Slashdot Mirror


How To Speed Up Linux Booting

An anonymous reader writes "A common complaint about Linux is the amount of time the operating system takes to start. Like Linux itself, there are plenty of options and lots of flexibility for boot-time optimization. From dependency-based solutions like initng to event-based solutions like upstart, there's an optimization solution that should fit your needs. Using the bootchart package, you can dig in further to understand where your system is spending its boot time to optimize even more."

301 comments

  1. -1st post by Looce · · Score: 5, Funny

    My Linux setup is so optimised that this first post is actually made before opening Firefox and typing slashdot.org.

    Ha!

    1. Re:-1st post by Anonymous Coward · · Score: 1, Funny

      I win: busybox with lynx in the initramfs.

      Oh, wait, this isn't NerdClue?

    2. Re:-1st post by Anonymous Coward · · Score: 0

      So you have SeaMonkey and a bookmark. Not all of us can be so lucky, richie rich.

      --
      b634560a51ba2fd6776f7cb99c41c92e

    3. Re:-1st post by Vulva+R.+Thompson,+P · · Score: 1

      Fwiw, you can optimize that further by using Opera and the "/." address bar shortcut to eliminate keystrokes.

      My "-2ND P5OT LOoSERs!" posts regularly beat out the rookie -1 fristpsotters with this trick.

    4. Re:-1st post by alanwj · · Score: 3, Funny

      and typing slashdot.org
      Slashdot isn't your home page?
    5. Re:-1st post by Looce · · Score: 5, Insightful

      Not really. Not even mozilla.org either. That would be about:blank.

      I don't need to download a page every time I start my browser, render it and slow it down, then replace it immediately with another page I want to visit. That's another part of system optimisation, and it avoids unnecessary strain on slashdot/mozilla/other servers, too.

    6. Re:-1st post by Anonymous Coward · · Score: 0

      1+ for about:mozilla !!!

    7. Re:-1st post by Anonymous Coward · · Score: 0

      If that was windows, we would all be saying that you were owned. Perhaps, you are.

    8. Re:-1st post by rabel · · Score: 1

      Slashdot isn't your home page? Umm, the 1990's called. They want their home page back.
    9. Re:-1st post by manifoldronin · · Score: 1

      and it avoids unnecessary strain on slashdot/mozilla/other servers, too.
      That's why I set my firefox homepage to http://www.microsoft.com/
      Oh, wait, never mind...
      --
      Tyranny isn't the worst enemy of a democracy. Cynicism is.
  2. Boot time not an issue. by AmIAnAi · · Score: 4, Interesting

    I thought one of the arguments for linux was that you didn't need to reboot - like you do with Windows. So the boot time should not matter :-)

    --
    Any sufficiently advanced bug is indistinguishable from a feature.
    1. Re:Boot time not an issue. by nbannerman · · Score: 4, Insightful

      I guess the point is that we *should* be switching our machines off whenever possible as opposed to leaving them running for no reason. The home user isn't going to be persuaded by Linux if he/she has to wait a long time to actually get a computer into a usable state*.

      To be fair, my Windows box boots pretty quick; I think the time between power on and desktop is somewhere in the region of 50 seconds. The method of loading the core services - desktop - additional services at least gives the impression of speed, even through the disk continues to thrash for another 45 seconds as applications load in the background.

      * Jokes about Windows never being usable even after booting can be inserted here as required! ;)

    2. Re:Boot time not an issue. by Anonymous Coward · · Score: 0

      Yes, but with a laptop?

    3. Re:Boot time not an issue. by arth1 · · Score: 5, Interesting

      Indeed. The boot time of a system you boot once a year is rather irrelevant.
      Laptop, you say? Hibernate, don't boot!

      What's more interesting is to reduce the login time and start-up time for applications. prelink is your best friend here. Make sure that all your apps are compiled for position independent code (PIC), and prelink them. Lots of time saved, at the expense of larger binaries.

      Regards,
      --
      *Art

    4. Re:Boot time not an issue. by Micah · · Score: 4, Interesting

      In my Kubuntu Feisty install, Hibernate takes about as long as shutting down and starting it back up about as long as a normal start.

      Of course, all the apps are still there so that helps. But it's not nearly as efficient as, say, a Mac where you can close the lid any time, open it back up again and have it right there in 2 seconds.

    5. Re:Boot time not an issue. by Khashishi · · Score: 1

      Unless you are running a server or something, there's no reason to leave your computer on when you aren't using it. It just wastes power.

    6. Re:Boot time not an issue. by GIL_Dude · · Score: 1

      Well I haven't used a Mac, so I am in a bit of ignorance here - but isn't that sleep or suspend? Which Windows has too (and I would have to imagine Linux does as well). Now I am regretting that I have never used Linux on a notebook (I have it on desktops; low end ones without power management features), but my notebooks are always Windows. The sleep mode is pretty darn quick as long as you are not going to be without power for like 48 hours in which case sleep will still draw the battery down a lot in that time. I would assume that Linux can do the same and hibernate is only used for times when you are going to have the machine off or off of power for a long long time? BTW, with disk encryption hibernate or sleep is usually seen as less secure by far than shutting down because the keys are already in memory.

    7. Re:Boot time not an issue. by ThisNukes4u · · Score: 1

      You're suspending to disk, which is why it takes so long. Try suspending to ram, which is what Macs do.

      --
      thisnukes4u.net
    8. Re:Boot time not an issue. by Bert64 · · Score: 4, Interesting

      But as you said, for the first few minutes after boot, windows is thrashing the disk and running slower... Sometimes it even does ridiculous things, like closing the start menu while your trying to select something from it, or ignoring some mouse clicks.
      You could make linux start in the same way, modify the init scripts to start XDM first, and everything else later, in which case you'd have the same appearance of fast booting.

      --
      http://spamdecoy.net - free throwaway anonymous email - avoid spam!
    9. Re:Boot time not an issue. by MasterPoof · · Score: 1

      Precisely, you may gain some time with a good boot script, but one of the tenets of Linux is the lack of need to reboot. ;) However, this is more useful to people with linux on laptops, who tend to reboot more often.

      --
      Using GNU/Linux -- Windows-free zone!
    10. Re:Boot time not an issue. by Columcille · · Score: 5, Informative

      suspend to disk = hibernate, suspend to ram = sleep. Sleep uses the battery, hibernate doesn't. Granted sleep mode doesn't use much, but it isn't altogether negligible. If you don't want to use any power while moving around, hibernate is the way to go. Perhaps that's his scenario.

      --
      I love my sig.
    11. Re:Boot time not an issue. by doodleboy · · Score: 1

      I guess the point is that we *should* be switching our machines off whenever possible as opposed to leaving them running for no reason.
      I care about saving electricity as much as the next guy (fuel efficient car, low-power lighting, recycling, etc.), but I draw the line with my computers. One of my linux boxes is a firewall, so that's up all the time. Another is my old desktop that has some extra drives in it which also I use to play with remote X, LDAP, all that other stuff. And my desktop box is always on. I walk in, I tap the spacebar, and I'm up.

      One tip that will save a lot of power: ditch the CRT.

      The home user isn't going to be persuaded by Linux if he/she has to wait a long time to actually get a computer into a usable state*.
      Other than for power outages or new kernels, I don't ever reboot. Problem solved :-)

      root@tux:~/.ssh# uptime
      14:28:03 up 18 days, 17:05, 3 users, load average: 0.52, 0.59, 0.50
      root@tux:~/.ssh# ssh archive "uptime"
      14:28:08 up 86 days, 19:17, 0 users, load average: 0.00, 0.00, 0.00
      root@tux:~/.ssh# ssh -p 222 ipcop "uptime"
      14:28:15 up 67 days, 14:32, 0 users, load average: 0.28, 0.32, 0.26
    12. Re:Boot time not an issue. by ThisNukes4u · · Score: 2, Informative

      Yes, but he was comparing linux's performance to how quick OS X works just by opening the lid, which is directly comparable to sleep, not hibernate.

      --
      thisnukes4u.net
    13. Re:Boot time not an issue. by germanStefan · · Score: 1

      yes, my desktops and servers never need reboots. Only power outages or kernel upgrades. However laptops need to be restarted whenever you move location and are without power. Sadly Linux sleep support is not where it should be (thank you proprietary and non-standard acpi motherboards) and hibernate works 75% of the time for me. Thusly if you travel a lot you will be rebooting a lot.

    14. Re:Boot time not an issue. by cmacb · · Score: 3, Interesting

      To be fair, my Windows box boots pretty quick; I think the time between power on and desktop is somewhere in the region of 50 seconds.


      The only reason you can say that 50 seconds seems pretty quick is that most of us remember when several minutes was he norm. The interesting thing is that as feature creep in all operating systems has continued (you can't have glass windows in one OS without users of all other OSs feel the "need" for it too after all) has kept boot time longer than we would like, even though hardware speeds have continued to increase by orders of magnitude.

      Can you imagine how long an XP boot (or Ubuntu for that mater) would take on an old 386 system with a sluggish hard drive and not a lot of memory (if such a thing were even possible)? Hours I would guess, and you would shake your head wondering if your hard drives MTBF would get you through the process.

      In the future will it take four or eight processors in a box to keep the lag down to 50 seconds? Should we take any delight in the fact that Windows boot will be sped up again only by special code to pre-load parts of the OS into flash ram before shutdown? I don't. I'd much prefer to see an almost-instant-on OS that didn't depend on special hardware tricks but rather because the architects actually designed the bloody thing for a change. Aint gonna happen though. If there are still any really smart people working at MS I'm sure they are working on the next great Google/Sony/IBM/Oracle killer or something. Faster boots would benefit ALL Windows users, not just MS only shops. We can't have that now can we?

      It's a good sign when an OS rarely needs to be booted, which is at least the case with Linux and OS X (can't speak for Vista). I leave my machines running all the time, even my desktop has laptop innards, so they go into a low power state when not used for a while. What I do to clean up any cruft that has built p running poorly behaved applications is to reboot when I am done using the computer for a while. That way I don't have to sit around and wait for the process to complete. the machine reboots, sits there for thirty minutes and then goes to sleep. It's ready to go and "freshly" booted the next time I need it. Of course if you like to keep a lot of memory hogs autoloaded and running in the background this system may not work so well.
    15. Re:Boot time not an issue. by init100 · · Score: 1

      However, this is more useful to people with linux on laptops, who tend to reboot more often.

      Power-wise, turning off my desktop is far more important than turning off my laptop. My laptop draws about 50 W under load, while my desktop (inkluding screens) draws som 250+ W. With a price of at least 1 SEK/kWh (7 SEK = 1 USD), the savings of turning off your desktop can be significant pretty fast.

    16. Re:Boot time not an issue. by hackstraw · · Score: 1

      I guess the point is that we *should* be switching our machines off whenever possible as opposed to leaving them running for no reason.

      I leave my Linux box at work on so that I don't have to reopen windows and whatnot every morning. That is my reason, but I'm also upset that my work did not buy me a Mac like I have at home. It goes to sleep, and only uses a few watts of power and takes seconds to wake up.

      My point is that in 2007, computers should be viewed as appliances, and that power management should be part of the OS.

      Most servers need to be on 24/7, but workstations should utilize power management. I can't comment on Windows because I'm clueless about that OS, but OS X is great for power management, and Linux as great as it is, is still predominantly a server/embedded OS, and little has been done to improve power management on desktop systems and low power systems like laptops.

      Oh, and my Mac boots very quickly as well, but I only boot it a handful of times a year.

    17. Re:Boot time not an issue. by danwat1234 · · Score: 1

      So, i guess you haven't heard of a reasonable Distributed Computing project to run?

    18. Re:Boot time not an issue. by StarkRG · · Score: 1

      The problem with Windows is that, while the desktop shows up really quick, the boot is not done yet and the system is so overloaded you can't do much until they're all finished loading. With my Linux install the boot takes a while but when the desktop shows up it's all done and ready for use.

    19. Re:Boot time not an issue. by ehrichweiss · · Score: 2, Informative

      Never have to reboot Windows?!?!? Can you please post your IP so these MS.Blaster worms will have something to feed upon? ;) Seriously, if you never have to reboot Windows, you're probably a security risk since you've likely not applied any security patches either.

      And just for the record. The point of the GP was that Linux machines tend to have much longer times before a boot is necessary. It wasn't bashing your precious Windows, it was stating a *generalized* fact. I can leave my Fedora server running for 2 months without a single reboot; if I tried that with Windows I'd be at the very least SLOW, and more than likely infected with malware.

      --
      0x09F911029D74E35BD84156C5635688C0
    20. Re:Boot time not an issue. by Rakshasa+Taisab · · Score: 1

      Yeah, cause writing all those gigs of data to disk is free, right? Right?

      So you want the computer to sleep/hibernate for an hour, which is cheaper? Well, don't ask me as i haven't done any calculations on it, but I'm sure you haven't either.

      --
      - These characters were randomly selected.
    21. Re:Boot time not an issue. by nairb774 · · Score: 1

      From pressing the power button to the KDE desktop in 47 seconds flat. This includes the POST time, the 2 second delay in GRUB, manually logging on at the command prompt, then typing startx. My machine never loaded XP so fast. Linux bca 2.6.18-gentoo-r6 #5 Wed Feb 7 09:15:07 CST 2007 x86_64 AMD Athlon(tm) 64 Processor 2800+ AuthenticAMD GNU/Linux Custom compiled programs is not the reason for the speed - Gentoo or not. Trim boot scripts, kernel w/ minimal modules, and most drivers compiled in. Oddly powering down takes longer then powering on. I will give XP credit for being much faster then previous versions - I have not tried Vista yet, probably won't.

    22. Re:Boot time not an issue. by smaddox · · Score: 1

      It only has to store the memory actually in use, which shouldn't be more than a couple hundred megs when you put it in hibernate. (why would you go into hibernate in the middle of using memory intensive app?)

      Also, as far as I know, sleep only puts the processor into sleep mode (as apposed to cutting power off entirely). In other words, it still draws a considerable amount of power. However, since theres no disk access, and the LCD is off, it is a LOT better than nothing. But if you are going to be gone for more than 30 minutes, put it in hibernate to keep the battery from getting drained.

    23. Re:Boot time not an issue. by neongrau · · Score: 3, Insightful

      i think all we need is an near 100% reliable "suspend-to-some-super-fast-non-volatile-ram"

      when we have that a reboot will be more like compiling the current state.
      as long as you don't change any configration no reboot should be necessary at all.
      just suspend and restore all the time.

    24. Re:Boot time not an issue. by arth1 · · Score: 1

      Unless you are running a server or something, there's no reason to leave your computer on when you aren't using it. It just wastes power.


      Of course, that little "or something" includes things like nightly backups (you do take backups, don't you?), defragmentations, log rotations and archival, AV definition updates, patch downloading and other things you don't want to be interrupted with during the day.

      Also, a system that's freshly started will be slower, both because it won't have the unused part of memory filled up with disk cache yet, and because it can't page out infrequently used kernel and application memory blocks to swap to free the memory up for disk cache.
          A Windows system might run faster right after booting, but this is not the case for a Linux system, which won't hit optimal memory/swap use until several days after it's started.

      An example might clarify. Entering "free" on my 512 MB laptop which has been on for a few days shows:

      total used free shared buffers cached
      Mem: 516440 509664 6776 0 132 127192
      -/+ buffers/cache: 382340 134100
      Swap: 1052248 23304 1028944
      That's 125 MB of disk cache that wouldn't be there for a freshly started machine. And 23 MB of data swapped out to the swap partition because it's never being called, and thus can better be freed up for disk cache and program usage. Having a faster system is a good reason to leave the machine on, isn't it? Especially considering how dreadfully slow laptop hard drives usually are. It's not like it's consuming a lot of power when it scales down its CPU to minimum, the LCD turns off and the disk enters sleep mode.

      And what energy it uses is renewable electrical energy (tell your electric company that you want to switch to a "green" plan -- it'll cost more, but you may sleep better at night), and during winter, if I turned it off, it'd just cause the oil heater to work that much harder.
          It might have saved me a few pennies to turn it off, but not any substantial amount, and I'd burn more oil to keep the room the same temperature, and that's not environmentally friendly, is it?

      --
      *Art
    25. Re:Boot time not an issue. by Anonymous Coward · · Score: 0

      Yeah, cause writing all those gigs of data to disk is free, right? Most memory management works by paging into swap aggressively (while not erasing the in memory version). Thus, most of memory should already be in swap and not need saved to disk again. So yes, it's free in that it has already been done.

      Please don't forget to include that when doing your comparisons.
    26. Re:Boot time not an issue. by rawtatoor · · Score: 1

      In my Kubuntu Feisty install, Hibernate takes about as long as shutting down and starting it back up about as long as a normal start.

      There is no reason it should. I have removed a lot of the preinstalled stuff on my ubuntu laptop, and it resumes in 20 seconds flat, and it was the same with debian on that laptop. I think linux more than competes with anyone else in this regard.

    27. Re:Boot time not an issue. by caseih · · Score: 1

      I think that it's more worthwhile to spend time making Linux suspend and hibernate correctly. My PowerBook, for example, hasn't been booted or rebooted in quite a few months. For a desktop user boot time is important, but hitting a button and being ready to go, as should happen when you wake from sleep, is even better. I'm amused that Windows users are happy with their 15 second wake-from-sleep times. My PowerBook (and now MacBook) are ready to go in about a second or two. I think Linux on laptops and desktops should be this way.

    28. Re:Boot time not an issue. by billcopc · · Score: 1

      I personally don't understand why the OS needs to load so much crap to just give me a GUI. We wait 30 seconds to a minute for the thing to load up, all it does is present us with a bare desktop. Dude, I could write a friggin Dos program that will show a Start menu and taskbar in less time than it takes for the display to resync. The problem is we've got a few dozen background processes that do FUCKALL, but are somehow desperately needed. And what's this trend of having things loaded 24/7 just to check for updates ? What ever happened to the simplicity of the cron job ?

      I look at my current (WinXP) process list. 41 Processes running (-1 for the task monitor), and all I've got open is Firefox. I see some stupid spyware for the Lexmark printer I love to hate, and 3-4 processes for the ATI video card whose drivers have always been shite (one Canadian company I'm NOT proud of). Moving along I've got a few VMware support services, which is dumb considering they're only useful when VMware is actually running. Then a little process for my sound card: 2.5mb just to display a 32x32 icon in a taskbar.. so 4096 bytes for the RGBA icon, and 2.496mb of some other junk... it actually launches a 30mb ugly skinned mixer when you click it. Then there's a background defrag process, even though I strictly disabled any background disk processes because it makes my boot drive chug. Add in a few SVCHOST entries that do god knows what, and the total allocated VM is nearing 400mb. Notice I didn't mention any virus/malware scanners, thank god I only run them on-demand.

      So why does it take 400mb of ram and enough cpu usage to hose a 486, just so a web page can sit idly on my display ? Linux may present fewer actual processes at boot, and certainly a lower memory footprint, but it takes a good 2-3 minutes to start up, most of it spent waiting on I/O while probing devices. I'm in no way a kernel hacker, in fact I wouldn't want to touch that code with a ten foot pole, but most of my tweaking efforts are spent trying to speed up booting. The way I see it, the OS needs to initialize the VM manager, a handful of urgently-needed devices like keyboard/mouse (which should be practically instant), give the NIC a few seconds to negotiate DHCP, and switch to graphics mode. It's been a few years since I've written system code, but I figure that should take a whopping 5 seconds to do, the slowest part being the DHCP request. We're talking about a strictly desktop system here, no web server or database running in the background. I'd expect a server to take a little longer to load, because it's essentially loading its "work" software as part of the boot process, whereas as desktop user would just launch apps as needed.

      I think we need to refocus the true purpose of a core operating system kernel. It doesn't need to do the dishes and julienne fries, it just needs to take a PC from the BIOS screen to a clickable desktop in as few steps as possible. Let the application-specific stuff be handled by *drumroll* APPLICATIONS!

      --
      -Billco, Fnarg.com
    29. Re:Boot time not an issue. by Hamman · · Score: 1

      I actually found myself booting more often when using Linux than with Windows, due to suspend not working properly. Might've been my motherboard, or the fact that my experience is a bit dated(it's been roughly one year since i kicked Ubuntu out).

    30. Re:Boot time not an issue. by mikael · · Score: 1

      Linux (at least Fedora Core 6) is starting to fill up with some extra baggage. My most recent discovery was that around 1 gigabyte of disk space had been gobbled up by my ".beagle" and ".thumbnails" directories. The first apparently attempts to create an index of every single text document on the file system, while the latter creates a thumbnail of every single image viewed.

      I also encountered an annoying pop-up application (scim) which allows unicode text to be entered (obviously essential for languages like Chinese/Arabic or Japanese).

      Doing a 'ps -ae | wc' reveals that there are around 130 processes running - the obvious ones are related to system input/output and file storage, console terminals and windows managers. Other processes have manual pages, so you can find out what exactly they do, but others like 'escd', 'puplet' and 'eggcups' need a google search to find out what they do.

      --
      Vintage computer adverts: http://www.vintageadbrowser.com/computers-and-software-ads
    31. Re:Boot time not an issue. by celkin · · Score: 0

      The word "reboot" is being thrown around very loosely now. What I meant was that my computer rarely freezes up, leading to a forced reboot (i. e. holding down the power button). I do turn my computer off every night and "reboot" the next day because hibernating makes my computer slower than usual if I go too long without a precious reboot.

      I'm just fed up with people saying that all PC's crash ALL THE TIME. I'm very close to calling the hyperbole police to come lock these people up in Exaggeration-traz (American Dad reference). My PC works just fine. It runs slowly, but it's intended for Windows 98, and I have XP. It gets the job done, though. Macs are mainly eye-candy, and the interface isn't very productive (you can't maximize windows unless you drag the window to one corner of the screen and resize the other corner, then you have to do it again next time you open the window). I have never touched a Linux machine, so I can't judge it.

      --
      "Oh c'mon, I wumbo, you wumbo, he/she/me...wumbo, wumboed, womboing...wombology? The study of wumbo? It's first grade,
    32. Re:Boot time not an issue. by mackyrae · · Score: 1

      I don't think SCIM is there by default in Fedora. I'm pretty sure my Fedora partition doesn't have it. I know I had to go through and set it up to make it go on Ubuntu.

      --
      look! it's a bird, it's a plane, it's....a girl? yes, a girl browsing Slashdot on Linux
    33. Re:Boot time not an issue. by xenocide2 · · Score: 1

      There IS a reason it should -- Ubuntu's done a good job of keeping the base working set small and the start up fast. However, that doesn't stop people from buying a system with gigs of RAM. Running firefox in your GNOME Ubuntu setup probably costs 200MB of what could be 2GB. In contrast, hibernate has to restore the entire state of RAM. It doesn't matter if you had 200MB or a 3 GB of applications, it's all gotta come back in before you start running the kernel. It's quite easy to imagine that pulling in 1 tenth the data off disk and doing some calculations in the meanwhile is faster than hibernate.

      --
      I Browse at +4 Flamebait

      Open Source Sysadmin

    34. Re:Boot time not an issue. by wasabii · · Score: 1

      Heh. I have your exact same processor and XP takes about 15 seconds to boot, that's excluding bios and stuff... once windows actually gets going, I have 15 seconds to the desktop.

      Sounds like your KDE needs some work.

    35. Re:Boot time not an issue. by mabinogi · · Score: 1

      (why would you go into hibernate in the middle of using memory intensive app?) Why not?
      In fact, I'd say the times you want to go in to hibernate when in the middle of using a memory intensive application are probably exactly the times when you want it to be as quick as possible.
      --
      Advanced users are users too!
    36. Re:Boot time not an issue. by Thomas+the+Doubter · · Score: 1

      You must have a very minimal configuration. Where I work, and this is the truth, it usually takes about 5 minutes to boot into a working application under XP. Of course the boot includes network authentication, hot-patch uploads, virus scans, etc. But, this is a major corporation, and this waste of time seems to be tolerable for thousands of people every day. May I go so far to say that boot-up time is a non-issue at many large companies?

    37. Re:Boot time not an issue. by TheNetAvenger · · Score: 3, Interesting

      guess the point is that we *should* be switching our machines off whenever possible as opposed to leaving them running for no reason. The home user isn't going to be persuaded by Linux if he/she has to wait a long time to actually get a computer into a usable state*.


      Yes, but rebooting should not be the option people are using. Go the power management route and do things like enter low power mode, or do a full suspend to disk (hibernate).

      Rebooting really should not the solution to using less power, especially with the Power management concepts can be automated based on idle usage, turn themselves back on to run tasks and back off.

      I got into this habit from carrying a couple of laptops full time, and it is something I carry over to my personal desktops as well.

      People need to just exchange Off with Hibernate in their minds. Even with Windows, which doesn't have bad boot times, it is just easier to tell the system your power button is the hibernate button, and hit it and go on my way. Then turning on the computer is a few seconds and I don't have to worry about what I left running ever. Especially with someone like me that has tons of crap running all the time, including several VMs which I don't have to individually suspend to disk if I were to shut down the computer everytime.

      And thank god OSX finally added a true hibernate in 10.4 on their notebooks, cause not having it about drove me flipping insane.

      So now that 99% of all OSes have ok power management, rebooting should be a thing of core updates only.

      PS
      The boot times on Linux are really not bad in a default install on most distributions. And people shouldn't take this article as evidence that Linux is slow or sucks at boot times.

      However, I do applaud the efforts to improve boot times, and wish there was a bit more generic optimization like people from the Windows world are use to.

      Vista for example monitors the last 5 boot times, and will continually adjust disk layout and process order, etc to continue to speed up boot times. There is no reason all OSes couldn't add a generic form of optimization like this.

    38. Re:Boot time not an issue. by poopdeville · · Score: 1

      Sleep's power draw is actually quite negligible. I've woken up my PowerBook G4 after sleeping, on battery power, for a week.

      --
      After all, I am strangely colored.
    39. Re:Boot time not an issue. by anton_kg · · Score: 1

      Well, it's true. I don't reboot my box. And it's nearly true with Vista too. Windows users (and everybody else with them) so used to reboot workstations. However there is no real reason to do so if it just works. I guess Windows(c) is much more stable now and the default action has been changed to suspend to ram in Vista. Once you are doing suspend/hibernate it shoudn't be big difference which OS you are using.

    40. Re:Boot time not an issue. by Dillon2112 · · Score: 1

      I have a Dell 700m running Kubuntu Edgy Eft.

      I use "sleep" mode exclusively by symlinking the sleep.sh script over to the the lid.sh script.

      I can keep my laptop in sleep mode for a little over 6 days with no external source of power. Traveling, I once used my computer for a total of 4 hours on a set of flights spanning 2 days with only sleep mode being engaged between uses. It was at 17% when I opening it up and plugged it in at the hotel.

      This is a 12" laptop, and I was using the extended battery available from Dell.

      I use sleep as my preferred option for periods of inactivity anywhere between 5 minutes and 24 hours. Under 5 minutes, I leave it running. Over 24 hours, I tend to shut it down if I know it will be that long. Sometimes, I guess wrong, and leave it in sleep mode for days. I've never lost data because I drained the battery.

    41. Re:Boot time not an issue. by bruno.fatia · · Score: 1

      Can you imagine how long an XP boot (or Ubuntu for that mater) would take on an old 386 system with a sluggish hard drive and not a lot of memory (if such a thing were even possible)? Maybe it isn't possible on a 386, but surely is possible on a 8MHz Pentium II
      Also, my Win2k3 workstation doesn't reboot unless I do some major update, say updating video drivers.
    42. Re:Boot time not an issue. by 1110110001 · · Score: 1

      (why would you go into hibernate in the middle of using memory intensive app?)

      Because in my case these apps are called Firefox and NetNewsWire and I have some pages open I want to read a little later.

    43. Re:Boot time not an issue. by fluffy99 · · Score: 1

      You may have a logon prompt in 50 seconds, but I guarantee its not fully booted as services are still starting up. Since most people equate the logon screen with booted, they allowed the logon to occur before the services are all fully started. For example, boot and logon immediately and see if you can actually get to shared drives. Terminal services starts pretty late in the game for Windows 2003 I've noticed.

      Reading the article, it looks like they are introducing the concept of dependencies and allowing things to start in parallel instead of sequentially as the sysinit process normally does. In other words, they've finally caught up with Windows NT4 which does exactly this and with a much better method of keeping track of the service dependencies.

    44. Re:Boot time not an issue. by Anonymous Coward · · Score: 0

      In late 2004 I installed Windows XP on an old Compaq Presario. The maximum processor speed was 75mhz, but I installed some super CPU that took it to 211mhz (233 is what it was supposed to be). It has 81-ish MB of ram, and a 1.5gb HD.

      It took about 5 minutes to load, that was without any applications installed. The GUI in "normal XP" was so sluggish it took a few minutes to even open IE.

    45. Re:Boot time not an issue. by drsmithy · · Score: 1

      In the future will it take four or eight processors in a box to keep the lag down to 50 seconds? Should we take any delight in the fact that Windows boot will be sped up again only by special code to pre-load parts of the OS into flash ram before shutdown? I don't. I'd much prefer to see an almost-instant-on OS that didn't depend on special hardware tricks but rather because the architects actually designed the bloody thing for a change. Aint gonna happen though. If there are still any really smart people working at MS I'm sure they are working on the next great Google/Sony/IBM/Oracle killer or something. Faster boots would benefit ALL Windows users, not just MS only shops. We can't have that now can we?

      The vast majority of bootup time is in hardware initialisation and probes, not OS overheads.

      It's a good sign when an OS rarely needs to be booted, which is at least the case with Linux and OS X (can't speak for Vista).

      It is also the case for Windows. If your Windows PC needs to be rebooted for any reason other than patching, it's broken and should be fixed.

    46. Re:Boot time not an issue. by Architect_sasyr · · Score: 1

      There was an article recently (was it on /.? I don't know) which gave a lot of tips for power saving and performance. To cover one or two:

      * As parent said, LCD screens. Not only do these reduce power they're almost instant on
      * More RAM. Every geeks dream, but more RAM means less paging. This also helps to increase boot time. After monitoring system performance and adding an additional 50% RAM to my file server (parent's "archive") for a total of 256MB RAM I managed to reduce the boot time to something in the order of 30 seconds.

      Obviously there is only so much hardware you can throw at a system, but it is definately worth looking at.

      Doing this helps the home user reboot easily (I mean, seriously guys, you're posting on slashdot... how can you call yourself a home user?!)

      my $0.02 AU

      --
      Me failed English...
      FreeBSD over Linux. If my comments seem odd, this may explain...
    47. Re:Boot time not an issue. by fmaresca · · Score: 1

      My main debian box (Athlon 64 3000 2GB RAM) boots up
      in less than 2 minutes with standard debian bootup
      scripts, and it's running:
      Apache 2
      Postgresql 7.4
      Postgresql 8.1
      nfsd
      Samba
      CUPS
      exim4 mail server
      fetchmail daemon
      iptables fw with NAT
      dns proxy/server
      ncp server
      dhclient
      ssh server
      and finally, Xorg system.
      So I think this is not too bad time for booting up.
      What kind of time Windorch needs to _complete_ booting
      up in a similar hardware and this or similar set of services?
      Can somebody point it out?
      Cheers,

    48. Re:Boot time not an issue. by Antique+Geekmeister · · Score: 1

      It's useful for laptops, where every erg of boot-time energy is essentially wasted. Keep an eye on the One Laptop Per Child project for related work. They're actually using a Linux based BIOS.

      It's also useful for micro-Linux implementations, for cell phones, routers, firewalls, etc.

      It's also useful for testing environments and co-location uses where the manpower wasted connecting up and doing boot time operations also costs a lot of work and money.

      And it's fun to show off while tools like Vista are booting in layer after layer of memory gobbling and unwanted "optimizers" and "pre-indexers" that mean the pretty login screen shows up, but you can't do any real work for another few minutes until everything is completely pre-loaded.

    49. Re:Boot time not an issue. by Antique+Geekmeister · · Score: 2, Insightful

      Parallel processing is fine and useful. It is also vastly, vastly, vastly tougher to fully test and support, especially in such a wild and woolly environment as system boot: people have been hand-inserting all sorts of oddnesses in there, and the robustness is a testament to the wisdom of keeping it simple and single-threaded.

    50. Re:Boot time not an issue. by EsbenMoseHansen · · Score: 2, Interesting

      hibernate has to restore the entire state of RAM. It doesn't matter if you had 200MB or a 3 GB of applications, it's all gotta come back in before you start running the kernel.

      Actually, it is only the application areas that has to come in... disk cache on so on doesn't. I sometimes wonder if there should be a "shutdown" signal in the XSMP setup that would enable applications to dump cached data that would probably be stale anyhow when the computer is brought back on line, such the browsing cache.

      --
      Religion is regarded by the common people as true, by the wise as false, and by rulers as useful.
    51. Re:Boot time not an issue. by m50d · · Score: 1
      I don't. I'd much prefer to see an almost-instant-on OS that didn't depend on special hardware tricks but rather because the architects actually designed the bloody thing for a change.

      If you want instant-on switch to DOS. You'll quickly realise why people don't. Much as you mock "glass windows", the cycles aren't being wasted - there are good reasons for most of what a modern OS does when booting.

      --
      I am trolling
    52. Re:Boot time not an issue. by BrokenHalo · · Score: 1

      Seems to me that whenever I boot a Windows box, it takes somewhere around 35 seconds to show up a desktop, then at least that again tipping up its hourglass while I wait for all the spyware and malware to load. ;-)

    53. Re:Boot time not an issue. by Anonymous Coward · · Score: 0

      I thought one of the arguments for linux was that you didn't need to reboot - like you do with Windows. So the boot time should not matter :-)
      You're forgetting that this is Slashdot and most of us live in our parents' basements. So we always turn our computers off during the day, because otherwise the noise keeps us awake.
    54. Re:Boot time not an issue. by peragrin · · Score: 1

      I do the same thing with my powerbook. you put it to sleep and forget. i don't even bother with shutting it down. most of the time I just put it to sleep and travel that way.

      Sleep mode. 10 seconds to wake up and connect to a network, and that's stumbling through which network to connect to part. Your dell is probably similar.

      I spend more time typing in my password to unlock the system than actually waiting for the machine to be available.

      --
      i thought once I was found, but it was only a dream.
    55. Re:Boot time not an issue. by ponos · · Score: 1

      I thought one of the arguments for linux was that you didn't need to reboot - like you do with Windows

      I have to agree with that. If your work session is, say, 4 hours or more, then it wouldn't matter if the boot process takes 1-2 min. Unless your work requires rebooting all the time (which is hard to imagine). Furthermore, as another poster noticed, windows gives the IMPRESSION of booting quickly, but needs another 30+ seconds after you get inside the desktop. Even though the system is theoretically available, windows keeps loading anti-viruses, firewalls etc for a while.

      Don't get me wrong, the idea of improving an aspect of linux is clearly welcome. However, boot time can easily be shortened with a few simple measures (custom kernel, removing unwanted services and libraries) and I don't see the reason why the broad majority of users should care for more elaborate methods.

      P.

    56. Re:Boot time not an issue. by jtev · · Score: 1

      No, reboot means that you start the computer up again. Powering down, and then powering up counts as a reboot. And Linux machines don't need that. In fact they perform better WITHOUT a reboot. And Linux has many nice features for increased productivity. Some of them are available as third party applications for windows. some aren't. I can shade, iconify, or I can use virtual desktops on my linux machine. I also have a wide variety of desktop environments available. I tend to stick to one environment, and one focus setting, but if I wanted to I could change at any time. My computer was recently rebooted for a failed attempt to upgade it. But even though I had a planned reboot recently. my uptime is now just shy of 11 days. Everything is running fine. In fact, it'll be running fine until I have a hardware failure, or another reason to reboot.

      --
      That which is done from love exists beyond good and evil
    57. Re:Boot time not an issue. by Anonymous Coward · · Score: 0

      You hit the nail on the head. I got up this morning. Turned on the monitor and logged in. So what slow boot times. It only gets rebooted every few months of so and as other posters have noted when I do get a login screen the system is up. I don't get a login screen with the hard drive thrashing away. Windows only "appears" to start fast.

    58. Re:Boot time not an issue. by Glonoinha · · Score: 1

      Obviously there is only so much hardware you can throw at a system...

      As someone running a four box cluster of HyperThreaded P4's and dreaming of replacing them with the new 8-way offerings from Intel and AMD ... oh I assure you that's not the case.

      --
      Glonoinha the MebiByte Slayer
    59. Re:Boot time not an issue. by celkin · · Score: 0

      Yes, Linux sounds very adventurous. I may consider it...

      --
      "Oh c'mon, I wumbo, you wumbo, he/she/me...wumbo, wumboed, womboing...wombology? The study of wumbo? It's first grade,
    60. Re:Boot time not an issue. by jtev · · Score: 1

      Well, It was pissing me off. And I had data that I couldn't lose yet that I need to back up. I'm getting tired of some of Red Hat's crap with Fedora. I should be able to run a distro for more than a year. So I guess I'll upgrade to a different Distro in the future. It's still working though. And I have those backups to make, as I get around to them.

      --
      That which is done from love exists beyond good and evil
    61. Re:Boot time not an issue. by HAKdragon · · Score: 1

      Can you imagine how long an XP boot (or Ubuntu for that mater) would take on an old 386 system with a sluggish hard drive and not a lot of memory (if such a thing were even possible)? Hours I would guess, and you would shake your head wondering if your hard drives MTBF would get you through the process.

      Well, there's always Windows XP on an 8Mhz Pentium with 20MB of RAM

      --
      "Our opponent is an alien starship packed with atomic bombs. We have a protractor."
    62. Re:Boot time not an issue. by Deliveranc3 · · Score: 1

      On my laptop hibernate doesn't work, neither does suspend.

      This is a pretty serious boogeyman in Linux's closet if people I know with linux laptops are any indication.

    63. Re:Boot time not an issue. by celkin · · Score: 0

      If you read my parent comments, you'll see that you're kind of off-topic--and yet YOU get a 2 and I have a freaking 0. F*** you, Slashdot. Wired is better. Go ahead, rate me now.

      --
      "Oh c'mon, I wumbo, you wumbo, he/she/me...wumbo, wumboed, womboing...wombology? The study of wumbo? It's first grade,
    64. Re:Boot time not an issue. by billcopc · · Score: 1

      Indexing could be useful, but not everyone wants to throw disk space away if they never use the feature. Hell, I just went through a plain CentOS install for a server machine, and spent 2 more hours ripping out a bunch of crap I didn't need. 2 gigabytes of files for a machine that doesn't even have a web server set up, that's pretty pathetic. Logically it should be possible to have a kernel, web server and PHP/Perl/Python/whatever fit in no more than 256mb, and I'm being generous. But then you have to throw in the sloppy package managers that do fuck all, yet eat up gigabytes of disk space for their caches... I love Gentoo, but do I really need a full portage tree on my system ? Thousands of packages I've never heard of, and will never use. You'd think it could be a lot simpler. Have the emerge/apt/yum script request a package from the server, which also sends back a dependency tree. Client checks the tree, requests any additional packages as needed, and gets to work. Total non-tarball traffic: a couple kb for all but the worst spaghetti packages. No more sync bandwidth wasted, no more crazy disk space usage.

      Hell, MS Windows Update does a weak form of this, they just don't do strong dependency checks. Works for them, and they certainly don't have any reservations about wasting disk space... the light-weight client just works better for them. Why wouldn't it work for Linux ?

      --
      -Billco, Fnarg.com
    65. Re:Boot time not an issue. by danheskett · · Score: 1

      There are two phases to booting Windows. The first gets you to a login prompt (if so equipped, say, for a machine joined to a domain).

      That is a really pretty damn fast considering Windows overall. This is because starting with Windows XP much of the boot process was made to be multi-threaded, meaining the boot process isn't waiting for hardware to init, slow starting services to complete, etc. You can get a login prompt on a typical Windows XP box fast.

      With Vista, they actually did something useful that make sense. Tons of crap is set to load "on login", and that's what causes your thrashing and all that stuff. One is that they properly allowed Windows Explorer (the shell) to threaded startup tasks across multiple cores or CPUs. Previously it wouldn't properly spawn startup threads correctly. But beyond that they have allowed you to mark services that start more intelligently - "delayed auto-start". This type of service basically starts when your system has calmed down. Marking more and more services into this category makes the time to get a usable desktop very minimal. It's a great idea. Do you really need Google to start indexing your files the first opportunity it gets? Nope. And that's why it works good.

      (http://download.microsoft.com/download/9/c/5/9c5b 2167-8017-4bae-9fde-d599bac8184a/Vista_Services.do c#_Toc142993760) has more.

      Frankly its one of the only worthwhile chnages that they maded in Vista.

    66. Re:Boot time not an issue. by Anonymous Coward · · Score: 0

      To be fair, my Windows box boots pretty quick; I think the time between power on and desktop is somewhere in the region of 50 seconds.

      50 seconds "pretty quick"?

      I had my Linux machine down to 15 seconds - 2 for the kernel and 13 for the boot scrips - until my ISP changed me from a static IP to DHCP, waiting for their stupid DHCP server adds another 5-6 seconds to the boot sequence.

      (The next boot scripts after DHCP need the network, so not waiting for DHCP would require a major redesign of the boot scripts. Some day I'll get around to try a dependency-based system...)

    67. Re:Boot time not an issue. by surprise_audit · · Score: 1

      My new work laptop (2GHz Dell D810) seems to bring up Windows pretty fast, but then it spends about 3 minutes "applying security policies", "running startup scripts", etc, before I even get the option to ctrl-alt-del to get the login screen. I have no idea what it's doing in the background, but it seems to have to do it every damn time it starts up... Still, it gives me time to go get coffee. :)

    68. Re:Boot time not an issue. by ehrichweiss · · Score: 1

      Upgrading from Fedora 5 to 6 wasn't a big deal. I changed a few lines in my yum.conf(I think), then it updated me in about an hour; rebooted and I was done. I know it SEEMS like a big deal but if you stick with the net-upgrades, you'll be much happier. Just search for "yum upgrade fc6" and you'll find lots of info that will save you time and energy.

      And as a matter of fact the only thing that kept me from upgrading sooner than I did was that I couldn't find a kernel that was specifically labeled for SMP; once I discovered they don't need a specific SMP kernel it was no problem.

      --
      0x09F911029D74E35BD84156C5635688C0
    69. Re:Boot time not an issue. by jtev · · Score: 1

      Yeah, but you can't upgrade from 4 to 6 without intermediary. And my YUM configuration is a little hozed.

      --
      That which is done from love exists beyond good and evil
    70. Re:Boot time not an issue. by internewt · · Score: 1

      Moving along I've got a few VMware support services, which is dumb considering they're only useful when VMware is actually running.

      Here's a workaround: Set the VMWare services to manual, and use this little script to start and stop everything (though turn off VM's before stopping VMWare).

      C:\Tools>type VMW.cmd
      @echo off

      if {%1}=={up} goto up
      if {%1}=={down} (goto down) else goto usage

      :up

      start /min net start "VMware Authorization Service"
      start /min net start "VMware DHCP Service"
      start /min net start "VMware NAT Service"
      start /min net start "VMware Virtual Mount Manager Extended"
      net start "VMware Registration Service"
      pushd "C:\Program Files\VMware\VMware Server\"
      start vmware.exe
      popd

      goto eof

      :down

      net stop "VMware Registration Service"
      start /min net stop "VMware NAT Service"
      start /min net stop "VMware DHCP Service"
      start /min net stop "VMware Virtual Mount Manager Extended"
      :: sleep 3
      start /min net stop "VMware Authorization Service"
      goto eof

      :usage

      echo This batch file will start or stop VMWare server
      echo Usage: VMW Up
      echo VMW Down

      goto eof

      :eof
      --
      Car analogies break down.
  3. Correct me if I'm wrong by Anonymous Coward · · Score: 0

    It seems to me that Linux checks my system for everything every time it boots. If it could remember the system configuration from time to time, things would speed up a lot. My Gentoo buddies tell me that Gentoo doesn't suffer from that but I still don't want to spend a week compiling my own Gentoo. I, for one, welcome my long Linux boot times 'cause I only boot once or twice a year anyway.

    1. Re:Correct me if I'm wrong by strstrep · · Score: 1

      Compile your own kernel with your hardware options selected as compiled in to the kernel, rather than as modules. Then you won't have to load the same set of kernel modules each time, and you won't need an initrd.

  4. Fixed in Gentoo by Wonko+the+Sane · · Score: 5, Informative

    in /etc/conf.d/rc:

    rc_parallel_startup="yes"

    (actually that should be in caps, but the lameness filter doesn't like it)

    1. Re:Fixed in Gentoo by Ant+P. · · Score: 1

      I used to use that, but less common initscripts like Folding@Home don't like it - sometimes things will just fail to start randomly on boot.

    2. Re:Fixed in Gentoo by Wonko+the+Sane · · Score: 1

      Sometimes you need to manually fix the dependencies.

    3. Re:Fixed in Gentoo by Nasarius · · Score: 1

      Ick. GP should file a bug; there are too many Gentoo devs that need a kick in the ass before they bother to do basic QA.

      --
      LOAD "SIG",8,1
    4. Re:Fixed in Gentoo by Ant+P. · · Score: 1

      I might do that, but my luck tends to be that I file a bug and it gets duped to an unfixed 5 month old one within minutes. Also Bugzilla's search really sucks.

    5. Re:Fixed in Gentoo by Antique+Geekmeister · · Score: 1

      It's not the Bugzilla search: it's the bug authors who cannot be convinced to use a consistent format on submitting bugs. "What's wrong with my system" is not a useful title for a bug.

  5. An Uninformed Question by AaxelB · · Score: 1

    I have very little experience with Linux, but is it actually all that slow in comparison to other OSs? Most of my school's computers will boot in either XP or Redhat, and booting in Linux takes a tiny fraction of the time as XP. Is that a well-optimized version of Redhat? Or is Linux slow just in comparison to OSX (or others)? Or do Linux users just want it to be as fast as humanly possible (A noble quest), and it will never boot soon enough?

    1. Re:An Uninformed Question by mashade · · Score: 1

      It all depends on what's going on at boot time. Linux isn't "slow" to boot in most cases, but some distributions are faster than others, as they perform tasks differently while the system is being brought up.

      It's mostly just a speed freak thing. And I, for one, like it.

      --
      Technology tips and tricks.
    2. Re:An Uninformed Question by Goaway · · Score: 4, Informative

      Comparing fresh Linux and XP and OS X installs, Linux is by far the slowest to boot. If your school's XP installs boot slowly, they are probably doing a lot of work like transferring lots of data over a slow network connection.

    3. Re:An Uninformed Question by CastrTroy · · Score: 3, Interesting

      Or maybe he's just counting the time until the computer is actually usable. With Linux once the computer is booted, and you see the desktop, it's ready to use. With windows, it takes another 20 seconds (at least) after the desktop appears for it to actually become usable.

      --

      Anthropic principle: We see the universe the way it is because if it were different we would not be here to see it.
    4. Re:An Uninformed Question by Anonymous Coward · · Score: 1, Informative

      I defy you to prove that with a Slackware installation. Maybe some shitty "let's load EVERY MODULE" systems are like that. Slackware is not.

    5. Re:An Uninformed Question by Anonymous Coward · · Score: 0

      My computer with XP doesn't take another 20 second when the desktop is shown. This only happens if you have a bunch of things in the startup group.

    6. Re:An Uninformed Question by init100 · · Score: 1

      This only happens if you have a bunch of things in the startup group.

      Like antivirus software, which for me is the primary reason Windows is unusable for a while even after the desktop is shown..

    7. Re:An Uninformed Question by MooUK · · Score: 1

      In some windows labs at my uni, it takes about twice as long to get from the LOGIN SCREEN to a usable desktop as is does to boot my own machine from scratch, including a ten second delay at GRUB to let me dual-boot, right through logging in to getting a usable ubuntu desktop.

      In the closest lab to my room, it even takes most of two minutes to simply start Word.

    8. Re:An Uninformed Question by value_added · · Score: 1

      Or maybe he's just counting the time until the computer is actually usable. With Linux once the computer is booted, and you see the desktop, it's ready to use. With windows, it takes another 20 seconds (at least) after the desktop appears for it to actually become usable.

      The irony, of course, that it's necessary to re-state that last point in the hopes that people might get it. I doubt they will. Microsoft has gone to great lengths to lull the user into thinking the boot process is fast, streamlined and problem free, and given the typical disk thrashing that goes on, it doesn't take a genius to figure out they've also gone to great lengths to "optimize" the process. The end user, when seeing his login screen, or his wallpaper and desktop icons, thinks everything worked and everything is done when nothing could be farther from the truth. There's no real logging of any sort on Windows, so the measure of things remains somewhere between superficial and obscured.

      Personally, I think a discussion of boot times for Linux systems is of interest mainly to Windows users. The rest of us typically don't care because we aren't in the habit of needing to reboot for installations, updates, or to "fix" things. In that respect, the article is as interesting as an article on stripping the kernel for faster boot times; there's lots of reasons to compile a custom kernel, but that one is at the bottom of the list.

      On my BSD laptop (generic kernel), I can boot to a login prompt in about 35 seconds and be sure that that it's ready for use. Resuming from sleep takes about 4 seconds. If things were any faster, I wouldn't have time to take a sip of coffee.

    9. Re:An Uninformed Question by massysett · · Score: 2, Interesting

      Or maybe he's just counting the time until the computer is actually usable. With Linux once the computer is booted, and you see the desktop, it's ready to use. With windows, it takes another 20 seconds (at least) after the desktop appears for it to actually become usable.

      For my machine and GNOME, not true. GNOME goes through its splash screen and shows a desktop, but it is still loading panel applets and the clock. It takes several more seconds. It's hardly usable because loading anything while it's at this stage takes forever.

      I used to think KDE starts slower than GNOME. Now I think it's a dead heat. KDE shows its splash screen and it's there longer, but KDE is almost (if not entirely) done loading when it says "KDE is up and running." GNOME seems to take the Windows approach: show the desktop to fool 'em, even though it is actually still loading.

    10. Re:An Uninformed Question by optimus2861 · · Score: 1
      20 seconds? I wish. Try running some specialized engineering software packages. There's one in particular by Rockwell Software that, as soon as it's been installed on my laptop, adds at least that much time to the "usable desktop" time by itself, as it bloats the registry like you wouldn't believe (it installs information about a raft of automation devices, all into the registry instead of its own database. So Windows of course seems to have to read it all in when it boots. At least I think this is what's happening).

    11. Re:An Uninformed Question by bigpicture · · Score: 1

      I have both Windows XP and Kubuntu on a PC, but they are both on different HDs. I change the BIOS when I want to boot either. But after post there does not seem to be a whole lot of difference, between the boot times of either OS.

      Now if they are on different partitions of the same HD that might be different, because there might be more HD thrashing involved. Also the file system used might be a factor, say if the NT File System is used for Linux then it would probably scan a lot more files on start-up, even on the Windows partition. I use the Reiser 3 file system on the Linux HD, and it seems to completely ignore the XP HD on boot.

    12. Re:An Uninformed Question by Verte · · Score: 0

      Hmm, I don't know. Debian boots so fast on my machine [20 seconds] I never bother with hibernate anymore. XP takes nearly a minute to get to the desktop.

      --
      We at slashdot are scientists, specialists and kernel hackers. Your FUD will be found out.
    13. Re:An Uninformed Question by IdolizingStewie · · Score: 1

      The long delays in labs at your university is probably due to the computer having to pull down your roaming profile.

    14. Re:An Uninformed Question by mackyrae · · Score: 2, Insightful

      I've noticed kernels booting faster over the last 8 months. I started with 2.6.15 and now I'm using 2.6.20. I went from a boot time that was about a minute to the current one which is probably 20-30 seconds. I'll rebo...

      Wow, I love the "restore session" on Firefox. Anyway, I just timed my reboot, and from the time I hit "enter" to choose a distro (multi-boot) and the time GDM finishes loading, it's 23 seconds.

      --
      look! it's a bird, it's a plane, it's....a girl? yes, a girl browsing Slashdot on Linux
    15. Re:An Uninformed Question by grumbel · · Score: 1

      ### and you see the desktop, it's ready to use.

      Not really, my gnome-panel takes a good 20-30 seconds to even react to the first mouse click, till Firefox and friends then start up can take quite a while longer again. Thats definitvly not what I would call instantly usable after boot.

    16. Re:An Uninformed Question by sulfur · · Score: 1

      I think it's because most likely they have a whole bunch of background services installed on their Windows systems, just like our college has. Currently, we have some heavy asset management and remote control clients, an antivirus, a couple of "load time optimizers"/"task managers"(qttask)/update schedulers(jusched) loaded during the startup, which definitely make the system start much slower.

    17. Re:An Uninformed Question by mackyrae · · Score: 1

      Ah, correction, I'm an Ubuntu user, and that means it's using Upstart instead of init. Maybe that's why Sabayon's boot feels sooo sloooowwww (really guys, over a minute?)

      --
      look! it's a bird, it's a plane, it's....a girl? yes, a girl browsing Slashdot on Linux
    18. Re:An Uninformed Question by vtcodger · · Score: 1
      A couple of actual quick and dirty measurements:

      Acer-Power 1000 (Athlon 64 3800+, 512mB memory) -- Windows XP SP2 - 90sec (Probably would be 20sec faster if I had gotten around to killing a lot of the junk loaded at boot)

      Same Acer Power 1000 - Kunbuntu (Unbuntu 4.1.1) - 50sec

      P166 64mb with fully patched Windows 95 - 60 sec

      The Win95 machine has been optimized a bit to speed up boot and has a fairly new 8GB hard drive which is probably faster than the HD in a typical P166. A more typical P166 would probably be 10-20 sec slower. Even then, it would probably be comperable to XP on much faster hardware.

      ===

      Being nice to Microsoft I did not time the other XP machines around here, one of which takes about 200 seconds to boot. (And don't even mention the subject of W2K boot times).

      --
      You can't see ANYTHING from a car, You've got to get out of the goddamned contraption and walk...Edward Abbey
    19. Re:An Uninformed Question by Goaway · · Score: 1

      You will notice, upon re-reading it, that the original claim was for freshly installed systems.

    20. Re:An Uninformed Question by Anonymous Coward · · Score: 0

      Well, I'm running Windows Xp, Vista, and Ubuntu Linux. Start up speeds are as follows:

      Windows XP +/- 45 seconds plus another minute to start the firewall and all that jazz. Total, 2 minutes unuseable.
      Vista 2 minutes to boot, then start the firewall and all that.
      Ubuntu, under 20 seconds and another 10 to be really usable.

      So far, Linux kills the competition by a long shot. Windows boot times are horrible.

    21. Re:An Uninformed Question by toddestan · · Score: 1

      You should also take into account what is installed on the computer. A fresh install of Windows XP is incredibly fast to boot, even on fairly old hardware. Once you've got all the patches in, all the proper drivers, and the software you want to use, it'll boot up considerably slower. On the other hand, a typical Linux distro is going to come with all the software and drivers you need right from the start. Though I will admit that in general, Linux is still slower than XP to boot.

  6. Customize the Bootscript, Trim the Fat by Old+Duck · · Score: 5, Interesting

    Most bootscripts are very generic in that they will try to load all sorts of RAID drivers, various services that are not needed, special fonts, etc.. I've gone in on my computers and wrote a very simple, quick, and to-the-point bootscript (easy to do with a little BASH knowledge), and my system boots up remarkedly fast. Granted, my bootscript isn't very portable, but one of the benefits of Linux is the ability to customize it.

    Another trick is to prelink files and let KDE (if that is what you use) know about it. Even the startkde script can be long and drawn out, so trimming the fat and only including what's needed on your system can make a big difference. I've shaved over 13 seconds off a boot sequence by writing a minimum bootscript for my hardware, and that was using a relatively fast distro to start with.

    -Mike

    --
    There are more things in heaven and earth, Horatio, Than are dreamt of in your philosophy.
    1. Re:Customize the Bootscript, Trim the Fat by Stonent1 · · Score: 1

      A few years back I did this with Gentoo, there was actually an option to initialize things in parallel during boot. It did help the speed a lot but a few things didn't seem to like it due to some dependencies not being initialized before other things came up.

    2. Re:Customize the Bootscript, Trim the Fat by Anonymous Coward · · Score: 0

      I always commented out shit I don't use in the init scripts, stuff like NFS and the old hotplug system. I thought everyone did this?

    3. Re:Customize the Bootscript, Trim the Fat by Wonko+the+Sane · · Score: 4, Informative

      You can fix those yourself. For any service in /etc/init.d you can put a file of the same name in /etc/conf.d. Among other things, that file can list dependencies.

      DEPEND="foo bar"

    4. Re:Customize the Bootscript, Trim the Fat by Anonymous Coward · · Score: 0

      How does this help the average user?

    5. Re:Customize the Bootscript, Trim the Fat by r3m0t · · Score: 1

      He never said it would help the average user. We aren't average users. It helps us.

    6. Re:Customize the Bootscript, Trim the Fat by jank1887 · · Score: 0, Flamebait

      Linux doesn't help the average user...

    7. Re:Customize the Bootscript, Trim the Fat by Eudial · · Score: 1

      Most bootscripts are very generic in that they will try to load all sorts of RAID drivers, various services that are not needed, special fonts, etc.. I've gone in on my computers and wrote a very simple, quick, and to-the-point bootscript (easy to do with a little BASH knowledge), and my system boots up remarkedly fast. Granted, my bootscript isn't very portable, but one of the benefits of Linux is the ability to customize it.

      Another trick is to prelink files and let KDE (if that is what you use) know about it. Even the startkde script can be long and drawn out, so trimming the fat and only including what's needed on your system can make a big difference. I've shaved over 13 seconds off a boot sequence by writing a minimum bootscript for my hardware, and that was using a relatively fast distro to start with.

      -Mike


      Did that once to a slackware setup, just to see just how fast I could get it to boot up. The system, an old-ish laptop, booted in 9 seconds, with 8 seconds of the kernel spewing messages and 1 second to go from init to the login box.
      --
      GAAH! MY PRINTER IS ON FIRE!!! PUT IT OUT! PUT IT OUT!
  7. This used to annoy me by linvir · · Score: 1

    Back when I was starting out with Linux, this used to annoy me and was a factor to consider when weighing up Linux vs Windows.

    Eventually, as I became more proficient in Linux, the longer boot time became less and less of a factor as Windows became less useful for me compared to Linux.

  8. The "bootchart" tool looks promising. by Animats · · Score: 2, Interesting

    The "bootchart" tool mentioned in the article looks promising. But it's mostly unlabeled bars. Until they figure out how to correctly identify all the processes running during boot, it's not too helpful.

    The CPU utilization during booting is much higher than I would have expected. That's interesting, and unexpected. For most of the first ten seconds of post-kernel startup, the system is CPU bound, while the disk is idle more than half the time. Where is all that CPU effort going?

    1. Re:The "bootchart" tool looks promising. by 644bd346996 · · Score: 1

      It's been a while since I analyzed the boot process of one of my Linux systems, but this is what I recall noticing:

      Immediately after kernel startup, there is a lot of hardware probing and configuration going on. On older systems, setting up ISA pnp cards can take quite a while. Along with that, udev uses a lot of cpu time getting things going. Dbus was also one of the slower services to start. Also, if your init scripts are really borked, you can waste several seconds waiting for a dhcp response.

      Also, the bars in a chart are all labeled. The sample in the article was cropped. Check out the bootchart website to see a complete chart.

    2. Re:The "bootchart" tool looks promising. by slamb · · Score: 1

      The "bootchart" tool mentioned in the article looks promising. But it's mostly unlabeled bars. Until they figure out how to correctly identify all the processes running during boot, it's not too helpful.

      As another poster said, the article's image was cropped. bootchart generates one line per process, and it tells you what the process is. It's a really nice tool. It's more aimed at distribution developers than end users, of course - silly to recommend to people that they change their whole system's init system without sharing the changes with everyone else. Neat educational exercise, but it'll break every time you download an updated RPM. Why not collaborate to make it last?

      The CPU utilization during booting is much higher than I would have expected. That's interesting, and unexpected. For most of the first ten seconds of post-kernel startup, the system is CPU bound, while the disk is idle more than half the time. Where is all that CPU effort going?

      I'm not sure since they've cropped it. In the full images, you can drill down more - the bar for each process shows CPU usage over time.

      Anyway, you must mean 5-15 seconds after init starts, not 0-5. I want to know what's happening in 0-5. There's no CPU utilization, no disk utilization. That's wasted time! I see the same thing in the two-year-old example on bootchart's webpage, and I would have expected it to be fixed by now. Is it probing hardware or something? Can't it be doing something else in parallel? Surely there's something that doesn't need the full hardware list to run.

    3. Re:The "bootchart" tool looks promising. by Animats · · Score: 1

      Ah. The charts on the real web site are much more useful.

      Looks like "usb-agent" is using way too much CPU time. The Linux hot-plugging system seems to consume excessive resources. Big jobs like file system recovery and GUI startup aren't the bottleneck. It's probably going to turn out that something simple like parsing the file of hot-pluggable devices and drivers is inefficient.

    4. Re:The "bootchart" tool looks promising. by slamb · · Score: 1

      Keep in mind when looking at the images on bootchart's webpage that they're two years old. The best boot chart for your system is one you've made yourself. The project's about the tool to make charts, not the sample results.

    5. Re:The "bootchart" tool looks promising. by xenocide2 · · Score: 1

      I have no idea where they found that version of bootchart, but there's a .deb that will generate those .pngs and save them in /var/log/. It's a good start at diagnosing how to make things go faster, but once the low hanging fruit (like 4 second system pauses) are gone it's not as informative. You'll have to start examining why a process is IO bound or CPU bound, and whether it makes sense. I think Dave Jones had a journal article about stupid things user programs do like stat the same file 400 times, once per video driver, etc. That analysis is primarily done via strace.

      Their example is pretty crappy -- I recommend gathering a few charts from your own install. It should generate more detailed pictures, and the cutoffs for being represented or not is tweakable (with no cutoffs you get a lot of shell scripts that execute for microseconds). One interesting aspect of newer bootcharts is that they attempt to analyze the difference between disk utilization and throughput. I don't know if the analysis is just wrong but it seems that the throughput only rarely makes it that fast.

      --
      I Browse at +4 Flamebait

      Open Source Sysadmin

    6. Re:The "bootchart" tool looks promising. by Mignon · · Score: 1
      you can waste several seconds waiting for a dhcp response

      In this Linux Journal article Lisa Corsetti describes writing a program to test if the NIC is actually connected, so that if not she could bypass the network initialization and in particular the DHCP request with its requisite timeout.

      Programming ioctl calls is beyond my comfort zone so I took the lazy approach of changing the default 60 second timeout to 10 with a "-t 10" argument to my dhcpd commands in my startup scripts, so instead of a Linux Journal article, I get a Slashdot post.

    7. Re:The "bootchart" tool looks promising. by Anonymous Coward · · Score: 0

      All bars are indeed labeled with the process name. They also change color when they have a high disk or cpu usage.
      Here is a bootchart created this afternoon on my laptop:

      http://www.chauveau-central.net/misc/bootchart1.pn g

      Don't look at the overall time (55s). I was playing with the services and the actual login windows appears around 38s. I also added some dummy sleeping services (S17WAIT and S18WAIT) to speedup the startup of X11/GDM.

      The only thing that bothers me is the reiserfsck that appears to be doing nothing for about 5s.

    8. Re:The "bootchart" tool looks promising. by slamb · · Score: 1

      The only thing that bothers me is the reiserfsck that appears to be doing nothing for about 5s.

      The second one? Yeah, that's weird. Simultaneously, the first fsck has finished, and logsave doesn't reap it until that five seconds is up! logsave's so simple that I couldn't understand how that could happen. I downloaded the source, and it must be in this loop:

      while (!(waitpid(pid, &status, WNOHANG ))) {
      do_read(fds[0]);
      }
      So it will try to reap again as soon as do_read returns. These are the only ways I could see do_read being slow to return:
      • some hardware detection or something has locked your whole system for five seconds, or caused the clock to jump ahead five seconds.
      • the scheduler's keeping it down - not allowing logsave to run for five full seconds after fsck died. (Maybe fsck closed its stdout/stderr significantly before exiting, logsave was busy-looping for a while, and it got penalized for that? Maybe...there is a CPU spike at that time, though the chart doesn't show what process caused it. but five seconds? that's a long time to be on the runnable queue...and nothing else's running. I don't buy it.)
      • something other than fsck still has the pipe open...but since fsck/reiserfsck didn't have any more children and has no code for anything weird like transferring its stdout to another process over a UNIX Domain Socket...that seems unlikely, to say the least
      • it's blocked in send_output, which gets called from do_read. Either on the write to console or (if the destination filesystem was already up before it called fsck, which seems unlikely) to the logfile. (If the destination filesystem wasn't up, it won't try again to open it until after the child's been reaped.) It would have to be the write to console, in fact, or the process's state would be "uninterruptible sleep" rather than normal sleep.

      It would be interesting to get stack traces of the logsave and reiserfsck processes during that mysterious five seconds. Is this reproducible? Seems like there are a bunch of ways you could get those, like OProfile or SystemTap, or even just rigging some gstack shell script to run based on bootchart's instrumentation.

  9. Linux boots under 30 secs. by skynare · · Score: 0

    "One of the biggest complaints about Linux®, particularly from developers, is the speed with which Linux boots."

    Not sure where the author got that from.
    Mine boots less than 30 secs into X window (WindowMaker).
    That's much faster than windows and/or mac.
    Using KDE or Gnome will probably add around 20 secs. But still under 1 min.

  10. How to Pimp Out Linux Booting by rwa2 · · Score: 2, Informative

    Fascinating article, but while searching through aptitude for some of those alternative init engines, I came across bootsplash instead and I couldn't resist!

    http://www.bootsplash.org/

    Uh, yeah, I guess I could make good use of bootchart from the article too... mmm... more eye candy.... and you can keep looking at / admiring your stats / comparing with you friends' stats long after after you've booted up anyway!

    Seriously, real Linux servers don't reboot :P

    (burned by playing with runit some time ago)

    1. Re:How to Pimp Out Linux Booting by henca · · Score: 1
      The fact that Linux servers does not reboot is not only because because Linux is reliable enough not to have to be rebooted every now and then. This is also becuase many Linux distributions like many other unix installations assume that the machine will be running 24/7.

      If you would shut a Linux server down over night it does not only mean its services like NFS exports, web pages, CPU number crunching or whatever will not be avaiable during that time. For many distributions it will also mean that different housekeeping tasks configured by cron jobs to run at night will not be performed. Such housekeeping tasks might be stuff like updating the locate database, cleaning up temporary directories and rotating system logs.

  11. Popular FUD. by twitter · · Score: 2, Insightful

    A common complaint about Linux is the amount of time the operating system takes to start.

    Actually, it's a common insult and FUD. Understanding your boot process is nice and all, but your distribution already does this and has come to reasonable compromises. If you want to tweak with it, more power to you but you won't really save much. With proper power management you don't have to boot at all. For instance, the laptop I'm using says:

    12:47:33 up 65 days, 15:12, 21 users, load average: 1.20, 1.50, 1.61

    I put it to sleep when I'm done and it wakes up when I need it. I can't tell you how much time I saved by not having to reopen all of my applications and remember what I'm working on every day. The price of booting is far greater than the minute or so it takes to get your desktop, it's a loss of placekeeping and continuity. If I were using Windoze, I'd probably have to wipe and reload by now.

    People who complain about long gnu/linux boot times have either not learned how to use their much better systems yet or are FUDing astroturfers.

    --

    Friends don't help friends install M$ junk.

    1. Re:Popular FUD. by zappepcs · · Score: 1

      I was hoping someone would do this. I don't have a Linux laptop yet. The one I use is from work where they insist on Windows XP. If I could get it to perform properly using sleep mode I'd be rather happy. Moving back and forth from VPN to wired connection just upsets XP beyond my ability to sort it out.

      My Linux boxes (and some Solaris boxes, no laptops) all manage to have ridiculous uptimes. Well, they did till the PHB said to reboot everything after DST patching regardless of requirement. We had a FC4 and a 'Redhat 9' system, both with over 18 months uptime. For these servers, boot time is not an issue.

      At home, I can't see much differernce between XP and Fedora boot times. Does anyone know what distro boots quickest? Maybe I will see something to notice.

    2. Re:Popular FUD. by Nicopa · · Score: 4, Insightful

      The world doesn't end just near your nose. You may have a laptop and you may be happy with "hibernating" it, but many of us need to power off PCs. An office PC I power on every day, my home's PC I power on and off when I get and leave home.

      And it's your comment the one that is insulting. You insult lots of experienced Linux users who do care about their machines booting several times slower than an XP pc.

      And why is that? Because Linux boots up with a slow and serialized process, in which the whole system (with hyperthreading, gigs of ram, dual core, etc.) sits idel waiting for a single stupid syslog daemon to start, or worse: waiting for a DHCP client to get an IP address!

    3. Re:Popular FUD. by Silver+Sloth · · Score: 5, Funny

      12:47:33 up 65 days, 15:12, 21 users, load average: 1.20, 1.50, 1.61 21 users on a laptop? Doesn't it get a bit crowded around the keyboard?
      --
      init 11 - for when you need that edge.
    4. Re:Popular FUD. by metamatic · · Score: 2, Informative

      If your PC is so old that it doesn't have a hibernate option, you're in a minority. Most systems have been Energy Star compliant for years.

      --
      GCHQ Quantum Insert installed. If only our tongues were made of glass, how much more careful we would be when we speak
    5. Re:Popular FUD. by cciRRus · · Score: 1

      21 users on a laptop? Doesn't it get a bit crowded around the keyboard?
      Maybe the users are not there physically? Good thing I'm using a secure OS like Windows XP. Whew!
      --
      w00t
    6. Re:Popular FUD. by Idaho · · Score: 3, Insightful

      Actually, it's a common insult and FUD. Understanding your boot process is nice and all, but your distribution already does this and has come to reasonable compromises.


      Denying the problem doesn't make it go away, really. All the hibernation and sleep modes in the world don't change the fact that Linux boot times are much longer than, say, Windows XP's.

      I'm sorry, but I run both Windows and Linux on my Dell Inspiron 6000 laptop, and booting Linux takes much, much longer. We're talking 2-3x longer, in fact.

      If you want to tweak with it, more power to you but you won't really save much. With proper power management you don't have to boot at all. For instance, the laptop I'm using says:


      Yeah, that's nice and all, assuming the power management actually works correctly.

      Which, for many laptops, it doesn't. For instance, some time ago I finally managed to get hibernation mode to work (after a lot of fiddling), but it was still experimental at the time (ca. half a year ago) and would crash on resume sometimes. Not good when you have some important applications still running.

      Now I'm running a different distribution (FC5) and it hangs after resuming from "normal" sleep mode (which is activated when I close the lid).

      So yes, boot time is quite relevant for me, thank you very much, and saying this is FUD is ... well... uninformed, at best.

      People who complain about long gnu/linux boot times have either not learned how to use their much better systems yet or are FUDing astroturfers.


      Or perhaps they are people who get tired of having to spend 2 weekends to get some stupid features, such as sleep mode and hibernation, to work correctly. Which then promptly breaks, of course, on the next kernel upgrade (which, in my case, was needed to get wireless to work).

      This is also why I won't bother with trying to speed up the boot-time of my laptop, I'm waiting for some decent distro which does it for me!
      --
      Every expression is true, for a given value of 'true'
    7. Re:Popular FUD. by gottabeme · · Score: 1

      Lots of system processes run as different users...on Debian, anyway.

      --
      "Those who consume the bulk of goods are those who make them. We must never forget this secret of our prosperity."
    8. Re:Popular FUD. by Anonymous Coward · · Score: 0

      Yup every linux, BSD I can optimize to boot in 30 seconds with full support for everything I need. And I can say the same about Windows. And as far as I can tell most of the people on this thread bitching about Window's startup time haven't even bothered to try to tweak it at all if takes them like 50 seconds! lol

    9. Re:Popular FUD. by izomiac · · Score: 1

      AFAIK hibernate is entirely done in the OS. I.e. instead of booting normally it partially boots then retrieves the contents of RAM from disk. Sleep requires that hardware is put in a low power state and such. Personal example: my Mom's computer came preinstalled with Windows 98, so obviously it shouldn't have any hardware support for hibernation, but it still manages to hibernate with Windows XP just fine.

    10. Re:Popular FUD. by izomiac · · Score: 2, Interesting

      How about duel boot laptops? After all, Linux's hardware support for laptops isn't that great, and many people still have things they need (or prefer) to do in Windows or some other OS. Hibernation doesn't remount partitions, so guess what happens if you hibernate, boot another OS, edit a file, then resume?

      If boot time didn't matter then people wouldn't complain about it. You could hibernate all the time, but hibernation isn't perfect (or safe in all situations), and some people just like to start with a "clean slate" in the morning. Boot time is actually one of the big things that keeps me from using Linux very often. When I wake up in the morning and want to check my RSS feeds I have three options. First I turn on my laptop (which is off to conserve energy, prevent my P4 from acting like a space heater for my dorm room, keep the room quiet, and cut down on EM interference that keeps my "atomic" clocks from syncing), and get to my boot manager. If I choose BeOS (my primary OS) I can wait 30 seconds for the OS to load (since I'm stuck with 1 MB/sec disk read speed, normally it would be ~7 seconds), immediately launch Firefox in ~5 seconds (same problem) and I'm done. If I choose Windows I wait about 40 seconds for my desktop, and about 20 seconds for Windows to finish so I can launch Firefox. With Linux I have to wait 2 - 3 minutes for it to boot and I can launch Firefox immediately. Hibernation wouldn't work since I use one Firefox profile on a shared drive.

      So, even though hibernation works for you, it's not a valid excuse for an unnecessarily long boot time. (Oh, and if you're wondering why I still use the BeOS despite crappy hardware support, it's because it tends to be an order of magnitude faster for opening folders, e-mail, and non-ported applications. I get pissed off if I have to wait for a program and there isn't a hardware or network bottleneck that's causing the delay.)

    11. Re:Popular FUD. by Anonymous Coward · · Score: 0

      You mean I am a FUDing Astroturfer because my laptop cannot suspend or hibernate in Linux ? Or I am plain dumb. Thanks.

    12. Re:Popular FUD. by sulfur · · Score: 1

      Yes, but they do not show up in "logged in users". You need to interactively login (via tty/ssh/xterm) to be counted.

    13. Re:Popular FUD. by zsau · · Score: 2, Insightful

      How about for those plenty of us for whom hibernate under GNU/Linux simply doesn't work? On my laptop it works about one in three times depending on the phase of the moon; on my desktop it never works. (This is irritating, because there's some program—probably soundcard driver—on my desktop which is buggy and causes the computer to hang for five minutes or until I press Ctrl-C during boot. I'd kill to get around that.)

      --
      Look out!
    14. Re:Popular FUD. by Anonymous Coward · · Score: 0

      I'm not sure WinXP and FC5 is a valid comparison - Fedora is well on the bloated side; From past experience I know that it (and Ubuntu) love to start every installed service (sshd, ftpd, httpd, etc.) at every boot, without failure. A more valid comparison would be WinXP and a Red Hat distribution released in the same year, or Fedora Core 6 and Windows Vista. In addition, how long after the boot do all of the applications finish loading?

      Running just the standard sysvinit on Archlinux gets me up, logged in, and with my default startup applications loaded from a cold boot in about 35 seconds (including a 5 second grace period for USB settling) on my old 1.3GHz system with 128MB of RAM, long before my 1.8GHz WinXP box finishes loading. I wonder if the amount of default daemons and services started in other distributions is actually the cause of this whole problem, instead of the init system.

    15. Re:Popular FUD. by metamatic · · Score: 1

      I've found that hibernate is more reliable when not using buggy closed-source video drivers.

      --
      GCHQ Quantum Insert installed. If only our tongues were made of glass, how much more careful we would be when we speak
    16. Re:Popular FUD. by Luyseyal · · Score: 1

      Gnome Terminal ups that count.
      -l

      --
      Help cure AIDS, cancer, and more. Donate your unused computer time to worldcommunitygrid.org. Join Team Slashdot!
    17. Re:Popular FUD. by zsau · · Score: 1

      My desktop is a PowerPC with an nVidia card so it can't use buggy proprietary drivers; my laptop has Intel graphics so I could get well-supported free drivers. Would that it were simple as that.

      (The latest Ubuntu Feisty beta seems to work on my laptop a lot better, though.)

      --
      Look out!
    18. Re:Popular FUD. by Anonymous Coward · · Score: 0

      Here's my pissing contest entry:

      [seguin@helios voevent]$ uptime
        16:32:53 up 834 days, 5:09, 15 users, load average: 0.06, 0.07, 0.01

  12. How about 1.1 seconds to the bash shell? by dattaway · · Score: 2, Interesting

    If it can be done on a 200MHz ARM9 processor in just over a second, anyone else can do it:

    http://www.embeddedarm.com/epc/ts7400-spec-h.htm

    1. Re:How about 1.1 seconds to the bash shell? by 644bd346996 · · Score: 3, Insightful

      That ARM board stores the entire operating system in flash. It uses busybox and pretty much nothing else, to get a shell up that fast. It doesn't have to wait for any hard drives or peripherals to initalize. LinuxBIOS can do similar things, but only on some machines. TFA is all about getting services to start quickly. Turning off everything is not an acceptable alternative.

    2. Re:How about 1.1 seconds to the bash shell? by gradedcheese · · Score: 1

      I have one of these boards sitting right here, their claim is true. However you should see the Linux it boots into in 1.1-1.3 seconds. Needless to say, it's extremely stripped down and it can't do a whole lot, but it's great for adding just the embedded software you need and having a real usable product. Boot that board off an SD card with Debian and you'll see a more typical boot time.

  13. Ubuntu by Anonymous Coward · · Score: 0

    No need, Ubuntu has upstart by default.

    1. Re:Ubuntu by delvsional · · Score: 1

      in both my newly installed feisty and the old edgy my computer hangs during boot for about 10 minutes, unless I disable apci. then it boots up in about 45 seconds. Noone has been able to tell me so far what to look for. and yes I've looked a dmesg I have the line that it hangs at.

      --
      Oh Crap, I'm an optimist.....
    2. Re:Ubuntu by Anonymous Coward · · Score: 0

      Show us that line! * and a few around it for context, don't just tease us or is this one of those "I have an elegant proof but this margin is too small to contain it..." type comments?

  14. Applications suck. by Anonymous Coward · · Score: 0

    The applications which start on boot really suck - many of them parse XML files over and over again, some of them try and exec a process 18 times before they finally succeed in finding the right path, some others do fcntl64 on 0 to ulimit -n File descriptors even if none of them are valid, Xorg finds PCI bus and devices in utmost inefficient ways and module loading subsystem loads dozens of useless modules).

    If they fixed all of this, the boot times would not be so good. Given that suspend/resume won't work reliably Linux needs to have fastest boot times - it's a pleasure to use OSX - boots fast and rarely needs to be rebooted - just close the lid and open it when you need.

  15. Ubuntu already uses Upstart by wile_e_wonka · · Score: 4, Informative

    When I went out to take advantage of the article's suggestions, I found that, according to this thread in the Ubuntu Forums that Ubuntu 6.10 already uses Upstart.

    I did used the "profile" command in my bootup once, after reading about it in another article recently. So I guess my Ubuntu is booting about as fast as it can (unless I have useless processes starting up. I'm still trying to figure that out--I'm a noobie).

    1. Re:Ubuntu already uses Upstart by makomk · · Score: 1

      Ubuntu does already use Upstart, but it's basically set up to imitate the aging traditional init system, so I doubt you'll get much of a speed up.

    2. Re:Ubuntu already uses Upstart by stevied · · Score: 1

      According to this, feisty will switch fully to the new system, but I haven't mention of it in more recent release notes ...

    3. Re:Ubuntu already uses Upstart by wile_e_wonka · · Score: 1

      That's very interesting--is there a way to "turn on" the benefits of Upstart in 6.10? I haven't seen anything in the Forums.

      As for it not being listed in the release notes, I'm hoping that maybe they just thought it was too obvious to note; or maybe the newer release notes are just an addition to the earlier releases (so, Herd 2 would mention the stuff that was not in Herd 1).

      I say this because I also notice the Beta release notes don't say what version of the kernel is in use for Fiesty, whereas Herd 1 said .19 was being used, with plans to use .20, along with ext4(?!).

    4. Re:Ubuntu already uses Upstart by Anonymous Coward · · Score: 1, Interesting

      The full implementation of upstart isn't in Ubuntu yet- It was delayed until Feisty +1. So yes, upstart can still be much faster than it is now.

      See http://ubuntuforums.org/showpost.php?p=2264992&pos tcount=8

    5. Re:Ubuntu already uses Upstart by stevied · · Score: 1

      The feature spec notes this as well. Might be a good page for monitoring ...

    6. Re:Ubuntu already uses Upstart by mhall119 · · Score: 1

      To "turn on" the upstart benefits in 6.10 you need to convert your init.d scripts over to upstart scripts. There isn't any automated method to do this as far as I'm aware. I think that has been the biggest delay in converting fully to upstart, just getting all the scripts converted with the right dependencies and what not. Feisty does seem to have more in upstart than Edgy did, I see a noticeably faster boot (I haven't timed it though).

      --
      http://www.mhall119.com
    7. Re:Ubuntu already uses Upstart by benow · · Score: 1

      35s boot for me under Feisty (amd4800+x2, 40G wd sata2 raptor). Tasks seem quite parallelized. I'm happy with it.

    8. Re:Ubuntu already uses Upstart by benow · · Score: 1

      It should be noted that this timing was also done after booting with the 'profile' kernel flag, which re-arranges load order for file prefetch. Prefetch is done during early boot to bring in as much of that will be required into cache during the rest of the boot process. Over time things drift on the disk and the distro fetch order becomes stale leading to head thrashing and lost seconds. Profile recalculates the read order and can speed up a used system. Of course, profile should only be used infrequently after large changes, as it adds time to the boot process.

  16. It's not a big deal by Cracked+Pottery · · Score: 1

    I managed to make Windows 98 go weeks between a reboot. Avoided the KAK worm because I saw it the Startup menu before a reboot. If standby and hibernate work, it's even more meaningless. I wish it were easier to get the video subsystem up to snuff with modern adapters and make the various proprietary web plug-ins work under Linux. For the most part it's doable, but not with always with a consistent method.

  17. Debian by Anonymous Coward · · Score: 0

    I installed debian on a rackmount raid box (configured the raid during install), and it had a boot time of maybe 30 seconds, 60 max. Not sure what other installs do, but for debian the gui takes another 10 seconds max on top of that, if you choose to install it. Now we're at 40-70 seconds for a full boot on a box that will be up for months...

  18. Some tips for speeding up Linux by itz2000 · · Score: 1, Insightful

    a. don't load all the modules on startup, there's no need! load only modules that you actually use!
    b. don't DHCPCD every startup
    c. don't run xserver on startup, simply type startx, there are lot's of times that you don't even need gui so there's no need to load it
    d. run e2fsck if it makes problems and writing that there are problems with certain inodes every boot.
    e. don't load on xserver everything you don't need! for example : you don't need gaim to be on your startup every time you open kde.
    f. load direct rending driver so your gui will load (and play) faster. (I had some problems with it cause I've forgotten to load amd64-agp, I hope some people can gain something from this post and actually load it). +5 insightful.

    1. Re:Some tips for speeding up Linux by fireboy1919 · · Score: 1

      a. We shouldn't have to fiddle with this when it comes to our devices. A quick check of PCI, USB, and IDE controllers should tell us what we have and therefore need to load. The only exception to this is RAID controllers...but loading extra ones of those takes almost no time at all.
          Any driver not tied to a device will load extremely fast, so that shouldn't be an issue either. It means that we're talking about less than a second worth of boot time when we've already got minutes of boottime hanging over our heads.
      b. Less than a tenth of a second if you've got your system behind a router.
      c. On a good boot system, this happens close to last, and it doesn't affect the running of other systems - i.e. it's already parallelized.
      d. Are you trying to make it longer rather than shorter? The filesystems we have are designed to have difficulty getting corrupted. The time to do this *far* exceeds the boot time. And it shouldn't be necessary most of the time. I have mine set to check once every 40 boots.
      e. KDE comes with *lots* of bloat that you don't need and can't easily turn off because its part of KDE. Unlike GAIM, which adds less than a second to the start time of KDE, starting the hardware manager, the window manager, the sound manager, etc. takes a looonnng time.
      The real solution is to just not use the big WMs that come with lots of stuff. Use FVWM, or IceWM, or something like that. Then it won't be a problem if you start up your favorite programs at bootup.
      f. While this does make the system more responsive, it has no effect whatsoever on bootup time.

      You didn't address at all the fact that people using linux often have *services* running on their computer that can be started in parallel (for example, I don't mind if apache starts the same time as tomcat). Solving this problem is by far more important than any of the things you've mentioned.

      --
      Mod me down and I will become more powerful than you can possibly imagine!
    2. Re:Some tips for speeding up Linux by zCyl · · Score: 1

      b. don't DHCPCD every startup

      I usually tweak the networking boot script to background the dhcpcd call, since this is such a time consuming part (particularly if you are not plugged into the network and have to wait for it to timeout). But a truly parallelized solution would of course be preferable.
  19. launchd by spoonist · · Score: 1

    OS X went to launchd which is an init.d, cron, inetd, etc replacement. I'd like to see how launchd compares to upstart.

    1. Re:launchd by spoonist · · Score: 2, Interesting

      Replying to my own post? Nice. I found this comparison between upstart and launchd. Long story short, launchd isn't event driven.

    2. Re:launchd by 93+Escort+Wagon · · Score: 1

      The original problem with launchd - from a Linux perspective - was the licensing. Now that Apple has changed the license for it, is anyone working on bringing it to Linux or the BSDs?

      Having been both a Linux user and now a Mac head (and having used a Mac prior to launchd's advent), I think the other *nixes should at least be looking at it because it sure seems like a significant improvement boot-wise. But I suppose purists will just say "not GPL, no thank you", and then go back to complaining how Linux on the desktop still isn't gaining any traction.

      --
      #DeleteChrome
    3. Re:launchd by makomk · · Score: 1

      launchd has no dependency support, which means you have to modify every service you want to use with it to wait for its dependencies to start up. Not fun.

  20. On Kubuntu.... by disturbedite · · Score: 1

    i've been using kubuntu for about a year (testing feisty beta now) and for quite a while now i have noticed a problem that this article might help with. i don't have to reboot very often, a heck of a lot less than on windows, but when i do, it takes about two minutes. doesn't sound like much, but that is about 2-3x longer than before. i was told that you can press F2 after you choose your grub entry to see the progress and what it going on, but that doesn't work for me. i press F2 but the screen just remains black for about two minutes then it gets to graphical login. does anyone know if that is an ubuntu/kubuntu "feature" or not?

    --
    http://www.ronpaul2008.com/ Ron Paul for President 2008 http://www.infowars.com/
    1. Re:On Kubuntu.... by Anonymous Coward · · Score: 1, Informative

      i'm going to go out on a limb and say that you have at least one external hard disk plugged in. if you would boot next time without your external(s) plugged in, your monitor should neither go black nor take near as long to boot.

    2. Re:On Kubuntu.... by disturbedite · · Score: 1

      unfortunately, i do not have an external hdd. and just to clarify, after grub starts loading the kernel and what not is when the screens goes blank for ~2 minutes till graphical login. as i said before, i do not know if the screen going blank between initial kernel load and graphical login is a (k)ubuntu feature or if there is something wrong. apparently, as i was told, you should be able to press F2 to see the loading process, but nothing happens when i try that...

      --
      http://www.ronpaul2008.com/ Ron Paul for President 2008 http://www.infowars.com/
    3. Re:On Kubuntu.... by delvsional · · Score: 1

      go to /boot/grub/menu.lst and find the kernel line that's not commented out. delete splash and quiet off the end and save. reboot. ta da you can see most of what's going on. i don't know why they don't call it grub.conf.

      --
      Oh Crap, I'm an optimist.....
    4. Re:On Kubuntu.... by disturbedite · · Score: 1

      oh sweet. thank you! i will try it.

      --
      http://www.ronpaul2008.com/ Ron Paul for President 2008 http://www.infowars.com/
  21. Kernel modules by harry666t · · Score: 1

    The easiest way to speed up the booting is to throw away kernel modules you don't use and to statically compile things you use often.

    Do a 'lsmod | grep -v Module | sort > modules.txt' and keep that list when selecting things to be compiled in. Then grep through your .config looking for things you can't find in menuconfig, and change stuff from "m" to either "y" or "n".

    I've managed to gain at least 10 seconds this way (unsure about the exact time, but the box is starting noticeably faster).

    AFAIK optimizing the kernel image for size doesn't help much.

    1. Re:Kernel modules by Anonymous Coward · · Score: 0

      No you don't need to use a monolithic kernel to speed up boot. You just need to not load the modules you don't use on boot. For BSD type of inits tweak /etc/rc.d/. Just go chmod -x what you don't want to load.

      It's the biggest myth from linux ricers that you need to recompile your kernel to increase speed. Let me explain how a modular kernel works-- the modules are only loaded when they're needed. You can have as many modules as you want and they won't hurt your performance. You should only recompile your kernel if (a) it lacks hardware support or a feature you need, (b) you want to harden it, i.e. patch it.

  22. Wait, I thought Linux wasn't "bloatware"? by Anonymous Coward · · Score: 0

    Huh. I must have read the article wrong. I thought it was *Windows* that took too long to boot, was a memory hog, crashed all the time, blah blah blah. Now you're telling me that it's actually *Linux* that plays the snail in the 0-to-useable time trials? And that if you diddle around enough you can maybe speed that up by a few percent and still have a semi-stable boot process?

    Sorry Charlie, these facts simply don't agree with my dogma. Now if you'll excuse me, it's spring, and I have to get back to watching my lawn grow whilst I wait for my Linux box to boot.

    1. Re:Wait, I thought Linux wasn't "bloatware"? by Anonymous Coward · · Score: 0

      You know, the only one you are trying to convince with that argument is yourself. Seriously. I can build a linux box from scratch (from all source), have a customized beryl desktop and all the eyecandy and get it to boot to a login 5 to 10 seconds quicker than even XP. Or, I could just strip processes from init.d on my Kubuntu box for nearly the same effect. Now, can you do that with windows?

  23. Re: Boot time not an issue on my Mac by Anonymous Coward · · Score: 1, Interesting

    Ma notebooks now use Safe Sleep, which is kind of a hibernate/standby hybrid. When you put it to sleep, it writes the RAM contents to disk as if it was hibernating, and that can take a few seconds depending on how much RAM is in use. Unlike hibernation, when you wake, it wakes instantly because it doesn't go to the RAM image on disk unless there has been a power failure during sleep. It just resumes from whatever was in RAM at sleepytime.

    This best-of-both-worlds approach actually makes it possible to change the notebook battery without an internal backup battery, and yet not lose your session. When you resume after a power cut, it loads the RAM image from disk and all your apps are as you left them running.

    But if I do have to do a cold boot, my Mac Pro reaches the desktop in 15 seconds. I am not a Linux user but I certainly hope that Linux has optimized startup with caches by now. (If I make a major change to the system, such as an update or add/remove kernel extension, startup takes much longer as the caches need to be rebuilt.)

  24. How fast do you need it? by janestarz · · Score: 1

    I'm very interested to try this...how many minutes can I get below one minute of booting?
    Seriously, between pressing the button and logging in there's only 60 seconds of boot. Should be enough time for any girl to get her morning coffee fix, I should think.

    1. Re:How fast do you need it? by Anonymous Coward · · Score: 0

      You can go only one minute below one minute. Any lower than that and you would be going backwards in time.

  25. Alas, I'm all alone. by twitter · · Score: 1

    21 users on a laptop? Doesn't it get a bit crowded around the keyboard?

    I use a lot of Gnome terminals.

    I log into it myself from other machines from time to time but dhcp networking makes sharing it with other people impractical.

    --

    Friends don't help friends install M$ junk.

  26. I was an initng user... by 26reverse · · Score: 1

    But it just seemed that the support was extremely lacking. I know, I know - if it doesn't do what you want, code it yourself. initng, however, suffers from redesign issues (like moving startup scripts into and out of the main distribution - at one point I'd upgraded to the next version, but couldn't log in because none of the startup scripts could be found because they'd been moved into a separate package).

    It's nice, and all. But various inconsistencies have made me move back to init.d. Hopefully this kind of publicity will get more people involved and make it more stable. It really does speed up the boot time (not that I ever thought Linux was slow either).

  27. Lots of always-on selfishness by linvir · · Score: 3, Interesting

    Anybody else appalled by the amount of people saying "just don't turn it off"?. I know I am. It's shocking for various reasons:

    1. It's an unnecessary, unethical waste of energy.
    2. It's an unnecessary security risk to leave so much of your sensitive data connected to the internet completely unattended.
    3. If you do get compromised, there is then the ethical issue of the amount of spam/DDoS your box can be used for by the time you realise what has happened.
    4. It's a poor solution to the problem of long boot times. A better solution would be to man up and learn some fucking patience.
    5. It uses up your machine's useful lifespan much more quickly at no significant gain to you.
    6. It's yet another electrical appliance always on and always ready to set off your smoke alarm or even start an electrical fire.

    If you leave your computer on 24/7 just for convenience, then perhaps you ought to consider the possibility that you are a spoilt rich selfish pussy and not the infallible sysadmin you undoubtedly believe yourself to be.

    1. Re:Lots of always-on selfishness by Aliencow · · Score: 2, Interesting

      1. It's an unnecessary, unethical waste of energy.
      I pay for my hydroelectricity thank you very much. I use CFLs and wash my clothes with cold water.

            2. It's an unnecessary security risk to leave so much of your sensitive data connected to the internet completely unattended.

      Oh yeah, if my machine is easily hackable, leaving it on 12 hours a day instead of 24 will make a huge difference.

            3. If you do get compromised, there is then the ethical issue of the amount of spam/DDoS your box can be used for by the time you realise what has happened.

      I haven't. And if I did, I'd notice, and if I didn't notice, my ISP would bill me for the bandwidth and then I'd notice pretty fast.
            4. It's a poor solution to the problem of long boot times. A better solution would be to man up and learn some fucking patience.
      It's not a solution to long boot time. It's a solution to keeping it online to download stuff, monitor stuff..
            5. It uses up your machine's useful lifespan much more quickly at no significant gain to you.
      I'm not sure...warming up and cooling down must not be all that good for the components.

            6. It's yet another electrical appliance always on and always ready to set off your smoke alarm or even start an electrical fire.
      That is one reason I never leave my laptop on alone though.

    2. Re:Lots of always-on selfishness by PoesRaven · · Score: 1

      Actually, rebooting your computer uses up its lifespan. The booting of a machine is far more stressful than leaving it on, so if you shut it off and turn it on every day (or worse, several times a day) it really will last for a shorter period of time than just leaving it on 24/7.

    3. Re:Lots of always-on selfishness by Khaed · · Score: 2, Insightful

      It's an unnecessary, unethical waste of energy.

      So what? With my monitor turned off, my desktop uses a lot less energy. I'm not running a giant 500w power supply though. It's no worse than leaving a few lights on. Also, these are Linux boxes were're talking about -- they're hardly sitting around idle. (Mine is generally running MythTV, downloading one thing or another, etc.)

      I like having it do things that I don't want to sit there and wait on.

      I don't think security is such a big issue it deserves two points on your list. Again, this is Linux and most people using it are going to make sure there's some extra security; we're not the average user.

      It's a poor solution to the problem of long boot times.

      That point, I'll concede totally. Leaving your box online just so you don't have to boot it up is a really lame way to deal with it. That said, my Ubuntu install takes roughly the same amount of time as XP, both with no tweaking to the boot process of either.

      It uses up your machine's useful lifespan much more quickly at no significant gain to you.
      It's yet another electrical appliance always on and always ready to set off your smoke alarm or even start an electrical fire.

      The first -- yeah. Definitely. But I'm not sure how long my computer is going to last anyway.

      The second -- I'd not considered that. It's a good point.

      As for the just for convenience thing and the slams -- Linux users aren't the only ones. And it's not just computers. How many people have lots of lights they leave on? A half dozen 60w bulbs burn a lot of energy (and put a lot of heat that has to be cooled off with AC).

    4. Re:Lots of always-on selfishness by jeorgen · · Score: 1

      It's an unnecessary, unethical waste of energy.
      So, raise the price of (fossil fuel) energy. Let individuals make their own decisions. Besides, hard disks last longer if in continuous operation, it is the spinning up below working temperature that kills them. You have to account for the energy needed to make a new hard disk (and other environmental load from making it and dispensing with the old one).

      It uses up your machine's useful lifespan much more quickly at no significant gain to you.
      No keeping it on prolongs the life span of the machine, see above. Capacitors may wear out quicker though.

      It's a poor solution to the problem of long boot times. A better solution would be to man up and learn some fucking patience.
      Are you trolling? What else should we man up about? Car queues, malaria? Boot times kills productivity. It is a nuisance and as long as we believe most people actually do something useful at work, it hinders them from doing it efficiently.

      spoilt rich selfish pussy
      Ok, you are a troll, and an angry one too.
    5. Re:Lots of always-on selfishness by kripkenstein · · Score: 1

      Fair enough, but it isn't all one way. Concerning #5 ("It uses up your machine's useful lifespan much more quickly at no significant gain to you."), shutting down and booting up may shorten lifespan faster than keeping it on - there is much more stress during those two operations than constant idle for several hours. Also, keeping your parts at a constant temperature may be better in some cases than cooling them down (when off) and re-heating them.

      But I admit that there are plenty of arguments the other way as well.

    6. Re:Lots of always-on selfishness by Anonymous Coward · · Score: 0

      What, do I need to buy some carbon offsets for my computer now? :-)
      Seriously, I leave my machine on 24x7x365...and I will continue to. As noted, better for the lifespan. But mostly it is convenient.

      It IS a solution to slow booting. If you don't have to boot very often, you're saving the time and the whole premise of the problem goes away. Done.

      Spoiled (I spelled it correctly BTW), rich selfish pussy? No, but I'm working on the spoiled rich and selfish part as best I can.

    7. Re:Lots of always-on selfishness by ceoyoyo · · Score: 1

      You're assuming the machine is on, running flat out, doing nothing useful.

      My two Macs never get shut off. But they sleep, addressing 1, 4, 5 and 6. I suppose you could still argue 1 if you're the kind of person who physically unplugs his television, radio and oven when not using them.

      As for 2 and 3, I think someone else already picked those two up.

      If you don't have a computer that can sleep reliably then perhaps you ought to consider the possibility that you're a cheap bastard who doesn't care about kittens, butterflies and trees. Or a Windows user.

      (Note: the above paragraph is satire)

    8. Re:Lots of always-on selfishness by Anonymous Coward · · Score: 0

      Fully agree ! Well said.

    9. Re:Lots of always-on selfishness by toddestan · · Score: 1

      This topic comes up from time to time. My experience is that computers that are on 24/7 have more problems than computers that run for only part of the day. I've found that they tend to blow power supplies and motherboards a lot faster, probably due capacitors, which generally have lifetimes measured in the thousands of hours, but could also be because of tin whiskers or simply dust build up. In addition, mechanical parts like the harddrives and fans are wearing out anytime they are on. While I do realize that things that get hot like the CPU/GPU/Northbridge don't like hot-cold cycles and would rather just be on all the time, I've also found those components to be extremely reliable in both cases.

    10. Re:Lots of always-on selfishness by Simon+(S2) · · Score: 1

      1. It's an unnecessary, unethical waste of energy.
      I pay for my hydroelectricity thank you very much. I use CFLs and wash my clothes with cold water.

      Even if you pay for it does not mean it is ethical to waste it.

      2. It's an unnecessary security risk to leave so much of your sensitive data connected to the internet completely unattended.
      Oh yeah, if my machine is easily hackable, leaving it on 12 hours a day instead of 24 will make a huge difference.

      The security risk is not leaving the data unattended, but if you never reboot, you never apply patches. That's the security risk.

      3. If you do get compromised, there is then the ethical issue of the amount of spam/DDoS your box can be used for by the time you realise what has happened.
      I haven't. And if I did, I'd notice, and if I didn't notice, my ISP would bill me for the bandwidth and then I'd notice pretty fast.

      All rhight. That would be 1 month after your box was used to serve child porn and as a mail relay.

      6. It's yet another electrical appliance always on and always ready to set off your smoke alarm or even start an electrical fire.
      That is one reason I never leave my laptop on alone though.

      good.
      --
      I just don't trust anything that bleeds for five days and doesn't die.
    11. Re:Lots of always-on selfishness by Aliencow · · Score: 1

      Who said I didn't reboot? I said I leave my computer on. Of course I keep it up to date and reboot if necessary.

  28. What's the problem? by pilsner.urquell · · Score: 1

    Entertain yourself or multi-task, go make a ham sandwich.

  29. No problems for the most part. by FunWithKnives · · Score: 2, Interesting

    I am running Kubuntu 6.10, and I have to say that I have been pleasantly surprised with Upstart, for the most part. It really is much quicker than the older boot sequences, as well as Windows XP, in my personal tests. I still have a couple qualms, however.

    The lack of a method to switch back to the traditional, detailed boot sequence is annoying. Editing /boot/grub/menu.lst and commenting out "quiet splash" takes care of it, but I don't always want to perform a status check when booting. This is a really simple thing to provide accomodation for, and I may just be unaware of a more simple solution that is already in place. If anyone knows of one, please enlighten me.

    The other problem comes as a result of the first. There is, with no option to switch to a detailed boot sequence, also no way to skip network detection. I am used to just issuing 'Ctrl+C' to skip it, and so it is quite frustrating sometimes. If you aren't connected to a network, or if you connect only after you have brought the system up, you are stuck waiting however long the default timeout is, unless you, again, perform a manual edit. This wouldn't be so much of an issue if arguments could be passed by default at boot-time, but as far as I know they cannot.

    Once again, if anyone knows of solutions to these issues that I am ignorant of, I'd be grateful to hear them. Other than what I have related, though, K/X/Ubuntu 6.10 boots very quickly, and I am quite happy with it.

    --
    "We may face a scorched and lifeless earth, but they're accountable to their shareholders first."
    1. Re:No problems for the most part. by Anonymous Coward · · Score: 0

      You can edit the line directly from GRUB, when you need to make a check for something. just press E for edit, and B for boot when you're satisfied. The edit is volatile, next reboot the line will be back to the entry in menu.lst.

    2. Re:No problems for the most part. by phsdv · · Score: 1
      Once again, if anyone knows of solutions to these issues that I am ignorant of, I'd be grateful to hear them. Other than what I have related, though, K/X/Ubuntu 6.10 boots very quickly, and I am quite happy with it.

      And as soon as ubuntu will switch to real upstart scripts, and is not using the old sysvinit scripts in upstart compatible mode as it is doing now, you will have an even faster booting machine.

  30. No it isn't. by caluml · · Score: 2, Insightful

    A common complaint about Linux is the amount of time the operating system takes to start.

    No it isn't. Of all the things I've heard people complaining about Linux about, the start up time isn't one of them.

  31. Seriously, what about Windows booting by Proudrooster · · Score: 1, Interesting

    I know this is off-thread, but is there anyway to speed up Windows booting? At least with Linux you can tune and serialize or parallelize the boot process for speed. Is there anyway to do this with Windows? The 35 processes that launch out of my system tray just don't want play nice together on boot :)

    Sorry for the offtopic, but this is a sincere post.

    1. Re:Seriously, what about Windows booting by dvice_null · · Score: 1

      > I know this is off-thread, but is there anyway to speed up Windows booting?

      Reinstalling helps for a while, but downgrading to older version of Windows seems to be the only solution. (You have to reinstall the older version also once in a while or it doesn't help either.)

    2. Re:Seriously, what about Windows booting by Anonymous Coward · · Score: 0

      The next time I M2, whoever moderated this offtopic will get the boot. Be sure of that. Hardly offtopic - if anything, a nice parallel analysis of another operating system _specifically_ tied to this conversation.

    3. Re:Seriously, what about Windows booting by SanityInAnarchy · · Score: 1

      What you can do is remove startup stuff. I used to do this with start->run->msconfig, but I'm told Vista has a better way of doing it.

      What you can't do is make it take less time with the bouncing "progress bar" between power on and actually starting up the GUI (and giving you the login page if you actually login). This can go very quickly, but it seems one of the inevitable little things about Windows that deteriorates -- after a few months, for whatever reason (I'm guessing the Registry), this step will take longer than a tuned Linux does to boot entirely to a usable desktop.

      Like everything else -- Windows, out of the box, may or may not be closer to what you want. But Windows can only be tuned so far. Same with OS X, or any proprietary system. And when you run into the limit of how far it can be tuned, you have to ask Microsoft or Apple for more power, or to make it better for you. On Linux, you can do it yourself, or fire off a message to an entire community, any of whom can solve your problem for free -- although usually you can tweak quite a bit more before having to dig in source code.

      --
      Don't thank God, thank a doctor!
    4. Re:Seriously, what about Windows booting by Verte · · Score: 0

      I use regcleaner to remove windows services. I don't recall how I got it, but I'm pretty sure it was adware/spyware free and it works a treat. Kill any processes you don't want to load [such as the realplayer and QT daemons, adobe loader, apple update, jsupdate etc] and then remove them using the 'startup' tab on regcleaner. You may have to do this once every few months, but it has been well worth it. You can get the same functionality by browsing and modifying your registry yourself, but that's just icky. The main difficulty I've found with getting windows to boot fast is to stop it trying to cache your whole hard disk to ram, while at the same time creating a giant page file. I've seen "sort of" ways to do that, but most of them seem too dangerous, difficult or obscure to me. In fact, computers with less ram and smaller hard drives seem to load windows a little faster than those with more. If you're running windows in a VM, however, you can probably let it view your local disks as a network disk. That way, there's little caching and no paging. I've not tried that, but it would be interesting to know if anyone has. BTW, unless you're a heavy gamer without PP [multicore or multiprocessor], you'd probably get better performance and more Linux-y functionality with a VM-ed windows anyway.

      --
      We at slashdot are scientists, specialists and kernel hackers. Your FUD will be found out.
    5. Re:Seriously, what about Windows booting by rts008 · · Score: 1

      Try going to "www.microsoft.com/whdc/hwdev/platform/performance /fastboot/bootvis.mspx"
      Download, install, shut down all possible background tasks, start bootvis.
      1.Go to the "Trace" menu> activate "Next Boot + Driver Delays"
      2.Follow the prompts to let the system reboot by itself.
      3.Once rebooted, wait until the bootvis screen pops up- it will be a box with charts and text reports, take your time exploring this- a lot of info (pop-ups with detailed info will pop-up when you rest your mouse curser on info) is in here.
      4.Go back to the "Trace" menu>activate "Optimize System"; this will reboot the PC and rearrange the sequence of driver, process, and programs loading to give the best boot times with WinXP.

      It's been 3-4 years since I messed with this or I would give better directions. Basically it will do some kind of scan, ask you to leave it running while you reboot, it will load first and scan the bootup, then offer you a report, both text and graphical charts showing your bottlenecks and then offering to "optimise" the bootup sequence of drivers and other startup items. I remember getting between 5-30% improvements, YMMV.
      Hope this helps.

      --
      Down With Slashdot BETA!!! I've been around the corner and seen the oliphant; you can only abuse me from your perspecti
    6. Re:Seriously, what about Windows booting by Anonymous Coward · · Score: 0

      start run msconfig startup tab btw there might be semi-important shit in there. only take out stuff that you are sure about.

    7. Re:Seriously, what about Windows booting by hobbesx · · Score: 1
      Your link now seems to be dead, but there is this little snippet from a link at Wikipedia:

      Please note that Bootvis.exe is not a tool that will improve boot/resume performance for end users. Contrary to some published reports, Bootvis.exe cannot reduce or alter a system's boot or resume performance. The boot optimization routines invoked by Bootvis.exe are built into Windows XP. These routines run automatically at pre-determined times as part of the normal operation of the operating system.

      If you are an end-user seeking to resolve issues for boot/resume performance on your PC, we recommend that you contact the vendor from whom you purchased the PC. For information from Microsoft on specific issues, you can search Knowledge Base for Windows XP product issues related to "resume time." Knowledge Base is a free information service available at: http://support.microsoft.com/?scid=fh;%5Bln%5D;kbh owto

      You can also post questions to the Microsoft Windows XP Newsgroups at: http://www.microsoft.com/communities

      If you are a software developer or system designer seeking assistance for using Bootvis.exe in your development efforts, please work with your usual Microsoft Product Support Services (PSS) contacts for developer support. To get developer support if you do not already have a contact, please see: http://msdn.microsoft.com/support/

      The Bootvis.exe tool is no longer available from this site.


      Seems that the optimization that it provides is little more than a GUI for triggering XP's own built in prefetching. Knowing driver load times could be handy though, at least you would have the option of weighing a driver's load time with it's functionality. It doesn't seem that fine grained, however.

      Arguments to the contrary of MS's statements (along with posted results of 2-4 second boot time improvements and reports of so-what-if-you-screwed-your-system-just-go-buy-Nor ton-Ghost) here, and download links both here and here.

      I also just stumbled accross a claim that the '-b' flag to defrag on the command line will trigger the equivalent optimizations, but it seems that XP does this on it's own every three days anyway. It does something on my system, without any displayed information even with the verbose flag.
      --
      This rating is Unfair ( ) ( ) Fair (*) Funny
      Sigh... If only. Modding would be so much more fun.
    8. Re:Seriously, what about Windows booting by rts008 · · Score: 1

      I should have tested that link first.
      I lost track of what was happening with Windows when WGA insisted my retail copy of XP Pro was pirated....
      Kubuntu for the WIN!!!

      --
      Down With Slashdot BETA!!! I've been around the corner and seen the oliphant; you can only abuse me from your perspecti
  32. Missing option by tomhudson · · Score: 1

    "Unless you are running a server or something, there's no reason to leave your computer on when you aren't using it. It just wastes power."

    My computer is my central heating system, you ignorant clod!"

    Think about it - during the heating season, if you use electric resistance heating, you can run your computer "for free" 24 hours a day - the heat it generates will be the same as the heat from that many watts going through your baseboards, so its a wash. Same thing with incandescent lighting.

    Of course, in the summer, you have to expend that much more in AC to remove the excess heat ...

    1. Re:Missing option by danwat1234 · · Score: 1

      Your so right! My system keeps my room a few degrees warmer than the rest of the apartment, which I would otherwise tap on the oil heat to keep my room a comfortable temperature.

    2. Re:Missing option by Khashishi · · Score: 1

      True, but heat pumps are more efficient than resistive heaters.

    3. Re:Missing option by tomhudson · · Score: 1

      "True, but heat pumps are more efficient than resistive heaters."

      Of course, but heat pumps can't seed debian isos for the last month, or suse 10.3 isos this month. (and no, I don't want to get into the whole "Novell is bad" thing - Novell has already gone on record that there are no infringing patents in linux, and I wish everyone who's freaking out would get a grip and stop repeating Monkeyboy's FUD/lies/slander).

  33. Linux boot time? How about hardware startup! by Ed+Random · · Score: 3, Interesting

    All in all, playing with parallelized OS startup is very nice but the real problem lies elsewhere...

    In my experience (data center, 350+ Intel-based servers, Linux + Windows plus a bunch of SPARC Solaris boxes), the OS boot time is negligible compared to the time needed for hardware initialization:

    - BIOS startup
    - Memory check
    - Remote Console init (DRAC/XSCF etc.)
    - RAID Controller(s) init, disk spinup
    - RAID Consistency Check, volume initialization
    - Start Boot Sequence

    Especially the disk subsystems cause large delays - most time is spent waiting for the GRUB screen.

    Parallelizing the hardware initialization is where we could make some significant progress.

    --
    -- Gxis! Ed.
    1. Re:Linux boot time? How about hardware startup! by SanityInAnarchy · · Score: 1

      Not even that. Aside from doubts as to whether you can actually parallize the hardware initialization, there's the fact that most of these are just stupidly written -- Grub included. Lilo had the right idea -- it can be configured to automatically boot, with no delay, unless you're holding alt. I want everything up to and including Grub to function that way. I honestly don't care about the memory check, and most of these things are disabled on my box, but many controllers will take over the whole screen to give me some tiny piece of information and wait 5-10 seconds in case I might want to configure something.

      That's what bugs the hell out of me. Even though my Grub is on a timeout of only five seconds, I very rarely want to do something strange and configure it, and I very rarely want to boot into Windows -- usually when I do, I reboot with a script in Linux which modifies the grub config to default to Windows anyway. So, 95% of the time, I have 5 seconds added to my boot to make it easy for the other 5% of the time when I might have to customize something -- and that's just Grub.

      My nVidia RAID screen, like Promise RAID and every other (shitty) RAID I've ever worked with, displays a screen showing me the status of my array and prompting me to press some key if I want to configure it. I know this is an unnecessary pause, because if I hit enter as soon as that screen displays, I'll skip straight to Grub -- but if I leave it alone, it'll wait 5-10 seconds just in case I want to configure my RAID this boot.

      Compare this to Macs. Without Yaboot (or Grub, or Bootcamp, but my last Mac was a Powerbook), you can simply power on and it will hit OS X as fast as it possibly can. If you need to do something strange, you hold down a key while booting, and you can get things like boot from CD, firewire target mode, etc. That's the way any boot choice should operate -- boot the default as fast as you possibly can, unless the user is holding down a key. You don't need the RAID configuration to be user-friendly, especially when a person might want to configure their RAID once in their computer's lifetime -- maybe -- even if they do it once a month, you're slowing down ~30 boot cycles to make it easier for that one.

      LinuxBIOS is where we might make some real progress here -- it might actually parallize some of this -- but there will always be the odd NIC BIOS that wants to do things its own way.

      --
      Don't thank God, thank a doctor!
    2. Re:Linux boot time? How about hardware startup! by imemyself · · Score: 1

      Since you're using hardware RAID controllers I'm assuming you're not talking about some little desktop systems, so my question would be, why on earth are you rebooting servers so often?

      I mean, I honestly wouldn't care if it takes 5 extra seconds when booting, having it tell me what key I need to press to configure it would save me a lot of time. I don't want to have to spend twenty minutes of my time looking for documentation for some obscure RAID controller in order to configure it. Its not like most people are going to be staring at a monitor connected to a server every day while they turn it on. If it wastes five or ten seconds of the computer's time when it reboots once in a blue moon, I really don't care. If five or ten seconds is a big deal to you as far as uptime goes, then you'll be using something like clustering anyway.

      --
      Every time you post an article on Slashdot, I kill a server. Think of the servers!
    3. Re:Linux boot time? How about hardware startup! by SanityInAnarchy · · Score: 1

      Since you're using hardware RAID controllers

      I'm not using hardware RAID controllers. It's a big desktop system, but it's ultimately just a gaming rig. nVidia RAID is fakeraid -- it's BIOS RAID.

      I mean, I honestly wouldn't care if it takes 5 extra seconds when booting, having it tell me what key I need to press to configure it would save me a lot of time. I don't want to have to spend twenty minutes of my time looking for documentation for some obscure RAID controller in order to configure it.

      If it takes you twenty minutes of looking through documentation for what key to press, you're not trying. Takes me only a few seconds of a Google search to find out what to do with a Powerbook, and it's not obscure at all -- hold T for firewire Target mode. Hold C to boot the Cdrom. Do this once and write it on the machine, if you're that worried.

      If it wastes five or ten seconds of the computer's time when it reboots once in a blue moon, I really don't care. If five or ten seconds is a big deal to you as far as uptime goes, then you'll be using something like clustering anyway.

      Which is still no excuse. Consider the case where five seconds (times four of five devices, making it closer to 30 seconds) is a fucking HUGE deal, but we might not have the money to build a cluster to cover it? Consider that even if you've got a second machine, every second that first machine is down is a second you could be knocked offline entirely?

      More importantly, consider that rebooting happens sometimes on servers (kernel upgrades, at least), and quite a lot on desktops. Because you're too lazy to check in a manual to find out what key to hold to bring up a RAID menu for the maybe one or two times in the lifetime of the machine that you'll have to, we all have to suffer on the once-a-month reboots. Maybe not a big enough deal for me to waste time here -- but hey, it's even more annoying on my desktop machine, where even if I reconfigured the RAID array once a month, it's still making me wait every single day to make that once-a-month easier -- assuming I don't know what I'm doing, which I do.

      But again, that assumes I'm talking about servers. I'm not, I'm talking about custom-built desktop machines. Network card capable of booting off the network? There's a chance it'll throw up one of those screens. Have a high-end video card? Maybe it'll give you a splash screen before even letting you into the BIOS. This kind of shit is especially annoying when some off-the-shelf Dell-like boxes just boot -- maybe two seconds from power-on to bootloader.

      --
      Don't thank God, thank a doctor!
  34. Here's some more advice to counter FUD. by twitter · · Score: 1, Interesting

    You insult lots of experienced Linux users who do care about their machines booting several times slower than an XP pc.

    That's an insult to everyone's intelligence. There is no such machine, unless you have serious hardware problems and the present article is the "more power to you" that I mentioned. Many live CDs boot faster than XP and most installed distributions boot faster than that.

    You may have a laptop and you may be happy with "hibernating" it, but many of us need to power off PCs. An office PC I power on every day, my home's PC I power on and off when I get and leave home.

    You should happy be like me. I'm aware of the issue. Ultimately, if you care about power consumption you are going to get a laptop or desktop with reasonable power management. Test it out before you buy it or send it back if you can't and it does not work. Microsoft continues to design complex and "extensible" non standards for power management, so it's not easy. Comfort yourself by knowing that M$'s dirty tricks make things harder for their own users than they are for you.

    Of course, there are benefits to having at least one machine that you never turn off and that never sleeps. I have contact information, archives of photos, music, movies and work that I can get to anywhere by sftp and KDE makes it all look local. This is much easier than syncing three or four laptops and desktops. Wake on lan might be nice, but my equipment is all so old, I've never bothered trying. For at least one machine, I never will. A modest server consumes 50 watts or less, kind of like some light bulbs I never turn off. I get more use by far out of the servers than I do light bulbs. Your business might also consider that a reasonable convenience.

    --

    Friends don't help friends install M$ junk.

    1. Re:Here's some more advice to counter FUD. by jb.hl.com · · Score: 2, Insightful

      That's an insult to everyone's intelligence. There is no such machine, unless you have serious hardware problems and the present article is the "more power to you" that I mentioned.

      Anecdotal evidence suggests otherwise. On my (old; I ditched it this morning for a Mac...that feeling, twitter, is called cognitive dissonance) PC, XP booted in about 20-25 seconds whereas Ubuntu took about 30-40 seconds. I'm not even going to start on the prepostorous LiveCD thing...how could a full Linux/GNOME desktop booting from a CD take less time to start up than Windows XP installed on a hard disk? That makes no sense. Every time I've tried an Ubuntu liveCD it's taken a few minutes to boot up.

      Microsoft continues to design complex and "extensible" non standards for power management, so it's not easy. Comfort yourself by knowing that M$'s dirty tricks make things harder for their own users than they are for you.

      You mean ACPI? The one developed by Microsoft, Toshiba, Intel, HP and Phoenix? The one that comes pretty much as standard on just about every motherboard? Please.

      Now, you always go on about Microsoft "blaming the user". But the subtext to your posts in this thread is "You shouldn't have to boot up at all, silly person, you should be using sleep modes. Look at my l33t uptimes." What use is it to say that? Is it just trying to divert criticism from the fact that yes, fresh stock installs of popular Linux distros do actually tend to take more time to boot up than fresh installs of Windows XP on comparable hardware?

      --
      By summer it was all gone...now shesmovedon. --
    2. Re:Here's some more advice to counter FUD. by VON-MAN · · Score: 1

      "that feeling, twitter, is called cognitive dissonance"

      No, Joe, it is disdain.

    3. Re:Here's some more advice to counter FUD. by jb.hl.com · · Score: 1

      Well...one or the other. ;)

      --
      By summer it was all gone...now shesmovedon. --
    4. Re:Here's some more advice to counter FUD. by dsmall · · Score: 1



      I believe that if you look for it, you'll find that most of the default pointers for system crashes under the Windows desktop simply restart "Explorer.exe" (where Explorer is the 'Windows Desktop' Explorer, not the 'Internet' Explorer). In fact, it's useful to remember that if the desktop gets hung up, you can CTRL-ALT-DEL to the Windows Task Manager, then Kill -9 ... errr ... End Process "Explorer.exe", then Run Process "Explorer.Exe", and the desktop will usually come back. I tend to clean up and reboot, according to my brother's Law of Windows: A Reboot Cures Many Ills Of Windows.

      From time to time using Windows, you'll see the desktop blank out, replot all the icons with just the name and a generic icon, then fill in the icon with the correct graphics. I believe this is Yet Another Windows Desktop Crash. Of course, this is all just my (first-amendment protected) personal opinion. Please note how careful I am being in specifying this personal opinion, there is a very good reason for this.

      I first came across this some time ago. My (new) HP laptop machine here crashes twice on every startup. This takes cubic time.

      Is this a particularly big deal? Oh, I don't know. I've seen other computers do this. I can recall tracing startup on the Macintosh and seeing stores to location 0 (Nil pointer usage) that made the startup a real "Hmmm, how did that work, anyway?" philosophical question. Finder 1.1 and System 1.1g were the cleanest in this regard. System 7 was a nightmare, because it was the first true multitasking OS, and lots of programs hammered on location 0, and the ones who had been getting away with it because they stored to, then read from, 0 conflicted with others and the operating system, causing maddening, hard-to-trace errors.

      I'd be interested in hearing if the current 10.x operating system has this fixed in hardware.

      Thanks, Dave Small

  35. Neat idea- disappointing real world results. by puddnhead7 · · Score: 1

    After reading the article, I decided to give initng a try. I timed my boot from a powered off state to the GDM login prompt so as to have a base-line, installed and configured init-ng 0.6.7, and then retimed from a power off. The result: initng ADDED 12 seconds to my boot time. Tried it again to see if it was a first boot fluke and got the same results to the second. This is on the latest feisty. Someone mentioned that Ubuntu 6.06 used upstart- it doesn't look like feisty is to me, but that might be my ignorance. Cest la vie.

  36. Lightning like Linux. by k1e0x · · Score: 1

    Linux could boot very fast if only code and programs and hardware that were required to run did. A lot of the checks it makes on booting (discovering new hardware, loading modules, populating udev, etc..)

    If this was done by the installer once and then never needed again then it could come up much faster.. it is mostly a moot point though because rebooting is something that should be really done infrequently.

    It can be very fast though.. I actually once made Linux a system that went from OFF to Google in 15 seconds. (2.2 kernel). It just depends on how the system is implemented. A fairer statement would be "..Common Linux desktop distros that take a long time to boot.."

    --
    Bringing liberty to the masses. - http://freetalklive.com/
    1. Re:Lightning like Linux. by Billly+Gates · · Score: 1

      WindowsXP does what you said and creates a custom hal after installation. Problem is after you change some bios settings or add hardware without first installing the driver can cause it to blue screen.

  37. Forget about that, use suspend by harry666t · · Score: 1

    Well, I've decided to try out software suspend.

    I'm not certainly sure which packages to install and how to configure them etc. (somehow I already had it all on my Debian box) but just compare:

    A "fresh" boot to 5th runlevel, with my customized kernel, fsck, only a few basic system services enabled + apache2 and mysql, took 42 seconds. 42 seconds from grub menu to login prompt (I do not use any of those crappy [xkgw]dms but a real login(1)).

    A suspend, done with 'sudo /usr/sbin/hibernate' not only managed to power off the computer a lot faster, but it took 10 seconds from grub to bash prompt.

    The only disadvantage of this solution is that hibernate and nvidia kernel module doesn't seem to like each other. When I issue # hibernate on an xterm,

    Some modules failed to unload: nvidia
    hibernate: Aborting suspend due to errors in ModulesUnloadBlacklist (use --force to override).

    so I've tried with --force and shit happened.

    What I need now is just good session management in my WM.

  38. Boot times don't bother me by DoctorPepper · · Score: 4, Funny

    Because I only reboot my computers when I have to, like when I get an update to the kernel. The rest of the time, they just waste electricity and CPU cycles, and generate excess heat my A/C unit has to deal with! :-)

    --

    No matter where you go... there you are.
    1. Re:Boot times don't bother me by ceoyoyo · · Score: 1

      Hear hear. Occasionally a Windows user will brag about how fast their boot time is (and how slow Linux and/or OS X is). My response is that I don't really know how long it takes to book my computer, and I don't really care. Last time I rebooted my MPB was... two weeks ago. And the mini by the TV gets rebooted whenever the power fails or I get around to installing an update that needs it.

    2. Re:Boot times don't bother me by Evil+Pete · · Score: 1

      Not sure why the parent is modded funny when it so true. I switch my PC on daily, but I don't care about the amount of time it takes to start up ... except if it took 5 minutes like my win2k box at work used to (grrr) ... but I digress. So my linux install with too much stuff installed takes 60 seconds to boot, I don't really notice ... if I really cared about it I'd just leave it on or stop the startup of so many things ... not a biggie.

      --
      Bitter and proud of it.
    3. Re:Boot times don't bother me by evilviper · · Score: 1

      they just waste electricity and CPU cycles, and generate excess heat my A/C unit has to deal with!

      You can get the benefits without the drawbacks... Suspend (S3) only takes a couple seconds (to suspend or resume), and your PC barely draws any more power in S3 than it does when entirely off... maybe 1W, and that's probably earned back from not having to wait 30 seconds for startup/shutdown.

      --
      Slashdot gets worse every day... Pipedot: News for nerds, without the corporate slant
  39. Get your money back. by twitter · · Score: 0, Troll

    Denying the problem doesn't make it go away, really. All the hibernation and sleep modes in the world don't change the fact that Linux boot times are much longer than, say, Windows XP's.

    I've never seen this happen. Reapeating this FUD does not convince me it is true.

    for many laptops [power management does not work]

    If your laptop does not work, you need to get your money back. Take it back and tell them why. Better yet, test it before you buy it. Don't buy things that waste your time.

    Thinkpads usually work out of the box. I usually turn off ACPI and use APM, but ACPI works too.

    --

    Friends don't help friends install M$ junk.

    1. Re:Get your money back. by jb.hl.com · · Score: 1

      I've never seen this happen. Reapeating this FUD does not convince me it is true.

      "FUD" refers to something that is untrue. What you mean is "inconvenient truth". For instance, here's a comparison of Linux and Windows boot times; it shows that in all but ONE of the tests (Windows with antivirus software installed using 96MBs of RAM) Windows trounces OpenSUSE and Linspire in boot times.

      You've got your head in the sand. What's wrong with just holding up your hands and saying "Yes, Linux boot times are shit, but we're working on it"? There's far less shame in that than in just making shit up.

      --
      By summer it was all gone...now shesmovedon. --
    2. Re:Get your money back. by AusIV · · Score: 1

      For instance, here's a comparison of Linux and Windows boot times; it shows that in all but ONE of the tests (Windows with antivirus software installed using 96MBs of RAM) Windows trounces OpenSUSE and Linspire in boot times.

      That chart compares two of the most bloated distros and is nearly a year and a half old. I use Kubuntu (admittedly not the most streamlined distro around), and I've timed boots at 45 seconds. A slackware or gentoo install could put that to shame. My Windows install, having turned off any unnecessary startup options, takes about 2 minutes to a usable desktop - a significant improvement from the 5 minutes it took before tweaking the startup options (something the average user wouldn't know how to do).

      But honestly, boot time is the last thing I concern myself with so long as I can hibernate and suspend. On either OS, I reboot maybe once a week, the rest of the time is hibernate and suspend.

    3. Re:Get your money back. by Eideewt · · Score: 1

      ZDNet's measurements are very far from what I get (and yes, I have measured). Windows and Linux actually take roughly the same amount of time to boot on my computers, except Linux doesn't make me wait for the system to become usable.

  40. 1st post!! by Nushio · · Score: 1

    My Linux Box is so optimized that this is the fastest I could post!

    --
    Check out Unsealed: Whispers of Wisdom! http://unsealed.k3rnel.net It's an action-RPG about Open Sourcerers.
    1. Re:1st post!! by renegadesx · · Score: 1

      I like having a slow booting Linux, gives me time to have a quickie with your mother

      --
      Make SELinux enforcing again!
  41. My FreeBSD system by davecarlotub · · Score: 1

    My freebsd home server running on an HP dc7100:
    From shutdown -r now to login: 67 seconds.
    And that includes starting several daemons like apache, vsftpd, mysql, ircd, and of course /usr/local/etc/rc.d, thank you.

  42. Who needs optimized booting? by onetwofour · · Score: 1

    You never need to reboot a good linux distro. In addition I have lots of debug info available if something isn't working, you just don't get that with other operating systems.

  43. Unethical? I think not.. by danwat1234 · · Score: 1

    Since when is running a computer 24/7 unethical? I think the 150watts of electricity usage from a computer is far from a waste when your computer is Folding, or any other distributed computing project. Just my .02 dollars

  44. about:blank and toolbar bookmarks by billstewart · · Score: 1
    Agreed - about:blank saves lots of time when opening pages, especially since my laptop is sometimes plugged into a work LAN or VPN where it needs a proxy and sometimes into the Internet where it doesn't, and this not only avoids waiting for complex pages to load, but avoids getting stuck by incorrect proxy settings.

    The Personal Bookmarks Toolbar makes Slashdot one mouse click away.

    --

    Bill Stewart
    New Fast-Compression-only CPR http://preview.tinyurl.com/dy575ks
    1. Re:about:blank and toolbar bookmarks by Sillygates · · Score: 1

      For those of us living in places where we have to use DHCP, thats one of those things which can usually be done in the background. If you are using fedora/redhat the easiest place to do this is in: /etc/sysconfig/ifup-eth . This script gets called at boot. echo $" failed; no link present. Check cable?" ip link set dev ${DEVICE} down >/dev/null 2>&1 exit 1 fi #changed below this line echo -n " (Background)" { if [ -n "$ETHTOOL_OPTS" ] ; then /sbin/ethtool -s ${REALDEVICE} $ETHTOOL_OPTS fi if /sbin/dhclient ${DHCLIENTARGS} ${DEVICE} ; then echo $'\n'$"DHCP Success for ${DEVICE}" doRest else echo $'\n'$"DHCP FAILED for ${DEVICE}" exit 1 fi } & #changed sbove this line The andpersan tells the shell to spawn a new instance, while I put everything that happens after the if-else block in a function I called doRest(). Make a backup of the script if you have not done much in terms of shell scripting. Try it out :) .

      --
      I fear the Y2038 bug
    2. Re:about:blank and toolbar bookmarks by Sillygates · · Score: 2

      echo $" failed; no link present. Check cable?"
      ip link set dev ${DEVICE} down >/dev/null 2>&1
      exit 1
      fi
      # edited below this
      echo -n $" (Background)"
      {
      if [ -n "$ETHTOOL_OPTS" ] ; then
      /sbin/ethtool -s ${REALDEVICE} $ETHTOOL_OPTS
      fi
      if /sbin/dhclient ${DHCLIENTARGS} ${DEVICE} ; then
      echo $'\n'$"DHCP Success for ${DEVICE}"
      doRest
      else
      echo $'\n'$"DHCP FAILED for ${DEVICE}"
      exit 1
      fi
      } &

      --
      I fear the Y2038 bug
  45. Boot faster by profiling your boot by AliasMarlowe · · Score: 1

    Some linux versions use readahead while booting for a specific set of files which will be needed during the boot. The set of files is stored in a boot profile. After upgrading your system or changing the services started at boot, you should re-profile the boot, so that the readahead files are the ones actually needed. It can make a remarkable difference to boot speed.

    Instructions at http://www.ubuntuforums.org/showthread.php?t=25426 3&highlight=grub+profile+speed+boot

    Note that the profiling boot takes longer, but subsequent boots should be faster, since most files needed for booting will already be loaded before they are required.

    --
    Those who can make you believe absurdities can make you commit atrocities. - Voltaire
  46. Windows - check out your Startup options, etc. by billstewart · · Score: 1
    Windows starts up a bunch of different processes and applications at and after boot time, and it keeps hiding them in more different places every new OS version. So you'll have to hunt them all down. The easy one is your Startup folder (especially if your machine is set up by a Corporate IT Droid Department). See what apps you do and don't need, and make yourself a Don't Start folder to drag the ones you don't usually need into. Then go find the other various startup scripts, directories, etc. (i.e. RTFM, if you can find a manual) and see what stuff you do or don't need from there. I found that getting rid of the Office startup stuff helped a lot.


    Also, if your PC comes from a major manufacturer like HP or Dell, see if it has a bunch of "helpful" manufacturer-installed software, and see what you can get rid of from that. I don't know if it's as bad as it used to be, but when my mother-in-law had a Compaq, there was a lot of stuff we had to trash to get the system to behave adequately; we eventually upgraded her to an XP-from-scratch installation which worked a lot better. (Of course, getting rid of all the spyware helped too :-)

    --

    Bill Stewart
    New Fast-Compression-only CPR http://preview.tinyurl.com/dy575ks
    1. Re:Windows - check out your Startup options, etc. by iangoldby · · Score: 1

      AutoRuns by Mark Russinovich and Bryce Cogswell is the ultimate tool for finding all of those startup applications and processes.

      I don't think there is anything much that gets past AutoRuns.

      The only thing is - it many items that are essential to the normal running of Windows and it can be quite hard to work out exactly what is essential and what is not. Give it a try though.

  47. Speeding up modern Ubuntu boot not easy... by Sits · · Score: 3, Informative

    Ever since Ubuntu Edgy much of the low hanging fruit in speeding up the Ubuntu boot has already been taken. Looking at the bootcharts for my system since then shows remarkably little time when the CPU is idle once the base kernel has finished loading. This means that running anything more in parallel simply won't net me anything (in fact scheduler overhead and disk thrashing may in theory make things slower).

    For example, there is an improvement in the time it takes for the clock to appear from "Ubuntu Dapper Flight 3 Default kernel" to "Ubuntu Feisty Herd 5 generic kernel". The Ubuntu folks worked hard to try an eliminate sleeps from their initscripts and when a sleep was unavoidable they would run other parts of the startup process in parallel. They also made changes to Xorg to prevent it (re)reading so much stuff on launch. There was also the introduction of the readahead script which tries to arrange for as much of the boot time reading to be done in one big chunk. Throughput is higher when the disk is only reading and can utilise it's readahead. An attempt is also made to try and request files in the order in which they are laid out on disk (to minimise disk seeks which hurt performance). In Feisty a move was made to using dash instead of bash for scripts because it was smaller and executes scripts faster.

    The only things that seem to win me any gain over the default Ubuntu Feisty install are turning off initscripts for services I absolutely won't use (e.g. ipv4 autoconfig via avahi) and reducing the number of restricted binary driver modules being probed (I have long noticed that the only benefit that recompiling the kernel gives to boot speed is that you can simply leave out features not on your computer making the initial kernel startup where it probes for things you might not have (like which software RAID is faster) a shade faster). It is also worth noting that Ubuntu starts X quite early and continues loading services afterwards which means the gain from disabling one of these "after X" services (like CUPS) isn't so noticeable (but might mean your desktop actually starts responding to clicks a bit sooner).

    Profiling the boot to try and improve the readahead takes a long time to run - the profile run seems to take three times as long as a regular boot. It could be argued that you will never gain back the extra time you waited on the profile run...

    I suspect reducing the boot further will start to need more complicated procedures, perhaps reordering modprobe.conf and reducing the amount of needless reading of files. Eventually you end up having to do the same tricks as Windows/OSX - e.g. working out where the fastest part of the disk is and copying every file needed to boot there, bringing up the network cardafter the desktop has started, periodically defraging bits of the disk, prelinking...

  48. Seriously? by Anonymous Coward · · Score: 0

    I can't discern if this is a serious question. Boot times? Seriously? Ok, XP gets you to the login screen quickly but it does so at the cost of delaying the startup of vital services that are necessary to actually use the computer. Really, how important is it that I can login but then still wait >60 seconds for all the extra, and necessary, services to start, like AV, firewalling, updating, an so on. It's in fact really annoying that I can login and *appear* to begin working, only to have my work delayed and the active focus change when other necessary services start. I think I'm typing in a window but really I'm not because some mysterious service or program has grabbed focus for its startup sequence. I'd rather have a longer boot time with the end result being that I'm able to use the computer as I want to rather than something that gets me to the login screen quicker. Wow, I can login quicker, it must be better.

    A measure of boot time is inevitably done in order to try to quantify the amount of productivity gained or lost while the computer is readying itself for work. If one wants to quantify boot time however, then one needs to quantify the time necessary for real work to begin, the time necessary for reboots for seemingly innocuous updates that really shouldn't require a reboot (whether to the OS or to extraneous programs), the time spent virus and malware definition updating and scanning, and related factors that determine how much time I spend actually using the computer.

  49. I used bootchart to shave 5 minutes off my boot by Chaoticmass · · Score: 1

    Went from a 5 minute boot down to a boot time of only 60 seconds by optimizing my boot scripts and trimming the fat. Used bootchart to find the slow downs and optimized from there.

    Of course this was on a ~15 year old 50mhz Sparcstation LX running Debian :)

    1. Re:I used bootchart to shave 5 minutes off my boot by Chaoticmass · · Score: 1

      er, that should be "shave 4 minutes". and I even used the preview... d'oh

  50. Now you're cooking with gas by tepples · · Score: 1

    Think about it - during the heating season, if you use electric resistance heating, you can run your computer "for free" 24 hours a day If. Natural gas heating is more efficient in most areas, as it doesn't lose power over generation and transmission lines.
    1. Re:Now you're cooking with gas by tomhudson · · Score: 1

      My electricity comes from hydro-electric dams, so its a lot "greener" than natural gas can ever be - zero C)2 emissions.

    2. Re:Now you're cooking with gas by Des+Herriott · · Score: 1
  51. Slamd64 by Anonymous Coward · · Score: 0

    Slamd64 11 shutdown -r to prompt ... 71 secs.

      Brings up a satellite card etc.

  52. Mine is too by commodoresloat · · Score: 1

    But I had to wait too long for the damn thing to boot, so I missed the opportunity. I wish there was information somewhere that could help me speed that process up.

  53. Boot to desktop by DrYak · · Score: 2, Insightful

    To be fair, my Windows box boots pretty quick; I think the time between power on and desktop


    The trick is just right there. It's the time until desktop shows up. Not until the system is usable.
    Windows XP (and I think Vista too, but the friend that was supposed to gime her unused Vista CD hasn't yet) tries to show desktop as soon as possible, even if all services didn't finish booting.

    My machine isn't brand new. I mean at all. It's an old Pentium III Tualatin with 440BX chipset and 1GB SDR 133.

    On linux (opensuse with old skool init), most of what need to be started is started during boot time (which includes clamav/freshclam daemons, a couple of hardware monitoring daemons, software update service, ssh, cups/samba/nfs, etc.) it's not fast but it's not taking hours either. Once I get a log-on screen, everything I have to wait for is Gaim starting. And it seems linux is good enough at multitasking to handle it in background well enough.

    On windows XP, the boot phase doesn't seem much faster. And then I have to log-in and ll hell break loose. Only after I have logged in, the system decide to rescan my USB and Firewire interface, plus my anti-virus has to terminate to load, plus it has to update its definitions, and a half a dozen of small applets has to load into the task bar, all of this constantly checked by the antivirus.
    All this can take up to 5-10 minutes.
    And I'm running just a minimal amount of task bar applets (Firewall, antivirus, hardware monitoring, control pannels for 3D card) I'm not running an additionnal crap (no OpenOffice.org fast loaders, nor MSOffice, no Acrobad preloader, nor Photoshop)...

    It looks to me that Windows just try to show the desktop earlier to show off and give a false impression of quick boot. Or maybe it's designed to run only on last generation hardware (Athlon / P4 and upward with DDR at least). But I can't say the delay is short before I can do anything in Windows (usually : start a game).
    --
    "Sufficiently advanced satire is indistinguishable from reality." - [Tips: 1DrYakQDKCQ6y52z6QbnkxHXAocMZJE61o ]
  54. You forgot.... by DrYak · · Score: 1

    Like antivirus software, which for me is the primary reason Windows is unusable for a while even after the desktop is shown..


    You also forgot the antispyware, both of their definitions updater, and the firewall.
    All of which are absolutely obligatory if you don't want you machine to be completely unusable after a couple of days connected on the net.
    At least, thanks to FireFox (and its adblock extension), you don't need to start an additional anti-popup software.
    --
    "Sufficiently advanced satire is indistinguishable from reality." - [Tips: 1DrYakQDKCQ6y52z6QbnkxHXAocMZJE61o ]
  55. Slow booting? by hitmanWilly1337 · · Score: 1

    My Widows box actually takes quite a bit longer to completely load than my Linux one. If you factor in the time it actually takes for the OS to finish loading, the boot times come down in favor of Linux.

    p.s. I run Ubuntu 6.10 and Windows XP on a 2.3 ghz AMD64 w/ 1 GB of RAM and a 1.8 dual core intel laptop w/ 2 GB RAM respectively.

    p.p.s. Yes i know this isn't a scientific comparison by any stretch of the imagination, im just saying that this is what happens for me.

    1. Re:Slow booting? by TheAwfulTruth · · Score: 1

      I work on kernel and X drivers as well as drivers for Windows. I have a machine I swap HDs in and out of for dev and testing, including many version of Linux, several very current, as well as most all versions of Windows including Vista.

      On this machine, a 2.8gig P4 with 2 gig of ram, Linux loses to every version of Windows. Linux runs about :50 to 1:30 in boot time to an X desktop. Windows from 30-45 seconds.

      Most of these are relatively clean and minimal installs, but even the XP box which is now three years old and loaded with crap still boots in less than 45 seconds.

      The difference is not big enough to really complain about but the fact is, stock installs of nearly every, (if not every) desktop Linux distro DO take longer to boot than any version of windows (when booting into a functionaly similar state, i.e. into X)

      I infact just tested this within the last month for S&G

      --
      Contrary to popular belief, coding is not all free blow-jobs and beer. Those things cost MONEY!
  56. Fast boot for embedded systems by WereRaven · · Score: 1

    For desktop systems an optimised boot process isn't much of an issue but for embedded systems it is critical. You don't want a client asking "Why does it take a minute for my VoIP server to reboot?" "Why does my multimedia player take two minutes before I can watch a movie?" These people can't sit at a console and watch all the pretty boot messages scroll by. All they can do is sit and wait and get annoyed because their system hasn't come back up yet. Articles like this are interesting reading for anyone wanting to squeeze every last second out of the boot time of their embedded linux device.

    Another big contributor to boot times that the article didn't cover is the amount of time spent sitting in the BIOS. If you're serious about shaving off a few more seconds you might also want to look at http://linuxbios.org/.

  57. Trying linux on a laptop before you buy by zsau · · Score: 1

    If your laptop does not work, you need to get your money back. Take it back and tell them why. Better yet, test it before you buy it. Don't buy things that waste your time.

    Serious question: How do you get a laptop reseller to let you install Linux on a laptop to try before you buy? Or alternatively, How do you test all hardware (including hibernate) with a live cd? Hibernate at least needs to write to a swap partition which needs repartioning... I'd love my next laptop to be a walk in the park, but I've never had much luck...

    --
    Look out!
  58. As A Gentoo User: by Anonymous Coward · · Score: 0

    I first took the time to decide exactly what my system would be used for, which was a desktop system not a server. The next step I did was build into the kernel all of my hardware instead of using modules. Last but certainly not least, I ensured that my desktop uses a static IP instead of dhcp to bring up the network. This alone saved me in excess of 20 seconds.

    Simply put, by taking the time to decide exactly how I was going to be using the damn system, I was able to optimize the boot sequency by not having unneeded services starting and I can go from a cold power up to KDE in 75 seconds and that's with having to log-in and type start-x.

  59. Get what you want or get your money back. by twitter · · Score: 1

    How do you get a laptop reseller to let you install Linux on a laptop to try before you buy? Or alternatively, How do you test all hardware (including hibernate) with a live cd?

    You use a live CD to test the hardware. If it does not work, don't buy it. If you don't want to risk writing out a file for hibernate suspend, take it home and do it there. It's really easier for everyone for you to just do it there with the demo machine, which they "restore" all the time anyway.

    The bottom line is getting what you want or not parting with your money. If something does not work the way you want it to, you should not buy it. If you buy something and are not happy with it, you should get your money back. If the store won't give your money back, you should not do business with them anymore. I pay a small premium to local mom and pop computer stores because they let me try things before I buy them and have been very cool about things that don't work for me. Even Walmart has the good sense to let people bring back electronic gadgets that don't work.

    I'd love my next laptop to be a walk in the park, but I've never had much luck..

    I've had good luck with used Thinkpads. Newer hardware is always a pain, until Dell and other major vendors get smart. The easiest way to tell is to just try it before you buy it.

    --

    Friends don't help friends install M$ junk.

    1. Re:Get what you want or get your money back. by Anonymous Coward · · Score: 1, Insightful

      You have an unrealistic viewpoint. The fact is that many people don't want to just preach to the choir. Yes, we can and should try and support hardware that works the best, but we don't just fix problems for the people currently using linux but also for the people who WILL be using linux. What about the new person trying to get linux working on their bestbuy dell laptop? Will suspend and hibernate work on that? Should we just say, "Linux is the best it can be, why improve it?" Or should we say, "There are some people who still need faster boot speeds, lets make this happen!"

      If Linux booted faster than Windows, I guarantee people like you would be touting about how great linux is just on that alone. Let's just grow up and except that there is a minor issue like this and know that there are people who are looking for solutions. Lately boot speeds have greatly improved and its often faster than an old windows install. But lets face it, it can get better! Stop trying to say "good enough" and start saying, "Let's make linux even better!"

      You people act like this is an insult, but this is the way open source should be. We admit our problems and we work to fix them. So instead of bitching about how stupid everyone else in the world is, lets actually try this and see if this is an improvement, or maybe we can suggest better ways of accomplishing this task. There are people who would like to see this happen, otherwise no one would have spent time finding a way to parallelize boot processes.

      To people who are just saying that this is unnecessary: Why question people's motives? If I wrote a shell script that double sided printed pdf files (I did) and wrote an article about it, would you call it useless just because you don't use pdf files? Would you say its unnecessary because you can do the same thing with adobe's acrobat reader? Maybe I have a personal need for something that you just don't have. People have the need for this article, so instead of insulting it because you don't need it, how about you ignore it because you don't need it?

    2. Re:Get what you want or get your money back. by dedazo · · Score: 1
      So twitter, are you going to reply to this soon? You seem to enjoy replying to ACs with your clever "An AC taunts" and whatnot opening lines, maybe you could do the same for us here? It would be much appreciated, as always. I mean, since you never got back to me on the topic of the supposedly mythic Linux botnets.

      Thanks!

      --
      Web2.0: I love when people Flickr my cuil and digg my boingboing until my google is reddit and I start to yahoo
  60. more nonsense and FUD by twitter · · Score: 1

    One of my dedicated fans calls me a snob:

    But the subtext to your posts in this thread is "You shouldn't have to boot up at all, silly person, you should be using sleep modes. Look at my l33t uptimes." What use is it to say that? Is it just trying to divert criticism from the fact that yes, fresh stock installs of popular Linux distros do actually tend to take more time to boot up than fresh installs of Windows XP on comparable hardware?

    No, I've said that long gnu/linux boot times are bullhit and that they don't matter. There's no diversion going on there, it's a denial of bullshit. Nothing boots faster for me than a default Etch desktop, not even your supposedly "fresh" XP install. Once you add on anti-virus and the inevitable mal and spyware, the Windoze system is a brick that takes forever to boot and has to be reinstalled, how convenient! There's nothing leet about using features out of the box like I do. APM is as easy to use as closing the lid on your laptop. I like to share that because it's a great time saver and advantage everyone can and should enjoy.

    --

    Friends don't help friends install M$ junk.

    1. Re:more nonsense and FUD by jb.hl.com · · Score: 1

      No, I've said that long gnu/linux boot times are bullhit and that they don't matter.

      They DO matter. Lots of people turn off their computers at night and switch them on again in the morning, sticking your fingers in your ears and going "Lalala use sleep lalalalala" isn't going to change the fact that Linux boots slowly.

      Or is it that they only "don't matter" because they're one of Linux's flaws? Be honest. If it's good, it's good, if it's bad it doesn't matter. You in a nutshell.

      Nothing boots faster for me than a default Etch desktop, not even your supposedly "fresh" XP install.

      There's no such thing as a "default Etch desktop". Etch comes with no GUI to speak of and few services enabled. Of course a new install of a text-mode OS with no services will boot faster than a graphical OS with many services enabled.

      Once you add on anti-virus and the inevitable mal and spyware, the Windoze system is a brick that takes forever to boot and has to be reinstalled, how convenient!

      The page I linked to a while back said otherwise, and that even with antivirus or whatever XP still booted faster under most conditions. Try again.

      (Not that I ran antivirus or ever got spyware of course. Seriously.)

      --
      By summer it was all gone...now shesmovedon. --
  61. Ubuntu Feisty takes 95 seconds on my decent laptop by Jamie+Lokier · · Score: 1

    Ubuntu Feisty boot time to usable desktop: 95 seconds.

    Well I'm using Ubuntu Feisty, presumably with Upstart, and it takes about 75 seconds to boot to the Gnome login screen, on my Core Duo laptop.

    The kernel is only a tiny fraction of that; the kernel starts very quickly.

    Most of it is surprising delays in some of the boot scripts. For example "Starting system message bus dbus" is followed by a 15 second delay before the next message, "Starting Hardware abstraction layer hald". "Starting kernel event manager" is followed by a 5 second delay before the next message, "Loading hardware drivers". "Assembling MD array mdarrays" takes 10 seconds before the next thing, "Setting up LVM Volume Groups". (Surely MD setup should be near-instant, especially as there are no MD arrays? And I shouldn't have to tweak it manually to get rid of that step.) Most things take about a second. They add up to about 75 seconds.

    Then, after logging in at the Gnome login screen, if the cache is cold following a boot, it takes about 20 seconds to start all the applets, before it's possible to use the menu or icons. The panel itself appears quite quickly, but it's not usable until everything has appeared on it.

    Shutting down is pretty slow too. I usually do Control-Alt-SysRq-S-U-B/H if I want to power off completely or reboot, to avoid the shutting down time.

    -- Jamie

  62. Re:Ubuntu Feisty takes 95 seconds on my decent lap by mackyrae · · Score: 1

    Starting with Edgy my shutdown time has been definitely less than 20 seconds, and the "after login" time is under 10 seconds. When it boots, I usually watch the bootsplash, but if I switch it to text, everything just flies past. What are your specs? I have a Core Duo too. It's 1.6GHz per core (they're not all the same) and 1GB of RAM. It shouldn't be nearly that slow. File a bug report. I'm having issues on Sabayon. There's a 2 minute pause after "parkbd: no such parport" which doesn't make one bit of sense since I don't have a parallel port at all (laptop!).

    --
    look! it's a bird, it's a plane, it's....a girl? yes, a girl browsing Slashdot on Linux
  63. Re:Ubuntu Feisty takes 95 seconds on my decent lap by Jamie+Lokier · · Score: 1

    Similar here. 1GB RAM. 2GHz per core according to /proc/cpuinfo when busy. (Yet strangely Gnome CPU Frequency Scaling Monitor shows 2GHz (100%) and 1GHz (50%), at the same time as cpuinfo shows 2GHz for both). Shutdown time may well be comparable with 20 seconds (though it feels longer, I haven't measured). I boot with text, because I have encrypted /home and text is needed to enter the passphrase. Each little bit of the boot process is quite fast, except for the long delays at things I mentioned and a few others. The slow bits add up, times clearly shown in /var/log/boot, which is where I get 75 seconds.

  64. Re:Ubuntu Feisty takes 95 seconds on my decent lap by mackyrae · · Score: 1

    My /var/log/boot is here: http://paste.ubuntu-nl.org/11949/ (it wouldn't let me post it in here cuz it said "too few characters per line)

    --
    look! it's a bird, it's a plane, it's....a girl? yes, a girl browsing Slashdot on Linux
  65. Says who? by Anonymous Coward · · Score: 0

    I've used everything from mainframes to calculator watches, and Linux boots fastest of all, sans optimizing. I have also heard plenty of flames about Linux, but never about the time to boot. Don't judge all of Linux by one drag-your-ass distro. (Ubuntu, of course!)

  66. Don't know how your machine is set up, but... by Anonymous Coward · · Score: 0

    My Slackware machine boots up and KDE is completely ready in about 50 seconds (not counting the POST and the time it takes me to log in). Just disable the services you don't need running at boot (Samba, CUPS, ldconfig, fc-cache, etc.).

  67. MOD PARENT UP by The+Bungi · · Score: 1

    I wish I had the time and eloquence to reply to the "M$ windoze open source is perfect and i hate you if you don't agree" trolls this way.

  68. minit, anyone? by quigonn · · Score: 1

    Apparently, this topic has been discussed on /. before, and the most intriguing tool I became aware then was minit (http://www.fefe.de/minit/). Here's a nice posting about its most important properties: http://developers.slashdot.org/comments.pl?sid=808 17&cid=7115996

    --
    A monkey is doing the real work for me.
  69. My results. by Frozen+Void · · Score: 1

    I just shutdown and turned on my comp,then timed how much it takes to get win98 running.
    result is 24 seconds(from pressing Power on to MSDOS console windows).
    windows boots into MSDOS console window(shell=command.com).

  70. Mandriva using a parallel init since 2007.0 by buchanmilne · · Score: 1

    One of the contributors to Mandriva developed a drop-in parallel init system, which uses LSB tags in the init scripts to determine dependencies. This was tested in the development version for a while, until it was selected as the default before the release of Mandriva 2007.0.

    See more information on prcsys

    The advantages with prcsys are:
    -traditional serial startup can be used
    -minimal changes are required to initscripts
    -LSB compliance for free

    The Fedora wiki has some good discussion of why some aspects of new fangled init systems may not be desirable.

  71. This was fixed how long ago? by thogard · · Score: 1

    The syntax for the name field for inittab included a make like dependency filed sometime before 1981.
    So why does everyone want to do it the hard way?

  72. Novell by Anonymous Coward · · Score: 0

    If you're talking about "Novell Linux", yes. That's the goal of this project anyway. To make windows look fast and stable. This is how these people make money which is used to fund the Development of Mono and its widespread amongst hispanic-speaking people.

    Something like a custom linux setup, or LFS can boot you in bash in 2 seconds.

  73. Run levels by Skapare · · Score: 1

    You guys are still using run levels? When was the last time you actually used a run level different than the default run level, or changed the default, aside from doing emergency maintenance which can be done well enough with init=/bin/sh when needed?

    --
    now we need to go OSS in diesel cars
    1. Re:Run levels by Anonymous Coward · · Score: 0

      Actually, I use two runtime runlevels on my laptop, a battery and a AC power runlevel. When I get an ACPI event that goes to battery power I shut down distcc, nfsd, etc., and when power is restored, they get started back up. It's not the only way to do it, but it works...

  74. Update kernel without reboot? by bill_mcgonigle · · Score: 1

    Seriously, real Linux servers don't reboot :P

    Send me your IP address next time there's a kernel update for a network security problem. ;)

    I agree though, real servers shouldn't *have* to reboot. Time for my semi-annual CS taunt:

    Are there no CS researchers up to the challenge of working out an in-situ kernel replacement mechanism for Linux?

    Yeah, there's alot of bookeeping code to crank out, but think of the fame and glory associated with putting an end to reboots.

    --
    My God, it's Full of Source!
    OUTSIDE_IP=$(dig +short my.ip @outsideip.net)
  75. My Linux machines only take a few seconds to boot by Anonymous Coward · · Score: 0
  76. oops by Wonko+the+Sane · · Score: 1

    That actually should be:

    RC_NEED="foo bar"

    or RC_USE

  77. Reduce start time *and* RAM usage by justthinkit · · Score: 1

    Boot optimization is a subset of overall system optimization. As has been said, we rarely boot up these days (for me it is every week or two). But we do use our systems, some more intensely than others. In our household, two systems have more than a gig of RAM, one has 1GB, and the other two are 256MB and 320MB low-enders. Machines with 512MB or less benefit greatly from process reductions.

    (1) To get started, click Start, then Run, then type in Services.msc

    (2) Go through the list. If you don't need a service, disable it. If you are not sure, turn it off for this session only (i.e. click Stop). For example I don't have a printer on this computer, so I don't need the print spooling service ever.

    (3) Know your services. Many web sites will detail the services for you. Here's the first link I got on Google to get you started.

    (4) As others have mentioned, remove the "free" trash your computer manufacturer bogged your machine down with.

    (5) Turn off the "helper" apps -- you don't need the Microsoft Office "quickstart" thing, the Real player nor the Quicktime icons in your system tray. Etc. Get to know your system tray icons and what they are supposed to do.

    (6) If you are having trouble turning off system tray thingies, or want to do further tweaks, try running MSConfig -- Start, Run, MSConfig, {enter}. The last tab lists some of the things that are starting at bootup. Go through the list (MS doesn't make this easy -- the MSConfig window can't be enlarged, so you have to increase the column sizes, then scroll to the right and back again, etc.). You can get online help via Google -- here's another starter search link.

    (7) Run Task Manager -- its free with 2000/XP and it comes in handy, yet uses little CPU or RAM. I run mine all the time, with the "minimized to system tray" checked. This way it shows me when something is pushing my CPU to 100%. Anyway, run it and click on the Processes tab. Then click on Mem Usage -- at the bottom will be the programs using the most memory. Many can't be stopped, but the biggest offenders should certainly be looked at more closely. Also, this list should not run much more than one screenful. If it does you definitely have some more pruning to do. Some processes don't list all the RAM they are taking up -- e.g. my Opera browser can use up to 200MB, but never lists more than 80 to 100MB for itself. So, to perk up your system (if the free RAM is low), try closing and restarting Opera (this would probably work for Firefox as well).

    --
    I come here for the love
    1. Re:Reduce start time *and* RAM usage by thejynxed · · Score: 1

      Bootvis is important to use. So is gettig rid of excess fonts out of your C:\Windows\Fonts (or C:\WINNT\Fonts) folder. Everytime Windows restarts (at least with WinME to Vista), the OS will scan the font folder, open and preload every single font in the folder (one of the reasons why Explorer.exe takes such a huge chunk of RAM unecessarily).

      Another trick that can help is to strip out unecessary parts of the OS (like the POSIX subsystem if you aren't using OS/2 or *nix compatible programs, etc). In fact, in all of my years of using Windows, I've NEVER had to use useless cruft like NetMeeting or the POSIX subsystem for ANYTHING.

      Again, I trim off useless services, etc. just like you suggested. If you are a "Power User", you might also want to strip out the Help System and get rid of all of the useless .chm files as well.

      Another useless thing to get rid of: If you are only going to have your particular machine setup in your native language, you can get rid of about 400 MB of useless crap in the form of all of the foreign language files that MS stuffed into a typical Win install (and yes, that that OS scans through and loads parts of every single time you boot).

      I got rid of Dr Watson, Dr Watson32, Error Reporting, etc. All crap I don't need for my system. I use my system for light word processing, light spreadsheets, light gaming, IRC, P2P and web browsing. What do I need any of that crap for.

      Another thing I've noticed with WinXP: If you happen to replace the Explorer.exe with a different shell like Aston, LiteStep or GeoShell, boot times decrease dramatically depending on how you have any themes, etc setup.

      I've trimmed my times down quite considerably using these tricks and quite a few more I've picked up over the years (mainly from trial and error).

      Currently it takes me 20-30 seconds from a cold boot to a fully functioning desktop in WinXP, and that is with: Boot wait for choosing recovery console or WinXP partition, Webroot Spysweeper, NOD32, Sygate Pro, my BrotherMFC printer monitor and my Logitech keyboard & mouse monitor programs all loading at boot time. And yes, this means I am clicking links in Firefox and opening .PDF files in Foxit within 20-30 seconds of pushing the power button.

      It can be done, if you take the time to research various tweaks and things you can do to speed it all up, and apply them with a bit of common sense.

      --
      @Mindless Drivel: 100% of Twitter posts ever Tweeted.
  78. I have a computer that boots in about 5 sec by gosand · · Score: 2, Funny
    The only reason you can say that 50 seconds seems pretty quick is that most of us remember when several minutes was he norm.

    And some of us remember when it was about 5 seconds.

    Of course, that was a TRS-80. :)

    --

    My beliefs do not require that you agree with them.

  79. lsb way by mennucc1 · · Score: 1

    according to this wiki, LSB-compliant Init Script should document runtime dependencies.... and so it should be easy to use a parallelized bootup system; at a certain point, it was even planned for the new Debian Etch ; checking in my /etc/init.d, I see that many scripts in Debian Etch indeed do sport the LSB header; but I cannot tell if that goal was achieved - last time I tried , I apt-get installed runit, and then my box did not boot.
    The funny fact is that I did not see, in the posted article, any reference to this fact: according to that article, seems that, whatever parallel-boot-rc system you use, you need to manually write down all init-scripts dependencies ... thus effectively reinventing the wheel. Or am I wrong? is any of those gizmos out there capable of understanding the LSB-mandated dependencies?

  80. How about running fewer services by scruffy · · Score: 1
    Red Hat Magazine has a nice article describing the daemons that might be started at init. You might consider not starting some of them.

    http://www.redhatmagazine.com/2007/03/09/understan ding-your-red-hat-enterprise-linux-daemons/

  81. you gotta be kidding by Larry_Dillon · · Score: 1

    When an OS is capable of 440 days of uptime on a P90, who cares how long it takes to boot???

    --
    Competition Good, Monopoly Bad.
  82. Your mileage may vary. by WK2 · · Score: 1

    A bunch of /.ers are commenting about how Windows XP takes x time to boot, and "Linux" takes y. There are hundreds of different distros, you can customize each one as much as you want. Boot times WILL vary.

    I'm running Fluxbox on Debian Etch (Testing). The only things that boot are the stuff that was booting when I installed. It takes about 45-50 secs. Windows 2K, on this same machine, when I was using that, took about a min. When I was using Damn Small Linux, with a frugal install, it took about 30 secs to boot.

    I've never thought "Linux is slow." I've always thought that fast boot up times were one of Linux's advantages.

    --
    Write your own Choose Your Own Adventure. http://www.freegameengines.org/gamebook-engine/