Slashdot Mirror
Server Uptimes Ranked
Bex writes "Ever wonder how your server uptime is when compared to others with
different operating systems? Ever wanted some hard numbers for the
Linux vs NT or FreeBSD debates? Check out uptime.hexon.cx
for a list of servers and some interesting number on uptimes. It looks like FreeBSD stomps all over everybody else, with a whopping 1994 days of uptime for one server, and a 138 day average uptime. NetBSD is second for max uptime, but better on average. Windows 2000 is second to last, just barely beating out BeOS with a paltry record of 49 days. Its average uptime was under a week!
Remember, downtime doesn't always mean a crash, but is is a good indication of how often a machine needs maintenance." Update: 12/30 10:33 by R : There's a new version of the uptime page here.
Why not? because NOT THAT MANY home machines are hosting websites, because Joe AHU (Average Home User) doesn't have a static IP address -- he's got a dial up connection. So virtually every point made in this post doesn't add up to a lower uptime for windows.
The uptime counter bug mentioned by other posters and the BSOD (blue screen of death or black screen of death) are what accounts for it. That and the fact that most of the Windows operating systems have historically had serious memory leakage problems... [memory leak: when the OS loses track of which RAM has been allocated and/or deallocated to or by a specific program)
...Open Source isn't the only answer -- but it's almost always a better value than the alternatives...
Given that IBM's OS/2 and Microsoft's various Windows operating systems shared a similar code base, doesn't it seem weird that OS/2 never seemed to suffer this problem?
In fact, it did.
In 1997, a team at Ford Motor Company had noticed that, after 6 to 8 weeks of operation, our OS/2 v2.11 machines would begin executing once-per-day tasks several times during a day, and after several such executions, the computers would crash. Further analysis showed that those tasks were being executed every 1h09m54s. I spent a day trying to manipulate that number into something meaningful, and gave up in frustration. We assumed that our own code was to blame, and rewrote it several times (to no avail).
After rewriting our code seemed to have no effect, we decided to install the latest set of O/S patches on our machines. On a Sunday, we moved between machines that were scattered over a several hundred acre manufacturing facility.
Black Monday
Seven weeks and one day later, the facility started building units. Within hours, the OS/2 systems started showing the symptoms that led to the crash, in the exact order in which we upgraded them!
The coincidence was too much to escape notice, so we called IBM Technical support. Their Level 1 guy spent about five minutes talking to us before he realized this was a deep O/S problem, and we were kicked to Level 2 support. The Level 2 person heard our version of the events ("The machine flakes out and every 1h09m54s executes a task that should only happen once a day"), and asked, "are the machines crashing 49.7 days after being rebooted?"
BINGO!
Apparently, someone inside IBM had noticed this problem a few weeks before we did, and had the patch in final testing when we called. (I think the patch was #XR2011 or #XR2014). However, since we were a customer, our bug report took priority over theirs.
The Problem
Someone used an unsigned 32 bit integer to record the number of milliseconds since the O/S was booted. That number rolls over after 49 days, 17 hours, 2 minutes, 47.296 seconds. The symptoms we saw began the day previous to crash day due to the rollover that occured when our code scheduled a task for "tomorrow".
The Moral
It's too bad that Microsoft and IBM were not on speaking terms at this point. If they had still been working together, MS would have had a fix two years ago.
Russ B.
Chivalry is not dead, it's just frequently misspelt. - M. Langley
What's that, a pico-HOWTO?
You think that the average workstation has the same uptime as the average server? I'm not refering to possibilities, I am refering to real world usage. Which is what I am questioning in this survey.
This "kid" does'nt need his 10+ years in electronics/computing to know the above.
So by your definition, a Workstation can instantly become a Server by just leaving it on. So I don't have to enable any services, no ftp or http daemons? No SMB or DECNet? Not even a little Appletalk or IPX? Nothing but an untouched power switch eh?
So you gunna give me more info on these 2 incredible OS' you have written or are you going to remain just another AC kiddie?
War crimes, torture, lies, illegal spying... Would someone give Bush a blowjob, already, so he can be impeached?
Check out their selection of clients...
No MacOS - I know it's not great, but I bet a machine running appleshare server would do pretty well...
Only machines which have either binaries available (SunOS, Linux, Windows, FreeBSD) or PERL are eligible to partake.
Would someone running OpenBSD be running PERL for instance? (is it even included???) Or run an unknown binary?
How about Irix boxes that just sit and calculate all day long... Why would they want perl on their system, let alone an unneeded program?
This project is a JOKE. There is no way in the world that this can be subjective... Notice the heavy skewing in the number of linux boxes on the list vs others... Now look at the real world... All those NT boxes are probably admin'd by people who have no business setting foot near the console of one of those boxes.
Next pointless survey, please?
Attacking a person on his abillity to produce texts in a language that isn't his first is below the low level of courage to post with your registered name.
Never underestimate the relief of true separation of Religion and State.
Other |96 days, 06:41m |769 days, 22:37m
There are 33 users of this OS which is doing quite nicely, wonder what it is (i.e. what is not mentioned in the list), could it be DOS??
Windows 9x users are quite likely going to turn their machines off at night, so I don't think comparing them against the likes of NT etc is really fair.
However, on the 95 vs 98 thing, I have noticed that if I leave my machine on overnight running 98 then it tends to have died by morning.
>I think the answer is that the average Solaris
>admin comes from an NT background
I'd like to know where you came up with that.
Also -- if, as you assert, certain OS's attract various levels of competence then wouldn't an IS decision maker who wants the >best uptime still choose the OS with the best track record? After all, if BSD sysadmins are so much better than Solaris admins (who are clearly just washed-out NT admins) then why not deploy BSD and hire the technically superior BSD admins?
[facetiousness intended]
-- "In order to have power, I must be taken seriously." -Mojo Jojo
Unfortunatelly, I know you are correct. :(. Let's hope in the future people don't have the possibility to use a desktop as a server.
Never underestimate the relief of true separation of Religion and State.
you think that the open source reverse engineered drivers for linux are better? cool
---
Quite recently, my FreeBSD-based development box was rebooted. Violently. The machine had been up for about 320 days, when somebody accidentally removed its power cord in a co-location facility. I had a ssh session to the box at the moment, and pretty damn near shit in my pants when the machine stopped answering. I got to sigh in relief in a few minutes, after the machine booted and got all its services running.
Which gets us to the subject of the day. The reason I was so worried was that after 320 days of tinkering, new IP numbers, new system software, new services, old services scrapped, etc, I had no idea if the server would come up on a reboot or not.
Frequent, comfortably scheduled reboots tell you whether your machine will come up after an unexpected interruption or not.
That's pretty much the only direct problem with long uptimes. However, glorifying long uptimes has other, well known drawbacks. The most important of them is, of course, putting up important kernel security updates in pursuit of long uptime.
Marko Karppinen
www.uptimes.net doesn't seem to be loading. Maybe their server went down. ;) (hahaha)
Quite. And you'd do it in a quiet period, as quickly as possible, having given as much notice as possible, and moved as many services as possible to alternative nodes.
If I wanted to win idiotic uptime competitions, i'd just leave a box in the corner doing nothing save being up. But uptime by itself is not a valid measure of the reliability of a system.
It measures nothing. In particular I really doubt that most Solaris people come from NT backgrounds. Rather, most people running serious machines will simply not bother to install some random daemon to let other people know information that they don't particularly want to give away. I can see Solaris machines with 285 day uptimes from here, and they aren't particularly special.
I also must take issue with the `uptime being a point of pride' thing. If the machine doesn't have any particular state (say it's an NFS server), and it's going to take some time to work out what's wrong with it, the professional thing to do is just reboot it: only some idiot who isn't accounting for their time properly is going to spend half a day trying to work out what's wrong with it without rebooting first to make sure it's not something transient
This survey is worthless.
Yes!!! The daemon rulez ;)
if you want to make God laugh, tell him your plans
HUH?!?! Any sysadmin worth his/her salt would know that you should only bring down a production server if NEEDED. If you don't need the latest kernal patch, say it fixes a bug in a driver for card XYZ, but you don't have card XYZ, do you need to patch? Heck no. Keep the baby running, and your clients/users happy, by NOT patching just for the sake of patching.
Maybe when you say "up to date patches" you mean patches actually needed, but if you really need good uptime you could design the system around drivers that are very stable and unchanging thus reducing the chance that you will need to patch.
IIRC you can still find some pre-1.0 linux kernels out there running. Why? Because the system requirements haven't changed, if it is stable and secure and does what it is supposed to do, don't mess with it.
--
?
First of all, I have been using Linux for 2 years as a hobby and DecUnix and DecVMS for more than 4 years whilst I worked as a PC/Telecomms tech programmer for my local Stock Exchange. I also used Microsoft products.
When I state service, I am refering to a service that is served back and/or forth between the server, other servers and their client machines via some transmission line, ethernet or whatever.
Sure, a Mac can be serving a user Photoshop but this is not what I am refering to as being a server.
The common defenition of a workstation is one that has its CPU and disks used by the person in front of that particular computer, usually by the same person and usually for 8 hours a day. For the other 16 hours of the day, these machine are usually turned off to conserve power. Unless of course they have shared file systems or printers that are required to be up 24/7. But for the most part, this is not the case. Any decent sys admin will have this type of task delegated to a machine that is specifically set up to serve these tasks, which is good for backup ease.
The servers on the other hand usually stay up for obvious reasons. Mail serving, file, print, web, db, whatever.
"All computers are nothing but mere terminals jacking into the great global network."? You've been reading too much Cyberpunk fiction buddy.
But if it impresses you, my videocard'less OpenBSD firewall/NAT gateway is accessed through our network via its serial port through a DECServer700 from anywhere in the network which has a ssh client. (Whoops I better be carefull with my usage of the word client there, I would'nt want to confuse another AC).
War crimes, torture, lies, illegal spying... Would someone give Bush a blowjob, already, so he can be impeached?
Wow. This sort of surprises me. I know some of you out there did lots of the set-it-and-forget-it Netware 3.12 systems like I did. ;) You never heard back from the client - the server just didn't go down. I have a Netware 5 box with 400 days of uptime back at work and another 4.11 that would be up there too if it weren't for a 7 hour blackout. Ah, but alas, this is /. Linux rules the day. :)
Someone hasn't screwed up.. as you can see.. IF you read the page there's absolutely nothing scientific about this uptimes project... IT's a project for fun.. not for research.. but most slashdot readers seem to be missing the point... COMPLETELY..... Tgm just wrote something that wasn't there yet.. and a lot of people have joined his small project.. If the slashdot readers would read the pages before commenting on it then the world would be a better place... I think! :)
Windows 2000 is second to last, just barely beating out BeOS with a paltry record of 49 days.
Let's at least be fair here. Windows 2000 has not even been released yet. 49 days ago they were at beta 2 or something like that. Let's at least wait until it's released before we start bitching about how much it sucks. I mean, we don't want to sink to their level and start spreading FUD now do we??
the server is available for download on the uptimes page
I don't know about others, but I've never had to reboot my BeOS system because of a crash. That being said, what exactly is uptime? I mean, I have had the network server crash on Be, but, because of the micro kernel design, I just restarted the network server. No reboot. Is my machine still "up"?
Also, I imagine most people running Be also dual or triple boot. This might explain the short uptimes listed. I'll usually leave my system in BeOS unless I'm banging out some Java or playing some games with a friend.
well the best test for BeOS uptime would be it doing real work such as maintaining a theme park http://www.lcsaudio.com/installations.html (sidenote: wow! I didn't know Hayden Planetarium will use BeOS. My brother wanted to go there on a date only to find that it was closed for renovations)
---
I don't think it is really fair to bag Windows 2000 for having an average uptime of 5 days. Don't forget this is a cutting edge MS operating system, and you are going to need reboots for upgrades. It would be fairer to judge that six months or so after the release.
I do wonder about the 49.2 max uptime for Win2000 & 95, though. There was a bug in Win95 that would crash it after roughly that amoutn of time (Can't remember teh exact number of days) - the Win2000 uptime looks suspiciously close to that, too.
I was a little surprised about the BeOS stats, too, until I realised there was only 4 BeOS machines in the survey. No Macs, either.
There is also no way to compare what the machines were doing. A hardcore development or games machine is much more likely to crash (or reboot) often than a machine doing nothing.
Conclusion? Interesting, but don't read too much into the results. It is nice to see some of the really high uptimes, though.
Please update the story on the main page.
The Uptimes project has moved to http://www.uptimes.net/
Also the protocol has changed, so everyone going to http://uptime.hexon.cx/ will be downloading and running old clients.
Come on, Slashdot people, research a story for 5 seconds before you post it.
Phillip
Here's what I see as the technical problems with the "uptimes.net" approach:
These are all fixable, and fixing them would answer most of the criticisms given in previous postings.
I'd just like to share that, ironically, I couldn't view the list due to an internal server error for the page :)
I suppose you are right (well, you just are:) but what I meant to say was that i think it would be great if the community would work together on a project which would do this as scientifically as possible.
indierock / punkrock band photos and more... http://www.digitaldefection.net
what are you incapable of downloading file off of the internet?
Gentleman, you can't fight in here, this is the war room..
who sez death can't be funny....www.endlesssorrow.com
what are you incapable of downloading files off of the internet?
Gentleman, you can't fight in here, this is the war room..
who sez death can't be funny....www.endlesssorrow.com
The operating system is only one (small!) part of what makes a reliable system. Staff, procedures, and methodology are all more important. The most reliable OS in world will still crash if in the process of rushing a system into production, someone decided to run the power cord across the floor and someone trips over it (don't laugh - this happened to one of the places I worked before). Don't get me wrong, it's cute and fun to compare uptimes (my NT PDC and BDC have been up for 169 and 183 days respectively), just don't expect those numbers to correllate into the real-world when it actually comes to getting work done. If you plan, test, and implement properly and consistiently, any modern server OS will perform adequately. Always remember - the right tool for the right job.
Correct.
VMS is by far the most stable general purpose operating system, and puts Unix and Windows to shame.
At DECUS '97 some power company announced that they had a VAX which was up since 1985, which easily stomps out FreeBSD and the other little Unix computers on this little survey. Show me a Unix machine which has been up for 12 years!
Just that the 49 day thing does not happen everytime (or at all for your boxes) does not mean it does not exist. I had it happen for my SMP workstation (NT 4 SP 5 (at the time)). To be fair, I've also had a UP Linux 2.0.34 box to crash after 497 days when its jiffies wrapped around, and that does not happen for every 2.0 box, either. (NT uses 1024Hz, Linux 100Hz, both have 32bit counter.)
-- v --
...is the Mac. I know they are not great for high-end stuff (but just wait for MOSX! MWAHAHHAA!!!!) but theyre really good for a cheap low to medium load server, be it mail, web, print, whatever. Set it and forget it. Almost no maintenance. I have an old 7100, on system 8, that has been up for at least 5 months, running my website. Last time I had to reboot it was, like so many others, a power outage. When I get my G4 in January, my G3 will replace the 7100, which will be cool.
The one and (thankfully) only,
LafinJack
we are building a religion
a limited edition
we are now accepting callers
for these pendant key chains
I haven't administered Novell boxen for about a year and a half, but with Intranetware 4.1, my uptimes were on the order of hundreds of days.
Novell 4.0 was a buggy damn thing that I had trouble keeping up for a week, but with 4.02 and later Intra 4.1 they got the stability thing down.
I haven't had any experience with 5.0, but I would hope they can do as well.
I wonder why we don't see any uptime info for Novell at this site.
Russ
War is Peace. Freedom is Slavery. Ignorance is Strength. - George Orwell or George Bush?
I'd just like to point out that BeOS hasn't been in existence for 1994 days yet. About 1990 days ago, it was running on the Hobbit processor, and people were talking about the soon to appear PowerPC version.
That, plus the tiny number of registered systems, the dual-booting nature of a brand-new OS, and the fact few use it as a server.
The info is just not useful.
Cam
- Cam MacLeod
I am a contractor at Microsoft, and I know the facts on this one.
1) Internally at MS, we got the Win2K builds as early as possible on the corporate network
2) RC3 was build 2183. RC3 is not 49 days old.
49 days ago, I believe we were still at build 2145 or so.
God damn it. All BeOS needs is jackasses like this Bex guy spreading FUD.
Wake up people!!! BeOS is a _DESKTOP_ OS, not a server OS. Uptimes are completely irrelevant.
Christ.
What OS have you written AC? It does'nt take an OS coder to know the difference.
You can't rope me into YOUR argument AC. Obviously to everyone but yourself, when I state "server" I am speaking of a computer which has the role of serving services 24/7, and when I state "workstation" I am speaking of a machine that is used by individuals on a daily basis which is usually switched off at night.
The relevance in what I have said, is not in the choice of OS but the application of it.
Get over it.
War crimes, torture, lies, illegal spying... Would someone give Bush a blowjob, already, so he can be impeached?
Running as a router and NFS server I got 90 days out of Linux before the power was shut off by construction. On SMP however, forget it. NFS crashes Linux SMP instantly.
I have two PC at home:
assert(expired(knowldege)); core dump
Keep in mind that "kill" doesn't kill a process. It tells a process to commit suicide. A process that's stuck waiting on I/O (which sounds like what you're describing) will not be killable on any Unix or Unix-like OS. A process must run to die.
The only thing special about -9 is that a process can't trap it.
Computers are useless. They can only give you answers. -- Pablo Picasso
You missed the point. When you fly off to a convention, I assume you take your computer with you (a laptop, presumabaly). While you and others might be in the fractional minority, most home machines are not in that category. For example, I'm temporarily stuck on a WinXX machine at work, and have both a Linux machine and a WinXX machine at home. (gotta take care of wife and little kids, and there's not that much good Linux educational software for seven year olds -- yet.) The Linux machine stays up because it is hosting the site(s). The WinXX machines get shut down at the end of the day or session. But since they are not serving websites they don't count against the uptime average for WinXX machines.
...Open Source isn't the only answer -- but it's almost always a better value than the alternatives...
Yes I believe you are right. In fact the posts are so bitter the poster just looks stupid instead of the more desirable superior that s/he was going after.
Also one has to look at the pure number of clients on uptime.hexon.cx for each OS. Linux has such a low average time because there are 200+ clients on uptime instead of the, what, 30-40 for the BSDs? This gives a much better sampling where a few abherrant ones don't skew the data to far off.
As for the top dogs I've heard that one of the two is just copying his uptime between reboots since the current version of his OS wasn't even out when his uptime supposedly started.
Finally, as for Windows, you won't find any windows above 49 days, 17 hours and some odd minutes. Remember, Windows has the 49 day bug!
-- Grey d'Miyu, not just another pretty color.
I was once led to believe they had a FreeBSD varian. Anyhow, I had a 9.X series code router that was up for about 5 years until it had to be swapped out for y2k compliance.
First off, I'd like to appologize for this slightly off-topic comment. However, a few months back, a friend of mine and myself were discussing how misguiding the term 'Uptime' is. Say I am running a corporate server on Linux or BSD. I change a setting on it and suddenly no one can get into webpages outside of the firewall. The computer is still on, but it's not doing jack. I don't consider that up.
I personally feel that uptime should be rewritten, to not just calculate how long the computer itself has been on, but how long the server has been able to provide the services it's designed to provide.
The glorious BeOS has never crashed on anyone I know of (perhaps 10 hardcore BeOS users). Besides the fact that only 4 BeOS boxen were in the survey, most people who use BeOS also use another OS. I for one have Winblows 98 on here so I can play some phat FPSs. My uptime is very seldom more than 24 hours, often its 19 just before I reboot. Do any other BeOS users feel this is why the BeOS uptimes are pretty low? I know if Q3A were out for BeOS (i've seen it running on BeOS at PC Expo, just waiting for it to be released) then I would almost never need to reboot, except to play Worms Armageddon.
Too bad they don't have a Netware client. I had Netware 3.xx servers up for years and our new Netware 5 server only went down after 6 months because I had to replace the batteries in the UPS.
Netware is truly an underappreciated OS.
VMS is the uptime king with systems out there that haven't gone down in years. The only time my personal VMS system went down was during a power hit or when I moved offices.
What's my Karma Mr. Burns? "Excellent"
I took the DeCSS stuff off of my site and replaced it with a link to OpenDVD.org, which has a much better tarball of _all_ the linux software they could scrounge up pertaining to DVDs. Hope it helps.
A lot of people are considering these results skewed. I would have to agree, but let me give a real-world example:
I have an older pentium box (P200 non-MMX) with some fairly standard hardware. Diamond Stealth II 220, Diamond Monster I, SB PCI 128, Intel Pro/100B, 64 M Ram, ~2 gigs of IDE and a 32x Pioneer CDRom. Tyan 1573-ATX Motherboard.
Nothing bleeding edge, fairly competent box for any OS. I installed Windows 98 on it, and it crashed MULTIPLE times EVERY day. At one point I removed just about every PCI card (except video) and tested that - got about 2 days worth of uptime before it barfed on me again. Installed RH6.0 on the SAME hardware and it NEVER, not ONCE dumped on me.
So, yes, take the statistics with a grain of salt but DO remember that there are REASONS why the Windows systems uptimes SUCK - and those of Linux and *BSD don't.
der dee der.
That comment isn't funny, informative, insightful, or anything other than childish.
Oh yeah, that's fair.
There are a shade under 300 machines registered to this project. Over 250 of them are linux boxes.
It would be interesting to see how Compaq's (nee Tandem) Non-stop servers fair, since supposedly they can have the OS upgraded without stopping.
Oh yes, even the kernel can crash at 497. Been there, had that. Of course, it won't do it every time. For me, it happened with 2.0.34 - Alan says it should be pretty much fixed in 2.2.x. As of the strange things in userland, they already begin taking place after 2^31 Hz (298 days), because some programs use jiffies value as signed. There were problems at least with times() (both in glibc and kernel), select() (kernel) and procutils .
-- v --
Did anyone else notice that number 1 ("topsecret") has been "up" longer (1994 days or about 5.463 years) than the FreeBSD (2.0) release which it runs has been available (according to http://www.freebsd.org/releases/ it's only been available since November, '94 or about 5.167 years)?
I think I remember reading somewhere the the engineers eliminated 200 instances in which a reboot is necessary with Windows 2000 (from firsthand experience, stuff like changing network settings no longer requires a reboot).
And there are now ways for software developers to dynamically replace DLL's without requiring a reboot. If a software package requires a reboot in windows 2000, then it wasn't designed propperly.
Hell, I've had to tape over the power switch on production Unix servers. Just depends on who your junior admins are. :-)
Neither is having the latest kernal patch. But your argument for why long uptimes are idiotic was so you can put in the latest kernel patch. I say if you pick your hardware right you can find very stable drivers that shouldn't require you to need to patch the kernel thus you get high uptimes. Which can be a vaild measurement when also looking at what else the server was doing. Especially if uptime is important for the application. The ability to move services helps make this less of a requirement, but also adds a lot of variables that can cause problems (Oops somebody forgot a little piece of the payroll system needed the machine, doh, nobody's getting paid this week). So isn't wise to take some preventive measure to reduce the likelyhood of needing a kernel patch?
Any measurement by itself is rarely valid proof of the greatness of the measured item. Oooh this plane does Mach 10, (yeah but does it kills the pilot?). Oooh this car does 0-60 in .00001 sec (yeah but can it stop?). Oooh I'm running kernel 100.75.99 (yeah but what did you have to go through to get there?). Oooh I have 9 million days of uptime (yeah but was it doing anything?)
They are all idiotic statistics when looked at individually, but by gathering various stats and looking at them as a whole you can get an idea of the reliability of a system.
--
?
Huh? I've run SP5 on two very heavily used SMP workstations (both have SP6 now) and SP6 on one more SMP machine. Had no problems with it. I think your problems are not related to SMP.
-- v --
Now compare this to the 49 day high of Win NT (40% improvement) achieved from a pool of seventy-two NT servers (1800% more machines) and BeOS is pretty much doing as well as NT. Especially when you consider the amount of NT trouble-shooting being done by both MS and the community in general - When was the last time you saw BeOS mentioned on Bug-traq? Nt does manage to improve on BeOS' average uptime though - twelve days... Must be why NT admins get uptight every fortnight... :)
Welllll... the uptime apps that run on NT/2000 tend to use the GetTickCount call, which wraps on its long-integer limit at 49 days...
... so the counter's buggy, that's it.
Sucky, sure, but this means that it doesn't reflect the real uptime of the machines AT ALL - these figures are bogus.
Simon
Coming soon - pyrogyra
The correct way of doing it is
1) Something's wonky? Let's reboot, see if this fixes it
2) Hmm, it DIDN'T fix it permanently, *now* I will spend some time trying to figure out what's really wrong.
PS: Next time you've got something worthwile to say, don't say it anonymously. It just makes you look like a dork looking for someone to flame. If you've got something to say, stand by it or shut the fuck up. Thank you.
-m
99 little bugs in the code,
99 bugs in the code,
fix one bug, compile it again...
Voting Moo Anyway!
You want analysis...
There are 400 people relying on this machine for name service. Something is wrong with named and/or NIS on the machine. None of those 400 people are doing useful work right now, and this is conservatively costing $40,000 an hour.
What do you do: sit for an hour (or 10 minutes) and try and work out what is wrong, or reboot it now with the strong expectation that the failure is transient and it will just come back, and then try and work out what was wrong.
It's extremely tempting to do the first, but experience shows that machines just get fouled up sometimes -- the implementation languages we choose make it very hard to write systems without memory leaks and so on which cause long-term lossage. Quite often the professional answer is `just reboot', because quite often that will fix the problem and give you a chance to file a bug report and/or grovel through the source without it costing thousands of dollars an hour.
I have a web server that is a PII 233 128 Meg RAM machine running Mandrake Linux 6.0 with kernel 2.2.9.
The web site is delivered via Php3 and MySQL and Apache 1.3.6.
Currently the server has been running for 121 days! To see for your self goto http://www.jhs.jordan.k12.ut.us/about.php here.
I have seen a 2.2.5 box last for 197 days running ftp, ssh, and samba. After 197 days, it was not responding, it worked at night, and was dead by morning. 197 days is better than the hard coded 49 days in Mucrosift Wenders.
Lars -
Does this take into account that Linux users must, at any cost, spin a new kernel and reboot their machines every time a new tarball appears on ftp.xx.kernel.org?
As for your post, I've said it before, and I'll say it again... If you've got something to say, either stand by it, or shut up about it. Posting anonymously makes you look like a tard.
-m
99 little bugs in the code,
99 bugs in the code,
fix one bug, compile it again...
Voting Moo Anyway!
Is it me, or is there a single anonymous coward running around and trying to insult each person who posts?
I say this because rhe insults aren't particularly inspired...
- Jeff A. Campbell
- VelociNews (http://www.velocinews.com)
- Jeff
Out of the four NT workstations we have here (used for normal desktop work) at least three have easily beaten the 49 day mark. And at least one '95 machine has been running since June.
These figures (the max ones) cannot be accurate.
I am a sysadm, having a few NT servers, couple AIX boxes, and a bunch of Linux boxes.
/etc/rc.d/init.d/whateverprogram restart
I agree that on NT, rebooting is often the only way to fix things. And in many cases when you change some configuration, you are forced to reboot. I see this forced rebot as a way to force the sysadm to clean up the leftover junk from MS apps.
On a unix box (including one Linux box running 358 days on a 486dx2 named 'crasher' when it was running Windows), reboot does NOT solve the problem. A reboot might delay the time it takes until the symptoms are there again, but for suredoes not solve the problem.
I do not see uptime as a pride thing. But I do see uptime as something related to general stability of the OS. If an application can force you to reboot, I would not call it a good multi-tasking protected memory OS. I especially hate how user space apps can kill NT (NT 4 and 2000 are worse than Win 3.51 as more junk drivers are running in priviledged mode).
On a unix box, doing a
or a kill -HUP usually solves the problem.
I see average uptime as some indication of OS stability. The only valid reasons for reboots are hardware upgrades and kernel upgrades. And kernel upgrades are usually a sign of OS instability or something else. So the only thing that is really an excuse for reboots are hardware upgrades.
I DARE you do put a NT4 machine in the closet for "an idiotic uptime" whilst doing nothing. It will die.
Lars -
:)
Just how many AC's are pulling my chain here?
War crimes, torture, lies, illegal spying... Would someone give Bush a blowjob, already, so he can be impeached?
A U*x admin that has managed to avoid reboot on a production system for so long most likely has not left any known or even supsected h0lez for R00ting...
I have no idea how any admin would do that when it comes to kernel vulnerabilities. Crystal ball, perhaps? Most exploit principles are not new, but the techniques certainly are.
Most holes are at the application level, which is all well and good, and as it should be. They can be easily fixed without a reboot. Exploits of kernel services (tcp stack, for example), require patching and rebooting, and they're not unknown - not to mention that fixing them is generally outside the job description and specialized knowledge of a sysadmin.
Now, it's possible to secure your boxes from outside attack through firewalls and the like. Perhaps even to the point that the weakness is unexploitable from the outside. The feeling following such work is known as hubris, the pride that goeth before the fall. Are you willing to make the claim of total security for every machine within your network, or every employee you have or have had? All these things are potential vulnerabilities; a machine secured to the outside, as you can see, isn't really secure at all.
--
--
There is no premature anti-fascism. -Ernest Hemingway
All the servers not using NFS for any reason have outstanding uptimes. Those that do use NFS, .. well, they don't.
I find myself constantly needing to reboot system because of some NFS snafu. The most common problems are occasionaly a server will stop accepting nfs mount requests from a client. The clients say RPC Timed Out. This prompts me to restart the NFS service on the server:
Sometime it works sometimes it doesn't. Sometimes it says "Bind Failed, Port already In Use". Huh?
Another favorite is trying to restart a system that has any NFS components running. On Redhat systems the shutdown comes to the point where it says Unmounting NFS volumes and just hangs, ... forever. Is it just me or does anyone think this is completley unaceptable.
Sometimes the NFS frustration on Linux is so much that I feel prompted to write a book and title it "Why Does NFS on Linux Suck So Badly", but I don't think it would get published. Maybe a book called "NFS Annoyances" and have O'Reilly publish it.
Moderate it as flame bait or whatever you like, I just had to get that off my chest.
Aaron Newsome
Having admined a few Solaris boxen for 5 years, one thing I found irritating was the way an errant TCP/IP application--say, Netscape Enterprise Server--could get stuck in the middle of handling a request and end up unkillable. In order to release the port, the only remedy--I swear, ask Sun--is to reboot. Nothing you can do with kill, with proc tools, or by restarting netorking services, will kill a process in such a state, at least through Solaris 2.6.
I agree with most of what you've said. I believe a system administrators job is to maximize the long-term availability of the service, not just the service for a particular day (unless there is something unusual about the day, such as a project nearing completion).
If the first thing you do is reboot, then yes, you'll get the service back up and running in the shortest amount of time. But unless you understand why there was a problem in the first place, you haven't really fixed anything.
It's amusing to hear people talk about "transient" failures. With computers, there is almost always an explaination for why something went wrong. It's been my experience that with Windows, it's hard to find that explaination. With Unix systems, it's easier. With OSS Unix systems, it's even easier still.
If your first instinct is to reboot a system, you may be destroying valuable information on the cause of the problem. Such things like what state various daemons were in, which processes were running, etc. While the system is still running, it's possible to extract more debugging information out of the system (which should be analyzed later). For example, you can cause named to dump state information to a file.
So in almost all cases, even if you need to reboot the system, you owe it to yourself to at least gather as much information as possible first.
Out of a sample of 12 clients, too.
Yes, a BSD/win98 machine is here for those who *require* win$ for some reason. I do not care, the user is happy and that is always the point. Is it not. -d
from dissing Windows 2000, even though it only RTM'd a few weeks ago, and isn't even commercially available.
Honestly, I can't think of any group that spreads as much FUD as do Linux zealots at Slashdot. And yes, that includes all corporations and most governments.
Other than that, since my RC-2 box had been up for about 85 days until I moved up to RC-3, I'd say that the guy needs a new reporting tool.
Cheers,
ZicoKnows@hotmail.com
I want an honest answer why someone would use Win9x/NT as a file and print server. You have to pay for the windows license, and a samba setup is free/works better. Score points for saving money for the company? When I see file/print server being done by NT, I guffaw in laughter.
Lars -
Show me a general-purpose UNIX/PC box that's been up for 2 years and I'll show you a box full of unpatched security holes.
It's a shame there's not a convenient command which prints out the last time the box crashed. That's at least got a hope of supplying useful information, modulo power outages.
--Dave
These are hardly what I would call "Hard numbers". They are even worse than a microsoft sponsered comparision.
Don't get me wrong, I love linux, and I highly believe that yes, windows IS miserable for uptime, and yes FreeBSD does kick ass with uptime. (Linux has more frequent kernel release, and we haven't yet figured a way to upgrade your kernel w/o rebooting).
I'm surprised that an article phrased in such a way that its sounds as if its suppose to be serious would be posted on slashdot.
Thats enough ranting, I'll probably get moderated down as a troll. But seriously, tune out anyone who quotes these numbers as reliable.
I use to have a funny sig, but slash cut it off, and I forgot what the punchline was.
(someone needs to tell these people slashdot is an american website that happens to have attracted some non-americans. dipshit.)
That is evident through the widepspread pig-ignorance of the majority of contributors, dipshit.
Stop looking like a fuckwad, and stop posting. Or do what Real Men (tm) do -- SIGN THEIR POSTS.
-m
99 little bugs in the code,
99 bugs in the code,
fix one bug, compile it again...
Voting Moo Anyway!
I've got it running on a WinNT 4.0 server with SP6 that has 86 days, 18 hours, 9 minutes, 5 seconds of uptime. Admittedly it doesn't get used for much, just a private file server for me and my fellow support people, and running RC5DES, but NT can stay up for more than 49 days.
jeff_C
I would just like to know where VMS fits into all this. It's much more stable than any of the *nixes. At Amherst, we've got one system that's been running for 3 years without a reboot, and I'm sure that someone somewhere has one that's gone much longer.
Alas, no one seems to remember VMS these days (other than a Userfriendly strip last month). A great OS in its time, and still much more stable than any *nix.
I can second this. I have a system running 2.0.36, and it has *never* crashed, and does a heck of a lot more than any machines which I run the 2.2.x kernel on. I have had 2.2.x machines completeley lock up (no telnet, keyboard, mouse), while saving a file in Gimp! The 2.2.x machines are all SCSI based with AHA-2940UW cards, so I'm not sure if the aic7xxx driver is at fault, because my logs never have anything about these lockups. Maybe I'll take your recommendation and trim up the kernel, and remove any IDE stuff from it, since I do not run IDE. What else would cause a lock up like this? I've always had the lockups when saving to disk, so I'm assuming its a driver/linux issue, since on the 2.0.x kernels, same machine, this never happened. I have even replaced all my SCSI cables, to ones with terminators built in, and re-verified all the basic SCSI issues that would lead to crashes.
2.2.x and SCSI on this machine have a long way to go as far as stability is concerned. Although sometimes I do get 30+ days of uptime on that machine - sometimes 1 day. FYI: Its RH 6.0, with the kernel from RH 6.1.
Perhaps its the Slashdot effect, but I tried both sites and the uptimes.net site returned: Site down This page is temporarily out of order. Please try again later. I tried again, and all I got was no response. I wonder what operating system they're on?
I have a linux 2.0.36 machine which has a current uptime of 112 days, 112 days ago ( while the machine was on around 100-ish days ) it needed rebooting so its power source could be relocated.
More interestingly I have a whole horde of 2.2.x machines most of which are running the stock kernel that was installed with which ever distribution seemed good at the time, these seem to crash a fair bit, I occasionally find them completely locked up, not responding to the keyboard, pings etc.. most of them have good motherboards, a mix of scsi, ide disks, some SMP, some uniprocessor.
I have upgraded the kernel on a couple of the machines ( guinea pigs if you like ) to remove alot of the useless things and try to get some rock solid reliability back.
I absolutely love showing off good uptimes to clients on machines that do alot of work but I just havent managed to make it happen with the current stable kernel.
I am a lawyer and this constitutes legal advice and I shall indemnify you against any losses arising from taking it.
You might think, at first, that this measures the reliability of the OS. However there are some other factors here besides the OS, the main one being the competence of the administrator.
The clue here is that Solaris has a much worse uptime than the other Unixes. Yet we all know that Solaris is a damn fine product, and I've seen some Solaris boxen with amazing uptimes.
So why does it perform so poorly here?
I think the answer is that the average Solaris admin comes from an NT background and believes that reboots solve a problem. You get some of these people in the Linux stats too.
Now look at BSD. Who runs BSD? Old guard Unix people, who generally have their sh*t together, and know the hell what they're doing. These are the kind of people for whom uptime is a point of pride, who take it as a grave personal failing if they have to reboot to solve a software problem.
So while I don't doubt that BSD is a robust and stable OS, I think that to some extent the uptime stats reflect the average level of experience of the admins, and not just the robustness of their OS.
I would guess Solaris makes a much better showing if you can eliminate this effect. BSD would still presumably edge out Linux (since uptime is what BSD developers and users strive for, I think the OS provides it), but not, I think, by a 2:1 margin.
actualy, BeOS isn't cheaper. I've seen it range from 50 bucks (local software store) to 79.99 (best buys) suse is about 35 dollars. and Mandrake isn't that much more
Gentleman, you can't fight in here, this is the war room..
who sez death can't be funny....www.endlesssorrow.com
If you have two FreeBSD boxes, sitting in the closet since 1995, they have significant uptime even averaged between themselves. If you have 5,002 Linux boxes, two sitting in the closet since 1995 and 5,000 rebooted on odd chance, you have a heavily skewed bias.
Without basis on why the machines are up/down and factoring that into the averages, it's merely pretty pictures.
I have Linux boxes filtered and firewalled that have been up for years. Due to denial of service attacks to the vulnerable kernels they are running, I can't and won't post them. I will however say that two of these boxes were listed as #1 and #3 on the previous uptime site a couple of years ago. #1 had an excess of 500 days when the site disappeared.
Depending my boxes' job, it may be rebooted several times a day or it may be up for months at a time. I do a lot of code development and testing in/out of the kernel so I have a lot of boxes that get rebooted. I also have a lot of boxes that gather dusty electrons month by month. A few of the boxes I build kernels for crash. Dev kernels do that once in a while. By far however, the systems are completely stable.
All of my machines that lost large uptimes lost it 100% due to power loss.
Let's try and view these figures with an understanding that the Linux boxes outnumber all the others combined by a large number. I'm willing to bet that most of these Linux boxes are personal machines rather than black box setups.
-d
who would have thought anyone would ever use BeOS for a server. The comparison between Linux and FreeBSd is really interesting, considering how much publicity Linux has been getting lately as being "ultra stable". Statistically FBSD is more than twice as stable as Linux. Realistically this probably has more to do with the huge number of different Linux kernels used in the tests and the comparitively few different FBSD kernels used. It is interesting to think about though, are the half dozen different Linux distros more or less stable than others? I personally think that FBSD's development (Open and Net also) makes for a bit more stable of an OS.
I'm a loner Dottie, a Rebel.
"Server" is a computer people either log on to and does work ON THERE, or people use stuff ON that computer.
"Workstation" is the computer users can PHYSICALLY see and touch (and stroke, for whoever gets that fancy). They run programs there, locally.
For your information, people used server/workstation before Microsoft started making a server-oriented OS. Heck, even before they started thinking about networking. Who? Novell. Are THEY the evil empire now?
Grow a brain, for fucks sake. And a spine. Posting anonymously is (for lack of a non-punny way of saying it) cowardly.
-m
99 little bugs in the code,
99 bugs in the code,
fix one bug, compile it again...
Voting Moo Anyway!
You know, looking at those stats a thought struck me - despite BeOS managing a high of 35 days uptime, the average of four machines was just five days, implying there were a number of instant deaths. Pretty bad yeah?
Now compare this to the 49 day high of Win NT (40% improvement) achieved from a pool of seventy-two NT servers (1800% more machines) and BeOS is pretty much doing as well as NT. Especially when you consider the amount of NT trouble-shooting being done by both MS and the community in general - When was the last time you saw BeOS mentioned on Bug-traq? Nt does manage to improve on BeOS' average uptime though - twelve days... Must be why NT admins get uptight every fortnight... :)
In the end of the day though, Yay for Chuckie!!! Bu-Wa-Ha-Ha-Ha-Haaa!
"A goldfish was his muse, eternally amused"
Vs lbh pna ernq guvf, ybt bss abj. Tb bhgfvqr. Syl n xvgr.
I feel the need to bitch about some of the numbers is see, I'd like to know what job those systems are doing, BSD computers more often than not are a network server or something similar, you don't just turn that off. On the other hand, I turn my win NT machine at work off every evening (uptime roughly 8.5 hours?) or does uptime count the number of hours the system is running between crashes? In that case my BeOS machine at home must now be somewhere around 500 days or so. I upgraded it a few times but it never ever crashed on me.
Putting win2k in this statistic is of course ridiculous, the OS that has been out for 50 days has a maximum uptime of 49 days well..*duh*
To make a statistcally valid comparison of uptimes you'd have to use the same number of systems for each OS, not well over 500 for one and just under 20 for several others. In a larger population you are naturally going to see more extremes. I bet the record for shortest uptime can also be found in either the linux or freeBSD groups. The averages of course tell us something, but in the really small populations they too are irrelevant.
I'd like to see this uptime project become bigger amongst users of less uptime centered operating systems so that the statistics become a bit more valid.
The only thing this chart tells me is that *BSD and Linux users are more concerned with statistics like this than users of other operating systems.
What would psychologists make of that?
-m
99 little bugs in the code,
99 bugs in the code,
fix one bug, compile it again...
Voting Moo Anyway!
Besides the poll lookalike behaviour of this contest, as pointed out by others, I wonder.. why is Windows 9x also in this list? it's a desktop OS... not a server OS. And yes, IMHO these 2 are different. On a server you'll do everything to prevent the server from being offline, on the desktop all you want is to not loose work you've done in a certain program.
;)
hehe and it's also a good laugh to see these people think win2000 can have an uptime of a year or so... it's not even released!
Never underestimate the relief of true separation of Religion and State.
The shocking part is how little Novell servers manage to stay up. How many goverment institutions do I know that just bought those? That's pathetic. I knew that Netware was a pain to run, but I didn't know is was that tought to maintain.
You must not be doing it right and/or had no idea what to do with it. Netware is still one of the most stable and secure file/print services available out there - see topic "No Netware?" at the top of the list. There is, however, argument for the fact that Netware 4.0 was unstable. Novell was in their very early testing of NDS and did have some rather large problems adjusting it to the idea of a directory.
But, this time spent tweaking and fixing has paid off in spades. I believe intraNetware 4.11 and higher is absolutely rock solid. I wouldn't trust any large user base (50+) on NT boxes. Use NT for what it is good for, the Microsoft catch-22 knows as Back Office.
Leave the file serving to the real file server.
I was curious to see these uptimes on a chart, but when I went to the URL to see it, I was got a message saying the server was down. I wonder what OS it was running... ;)
Karma: Good. I'm hoping in the same way as pizza is 'good'...
This is odd... Scroll down the page, and look at the `Linux Kernel Usage' table. Why are FreeBSD, OpenBSD and Solaris listed in this table? And what impact does this have on the `maximum uptime' score for Linux?
--frank[at]unternet.org
If a production system is having a problem, your job as an admin is to bring it back to a working state in the quickest possible time.
The quickest thing to do frequently is just reboot the sucker. Are you instead going to tell your endusers that they system will be up in 8 hours because your debugging a problem? If so, do you mind being hung with a old rope, or should they go buy a new one?
If the reboot doesn't fix the problem, then you look at it from a different angle.
If it's a recurring problem... It's doing it every day, or every couple weeks or something. Then you might spend 5 minutes seeing if you can determine what's actually happening. But if you can't fix it in a couple minutes, and you know a reboot will solve it... then reboot.
In the meantime you'll try to reproduce the issue on your test environment.
Once you have identified a recurring problem, you either fix it(if you can)... or you may just throw together a script which stops and restarts the processes effected, as this is quicker than rebooting.
And sometimes those scripts get thrown together with a monitoring service which automatically runs the script if it senses the process has failed again. Or maybe a cron job at 4am everyday as "preventative" maintenance.
This is the life of a sysadmin. If you don't like it, then maybe you should stay out of the business, or buy stock in rope making companies.
And it doesn't matter whether your production systems are running Linux, Solaris, BSD, or NT... they all basically operate the same way, and you react the same way.
Uptime is irrelevant for home users because of a few simple facts:
Home users turn off their computers at night. Most of the Windows users I know aren't running mail servers or FTP servers that require constant uptime, so they power down at night to save some pennies on the juice bill.
Home users don't have uninterruptible power supplies. If the power goes out, the last thing they want to be doing is sitting in front of their computer. The $100 investment just doesn't make sense for them, and thus, they experience downtime with every power drop.
Home computers are used by children. Your spiffy FreeBSD machine is probably locked in a wiring closet somewhere, well away from six year olds with a penchant for DirectX games and dripping their Cokes on the keyboard.
Home computers are moved around. It might sound odd, but you're much more likely to shut down and pull the plug on a home system than a server just to move it over a few feet or to clean underneath it.
I'm not meaning to slam Windows as a home operating system, but isn't it fair to say that Windows (all flavors, even NT) has more home users than FreeBSD? Isn't it thereby, safe to assume, that if you really have an accurate survey of uptime, Windows will naturally be lower? Just something to keep in mind.
What's your damage, Heather?
I have no idea about FreeBSD but i am guessing that either it has less time between kernel updates, the version our leader is running is the last in the stable series and FreeBSD has moved on to a new series or it too is vulnerable to any bugs or attacks that have been fixed in newer kernels.
I am proud of the fact that my servers have an uptime of only 30 days or so. Because i know that I am performing regular maintenence on them. They crash rarely, usually due to hardware failure, but I reboot them frequently to make sure they are running all the latest fixes eg a new kernel install.
This is like saying "WooHoo my bog standard RedHat 5.0 box has been up for 2 years!!" Crackers ahoy! Vulnerable target sited!! A quick search of any crack DB will give you root access in less time then it will take you to make a cup of coffee.
I would expect any NT box should have a maximum uptime dating back to the release of Service pack 5. (Dont know about 6. A few admins i know are avoiding that like the plague)
The same applies to Linux and or FreeBSD or whatever. If you fail to apply critical patches to your system, most likely in production use, why in god's name should you get kudo's off the hacker/admin community for a job well done??
Kaptain Krash.
My other .sig is a 4000 line perl program.
Good night. You have provided much entertainment and enlightenment on my quest for knowledge on the little known Microsoft concepts of Servers and Workstations and the finer art of OS programming.
I bow to your intellect.
War crimes, torture, lies, illegal spying... Would someone give Bush a blowjob, already, so he can be impeached?
I agree with you. I wish there was more out there for BeOS. Yes, it isn't completely POSIX compliant. and Yes, telnet hasn't been ported. (To my knowledge. I looked for it a little while back.) It is silly to try and compare BeOS to NT server. The two OSes have different goals, and intended users. I am not going to install NT server in a home box that isn't connected to anything, and I wouldn't use BeOS as a mail server.
[john@darwin ~]$ uptime
7:10am up 141 days, 21:25, 1 user, load average: 0.02, 0.02, 0.00
141 days = 8,000 emails passed, 3,000 web hits, and around 250 SQL queries. Not too bad for an old celeron desktop machine....
-m
99 little bugs in the code,
99 bugs in the code,
fix one bug, compile it again...
Voting Moo Anyway!
-m
99 little bugs in the code,
99 bugs in the code,
fix one bug, compile it again...
Voting Moo Anyway!
www.uptime.net - no response, server could be down :-((. Win2K pre-release test site??
Uptime doesn't matter to home users, so they won't be interested in supplying statistics to the uptime site so they won't be included.
The ony people who will supply statistics to the site are those people who are interested in uptimes - administrators mainly and for systems that they expect to stay up for a significant period.
BTW, you need 30-40 samples to be significant.
Deleted
2*2*2*2*2*2*2*2*2*2*2*2*2*2*2*2*2*2*2*2*2*2*2*2*2* 2*2*2*2*2*2*2ms = 4 bil ms 4 bil / 1000(ms/s) = 4 mil 4 mil / 3600(s/h) / 24(h/d) = 49 days LONG GetTickCount(VOID)
The answer to the Universe and everything is "64" according to the Hitch-hikers guide to the Galaxy.
Shoot the Trollisms! Intelligence rules!
-m
99 little bugs in the code,
99 bugs in the code,
fix one bug, compile it again...
Voting Moo Anyway!
IMHO, drivers are the reason that the windows 2000 machines have such short uptimes, or more specifically, the fact that new drivers are still being released constantly.
I have been running RC2 since it was released, and my computer has never crashed. However, it has never been up for more than about 15 days at once, because I have wanted to try some new TNT drivers, or some new ATA-66 controller drivers, etc.
Once I am satisfied that I have all the right drivers (and there is no reason that I should be satisfied yet, since windows 2000 has not been released yet!) I will stop rebooting to change drivers -- and if my past experience is any indication, that means that my computer's uptime will keep on increasing at a rate of 24 hours/day.
Tarsnap: Online backups for the truly paranoid
www.uptimes.net appears to have been slashdotted.
at the time of writing, the home page reads:
>
I think was the limit on a Windows 95 machine. There is (was?) an internal counter that rolled over after 47.5 days causing the machine to lock.
May of been fixed by now.
While that's true, this kind of survey does give us maximum runtimes, and I don't think that's available anywhere but here.
For example, maybe a few posters could close their blathering pieholes long enough to see that the 49 -day figure applies to Windows _anything_, not just Win2k. For a startling revelation, go to The List and click "All" under "Alltime".
There are dozens of 49-day, 17h02m uptimes for Win32, and none longer. Obviously, either the OS or some popular [driver|service|screensaver] is broken [insert dumb "already knew that" joke here]. I dimly recall Microsoft claiming this was fixed in an NT service pack; obviously that's not the case.
For a more subtle trend, you can see a clump of Linux boxes topping out at 497 days, 02h27m. This is 2**32 Intel jiffies (100ths of a second; Alpha jiffies are 1024ths of a second) -- if you're running a module that assumes the jiffy count is always increasing, you'll get weird happenings when the counter rolls over. Again, I dimly recall one of the kernel people suggesting the jiffy counter be initialized (at boot) to MAX_JIFFIES - 3600, so that every module author writes code that will handle a rollover.
Faults tht only appear after a long runtime are typically easy to fix, but almost impossible to detect. Right now, the survey doesn't filter out shutdowns for known reasons, or collect enough info from the client (what modules are running, etc.).
If that changed, it could be a real goldmine, both to software maintainers, and to those who want to know when their system is due for its next crash.
cheers,
mike
To a certain extent this also measures the frequency of major kernel upgrades. While Unices can stay up with virtually any piece of software being installed or upgraded, everybody reboots for a new kernel. I don't think that installing a new kernel a couple times a year is necessarily a sign of weakness in an OS. There will be security patches, bug fixes and the occasional new feature.
The net will not be what we demand, but what we make it. Build it well.
Look for event 6009 in the event log, that's the NT ver number that gets written just after the eventlog service starts - a MUCH better indicator than all the uptime POS apps that clock out after 49 days (I have an NT server that, since SP5 was applied, has been running for 80 days).
Anyone who criticizes Windows 2000 for having a short uptime right now is clueless. How does a non-released, updated tri-weekly OS POSSIBLY maintain an uptime of more than a few weeks???
Watch after people deploy it next year - THAT will be the true test.
the site is down atm because tgm is trying to stop his box from crashing under the load of a hell of a lot of slashdotters...
Try again in a few hours.. and things should be cleared up again!
Or haven't you noticed?
--
My comments and opinions completely reflect those of anyone and anything I am remotely associated with.
Not knocking Linux - it's a fine OS, but did want to point out that I've had two NT 4.0 servers running an intranet site for more than six months w/o a single re-boot.
They aren't the exact same thing, and they actually share little code (Linux and FreeBSD) so why should they be the same? I bet HP-UX and Solaris don't both get exactly the same "longest uptime streak".
Ever thought of replying "No" when prompted with these reboot messages and wait until you have done all your changes before sinking the beast known as NT? Isnt pretty but still brighter than doing each task one at a time :)
I'm not an expert of FreeBSD, but 1994 days of uptime for a OS from october 1995? 1994 days are about 5,5 years, and 2.0 release is said to be from nov '95... For those of you that don't want to do the math, it seems to mean he have been running the release before it was released... Or anything else you want it to mean. I could be wrong.
Windows NT can stay up for more than 49 days, but since GetTickCount() only returns a DWORD... there's no way to check uptimes > 49.x days.
(GetTickCount() retrieves the number of milliseconds that have elapsed since Windows was started)
Does anyone know of a better function to return the uptime?
Hello all,
:-)
The new version of this uptime project can be found at http://www.uptimes.net and there you can see that Linux rules!
Groeten Raimond Kollman
My BSD router, server and workstation proudly stand up against my daughter in the open. It's the Win98 box I worry about her messing up. Only thing odd I've had to do is tape over the power switches. The BeerBarn
Who the hell marked this as insightful?!?
That is the difference between a U*x and a Windows. On a good unix you need to reboot only for the most crytical fixes. In other words you do not need to reboot unless it is: oh well, kernel change time again.
Unless you have thy one and only internet facing server in your company you quite often do not need to apply fixes for lots of the kernel stuff (you are already behind a firewall that can be kicked and rebooted any time). Than it comes to userland. Yes, there have been lots of fixes for userland since BSD 2.0 or Linux 2.0.18. If you know what are you doing you could have applied most off them without reboot and still run with linux 2.0.18 happily. Overall, on a Linux or BSD behind a proper firewall you should have been happy without a reboot for the last year or so...
In other words:
1. You are mistaking server uptime for a properly configured system on a properly designed network with the uptime of a device that faces the net directly. These should be different devices altogether. For example a linux based firewall booting via network into ramdisk costs under 300$ and usually boots from cold in less than 15 seconds. Skipping a device like that out of the spec for financiall or whatever other reason (if you are running a service of course) is sheer idiocy...
2. You may be right about the crack DB for a home machine (which is least likely to have such an uptime anyway). You are utterly wrong for a production system. A U*x admin that has managed to avoid reboot on a production system for so long most likely has not left any known or even supsected h0lez for R00ting...
Baker's Law: Misery no longer loves company. Nowadays it insists on it
http://www.sigsegv.cx/
I admin 4 *BSD machines and 2 NT Servers.
.oO0Oo.
The NT needs rebooting for just about any change you make to it - even changing the screen res.
Make a change to the registry - reboot.
It's just so poorly designed as a server.
I cast my mind back to the day when I upgraded a box with Exchange server on it. Each shutdown took 15-20 minutes and boot time was about 5-10. I was installing IIS on to it with a few modules. 7 reboots - 4 dull hours.
Stick with ANY *nix
There are places where the networks are not touching,and there are places where they are-Boeing's Lori Gunter
We don't need convincing!! We run both FreeBSD and Linux.
realkiwi
NT4 has the rather lame feature of asking you to reboot whenever you visit the networking ui even when you undo changes that you made, and even when you make changes that do not require rebooting. Hardly any of the networking changes actually require rebooting.
Does this qualify as "show-stopper"?
- A.P.
--
"One World, one Web, one Program" - Microsoft promotional ad
"Remember when the U.S. had a drug problem, and then we declared a War On Drugs, and now you can't buy drugs anymore?"
Because the fool who wrote the daemon uses a Win32 call that returns a DWORD?
When I got a (then, to me) brand new Ultra about
3 years ago, I installed Slowlaris 2.5.1. The only
time I had to reboot it was about 10 days ago,
when I installed Sun's recommended Y2K patches.
Since then, I never had to reboot. To me, nearly
1100 days is still pretty impressive.
Not that I give a damn about uptimes...
Shame on me, I thought about a NT machine, where a rogue program can take the whole system down
;-)))
An old version is not necessarally bad. I have some old versions that are on an intranet. Here is what I get:
bruser:~$ uname -a ; uptime ; date
Linux bruser 2.0.27 #7 Fri Oct 9 14:25:53 EDT 1998 i586
8:24am up 382 days, 18:31, 18 users, load average: 0.00, 0.00, 0.00
Thu Dec 30 08:24:29 EST 1999
I even have a 1.2.8 running (It's standalone)
You are talking total crap. BeOS is so low on statistics because *almost* noone is using beos as a server and noone leaves a beos machine on. BeOS is installed usually in a dual or triple boot PC. This means that everyone is rebooting in order to play games on windows or reboot back to linux/bsd. BeOS had NEVER crashed on me, but it never had more than 12 hours uptime too! I am always rebooting back to another OS sooner or after. If there was a reason to leave BeOS on,(by being a server or something), then BeOS would have incredible uptimes too. Eugenia.
http://www.distributed.net
I meant the distributed effort known as RC5, it was a play on words if you will. I thought your god's company encouraged thinking outside the box, something that you seem to lack skill in.
Lars -
>Think about this. Some person is amazingly proud of the fact they are running kernel 2.0.18.
:)
Correctomundo. An uptime of > 1000 days is roughly 3 years. Using such an old kernel makes your machine vulnarable to any script kiddie being able to blast your server away.
There are other things in this summary that pose me problems :
1. Comparing uptimes between different OSes is one thing, but you should mention also the different hardware of these systems. Comparing a i386 and a RS6K both running Linux will give you quite different results.
2. Besides, as the Linux kernel works with jiffies, on a 32 bit arch, your uptime is limited to 497 days, not ?
(A 64 bit system would give you an uptime of 5.84 billion years - now, that's the stuff you can run universes on
Kristof
"Nr. of hosts registered per OS"
This tells me that they aren't using a random sample, but rather you have to activly register your box to have your uptimes scored, which doesn't indicate that the sample of people is very diverse first of all. Also, the sample size for some of the OS's is horrible. A few examples:
Windows NT 71
Windows 95 30
BeOS 5
And people are actually making remarks about BeOS's performance when only 5 people have contributed their uptimes to this study? As far as I'm concerned, the only samples that are worth jack are Linux and FreeBSD (590 and 137, respectivly), and even then I don't trust the results because of the first point I brought up.
Man, sometimes you just gotta look at the numbers.
Just figured I'd mention it..
--
Never hit your grandmother with a shovel, for it leaves a bad impression on her mind...
Sure, but if an OS has a commonly-known flaw, good software will know how to get around the flaw.
- In Capitalist America, law violates YOU!
> 2. Besides, as the Linux kernel works with jiffies, on a 32 bit arch, your uptime is limited to 497 days, not?
Actually your box won't crash when your jiffies wrap - the kernel will mostly keep on trucking. But 'strange and mysterious things' will begin to happen in user land. Most obviously 'ps' and its friends will get stuck at the momment of wrap. Hmm... sometime ago, when the idea of setting your HZ=1000 was trashing about the kernel dev list some posted a patch for all those user land things. That was diffinitly pre 2.0.0 though so who knows what state its in today.
... with eskimo chains i tatto my brain all the way...
From their site :
"Everything is still in very beta-alpha-test-mode, so don't expect it to work."
Who can be stupid enough to run such code on a Production server, where UPTIME is an issue ???
As *BSD's tend to be servers, and Linux tends to have a much higher number of workstation and newbie use, plus the fact that Linux boxes probably get tinkered with a lot more than *BSD boxes. The average uptime for a give OS/kernel version is completely irrelevant if they should be compared with reliability in mind.
What is important is the maximum uptime for a given OS version. Notice that the maximum uptimes for the various Linux kernels are close to the actual age of those kernels! What does that tell you?
The fact that Microsoft could'nt break 7 weeks reflects what I have witnessed in every contract I have had.
War crimes, torture, lies, illegal spying... Would someone give Bush a blowjob, already, so he can be impeached?
So how does the uptime client work? Does it actually send the complete uptime data back to the main server?
/proc/uptime file. Now how easy it would be to modify the source to point to your own 'hand-crafted' uptime file?
There could be a potential hole for cheating, where a hacked client could be used to send fake uptime data back to the server. (This is just like the problem John Carmack was expressing not long ago with the Open Source Quake)...
Having a closer look, I went to the new 'uptimes' site, http://www.uptimes.net/ and it had the source for the client available for download.
Just breifly looking at the client source, I noticed that it fetches the uptime from the good old
if You sort the complete list by OS, several entries of NT with 49 days uptime show up, indicating that it's so unusual to reach those 49 days and more.
remember also that NT has to be rebootet more often because of silly things, like changing the default gateway. on the other hand those intentional reboots might lower the probability of a crash.
Interesting to note that FreeBSD is sitting in 5 of the top 10 spots, including the top spot.
Are registering exactly the same record uptime. I know for a fact that a Windows machine can stay up longer than the maximum uptime listed; I've seen it with my own eyes. I wonder if the software that's used has a bug in that it forgets to take into account the fact that the Windows uptime clock wraps at 49 days.
- In Capitalist America, law violates YOU!
OK, you always see the server uptimes, but for people who actually do day to day work on a machine's console, how much uptime do you have? My average uptime on my Linux box at home is probably 3 days, because I am always fiddling with the hardware or software. I've only had a handful of crashes, though. One was because of a JAZ drive that I have never gotten to work with a cheap SCSI card I bought, and the others were because of something I broke in X, i.e., I fiddled with the hostname and X would lock up the system, or I changed some configuration file and hosed things up. I sometimes wonder how much uptime I would have if I just left well enough alone. But that wouldn't be any fun, would it?
Logic ... merely enables one to be wrong with authority. -- Doctor Who
How about telling us what all those machines are used for. Anyone can start up a server and sit it in a closet for a year. If nobody uses the server it will probably stay up until the power is interrupted. Now, get a big box with a lot of DB activity & users pounding away at it and let's see how strong it is. It's also worth noting that some people believe that theraputic reboots can help prevent problems before they occur. I'd rather take a planned outage window every Saturday night than crash one Monday morning.
Uptime means nothing. It's what you do with the box that counts.
You see what stuff the top 10 is running? Keeping those uptimes mean that you never upgrade your kernel etc... that is often suggested for security and other things. The linux one in the top 10 runs a 2 and a half year old kernel. FreeBSD has slower releases, and would, in general, have a better uptime for it. I am impressed with the 61 day Windows 98 max uptime though. Strangely enough, Windows 98 has the best max uptime among all the Microsoft OSes on there
All this serious talk about uptimes..
:)*
sure you can learn some things if you look at the list.. mostly that nt and win are crap os'es when it comes to obligatory reboots etc...
Hasn't anyone thought about it that it can be fun
to have the highest uptime? I mean.. it's just like trying to get the character with the highest nr of hitpoints in your local mud, or playing text only versions of planets... totally useless on it's own.. but fun nevertheless...
Also on a note about the stats not being really usefull... www.uptimes.net has only recently been activated... so it's normal that there are not much beOs users etcetcetc... the old page had a lot more users.. and i guess most will migrate to www.uptimes.net soon.. But... if you're not satisfied.. just dl the client and add your beOs box to the list!
I personally know the guy from uptimes...
*We even had a "few" beers together last nite
And he himself sees it as it is.. a small bit of fun.. and a nice bunch of statistics to show of his php skills.
So.. try to look at things his way.. don't start os wars over the page and the stats.. but register and contribute to the overal stats..
BeOS was never intended to be a server OS. Why don't they be fair and see how well it works compared to the others as an internet appliance OS or a workstation OS.
I think the stats on Windows 2000 and BeOS should be looked at very carefully. These two OSes will go down not necessarily due to instability but due to sysadmins playing with them.
With Windows 2000, the system people have isn't finished, and as patches have come out, well, they get applied, and as they get applied, well, the OS must restart (it IS Windows).
Now, for Be, there are very few people I know who would run it as a standard server. Plus, on top of that, there are lots of new programs coming out. I don't think they require a reboot, but, well, we'll see. At any rate, BeOS isn't really designed to be a server (yet).
The shocking part is how little Novell servers manage to stay up. How many goverment institutions do I know that just bought those? That's pathetic. I knew that Netware was a pain to run, but I didn't know is was that tought to maintain.
The power of accurate observation is commonly called cynicism by those who have not got it. - G.B. Shaw
Sounds like Microsoft never thought that anybody would ever have a machine that would ever be up and runing for more than 49 days. That's an OS bug.
Deleted
Ok, not to be "pro-microsoft" or anythign here, but comparing the uptime of windows 2000 with others really doesnt make much sense now does it? Windows 2000 has been available for how long? certainly not 1994 days :P
If you look at their "Linux Kernel" stats, some of their Linux boxes are running BSD and Solaris kernels? Is this Debian GNU/BSD or someone sending in deliberately broken info?
Windows 98 2 days, 21:15m 61 days, 13:52m
...sigh...
2 days? I have had more uptime on my coffee pot! : ) 2 days is the AVERAGE, that means a decent amount of the machines had to have lower uptimes then 2 days to bring the average down.
If these Win98 users downloaded the uptime program, that means that they are concerned about their uptimes. Most users shut there PC off when they are done with it, but if you download a program to keep track of your uptime, that means you are TRYING to get a decent uptime on your machine, this is just sad.
So are my tax dollars still paying for goverment offices to use Windows in their workplace?
"`Ford, you're turning into a penguin. Stop it.'" -THHGTTG
......for those of us on dual cpu machines. I have two roomates who have said that SP5 makes it impossible to install ANY software. I run on SP4. The funny thing with service packs is that it's a game of chance whether it'll screw up your box or not. So the best advice is to only upgrade if you have some pressing need.
OS -- avg -- max
Windows NT -- 12 days, 17:46m -- 49 days, 17:02mWindows 95 -- 6 days, 15:23m -- 49 days, 17:02m
Windows 2000 -- 5 days, 10:38m -- 49 days, 17:02m
Aside from Win98, which strangely managed to have an uptime of 61 days, 13:52m (I'd really like to know how, the best I could get out of it was 4 days on at least 6 different systems), They all have the exact same max uptime. Now, I remember something like this being mentioned awhile back and microsoft acknowledging the mistake, but then why is Windows 2000 in the same boat as 95 and NT?
I'm not a big Microsoft fan, but I have to argee with you...I have had NT servers stay up for 6+ months without reboot.
This link isn't to Microsoft, so no need to register...
It's the Win98 box I worry about her messing up. Only thing odd I've had to do is tape over the power switches.
Absolutely true... I've have to tape a peice of cardboard over the front of my computers to protect not only the power switch, but floppy and CD drive. I run Linux on my system, and Win 98 on my wife's system. I am constantly having to fix the 98 system mainly due to my 3 year old reconfiguring it. I have yet to have either child come close to harming the Linux machine. They have their own accounts, and I don't have to worry about them messing with things I don't want them to mess with in Linux. I expect to have to reinstall Win98 within a few weeks if the pattern holds up.
When will Windows be ready for the desktop?
This discussion will be null and void in about 32 hours because the world is going to lose power and blow up.
Or so the media wants us to think.
So the real statistic should be:
How many FreeBSD machines are connected to an EPS (eternal power supply) and can update uptimes.net via floppy disk, since the internet won't work either.
ok
Didn't anyone notice that the Windows NT, 2000, and 95 maximum uptimes were all EXACTLY the same?
Seems pretty fskd up to me.
Also, what the hell is "multiple"? I assume its dual boot, but that shouldnt be an option in the client.
Then there's the Linux kernel records. Since when is "Solaris" a Linux kernel.
On top of that, there just isn't enough information to be scientific. 7 BeOS Machines is not a big enough number to calculate an average uptime for that operating system.
Something is obviously screwed up and i think it would do our community a great deal of good if someone would either write a similar application (that works), or give some help to this project.
indierock / punkrock band photos and more... http://www.digitaldefection.net
Is this a joke? Using "Wrong.", "NEWS FLASH:" and "Here's the secret:". Are you following the step by step jack ass HOW-TO. You claim you've written 2 operating systems to support your argument.... and those would be? And you're referring to him as "kid" but writing more like a child.
His point is that a server provides services and a workstation does things like CAD work. You could turn off a workstation at night and no one would care. The same could not be done with a server.
Thats great if you disagree with this having a significance, but grow up.
---- sonoffreak
Hold on a second....I never said that SP6 caused problems, I said 5 caused problems on DUAL cpu machines. I have no idea about sp6, i just said that needlessly upgrading to another sp is not a good idea.
This time, it seems to be a bug in the uptime client for Windows, not Windows itself:
:->
I've got an NT Server that's been up for ages; Performance Monitor shows 21171900 seconds uptime, which translates to about 245 Days; this should be about right.
The Windows Uptime Client however only shows 46 Days!!
Same issue with Version 3 and 4 of the Windows Clients running on NT 4.0; Uptime of a 2nd system running for just 16 Days is displayed OK.
If this was a Microsoft program for display of Linux uptime, I'm sure everyone would have a perfect explanation for the failure
you have moved your mouse, please reboot to make this change take effect
This is similar to the reason that none of the Windows machines (baring Win98) report uptimes in excess of 49 days. Their numbering syscall stops at that point (and caused the kernel to crash in Win95). Apparently there is annother facility to get this information but the Windows uptimed is not using it.
-- Remember: Wherever you go, there you are!
I just love these AC's that start a reply with a "bzzzt" or a "wrong!".
My point is that a server is much less likely to be periodically rebooted than a workstation is.
You're gunna argue with that because of what YOU do? I'm speaking for the average situation.
It's not stupid. You think I believe that a server is any computer that is not switched off? Maybe you should re-read what I wrote.
War crimes, torture, lies, illegal spying... Would someone give Bush a blowjob, already, so he can be impeached?