Re:Recycling processes is normal for windows
on
Debugging Microsoft.com
·
· Score: 5, Interesting
IIS just can't hold up by itself, if you just serve static pages you are ok, but when people starts using that asp + odbc shit, you have to restart IIS every 5 fucking minutes.
That's not because of IIS; it's because of the people writing the ASP apps and stupid admins not configuring IIS correctly. If you have stupid people writing applications, those applications have a tendency of doing stupid things. Combine that with admins who don't properly isolate that applications running on IIS and you've got a recipe for requiring an IIS restart "every 5 fucking minutes".
Give me 5 minutes and I can write a nice app that takes down Apache no problem. A few infinite loops, perhaps each creating a dozen new database connections and allocating a massive string buffer in memory.
IIS 6.0 has a lot of features built into it that allow for admins to configure application pools to more effectively isolate applications. You can configure those application pools to recycle automatically given certain criteria (like memory usage, CPU usage, # req/sec, @ req/total, etc.), and the pools are isolated from each other so that if one dies due to a misbehaving application, the other applications on the system are not affected.
We used to receive a stupid "too many conections" from ODBC in our log, and restarting the stupid services woudln't do a damn thing, all you could do was restart the machine, Yes, restart a SERVER.
Perhaps that's all you could do, but somebody who spent more than 10 minutes reading about administering IIS would know to recycle the ODBC COM+ application to clear out the connection pool. Then they would find the stupid people writing that crappy applications and fire them, or at least isolate their applications in a separate app pool or worker process. (Dllhost.exe.)
Spare me the anecdotal stories of your LAMP solutions doing so much better than your Windows solutions. You have absolutely no credibility given your complete ignorance.
Got any links? I suspect that has more to do with kernel mode listeners (which at the time Tux had and IIS 5.0 didn't) than the TCP stack, and since IIS 6.0 has a kernel mode HTTP listener, it's probably not an issue anymore.
Regardless, that has little to do with the problem Microsoft encountered in connecting two datacenters that where phsyiscally seperated by a long distance but connected with a high bandwidth pipe. See this research paper.
If the new tcp stack retains compatability with the old one, then how are the innovations in the way it interacts with the network? For it to be compatible it has to "interact" with the network like any other stack, so it can't gain any leverage in that level, can it? Or are you saying that the new stack is only advantegeous if the machines are all vistas? Just curious how that works out.
No, the most benefit certainly comes when both ends are using CTCP (Compound TCP... the Vista TCP stack), but some benefit is still seen when only one end is using the new stack. This is possible because of some of the built in flexibilty of the TCP design. Just because the standard implementation sends packets and waits for confirmation in a certain way (or, more specifically, for a certain amount of time) doesn't mean that a new implementation has to as long as it follows the same basic rules when doing that sending.
Opps... I posted the wrong link. Somebody later on in the comments posted the correct link to the document that describes the new Microsoft TCP stack called Compound TCP (CTCP).
That is absolutely not what Microsoft has done. You honestly thing the reason they were only getting 10 Mb/s between their datacenters was due to memory copying? And when I said "implementation" I wasn't talking about the internal code, I was talking about the way it interacts across the wire.
Microsoft has completely revamped the way that TCP handles packet transmission.
The limitations discussed in the video of the Windows TCP stack are not limited to Windows. These are limitations imposed by a to-the-spec implementation of TCP. TCP is 30+ years old, and it wasn't designed for the kinds of networks it runs on today.
The new TCP stack in Vista effectively implements TCP is such a way that it removes these limitations while preserving compatibility with old stack implementations.
About 50% of companies use Exchange/Outlook (or OWA) for their e-mail. In order for Linux desktops to become more widespread they need to be able to seamlessly integrate with Exchange. Period.
Ironically, the next version of OWA will be so good that the Outlook rich-client will become more or less optional. As long as your Linux machine has a browser capable of displaying OWA, you've solved your e-mail problems.
The current version of OWA already has decent support for non-IE browsers, and they're apparently going to improve that a great deal in the next release.
I guess "captialism" is another word for "government enforced monopolies"... on bandwidth, on protocols and interfaces (though software and other process patents), and in some cases subsidies and enforced franchises.
Government enforced monopolies? How many monopolies can a particular market have? Because from what I see, there is Verizon, Cingular, AT&T, T-Mobile, Sprint/Nextel, Dobson, Bell Mobility, and at least a dozen other smaller companies competing in the cellphone service space.
True, these companies all lease ranges of the spectrum from the government, but how does that equate to a government "enforced" monopoly?
Last I checked, nobody forces Verizon to use CDMA and Cingular to use GSM. And what subsidies / enforced franchises are you talking about?
The other is the cellphone carriers, or, as I like to call them, "the Soviet ministries," which too often treat their customers as captive and refuse to allow open competition for services they offer over their networks.'
Does anybody else see the irony in this statement? The cellphone companies have built their networks through the process of captialism, and now many, in the name of capitalism, advocate the forceful opening of these networks so that 3rd parties can take advantage of the work done by the cellphone companies.
The concept of a "Type Manager" is essentially what Microsoft is going for with the Vista + WinFS combination.
WinFS will allow for developers to store easily accessible meta-data about any kind of file they want. They can publish these schemas (all XML), and allow other developers to utilize this meta-data in their own applications.
Vista's explorer allows developers to extend the display of a file or folder to include this meta-data, as well as implement their own UI components for displaying non-textual meta-data, such as album art or video icons or whatever.
The key thing that Microsoft is doing that others don't seem to be (perhaps because they aren't in a position to mandate it) is that the extensibility should not come at the cost of UI consistantcy. Yes, it's awesome to be able to open a folder full of music on your computer and have that inteface be taylored to dealing with music, but it shouldn't be so different that common tasks that apply to all file types aren't possbile or are convoluted.
While these videos are somewhat outdated, they show the direction Microsoft is trying to take things and it's quite innovative:
Actually, that would be looking at quantum physics in an extremely wrong light.
Unless the whales existing in a completely isolated environment, even isolated from all forms of energy from the outside environment, they very much do exist regardless of whether or not somebody is "looking" at them.
Last time I checked, most whales live in the ocean where there are plenty of things "looking" at them all the time. (Engery from a wide variety of sources, water, etc.)
The cat-in-a-box experiement, which you are undoubtedly referencing, is a *thought* experiement. It cannot be taken literally.
I've never, ever seen a blue whale in person. Never! In fact, I don't know anybody who has! Sure, I've seen pictures... but those could be fake.
My personal experiences and the hearsay of my friends are the only accurate way to measure the world around me. If I personally don't know somebody who has seen a blue whale, they can't possibly exist.
Most people buy iPods because they're sleek and pretty and well marketed and do what most people want them to do. These are "typical consumers". Basically none of these people care if the products they use are open source or not. Most don't even know what open source is.
Sure, there is a small segment of people who love to tweak and/or love the idea of open source, but those people don't matter. Why? Because they are more or less a statistical anomaly as far as sales figures go.
That's right people! You don't matter! Your passion for open source everything makes little or no difference, especially in the target market for iPods.
If anybody ever beats the iPod it will be because of better marketing, or maybe even better technology (as long as it's combined with better marketing)... being open source will never, ever make any damn difference.
If you're going to make claims like that the burden is on you to provide proof.
Provide evidence that Microsoft has a strategy to leave bugs in Windows so it can sell anti-malware software. Provide proof that McAfee writes viruses so it can sell software to fix them.
Microsoft offering anti-virus or anti-malware for Windows does not mean that they will stop fixing bugs in Windows.
No more than the fact that McAfee or Symantec offers antivirus software means they active release viruses to spurn the adoption of their software.
Microsoft is being pro-active about security by trying to get software into Windows that will stop undiscovered bugs from making systems expoitable. This will make users safer in the long run, and eventually (probably) will be included in every copy of Windows.
It is not the record labels that make ringtones expensive. Typically around 50% of the cost of the ringtone goes directly to the phone carrier. Some charge as high as 60%.
Most ringtones fees are billed directly to a user's cellphone bill. This dramatically increases sales because people buy more impulsively. To have this privilege, companies that sell ringtones must give a MAJORITY of the revenue from the sale directly to the cellphone company itself.
Of the remaining dollar, about 20 cents goes to aggregators that provide SMS/PSMS (premium sms... billing messages) integration (the guys who let you send text messages to cellphone networks.)
Of the 80 cents that remains after that, 20 cents will usually go to the content providers... or, for the lucky ones who have the resources to create and managing the licensing of their own ringtones, they get to keep that 20 cents.
Around 30 cents of the remaining 60 to 80 cents go to the record labels.
In the end, the people who actually run the ringtone site get between 30 and 50 cents per ringtone sold. Minus advertising. (Which is almost always a LOT more than 30 to 50 cents per sale.)
That is why almost all ringtone sites sell subscriptions, not single ringtones. They're hoping you don't use all your ringtones that come with your monthly subscription, and that you continue your subscription for more than one month. Otherwise, it's just not profitable.
But the point is that record labels only get between 10% and 15% of the cost of a ringtone. Of that 10% - 15%, the artists get some portion. (Usually 2% to 5%.) It's the carriers that take the VAST majority of the money.
If you want ringtones to be cheaper, bitch to the cellphone companies not the record labels.
You have the right to choose whether or not to purchase the CD. Period. End of story. That is where your "rights" end.
You do not have the right to force a company to allow you to play the CD here or there or anywhere. In fact, the company who sells you the CD could make it so the CD doesn't play anywhere at all, and you would have no right to force them to do otherwise.
Of course not too many people will buy a CD that can't play at all. The question is how far can a company go in restricting the use of the CD until enough people stop purchasing them that it makes the measures financially detrimental.
If the company does not make it clear that the CD won't play on a computer then there might be some issue with misrepresentation of the product. All the company has to do, however, is put a label on the CD stating that there is DRM that prevents the playback of the CD on a PC. In fact, I've seen several CDs with exactly this label.
At any rate, you have the right to buy or not to buy. That's it. Any other "rights" are figments of your imagination.
Once you start handling millions of requests per day, bandwidth is your single largest reoccuring cost. The application that I build, and that runs on two Win2k3 machines (each of which cost about $4000, including the cost of Windows), costs about $50,000 a month in bandwidth costs.
The cost of the license for Windows is almost never a big factor when you're talking about large scale applications. Small apps, like home-brew ones, are affected far more by these costs.
The browsercaps file isn't updated enough for my taste, but it uses regular expressions and matches on general things like browser type (Openwave, etc.) so that it renders as best it can when it doesn't have an exact match.
You can easily update this yourself, and there is a device profiler that allows you to easily add support for additional phones.
Can you suggest a better way to do this? Mobile devices follow basically no standards and each have their own quirks. How do you suggest targeting these devices without a browsercaps file or database?
Slashdot Mirror
Sorry, that chimney article was the wrong link. I really need to read things more carefully before hitting post.
Read this instead.
Read this. It should clear things up for you.
Give me a break. Microsoft uses Akamai to do DNS load balancing.
100% of Microsoft.com runs on Windows.
IIS just can't hold up by itself, if you just serve static pages you are ok, but when people starts using that asp + odbc shit, you have to restart IIS every 5 fucking minutes.
That's not because of IIS; it's because of the people writing the ASP apps and stupid admins not configuring IIS correctly. If you have stupid people writing applications, those applications have a tendency of doing stupid things. Combine that with admins who don't properly isolate that applications running on IIS and you've got a recipe for requiring an IIS restart "every 5 fucking minutes".
Give me 5 minutes and I can write a nice app that takes down Apache no problem. A few infinite loops, perhaps each creating a dozen new database connections and allocating a massive string buffer in memory.
IIS 6.0 has a lot of features built into it that allow for admins to configure application pools to more effectively isolate applications. You can configure those application pools to recycle automatically given certain criteria (like memory usage, CPU usage, # req/sec, @ req/total, etc.), and the pools are isolated from each other so that if one dies due to a misbehaving application, the other applications on the system are not affected.
We used to receive a stupid "too many conections" from ODBC in our log, and restarting the stupid services woudln't do a damn thing, all you could do was restart the machine, Yes, restart a SERVER.
Perhaps that's all you could do, but somebody who spent more than 10 minutes reading about administering IIS would know to recycle the ODBC COM+ application to clear out the connection pool. Then they would find the stupid people writing that crappy applications and fire them, or at least isolate their applications in a separate app pool or worker process. (Dllhost.exe.)
Spare me the anecdotal stories of your LAMP solutions doing so much better than your Windows solutions. You have absolutely no credibility given your complete ignorance.
Got any links? I suspect that has more to do with kernel mode listeners (which at the time Tux had and IIS 5.0 didn't) than the TCP stack, and since IIS 6.0 has a kernel mode HTTP listener, it's probably not an issue anymore.
Regardless, that has little to do with the problem Microsoft encountered in connecting two datacenters that where phsyiscally seperated by a long distance but connected with a high bandwidth pipe. See this research paper.
Ya, sorry about that. I posted the wrong link. This is the correct one.
If the new tcp stack retains compatability with the old one, then how are the innovations in the way it interacts with the network? For it to be compatible it has to "interact" with the network like any other stack, so it can't gain any leverage in that level, can it? Or are you saying that the new stack is only advantegeous if the machines are all vistas? Just curious how that works out.
No, the most benefit certainly comes when both ends are using CTCP (Compound TCP... the Vista TCP stack), but some benefit is still seen when only one end is using the new stack. This is possible because of some of the built in flexibilty of the TCP design. Just because the standard implementation sends packets and waits for confirmation in a certain way (or, more specifically, for a certain amount of time) doesn't mean that a new implementation has to as long as it follows the same basic rules when doing that sending.
Read the research paper for a complete description.
Opps... I posted the wrong link. Somebody later on in the comments posted the correct link to the document that describes the new Microsoft TCP stack called Compound TCP (CTCP).
That is absolutely not what Microsoft has done. You honestly thing the reason they were only getting 10 Mb/s between their datacenters was due to memory copying? And when I said "implementation" I wasn't talking about the internal code, I was talking about the way it interacts across the wire.
Microsoft has completely revamped the way that TCP handles packet transmission.
Nice try with the ignorant MS bashing though.
The limitations discussed in the video of the Windows TCP stack are not limited to Windows. These are limitations imposed by a to-the-spec implementation of TCP. TCP is 30+ years old, and it wasn't designed for the kinds of networks it runs on today.
The new TCP stack in Vista effectively implements TCP is such a way that it removes these limitations while preserving compatibility with old stack implementations.
About 50% of companies use Exchange/Outlook (or OWA) for their e-mail. In order for Linux desktops to become more widespread they need to be able to seamlessly integrate with Exchange. Period.
Ironically, the next version of OWA will be so good that the Outlook rich-client will become more or less optional. As long as your Linux machine has a browser capable of displaying OWA, you've solved your e-mail problems.
The current version of OWA already has decent support for non-IE browsers, and they're apparently going to improve that a great deal in the next release.
See: Exchange 12 Channel 9 Video
I guess "captialism" is another word for "government enforced monopolies"... on bandwidth, on protocols and interfaces (though software and other process patents), and in some cases subsidies and enforced franchises.
Government enforced monopolies? How many monopolies can a particular market have? Because from what I see, there is Verizon, Cingular, AT&T, T-Mobile, Sprint/Nextel, Dobson, Bell Mobility, and at least a dozen other smaller companies competing in the cellphone service space.
True, these companies all lease ranges of the spectrum from the government, but how does that equate to a government "enforced" monopoly?
Last I checked, nobody forces Verizon to use CDMA and Cingular to use GSM. And what subsidies / enforced franchises are you talking about?
The other is the cellphone carriers, or, as I like to call them, "the Soviet ministries," which too often treat their customers as captive and refuse to allow open competition for services they offer over their networks.'
Does anybody else see the irony in this statement? The cellphone companies have built their networks through the process of captialism, and now many, in the name of capitalism, advocate the forceful opening of these networks so that 3rd parties can take advantage of the work done by the cellphone companies.
The concept of a "Type Manager" is essentially what Microsoft is going for with the Vista + WinFS combination.
WinFS will allow for developers to store easily accessible meta-data about any kind of file they want. They can publish these schemas (all XML), and allow other developers to utilize this meta-data in their own applications.
Vista's explorer allows developers to extend the display of a file or folder to include this meta-data, as well as implement their own UI components for displaying non-textual meta-data, such as album art or video icons or whatever.
The key thing that Microsoft is doing that others don't seem to be (perhaps because they aren't in a position to mandate it) is that the extensibility should not come at the cost of UI consistantcy. Yes, it's awesome to be able to open a folder full of music on your computer and have that inteface be taylored to dealing with music, but it shouldn't be so different that common tasks that apply to all file types aren't possbile or are convoluted.
While these videos are somewhat outdated, they show the direction Microsoft is trying to take things and it's quite innovative:
Vista Concept Videos
In particular, the Higher Education video shows the potential of WinFS.
Actually, that would be looking at quantum physics in an extremely wrong light.
Unless the whales existing in a completely isolated environment, even isolated from all forms of energy from the outside environment, they very much do exist regardless of whether or not somebody is "looking" at them.
Last time I checked, most whales live in the ocean where there are plenty of things "looking" at them all the time. (Engery from a wide variety of sources, water, etc.)
The cat-in-a-box experiement, which you are undoubtedly referencing, is a *thought* experiement. It cannot be taken literally.
I've never, ever seen a blue whale in person. Never! In fact, I don't know anybody who has! Sure, I've seen pictures... but those could be fake.
My personal experiences and the hearsay of my friends are the only accurate way to measure the world around me. If I personally don't know somebody who has seen a blue whale, they can't possibly exist.
Right?
Apple should understand that most people want OS X for the operating system's technical superiority, not the beautifulness of Apple's hardware.
BWHAHAHAHAHHAAAH.... . hAHHAHHAhahaha.... gurgle... AHHAHAHAAHAHAHAH!
Yet another example of a Slashdot user who is completely detached from reality and has absolutely no idea how most people choose/use computers.
People buy Apples because of marketing. Macs look great. People love that. Nough said.
Most people buy iPods because they're sleek and pretty and well marketed and do what most people want them to do. These are "typical consumers". Basically none of these people care if the products they use are open source or not. Most don't even know what open source is.
Sure, there is a small segment of people who love to tweak and/or love the idea of open source, but those people don't matter. Why? Because they are more or less a statistical anomaly as far as sales figures go.
That's right people! You don't matter! Your passion for open source everything makes little or no difference, especially in the target market for iPods.
If anybody ever beats the iPod it will be because of better marketing, or maybe even better technology (as long as it's combined with better marketing)... being open source will never, ever make any damn difference.
If you're going to make claims like that the burden is on you to provide proof.
Provide evidence that Microsoft has a strategy to leave bugs in Windows so it can sell anti-malware software. Provide proof that McAfee writes viruses so it can sell software to fix them.
Until you do, your claim is meritless.
Microsoft offering anti-virus or anti-malware for Windows does not mean that they will stop fixing bugs in Windows.
No more than the fact that McAfee or Symantec offers antivirus software means they active release viruses to spurn the adoption of their software.
Microsoft is being pro-active about security by trying to get software into Windows that will stop undiscovered bugs from making systems expoitable. This will make users safer in the long run, and eventually (probably) will be included in every copy of Windows.
It is not the record labels that make ringtones expensive. Typically around 50% of the cost of the ringtone goes directly to the phone carrier. Some charge as high as 60%.
Most ringtones fees are billed directly to a user's cellphone bill. This dramatically increases sales because people buy more impulsively. To have this privilege, companies that sell ringtones must give a MAJORITY of the revenue from the sale directly to the cellphone company itself.
Of the remaining dollar, about 20 cents goes to aggregators that provide SMS/PSMS (premium sms... billing messages) integration (the guys who let you send text messages to cellphone networks.)
Of the 80 cents that remains after that, 20 cents will usually go to the content providers... or, for the lucky ones who have the resources to create and managing the licensing of their own ringtones, they get to keep that 20 cents.
Around 30 cents of the remaining 60 to 80 cents go to the record labels.
In the end, the people who actually run the ringtone site get between 30 and 50 cents per ringtone sold. Minus advertising. (Which is almost always a LOT more than 30 to 50 cents per sale.)
That is why almost all ringtone sites sell subscriptions, not single ringtones. They're hoping you don't use all your ringtones that come with your monthly subscription, and that you continue your subscription for more than one month. Otherwise, it's just not profitable.
But the point is that record labels only get between 10% and 15% of the cost of a ringtone. Of that 10% - 15%, the artists get some portion. (Usually 2% to 5%.) It's the carriers that take the VAST majority of the money.
If you want ringtones to be cheaper, bitch to the cellphone companies not the record labels.
Actually that 4 point disclaimer counts just fine. That's why it is there... to protect them legally.
You have the right to choose whether or not to purchase the CD. Period. End of story. That is where your "rights" end.
You do not have the right to force a company to allow you to play the CD here or there or anywhere. In fact, the company who sells you the CD could make it so the CD doesn't play anywhere at all, and you would have no right to force them to do otherwise.
Of course not too many people will buy a CD that can't play at all. The question is how far can a company go in restricting the use of the CD until enough people stop purchasing them that it makes the measures financially detrimental.
If the company does not make it clear that the CD won't play on a computer then there might be some issue with misrepresentation of the product. All the company has to do, however, is put a label on the CD stating that there is DRM that prevents the playback of the CD on a PC. In fact, I've seen several CDs with exactly this label.
At any rate, you have the right to buy or not to buy. That's it. Any other "rights" are figments of your imagination.
Quantify your claims.
Once you start handling millions of requests per day, bandwidth is your single largest reoccuring cost. The application that I build, and that runs on two Win2k3 machines (each of which cost about $4000, including the cost of Windows), costs about $50,000 a month in bandwidth costs.
The cost of the license for Windows is almost never a big factor when you're talking about large scale applications. Small apps, like home-brew ones, are affected far more by these costs.
The browsercaps file isn't updated enough for my taste, but it uses regular expressions and matches on general things like browser type (Openwave, etc.) so that it renders as best it can when it doesn't have an exact match.
You can easily update this yourself, and there is a device profiler that allows you to easily add support for additional phones.
Can you suggest a better way to do this? Mobile devices follow basically no standards and each have their own quirks. How do you suggest targeting these devices without a browsercaps file or database?