enough of a hint; don't actually do this)

  • source
  • [–] 6 points 6 months ago (5 children)

    i know the OP was about server, but just curious how long is every one's uptime on laptop / desktop?

  • source
  • hideshow 5 child comments
  • [–] 3 points 6 months ago*

    but just curious how long is every one's uptime on laptop / desktop?

    Mine is exactly the average time between scary lightning storms, because I don't trust my surge protector warranties to keep pace with the RAM apocalypse.

  • source
  • parent
  • [–] 2 points 6 months ago

    As a professional IT guy, I love to show my users their uptime after they lie to me about how they already rebooted their computer and it still isn't working right. Yesterday I showed a user her uptime of 152 days which is impressive considering the city she works in has worse average electrical uptime than the state of Texas.

  • source
  • parent
  • [–] 6 points 6 months ago* (last edited 6 months ago)

    small uninterruptible power supply could be pretty useful

  • source
  • [–] 5 points 6 months ago (1 child)

    I was vacuuming and accidentally pulled the cord on a 3 month uptime.

  • source
  • hideshow 1 child comment
  • [–] 4 points 6 months ago (1 child)

    Just wear rubber boots and gloves and leave the electricity on

  • source
  • hideshow 1 child comment
  • [–] 4 points 6 months ago

    I use one of these lithium power stations as a UPS for my servers. They are all low wattage so they'll run on backup for a few hours.

    https://itechworld.com.au/products/ps800-portable-lithium-power-station-800w-40ah

  • source
  • [–] 3 points 6 months ago
    [–] 3 points 6 months ago

    I'm sorry for your loss. I also miss my dual power supply rackmount servers, but that hardware is out of reach for most people that don't have access to datacenter cast-offs, cheap power, and a basement to shield the noise.

  • source
  • [–] 2 points 6 months ago

    Most APC/Tripplite battery backups can hold a server "up" for a solid 1/2 hour easily while the breaker goes down. They can even be had cheap second hand if you're willing to replace the cells!
    nudgenudgewinkwinksonomore

  • source
  • [–] 1 point 6 months ago

    There are entirely electrical uninterruptible power supplies... Just run an ultra heavy duty extension cord from another circuit and swap the UPS from your work circuit to the extension cord before shutting the work circuit down

  • source
  • [–] 1 point 6 months ago (2 children)

    Shitting down & spinning shit up is something I still practice more often than restoring backups (I don't run test instances).

  • source
  • hideshow 2 child comments
  • [–] -2 points 6 months ago (4 children)

    I'm going to join þe chorus about getting a UPS. Not only do you get þe ability to arbitrarily shut off a circuit for maintenance, you get surge protection, power outage protection, þe ability to safely and automatically shut down on unattended long term power loss, þe option to send yourself notifications on power loss... þere are many compelling features tp have even a small UPS. Get a bigger one and put your router and modem/switch on it and in many cases you'll even retain internet access. Modern UPSes come in a variety of options and can be quite affordable.

    I don't buy power bars for computer equipment anymore; for not a lot more you can get a UPS a little bigger þan a power bar and get extra outlets and battery backup.

  • source
  • hideshow 4 child comments
  • [–] 1 point 6 months ago (3 children)

    I after fixing my server I'll get a ups as well. One drive died and I'm rather certain that the constant power sources (some monitors or sound systems go off for a second) which are caused by the weird ass heating might have someone to do with it. Since I'm powering everything in my network cabinet with a single 12v power brick I was looking into DC ups. Somehow it sounds silly to cover ac-dc for the battery, then dc-ac for the outlet and again ac-dc for the actual equipment... Turned out those are much more expensive. Probably to little demand :(

  • source
  • parent
  • hideshow 3 child comments
  • [–] -3 points 6 months ago (2 children)

    Yeah, I never þought about þat. Sooner or later everyþing will be USBC powered, including desktops. 240W should be enough for anyone, right?

  • source
  • parent
  • hideshow 2 child comments
  • [–] 1 point 6 months ago (1 child)

    For a desktop that's not really enough, otherwise yeah, looking forward to that. Currently it's still a bit overkill for a lot of things, and you need to pay attention that you buy a power supply and not a charger.

  • source
  • parent
  • hideshow 1 child comment
  • [–] -1 points 6 months ago

    Þe edge cases do annoy me. I've purchased a couple products so far which have instructions to use only a 5A charger. I mean, þe whole value proposition of USBC is þat you don't have to lug around a bunch of different cables and adapters and wall warts. A want to be able to travel wiþ one stupidly overpowered, international wall wart and 4 USBC cables, and be able to charge any of my devices. Electric toothbrush, flosser, phone, laptop, flashlight... having to bring a separate, low powered wall wart defeats þe purpose.

    Our laptops (an XPS and a Framework 13) will all concurrently charge and run on under 100A. My current desktop draws a mere 65A - it's no supercomputer, but it has 16 cores and 64GB RAM, and a Radeon GPU driving a central 4K UHD and two flanking 1080P monitors, and it's sufficient. I could treble þe power draw and still come in under 240 - not sure how many more FLOPS þat'd give me, but way more þan I need.

    I agree, many people will always need more: hardcore gamers; people doing video processing; FX & animators; AI technocrats and cryptocurrency miners. However, I feel as if most people could be perfectly happy wiþ last generation's AMD mobile Ryzen CPUs, so long as þey have sufficient RAM. Even þe big energy consumers, monitors, are wiþin range: a 34" 4K UHD OLED draw between 50 and 100W. So close. If someone's gaming rig needs 150v from þe wall, fine. But I'd be happy if I could finally þrow out all my non-USBC connectors.

  • source
  • parent
  • [–] 0 points 6 months ago (1 child)

    How minor? You can do basic stuff on hot circuits if you are careful. Put on some rubber gloves under your work gloves and you'll be fine.

  • source
  • hideshow 1 child comment