You don't need root (which they have said), because root can simply create a "web" or "www" group on the system (in the highly unlikely event one doesn't already exist), make you a member of the group, and give the group full privs on the stuff the webmaster needs privs on, e.g.:
/etc/httpd/ and things under it, so you can tweak httpd.conf
/var/log/httpd/ and things under it, so you can munge logs
wherever Postgres lives
whatever the web docroot is.
That's the course I'd probably take, in your root's shoes.
Kinda need more zoom (500mm is good for most kinds of eclipses, I think), though. A nice telescope with a guider on it would work wonders for this sort of thing.
Another timelapse of the same eclipse, from 2400 miles WSW of San Francisco, is at http://www.astroday.net/LunarEclipse.html in QuickTime format... I hope my friend doesn't mind the slashdotting. Still not much in the way of telephoto or zoom, but nifty cloud patterns and stuff.
Okay, yes, an Itanium 2 at certain clock speeds is faster than a PPC440GX at certain clock speeds. It's also faster than a 286-12, the 1GHz Motorola 6502 out of a Commodore 64, and an ENIAC. But anway.:)
It's important to remember that Hawaii was colonized twice -- once (before 600 AD) by small, slight people from the Marquesas islands, and then later (1200-1300 AD) by taller people from Tahiti.:)
It was not chosen blindly for the Altix. It was a "we want the absolute best for this system, money no issue" decision.
Uh, you almost got that right. I think what you meant was "we want the absolute best IA-64 compatible processor for this system, money no issue."
After all, having not learned from the pain of going from 32-bit IRIX 5.x to 64-bit IRIX 6.x and issuing buffer-overflow patches for just about every command in UNIX in the mid-1990s (at which time I was an IRIX admin), SGI threw out its management in 1998, brought in a guy from Compaq, and promptly committed to a cutover from MIPS to Intel, starting with Xeon boxes that fall, and in early 2000, the Itanium. Of course there were
delays and problems and more delays while SGI announced MIPS-and-someday-Itanium systems and tried one Pentium system after another, at least having the sense to drop Windows NT for Linux in the process. Finally, in early 2003, having apparently failed to ever sell even a single Itanium system, mind you, SGI shipped Itanium 2 systems.
My take? Well, I like SGI. They've always made really cool shiny toys, and I really expect them to continue doing that, regardless of CPU architecture. And I'm glad to see them at the top of the Top500 list - a position that, despite all their shiny toys and sometime ownership of Cray, they've never occupied before, I might add - since even ASCI Blue Mountain was down to something like #30 a year ago after dropping out of the top 10 in 2002 (and I don't think it was even on this spring's list; maybe it got decommissioned).
But IMO, Intel has SGI seduced, fooled, and generally whipped. Even though SGI still has more model lines with MIPS processors in them, and has committed to keeping MIPS around until at least 2006 (and good luck getting rid of it then at this rate), Intel got SGI to sell its soul for a basket full of promises. Either SGI loathes Intel for this but is afraid to do anything about it (after all, switching architectures again would be a pain, even if they are now a Linux company and Linux runs on everything), or SGI has a nasty case of Stockholm Syndrome and thinks that Intel is actually on its side. Anyway, SGI has tied itself to Itanium for over a half-decade now, even if it didn't ship jack for most of that time, and it's unwilling to change horses.
Intel, of course, doesn't care, because it knows that if it's not SGI building that 10240-Itanium2 supercomputer, it'll be Hewlett-Packard (stripped of Alpha and PA-RISC) or some other Itanium server vendor. And with Alpha gone, MIPS on the way out, PA-RISC gone or going and Sun screwing around with AMD, Intel only has to worry from competition from IBM Power and traditional supercomputers from Cray and NEC and folks like that.
What about the 16000+ PPC cluster that was announced a short while ago? Oh yes, it was slower than this 8000+ I2 cluster.
You wouldn't be spouting bullshit, would you?
Nope - and your unfamiliarity with the PowerPC lineup is showing.:)
Blue Gene/L used chips based on the PowerPC 440GX. The 400-series PowerPC chips are designed for lower power usage in embedded applications. For x86 folks, think of something like a Via C3 or a Transmeta Crusoe. It takes a LOT of them to add up to serious computational power, but they don't take a lot of juice to do it - that's the big deal in the case of that system.
Each 2.3GHz PPC970FX processor in the Xserve G5s now used in Virginia Tech's System X is about 2.5x as powerful as a PowerPC 440 GX, so you could theoretically get around the same performance as those 16,000 440GXes with "only" 6,600 960FXes.
Of course,the 2.3GHz Xserve G5s aren't even the fastest G5 systems out there - but the 2.5GHz Power Mac G5s take up more space and require more power.
Oh, and I don't dispute that the Itanium 2 is a better chip than the original Itanium (which I think was an albatross, big-time). But in the time it took Intel to get around to it, pretty much everybody else came out with 64-bit chips, and a fair number of them can do more with less.
I'm now on my second Bluetooth phone from AT&T, and both have synced just fine with my Macs. I haven't tried using them in a Bluetooth-enabled car, but they're by no means limited to use with headsets, for example.
This one's a Sony-Ericsson T616; its predecessor was a Nokia 3650, "world phone" but unfortunately AT&T "built out" their network in the rural areas around me using a different GSM frequency band than any of the ones supported by that tri-band phone. (It was 900/1800/1900MHz; they had 900MHz in town, but out of town deployed somewhere around 800-850Mhz. Wankers.)
Here, Cingular is inheriting a pretty good network; FCC databases show that AT&T's tower count in this county is about double anyone else's. I think some of the others are trying to build out and catch up, so maybe in a few years I'll have other feasible options.
I'm not the one who made the comparison to fine art, but I certainly think it's "art" in the broader sense. I'm guessing that either you're some sort of fine-art snob, or you've honestly never encountered the concept of "installation art" or "performance art."
Making a horribly old machine run a modern OS via emulation, and more importantly posting a log of how long it takes to do seemingly meaningless tasks, is "art" in the same way as hooking up all the lights in a tall building to a computer and setting up a web script to let people display messages on the building.
(I believe that latter one is done in Germany or somewhere else in Europe, each year.)
I don't think you, I, or anyone else in the world -- especially any of us Slashdot geeks -- is remotely qualified to dictate to others what is, and isn't, art. Art is part of culture, and cultural absolutism isn't a very healthy thing in a heterogenous world.
"...with Columbia, scientists are discovering they can potentially predict hurricane paths a full five days before the storms reach landfall."
You don't live somewhere that gets hurricanes, do you? 'Cause scientists can already "potentially predict hurricane paths a full five days before the storms reach landfall." Hell, I can do that. A freakin' Magic 8 Ball can potentially do that.
Maybe they're trying to say something about doing it with a better degree of accuracy, or being right more of the time, or something like that, but it doesn't sound like it from that quote.
"Hey, guys, look at this life-sized computer-generated stripper I'm rendering in real-ti... oh, what? Um, tell the reporter we think it'd be good for hurricane prediction."
...is that a lot of people are going to say "Oooh, shiny! The world's fastest computer uses Itaniums, so they must be good!" failing to notice that the system requires huge quantities of Itaniums, or that just about any system with the same number of... well, gosh, almost any processor except an Itanium would be even faster, or that, well, the Itanium in general has been a horrendously expensive fiasco for Intel.:)
But I'm sure in 6 months someone will go build a 10,000-CPU PPC or AMD-whatever cluster that'll eat this one for breakfast and still be hungry enough to eat the team that came up with it.
For the last few years, various companies have been coming out with devices that store and/or display photos. Some have viewing screens, others don't. Most have card readers, USB or Firewire connections. Some offer features like video out or CD-burning, some are primarily MP3 or video players with photo storage as an additional feature.
Right now, today, the new iPod Photo isn't destroying their market share. But as of today, we're at the point where we can buy an iPod and a little gizmo (like the SanDisk one) we stick flash cards into for display on a TV... or just buy an iPod Photo. With that Belkin attachment, any iPod can be your place to dump photos in the field. And other than adding card slots, most of the other features other products have that the iPod Photo doesn't offer can be added in firmware updates. One at a time. Step by step. Until another market segment is overrun by white-earbudded iPod people.:)
But by the time that happens, the iPod Photo will probably have video playback capability, since again, that's totally just a matter of adding the capabilities through a firmware upgrade. Sure, it's not a top priority for Apple right now, but they've got the hardware now, and just have to code the functionality in the firmware.
In a year or two, will we all be saying "iPod uber alles" with regard to things other than music? Dunno. But if it happens, I won't be surprised.
This is the same thought I had... by Christmas all the "hip" speakers in New York, San Francisco and the Bay Area will be showing up for presentations with iPods instead of laptops.;)
By the late 1980's, the Citadel BBS program originally written around 1980 for CP/M by Jeff Prothero (aka Cynbe Ru Taaren) had been ported to the DOS PC, Mac, Atari ST, Amiga, C-64, C-128, and (courtesy of Ignatius T. Foobar) UNIX. Citadel/UX ran on QuartzBBS at quartz.rutgers.edu from 1989-1994, and still runs at bbs.fdu.edu, bbs.k2nesoft.com, and bbs.quartz.org (Quartz II). A derivative, DOC (Dave's Own Citadel) is used for the ISCA BBS at bbs.isca.uiowa.edu. If you used a Citadel on any other platform way back when, you should feel right at home.
If this is the "secret summit" I'm aware of, it's really old news - it happened almost a year ago. It wasn't totally a CAUCE thing - other antispammers were involved as well. I didn't go, but was a part of the discussions that led up to it. Both sides reached some common ground and put out a press release afterward. The DMA has since broken their word on every point they agreed to, if I recall correctly.:)
It's kinda strange to think it, but *every* girl I ever dated or did anything with, I either met online (BBS, QuantumLink, Bitnet, Internet) or because of a common interest in computers/online stuff. And that's, um, not exactly a small number of people.
I first ran across my wife when I was an op on an Internet BBS and she was a bewildered new user, back in 1992. We didn't meet in person 'til 1993 (by which point we'd talked enough to know each other's gender and geographic location), and yah, it involved some travel initially. We've been an item for 6 years now, though, and married half that.
We even have a baby (her name is Tera, but we call her Terabit in true/. fashion), who was fortunate enough to load up on cuteness from her mom (my wife has a degree in dance and also modeled during college, looks good).:)
"The SID - Security Identifier - is a 128-but GUID created during the install of NT to uniquely identify that machine to the domain that it will become a member of. Therefore, if you GHOST one install to another machine, both will have the same SID, and unpredictable behaviour will arise. If your not using NTs domain security, then it doesn't matter."
All true. I'm a little curious as to what the SID gets you, though. Isn't it just as easy to refer to machines by their MAC addresses, names within the domain, or IP addresses if you're using IP, with all those pieces of data cross-referencing like they do with ARP/RARP? I'm not sure why a fourth piece of data should be necessary.
NT invented ACL's? Huh? I thought Apollo's DomainOS had them all along. And didn't HP/UX have them too? I know I've encountered and used them on DomainOS, and that was on seriously crufty, first-half-of-the-decade hardware.:)
Adaptec and BusLogic both are, and have been for years, pretty well supported by Linux.
As far as your specific question (limiting the access of one user within a group to a certain file in a certain directory where the group otherwise has full access)... uh, I have my doubts about whether that can even be made to work in such a way that the user cannot override it due to having full access to the directory. And I have a hard time figuring out what kind of scenario would warrant this particular configuration. And yes, I know ACL's.:)
NEC has (or had) an article on their website about a fellow at a.gov installation who lost his bag (with his laptop in it, foolish man); said bag (with the laptop still in it) turned up in the parking lot of the facility, was deemed "suspicious" and taken out by a bomb squad robot with a shotgun. The laptop's screen was trashed, but the machine still booted and he was able to retrieve his data. Woo.
Salamander said that the state of HA for Linux wasn't much different than the state of HA for NT. Beg to differ just a little.
I've got a cluster up and running a beta of TurboCluster (from our buddies at TurboLinux) with Apache atop it. This is a product that takes multiple machines (they've tested up to 12 nodes, I think) and makes them look like one big one. One node serves as a "router" (or "cluster controller" as some might call it) as well as being a node, and if that one goes down, another takes over. We've got the heartbeats and all that stuff.
It's - dare I say it? - EASY to set up, and it works. And that's the BETA code. They're getting ready for the official release of it later this month, I believe, and by then, I plan on having some "name brand" websites running on it for them to point to.
The only stuff it doesn't handle on its own yet is the synchronization of data and logfiles between the servers - but hey, that's a pain on ANY cluster, and there are a "metric shitload" of ways to deal with that under Linux. Everything from rsync and rdist to NFS and FibreChannel-based GFS. And VirtualFS, and Coda, and... did I miss any?:)
Imagine, if you will, a server cabinet containing a few Penguin Computing rackmounts running TurboCluster, a switch, and some sundry other stuff - maybe an NFS server, maybe an SQL server - and you've got one kickass server setup.
Slashdot Mirror
Okay, FCC folks, I'm waiting for the WiFi mesh...
Scarily so, IMO. From the side, except for the extra wheels, the lines are nearly identical.
You don't need root (which they have said), because root can simply create a "web" or "www" group on the system (in the highly unlikely event one doesn't already exist), make you a member of the group, and give the group full privs on the stuff the webmaster needs privs on, e.g.:
That's the course I'd probably take, in your root's shoes.
Another timelapse of the same eclipse, from 2400 miles WSW of San Francisco, is at http://www.astroday.net/LunarEclipse.html in QuickTime format... I hope my friend doesn't mind the slashdotting. Still not much in the way of telephoto or zoom, but nifty cloud patterns and stuff.
Geez, and I thought I was a pedant. ;)
:)
Okay, yes, an Itanium 2 at certain clock speeds is faster than a PPC440GX at certain clock speeds. It's also faster than a 286-12, the 1GHz Motorola 6502 out of a Commodore 64, and an ENIAC. But anway.
It's important to remember that Hawaii was colonized twice -- once (before 600 AD) by small, slight people from the Marquesas islands, and then later (1200-1300 AD) by taller people from Tahiti. :)
Uh, you almost got that right. I think what you meant was "we want the absolute best IA-64 compatible processor for this system, money no issue."
After all, having not learned from the pain of going from 32-bit IRIX 5.x to 64-bit IRIX 6.x and issuing buffer-overflow patches for just about every command in UNIX in the mid-1990s (at which time I was an IRIX admin), SGI threw out its management in 1998, brought in a guy from Compaq, and promptly committed to a cutover from MIPS to Intel, starting with Xeon boxes that fall, and in early 2000, the Itanium. Of course there were delays and problems and more delays while SGI announced MIPS-and-someday-Itanium systems and tried one Pentium system after another, at least having the sense to drop Windows NT for Linux in the process. Finally, in early 2003, having apparently failed to ever sell even a single Itanium system, mind you, SGI shipped Itanium 2 systems.
My take? Well, I like SGI. They've always made really cool shiny toys, and I really expect them to continue doing that, regardless of CPU architecture. And I'm glad to see them at the top of the Top500 list - a position that, despite all their shiny toys and sometime ownership of Cray, they've never occupied before, I might add - since even ASCI Blue Mountain was down to something like #30 a year ago after dropping out of the top 10 in 2002 (and I don't think it was even on this spring's list; maybe it got decommissioned).
But IMO, Intel has SGI seduced, fooled, and generally whipped. Even though SGI still has more model lines with MIPS processors in them, and has committed to keeping MIPS around until at least 2006 (and good luck getting rid of it then at this rate), Intel got SGI to sell its soul for a basket full of promises. Either SGI loathes Intel for this but is afraid to do anything about it (after all, switching architectures again would be a pain, even if they are now a Linux company and Linux runs on everything), or SGI has a nasty case of Stockholm Syndrome and thinks that Intel is actually on its side. Anyway, SGI has tied itself to Itanium for over a half-decade now, even if it didn't ship jack for most of that time, and it's unwilling to change horses.
Intel, of course, doesn't care, because it knows that if it's not SGI building that 10240-Itanium2 supercomputer, it'll be Hewlett-Packard (stripped of Alpha and PA-RISC) or some other Itanium server vendor. And with Alpha gone, MIPS on the way out, PA-RISC gone or going and Sun screwing around with AMD, Intel only has to worry from competition from IBM Power and traditional supercomputers from Cray and NEC and folks like that.
You wouldn't be spouting bullshit, would you?
Nope - and your unfamiliarity with the PowerPC lineup is showing. :)
Blue Gene/L used chips based on the PowerPC 440GX. The 400-series PowerPC chips are designed for lower power usage in embedded applications. For x86 folks, think of something like a Via C3 or a Transmeta Crusoe. It takes a LOT of them to add up to serious computational power, but they don't take a lot of juice to do it - that's the big deal in the case of that system.
Each 2.3GHz PPC970FX processor in the Xserve G5s now used in Virginia Tech's System X is about 2.5x as powerful as a PowerPC 440 GX, so you could theoretically get around the same performance as those 16,000 440GXes with "only" 6,600 960FXes.
Of course,the 2.3GHz Xserve G5s aren't even the fastest G5 systems out there - but the 2.5GHz Power Mac G5s take up more space and require more power.
Oh, and I don't dispute that the Itanium 2 is a better chip than the original Itanium (which I think was an albatross, big-time). But in the time it took Intel to get around to it, pretty much everybody else came out with 64-bit chips, and a fair number of them can do more with less.
I'm now on my second Bluetooth phone from AT&T, and both have synced just fine with my Macs. I haven't tried using them in a Bluetooth-enabled car, but they're by no means limited to use with headsets, for example.
This one's a Sony-Ericsson T616; its predecessor was a Nokia 3650, "world phone" but unfortunately AT&T "built out" their network in the rural areas around me using a different GSM frequency band than any of the ones supported by that tri-band phone. (It was 900/1800/1900MHz; they had 900MHz in town, but out of town deployed somewhere around 800-850Mhz. Wankers.)
Here, Cingular is inheriting a pretty good network; FCC databases show that AT&T's tower count in this county is about double anyone else's. I think some of the others are trying to build out and catch up, so maybe in a few years I'll have other feasible options.
I'm not the one who made the comparison to fine art, but I certainly think it's "art" in the broader sense. I'm guessing that either you're some sort of fine-art snob, or you've honestly never encountered the concept of "installation art" or "performance art."
Making a horribly old machine run a modern OS via emulation, and more importantly posting a log of how long it takes to do seemingly meaningless tasks, is "art" in the same way as hooking up all the lights in a tall building to a computer and setting up a web script to let people display messages on the building.
(I believe that latter one is done in Germany or somewhere else in Europe, each year.)
I don't think you, I, or anyone else in the world -- especially any of us Slashdot geeks -- is remotely qualified to dictate to others what is, and isn't, art. Art is part of culture, and cultural absolutism isn't a very healthy thing in a heterogenous world.
Ah, but I've potentially predicted it. :)
:)
Throw in enough adverbs, and no one has to actually do anything, ever.
I thought the people who don't have a clue what art is were the ones who went around trying to tell other people what was, and wasn't, art. :)
You don't live somewhere that gets hurricanes, do you? 'Cause scientists can already "potentially predict hurricane paths a full five days before the storms reach landfall." Hell, I can do that. A freakin' Magic 8 Ball can potentially do that.
Maybe they're trying to say something about doing it with a better degree of accuracy, or being right more of the time, or something like that, but it doesn't sound like it from that quote.
"Hey, guys, look at this life-sized computer-generated stripper I'm rendering in real-ti... oh, what? Um, tell the reporter we think it'd be good for hurricane prediction."
But I'm sure in 6 months someone will go build a 10,000-CPU PPC or AMD-whatever cluster that'll eat this one for breakfast and still be hungry enough to eat the team that came up with it.
Companies involved in various aspects of this market include Archos, Nikon, Sony, SanDisk and Epson, as well as a whole slew of smaller names like I/OMagic, Sima, Transcend, Vosonic, Innoplus, Digi Magic and Delkin.
Right now, today, the new iPod Photo isn't destroying their market share. But as of today, we're at the point where we can buy an iPod and a little gizmo (like the SanDisk one) we stick flash cards into for display on a TV... or just buy an iPod Photo. With that Belkin attachment, any iPod can be your place to dump photos in the field. And other than adding card slots, most of the other features other products have that the iPod Photo doesn't offer can be added in firmware updates. One at a time. Step by step. Until another market segment is overrun by white-earbudded iPod people. :)
But by the time that happens, the iPod Photo will probably have video playback capability, since again, that's totally just a matter of adding the capabilities through a firmware upgrade. Sure, it's not a top priority for Apple right now, but they've got the hardware now, and just have to code the functionality in the firmware.
In a year or two, will we all be saying "iPod uber alles" with regard to things other than music? Dunno. But if it happens, I won't be surprised.
This is the same thought I had... by Christmas all the "hip" speakers in New York, San Francisco and the Bay Area will be showing up for presentations with iPods instead of laptops. ;)
By the late 1980's, the Citadel BBS program originally written around 1980 for CP/M by Jeff Prothero (aka Cynbe Ru Taaren) had been ported to the DOS PC, Mac, Atari ST, Amiga, C-64, C-128, and (courtesy of Ignatius T. Foobar) UNIX. Citadel/UX ran on QuartzBBS at quartz.rutgers.edu from 1989-1994, and still runs at bbs.fdu.edu, bbs.k2nesoft.com, and bbs.quartz.org (Quartz II). A derivative, DOC (Dave's Own Citadel) is used for the ISCA BBS at bbs.isca.uiowa.edu. If you used a Citadel on any other platform way back when, you should feel right at home.
If this is the "secret summit" I'm aware of, it's really old news - it happened almost a year ago. It wasn't totally a CAUCE thing - other antispammers were involved as well. I didn't go, but was a part of the discussions that led up to it. Both sides reached some common ground and put out a press release afterward. The DMA has since broken their word on every point they agreed to, if I recall correctly. :)
I first ran across my wife when I was an op on an Internet BBS and she was a bewildered new user, back in 1992. We didn't meet in person 'til 1993 (by which point we'd talked enough to know each other's gender and geographic location), and yah, it involved some travel initially. We've been an item for 6 years now, though, and married half that.
We even have a baby (her name is Tera, but we call her Terabit in true /. fashion), who was fortunate enough to load up on cuteness from her mom (my wife has a degree in dance and also modeled during college, looks good). :)
So yeah, it can definitely work.
No, no, it's "Buy more hardware, buy more software, turn off security, and everything will be just fine." :)
"The SID - Security Identifier - is a 128-but GUID created during the install of NT to uniquely identify that machine to the domain that it will become a member of. Therefore, if you GHOST one install to another machine, both will have the same SID, and unpredictable behaviour will arise.
If your not using NTs domain security, then it doesn't matter."
All true. I'm a little curious as to what the SID gets you, though. Isn't it just as easy to refer to machines by their MAC addresses, names within the domain, or IP addresses if you're using IP, with all those pieces of data cross-referencing like they do with ARP/RARP? I'm not sure why a fourth piece of data should be necessary.
NT invented ACL's? Huh? I thought Apollo's DomainOS had them all along. And didn't HP/UX have them too? I know I've encountered and used them on DomainOS, and that was on seriously crufty, first-half-of-the-decade hardware. :)
:)
Adaptec and BusLogic both are, and have been for years, pretty well supported by Linux.
As far as your specific question (limiting the access of one user within a group to a certain file in a certain directory where the group otherwise has full access)... uh, I have my doubts about whether that can even be made to work in such a way that the user cannot override it due to having full access to the directory. And I have a hard time figuring out what kind of scenario would warrant this particular configuration. And yes, I know ACL's.
NEC has (or had) an article on their website .gov installation who lost
about a fellow at a
his bag (with his laptop in it, foolish man);
said bag (with the laptop still in it) turned
up in the parking lot of the facility, was
deemed "suspicious" and taken out by a bomb
squad robot with a shotgun. The laptop's screen
was trashed, but the machine still booted and he
was able to retrieve his data. Woo.
I've got a cluster up and running a beta of TurboCluster (from our buddies at TurboLinux) with Apache atop it. This is a product that takes multiple machines (they've tested up to 12 nodes, I think) and makes them look like one big one. One node serves as a "router" (or "cluster controller" as some might call it) as well as being a node, and if that one goes down, another takes over. We've got the heartbeats and all that stuff.
It's - dare I say it? - EASY to set up, and it works. And that's the BETA code. They're getting ready for the official release of it later this month, I believe, and by then, I plan on having some "name brand" websites running on it for them to point to.
The only stuff it doesn't handle on its own yet is the synchronization of data and logfiles between the servers - but hey, that's a pain on ANY cluster, and there are a "metric shitload" of ways to deal with that under Linux. Everything from rsync and rdist to NFS and FibreChannel-based GFS. And VirtualFS, and Coda, and... did I miss any? :)
Imagine, if you will, a server cabinet containing a few Penguin Computing rackmounts running TurboCluster, a switch, and some sundry other stuff - maybe an NFS server, maybe an SQL server - and you've got one kickass server setup.