The provider I'm with, ServerMatrix, has Celeron 1.7s available for 49$/mth, with 1200GB/mth of transfer.
They have, last I checked, 12 GigE connections, more than just a few DS3s. Those 12 GigE connections are only partially utilized to boot. And none of them are Cogent.
Servers come with 10mbit connections (And there is no problem maxing out that 10mbit, trust me), but if you desperately need it you can get upgraded to a 100mbit card for 10$/mth.
Several other providers also have similar deals, but I've been with ServerMatrix a long time so I have done but minor research into other providers. I understand ServerBeach offers 2000GB/mth for 99$, but I have never used their service. I also understand 1and1 offers no-setup servers for 49$/mth, but those only have 500GB of bandwidth. P4 2.4s though.
Keep in mind that 49$ figure is with a 299$ setup fee... ServerMatrix has a setup pricing structure. For the server I mentioned:
There are, however, occasional specials for machines, that give you (permanently) the $199 setup level for 0$ setup.
All that said, I agree with you about server management. You can get that for 59$/mth from 1and1, but I'm not so sure I'd want to rely on somebody else to make every single config change to my box. What if some urgent change is needed? Do you simply wait on their techs to do it?
With dedicated servers starting at $49/mth US (I haven't seen any cheaper yet), considering the benefits they have over shell accounts, it's not surprising that unix shell accounts are losing popularity.
Most dedicated servers these days have over 1TB of bandwidth to boot, even at the 49$ level.
Some companies have return, replacement, or warranty policies that say they'll pay the shipping in certain circumstances. Since shipping to canada costs a lot more (fees, tarrifs, etc), that can cost them a lot of money.
Wouldn't it be grand if the developers behind SFU got together with the developers behind WINE and they shared knowledge of their respective OS, further enhancing the quality of both products?
Oh wait, it'll never happen, because while Microsoft has access to linux/UNIX sourcecode, the WINE team doesn't have access to Windows sourcecode...
That's true. Makes me wish that one could modify the TTL for domains hosted at GoDaddy. Their "total control" interface doesn't seem to list it, last I checked. If you want your GoDaddy domain to have a custom TTL, you have to host it yourself:(
Or you could just buy a Cantenna. It's based on the pringles-can antenna, but it's bigger, better, weather-proofed, and takes a lot of the hassle out of making an antenna out of a pringles can. Not to mention the fact it's better.
I never said that was a law, I'm only bringing up a point. If you don't like that example, see the earlier example I used with RedHat.
While I dislike cheaters, that doesn't mean I'm not capable of dealing with them. The game community I admin for has a very reasonable policy towards cheaters, including warnings, kicks, and bans. Our servers are, as a result, cheat-free, for the most part. Such is the case of any well-admin'd server. If I am stuck on a server with a cheater and nobody can do anything, the simple solution is to find another server. If a cheater wanders onto a server while I'm admining, then I'll take appropriate action.
To say that I should only play single player because I can't deal with other humans or cheaters is a laughably uninformed statement. Taking a strong stance against cheaters does not infer an inability to deal with cheaters.
So, a company (Redhat) produces a product that has a bug in it. That bug causes a security hole. Some asshole exploits the security hole and causes millions of dollars in damages. Are you saying that RedHat shouldn't be able to sue that person, or that anybody shouldn't be able to sue that person, because it's RedHat's fault for not fixing the bug? Your logic doesn't hold up.
In my example, a malicious user taking advantage of an exploit cost financial damage to companies. Cheaters cause financial damage to companies too; I'm sure there are people out there who have decided not to buy one game or another after having a bad experience with cheats. Perhaps a cheat-ridden demo put them off a final purchase.
Cheaters also cause other damage; it's no fun playing with a cheater. And while ruining other people's fun isn't illegal, it should damn well be something to sue over.
I'm not an American, but I understand that American culture is big on freedom. Shouldn't the freedom to be entertained be something that you shouldn't be allowed to violate for your own entertainment?
That is most unfortunate, as they appear to have marked up the drives to well over 200% their retail price.
Ah well, the option remains, it seems, to purchase a 14-drive machine and bump it up to a 5.6TB system by purchasing your own drives. That is, if the drives don't have custom firmware, as the earlier troll suggested.
I understand them taking up more space, but the dedicated server providers seem to have solved that problem. For ~1600 servers, you wouldn't need a very large datacenter at all... But I suppose even a very small datacenter is a lot more expensive than a place to house 10-11 racks for the 1u server versions (XServe).
I'm confused about cabling, however. Whitebox machines would surely have the exact same cabling requirements as an XServe (One power, one or more network), albeit a bit longer...
Is this really a supercomputer? Sounds more like a... supercluster to me.
At what point does linking together a bunch of off-the-shelf fully-self-contained PCs become a supercomputer? If doing so is the case, wouldn't it be a heck of a lot cheaper to link together whitebox machines, much as datacenters (the type that rent servers) tend to use whitebox servers rather than rackmount boxes?
I just feel like the term "supercomputer" is being sullied by so-called supercomputers that are nothing more than a simple cluster. Of course, I'm probably a moron, as I said earlier.
People who refuse to identify themselves, even if they are not suspected of a crime, will be arrested. Sound Orwellian? The Supreme Court also said people who are suspected of another crime might not be subject to arrest for not revealing their name.
So, lemme get this straight. You're NOT suspected of a crime and refuse to identify yourself, you get arrested. You ARE suspected of a crime and refuse to identify yourself and you DON'T get arrested? That's pretty fucked up.
Re:Dangers of using ATA or SATA for Raid
on
SATA vs ATA?
·
· Score: 1
It doesn't matter HOW long the power is down. The post this is all about is claiming that power failures can cause data loss due to cache loss. But if you have a 5 day power outage, obviously your UPS/generators won't last that long. But even if you only have a UPS, that's time enough to shut down the servers, avoiding data loss. The guy was saying you shouldn't use ATA/SATA because if the power fails your data will go poof, I'm saying that's not an issue.
Besides, 5 days of down time is nothing compared to the 2 weeks plus that Quebec and parts of Ontario were without power due to the Ice Storm in, I think it was, 1998. Some places were without power for over a month. In the middle of the Winter. In Canada.
I'm very surprised about that... Any.NET application designed with the integrated GUI designer would be done with Windows.Forms, no? Why would somebody writing a GUI app in a.NET language do the GUI from scratch when there's already a perfectly good (and powerful!) GUI designer/toolkit provided?
Re:Dangers of using ATA or SATA for Raid
on
SATA vs ATA?
·
· Score: 1
That's exactly what Apple's XServe RAID does; it disables the onboard cache on the drives and uses the RAID controllers' cache. That cache is what is battery backup, if I read it correctly.
The drives also appear to be ATA.
Re:Buy a RAID
on
SATA vs ATA?
·
· Score: 0, Flamebait
I don't recall asking what you like. There's a right and a wrong, and you're wrong.
And I don't recall telling you that I care. You're free to whine about it, of course, but I'm not quite sure what point that serves.
Firmware. If you require more explanation than this, then you're so far out of your depth that I can't even see you from here.
Try not to be quite so patronizing, it doesn't fit you well. I'm going to combine this with your next comment and ask you to provide proof. Your "Because I said so" response isn't sufficient. If you are going to insist that I'm incompetent, by all means, I may very well be. So PROVE it. Link to official Apple content showing that the drives use custom firmware, and that you can't place any other drive inside of an Apple Drive Module. You do that and I'll freely admit that I'm ill informed. I will, of course, be able to take solace in the fact that you have proven yourself to have the maturity of a young child, despite your superior knowledge.
Do you intentionally try to come across as haughty and insolent?
The dollar sign goes in front of the figure, not behind it. Basic literacy isn't too much to ask, I hope?
I happen to like writing currency how it's read. Is that too much of a problem for you? Does your brain shut down when you see it written that way, and you can no longer understand? Didn't think so.
Heh. That's good. That's funny. Run along and play now, you fucking amateur. Leave the discussion of business-class RAID systems to the grown-ups.
What a nice elitist-bastard attitude you've got there. Pray tell, what do you find so fucking amateurish about populating the drive bays after-market? You're not sacrificing reliability in the slightest, as I'm sure Apple uses off-the-shelf drives. You're simply using superior components, which seems to be what you are insisting is required. Fact is it remains cheaper (and better, through faster, higher capacity drives) to purchase Apple's 4x250 model and populate the device ourselves with 400GB drives. The extra 250GB drives can be used elsewhere, or sold off to cut costs.
Now, if you're done with your "I'm going to pretend I can't argue effectively as an excuse to hurl insults!" riff, I'd like to know what's so taboo about this.
I never said it wasn't reliable. I just said it wasn't inexpensive.
I had budgeted 1000$ for the rest of the server. A fast processor isn't required since we have all-hardware controllers. As long as the CPU isn't saturated under heavy load, all is good.
You think you can't fit redundant power, cooling, and controllers into that 1000$? Fine, add another 1000$. 2000$ even. That's 3000$ total for the server plus the 3600$ for the drives. That's 6600$ in total, still a good 5500$ cheaper than Apple's solution. Fine, it's not less than half the price, it's only 60% of the price.
The reason Apple's machines are so expensive is that they overcharge horrendously for the drives. They charge 5999$ for a model with 4x250GB drives, and 10999$ for a model with 14x250GB drives. That's 500$ per drive, or twice as expensive as the 400GB Hitachi drives (per GB) I used for my calculations. The drives Apple is using go for about 225$ US around where I live, so their markup is where the cost is coming from.
Some quick math: From the difference between their 4 drive and 14 drive model, we've determined that Apple charges 500$ per drive. This also holds true for their 7 drive model. Now, if Apple were to sell a model with NO DRIVES (So we could fill it ourselves), the XServe RAID would cost 3999$ US. Not so unreasonable any more for what you're getting.
Then, a 3.6TB XServe RAID with our own drives would cost 7599$, a good 3500$ less than Apple. Or if we wanted we could fill up the XServe with 14 drives, for a 5.6TB XServe RAID for 9599$ US, still 1500$ less than Apple, with over 2TB more storage capacity.
Heck, even if you buy the low-end XServe (the 4x250GB model), and THROW AWAY the drives (Or ebay them?), and buy 9x400GB drives, your total cost is 5999$ + 3600$ = 9599$.
So it's cheaper to buy an xserve, sell the drives, and use your own drives. Apple's pricing is a bit odd.
You can't really stop them from using P2P wireless short of jamming them, which is probably illegal.
Re:Dangers of using ATA or SATA for Raid
on
SATA vs ATA?
·
· Score: 3, Insightful
This would be why professional ATA RAID solutions have battery backup. Somebody previously linked to Apple's XServe solution. It has enough battery backup power built in to keep the caches going for 24 hours. If you can't find a power source for your server within 24 hours of a power failure, your data obviously isn't that important.
First off I'd assume if your data is so important you're going to have UPS and generators. If you don't have a generator, and the power fails, great, you've got 24 hours to purchase one. A 1500W generator costs about 450$ US, and should be more than powerfull enough to run your server, AND network connectivity. You'll not only keep your server happy during a power failure, you'll be able to keep using the server.
Anyhow, this post started out about the battery backup. What you stated as a major problem isn't one, since serious ATA RAID solutions have battery backup.
Re:Buy a RAID
on
SATA vs ATA?
·
· Score: 4, Informative
You may be right about building a system as reliable, and it'd certainly be hard to compete with it from a size standpoint, you are totally wrong about it being inexpensive.
Apple's 3.5TB system costs $10,999 US. If you were to build a system that comprised 9 Hitachi 7200RPM 400GB drives, you would acheive 100GB more storage space for 3,600$ plus the cost of the server it was hosted in. Throw in 750$ for a high-end RAID card and 1000$ for a server to enclose and handle it, and you're still priced at under HALF the price of Apple's solution.
So, in conclusion, Apple's solution is many things, and is certainly VERY sexy and attractive. But inexpensive compared to a self-built solution it is NOT.
Re:Connectors are poor on SATA
on
SATA vs ATA?
·
· Score: 2, Interesting
IIRC, I read something about a certain drive that had some sort of retention clip system. So it seems that the falling-out problem has already been solved by at least some manufacturers.
Slashdot Mirror
Yes, sorry, transfer.
The provider I'm with, ServerMatrix, has Celeron 1.7s available for 49$/mth, with 1200GB/mth of transfer.
They have, last I checked, 12 GigE connections, more than just a few DS3s. Those 12 GigE connections are only partially utilized to boot. And none of them are Cogent.
Servers come with 10mbit connections (And there is no problem maxing out that 10mbit, trust me), but if you desperately need it you can get upgraded to a 100mbit card for 10$/mth.
Several other providers also have similar deals, but I've been with ServerMatrix a long time so I have done but minor research into other providers. I understand ServerBeach offers 2000GB/mth for 99$, but I have never used their service. I also understand 1and1 offers no-setup servers for 49$/mth, but those only have 500GB of bandwidth. P4 2.4s though.
Keep in mind that 49$ figure is with a 299$ setup fee... ServerMatrix has a setup pricing structure. For the server I mentioned:
- $49 Monthly / $299 Setup
- $59 Monthly / $199 Setup
- $79 Monthly / $99 Setup
- $99 Monthly / $49 Setup
There are, however, occasional specials for machines, that give you (permanently) the $199 setup level for 0$ setup.
All that said, I agree with you about server management. You can get that for 59$/mth from 1and1, but I'm not so sure I'd want to rely on somebody else to make every single config change to my box. What if some urgent change is needed? Do you simply wait on their techs to do it?
With dedicated servers starting at $49/mth US (I haven't seen any cheaper yet), considering the benefits they have over shell accounts, it's not surprising that unix shell accounts are losing popularity.
Most dedicated servers these days have over 1TB of bandwidth to boot, even at the 49$ level.
Some companies have return, replacement, or warranty policies that say they'll pay the shipping in certain circumstances. Since shipping to canada costs a lot more (fees, tarrifs, etc), that can cost them a lot of money.
Wouldn't it be grand if the developers behind SFU got together with the developers behind WINE and they shared knowledge of their respective OS, further enhancing the quality of both products?
Oh wait, it'll never happen, because while Microsoft has access to linux/UNIX sourcecode, the WINE team doesn't have access to Windows sourcecode...
That's true. Makes me wish that one could modify the TTL for domains hosted at GoDaddy. Their "total control" interface doesn't seem to list it, last I checked. If you want your GoDaddy domain to have a custom TTL, you have to host it yourself :(
Yes, except those people relying on old cached info will be... everybody, since the TTL is still 48 hours.
As I read it, this changes nothing, because updates will still take days to get to users.
Or you could just buy a Cantenna. It's based on the pringles-can antenna, but it's bigger, better, weather-proofed, and takes a lot of the hassle out of making an antenna out of a pringles can. Not to mention the fact it's better.
I never said that was a law, I'm only bringing up a point. If you don't like that example, see the earlier example I used with RedHat.
While I dislike cheaters, that doesn't mean I'm not capable of dealing with them. The game community I admin for has a very reasonable policy towards cheaters, including warnings, kicks, and bans. Our servers are, as a result, cheat-free, for the most part. Such is the case of any well-admin'd server. If I am stuck on a server with a cheater and nobody can do anything, the simple solution is to find another server. If a cheater wanders onto a server while I'm admining, then I'll take appropriate action.
To say that I should only play single player because I can't deal with other humans or cheaters is a laughably uninformed statement. Taking a strong stance against cheaters does not infer an inability to deal with cheaters.
So, a company (Redhat) produces a product that has a bug in it. That bug causes a security hole. Some asshole exploits the security hole and causes millions of dollars in damages. Are you saying that RedHat shouldn't be able to sue that person, or that anybody shouldn't be able to sue that person, because it's RedHat's fault for not fixing the bug? Your logic doesn't hold up.
In my example, a malicious user taking advantage of an exploit cost financial damage to companies. Cheaters cause financial damage to companies too; I'm sure there are people out there who have decided not to buy one game or another after having a bad experience with cheats. Perhaps a cheat-ridden demo put them off a final purchase.
Cheaters also cause other damage; it's no fun playing with a cheater. And while ruining other people's fun isn't illegal, it should damn well be something to sue over.
I'm not an American, but I understand that American culture is big on freedom. Shouldn't the freedom to be entertained be something that you shouldn't be allowed to violate for your own entertainment?
That is most unfortunate, as they appear to have marked up the drives to well over 200% their retail price.
Ah well, the option remains, it seems, to purchase a 14-drive machine and bump it up to a 5.6TB system by purchasing your own drives. That is, if the drives don't have custom firmware, as the earlier troll suggested.
*slaps self* What am I talking about with 10-11 racks, that'd be for blade servers. 1u, that'd be 37-38 racks.
I understand them taking up more space, but the dedicated server providers seem to have solved that problem. For ~1600 servers, you wouldn't need a very large datacenter at all... But I suppose even a very small datacenter is a lot more expensive than a place to house 10-11 racks for the 1u server versions (XServe).
I'm confused about cabling, however. Whitebox machines would surely have the exact same cabling requirements as an XServe (One power, one or more network), albeit a bit longer...
Is this really a supercomputer? Sounds more like a... supercluster to me.
At what point does linking together a bunch of off-the-shelf fully-self-contained PCs become a supercomputer? If doing so is the case, wouldn't it be a heck of a lot cheaper to link together whitebox machines, much as datacenters (the type that rent servers) tend to use whitebox servers rather than rackmount boxes?
I just feel like the term "supercomputer" is being sullied by so-called supercomputers that are nothing more than a simple cluster. Of course, I'm probably a moron, as I said earlier.
I wonder how this will affect Half-Life 2. Oh well, at least VALVe still has STEAM to release HL2.
People who refuse to identify themselves, even if they are not suspected of a crime, will be arrested. Sound Orwellian? The Supreme Court also said people who are suspected of another crime might not be subject to arrest for not revealing their name.
So, lemme get this straight. You're NOT suspected of a crime and refuse to identify yourself, you get arrested. You ARE suspected of a crime and refuse to identify yourself and you DON'T get arrested? That's pretty fucked up.
It doesn't matter HOW long the power is down. The post this is all about is claiming that power failures can cause data loss due to cache loss. But if you have a 5 day power outage, obviously your UPS/generators won't last that long. But even if you only have a UPS, that's time enough to shut down the servers, avoiding data loss. The guy was saying you shouldn't use ATA/SATA because if the power fails your data will go poof, I'm saying that's not an issue.
Besides, 5 days of down time is nothing compared to the 2 weeks plus that Quebec and parts of Ontario were without power due to the Ice Storm in, I think it was, 1998. Some places were without power for over a month. In the middle of the Winter. In Canada.
I'm very surprised about that... Any .NET application designed with the integrated GUI designer would be done with Windows.Forms, no? Why would somebody writing a GUI app in a .NET language do the GUI from scratch when there's already a perfectly good (and powerful!) GUI designer/toolkit provided?
That's exactly what Apple's XServe RAID does; it disables the onboard cache on the drives and uses the RAID controllers' cache. That cache is what is battery backup, if I read it correctly.
The drives also appear to be ATA.
I don't recall asking what you like. There's a right and a wrong, and you're wrong.
And I don't recall telling you that I care. You're free to whine about it, of course, but I'm not quite sure what point that serves.
Firmware. If you require more explanation than this, then you're so far out of your depth that I can't even see you from here.
Try not to be quite so patronizing, it doesn't fit you well. I'm going to combine this with your next comment and ask you to provide proof. Your "Because I said so" response isn't sufficient. If you are going to insist that I'm incompetent, by all means, I may very well be. So PROVE it. Link to official Apple content showing that the drives use custom firmware, and that you can't place any other drive inside of an Apple Drive Module. You do that and I'll freely admit that I'm ill informed. I will, of course, be able to take solace in the fact that you have proven yourself to have the maturity of a young child, despite your superior knowledge.
Do you intentionally try to come across as haughty and insolent?
The dollar sign goes in front of the figure, not behind it. Basic literacy isn't too much to ask, I hope?
I happen to like writing currency how it's read. Is that too much of a problem for you? Does your brain shut down when you see it written that way, and you can no longer understand? Didn't think so.
Heh. That's good. That's funny. Run along and play now, you fucking amateur. Leave the discussion of business-class RAID systems to the grown-ups.
What a nice elitist-bastard attitude you've got there. Pray tell, what do you find so fucking amateurish about populating the drive bays after-market? You're not sacrificing reliability in the slightest, as I'm sure Apple uses off-the-shelf drives. You're simply using superior components, which seems to be what you are insisting is required. Fact is it remains cheaper (and better, through faster, higher capacity drives) to purchase Apple's 4x250 model and populate the device ourselves with 400GB drives. The extra 250GB drives can be used elsewhere, or sold off to cut costs.
Now, if you're done with your "I'm going to pretend I can't argue effectively as an excuse to hurl insults!" riff, I'd like to know what's so taboo about this.
I never said it wasn't reliable. I just said it wasn't inexpensive.
I had budgeted 1000$ for the rest of the server. A fast processor isn't required since we have all-hardware controllers. As long as the CPU isn't saturated under heavy load, all is good.
You think you can't fit redundant power, cooling, and controllers into that 1000$? Fine, add another 1000$. 2000$ even. That's 3000$ total for the server plus the 3600$ for the drives. That's 6600$ in total, still a good 5500$ cheaper than Apple's solution. Fine, it's not less than half the price, it's only 60% of the price.
The reason Apple's machines are so expensive is that they overcharge horrendously for the drives. They charge 5999$ for a model with 4x250GB drives, and 10999$ for a model with 14x250GB drives. That's 500$ per drive, or twice as expensive as the 400GB Hitachi drives (per GB) I used for my calculations. The drives Apple is using go for about 225$ US around where I live, so their markup is where the cost is coming from.
Some quick math: From the difference between their 4 drive and 14 drive model, we've determined that Apple charges 500$ per drive. This also holds true for their 7 drive model. Now, if Apple were to sell a model with NO DRIVES (So we could fill it ourselves), the XServe RAID would cost 3999$ US. Not so unreasonable any more for what you're getting.
Then, a 3.6TB XServe RAID with our own drives would cost 7599$, a good 3500$ less than Apple. Or if we wanted we could fill up the XServe with 14 drives, for a 5.6TB XServe RAID for 9599$ US, still 1500$ less than Apple, with over 2TB more storage capacity.
Heck, even if you buy the low-end XServe (the 4x250GB model), and THROW AWAY the drives (Or ebay them?), and buy 9x400GB drives, your total cost is 5999$ + 3600$ = 9599$.
So it's cheaper to buy an xserve, sell the drives, and use your own drives. Apple's pricing is a bit odd.
You can't really stop them from using P2P wireless short of jamming them, which is probably illegal.
This would be why professional ATA RAID solutions have battery backup. Somebody previously linked to Apple's XServe solution. It has enough battery backup power built in to keep the caches going for 24 hours. If you can't find a power source for your server within 24 hours of a power failure, your data obviously isn't that important.
First off I'd assume if your data is so important you're going to have UPS and generators. If you don't have a generator, and the power fails, great, you've got 24 hours to purchase one. A 1500W generator costs about 450$ US, and should be more than powerfull enough to run your server, AND network connectivity. You'll not only keep your server happy during a power failure, you'll be able to keep using the server.
Anyhow, this post started out about the battery backup. What you stated as a major problem isn't one, since serious ATA RAID solutions have battery backup.
You may be right about building a system as reliable, and it'd certainly be hard to compete with it from a size standpoint, you are totally wrong about it being inexpensive.
Apple's 3.5TB system costs $10,999 US. If you were to build a system that comprised 9 Hitachi 7200RPM 400GB drives, you would acheive 100GB more storage space for 3,600$ plus the cost of the server it was hosted in. Throw in 750$ for a high-end RAID card and 1000$ for a server to enclose and handle it, and you're still priced at under HALF the price of Apple's solution.
So, in conclusion, Apple's solution is many things, and is certainly VERY sexy and attractive. But inexpensive compared to a self-built solution it is NOT.
IIRC, I read something about a certain drive that had some sort of retention clip system. So it seems that the falling-out problem has already been solved by at least some manufacturers.