as 'outside government jurisdiction'. A ship HAS to be registered and carry the flag of SOME nation, and it will be subject to the laws of that country. So a ship is no more or less outside the law than if you built your data center in that country.
Supposing someone has a ship which is NOT registered anywhere, then it is essentially 'fair game'. If say the US didn't like what you're doing they can just sail on up and do whatever they want with you. They could certainly board and seize any such vessel, after all who's going to object? In theory there might be some construction of maritime law that provides some protections, but without a government capable of objecting you're basically SOL.
So, there would be no consideration on Google's part of evasion of law. Possibly a way to choose a regulatory regime you like, but that's about it. Plus remember any large corporation is pretty much held hostage to its investors, insurance requirements, financing, and ultimately to whatever nations it has substantial business interests in.
Yeah, that's what I was about to say. Any "IT company" that doesn't even review code, doesn't write requirements and/or tests that are thorough enough to catch that kind of stuff, and puts each programmer entirely on their own on a given project deserves what it gets.
Lawyers, accountants, etc ARE professionals, they are regulated by professional organizations and have to prove their qualifications in order to be licensed to practice. IT workers have no such regulation. In effect the State tests your lawyer for you.
I agree the tests for programmers are almost universally stupid and worthless, but if it makes some PHB someplace happy, who cares?
code and the process of coding. OK, code is like air, you got nothing without it, obviously. And bad code is certainly a minus for a project, but less of a minus than no code at all...
What I mean is if you go to SourceForge and poke around you'll find that there are a really large number of nascent projects that are basically no more than a name, a description of an idea, and nothing else. Rare is the instance where such a project attracts any attention. People are usually looking for solutions, not so much ideas. If I need something those empty projects really don't offer me anything. There is precious little motive for OSS developers to 'join' such a project, they can simply set up their own project, one that DOES have whatever code they came up with, and at least that project will offer some sort of technical starting point.
You'll also find that the process of implementation itself often serves to help focus and refine a raw idea. Even more valuable in that regard is the input of other people who are actually working on the implementation and the idea with you.
Projects succeed or fail for a wide variety of reasons, most of which, especially at the beginning, are not really technical in nature. Just as in the commercial world. For every Linux Kernel, or Apache, or whatever there are or were probably a 100 people who set out to build a POSIX compliant OS kernel or a high performance web server. Again the same sort of examples can be drawn from the commercial software world. Success comes from timeliness, luck, savvy promotion, political/managerial skill, determination, quality, technical excellence, and probably many other factors.
To focus more on the question at hand, I would say that producing a mediocre initial implementation of an idea yourself is not necessarily a bad idea. If, as you say, it is not really a highly difficult idea to implement then chances are you CAN produce something yourself. Maybe it isn't great code, and maybe your prototype won't much resemble the eventual mature project down the road, but it will provide some kind of starting point. Something people can look at and play with and improve on, and something they can use to get a handle on the concept and understand what it is you ultimately want to do.
I don't know what your idea is, and I don't know how fully formed it is. Thus I can't really say whether or not it would make sense to pay someone to work on it. Very few software projects are successful when the customer has less than a precise idea of what they want code to do. If you can articulate the goals of the project, what the code needs to do, and some vision of what it should look like from the perspective of various stakeholders (users/admins/developers/business/etc) then it might be worth paying someone to do it. But if you go that route really make sure you go through the process of articulating all these things, write them down, try to discuss them with others who might be interested.
If you can't articulate things at that level, then chances are anyone you hire to work on the thing will at best end up spending a lot of extra effort, time, and money, and chances are slim that the results will be satisfactory.
The other issue with say using a 'rent-a-coder' is that you really have little idea of the sort of quality of person you will get. They may well not be any more skilled at coding than you are yourself. Maybe worse. Sure, you can check their past work history and talk to them and maybe look at samples of their work, but if it were simple to pick out the good developers from that crowd then everyone would have crackerjack dev teams. Also I think you'll find the really good people you CAN find that way are either booked solid, or they quickly end up permanently attached to some team someplace and what is left in rent-a-coder land are the ones that aren't so great. Plus a lot of those type guys ARE good in the sense that they are quite skilled at quickly knocking off bits of code that do some little task, but they mostly aren't good
This is a fairly common fallacy which most less experienced sysadmins fall into. You can only optimize a particular software stack performing a particular task with particular load characteristics in a particular server/network environment. In other words you could optimize your CentOS 5.2/Apache/TWiki corporate intranet running on hardware XYZ on whatever the specific network configuration is, but you CANNOT simply 'optimize a server'.
In general any competent Enterprise Class Linux distro (CentOS, SLES, RHEL) will out of the box be configured for reasonably optimum overall performance on common hardware. If it wasn't, they'd tweak it so it was. Not to say that whatever the defaults are ARE optimum for your given hardware, there could be driver settings or similar things that may even be horribly bad once in a while. No vendor can anticipate all configurations, but until you have specific software you know you will be running and the specific hardware it will run on, tweaking is possibly educational in a basic sense, but that's about all.
Similar comments could be made about security, although there are some pretty good general rules that will apply in most situations. Again, a competent Enterprise Class distribution WILL be pretty much locked down. The best thing to do there is obviously patch up to the latest patch levels and review the vendor's security update database to ensure there aren't any known gotchas that the patches don't cover, like sometimes an app will be fine, but the supplied default configuration may present vulnerabilities in certain situations.
I think overall your plan is not a bad one. Most of the people here (that DO know anything) learned it 'hands on' like that.
Overall security is really the sysadmin's most challenging task. I don't know if you'll be in charge of an entire network, but either way you'll want to gain an in depth understanding of things like DOS/DDOS prevention techniques, DNS security considerations, Proxies, SSH, IDS systems, etc. If you have input or responsibility for security policies and/or enforcement then there could be a LOT more to know/do. Actually a lot of that has little to do with specific applications, or even really technology itself.
There are a few dozen other things you may well want to know all about. I'd suggest maybe doing some extensive research online and find out what you do and don't know about, then you can start to fill in where you may be lacking.
Heh, there is no more 'long run', this civilization has hosed itself. Might as well be the captain of the Titanic right about the time the lookout said 'Iceburg ahead'. Already screwed. Typical humans, always reacting after it is too late.
I don't use Suse, but my understanding was they were dropping support for it. Mandriva definitely did and it was also their default fs. Dunno about any of the smaller distributions.
Sounds like that is likely to change. No mainstream distribution is supporting it at this point as a first line file system AFAIK. The handwriting seems to be on the wall.
In my experience not all combinations of these worked so good. I had horrible performance with striping via MD and putting ReiserFS on the resulting raid device. After experimenting some what I found was that arrangements like that are very hardware dependent, and the possible 'good' configurations are not really obvious. It probably has to do with implementation details of the host interface, the drives, driver and kernel issues, etc. No doubt someone with the time to sit down and really analyze a particular configuration could come up with explanations.
The upshot is I wouldn't necessarily say that any two hardware/software configurations are 'pretty much equivalent' just because the hardware specs are fairly similar. One configuration might make excellent use of performance enhancing features, while the other could exhibit pathological behavior that kills performance. Obviously what actually works in a given real world configuration is what matters to people, so they should use what works well for them, it just may not be the same thing for everyone.
Sure, but as soon as you say 'command line', it makes little difference how simple the commands are. Your average person is already lost. I don't think ZFS is hard to manage either, the majority of the world just isn't even sure what a file is, let alone a partition, etc. Getting them to grasp concepts like storage pools etc is probably expecting too much. This isn't a ZFS problem, it is just a challenge to present things in a way that is comprehensible to everyone and gives them useful options.
In any case, ZFS on Linux is a whole other beast, setting up a userland file system is yet another magilla, lol. My advice to ZFS supporters is "figure out how to build management tools that both hide the complexity, and allow access to the cool functions in a useful way." It will be a bit challenging, but sure can be done.
my experience is totally different. with big filesystems (1-4TB), doing a full copy or rsync from one to the other completes in nearly the same amount of time, but if either source or destination was ReiserFS (or worse, both) other accesses to any other filesystem on that machine get terribly slow. also, just mounting a Reiser FS over a few hundred Gigs (cleanly unmounted) is a multi-minute task, while on ext3 it's instantaneous.
one possible explanation might be that ReiserFS (3.6, at least) uses the BKL (big kernel lock), which is a huge obstacle to real scalability. there's currently a lot of work in the kernel to remove last uses of that... who knows what that could mean for ReiserFS support?
Hehe, I have a 1 TERABYTE Reiser3 partition (spanned on 2 500 gig drives). Works fine. mounts instantly, fsck takes maybe 4-5 seconds. Not really sure why you see slow mounting, I've used it on probably 20 machines at least in the past and never had that problem.
Speaking for myself I don't know enough about the details of kernel internals and impact on performance. It seemed to be reasonably fast, let us say fast enough. My guess is in general it worked pretty well for the time frame when it was developed, but the Namesys people were never really kernel insiders. I don't get the impression that there was a lot of concern about what the impact of big kernel architecture changes would be on ReiserFS, nor did those guys have a lot of input on kernel issues. It was kind of the poor stepchild of fses. Once active development wound down it has kind of drifted off into the shadowland of code that may work, but nobody is going to worry about it. Kind of too bad really, it did have some interesting potentialities even if they were mostly never realized in actual use.
That's basically it. I liked ReiserFS fine and used it extensively. But sooner or later kernel changes are going to break it, and nobody is going to fix it apparently. If a group of developers appears that supports it and carries it forward, then nothing is really wrong with it at all. Personally I've just decided that it is probably best overall to phase out use of it before it goes away entirely.
Yeah, except as I pointed out above, userland file systems suck CPU performance (due to context switching and the resulting TLB invalidations amongst other things). An HTPC usually already has significant real time demands on its CPU, and most of them are built using lower end components that don't require a lot of noisy fans. So you generally don't have those CPU cycles to spare.
In other contexts userland ZFS might be appropriate, and it might even be fine for a given MythTV box depending on what the hardware is and what is being done with it, but it still isn't a suitable general recommendation for that application at this time IMHO.
No, I am afraid it goes even deeper than that. I wouldn't want surveillance even in a "perfect democracy". Majority rule and surveillance make bad bedfellows. There is nothing inherently "good" about majority rule, it just happens to be more just than the alternatives. In a majority rule state, you can still have 49% of the population enslaved by other 51%.
IMHO this is a misunderstanding of the dynamics involved in a representative system. 51% don't 'tyrannize' the other 49% (well, it can happen, but that isn't the primary problem). The real problem is that a small minority can simply play various segments of the people off against each other. If the people don't adhere to the principle that any violation of the rights of some must be considered a threat to the rights of the whole, then tyrants will inevitably undermine everyone's rights. They will simply appeal to a plurality of the public to subvert the rights of one small segment, and then turn around and do the same thing to another segment of the population until they have concentrated so much power in their own hands that nobody effectively has the exercise of any of their rights anymore. I know of no defense against this except that the vast majority of the people must steadfastly refuse the abridgment of the rights of each and every other citizen.
I specifically want a NON-Democratic government. I am not asking for a tyranny or a dictatorship or anything of that nature, but a government where the whim of the majority isn't law. The things I love best about the American government are its non-democratic parts. The Bill of Rights in particular is a document that sets out exactly what is beyond the realm of majority rule. It sets down rights that are (supposed) to be kept, majority be damned.
You refer to the Constitution, but that is only a piece of paper. It neither guarantees nor defends anything. In the long term the choice is only between a body politic which defends the rights of every individual against any other combination of individuals, or tyranny. What you want is to have all the benefits and none of the responsibilities of freedom. Sorry, not an option.
How do you set up a just liberal non-democratic government where surveillance doesn't send shivers down my spine? The hell if I know. I just know that surveillance and democracies do not go together. I would rather suffer higher crime rates than suffer a surveillance society in a society filled with literally countless laws.
The answer is your desire is simply an impossibility. You might as well be lamenting the fact that you cannot buy a perpetual motion machine. Either we all assert each other's rights against any abridgment, or eventually we will all be enslaved. United we stand, divided we fall. Few and far between are those who seem to truly understand what that means...
ReiserFS was a great filesystem. Certainly isn't a WORSE one now than it was before in and of itself. It is just a matter of ongoing support. Whatever the reason people are dropping support and inevitably an unmaintained file system is going to become at best a marginal legacy tool. Given that it isn't even a default supported Linux fs chances are it will be broken as of kernel X, and then you will be pretty much forced to migrate, so why bother with it?
As to WHY, that I have to assume is political, there was never anything technically wrong with it. In fact there was pretty much everything technically RIGHT about it... Reading between the lines my hunch would be
Hans Reiser wasn't what you would call a 'team player' and he never quite got people on board with his code base.
The developers never quite had a production mentality. Not that it wasn't production quality code, but they didn't make it easy to support
Reiser's conceptions about file systems just didn't match with what the rest of the world thought was most important. He may well have been right about a lot of things, but that mismatch pretty well prevented effective use of a lot of the more interesting potentials in his code.
In some ways ReiserFS is now obsolescent, much like ext3 is obsolescent. The state of the art in fs design is moving on. No doubt ReiserFS could support a lot of what people are now desiring in a filesystem, but it doesn't.
Finally, it was a project with only a small developer base, and now with that group essentially scattered to the four winds my guess would be nobody who's interested in working on it is around who understands the code base. By the time you figure out something that complex, you can often write a replacement yourself in a similar amount of time. So the value becomes marginal unless there is a real serious need for that specific code base.
Most people are using Linux for HTPC applications, BSD is a whole other kettle of fish...
I'm not saying ZFS is HARD to admin, just that a basic ext3 fs is nothing at all to admin. If you know ZFS, it is no big deal. For your average garden variety user they will never take advantage of the ZFS features anyway, so they'd be better of just going with ext3. There is a lot better chance their recovery tools support it, etc.
So, I'll agree with you, if the user happens to be sophisticated and using BSD, then go for ZFS, why not?
The critical factor being CPU overhead. Fuse based file systems are nice and you can solve a lot of problems with them, and they certainly can exhibit good throughput but the situation with an HTPC is that generally you have limited CPU resources and you definitely have significant CPU demands in most cases.
The best case scenario is you have hardware MPEG decoding and all you're doing is watching a stream that is already on disk. In that case you're probably fine with most anything, and even antique machines will usually work fine.
But commonly you might see something like an HTPC running a fairly low power CPU where you're pulling data off some source and meanwhile watching something else, and frequently decoding the incoming stream, transcoding it and at the same time decoding what you're watching (which may be HD content for that matter). In those cases every CPU cycle counts and the overhead of a user land file system is going to burn you.
The upshot being ext3 still ends up in general being the best available recommendation.
ZFS isn't available on Linux. It is a great enterprise class file system, but for the type of application discussed here it is probably overkill in terms of management complexity, etc. Not that it would be bad or wrong to use, and you could benefit from some of the features of course, but it really shines in demanding environments. In any case unless you're running openSolaris it isn't an option.
Ext3 in my experience is just plain inferior to ReiserFS. Recovery and formatting are both slow as death. Like the OP I have yet to suffer any data loss on a ReiserFS since way back in the early days when it first came out. Ext3 seems pretty reliable as well, but the slow recovery times are annoying and once in a while it seems like a whole filesystem just plain becomes irretrievably corrupted. OTOH it does demand less CPU overhead. Rarely a BIG issue, but can be with HTPC type systems.
Overall though I don't think you have a lot of choice. XFS or JFS might be perfectly good solutions, not really had a need to mess with either of them myself so I can't comment. Obviously ReiserFS looks like it has about reached the end and that pretty much leaves Ext3 as the only man left standing in the ring at this point. Cheer up, it works well enough, you'll just have to live without the shrink functionality...;).
works in an environment where the vast majority of 'citizens' are willfully ignorant fools who can't get off their couches let alone carry out their duties as citizens. It cannot...
The fault, my dear Brutus, lies not in the stars, but in ourselves.
No matter how you construct it, a population of sheep are only fit to be sheared. If people will not be bothered to go to even the most minimal amount of effort to maintain themselves in a free state, then they will be slaves. And when they are unwilling to all stand up mutually for the rights of every one else's rights as well as their own, then the only inevitable result will be some form of tyranny because the collective authority of society will be exercised by someone, and if it is not everyone collectively then it will be the few, and they will quickly become the privileged.
The way I see it surveillance and universal identity databases etc are inevitable products of modern technology coupled with human nature. It isn't something we can just close the door on and expect the result to be that such things will simply go away.
It seems more like a choice between acknowledging that we WILL be surveilled, and that there WILL be such databases or sticking our heads in the sand and denying it. Thus two potential situations can arise. Either the surveillance and data acquisition are surreptitious, covert, and beyond the supervision of the body politic, or they are overt and can be subject to certain oversight and control.
I would also observe that there are undeniable practical benefits which could be realized by such technologies. Many of these benefits will not be fully realized if they exist in secret and can thus be put to only a limited set of uses. If they are acknowledged and in the open, subject to regulation and control, and available for certain legitimate uses then we may be able to reap great benefits while mitigating the most serious dangers.
The wise man understands that progress is made in the nature and structure of society, not that of the world. Therefore all progress depends on the wise;)
Is to not buy cutting edge hardware. Any time you go out and buy the latest super duper brand new shiny high performance stuff you're a lot more likely to get something unstable or potentially defective than if you go for the middle of the road stuff.
The high end stuff is always by its very nature pushing the hairy edge of the power, thermal, process, etc envelope. Thinking back I've purchased 5 video cards in the last 12 years or so. All of them worked flawlessly. For that matter I still own 4 of the 5 and 3 of them are still in service and working fine. None of them ever failed.
Seems to me I never had really serious driver issues either. Certainly the drivers always steadily improved to the point of being pretty near 100% reliable. The NVIDIA Quattro card I'm using now works great. I had maybe a half dozen crashes really early on. Nothing since then.
I think hardware manufacturers are certainly obliged to TRY their best to make a reliable product, but if people want to be able to play really demanding 3D games at high res then they're also going to have to be willing to deal with the fact that cards required to do that are going to be less stable. It is a tradeoff. Nobody forces people to buy a 9800 GTX, they could settle for a lower end card if they wanted better stability. That being said obviously it isn't either good business practice nor good ethics to sell a product which is useless or known to be seriously defective.
Sounds to me like NVIDIA ran into an engineering problem with certain products which I am guessing was pretty difficult to uncover before the fact, and now they're doing the right thing and backing those products. I guess it remains to be seen as to how well they do that.
Which do you want? A $3000 perfectly reliable video card or a $150 video card that is 99% reliable? You probably don't have $3000 to spend on a video card. I know I don't. If you do, then chances are you can get what you want, but also be prepared for it to be 5 years out of date, because no amount of NRE will make such a product come out of the pipeline fast.
Beyond that it is really a false choice, because the hypothetical $3000 card will only BE 100% reliable if you also buy only the motherboard, power supply, etc that it can be guaranteed to BE 100% reliable with, and only run the OS and apps it is guaranteed to work reliably with. Even then nothing is 100% reliable. Eventually your perfect card will drift out of spec just a bit and 1 time out of a billion some edge of some signal won't hit the bus inside some timing window in some obscure condition and you'll have a crash.
I helped design, test, and verify avionics systems on practically all major commercial airliners put in service in the '90s. The systems we designed and built are safety of flight critical systems. Those systems are designed, verified, built, and tested to literally the highest standards known to engineering, every bit as stringent as anything you'd find inside a nuclear reactor. Yet STILL we occasionally ran into scenarios with equipment in use in the field where the function of the system was compromised by failure scenarios which simply could not be fully anticipated. The cost of those systems was enormous. No retail grade product could ever conceivably achieve even that level of reliability and still be worth the cost.
So really, what you're effectively saying is 'if we can't have a perfect video card, we should have none at all'.
I agree with you, but 'Science History' or 'Whatever History' is not a science class. I have no use for ID in the science classroom, but the history of human thought certainly is a legitimate and IMHO valuable thing to teach.
Part of the whole problem I think is that the fundies have been ramming ID etc down our throats for years now and at this point the rest of society just wants to be done with the whole thing and are REALLY suspicious of ANY material relating to religion in schools.
In the long run the fundies are going to find out, I suspect, that their tactics have done more damage to their own views than anything else.
Slashdot Mirror
as 'outside government jurisdiction'. A ship HAS to be registered and carry the flag of SOME nation, and it will be subject to the laws of that country. So a ship is no more or less outside the law than if you built your data center in that country.
Supposing someone has a ship which is NOT registered anywhere, then it is essentially 'fair game'. If say the US didn't like what you're doing they can just sail on up and do whatever they want with you. They could certainly board and seize any such vessel, after all who's going to object? In theory there might be some construction of maritime law that provides some protections, but without a government capable of objecting you're basically SOL.
So, there would be no consideration on Google's part of evasion of law. Possibly a way to choose a regulatory regime you like, but that's about it. Plus remember any large corporation is pretty much held hostage to its investors, insurance requirements, financing, and ultimately to whatever nations it has substantial business interests in.
Yeah, that's what I was about to say. Any "IT company" that doesn't even review code, doesn't write requirements and/or tests that are thorough enough to catch that kind of stuff, and puts each programmer entirely on their own on a given project deserves what it gets.
Lawyers, accountants, etc ARE professionals, they are regulated by professional organizations and have to prove their qualifications in order to be licensed to practice. IT workers have no such regulation. In effect the State tests your lawyer for you.
I agree the tests for programmers are almost universally stupid and worthless, but if it makes some PHB someplace happy, who cares?
code and the process of coding. OK, code is like air, you got nothing without it, obviously. And bad code is certainly a minus for a project, but less of a minus than no code at all...
What I mean is if you go to SourceForge and poke around you'll find that there are a really large number of nascent projects that are basically no more than a name, a description of an idea, and nothing else. Rare is the instance where such a project attracts any attention. People are usually looking for solutions, not so much ideas. If I need something those empty projects really don't offer me anything. There is precious little motive for OSS developers to 'join' such a project, they can simply set up their own project, one that DOES have whatever code they came up with, and at least that project will offer some sort of technical starting point.
You'll also find that the process of implementation itself often serves to help focus and refine a raw idea. Even more valuable in that regard is the input of other people who are actually working on the implementation and the idea with you.
Projects succeed or fail for a wide variety of reasons, most of which, especially at the beginning, are not really technical in nature. Just as in the commercial world. For every Linux Kernel, or Apache, or whatever there are or were probably a 100 people who set out to build a POSIX compliant OS kernel or a high performance web server. Again the same sort of examples can be drawn from the commercial software world. Success comes from timeliness, luck, savvy promotion, political/managerial skill, determination, quality, technical excellence, and probably many other factors.
To focus more on the question at hand, I would say that producing a mediocre initial implementation of an idea yourself is not necessarily a bad idea. If, as you say, it is not really a highly difficult idea to implement then chances are you CAN produce something yourself. Maybe it isn't great code, and maybe your prototype won't much resemble the eventual mature project down the road, but it will provide some kind of starting point. Something people can look at and play with and improve on, and something they can use to get a handle on the concept and understand what it is you ultimately want to do.
I don't know what your idea is, and I don't know how fully formed it is. Thus I can't really say whether or not it would make sense to pay someone to work on it. Very few software projects are successful when the customer has less than a precise idea of what they want code to do. If you can articulate the goals of the project, what the code needs to do, and some vision of what it should look like from the perspective of various stakeholders (users/admins/developers/business/etc) then it might be worth paying someone to do it. But if you go that route really make sure you go through the process of articulating all these things, write them down, try to discuss them with others who might be interested.
If you can't articulate things at that level, then chances are anyone you hire to work on the thing will at best end up spending a lot of extra effort, time, and money, and chances are slim that the results will be satisfactory.
The other issue with say using a 'rent-a-coder' is that you really have little idea of the sort of quality of person you will get. They may well not be any more skilled at coding than you are yourself. Maybe worse. Sure, you can check their past work history and talk to them and maybe look at samples of their work, but if it were simple to pick out the good developers from that crowd then everyone would have crackerjack dev teams. Also I think you'll find the really good people you CAN find that way are either booked solid, or they quickly end up permanently attached to some team someplace and what is left in rent-a-coder land are the ones that aren't so great. Plus a lot of those type guys ARE good in the sense that they are quite skilled at quickly knocking off bits of code that do some little task, but they mostly aren't good
This is a fairly common fallacy which most less experienced sysadmins fall into. You can only optimize a particular software stack performing a particular task with particular load characteristics in a particular server/network environment. In other words you could optimize your CentOS 5.2/Apache/TWiki corporate intranet running on hardware XYZ on whatever the specific network configuration is, but you CANNOT simply 'optimize a server'.
In general any competent Enterprise Class Linux distro (CentOS, SLES, RHEL) will out of the box be configured for reasonably optimum overall performance on common hardware. If it wasn't, they'd tweak it so it was. Not to say that whatever the defaults are ARE optimum for your given hardware, there could be driver settings or similar things that may even be horribly bad once in a while. No vendor can anticipate all configurations, but until you have specific software you know you will be running and the specific hardware it will run on, tweaking is possibly educational in a basic sense, but that's about all.
Similar comments could be made about security, although there are some pretty good general rules that will apply in most situations. Again, a competent Enterprise Class distribution WILL be pretty much locked down. The best thing to do there is obviously patch up to the latest patch levels and review the vendor's security update database to ensure there aren't any known gotchas that the patches don't cover, like sometimes an app will be fine, but the supplied default configuration may present vulnerabilities in certain situations.
I think overall your plan is not a bad one. Most of the people here (that DO know anything) learned it 'hands on' like that.
Overall security is really the sysadmin's most challenging task. I don't know if you'll be in charge of an entire network, but either way you'll want to gain an in depth understanding of things like DOS/DDOS prevention techniques, DNS security considerations, Proxies, SSH, IDS systems, etc. If you have input or responsibility for security policies and/or enforcement then there could be a LOT more to know/do. Actually a lot of that has little to do with specific applications, or even really technology itself.
There are a few dozen other things you may well want to know all about. I'd suggest maybe doing some extensive research online and find out what you do and don't know about, then you can start to fill in where you may be lacking.
Heh, there is no more 'long run', this civilization has hosed itself. Might as well be the captain of the Titanic right about the time the lookout said 'Iceburg ahead'. Already screwed. Typical humans, always reacting after it is too late.
All go have a cold one, relax...
I don't use Suse, but my understanding was they were dropping support for it. Mandriva definitely did and it was also their default fs. Dunno about any of the smaller distributions.
Sounds like that is likely to change. No mainstream distribution is supporting it at this point as a first line file system AFAIK. The handwriting seems to be on the wall.
In my experience not all combinations of these worked so good. I had horrible performance with striping via MD and putting ReiserFS on the resulting raid device. After experimenting some what I found was that arrangements like that are very hardware dependent, and the possible 'good' configurations are not really obvious. It probably has to do with implementation details of the host interface, the drives, driver and kernel issues, etc. No doubt someone with the time to sit down and really analyze a particular configuration could come up with explanations.
The upshot is I wouldn't necessarily say that any two hardware/software configurations are 'pretty much equivalent' just because the hardware specs are fairly similar. One configuration might make excellent use of performance enhancing features, while the other could exhibit pathological behavior that kills performance. Obviously what actually works in a given real world configuration is what matters to people, so they should use what works well for them, it just may not be the same thing for everyone.
Yeah, Ironically I got hit with some mod points right after I posted, lol. Taco must love me, I can hardly log in these days and NOT have some, hehe.
Sure, but as soon as you say 'command line', it makes little difference how simple the commands are. Your average person is already lost. I don't think ZFS is hard to manage either, the majority of the world just isn't even sure what a file is, let alone a partition, etc. Getting them to grasp concepts like storage pools etc is probably expecting too much. This isn't a ZFS problem, it is just a challenge to present things in a way that is comprehensible to everyone and gives them useful options.
In any case, ZFS on Linux is a whole other beast, setting up a userland file system is yet another magilla, lol. My advice to ZFS supporters is "figure out how to build management tools that both hide the complexity, and allow access to the cool functions in a useful way." It will be a bit challenging, but sure can be done.
my experience is totally different. with big filesystems (1-4TB), doing a full copy or rsync from one to the other completes in nearly the same amount of time, but if either source or destination was ReiserFS (or worse, both) other accesses to any other filesystem on that machine get terribly slow. also, just mounting a Reiser FS over a few hundred Gigs (cleanly unmounted) is a multi-minute task, while on ext3 it's instantaneous.
one possible explanation might be that ReiserFS (3.6, at least) uses the BKL (big kernel lock), which is a huge obstacle to real scalability. there's currently a lot of work in the kernel to remove last uses of that... who knows what that could mean for ReiserFS support?
Hehe, I have a 1 TERABYTE Reiser3 partition (spanned on 2 500 gig drives). Works fine. mounts instantly, fsck takes maybe 4-5 seconds. Not really sure why you see slow mounting, I've used it on probably 20 machines at least in the past and never had that problem.
Speaking for myself I don't know enough about the details of kernel internals and impact on performance. It seemed to be reasonably fast, let us say fast enough. My guess is in general it worked pretty well for the time frame when it was developed, but the Namesys people were never really kernel insiders. I don't get the impression that there was a lot of concern about what the impact of big kernel architecture changes would be on ReiserFS, nor did those guys have a lot of input on kernel issues. It was kind of the poor stepchild of fses. Once active development wound down it has kind of drifted off into the shadowland of code that may work, but nobody is going to worry about it. Kind of too bad really, it did have some interesting potentialities even if they were mostly never realized in actual use.
That's basically it. I liked ReiserFS fine and used it extensively. But sooner or later kernel changes are going to break it, and nobody is going to fix it apparently. If a group of developers appears that supports it and carries it forward, then nothing is really wrong with it at all. Personally I've just decided that it is probably best overall to phase out use of it before it goes away entirely.
Yeah, except as I pointed out above, userland file systems suck CPU performance (due to context switching and the resulting TLB invalidations amongst other things). An HTPC usually already has significant real time demands on its CPU, and most of them are built using lower end components that don't require a lot of noisy fans. So you generally don't have those CPU cycles to spare.
In other contexts userland ZFS might be appropriate, and it might even be fine for a given MythTV box depending on what the hardware is and what is being done with it, but it still isn't a suitable general recommendation for that application at this time IMHO.
No, I am afraid it goes even deeper than that. I wouldn't want surveillance even in a "perfect democracy". Majority rule and surveillance make bad bedfellows. There is nothing inherently "good" about majority rule, it just happens to be more just than the alternatives. In a majority rule state, you can still have 49% of the population enslaved by other 51%.
IMHO this is a misunderstanding of the dynamics involved in a representative system. 51% don't 'tyrannize' the other 49% (well, it can happen, but that isn't the primary problem). The real problem is that a small minority can simply play various segments of the people off against each other. If the people don't adhere to the principle that any violation of the rights of some must be considered a threat to the rights of the whole, then tyrants will inevitably undermine everyone's rights. They will simply appeal to a plurality of the public to subvert the rights of one small segment, and then turn around and do the same thing to another segment of the population until they have concentrated so much power in their own hands that nobody effectively has the exercise of any of their rights anymore. I know of no defense against this except that the vast majority of the people must steadfastly refuse the abridgment of the rights of each and every other citizen.
I specifically want a NON-Democratic government. I am not asking for a tyranny or a dictatorship or anything of that nature, but a government where the whim of the majority isn't law. The things I love best about the American government are its non-democratic parts. The Bill of Rights in particular is a document that sets out exactly what is beyond the realm of majority rule. It sets down rights that are (supposed) to be kept, majority be damned.
You refer to the Constitution, but that is only a piece of paper. It neither guarantees nor defends anything. In the long term the choice is only between a body politic which defends the rights of every individual against any other combination of individuals, or tyranny. What you want is to have all the benefits and none of the responsibilities of freedom. Sorry, not an option.
How do you set up a just liberal non-democratic government where surveillance doesn't send shivers down my spine? The hell if I know. I just know that surveillance and democracies do not go together. I would rather suffer higher crime rates than suffer a surveillance society in a society filled with literally countless laws.
The answer is your desire is simply an impossibility. You might as well be lamenting the fact that you cannot buy a perpetual motion machine. Either we all assert each other's rights against any abridgment, or eventually we will all be enslaved. United we stand, divided we fall. Few and far between are those who seem to truly understand what that means...
ReiserFS was a great filesystem. Certainly isn't a WORSE one now than it was before in and of itself. It is just a matter of ongoing support. Whatever the reason people are dropping support and inevitably an unmaintained file system is going to become at best a marginal legacy tool. Given that it isn't even a default supported Linux fs chances are it will be broken as of kernel X, and then you will be pretty much forced to migrate, so why bother with it?
As to WHY, that I have to assume is political, there was never anything technically wrong with it. In fact there was pretty much everything technically RIGHT about it... Reading between the lines my hunch would be
Most people are using Linux for HTPC applications, BSD is a whole other kettle of fish...
I'm not saying ZFS is HARD to admin, just that a basic ext3 fs is nothing at all to admin. If you know ZFS, it is no big deal. For your average garden variety user they will never take advantage of the ZFS features anyway, so they'd be better of just going with ext3. There is a lot better chance their recovery tools support it, etc.
So, I'll agree with you, if the user happens to be sophisticated and using BSD, then go for ZFS, why not?
The critical factor being CPU overhead. Fuse based file systems are nice and you can solve a lot of problems with them, and they certainly can exhibit good throughput but the situation with an HTPC is that generally you have limited CPU resources and you definitely have significant CPU demands in most cases.
The best case scenario is you have hardware MPEG decoding and all you're doing is watching a stream that is already on disk. In that case you're probably fine with most anything, and even antique machines will usually work fine.
But commonly you might see something like an HTPC running a fairly low power CPU where you're pulling data off some source and meanwhile watching something else, and frequently decoding the incoming stream, transcoding it and at the same time decoding what you're watching (which may be HD content for that matter). In those cases every CPU cycle counts and the overhead of a user land file system is going to burn you.
The upshot being ext3 still ends up in general being the best available recommendation.
ZFS isn't available on Linux. It is a great enterprise class file system, but for the type of application discussed here it is probably overkill in terms of management complexity, etc. Not that it would be bad or wrong to use, and you could benefit from some of the features of course, but it really shines in demanding environments. In any case unless you're running openSolaris it isn't an option.
Ext3 in my experience is just plain inferior to ReiserFS. Recovery and formatting are both slow as death. Like the OP I have yet to suffer any data loss on a ReiserFS since way back in the early days when it first came out. Ext3 seems pretty reliable as well, but the slow recovery times are annoying and once in a while it seems like a whole filesystem just plain becomes irretrievably corrupted. OTOH it does demand less CPU overhead. Rarely a BIG issue, but can be with HTPC type systems.
Overall though I don't think you have a lot of choice. XFS or JFS might be perfectly good solutions, not really had a need to mess with either of them myself so I can't comment. Obviously ReiserFS looks like it has about reached the end and that pretty much leaves Ext3 as the only man left standing in the ring at this point. Cheer up, it works well enough, you'll just have to live without the shrink functionality... ;).
works in an environment where the vast majority of 'citizens' are willfully ignorant fools who can't get off their couches let alone carry out their duties as citizens. It cannot...
The fault, my dear Brutus, lies not in the stars, but in ourselves.
No matter how you construct it, a population of sheep are only fit to be sheared. If people will not be bothered to go to even the most minimal amount of effort to maintain themselves in a free state, then they will be slaves. And when they are unwilling to all stand up mutually for the rights of every one else's rights as well as their own, then the only inevitable result will be some form of tyranny because the collective authority of society will be exercised by someone, and if it is not everyone collectively then it will be the few, and they will quickly become the privileged.
The way I see it surveillance and universal identity databases etc are inevitable products of modern technology coupled with human nature. It isn't something we can just close the door on and expect the result to be that such things will simply go away.
It seems more like a choice between acknowledging that we WILL be surveilled, and that there WILL be such databases or sticking our heads in the sand and denying it. Thus two potential situations can arise. Either the surveillance and data acquisition are surreptitious, covert, and beyond the supervision of the body politic, or they are overt and can be subject to certain oversight and control.
I would also observe that there are undeniable practical benefits which could be realized by such technologies. Many of these benefits will not be fully realized if they exist in secret and can thus be put to only a limited set of uses. If they are acknowledged and in the open, subject to regulation and control, and available for certain legitimate uses then we may be able to reap great benefits while mitigating the most serious dangers.
The wise man understands that progress is made in the nature and structure of society, not that of the world. Therefore all progress depends on the wise ;)
True, but it REALLY beats raising that remote to tune into John McAncient and crew bullshitting the public AND THEN SWALLOWING IT.
Pardon me while I go gag for a while... ;)
Is to not buy cutting edge hardware. Any time you go out and buy the latest super duper brand new shiny high performance stuff you're a lot more likely to get something unstable or potentially defective than if you go for the middle of the road stuff.
The high end stuff is always by its very nature pushing the hairy edge of the power, thermal, process, etc envelope. Thinking back I've purchased 5 video cards in the last 12 years or so. All of them worked flawlessly. For that matter I still own 4 of the 5 and 3 of them are still in service and working fine. None of them ever failed.
Seems to me I never had really serious driver issues either. Certainly the drivers always steadily improved to the point of being pretty near 100% reliable. The NVIDIA Quattro card I'm using now works great. I had maybe a half dozen crashes really early on. Nothing since then.
I think hardware manufacturers are certainly obliged to TRY their best to make a reliable product, but if people want to be able to play really demanding 3D games at high res then they're also going to have to be willing to deal with the fact that cards required to do that are going to be less stable. It is a tradeoff. Nobody forces people to buy a 9800 GTX, they could settle for a lower end card if they wanted better stability. That being said obviously it isn't either good business practice nor good ethics to sell a product which is useless or known to be seriously defective.
Sounds to me like NVIDIA ran into an engineering problem with certain products which I am guessing was pretty difficult to uncover before the fact, and now they're doing the right thing and backing those products. I guess it remains to be seen as to how well they do that.
Which do you want? A $3000 perfectly reliable video card or a $150 video card that is 99% reliable? You probably don't have $3000 to spend on a video card. I know I don't. If you do, then chances are you can get what you want, but also be prepared for it to be 5 years out of date, because no amount of NRE will make such a product come out of the pipeline fast.
Beyond that it is really a false choice, because the hypothetical $3000 card will only BE 100% reliable if you also buy only the motherboard, power supply, etc that it can be guaranteed to BE 100% reliable with, and only run the OS and apps it is guaranteed to work reliably with. Even then nothing is 100% reliable. Eventually your perfect card will drift out of spec just a bit and 1 time out of a billion some edge of some signal won't hit the bus inside some timing window in some obscure condition and you'll have a crash.
I helped design, test, and verify avionics systems on practically all major commercial airliners put in service in the '90s. The systems we designed and built are safety of flight critical systems. Those systems are designed, verified, built, and tested to literally the highest standards known to engineering, every bit as stringent as anything you'd find inside a nuclear reactor. Yet STILL we occasionally ran into scenarios with equipment in use in the field where the function of the system was compromised by failure scenarios which simply could not be fully anticipated. The cost of those systems was enormous. No retail grade product could ever conceivably achieve even that level of reliability and still be worth the cost.
So really, what you're effectively saying is 'if we can't have a perfect video card, we should have none at all'.
I agree with you, but 'Science History' or 'Whatever History' is not a science class. I have no use for ID in the science classroom, but the history of human thought certainly is a legitimate and IMHO valuable thing to teach.
Part of the whole problem I think is that the fundies have been ramming ID etc down our throats for years now and at this point the rest of society just wants to be done with the whole thing and are REALLY suspicious of ANY material relating to religion in schools.
In the long run the fundies are going to find out, I suspect, that their tactics have done more damage to their own views than anything else.