I think this is a huge issue, it undermines any effort that security could do to prevent another plane going into a building. The previous terrorists trained to fly 747s before smacking them into the towers for years. It wouldn't take them much more effort to obtain their license and get a job flying them.
So, what, you think you can just walk in off the street with a pilot's license and 6 months of experience and get a job flying large commercial jets unsupervised ?
They're not waiting for Terminal 5, I was photographed and fingerprinted like a criminal today on my way home from a meeting in Hamburg, via Heathrow Terminal 1. I wasn't happy, why should I as a UK passport holder have my fingerprints taken? It's a police state.
I wasn't photographed and fingerprinted, but I did get to have a full-body X-ray on the way through T4 last weekend. Even after that, I STILL had to walk through the damn metal detector, etc.
That happened the second time through security. I was coming through again because the first time they searched my carry-on bag and found the cable crimper I'd brought over to help a friend with some DIY. Despite pointing out about the nastiest thing I could do with that was bruise someone's finger, I still had to go and check my carry-on, resulting in an 15-20 minutes hanging around Zurich airport waiting for it at 10:30 that night (thankfully my Qantas Club membership let me scoot through the business check-in to check my bag in only a couple of minutes).
The issue that you're missing is that the pilot could be the damn terrorist in the first place. He could announce to all the passengers that "All your live are belong to me, hahaha make your time" and because the damn door is locked all you can do is watch yourself die.
This is not "big". This is "tiny". The probability of that happening is so small it's irrelevant.
Or what happens if the pilot has a heart attack or something which if movie plots is anything to go by happens on every flight.
But I still don't think that answers GP's question: why is this happening now instead of back when IRA blew up bomb and killed people pretty much weekly?
The most obvious reason is because it's only really become technologically and economically feasible in the last decade.
There are social factors as well, to be sure, but advances in technology are a pretty major part.
I have 6 tabs open on Firefox, and I've been using it continuously since this morning, and it's only taking up 70 MB of RAM. I never understood how people got it up to 200 MB. Maybe if you leave it open for weeks at a time, then eventually it piles up, but for the most part, I've never experienced Firefox going up to 200 MB.
This is how you do it:
Home PC:
5 windows, 73 tabs, 420M memory usage.
Work PC:
8 windows, 110 tabs, 550M memory usage.
The point I'm trying to make though, is that just running vista and browsing your files with explorer requires at least 1 GB to do smoothly. Wouldn't it be nice if it required a more sane amount of memory, for doing such trivial things, so that applications that needed memory, like MMO, Office suites, browsers, and image/video editing could make use of the memory you had?
The OS gets out of the way when the memory is required by apps, as designed. Further, the extra memory is so cheap it's practically free. Personally I don't see it as important - I haven't had a PC with less than 2G of RAM for ~4 years now and hadn't had one with less than 1G for ~4 years before that.
Linus Torvald already dumped on Microsoft's kernel model a couple of months ago.
While it is indisputable that Linus is a good programmer, one has to take any of his comments on new designs with a grain of salt, given that his main claim to fame is reimplementing other people's ideas.
Linus is not an authority on OS design, especially the field of new OS design, and even more so when talking about anything that isn't like UNIX.
When rebuilding a system from the ground up for security, these issues need to be hashed out first.
Vista wasn't rebuilt "from the ground up" in any way, shape or form.
The fact that the security and driver models were changing significantly shortly before launch is a sign of bad design.
Or a policy change like, say, "64 bit Windows won't be allowed to load unsigned drivers".
Or at the very least horrible project management. If Vista was in the works for over 5 years, and it was designed properly from the start, 3rd parties should have had plenty of time (years) to conform to new models.
Amazing how some vendors managed to have working drivers, though, despite these "last minute changes". Taking that into account, I know where I'm pointing the finger at "bad design".
Oh, and Vista wasn't "in the works for over 5 years" in any technical (ie: relevant) sense. The "Vista reboot" resulted in an actual development time of 2-3 years.
So, why then is Vista so much slower then XP even with all the extra eye-candy and features turned off? Either MS can't program a decent OS which could be true, or there is some hidden thing going on such as DRM or WGA. So there are two logical choices, MS can't program so don't use Vista or MS is using DRM so don't use Vista either way, Vista is a failure of an OS and you must agree with that.
Or (c) it's not actually slower at all.
Or (d) it's doing things that have nothing to do with DRM.
As for lack of support, where else are the media companies going to go if MS says no to DRM?
"Oh damn, we'll just have to cope with the standalone appliances that 99% of people use for consuming their media."
DRM is a non-argument against Vista. Either you don't have DRM-encumbered media, and the presence of DRM is irrelevant, or you do have DRM-encumbered media and its presence is an advantage.
Actually the main complaint I have on Lepoard is that the dock does not work as well as the Vista equivalent and it takes up a huge amount of real estate.
This isn't really a Leopard-specific complaint, either. The Dock has been a UI train wreck since its first release (although various OS X releases have improved and/or worked around some of the problems).
On slashdot people like to reply to individual paragraphs with short and witty comments. And forget about the original point. The original point was that MS Windows makes rebooting the "default" fix, regardless of the problem, whereas in Linux there are usually other valid options. We're talking about technical merits here, for discussion sake (we're on slashdot after all), and I don't care what your PHB boss thinks.
I hope you're not the ignorant one here. Are you sure you understand the fundamental differences in handling libraries (DLL's/shared objects) in Windows and Linux?
Yes.
The architecture of Windows makes it extremely difficult to upgrade libraries safely without a reboot, whereas in Linux there's basically no special handling required. An "ignorant user" in Linux can simply type "apt-get upgrade" (or yum or whatever) and expect things to be done perfectly. Without a reboot. This simply isn't the case for Windows.
The ignorant end user is still going to have to restart anything using shared libraries that were updated. Since the difference for most people between doing that and rebooting is semantics, and it's unreasonable to assume the user has the knowledge to identify which programs do and don't need restarting, Windows (and OS X, for that matter) simply take the easier, more reliable option and ask the user to reboot.
Look, if Grandma uses a computer preinstalled with, say, Ubuntu, and enables some auto update program, she doesn't get prompted to reboot the computer every other week just to fix a web browser vulnerability.
Grandma turns her computer off every night anyway, so it's irrelevant.
False dichotomy. There aren't only two groups of people in the world where one wants 99.99999% uptime and the other is happy with 90% uptime. (Of course after you drink a sufficient amount of cool aid you might actually believe this). Some people are actually interested in (reasonably) maximizing the uptime of a machine, instead of having it rebooting every other day.
No-one has to reboot "every other day". Windows patches are released monthly unless it's a very significant issue.
There is no false dichotomy here. Either you need 24/7 uptime or you don't. If you do, then you architecture has to be able to handle individual machine outages. If you don't, then you simply schedule reboots in off-hours.
In short, scheduled reboots of machines should not affect your availability requirements, no matter if they're 90% or 99.999%.
The GP is saying, where in Linux it's easy to achieve, say 99% uptime, in similar situations Windows through bad design makes it (for example) 95%.
And the GP is wrong.
The time to start these applications, set them to the desired state, etc. is non-trivial. What a sane user would do is simply to leave the computer on when he's away, and expecting the computer to be in the state where he left it. If your computer HAS TO reboot every other day then this type of work flow would be severely affected. (And before you start I'm not talking about data loss since every sane user would save their work anyway. I'm talking about the TIME lost for closing all these applications, rebooting and starting them over again.)
So don't reboot when you've got lots of stuff open. Simple.
PS: What am I doing responding to a troll anyway.....:-(
WTF ? You're the guy spinning bullshit about people having to reboot "every other day" and I'm supposed to be the one trolling because I point out that there's no reason for scheduled reboots to affect availability ? GTFO.
I used windows as a kid from 10-18, i know nothing about coding, i know little about linux (to this day). As soon as i switched i found that my crashes were less sever, ctrl-alt-bkspace, hell event alt+print scrn+k isnt a system reboot.
For most people it's the equivalent thereof. The difference between "killing everything you're working on" and "rebooting" is mostly semantics.
Once i moved a a laptop which had wireless problems, i used modrpobe to 'restart' drivers, since going to uni ive had a few problems with a cisco module that i dont know how to fix, but instead of restarting as windows users have to, if i have a problem i 'restart' my networking stuff. Ive now used likes for about 2 years and only restart when my system crashes (ive managed to mess something up with power management).
The typical user's eyes would have glazed over at about 'modprobe' above. That is why they are told to reboot.
Who mentioned end users? hes talking about installing software. On linux i simply install a program and run it, under windows that wasn't what happened.
Who do you think installs the software on most of the world's computers if not the end users ?
Your conflating software actions and end user action, (by critical shared library i assume he means libqt or whatever flashes up during upgrades) the system will sort all of this out not me, the user just clicks update, then after the update the system runs the newer software. You don't even need to close the software being updated in most cases.
If something is using a shared library (or other component) that is upgraded, you most certainly should shut down that program and restart. If for no other reason than it won't use the updated library until you do.
Since, again, knowing which things need to be restarted and which don't is knowledge beyond most users, they are simply told to reboot. It's quicker, easier and more reliable.
Because a server shouldn't NEED to be rebooted to get stuff working, the only time i need to reboot my desktop is when the kernel gets upgraded, and people are working on that!
I'm not quite sure what question you think you're answering here, but the actual question was: why do you think people will treat their computers any differently to any of their other appliances ?
Theres a difference between being stupid and having an operating system that needs you to restart. I doubt a server has ever gone down because the monitor power-savig kicked in, but ive seen desktops trashed by a windows update so image servers have done the same.
And I've seen Linux systems trashed by broken RPMs. Your point ?
What assumptions? that if you might want to not reboot? that you want a stable system? If you WANT to reboot to fix a problem, go for it but you shouldn't have to.
It assumes the end users have a relatively high amount of knowledge. Hence, you get directions like "restart service X" or "reload module Y" or "restart any programs that use library Z" because it is assumed that the end users will have the knowledge necessary to do this. On Windows (and OS X for that matter) you get "reboot" because the assumption is that the end users cannot (and should not) be expected to have the necessary knowledge.
If you know what you're doing, you can perform most Windows updates without rebooting.
again why? why should you NEED to reboot?
Because it's quicker, easier and more reliable.
scheduling for patching is surely a dangerous game to play? "sorry we cant patch the vulnerable version of sever.exe until as we dont have an opening in out patching schedule till next tuseday".
Uh, no. Scheduled patching is an integral part of properly run IT infrastructure.
While kernel* / windows vulnerabilities do require a restart, why should anything else?
Why does it matter ? Nobody important cares about server uptimes. They're interested in availability.
You're pretty messed up. I wonder why you get modded up.
Because there are still somepeople on Slashdot who have actual real-world knowledge and experience.
I know most businessmen are not interested in "wtf is wrong". And I agree. They shouldn't be. That's not their business. It's the engineers and technicians to find out the cause of the problem, and to avoid having the problem again. If nobody finds out the cause of the problem, chances are that the problem will come up again and again, sometimes with increasing frequency, that your supposed 99.9999% slowly becomes 99.999% then 99.99%, etc...
Here's the problem. Since "engineers and technicians" are loathe to just get something working again without understanding what's wrong, they have a tendency to ignore the most important issue (downtime) in favour of troubleshooting.
The time to spend hours or days figuring out what went wrong is afterwards, on testing or development systems, not during people production hours on live boxes.
Throwing more machines (or money) to the problem doesn't always work, despite what they taught you in business school. If the cause is a software error, and if it can be triggered simutaneously across a couple of so-called "redundant" machines, then you're fscked anyway. Let's say a bug makes the machine fail at Monday midnight (maybe a messed up cron job?). Do you simply "DO IT", reboot and forget about the problem?
Of course you don't "forget about the problem". But when a reboot will fix the problem in a couple of minutes, vs tens of minutes to do advanced troubleshooting, the reboot is the appropriate step to take.
And this.... I have no idea how you got to this conclusion:
Because it's the truth. No-one important cares about your system uptime. All they care about is availability.
When thousands of dollars are being lost every minute that a service is unavailable, you just DON'T TOUCH THE DAMN THING FOR NO GOOD REASON. Rebooting the machine because a minor update fixes a BROWSER bug is a fscking bad reason to reboot a machine, even if you have 10 of them chugging along.
Like I said. If individual machine reboots concern you, your architecture is broken.
Yes you just reboot it (i.e. "DO IT") if you're messed up enough to rely on inherently unreliable systems under critical environments, but the discussion we're having now is not "how to respond to a failing machine on critical systems" but how to avoid these situations in the first place.
You "avoid these situations in the first place" by having a properly designed infrastructure where individual machine downtime is unimportant.
You're right. But there's no necessary conflict. The right thing to do is not "reboot the damn thing and forget about it", the right thing is to "reboot the damn thing and find out the cause of the problem afterwards".
At no point have I even *suggested* you "reboot the damn thing and forget about it". You appear to be attacking a straw man.
That's three weeks of hard core debugging, tweaking, and hair pulling.
The fact that you were able to wait *three weeks* demonstrates that the problem was, at most, insignificant.
When thousands of dollars (or more) are being lost every minute that a service is unavailable[0], you don't fuck around with idiotic philosophising about how "its UNIX, I shouldn't need to reboot for anything"[1], you just DO IT.
[0] We shall ignore here for a minute the false economy of not just investing in a properly redundant architecture where individual machine outages do not impact availability.
[1] I've been there myself and had arguments with my (at the time) boss about it. It is the difference between how geeks think and how businesspeople think. The geek is interested in figuring out wtf is wrong. The businessman is interested in whether or not his business is still operating.
Yes, it is. People who use Windows, when using Linux, are going to respond exactly the same way to problems - by rebooting.
Anyone that's ever installed software, or run "windows update" knows that rebooting is a very likely part of this process. The dependencies and non-modular approach of Windows are quite apparent. Software vendors say "just reboot" because of all the complexities and dependencies within windows.
No, they do it because it's a simple step for the ignorant end user to understand.
The same simply isn't true for Linux. Replace a critical shared library? No problem, running programs still have a hook to the old version. Any new process that starts will get the new version of the library. Why reload the whole damn OS when restarting a process will do the same thing?
Because for people who don't know that, it's easier to say reboot.
You are conflating knowledgable end users with typical end users. This is at best naive and at worst deliberately deceptive.
You're trying to tell me with a straight face that the BBS market influenced Microsoft? (Which flies in the face of what we've all experienced with Windows).
No, I'm telling you that a random individual's attitude towards rebooting is going to be vastly more influenced by their skill level ad what they're using their computer for than the OS it runs.
No, the reason people have this attitude is because it freaking works.
Exactly. Now, again, why do you think they're going to treat computers any differently ?
I've been administrating Linux machines for 13+ years. I can count on one hand the number of times a reboot solved any problem. The only class of problem this solved is a kernel bug, or the kernel crashing (usually from a hardware problem).
Not done much work with NFS then, I take it ? Or services that have long timeout periods and don't die nicely ?
I struggle to believe anyone has been using Linux for "13+ years" and can only "count on one hand the number of times a reboot solved any problem". Either you've not used Linux anything close to "13+ years" or you've not used it in a very wide range of situations.
Why would anyone reboot without a "good reason"?
The fact that you even need to ask disqualifies you from any useful input to this discussion. Fucking hell. People hit the rest button on their PCs because the monitor power-saving kicked in and for dozens of other reasons that aren't even that good.
The point is that Linux simply has less "good reasons", and requires less reboots. Linux requires FAR less reboots for "patching".
Linux also makes a lot more assumptions about its users (and "users" in this sense reaches from Grandma to software developers).
Wow. Now I know you've really drank the Microsoft kool-aid. Not everyone can afford multiple machine redundancy just to fix the endemic problems of Microsoft who advocate "Just reboot!" to fix so many problems. There's really no reason why I need to reboot just to update what's essentially some new versions of DLLs. The Microsoft architecture is essentially broken if you have to buy another damn machine for the SOLE purpose of maintaining high availability.
Yeah, like I thought. "13+ years" and 12 of those were probably using it on your home PC.
The only meaningful difference between a "reboot" and a hardware failure is the amount of warning. I'll say it again. If your business continuity is vulnerable to individual machine outages (be they from reboots or motherboards going up in smoke), then it's broken. Period. If you can't afford "multiple machine redundancy" then you don't need 24/7 uptime. If you don't need 24/7 uptime, then either scheduled machine reboots (eg: for patching) are irrelevant, or brief outages are acceptable.
Any sysadmin who thinks he can run a high-availability operation without multiple machine redundancy is incompetent. Any sysadmin who is purporting to do so, is grossly negligent. The fact that there's a hell of a lot of people whose Linux (and UNIX in general) bias puts them into these categories, does not make them any less incompetent or negligent.
Exactly. I'm desperately unhappy with the horrible uptime of my home ISP. Forget about five nines, or even three nines -- I don't even get one nine. I'd say my uptime is roughly 80%.
Really ? Your ISP is down for nearly 5 hours out of every single day ?
The origins of an OS really show through a lot of the time. Windows started out as a single user OS, so rebooting was OK because the only person you messed up was the guy sitting in front of the screen. It eventually evolved into a multi-user OS, but the "just reboot!" mentality persists to this day.
Windows NT (ie: contemporary Windows) has been a multiuser OS since it's first release.
The reason the "just reboot" mentality persists is simply becaus e99% of the time it *is* used as a single-user OS, and no-one else is impacted. This has _zero_ to do with the architecture and everything to do with the user. Linux would be (and is) treated in the same way in similar situations.
Linux/Unix on the other hand started out life as a multi-user OS. Rebooting was a big no-no, because you'd affect countless people logged in, and you'd get yelled at for ruining someones work.
UNIX actually started out as a single-user OS and the multiuser aspect was bolted on later. Linux didn't, of course, because by the time Linus banged together his UNIX rip-off, UNIX had been multiuser for quite a while.
However, again, the attitudes towards how their relevant users treat servers and workstations have about 10% to do with their architectures and 90% to do with their knowledge. DOS and OS/2 were single user, yet frequently had BBSes and similar running off them. You can be assured the people running those BBSes were far less like to have the "just reboot" mentality.
Further, the other reason most people have that attitude is because to them a computer is just another appliance. When other appliances act up, pretty much the first thing _everybody_ does is turn it off and back on again. Why on Earth would you expect them to treat a computer any differently ?
Windows administrators categorically will try rebooting the damn thing first to fix any problem (and it usually works). Linux administrators will only try this as a last resort (and it almost never works).
No. Inexperienced admins will try rebooting first, regardless of platform. Experienced admins will not. Incidentally, there are numerous classes of problems on Linux (and UNIX in general) which are more quickly and easily "fixed" with a reboot.
Anyway, at Microsoft the idea that you can somehow tweak windows just right so rebooting isn't necessary is crazy.
I can't even remember the last time I had to reboot any of my Windows machines without a good reason (eg: patching).
Finally, there's nothing wrong with rebooting _anyway_. If your service uptime requirements are affected by a single machine rebooting, your architecture is broken. All the reboot does is demonstrate that it's broken without a real problem actually occurring.
Sysadmins comparing machine uptimes is like ricers comparing spoilers.
A very illuminating Microsoft Confidential presentation from the antitrust discovery process. If you're in a hurry start with the slides at page 9. This is what he should have been asked about...
So I read through to page 45 or so and I'm wondering exactly why I'm supposed to be surprised or concerned that a business is operating just like every other business, and exactly how I would expect it to.
Or are you seriously try to suggest that these sort of discussions and attitudes don't happen within Apple, IBM, HP, Sun, et al ?
Eventually it will believed to be true. I think even the liar will start believing it.
Indeed. For example, lies like these:
Sadly many IT professionals believe Windows saves money because its an integrated platform. But ignore the reboots and being forced to buy alot more servers as Windows is not friendly with using one or 2 more apps on a single server compared to Unix.
Oh and lets not forget about the blanket licensing fees. What is the average? $12,000 per year for licensing and support per desktop? Uh yeah thats true TCO.
If it were not for Microsoft already setting the standards for Office the corporate world would have abandonded them years ago. Linux is alot cheaper and has 1/10th of the issues if only it could the VB apps and Office.
If Linux were even half as much better as people like you thought it was, business would be falling over themselves trying to save money using it.
I've just been reading about some of the win2k8 stuff, and Hyper-V. It looks like MS are tending towards high availability stuff by using multiple lightweight VMs in a cluster, rather than having fancy stuff like hot-patching the OS.
Which isn't a bad thing anyway. If your (24/7) services are affected by taking an individual machine offline (eg: for patching), then your architecture is broken.
Slashdot Mirror
I think this is a huge issue, it undermines any effort that security could do to prevent another plane going into a building. The previous terrorists trained to fly 747s before smacking them into the towers for years. It wouldn't take them much more effort to obtain their license and get a job flying them.
So, what, you think you can just walk in off the street with a pilot's license and 6 months of experience and get a job flying large commercial jets unsupervised ?
They're not waiting for Terminal 5, I was photographed and fingerprinted like a criminal today on my way home from a meeting in Hamburg, via Heathrow Terminal 1. I wasn't happy, why should I as a UK passport holder have my fingerprints taken? It's a police state.
I wasn't photographed and fingerprinted, but I did get to have a full-body X-ray on the way through T4 last weekend. Even after that, I STILL had to walk through the damn metal detector, etc.
That happened the second time through security. I was coming through again because the first time they searched my carry-on bag and found the cable crimper I'd brought over to help a friend with some DIY. Despite pointing out about the nastiest thing I could do with that was bruise someone's finger, I still had to go and check my carry-on, resulting in an 15-20 minutes hanging around Zurich airport waiting for it at 10:30 that night (thankfully my Qantas Club membership let me scoot through the business check-in to check my bag in only a couple of minutes).
Fuckers.
The issue that you're missing is that the pilot could be the damn terrorist in the first place. He could announce to all the passengers that "All your live are belong to me, hahaha make your time" and because the damn door is locked all you can do is watch yourself die.
This is not "big". This is "tiny". The probability of that happening is so small it's irrelevant.
Or what happens if the pilot has a heart attack or something which if movie plots is anything to go by happens on every flight.
That's why you have a co-pilot.
But I still don't think that answers GP's question: why is this happening now instead of back when IRA blew up bomb and killed people pretty much weekly?
The most obvious reason is because it's only really become technologically and economically feasible in the last decade.
There are social factors as well, to be sure, but advances in technology are a pretty major part.
I have 6 tabs open on Firefox, and I've been using it continuously since this morning, and it's only taking up 70 MB of RAM. I never understood how people got it up to 200 MB. Maybe if you leave it open for weeks at a time, then eventually it piles up, but for the most part, I've never experienced Firefox going up to 200 MB.
This is how you do it:
Home PC:
5 windows, 73 tabs, 420M memory usage.
Work PC:
8 windows, 110 tabs, 550M memory usage.
The point I'm trying to make though, is that just running vista and browsing your files with explorer requires at least 1 GB to do smoothly. Wouldn't it be nice if it required a more sane amount of memory, for doing such trivial things, so that applications that needed memory, like MMO, Office suites, browsers, and image/video editing could make use of the memory you had?
The OS gets out of the way when the memory is required by apps, as designed. Further, the extra memory is so cheap it's practically free. Personally I don't see it as important - I haven't had a PC with less than 2G of RAM for ~4 years now and hadn't had one with less than 1G for ~4 years before that.
Linus Torvald already dumped on Microsoft's kernel model a couple of months ago.
While it is indisputable that Linus is a good programmer, one has to take any of his comments on new designs with a grain of salt, given that his main claim to fame is reimplementing other people's ideas.
Linus is not an authority on OS design, especially the field of new OS design, and even more so when talking about anything that isn't like UNIX.
When rebuilding a system from the ground up for security, these issues need to be hashed out first.
Vista wasn't rebuilt "from the ground up" in any way, shape or form.
The fact that the security and driver models were changing significantly shortly before launch is a sign of bad design.
Or a policy change like, say, "64 bit Windows won't be allowed to load unsigned drivers".
Or at the very least horrible project management. If Vista was in the works for over 5 years, and it was designed properly from the start, 3rd parties should have had plenty of time (years) to conform to new models.
Amazing how some vendors managed to have working drivers, though, despite these "last minute changes". Taking that into account, I know where I'm pointing the finger at "bad design".
Oh, and Vista wasn't "in the works for over 5 years" in any technical (ie: relevant) sense. The "Vista reboot" resulted in an actual development time of 2-3 years.
So, why then is Vista so much slower then XP even with all the extra eye-candy and features turned off? Either MS can't program a decent OS which could be true, or there is some hidden thing going on such as DRM or WGA. So there are two logical choices, MS can't program so don't use Vista or MS is using DRM so don't use Vista either way, Vista is a failure of an OS and you must agree with that.
Or (c) it's not actually slower at all.
Or (d) it's doing things that have nothing to do with DRM.
As for lack of support, where else are the media companies going to go if MS says no to DRM?
"Oh damn, we'll just have to cope with the standalone appliances that 99% of people use for consuming their media."
DRM is a non-argument against Vista. Either you don't have DRM-encumbered media, and the presence of DRM is irrelevant, or you do have DRM-encumbered media and its presence is an advantage.
Actually the main complaint I have on Lepoard is that the dock does not work as well as the Vista equivalent and it takes up a huge amount of real estate.
This isn't really a Leopard-specific complaint, either. The Dock has been a UI train wreck since its first release (although various OS X releases have improved and/or worked around some of the problems).
Why doesn't it have fast, easy keyboard shortcuts for most tasks?
Uh, what ? Windows is one of the *best* out there for keyboard access. It's _leagues_ ahead of any version of MacOS ever released.
This is not true.
Yes, it is.
On slashdot people like to reply to individual paragraphs with short and witty comments. And forget about the original point. The original point was that MS Windows makes rebooting the "default" fix, regardless of the problem, whereas in Linux there are usually other valid options. We're talking about technical merits here, for discussion sake (we're on slashdot after all), and I don't care what your PHB boss thinks.
And that "original point" was wrong.
I hope you're not the ignorant one here. Are you sure you understand the fundamental differences in handling libraries (DLL's/shared objects) in Windows and Linux?
Yes.
The architecture of Windows makes it extremely difficult to upgrade libraries safely without a reboot, whereas in Linux there's basically no special handling required. An "ignorant user" in Linux can simply type "apt-get upgrade" (or yum or whatever) and expect things to be done perfectly. Without a reboot. This simply isn't the case for Windows.
The ignorant end user is still going to have to restart anything using shared libraries that were updated. Since the difference for most people between doing that and rebooting is semantics, and it's unreasonable to assume the user has the knowledge to identify which programs do and don't need restarting, Windows (and OS X, for that matter) simply take the easier, more reliable option and ask the user to reboot.
Look, if Grandma uses a computer preinstalled with, say, Ubuntu, and enables some auto update program, she doesn't get prompted to reboot the computer every other week just to fix a web browser vulnerability.
Grandma turns her computer off every night anyway, so it's irrelevant.
False dichotomy. There aren't only two groups of people in the world where one wants 99.99999% uptime and the other is happy with 90% uptime. (Of course after you drink a sufficient amount of cool aid you might actually believe this). Some people are actually interested in (reasonably) maximizing the uptime of a machine, instead of having it rebooting every other day.
No-one has to reboot "every other day". Windows patches are released monthly unless it's a very significant issue.
There is no false dichotomy here. Either you need 24/7 uptime or you don't. If you do, then you architecture has to be able to handle individual machine outages. If you don't, then you simply schedule reboots in off-hours.
In short, scheduled reboots of machines should not affect your availability requirements, no matter if they're 90% or 99.999%.
The GP is saying, where in Linux it's easy to achieve, say 99% uptime, in similar situations Windows through bad design makes it (for example) 95%.
And the GP is wrong.
The time to start these applications, set them to the desired state, etc. is non-trivial. What a sane user would do is simply to leave the computer on when he's away, and expecting the computer to be in the state where he left it. If your computer HAS TO reboot every other day then this type of work flow would be severely affected. (And before you start I'm not talking about data loss since every sane user would save their work anyway. I'm talking about the TIME lost for closing all these applications, rebooting and starting them over again.)
So don't reboot when you've got lots of stuff open. Simple.
PS: What am I doing responding to a troll anyway..... :-(
WTF ? You're the guy spinning bullshit about people having to reboot "every other day" and I'm supposed to be the one trolling because I point out that there's no reason for scheduled reboots to affect availability ? GTFO.
I used windows as a kid from 10-18, i know nothing about coding, i know little about linux (to this day). As soon as i switched i found that my crashes were less sever, ctrl-alt-bkspace, hell event alt+print scrn+k isnt a system reboot.
For most people it's the equivalent thereof. The difference between "killing everything you're working on" and "rebooting" is mostly semantics.
Once i moved a a laptop which had wireless problems, i used modrpobe to 'restart' drivers, since going to uni ive had a few problems with a cisco module that i dont know how to fix, but instead of restarting as windows users have to, if i have a problem i 'restart' my networking stuff. Ive now used likes for about 2 years and only restart when my system crashes (ive managed to mess something up with power management).
The typical user's eyes would have glazed over at about 'modprobe' above. That is why they are told to reboot.
Who mentioned end users? hes talking about installing software. On linux i simply install a program and run it, under windows that wasn't what happened.
Who do you think installs the software on most of the world's computers if not the end users ?
Your conflating software actions and end user action, (by critical shared library i assume he means libqt or whatever flashes up during upgrades) the system will sort all of this out not me, the user just clicks update, then after the update the system runs the newer software. You don't even need to close the software being updated in most cases.
If something is using a shared library (or other component) that is upgraded, you most certainly should shut down that program and restart. If for no other reason than it won't use the updated library until you do.
Since, again, knowing which things need to be restarted and which don't is knowledge beyond most users, they are simply told to reboot. It's quicker, easier and more reliable.
Because a server shouldn't NEED to be rebooted to get stuff working, the only time i need to reboot my desktop is when the kernel gets upgraded, and people are working on that!
I'm not quite sure what question you think you're answering here, but the actual question was: why do you think people will treat their computers any differently to any of their other appliances ?
Theres a difference between being stupid and having an operating system that needs you to restart. I doubt a server has ever gone down because the monitor power-savig kicked in, but ive seen desktops trashed by a windows update so image servers have done the same.
And I've seen Linux systems trashed by broken RPMs. Your point ?
What assumptions? that if you might want to not reboot? that you want a stable system? If you WANT to reboot to fix a problem, go for it but you shouldn't have to.
It assumes the end users have a relatively high amount of knowledge. Hence, you get directions like "restart service X" or "reload module Y" or "restart any programs that use library Z" because it is assumed that the end users will have the knowledge necessary to do this. On Windows (and OS X for that matter) you get "reboot" because the assumption is that the end users cannot (and should not) be expected to have the necessary knowledge.
If you know what you're doing, you can perform most Windows updates without rebooting.
again why? why should you NEED to reboot?
Because it's quicker, easier and more reliable.
scheduling for patching is surely a dangerous game to play? "sorry we cant patch the vulnerable version of sever.exe until as we dont have an opening in out patching schedule till next tuseday".
Uh, no. Scheduled patching is an integral part of properly run IT infrastructure.
While kernel* / windows vulnerabilities do require a restart, why should anything else?
Why does it matter ? Nobody important cares about server uptimes. They're interested in availability.
You're pretty messed up. I wonder why you get modded up.
Because there are still some people on Slashdot who have actual real-world knowledge and experience.
I know most businessmen are not interested in "wtf is wrong". And I agree. They shouldn't be. That's not their business. It's the engineers and technicians to find out the cause of the problem, and to avoid having the problem again. If nobody finds out the cause of the problem, chances are that the problem will come up again and again, sometimes with increasing frequency, that your supposed 99.9999% slowly becomes 99.999% then 99.99%, etc...
Here's the problem. Since "engineers and technicians" are loathe to just get something working again without understanding what's wrong, they have a tendency to ignore the most important issue (downtime) in favour of troubleshooting.
The time to spend hours or days figuring out what went wrong is afterwards, on testing or development systems, not during people production hours on live boxes.
Throwing more machines (or money) to the problem doesn't always work, despite what they taught you in business school. If the cause is a software error, and if it can be triggered simutaneously across a couple of so-called "redundant" machines, then you're fscked anyway. Let's say a bug makes the machine fail at Monday midnight (maybe a messed up cron job?). Do you simply "DO IT", reboot and forget about the problem?
Of course you don't "forget about the problem". But when a reboot will fix the problem in a couple of minutes, vs tens of minutes to do advanced troubleshooting, the reboot is the appropriate step to take.
And this.... I have no idea how you got to this conclusion:
Because it's the truth. No-one important cares about your system uptime. All they care about is availability.
When thousands of dollars are being lost every minute that a service is unavailable, you just DON'T TOUCH THE DAMN THING FOR NO GOOD REASON. Rebooting the machine because a minor update fixes a BROWSER bug is a fscking bad reason to reboot a machine, even if you have 10 of them chugging along.
Like I said. If individual machine reboots concern you, your architecture is broken.
Yes you just reboot it (i.e. "DO IT") if you're messed up enough to rely on inherently unreliable systems under critical environments, but the discussion we're having now is not "how to respond to a failing machine on critical systems" but how to avoid these situations in the first place.
You "avoid these situations in the first place" by having a properly designed infrastructure where individual machine downtime is unimportant.
You're right. But there's no necessary conflict. The right thing to do is not "reboot the damn thing and forget about it", the right thing is to "reboot the damn thing and find out the cause of the problem afterwards".
At no point have I even *suggested* you "reboot the damn thing and forget about it". You appear to be attacking a straw man.
That's three weeks of hard core debugging, tweaking, and hair pulling.
The fact that you were able to wait *three weeks* demonstrates that the problem was, at most, insignificant.
When thousands of dollars (or more) are being lost every minute that a service is unavailable[0], you don't fuck around with idiotic philosophising about how "its UNIX, I shouldn't need to reboot for anything"[1], you just DO IT.
[0] We shall ignore here for a minute the false economy of not just investing in a properly redundant architecture where individual machine outages do not impact availability.
[1] I've been there myself and had arguments with my (at the time) boss about it. It is the difference between how geeks think and how businesspeople think. The geek is interested in figuring out wtf is wrong. The businessman is interested in whether or not his business is still operating.
This is simply not true.
Yes, it is. People who use Windows, when using Linux, are going to respond exactly the same way to problems - by rebooting.
Anyone that's ever installed software, or run "windows update" knows that rebooting is a very likely part of this process. The dependencies and non-modular approach of Windows are quite apparent. Software vendors say "just reboot" because of all the complexities and dependencies within windows.
No, they do it because it's a simple step for the ignorant end user to understand.
The same simply isn't true for Linux. Replace a critical shared library? No problem, running programs still have a hook to the old version. Any new process that starts will get the new version of the library. Why reload the whole damn OS when restarting a process will do the same thing?
Because for people who don't know that, it's easier to say reboot.
You are conflating knowledgable end users with typical end users. This is at best naive and at worst deliberately deceptive.
You're trying to tell me with a straight face that the BBS market influenced Microsoft? (Which flies in the face of what we've all experienced with Windows).
No, I'm telling you that a random individual's attitude towards rebooting is going to be vastly more influenced by their skill level ad what they're using their computer for than the OS it runs.
No, the reason people have this attitude is because it freaking works.
Exactly. Now, again, why do you think they're going to treat computers any differently ?
I've been administrating Linux machines for 13+ years. I can count on one hand the number of times a reboot solved any problem. The only class of problem this solved is a kernel bug, or the kernel crashing (usually from a hardware problem).
Not done much work with NFS then, I take it ? Or services that have long timeout periods and don't die nicely ?
I struggle to believe anyone has been using Linux for "13+ years" and can only "count on one hand the number of times a reboot solved any problem". Either you've not used Linux anything close to "13+ years" or you've not used it in a very wide range of situations.
Why would anyone reboot without a "good reason"?
The fact that you even need to ask disqualifies you from any useful input to this discussion. Fucking hell. People hit the rest button on their PCs because the monitor power-saving kicked in and for dozens of other reasons that aren't even that good.
The point is that Linux simply has less "good reasons", and requires less reboots. Linux requires FAR less reboots for "patching".
Linux also makes a lot more assumptions about its users (and "users" in this sense reaches from Grandma to software developers).
Wow. Now I know you've really drank the Microsoft kool-aid. Not everyone can afford multiple machine redundancy just to fix the endemic problems of Microsoft who advocate "Just reboot!" to fix so many problems. There's really no reason why I need to reboot just to update what's essentially some new versions of DLLs. The Microsoft architecture is essentially broken if you have to buy another damn machine for the SOLE purpose of maintaining high availability.
Yeah, like I thought. "13+ years" and 12 of those were probably using it on your home PC.
The only meaningful difference between a "reboot" and a hardware failure is the amount of warning. I'll say it again. If your business continuity is vulnerable to individual machine outages (be they from reboots or motherboards going up in smoke), then it's broken. Period. If you can't afford "multiple machine redundancy" then you don't need 24/7 uptime. If you don't need 24/7 uptime, then either scheduled machine reboots (eg: for patching) are irrelevant, or brief outages are acceptable.
Any sysadmin who thinks he can run a high-availability operation without multiple machine redundancy is incompetent. Any sysadmin who is purporting to do so, is grossly negligent. The fact that there's a hell of a lot of people whose Linux (and UNIX in general) bias puts them into these categories, does not make them any less incompetent or negligent.
Exactly. I'm desperately unhappy with the horrible uptime of my home ISP. Forget about five nines, or even three nines -- I don't even get one nine. I'd say my uptime is roughly 80%.
Really ? Your ISP is down for nearly 5 hours out of every single day ?
The origins of an OS really show through a lot of the time. Windows started out as a single user OS, so rebooting was OK because the only person you messed up was the guy sitting in front of the screen. It eventually evolved into a multi-user OS, but the "just reboot!" mentality persists to this day.
Windows NT (ie: contemporary Windows) has been a multiuser OS since it's first release.
The reason the "just reboot" mentality persists is simply becaus e99% of the time it *is* used as a single-user OS, and no-one else is impacted. This has _zero_ to do with the architecture and everything to do with the user. Linux would be (and is) treated in the same way in similar situations.
Linux/Unix on the other hand started out life as a multi-user OS. Rebooting was a big no-no, because you'd affect countless people logged in, and you'd get yelled at for ruining someones work.
UNIX actually started out as a single-user OS and the multiuser aspect was bolted on later. Linux didn't, of course, because by the time Linus banged together his UNIX rip-off, UNIX had been multiuser for quite a while.
However, again, the attitudes towards how their relevant users treat servers and workstations have about 10% to do with their architectures and 90% to do with their knowledge. DOS and OS/2 were single user, yet frequently had BBSes and similar running off them. You can be assured the people running those BBSes were far less like to have the "just reboot" mentality.
Further, the other reason most people have that attitude is because to them a computer is just another appliance. When other appliances act up, pretty much the first thing _everybody_ does is turn it off and back on again. Why on Earth would you expect them to treat a computer any differently ?
Windows administrators categorically will try rebooting the damn thing first to fix any problem (and it usually works). Linux administrators will only try this as a last resort (and it almost never works).
No. Inexperienced admins will try rebooting first, regardless of platform. Experienced admins will not. Incidentally, there are numerous classes of problems on Linux (and UNIX in general) which are more quickly and easily "fixed" with a reboot.
Anyway, at Microsoft the idea that you can somehow tweak windows just right so rebooting isn't necessary is crazy.
I can't even remember the last time I had to reboot any of my Windows machines without a good reason (eg: patching).
Finally, there's nothing wrong with rebooting _anyway_. If your service uptime requirements are affected by a single machine rebooting, your architecture is broken. All the reboot does is demonstrate that it's broken without a real problem actually occurring.
Sysadmins comparing machine uptimes is like ricers comparing spoilers.
I admire their stand against premarital sex [...]
Why ?
Well, I would say you were really uninformed if you bought a low-end PC to run full Vista on. Everyone knew it would require a badass PC to run on.
No, it didn't. At release Vista ran fine on sub-US$1k machines.
The only "badass" PC Vista requires, is one that was "badass" about 5-6 years ago.
A very illuminating Microsoft Confidential presentation from the antitrust discovery process. If you're in a hurry start with the slides at page 9. This is what he should have been asked about...
So I read through to page 45 or so and I'm wondering exactly why I'm supposed to be surprised or concerned that a business is operating just like every other business, and exactly how I would expect it to.
Or are you seriously try to suggest that these sort of discussions and attitudes don't happen within Apple, IBM, HP, Sun, et al ?
The big question is: Why does anybody actually believe MS Windows is a genuine server platform?
Because not everyone is as stupid and ignorant as you are.
Eventually it will believed to be true. I think even the liar will start believing it.
Indeed. For example, lies like these:
If it were not for Microsoft already setting the standards for Office the corporate world would have abandonded them years ago. Linux is alot cheaper and has 1/10th of the issues if only it could the VB apps and Office.
If Linux were even half as much better as people like you thought it was, business would be falling over themselves trying to save money using it.
I've just been reading about some of the win2k8 stuff, and Hyper-V. It looks like MS are tending towards high availability stuff by using multiple lightweight VMs in a cluster, rather than having fancy stuff like hot-patching the OS.
Which isn't a bad thing anyway. If your (24/7) services are affected by taking an individual machine offline (eg: for patching), then your architecture is broken.