Windows Fails 8% of the Time 913
descubes writes "A Journal du Net article reports that about 8% of Windows sessions require a machine reboot.
The relevant quote (translated from french) is: "The average rate of failures requiring a system reboot has been measured at around 8% per session. This number varies widely depending on the version of Windows. Windows 2000 has a failure rate of 4%, and NT4 is at 3%, whereas Windows XP is close to 12%." The study was originally made by Acadys and Microcost and gathered data from 1.2M machines belonging to about one thousand companies over a period of one month in seven different countries."
Perhaps is the user base of those versions? (Score:4, Insightful)
So, maybe the article tells more than the blurb, but it would appear to me that the reason that XP crashes more is that the people who are running it could be partly at fault (ie worms, trojans, poor hardware choices with outdated drivers).
Personally I use 2k at work and XP at home (for my Windows machines) and I can't remember a crash for either. Work is a bit of a stretch as I do shut it down daily but the XP machine hums along just fine without problems.
YMMV.
Somewhat misleading (Score:5, Insightful)
Re:Perhaps is the user base of those versions? (Score:2, Insightful)
Re:Perhaps is the user base of those versions? (Score:5, Insightful)
Re:Perhaps is the user base of those versions? (Score:3, Insightful)
I think at one point I had the server up for ~180 days straight, I was amazed at the totals in the "process run-time" in Task Manager.
Windows works great, for people who know how to use them. (Same can be said for Linux, Mac, etc).
reboot is necessary (Score:1, Insightful)
One of my friend had an uptime of 1 month with is XP box.
Considering he does a lot on it and that he was able to last that long without being forced by "setup wizards" to reboot, this is a record.
But when XP runs that long without reboot, it REALLY becomes unstable.
He showed me and I had never seen so much unstability. Every progs crash. That was terrible.
I do believe you don't need to reboot often, but it still is necessary.
Re:Perhaps is the user base of those versions? (Score:1, Insightful)
True, very very true, but honestly my Linux machine had crashed twice in August whereas my XP machine hasn't skipped a beat.
No way... (Score:4, Insightful)
At home, I play games, surf the web, write in MS Office...all of the typical things a normal user would do. Plus I do things that a "power user" might do. Newsgroups, Irc, nothing too great...and I NEVER reboot. I would say on average I need to reboot about once a month when Seti@home decides to get flakey or something. Does that count as needing to reboot...after a month!!?? Then I guess it needs to 100% of the time.
If people need to reboot 12% of the time, then they are doing something wrong. It's not the OS, but more the user in my opinion. XP is a stable system, and does a good job of keeping my machines running.
Win98, however, I would say needs a reboot 50% of the time. The other 50% you have no choice and it dies without a reboot.
Only 8%? (Score:4, Insightful)
Of course the big difference is uptime. My Windows (98) box has been up for 48 hours and is starting to feel sluggish, whereas my Linux box has been running for 4 months.
Re:A bit too high.. (Score:5, Insightful)
Given that XP isn't just Win2K SP5 but is in fact Win2K with an awful lot of extra chrome tacked on, it was never going to be more stable to begin with.
Re:Perhaps is the user base of those versions? (Score:2, Insightful)
Re:Perhaps is the user base of those versions? (Score:4, Insightful)
You say that like it's some accomplishment, well I guess it is for a WinBox, but in My World (*nix world) I would be very disapointed if my boxes where up any less that 180+ days!
My French is rusty .... (Score:5, Insightful)
Re:reboot is necessary (Score:1, Insightful)
Wait, I take that back. I probalby have rebooted once or twice in the past year with XP when things got werid. I'd hardly call that 12%.
Too bad the article is in French and I cannot know the polling criteria.
Define "require" (Score:2, Insightful)
Quite often it was an issue of restarting a service that "required" a reboot.
Then there are the times when the "required" reboot can be achieved by (heaven forfend) logging off and logging on again.
Windows 2000 was definitely better at cutting out spurious reboots than XP. Someone made a point about the user bases for the OSs being different... I would point out that a fair number of large corporations use XP Pro on the desktop, primarily because it is even more manageable than Win2K Pro under AD, which kind of sinks the idea that XP was designed as a home user's OS.
What really mystifies me is the low percentage of Windows NT4 sessions that require reboots... WTF.
I worked with that OS for years and that just doesn't seem right to me.
Don't mod this insightful! (Score:2, Insightful)
"needs a reboot" on a UNIX machine usually means "they released a security update for the kernel", or "the power went out". "needs a reboot" on a Windows machine (yes, I use one) usually means "it bluescreened", or "things are getting slow and weird". At work, it's a pain to get everything running again so I try to avoid it, but I still need to reboot about every 2 weeks or so.
To be fair, much of the instability is caused by shitty 3rd party drivers, but that still doesn't address the root problem of resource leaks and other bugs.
XP unstable? I think not. (Score:1, Insightful)
Nice title Mr.Taco (Score:5, Insightful)
No. 8% of Windows failures require a reboot. Big difference.
But what were they running? (Score:2, Insightful)
I'm fairly sure that if you left a Windows box up without ever touching it or running anything on it it'd work 100% of the time. It's all down to circumstances.
------
Guns don't kill people, rappers do!
Re:Puh-lease (Score:3, Insightful)
I think it's all how much you use it and how you use it. I run windows 2k servers and linux servers at work, and the win2k servers are fine as long as you don't have to touch them. That conflicts with MS's bug releases though. Everytime I update, I have to reboot. 9 time out of 10, the servers don't have a problem rebooting, but every now and then there's some failure that prevents it from operating correctly. I have had the same problem with linux also, but those are usually much easier to fix, since you can just pop out the drive and plug it into another machine (and not have to go through hardware detection again).
win2k3server is much more stable than win2k, but you still have the same problem with the updates. Rebooting a server to apply a security patch might not be a problem if you have one or two servers, but if you have a room full of servers, windows patching is your full time job.
At least linux will allow you to stop a single service, reconfigure or upgrade it, then restart the service. There should not be a reason to reboot a server to apply an Internet Explorer patch.
Re:Perhaps is the user base of those versions? (Score:3, Insightful)
- "The study was originally made by Acadys and Microcost and gathered data from 1.2M machines belonging to about one thousand
- companies over a period of one month in seven different countries."
These aren't home users, these are businesses that should have IT staff monitoring or looking at them regularly.Re:Biased (Score:2, Insightful)
Running an apache web server on a barebones linux box is very different from playing Doom 3 on an XP box with an Audigy soundcard and ATi 9800 Pro while also streaming iTunes over the network, etc etc...
Re:Perhaps is the user base of those versions? (Score:1, Insightful)
I have two XP boxes. The failure rate is probalby 90% of sessions or more. However, the only time I ever reboot is when there are problems. Otherwise they just stay running and logged in 24x7. It's like saying 100% of marriages end in death or divorce. I mean... duh.
Re:Perhaps is the user base of those versions? (Score:2, Insightful)
Leave the PC on but set power saving to something sensible, that way you can save money on electricity and be up and running in the morning without the obligatory 5 min wait!
Re:Perhaps is the user base of those versions? (Score:4, Insightful)
However, I've had a few servers that stay on all the time lose a hard disk after a restart due to power failures, or other infrequent power downs.
Re:Perhaps is the user base of those versions? (Score:2, Insightful)
100% of the time (Score:2, Insightful)
Yes, but what was running on the machines? (Score:4, Insightful)
Windows 98 (not SE) was less than this, I only rebooted my Windows 98 box every 2-3 months. About 2.5% of the time in that case. Windows 95 crashed 3 or 4 times a day
So, if you factor in adding patches, I maybe loose 1 hour of work per month due to faults with the OS.
I think the main reason my Windows boxes stay fairly stable is because I don't install a great deal of software on them. I only install Office (Microsoft), A virus scanner, Gaim, Firefox, Thunderbird and a few apps I need for my job. I also keep up to date on patches, and do housekeeping tasks like keeping my disks defragmented.
Most of the unstable Windows boxes I've seen are the ones that have been overloaded with a ridiculous number of apps, most of them the silly ones that come on cereal packets
I'm not saying Windows doesn't have its flaws (I think everyone would be happy to forget Me!), but if used sensibly it's not *that* unreliable.
As a comparison, my Linux servers have maintained a 100% uptime so far as crashes are concerned. The only thing that's knocked them out in the last 12 months has been due to Hurricanes. My Linux desktop (KDE), however, crashes about once every 2 months. So, from a desktop perspective at least, Linux is about as reliable as Windows XP.
Yeesh (Score:2, Insightful)
Windows (especially XP) is damned stable if set up right.
-ANY- OS is damned unstable if not set up right.
Re:Nice title Mr.Taco (Score:3, Insightful)
-
"Windows Fails 8% of the Time"
Well that's one interpretation, but I don't think that's the most direct one. Reading the article segment again ("Ainsi, le taux de panne moyen nécessitant un redémarrage du système est mesuré autour de 8% par session"), I would parse it as follows: "Ainsi, le taux de (panne moyen nécessitant un redémarrage du système) est mesuré autour de 8% par session". In english: the average rate of (bugs requiring a reboot) is 8% per sessions.No. 8% of Windows failures require a reboot. Big difference.
I think that means that 8% of the sessions encountered a bug that made a reboot necessary. After all, if they were measuring which fraction of bugs make a reboot necessary (as you are suggesting), why would they measure that "per session"? In that case they would not say "par session" but "par bug".
And, before you flame -- yes, I do speak the language.
Re:Perhaps is the user base of those versions? (Score:2, Insightful)
What's really amazing is how users who consistantly move their task bar to another edge of the screen and can never remember how to move it back, can get around all the safeguards of a managed system and screw it all up.
The computers we have here at the jail are locked out of almost everything. The only two things the average user can access is a terminal program and MS word. Yet these same people, who cannot tell the difference between a monitor and a CPU, manage to find every chink in the armor. They manage to gain access to the internet, play games, and in general screw up the whole operating system.
They may be idiots when it comes to computers, but never underestimate the users abilities to circumvent all your best laid plans of managing the computers they use.
Re:Perhaps is the user base of those versions? (Score:3, Insightful)
That is a funny joke, yes?
Sadly this is more true than anything. Everyone expects that the IT depts. are watching over everything and will fix all ills. If the (general non-geek) userbase realized how much they hold their own computing destiny in their own hands they would have far less problems.
-nB
You are wrong (Score:3, Insightful)
Now, upon what are you basing the assertion that XP is 1000 more stable than Win2K? My understanding is that both have a similar kernel design / driver interface. In fact, many Win2K drivers work fine on XP and vice-versa. It seems probable that Win2K is actually *more* stable, since it has had longer to mature and has had more service packs. Granted, most of those fixes have probably gone into XP, too, but the newer features of XP may not be as clean.
I have to agree with you about drivers in general, however. They are pretty much the only thing that has ever caused me problems with Win2k / XP. The one thing about XP that seems worse is its scheduler, which seems to lock up the system occasionally for about 5-10 seconds while using explorer.
Maybe you meant 2000/XP are 1000 times more stable than Me/98? Because that makes a great deal of sense. 2000 has never been considered an unstable OS, IMO, by those who know how to use it. XP simply continues the tradition, although I think it has dropped back a bit.
-Dan
Re:This has become accepted (Score:4, Insightful)
Easy is in the eye of the beholder.
People don't use Windows because it's better. They use it because it's easier. It's easier than having to learn something new. It's easier than having to install new software. It's easier than having to think about choices. It's just easier.
It's easier to reboot 12 times. Easier to just use Office. Easier to just reinstall the OS. Easier to just not care.
People don't vote because it's easier not to vote. Easier not to make up their minds... easier to just complain.
Change is hard work. Even if it's good change. Change is stressful even if it's change for the better. Change is not easier than just suffering with what you know. Learning is hard work.
Re:Perhaps is the user base of those versions? (Score:3, Insightful)
I would disagre, as most IT people have little more clue than users. I say this not to be a prick, but because it's true. I've met so very few people in the last 20 years that really, really know what they're doing and have good troubleshooting skills. How often do you hear from an IT person "Ooops, it bluescreened, that means it's time to reboot!"? No, if you got a bluescreen, that's a friggin' error message. Read it and find out what went wrong. I'd also argue that these clueless masses of IT folks think that reboots are the cure for all problems. If you're forced to reboot a machine, that machine has a problem....FIX IT! If you can't fix it, it's not MS fault, it's yours. You don't need the source code to figure out what's wrong, just a brain.
Boy, that ended up being more ranty than intended.
Re:Perhaps is the user base of those versions? (Score:2, Insightful)
Re:Perhaps is the user base of those versions? (Score:3, Insightful)
No, most normal people turn their appliances off when they're not being used. Home computers are no exception.
Re:Perhaps is the user base of those versions? (Score:2, Insightful)
Re:Perhaps is the user base of those versions? (Score:3, Insightful)
User Error (Score:2, Insightful)
Re:Perhaps is the user base of those versions? (Score:2, Insightful)
Re:Perhaps is the user base of those versions? (Score:3, Insightful)
Is there a babelfish something somewhere that'll translate BSOD into English? Because most of the error messages I've seen there have been spectacularly unhelpful.
Yes, there are two easily accessible: called "Google" and "The MS Support KB". I'll give you a hint: that long string of numbers in the upper left-hand corner is your error. Put that in the KB, and you'll generally get an answer or 25. You don't have to put in all the leading zeros (0x0000007b becomes 0x7b). Put that in with the word "Stop".
Well, the machines problem is probably simply the fact that it's running Windows.
No, the machine's problem is simply that its support staff has no idea what it's doing.
If a user is getting bluescreened once a weeek, it seems kind of silly for me to take the machine offline for half a day,
And thus the reason you're fired. If I had you on my staff and you uttered those words to me, you'd be out of there so fast your head would spin. How long does the user suffer every week because of those blue screens? 10 minutes? 20? Add it up and you'll see that a few hours work is worth the time. Regardless, if it's taking you half a day to solve a bluescreen issue, you shouldn't be working second level support.
Sure, if you're blueing regularly
Define "regularly". I would define regularly when it comes to blue screens as "once". Weekly is unacceptable and constant. Blue screens are a Windows last resort. Something has happened on the machine so bad that the OS has to shut it down lest any further damage takes place. That doesn't happen with a proper install. It's up to YOU to figure out what happened, and fix it. I've never once encounted a bluescreen that I couldn't fix, therefore the problem is with you.
chalk it up to Windows overhead, remember to save your work regularly, and get on with things.
Spoken like a true professional. I'm sure your boss goes to bed every night and thanks his stars he was lucky enough to land someone like you. It's your responsibility to fix problems. When problems come up, and you don't fix them, you're shirking your responsibilities. Grow up, learn and take some responsibility for a change.
Re:This has become accepted (Score:2, Insightful)
Macintosh? Pay waaaay more and can't run most wal-mart/etc software.
It's all well and good to delude ourselves into thinking there's a viable alternative, but for most people there simply isn't. How about focussing the energy spent bashing windows into making linux useable?
Re:Mod parent way up Up UP! (Score:2, Insightful)
How the hell do we force windows (xp/2k) to stop paging apps out to SWAP when it really doesnt need to, and also how to tell it NOT to cache so damn much, id like to configure caching based on folders/applications to define inclusions/exclusions just like a firewall. I want a firewall for my ram
I insanely HATE how firebox gets paged out to swap when not used and minimized, can mozilla team just hack/tweak their code someone to force most of it not to swap out, or use none-pageable ram allocations?
Situation 2.
XP with 256 meg ram, ZERO swap/vm. Boot up minimum services/setup, using 150meg free unused).
Why not just leave everything in ram, and page out stuff thats used least often based on historical usage not just the last few hrs. Dont cache everything from the HD, only really frequently used stuff and *ALL* desktop/menu ICONS, damn why is a 3ghz PC load 50 icons worth 100kb so slowly? Pathetic C++ code????? what is it?
Is it a case of "bugger it, 1gig ram is $100, cheaper than using good coding/design" ?
Re:Perhaps is the user base of those versions? (Score:2, Insightful)
Re:Perhaps is the user base of those versions? (Score:3, Insightful)
No, I use BeOS on every machine, both private and corporate.
Yes - but the blue screen is indicating that the OS is no longer stabile - you DO NEED to reboot.
No, the bluescreen is not "indicating", it is TELLING you that the machine (OS, software & hardware) is unstable and you need to find out what it is and fix it. Rebooting only gets rid of a bluescreen for a short while. If a machine bluescreens, the bluescreens will come back until you fix them. Prime example: my home machine started bluescreening. Read the error, did some research, and confirmed that my HD was dying. Drive Imaged over to a new drive, bluescreens stopped. Interestingly...I installed Linux on the old drive to see if the drive was recoverable, and the machine core dumped. What a surprise, not a problem with the OS at all!
MS error messages contained in blue screens is sometimes cryptic at best.
So? Most error messages are cryptic at best, that doesn't mean you don't take the time to learn how to read them, or figure out how to fix the problems they're indicating. To do so is simply lazy. And, MS doesn't hold a patent on cryptic error messages. Those from Linux can be twice as obtuse. Of course, their support is usually better: "If you can't figure it out, you shouldn't be using Linux" always solves problems.
Researching a blue screen message can often point to a bug in an application, or in the OS itself.
Well, yes, that would be the point of an error mesage. Funny that, huh?
I've had MS give me indications that I may have older drivers. Downloaded the newest vesions just a day or so prior. Sometimes the advice is just wrong.
Yes, it is. So? Then stop taking their advice and learn to fix problems on your own. I've called MS 3 times in the last 12 years. All three times they were a last resort and all three times they couldn't give me an answer. Save the dollars and learn.
There's not always something else that CAN be done on a Win box.
That's almost true. In the past 20 years, I've encountered exactly ONE problem I couldn't fix, and to this day have no idea what the problem was. It actually occured on two machines:
The machines were running Windows 98. The user would log in, and then get an error: "SHLWAPI.DLL has caused and invalid page fault in explorer.exe". Explorer would never load, and the user couldn't use the machine. (I'll save you the time, shlwapi.dll is a component of IE). Here's where it gets interesting: if another user logged into the machine, the machine worked fine. If the user logged into another machine, the machine worked fine. If I made the user a domain admin (temporarily), the machine worked fine. It was only if THAT user logged into THAT machine as a user. Profiles were not enabled in Windows. We did end up reimaging the machine, and she was the first to log into it, and got the same error. All previous conditions applied. I ended up just swapping her machine with her neighbor's since they were imaged and identical. Never did figure that one out, even when it occured to another user. Same solution, and it never happened again in the two years I was there.
In any case, that was an isolated and singular situation, every other problem, I've fixed. I don't leave a machine until it's FIXED. Reboots and defrags are not repairs.
You're certainly right, reboots are not the answer for ALL problems. But they are solutions (and pretty darm good ones at times) for the weird, occassional problem. Reboot, log-in, run your app, and you're fine for days.
No. You still have a problem, you haven't fixed anything. I tell my users: "if something happens, reboot once. If it happens again, call us. If it happens again within the next six months, call us. Otherwise, it's a one-time random thing. Anymore frequently than every six months is not rand
Re:Perhaps is the user base of those versions? (Score:5, Insightful)
It is essential that our servers stay up all the time, that goes without saying, but I also leave my laptop on all the time. I hate having to reboot as I lose the state of my work environment: what files I had open and where, what logs I was tailing, the specific command history for terminals, what web pages I had open, etc. I reboot typically on average about avery 3-4 weeks, usually because of an OS patch. I could use some other technologies (like using VNC connection to a server that is stable) but I don't have to and I really like it.
HTPCs (Home theater PCs) also require almost indefinite uptime - you don't want to have to boot a PC to watch TV or a movie, and a crash while watching a movie is not acceptable. I've heard many stories on HTPC forums of people spending days and weeks attempting to track down causes of intermittent crashes. As HTPCs get more prevalent I can only see long uptimes getting more important.
Re:No way... (Score:2, Insightful)
The story showed the facts - that a research carried out (with the help of MS) on 1285000 computers showed that among those with WinXP 12% of sessions required a reboot, while the numbers for other OSes was less. That was a fact. Now your comment that you "find it hard to believe", followed with some personal anecdotes, is idiocy. There is time and place for personal experiences and this time is not when you discuss a large survey. There is zero relevant information in your post, because it's obvious to everyone that there are always probability distributions - there are computers that require a reboot more often and those that require it less often. Any intelligent person realises that there are more factors than one (OS) and that hardware, software, usage patterns, networking environment, etc. are relevant too.
But facts are facts - if a random user (assuming the sample was representative) has WinXP, he is likely to reboot in about 12% sessions. A Win2k user is likely to reboot in about 4% sessions. Your comment is irrelevant. Period.
Re:Perhaps is the user base of those versions? (Score:5, Insightful)