Well, in one case it does: when you're trying to run a different operating system simultaneously on the same machine. But in most "enterprise" scenarios, you just want to set up several isolated environments on the same machine, all running the same operating system. In that case, virtualization is absofuckinglutely insane.
Operating systems have been multi-user for a long, long time now. The original use case for Unix involved several users sharing a large box. Embedded in the unix design is 30 years of experience in allowing multiple users to share a machine --- so why throw that away and virtualize the whole operating system anyway?
Hypervisors have become more and more complex, and a plethora of APIs for virtualization-aware guests has appeared. We're reinventing the kernel-userland split, and for no good reason.
Technically, virtualizaiton is insane for a number of reasons:
Each guest needs its own kernel, so you need to allocate memory and disk space for all these kernels that are in fact identical
TLB flushes kill performance. Recent x86 CPUs address the problem to some degree, but it's still a problem.
A guest's filesystem is on a virtual block device, so it's hard to get at it without running some kind of fileserver on the guest
Memory management is an absolute clusterfuck. From the point of view of the host, each guest's memory is an opaque blob, and from the point of view of the guest, it has the machine to itself. This mutual myopia renders the usual page-cache algorithms absolutely useless. Each guest blithely performs memory management and caching on its own resulting in severely suboptimal decisions being made.
In having to set aside memory for each guest, we're returning to the OS9 memory mangement model. Not only are we reinventing the wheel, but we're reinventing a square one covered in jelly.
FreeBSD's jails make a whole lot of sense. They allow several users to have their own userland while running under the same kenrel --- which vastly improves, well, pretty much everything. Linux's containers will eventually provide even better support.
Right and wrong have nothing to do with it. The seminal question is whether a given policy produces the desired outcome. Supposing that that we want as much happiness as possible for as many people as possible, we can conclude that for conservative approaches is "no".
Of course, if what you want is feudalism, then any discussion is moot.
(I jest, but no politician that opposes coal will ever be elected, or continue to hold his office after his term expires. Ditto for corn.)
I wouldn't be so downbeat, but I agree that it's incredibly hard for an elected official to resist the siren's call of powerful lobbyists.
Breaking the power of special interests is the most important task for ensuring the long-term viability of our nation. We can talk about the best way to do that, but my favorite publicly funding elections and banning all private campaign money. Unfortunately, doing that probably requires a constitutional amendment.
Another, less direct way is to jack up the income tax at the top end and reduce income inequality. With a more equitable distribution of wealth, the rich won't be as able to use their resources to distort the political process.
We can't have a sustainable electricity grid when lightbulbs are using 100w or even 60w.
France manages to have a Western-style power grid while generating a majority of its power from sustainable sources. I count nuclear as sustainable because it produces no emissions and because we have a stupidly huge amount of fuel for it.
Conservation is futile in the presence of geometric population growth. Do you propose regulating reproduction as well?
As for energy: we have enough nuclear energy to last at least a thousand years with thorium and breeder reactors. If you don't want to go that route, there's a lot of uninhabited land to harness for solar, even at relatively low efficiency. And if that isn't good enough for you, there's orbital solar with practically unlimited potential.
Why not *both* conservation *and* increased generation?
Conversation is good, but it happens automatically when resources are priced appropriately. Look at how Prius sales went through the roof when gasoline passed $4 per gallon.
Any time you need to set explicit efficiency regulations (like CAFE limits for automobiles, the incandescent bulb phase-out, and low-flow toilets), it's the result of an insufficiently-regulated market.
Like nuclear. Cleanest and most manageable power source per unit of energy produced.
Nuclear is my favorite too, but I'll take pretty much anything over coal. Coal is about as bad as it gets: to do worse, you'd have to start clear-cutting forests and burn the charcoal. (And at least that is carbon-neutral!)
You're assuming CFLs are just thrown in the garbage, a fallacy.
As James Madison wrote, "if men were angels, no government would be necessary."
We both know that people just throw the things out. You could label a product with "IF PUT IN THE GARBAGE, THIS PRODUCT WILL TURN YOUR HOUSE INTO A SMOLDERING, RADIOACTIVE CRATER AND REDUCE THE LENGTH OF YOUR PENIS BY TEN FEET" and people will still throw it out. That won't change until you start inspecting household garbage, and I'd love to see a plan to do that cheaply.
In creating any plan, we must first acknowledge that human nature is what it is, and that we need to work with it, not against it.
But what if generating that extra watt is cheaper than saving it? (Or equivalently, what if what you have to sacrifice to save that watt is worth paying for the extra power?)
(If by "banned" I mean you can't market them for general-purpose lighting.)
Banning things in general as a terrible, heavyhanded policy tool. If you must discourage something, it's far better to simply tax it so that people who really want or need something can get it.
If you outright ban certain items, you create ludicrous crimes like toilet smuggling. Where there's a demand, there's a market, and I'd much prefer a functioning market to a black one.
Markets are wonderful when externalities are accounted for: instead of playing whack-a-mole with polluting technologies, just tax them according to how much pollution they produce. Doing that, you'll find that the market will automatically find the right level of use for these technologies. You'd be hard pressed to strike this balance manually.
Conservation is a red herring: population growth will outstrip any resulting savings. Instead, we should focus on generating energy sustainably. We can do that today with a combination of wind, hydroelectric, and nuclear power.
Conservation almost always reduces our quality of life. Why should we do that when we have the technology to not only save the environment, but improve our lives as well? We should be encouraging people to use more energy when that power makes life easier. By all rights, electricity should be cheap and plentiful.
I can't help but wonder whether conservation advocates feel guilt over civilization itself. I certainly don't. There's no shame in using technology to make our lives better.
Incandescent bulbs are more environmentally friendly than CFLs if you use nuclear energy to power them. Both CFLs and incandescent would be using green energy, but only one would add mercury to the environment.
It used to be a lot easier to be a good assembly programmer. Just consider the implications of modern memory models. Just consider that Ulrich Drepper had to write a seven-part series just on memory.
Programming languages don't have attributes like size and speed: implementations of these languages do. Take Common Lisp for example: SBCL is blazing fast, while CLISP is rather pudgy (albeit smaller). Any conforming Common Lisp program will run on both. Or consider Python --- IronPython and CPython have different performance characteristics. (I'm too lazy to link these now.)
Point being, describing a programming language as "fast" makes about as much senese as describing a natural, human language as "smart".
Obviously, in that case, creating foo.tmp fails --- or if you create foo.tmp elsewhere (not recommended), then the final rename from foo.tmp to foo will fail.
You don't actually need write permission on foo at all --- only on the directory containing it. (Sticky bit aside aside, of course.)
Why should POSIX care which order it happens in? Allowing these two separate operations to be reordered could (in theory) allow a filesystem driver to increase performance by ordering the operations such that the drive head travels the shortest distance.
We're talking past each other. Of course POSIX defines an ordering of operations on a running system. You're talking about writes to the backing disk being reordered, which is outside the scope of POSIX, and about crash recovery, which is also outside the scope of POSIX. My argument is that a reasonable system ought to maintain at least some of the POSIX ordering guarantees across crashes.
My Google skills are apparently lacking, as I'm unable to locate the POSIX specs for file system operations.
I meant write, close, rename, of course. The fsync is unnecessary unless you need to guarantee that to some third party (an SMTP client, a user, an NFS client, etc.) that the data are on the disk after the operation is complete.
Slashdot Mirror
I'm advocating a dilution of power, not a concentration, you ignorant bumpkin.
Or Linux containers for that matter.
(Or for something more mature today, but implemented as a large out-of-tree patch, OpenVZ)
I hate to link to my own comment, but it seems particularly relevant here.
"Here we go again" indeed. Hell, I wasn't around for the first go-round and I recognize it when I see it.
Well, in one case it does: when you're trying to run a different operating system simultaneously on the same machine. But in most "enterprise" scenarios, you just want to set up several isolated environments on the same machine, all running the same operating system. In that case, virtualization is absofuckinglutely insane.
Operating systems have been multi-user for a long, long time now. The original use case for Unix involved several users sharing a large box. Embedded in the unix design is 30 years of experience in allowing multiple users to share a machine --- so why throw that away and virtualize the whole operating system anyway?
Hypervisors have become more and more complex, and a plethora of APIs for virtualization-aware guests has appeared. We're reinventing the kernel-userland split, and for no good reason.
Technically, virtualizaiton is insane for a number of reasons:
In having to set aside memory for each guest, we're returning to the OS9 memory mangement model. Not only are we reinventing the wheel, but we're reinventing a square one covered in jelly.
FreeBSD's jails make a whole lot of sense. They allow several users to have their own userland while running under the same kenrel --- which vastly improves, well, pretty much everything. Linux's containers will eventually provide even better support.
Right and wrong have nothing to do with it. The seminal question is whether a given policy produces the desired outcome. Supposing that that we want as much happiness as possible for as many people as possible, we can conclude that for conservative approaches is "no".
Of course, if what you want is feudalism, then any discussion is moot.
I wouldn't be so downbeat, but I agree that it's incredibly hard for an elected official to resist the siren's call of powerful lobbyists.
Breaking the power of special interests is the most important task for ensuring the long-term viability of our nation. We can talk about the best way to do that, but my favorite publicly funding elections and banning all private campaign money. Unfortunately, doing that probably requires a constitutional amendment.
Another, less direct way is to jack up the income tax at the top end and reduce income inequality. With a more equitable distribution of wealth, the rich won't be as able to use their resources to distort the political process.
We'll hit the wall on food before we have to start worrying about electricity: topsoil is a non-renewable resource, and we're squandering it .
(Some people worry about fresh water, but I don't: we can build as many desalination plants as we can power.)
France manages to have a Western-style power grid while generating a majority of its power from sustainable sources. I count nuclear as sustainable because it produces no emissions and because we have a stupidly huge amount of fuel for it.
Conservation is futile in the presence of geometric population growth. Do you propose regulating reproduction as well?
As for energy: we have enough nuclear energy to last at least a thousand years with thorium and breeder reactors. If you don't want to go that route, there's a lot of uninhabited land to harness for solar, even at relatively low efficiency. And if that isn't good enough for you, there's orbital solar with practically unlimited potential.
Conversation is good, but it happens automatically when resources are priced appropriately. Look at how Prius sales went through the roof when gasoline passed $4 per gallon.
Any time you need to set explicit efficiency regulations (like CAFE limits for automobiles, the incandescent bulb phase-out, and low-flow toilets), it's the result of an insufficiently-regulated market.
Nuclear is my favorite too, but I'll take pretty much anything over coal. Coal is about as bad as it gets: to do worse, you'd have to start clear-cutting forests and burn the charcoal. (And at least that is carbon-neutral!)
As James Madison wrote, "if men were angels, no government would be necessary."
We both know that people just throw the things out. You could label a product with "IF PUT IN THE GARBAGE, THIS PRODUCT WILL TURN YOUR HOUSE INTO A SMOLDERING, RADIOACTIVE CRATER AND REDUCE THE LENGTH OF YOUR PENIS BY TEN FEET" and people will still throw it out. That won't change until you start inspecting household garbage, and I'd love to see a plan to do that cheaply.
In creating any plan, we must first acknowledge that human nature is what it is, and that we need to work with it, not against it.
But what if generating that extra watt is cheaper than saving it? (Or equivalently, what if what you have to sacrifice to save that watt is worth paying for the extra power?)
Banning things in general as a terrible, heavyhanded policy tool. If you must discourage something, it's far better to simply tax it so that people who really want or need something can get it.
If you outright ban certain items, you create ludicrous crimes like toilet smuggling. Where there's a demand, there's a market, and I'd much prefer a functioning market to a black one.
Markets are wonderful when externalities are accounted for: instead of playing whack-a-mole with polluting technologies, just tax them according to how much pollution they produce. Doing that, you'll find that the market will automatically find the right level of use for these technologies. You'd be hard pressed to strike this balance manually.
Coal power plants, not light bulbs, are the problem.
We need a sustainable electric grid, and the best way to create one right now is to tax coal and subsidize alternative power sources.
Conservation is a red herring: population growth will outstrip any resulting savings. Instead, we should focus on generating energy sustainably. We can do that today with a combination of wind, hydroelectric, and nuclear power.
Conservation almost always reduces our quality of life. Why should we do that when we have the technology to not only save the environment, but improve our lives as well? We should be encouraging people to use more energy when that power makes life easier. By all rights, electricity should be cheap and plentiful.
I can't help but wonder whether conservation advocates feel guilt over civilization itself. I certainly don't. There's no shame in using technology to make our lives better.
Incandescent bulbs are more environmentally friendly than CFLs if you use nuclear energy to power them. Both CFLs and incandescent would be using green energy, but only one would add mercury to the environment.
It used to be a lot easier to be a good assembly programmer. Just consider the implications of modern memory models. Just consider that Ulrich Drepper had to write a seven-part series just on memory.
Huh? If you need to lock your program into memory, just call mlockall(2) when your program starts up (and then drop privileges, of course).
Theodore Ts'o' actually suggests there be a binary that just opens a file and calls fsync on it, and that all shell scripts use it.
Talk about being out of touch with the actual requirements of users...
Programming languages don't have attributes like size and speed: implementations of these languages do. Take Common Lisp for example: SBCL is blazing fast, while CLISP is rather pudgy (albeit smaller). Any conforming Common Lisp program will run on both. Or consider Python --- IronPython and CPython have different performance characteristics. (I'm too lazy to link these now.)
Point being, describing a programming language as "fast" makes about as much senese as describing a natural, human language as "smart".
Obviously, in that case, creating foo.tmp fails --- or if you create foo.tmp elsewhere (not recommended), then the final rename from foo.tmp to foo will fail.
You don't actually need write permission on foo at all --- only on the directory containing it. (Sticky bit aside aside, of course.)
Yep. HTML5 blesses and standardizes a lot of the compatibility workarounds everyone is already using.
We're talking past each other. Of course POSIX defines an ordering of operations on a running system. You're talking about writes to the backing disk being reordered, which is outside the scope of POSIX, and about crash recovery, which is also outside the scope of POSIX. My argument is that a reasonable system ought to maintain at least some of the POSIX ordering guarantees across crashes.
It's not free.
I meant write, close, rename, of course. The fsync is unnecessary unless you need to guarantee that to some third party (an SMTP client, a user, an NFS client, etc.) that the data are on the disk after the operation is complete.