Follow Slashdot blog updates by subscribing to our blog RSS feed

 



Forgot your password?
typodupeerror
Check out the new SourceForge HTML5 internet speed test! No Flash necessary and runs on all devices. ×
Slashback

Slashback: Plexion, Kernelism, Salaryness 87

The list grows of how many OSes Plex86 can boot. Soon you may have an easier time of installing a new kernel (besides turning to page 207 of Running Linux). SAGE wants to know the intimate, personal, steamy details of how much you earn as a SysAdmin. Also, not everyone trusts the proposed data-escrow deal involving Celera and Science. All below, in this episode of Slashback.

Plex those muscles, yeah, and one and two and three ... dhunley expressed himself thusly: "A handful of days after getting Dos, Win95, and Linux to boot, Plex86 does it again! Initial support for booting QNX now works! Screenshot is here!" And shortly thereafter, the very same dhunley wrote: "Plex86 boots NT 4.0! Well, at least to the login dialog box. This is an old bochs disk image file. The mods will be uploaded to CVS soon."

Thanks, d. Watch out soon for an interview with bochs and plex86 lead Kevin Lawton, who promises to take time out from his busy schedule to tell you what's up in the worlds of emulation and virtualization.

While you're here, why not take a tour of the whinery? Apropros the wacky things that people have decided to put in the kernel lately (like GNOME ORBit), Booker writes: "Every time there's a major new kernel release, there's whining on Slashdot about killing uptime, and how much it sucks to have to reboot to get a new kernel. Well, whine no more.

Erik Hendriks at scyld.com brings us Two Kernel Monte, a 'kernel module which allows Linux to load another kernel image into RAM and restart the machine from that kernel.' The only major limitation appears to be that it will not work with SMP machines. Apparently Erik got the idea from the Linux Bios Project."

Now if some smart distribution maker (anyone, anyone) were to integrate that into their updates system, would that be so wrong?

We join this survey already in progress: Marketing Manager writes: "Anyone who's been in the job market knows how tough it is to find information about salaries and compensation. For system administrators, the search is compounded by the varieties of tasks, talents and responsibilities required to get the job done. Now there's something you can do to fix this.

SAGE invites you to participate in the 2000 System Administrator Salary Survey starting December 1, 2000. This annual survey is part of SAGE's ongoing effort to advance Systems Administration as a profession through information and advocacy. By participating in this survey, you join thousands of system administrators in examining the market and defining compensation according to your talents, your location, and your technical prowess."

So now you can find out where you stand in comparison to others with similar positions, perhaps a valuable bargaining chip come contract renewal (or incentive to check out a new city). The results will be available early next year -- automatically sent to SAGE members, available by request to everyone else.

Number One, can't you do anything about that rabble? bluets writes: "Some leading scientists and open-source advocates are attempting to raise awareness that the Celera/Science Magazine deal is a 'Big Mistake.' More details [here]."

And if you didn't know about this, it's only because you're not opening the barrel-of-monkeys Slashdot Science Section often enough. We're considering an experiment where everyone who opens it gets a food pellet, and everyone who doesn't ... well, do we have to bring out the Punishment Stick?

This discussion has been archived. No new comments can be posted.

Slashback: Plexion

Comments Filter:
  • Can someone please write a module to handle all those unpopped kernels... I really hate biting down on them - though my dentist wouldn't mind more uptime...

  • by Anonymous Coward
    The time that a W2K box is hibernated is still counted towards the uptime of the computer.

    ie. Use computer for 8 hrs, hibernate for 12 hrs, use for 6 hrs - uptime reads: 1 day 2 hrs.

    Bug, or 3v1l m1cr0$0f7 k0n5p1r4cy?
  • I had understood (well, I haven't read the dox yet, so my ass is talking here) that the state of processes was saved and hot-swapped into the newly started kernel.

    I thought that sounded reasonably cool.

    If it is just a way to avoid bois restart, that is less cool, but definately beats running lilo to remotely restart with a new kernel.
  • Just installed it for 2.4.0-test8
    You are right, it copied itself as "misc" under /lib/modules/2.4.0...
    I just renamed it monte.o, created the directory misc, mv it there, did a depmod and loaded it with a modprobe without any problem.

    --
  • Anyone who believes that reboots are somehow "necessary" in the course of normal system operation have probably never written a line of code in their life. IMO, excessive rebooting can mean one of two things:
    1. The system is not optimally designed. For example, there is no good reason why a system should have to reboot to load a library or run a new program. Except for the kernel, of course.
    2. The system is not properly designed. In other words, the code is technically incorrect and contains serious bugs (hence the famous BSOD).

    Using Windows you will find a wide variety of both types of error. Using Linux you will find hardly any problems of type #1, but occasionally some of type #2. Not very often, however. Using *BSD you will find even less of #2.

    So overall, I feel that rebooting is a serious issue indeed. It can be a good indicator of the general "correctness" of system design. I am no systems programmer, but this is my experience in running the different operating systems.

  • Hmmm...let's see...
    # kill 23 74
    # //2/bin/disktrap
    # //2/bin/nettrap
    :-)
  • by antdude ( 79039 ) on Monday December 11, 2000 @03:38PM (#566473) Homepage Journal
    Salary.com [salary.com]
    I am not sure how accurate they are:

    JobSmart's Salary Surveys [jobsmart.org]

    Pencom Systems Incorporated: Interactive Salary Guide [pencomsi.com]

    SalaryZone [zdnet.com]

    http://www.realrates.com [realrates.com]

    Are there any others and how accurate are they? Thanks. :)

  • Exactly. Why pay $300 for software that emulates another machine when for $400 I could buy a Windows machine, put it under my desk and VNC to the thing? Two machines are almost always better than one.

    Of course, with the added competition of Win4Lin and now Plex86 there are going to be a whole lot of geeks that re-think their future VMWare purchases. So VMWare will do poorly, and they will probably go bankrupt, and then they will blame the Free Software community for making it impossible to make a living selling Linux software. There will be a whole new round of articles entitled "Linux Software Companies Deemed Severely Unprofitable."

    Grumble, Grumble.

  • I, and I'm sure many others, submitted this letter to /. as soon as it plunked into my inbox from the bioxml mailing list. I wasn't concerned about the karma aspects of getting a story posted to slashdot--just that the story GET POSTED in a public way. I thought to myself:

    "Slashdot *has* to accept this story from someone. It represents the *informed* opinions of a *prominent* bioinformatics researcher, as well as that of a large portion of the bioinformatics research community."

    Of course, it was rejected.

    I emailed Ewan, and told him to submit the letter as a story. He did. I figured:

    "Maybe /. doesn't want to post secondhand messages (though they've done so many times in the past). But if the *author* of the letter submits, surely /. will post the story, because of it's *obvious* importance."

    Again, no story.

    Instead, what do we get? A blurb at the bottom of a slashback article, hidden amongst various pieces of random Linux cruft.

    This is getting ridiculous.

    Rob Malda et al., you guys have been hiding behind the "this is our website, and we'll put up things that we find interesting" line for too long now. Especially when slashdot *routinely* posts factually incorrect and uninformative stories based on heresay and speculation.

    Yet the opinions of a Howard Hughes Invesitgator in bioinformatics get ignored. Dare I ask why?? Perhaps it's too informative?? Maybe just a little too factual??

    Slashdot, when you sold the site, and became a commercial enterprise, you ceased to be Rob's homepage. Your public perception as journalism leaped. Unfortunately, your journalistic standards have not.

    The saddest thing about all this is that posting the letter on the front page of slashdot could have made a real social impact on the editors of Science. A lot of biological researchers read slashdot. Posting this letter as a story could have provided a nice focal point for community organization. As a second-line posting to slashback, I doubt that will happen.

  • Actually it does do some emulation and more is on the way from what I've heard. As of now it emulates the BIOS calls, which could be called virtualization. However, they are going to have the ability to "emulate" hardware eventually. So you can tell you OS you have hardware devices that you don't have.

    Can you imagine how great it would be for testing device drivers if you could just do it inside a plex86 instance of your OS, cause then you could change your hardware config to whatever you wanted and see how it performs.

    Cool stuff.

    Also, as for the performance issue and how useless it makes the product. That is a very nearsighted point of view. What do you think that 10ghz machine is gonna use all it's power for? Abiword and Gnumeric? :-). Hell no. You'll be able to run three or for plex86 session at FULL speed. I can't wait.

    Justin Dubs
  • Sun Solaris already does this and Solaris is slower than Linux... >Additionally, this project is a perfect example >of why businesses don't trust Linux. Not only >does this tweak the statistics to make Linux >look more reliable than it is, but it does >something much worse in the eyes of business.

    What exactly does it do thats bad in the eye of business? The worst thing I know very much about (unfortuantly!) is that there are plenty of hacker targets out there purely down to the fact that there servers are felt so mission critical tehy will not take them down for a Kernel upgrade.

    But the point about uptime being tweaked may show why Solaris seemingly has great uptime :-).

    I take downtime to really mean server crashes, rather than upgrades or improvements.

    You raised some good points, but I wanted to add more. One final note lets not forget where Linux is already succeeding...In the Server Space.

  • It's "News for Nerds", allegedly, not News for Bioinformatics Investigators. Since the article was posted elsewhere, it can be linked to. Slashdot has no responsibility to post the letter on the front page, or anywhere else.

    Who the fuck are you to throw a tantrum, just because it fits your definition of "Stuff that Matters" but not everyone (anyone?) elses? If you don't like it, take your money elsewhere. Oh yes, you don't spend any money here.

    I don't know a Howard Hughes investigator from fucken nobody, and after reading the article, my interest has not been piqued. I can assume that if the letter itself was posted on the front page in 48 pt. Impact with a <BLINK> tag, it would not make me any more interested. It's not like somebody is inventing carnivorous vegetables or a chicken that cooks itself, it's a mainly semantic dispute between one or more bodies in a field that doesn't excite me, or about 99% of the /. populace.

    So before you start slinging more daggers about Malda and Homos not having journalistic integrity, just shut the fuck up and stop crying; your mascara is running. You got your mention, which is a lot more than most stories get. Be happy with that, or I'll give you something to cry about.

    --

  • How is this comment "insightful?"

    To begin, when the new kernel is loaded, the uptime goes back down to zero. Secondly, this software was written by Erik Hendriks, who works for Scyld Corporation, not Redhat.

    Go play on your Windows box. Yes, I know you're running Windows, because if you were running Linux, you would have downloaded monte, installed it, run it, and seen that your uptime got reset.

  • Of course, anyone can add this themselves just by hacking the kernel. You wouldn't even need to go to much effort, since you could just make the uptime start at whatever value you want, rather than 0.

    The better solution is to ignore the uptime-dick contests and get on with your life, rather than lowering the kernel's quality as a piece of software.

    Hell, a moderately useful feature would be to make the uptime start at what it was last time the system was up, keeping a cumulative uptime as well as a standard one. If you also store when the cumulative uptime started, then you could store uptime as a percentage of real time. This would put some meat behind the 99.9% uptime claims. You could also have dick-size contests about how much your machine is up, rather than the longest particular consecutive stretch. Kernel upgrading reboots are then less of a problem, because you can still say things like "My machine was only down for 60 seconds last month!".
  • I use linux for a good time and I know that if you start taking your system down to install a new kernel every time they do a new one you will end up with 0 uptime. The downtime is the time the machine stop doing what it is supposed to be
  • I am waiting for kernel 2.4 to come out (and then even wait for SuSE to release there one). Should not be long now. The 2.2* line is basically a dead line and I imagien that 2.2.18 will be its last ever release.
  • by Anonymous Coward
    this project is a perfect example of why businesses don't trust Linux. Not only does this tweak the statistics to make Linux look more reliable than it is, but it does something much worse in the eyes of business.

    I think you are quite correct with this. I work at a Fortune 500 company, in the IT procurement division. About Six Months ago, we were considering upgrading out 500 Server machines from NT. We considered Red Hat Linux as a possiblity.

    Anyway, we contacted Red Hat and they sent round some of their executives to convince us. What we wanted from them was an assurance that the information we had about Linux was trustworthy, and furthermore that there would not be a code fork in the future, and that we would be able to 'morph' Linux to our needs.

    Amazingly, the executives they sent were unable to satisfy our demands. Although they told us that Alan Cox works for Red Hat, and thus that Red Hat is largely responsible for Kernel development, they were unable to provide the special type of install we required.

    When we gave our bosses a summary of their position, they saw it as a no-brainer. We ended up installing Windows 2000.

    I am an avid Linux supporter, and I would like to see all statistics regarding Linux (corps are mad on stats) verified by an outside body. The problem just now is that the stats are done by those in the Linux Industry, and are therfore skewed and biased.

    Just my halfpenny.

  • Care to spell out how the Open Source model applies to what is essentially data? That would be like a graphic artist working on the Open Source Model...
  • Preposterous! Pay to view a survey they collected info on for free, I don't fucking think so!

    Given this, I wish to be paid for my info....

    RB
  • by Seumas ( 6865 ) on Monday December 11, 2000 @02:29PM (#566486)
    I'm really surprised at the lack of media attention (that is, NONE) given to the purchase of the rights to index full genetic information for an entire population -- liberties granted by their own government for a fee.

    Outside of specific groups like Slashdot and a few scientific journals, I have heard ZERO conversation on this in meat-space or meat-media.

    Come on people -- real life makes The 6th Day look like a fucking walk through Disneyland.
    ---
    seumas.com

  • by Frac ( 27516 ) on Monday December 11, 2000 @02:31PM (#566487)
    Somebody finally explained the word warez to me! I always thought they were some kind of tupperware!
  • I am a sysadmin for Sun. I only get $65k. You must not work directly for Sun, hrm? Either that or you're pretty senior or something.
  • VMware will be discontinuing hobbyist pricing on December 16th, meaning that if you want a copy, but don't qualify for the academic price, you're going to pay $299 instead of $99. I like VMware, but if I had to pay full price for it I wouldn't have done it.

    Plex86 will start to be a more attractive option after that day. If it also means more development activity behind it, that's all for the better.
  • i still dont see the point of virtual machines...

    sure they're usefull for emulating a completely different system (as long as it doesnt need as many system resources as the machine you're running it on) but whats the point of emulating x86 on x86...

    the emulated system runs a lot slower than it would natively, as does the system doing the emulation, plus, you're not saving a whole lot of time... i mean, when you start the new system, the virtual bios has to start, then whatever os you're running has to boot before you can use it. so the only time you're saving is shutting down the currently running system...

    plus, there's the performance issue. you cant just leave the thing running, and expect decent performance, so the only thing its really good for is compiling and testing on another os, but since it runs so much slower, the compile takes that much longer, probably longer than the time you would save not having to reboot. you also dont have the ability to test the program properly, because you dont know how fast the program will really run, nor can you properly determine the bare system your program requires...

    so basically, the only thing its good for is being a neat toy for lazy bums with ultra-fast hardware...

  • What's wrong with you? He answered a legitimate question some would have when deciding whether to fill out this survey. Though he contradicted himself, it does not mean there was no information. If his original post wasn't modded up, I wouldn't have read the next comments -- moderation is a tool to find information, but once I acquire the information, I look through the replies for clarification/contradiction.
  • From the w3 page:
    This version supports Linux 2.2.x and Linux 2.3.x.


    {|}---Tony Hagale -- tony@hagale.net -- http://tony.hagale.net
  • Moderators -- crack ain't enough eh? Ya gotta speedball crystal meth and heroin too? You never should have touched that post. If anything you should have given him +1 Funny.

  • Correction notwithstanding, you wouldn't have to pay to view your own data, you already know what your salary is!
  • > If it is just a way to avoid bois restart, that is less cool, but definately beats running lilo to remotely restart with a new kernel.

    Even if it doesn't doe Everything You Could Possibly Hope For (tm), it might be a starting point for evolution in that direction.

    --
  • yeah, but does it reset the uptime in top? ;-)

    The cynic in me says the 'whiners' care more about the stat than any actual interruption in the machine's service.

    Personally, a reboot isn't that big of a deal for something like a kernel upgrade. Beats the hell out of having to reboot for an IP change or the like in versions of NT prior to 5/2000.
  • > Somebody finally explained the word warez to me! I always thought they were some kind of tupperware!

    I always thought it was a kiddie's way of spelling Juarez.

    --
  • Oh yeah, WIN 2k, lets see

    Add RAS : Reboot

    change route in RAS through NAT : Reboot

    CHANGE AN IP ADDRESS (like the gateway) WHEN YOU HAVE 2 NICS : Reboot

    I enjoy win2k, don't get me wrong, for a lot of situations it is an acceptable and worthy solution. BUT PLEASE Stop spreading FUD. It still requires a LOT of reboots. IT IS however very much improved over NT/98.

  • You have a problem with being a member of usenix/sage? Perhaps you might consider that salary survey results and such are a benefit of membership in usenix/sage.

    It *does* cost money. Their surveys can be taken seriously; they aren't half-cocked surveys by some website. They want real results.

    And if you don't like the arrangement, by all means, don't submit any data.

  • There was a story [abc.net.au] about the sale of the Tonga genome and follow-up interview [abc.net.au] with Joe Gutnick (chairman of the purchasing company, Autogen) on ABC's Lateline last month.

    Isn't public broadcasting wonderful?

  • I've looked at the config screen, and no, it does not seem possible to get all headlines (or maybe i should get them, but are missing a random subset for some reason).

    Yep, BTW, it is from Paranoia.
  • Anyway, we contacted Red Hat and they sent round some of their executives to convince us. What we wanted from them was an assurance that the information we had about Linux was trustworthy, and furthermore that there would not be a code fork in the future, and that we would be able to 'morph' Linux to our needs.

    Amazingly, the executives they sent were unable to satisfy our demands. Although they told us that Alan Cox works for Red Hat, and thus that Red Hat is largely responsible for Kernel development, they were unable to provide the special type of install we required.

    When we gave our bosses a summary of their position, they saw it as a no-brainer. We ended up installing Windows 2000.

    So, because the RedHat execs can't predict the future, your bosses went with Windows?

    Was Microsoft able to provide the "special type of install we required" or did this requirement vanish suddenly?

    Why was the posibility of a code fork such a worry?

    Will Microsoft allow you to morph windows to your needs?

    Why the double standards?

    I realise I am asking the wrong person, but your company management aren't available.

  • The human genome is the property of the human race. Anyone attempting to 'copyright' such information is stealing from all of his fellow humans.

    So all I can say is ....

    "Stop 'copyright' theft".

  • VMWare is too expensive. Plex86 is looking VERY nice. Once it can boot Win2k, I'll have replaced VMWare. Keep up the good work.
  • Just a note, you can contribute, but you must be a paying member to view the salary surveys.

    Frankly, I think the idea of a UNIX sysadmin salary survey is great (being one myself), but I don't feel paying for the data that I submitted is fair.

    Perhaps if you submit data, you can get access to the results free. You listening USENIX?
  • > This modification shows the business community that Redhat is only developing more toys for their toy operating system

    Yep. They should be busy integrating the GDI in the kernel, like any mature OS does.

    Come on, tI am sure that there are better ways to troll this story.

    Cheers,

    --fred
  • > Not to mention of course that if you are not using an obsolete version of Windows, changing IP addresses and adding protocols do not require a reboot

    What is your definition of obsolete ?

    Got a HP PIII 600 delivered here 1 month ago (to configure for a bank). The beast came with 2 preinstalled OS (that you could choose between at first boot).

    The choices were:

    1/ Windows 95 (First edition)
    2/ Windows 3.11 for worksgroup

    Talk about obsolescence. I am still rotfl. About no-one is using your latest w2k OS in the real world.

    Cheers,

    --fred
  • > -- I use Macs for work, Linux for education, and Windows for cardplaying.

    Then, in your user bio you claim that you use w98 at work.

    Mmmm.

    Cheers,

    --fred

  • by Anonymous Coward on Monday December 11, 2000 @02:07PM (#566509)
    This project to change your kernel without losing your uptime raises some very interesting points.

    How is shutting your machine down to a minimal state and bringing it back up NOT downtime? What exactly is the point of uptime if it's not that?

    That's kind of like cheating, if you ask me. Why not just create a kernel module that will log your uptime when you shutdown and start it at that point when you boot up?

    Additionally, this project is a perfect example of why businesses don't trust Linux. Not only does this tweak the statistics to make Linux look more reliable than it is, but it does something much worse in the eyes of business.

    This modification shows the business community that Redhat is only developing more toys for their toy operating system, instead of making important advancements that are necessary to let the operating system finally break through on the desktop. These kernel developers need to get in line if they want users.

    -lb

  • by imac.usr ( 58845 ) on Monday December 11, 2000 @02:36PM (#566510) Homepage
    from the survey:

    62.Does your organization provide any special benefits or working conditions that you particularly like? (please specify)


    Sure; for our company Christmas gift we got a bag of microwave popcorn and a hockey puck with our company logo on it, since apparently our big marketing coup this year was to sponsor a "play of the game" segment for Washington Capitals games. Bonus? What's that?

    Looking for a buzzword-heavy sinking ship? come work with me! [sytel.com]

    Assholes. I can't wait until my Oracle DBA class is finished and I can get the fuck out of technical support.


    --
  • by Anonymous Coward
    Celera should not be trusted. They are attempting to bypass the normal publishing standards for Genetic Data. As a gene carrying human you should be concerned about this. We cannot allow companies to close source your Genetics. Publication is the closest thing to GPL they have in biotech and this is moving the wrong direction.
  • Do you have to restart to install 2 Kernel Monte?

    Damn!

  • Its difficult for any multi-architecture distribution to use a Linux kernel monte as introduces an basic inconsistancy between how the system behaves on different architectures.

    Good idea though, which it was arch independent.
  • Which is what to do if I find hardware I don't understand? Stop, an offer the option to boot a more generic (install/config/debug) kernal.
  • You wouldn't believe how many surveys I've filled out where they said they'd email me the results. Of course, they never did...maybe they'll prove to be the exception to the rule.
    ---- Sigs are bad for your health ----
  • I think this is a reboot. Just as using one Linux Kernel to load another is, just as Alphas have been doing. Ever heard of MILO-Linux Mini-loader? Milo is one of the ways to boot AlphaLinux (and my favorite one) Milo is a linux kernel designed to boot another Linux kernel. It has a shell (supporting only a few commands), and can boot from anything the kernel supports. (Have to try it with parport zip disks, just for fun :) Anyway, This new x86 thing sounds just like milo, with the ability to load from a running kernel.
  • That's my point, it's totally equivalent to a warm boot. The kernel uptime goes back to zero.
  • by chris.dag ( 22141 ) on Monday December 11, 2000 @03:30PM (#566518) Homepage

    The open letter from Ewan Birney and Sean Eddy that the genomeweb article talks about can be read in full at:
    http://bioperl.org/pipermail/bioperl-l/2000-Decemb er/001826.html [bioperl.org]

    If you plan to write to Donald Kennedy check the listserv email thread to read about a correction to his email address.

    -chris

  • by stripes ( 3681 ) on Monday December 11, 2000 @04:50PM (#566519) Homepage Journal
    It may be downtime but it is much less then if you had to rebbot and wait for the BIOS, hardware detection etc.

    Hopefully that is skipping the BIOSes hardware detection, not Linuxes. If it skips Linuxes then:

    • You can't add support for a new device because the old kernel didn't detect it (somewhat minor)
    • You might install a kernel that fails to recognise a device you need, but still can run it. A later reboot will fail to get you a working system. It could be many months and sever kernel upgrades before the problem shows up, making it extremly hard to track down (not at all so minor)
  • by Trepidity ( 597 ) <delirium-slashdot AT hackish DOT org> on Monday December 11, 2000 @03:30PM (#566520)
    Of course since you have to unmount the filesystems to use this Linux kernel swapper it really is effectively a reboot. The system will still be up for the upgrade, but anything anyone is trying to do on it will be restarted, so you can still strike fear into the hearts of users. Not to mention of course that if you are not using an obsolete version of Windows, changing IP addresses and adding protocols do not require a reboot - stop spreading FUD.
  • I have said it before, and I'll say it again; the only way to put a stop to all these stupid uptime dick-size contests is to put in a writable /proc/uptime.

    If anybody can make their damn uptime say whatever they want, it'll stop.

    -
  • Well, the best salary resource on the web I have found is salary.com, give it a try.
  • The point is that it isn't emulating anything. It's a virtualization, not emulation.

    It's nice in a lot of situations where you want to easily be able to pop into an OS and test something without constantly rebooting and shutting down your email, etc.

  • Yeah, then they'll reason they really only can afford ONE of you, which does ALL the work.
  • interestingly, microsoft will install datacenter to your needs and specified hardware, afaik.

    in fact, you can not buy win2k datacenter on its own- must be installed by a retailer (and not your local store, we're talking behemoths here). why that is, i'm not completely clear. from what i understand tho, it is written very specifically to follow some sort of guidelines with high end hardware, and every implementation is different.

    as for why there's a double standard, well... corporate culture sometimes (whether justified or not) seems to prefer staying with the tried-and-true. yes, linux might have .3% more uptime than windows, but those corp whores above you freak out when they see you hacking at their systems in an effort to make them run better.

    "open source" they think, "that means ANYONE can look at the code, and how secure is that?" talk until you're blue in the face, you won't convince them.

    fact is, nt worked. and worked reasonably well for a long time. 2000 works. and works better, and probably will be around longer than nt and in a wider variety of uses.

    whether or not linux is a better os doesn't matter because of the previous paragraph. it's like my dad and ford. he buys the damned things because he had one as a kid that lasted for 10 years without a problem. so now he swears by it. (freak:)

    (((look out, a not-anti-m$-stance on slashdot! duck and cover! flame away!)))

  • I don't know too much about Plex86...How stable is it?
  • I participated last year, at which time I was not a member. They do, in fact, give you the results if you participate and are not a member.
  • by RJ11 ( 17321 ) <serge@guanotronic.com> on Monday December 11, 2000 @02:13PM (#566528) Homepage
    I haven't played with it yet, but I heard that the newest Solaris revision supports live patching of the kernel. That's very cool. Imagine never having to reboot for a software related thing again? Hell, on some of the E class Sun boxen, you can hot swap memory, CPUs, and other I/O devices. About time someone started working on technology like that for the PC (granted it's a different market and less demand).

    Speaking of which, it would also be nice to have an x86 hardware bootloader instead of the outdated BIOS. Being able to load an OS or boot from a differnt device via a commandline would be a great feature for the home PC.
  • > -- I use Macs for work, Linux for education, and Windows for cardplaying.

    Then, in your user bio you claim that you use w98 at work.

    Mmmm.


    Technically, they're both true. My job is to do tech support for the Macs at our location. However, the company uses a total POS call-tracking system that only runs under Windows (no, it doesn't have any web interface) and uses Outlook to schedule the calls. So, I had to have them buy a copy of Virtual PC for me to use solely with this one program and Outlook. Hence the "forced to use" part.

    You're right, though, I should probably make that a little less contradictory.


    --
  • Perhaps you should have gone through the survey:

    "Use this form if you are NOT a member of SAGE but have submitted data to the SAGE System Admnistrator Salary Survey 2000 and wish to have the results posted to you. We have made this form using 'USENIX look and feel' to..."
  • eeek. i would choose 3.11... get the larfs out.

    anyways, in rgards to w2k - yes, it is being used. it is rare, but ramping up. i just installed a 2k server as part of our nt network emulating a pdc we took down, and it works beautifully. it is one of the 2 dns servers we have, and in the 4 months since i installed it, we have had 100% uptime on that system.

    my school just upgraded all (read: ALL) of their servers to 2000. this could maybe be because they teach 2000 cert, but hey. :)

    2000 is gaining acceptance as ppl realize it isn't 95, 98, or nt. it's better than all of them. it's gonna be around a while, linux fans... i know, it must suck to admit, but ms actually put together a good os for a change.

    now if only we could get a beowulf clu- oh, sorry.
  • by modus ( 122983 ) on Monday December 11, 2000 @02:14PM (#566532)
    Two Kernel Monte stops you from a having to do a warm reboot of your system, but it is 100% equivalent to it, from the user perspective. The entire system, including the previously running kernel, is killed off and the new kernel is booted up. All you save is the 10 (or more) seconds waiting for your system to do a reboot. So, the whining will continue :-).
  • I usually read /. through an automagic headline grabber -- what can I say?

    I would wish that there was a way to configure /. to show all headlines within a given section on the main page.

  • by Hollins ( 83264 ) on Monday December 11, 2000 @02:51PM (#566534) Homepage
    Get us all a raise.

    Post inflated numbers to SAGE and then direct the boss to the site during your next salary negotiation.

    C'mon, everyone. Chip in!
  • by cluge ( 114877 ) on Monday December 11, 2000 @02:14PM (#566535) Homepage
    The ultimate advantage over NT. New kernel no reboot. That way you can compile a the kernel a couple of different ways and then test for performance advantages right then and there. Or is it, lets consider.....

    Linux : built to your spec, optimize to your hearts content. Preserve uptime(ugh)

    Windows : Change IP address reboot. Add protocol : reboot. Install program : reboot. Upgrade : reboot

    MS has has won an important aspect in the ever lucrative server market. That is the "reboot" aspect. Reboots are becoming more and more popular. Reboots are now considered fun, they give sysadmins a good excuse for a traditional "coffee break". Reboots are familliar. What will sysadmins do without them? How will a sysadmin be able to wield his usual power. The phrase "We will be down for an upgrade!" no longer holds water! ACK what is Linux doing?? What are they thinking?? Are they putting us SYSADMINS out of business? Reducing our free time? I no longer strike fear into the hearts of lusers by saying "we have to upgrade the software on the server"

    Do you realize just how difficult it is to write with my tongue so firmly stuck in my cheek?

  • Whoooo... you know that a lot of nerds have to be into S&M...

    The computer is the ultimate 'bottom'. It will do whatever you tell it to, but only if you tell it every single last detail. And if you tell it the slightest thing wrong, at best, it will stop and do nothing. At worst, it will look like it's doing everything perfect, until it can cause the most possible damage...

  • Here's a User Friendly strip related with what you said:
    http://ars.userfriendly.org/cartoons/?id=19990302 [userfriendly.org]

    ---
    Guillaume

  • Running monte is similar to /sbin/reboot -f. You should unmount (or remount read-only) and sync all your file systems before using monte.

    In my opinion just the fact that you're changing the kernel means you just rebooted. Whats the difference between "loading" a new kernel and "booting" a new kernel? All the initialization (bios, pci, scsi, etc.) maybe?

    Running monte in its default mode is preferable since the setup code performs some important functions like video detection/initialization and APM bios initialization.

    Changing the OS (kernel), re-initializing some things here and there, 0.k. at best this is half of a reboot. You may choose to call this a "reb", a "oot" or any C(6, 3) conbinations ;-p

    Great work anyways Erik! I'll try it ASAP!

    --
  • Well I admin about 20 machines running five different operating systems on three different networks. And I make ...

    $10k


    I'm a college student. YAY


    --
  • by Dominic_Mazzoni ( 125164 ) on Monday December 11, 2000 @05:30PM (#566540) Homepage

    I'm currently using VMWare and I actually think it's very worth having. I'm developing cross-platform (Linux/Windows/Mac) software, and while I do 90% of the coding in Linux, I need to recompile the Windows version all the time. It's much faster to resume my VMWare session and do the Windows compile there than it is to reboot into Windows. Plus, I can continue using Linux while it compiles, and the system is totally responsive. When I'm done using Windows for a while, I can just suspend my VMWare session and free up all of its memory.

    The other reason I use it is that there are still a fair number of fun web sites out there that only really work in IE. I do 90% of my browsing using Konqueror or Netscape in Linux, but if my AOL-using friend sends me a link to a cool game that requires a Windows plug-in, I like being able to check it out without having to reboot.

    So I'm all in favor of Plex86. As soon as it can run VC++ and IE, I'll be switching.

  • Beos had versions for windows, as well as linux and many other platforms..
    Where's my Plex86 for windows? :P

    -since when did 'MTV' stand for Real World Television instead of MUSIC television?
  • Well, if nothing else, it does eliminate some of the time spent during a reboot process. And I could be wrong, but even on a warm reboot when the PC does BIOS init, doesn't that cause the hard drives to reset, thereby losing their internal read cache? so, no BIOS init, and your hard drive caches are still nice and full and stuff like that. Anyway, I think it sounds just as useless as everyone else, except to speed boot time barely. (back in the OS/2 days I used to sit on #os/2 on one puter and reboot the other one over and over after changing various configurations to see how to make booting faster. man that was dumb!)

    Seriously, I'd be willing to bet that some smart induhvidual out there will take this and maybe combine something like the kernel module concept, and the concept of being able to dynamically load parts of executeables (what is that, dload()? i only remember dealing with it while working with MudOS on Linux), merge it all together, and come up with a kernel that you can reload 99% of the whole thing without disturbing other pieces of it. Sounds like an interesting beginning.

    Remember, all the useful projects have humble beginnings. Just like when Linux started, it was a boot block program that could print "A" and "B" simultaneously to the screen. LOL
  • oh oh oh! Another useful thing is that people developing things inside the kernel would be able to test them out a lot faster. Perhaps a first step modification to this would be some way to save the state of the system (at least the executeables running and their parameters) and make an attempt to restore it.. then you could, as a developer, wire up a quick shell script to compile your development kernel, monte over to it, with much less wait time than the complete and total reboot. Also, looking at the docs, I see that it says that it's preferable to have the system go back to real_mode and run the linux init code all over again, but it has the option of starting right out in protected mode.
    This gives me the idea that with some kernel modifications behind this module, it could well be possible to just kickstart a new kernel while one is running, and just transfer control to the new one... hmm... if I understood how the modern internals of Linux worked, this would be something I would look into.. but i'm about 8 years out of it, so I'll leave it up to the younger hackers with more stamina than me.

  • You can find a link to the 1999 SANS [sans.org] (System Administration, Networking, and Security) salary survey right here [sans.org]. You can give them an email address to receive the survey (.pdf) by mail. If you don't want to give them an email address, I've posted a copy on The Linux Pimp [thelinuxpimp.com], which you can view right here [thelinuxpimp.com].
  • I agree. This isn't HA at all. You need clustering to get _true_ non-downtime, and be able to upgrade kernels. One part of the cluster upgrades while the other handles the load, then you switch load and the other part upgrades.

    Sun gets it right.
    http://www.sun.com/fullmoon/
  • Every time there's a major new kernel release, there's whining on Slashdot about killing uptime, and how much it sucks to have to reboot to get a new kernel. Well, whine no more.

    Erik Hendriks at scyld.com brings us Two Kernel Monte [scyld.com], a 'kernel module which allows Linux to load another kernel image into RAM and restart the machine from that kernel.' The only major limitation appears to be that it will not work with SMP machines. Apparently Erik got the idea from the Linux Bios Project.


    This is not an in-place kernel upgrade tool. (That's very difficult, especially without stable data structures and APIs. It's possible, but don't hold your breath.) You are avoiding the BIOS initialization sequence (thus saving some time), but you're still rebooting. The new kernel doesn't know that it was loaded by a Linux system rather than LILO. The old kernel is clobbered (including all its data structures, device drivers, running processes, etc.) and is no longer running. This still kills your uptime and requires a reboot to get a new kernel. (Feel free to keep whining.)

    This project is the Linux equivalent to the DOS program LOADLIN, which loads a Linux kernel (like LILO does) and executes it. The kernel then takes over the machine and starts booting. "Two Kernel Monte" does exactly the same thing, complicated by the need to disable paging, relocate the newly-loaded kernel into a contiguous memory block (in protected mode), return to real mode, and ask the video BIOS to reset the video mode to 80x25 text, all before executing the new kernel. As with LOADLIN, the new kernel will take over the machine and begin booting. Whether started by LILO, LOADLIN or Two Kernel Monte, the end result is much the same -- the new kernel ends up with complete control over the machine (without knowing or caring how it got there), and the old OS (if any) is moot.

    This may seem useless at first glance, but it's a perfect tool to test a new kernel before installing it. Also, the full power of a running Linux system would be available to obtain (or create!) the kernel image to be booted, which could enable some interesting tricks that would be difficult (or impossible) to implement in LILO. It should also be fairly easy to boot a kernel from a complex filesystem. (e.g. NFS, Coda, AFS, etc.) It might also allow for more complex boot-time decisions, such as dynamically selecting the best kernel based on complex criteria. In principle, it should be as easy to execute alternative operating systems in place of Linux, but I don't know if this works. Any form of compression or encryption (no matter how complex) could potentially be used, even if virtual memory is required to decrypt or uncompress the kernel because of limited RAM! (Try that in LILO!)

    In short, this is a cool project, and potentially more useful than it may first appear.
  • They could *really* be dicks and tell you to go generate the data yourself. They're not stopping you from doing that, they're just trying to make some money for all their work. Why ever would they give the data away for free? What kind of business model is that? Sheesh.
  • I just tried installing kmonte with linux 2.4 test7 (which I am still running) and it installed itself like this:

    install -m 644 kmonte.o //lib/modules/$kver/misc

    since 2.4 has no /lib/modules/*/misc/ directory, the module itself becomes "/lib/modules/$kver/misc", and doesn't load anyway because of unresolved symbols. I suppose we will see it for 2.4 some day, but not today!
  • If anybody can make their damn uptime say whatever they want, it'll stop.

    The better solution is to ignore the uptime-dick contests and get on with your life, rather than lowering the kernel's quality as a piece of software.

    Of course, there is historical precedent for such a solution. As I recall, the users of the old Incompatible Timesharing System had problems with some of the newer hackers trying to crash the system, as sort of a coming-of-age ritual. The authors of ITS added an explicit kill command, so that any user could crash the system at any time, simply by typing "KILL" at a command prompt. Suddenly, there were no more attempts to crash the system.

  • Disclosure: I run the Real Rate Survey/Real Salary Survey found at http://www.realrates.com cited in the previous message.

    Salary.com looks pretty lame to me. Their explanation of how they get their figures is high on mystery and "secret sauce." An HR professional told me that the word in the industry is that they merely multiply some generic job title numbers by a factor that is the Census bureau's cost of living figure for the area. Not very accurate.

    Indeed, after years of looking at surveys, I don't have much respect for any summary numbers purporting to tell you what salaries are in some niche.

    There is no such thing as "the salary" for a Unix Sysadmin. There's all different levels of responsibility and all kinds of complexity in the work environment, all kinds of employers in all kinds of locations and industries, but this detail gets smooshed out in surveys that only show you averages or medians.

    Any salary survey that combines the salaries of Java programmers in Silicon Valley with that of Cobol programmers in Des Moines to come up with a salary for a "Sr. Programmer" is garbage--and alas, that is what most of them do. That's about as useful as an average "American" salary figure computed with my salary, yours, and Bill Gates'.

    At Realrates.com All our data is viewable on the site, using the SEARCH. You don't have to rely on medians, averages and other boiled down useless figures because we show you all the data points that describe the job--where, how long, what industry, what software, what credentials etc. etc.

    I wish more surveys did the same as it is the only way that people can really figure out what they can expect to earn.

  • by n3rd ( 111397 ) on Monday December 11, 2000 @02:22PM (#566551)
    Whoops, I need to take that back.

    You aren't able to view the results, but if you fill out the survey, you can submit information and the results will be sent to you.

    I was under the impression that one could not view the results due to the surveys being password protected on the web site.

    My most sincere apologies.
  • in fact, you can not buy win2k datacenter on its own- must be installed by a retailer (and not your local store, we're talking behemoths here). why that is, i'm not completely clear

    It has to come with a 4 9's uptime gurantee that must be supplied by the retailer. That's the whole point of datacenter. Does it work like that? No idea. And 4 9's isn't even that impressive IMNSHO.

  • It may be downtime but it is much less then if you
    had to rebbot and wait for the BIOS, hardware detection etc. If you can afford a cluster or another HA solution this is pur magic. Think a
    single server that has to have a new kernel for whatever reason, this could allow it to be back up in 10 seconds instead of 1 to 3 minutes (Ive seen
    machines slow on boot beacuse of slow scsi detection etc)

"No, no, I don't mind being called the smartest man in the world. I just wish it wasn't this one." -- Adrian Veidt/Ozymandias, WATCHMEN

Working...