I read it too, and it's true - they DO say all of the machines need access to the source which they do not.
Maybe there's some special cases, but I've never had to have a shared source repository in order to use distcc.
They also say the machines need to be exactly the same configuration, and they do elaborate on that a little bit, but it's not strictly true. Depending on the source you're compiling, you might only need to just have the same major version of GCC.
Yea. I used to use distcc a lot about five years ago. It doesn't help with all compile functions but it can help if you're compiling something big like X11 or KDE.
When I first used Gentoo several years ago it was with a 950Mhz Athlon CPU and it wasn't too bad.
Now, with quad-core CPU's becomming the norm even in Desktop machines, the compiling thing is even less of an issue.
I was able to run software on Gentoo that I could never get to run well together on any other distribution. You can almost always get the latest and greatest versions of everything with Gentoo. With Kernels taking almost no time to compile these days, the source distribution excuse just keeps getting weaker and weaker.
The main problem with Gentoo was it's moving target when it came to core system tools. They'd upgrade the ebuild system and it could sometimes hose your system. I haven't used Gentoo in a couple years because of it, but I'd be willing to give it another shot eventually.
I was a lego boy when I grew up. I had a lot of legos.
I also had some imitation blocked made by Tyco. These immitation blocks never fit together well. You'd build something and it would fall apart. Although the bricks looked almost identical, the Tyco bricks just sucked.
So, I do worry about imitation blocks. Lego blocks are the best because they have impossibly high standards during manufacturing in order to avoid the frustration I experienced with the Tyco blocks.
If someone else is going to start really making sets to compete with Lego, let's hope they go the distance and implement quality control like Lego does.
Wait, you said "obviously, they did something wrong if their economy collapsed" - but who's they? The government?
I thought the solution was no government involvement at all?
There's a big problem with no government oversight, and that's accountability. With no controls, no oversight, and no laws - nobody goes to jail when these big companies screw people into the ground.
Predicting a big failure in a system is easy. I predict that the Golden Gate bridge will someday collapse or be town down. It might not happen for 200 years but it will collapse I promise you!
It would be like another attack on the USA two months after Obama is elected; one that was planned for five years. Would that give McCain the right to say "TOLD YOU SO!!"
Coming up with a reasonable solution is entirely different to predicting failure, and it doesn't prove anything. Saying "This economic crisis is BECAUSE of the government!" can't be proven, and indeed this particular issue could actually prove otherwise. Because there wasn't ENOUGH regulation, this happened.
I really hate this type of "pop economics" crap that comes up whenever there's any problems. Free markets require regulation because individuals need to be protected. This isn't a strip mall where I can choose to buy a loaf of bread from three different stores, where it makes them set prices to compete. This is big banks, big economics, and we as individuals do not have the ability to force companies to compete for our business simply by choosing one over another. It's much too complicated and big, and these companies have too much power over our economy and therefor country to be left to their own devices.
Yea because we need an entire data center up and running to service all 20 million people in Australia. What is it with the Australians anyways? They're always bitching because they're not considered some huge superpower that everyone caters to. Get over it, you don't have the population. Australia is great but it's very small (in population.)
Australia has only slightly double the population New York City.
The idea was that you don't need to keep a huge web farm running outside of US/UK timezone areas because usually, an English web site will be hit during those times and mostly idle at other times.
Sure, but that's just trying to rationalize a dumb argument that every web site is used all over the world, and that's simply not true. English speaking sites usually only get hit from the North America and the UK regions.
Software RAID is a lot more flexible. You can plug the disks into ANY machine running the right OS and get your data up and running.
You can also RAID just PARTS of your disk, like with Linux LVM.
You can get new features with a software patch.
There's a lot of software out there to recover data from failed Windows software RAID sets.
Versus hardware, where you have a single vendor that you need to deal with to help you, and that you might need to acquire the same exact model controller if yours fails. This could be a problem if your controller is several years old.
Hardware RAID is a lot faster (in most cases) but that's really the ONLY benefit of it.
Well, because it's still too expensive for a lot of businesses to go full RAID 10 on their main storage system.
The disks you buy at NewEgg are cheap, but the disks you buy for your SAN are not as cheap. They might be the same disks, but that's just the way it is. And, the big costs come in the form of cost per slot, not necessarily the disk that plugs into it.
RAID-5 doesn't suffer from any real performance issues. Not for the last 10 years, anyway. Read speed is as fast as a stripe set, and write performance hits are easily mitigated by on-board cache. I kinda thought this was common knowledge..
Replication can be done to a much cheaper unit or DAS, and/or can be sent to an off-site location for better recoverability in the case of a real disaster.
Daily KOS doesn't report anything that CNN doesn't. Daily KOS isn't exactly a Blog. It's an entire web site, not just some guy or group of people ranting away like most of them.
Ohh, wait. Do you actually think that all web sites are blogs?
I'm not saying that Blogs aren't interesting to read sometimes or that they are no good at all, but they aren't a reliable news source and they WON'T replace main media.
There's no revolution here. Web news simply adds to the overall picture, just like TV did - except that Blogs are notoriously biased and/or extremely poorly researched (or not researched at all.)
I know it's popular around the Intarwebs to accuse the so-called "main stream" media of being nothing but crap, but it's just not true.
You have to get your news from multiple sources, but I'll trust the so-called mainstream media sources a lot more than "Bob's Obama Blog" or "Julie's McCain Blog."
There's some good blogs, yes. But if you notice, a lot of those people ALSO work with the larger media outlets or newspapers also.
It's better than some random blog. Blogs are - nearly all of them - just random musings from every day people. Usually, they will have a severe bias one way or another, and never (practically speaking) check facts.
No single news source should be taken as 100% accurate or unbiased, but I do trust that what they're telling me on CNN is accurate to the best of their knowledge, or that they big newspapers (sans editorials) will try their best to make sure that they are printing is accurate.
But hey, if you want want to get all your news from angry bloggers spewing nonsense with absolutely zero credibility, go for it! I'm sure you can get your conspiracy theory fix really easy there.
Wow, you really don't know what you're talking about!
MTBF is mostly marketing. It's not a guarantee, and drives will go above or below the MTBF. There's no possibly way for a manufacturer to know when these drives will fail. It's an estimate. They give "high end" disks a higher number but the desktop drives are built with exactly the same quality control.
"Server" drives are absolutely "bleeding" edge. The reason they don't have as big of capacity (usually) is because of the spindle speeds. Your average server disk is a 10k or 15k disk. It's more difficult to make higher capacity disks with spindle speeds that high. So, you end up with faster spindles, and lower capacity. As the manufacturing process improves, you see larger capacities in higher spindle speeds.
SAS and SATA use the same connector and are electrically compatible. I guess you didn't know that. So there goes that theory.
I don't love the fragility of the SATA/SAS connector but I love the interface. By doing what they did for standard connector between disks, standard positioning on the disk itself, and electrically compatible means that almost all SAS controllers and SAS drive enclosures can also use SATA disks. You can have one array with SAS, one array with larger SATA disks, all in one enclosure with one controller.
The warranties aren't always better. Seagate, for instance, warranties all of their desktop drives for 5 years, and they have an excellent returns system.
I have no idea what you're talking about for data recovery. No hard drive manufacturer I know of offers data recovery for warranty work. They warranty the part, not the data - it's up to YOU to provide yourself with recoverability of failed disks.
I wouldn't use ExcelStore drives, but I would, and DO, use Seagate, Samsung, and Hitachi SATA drives in production.
It seems as though you've picked up on the "common" thoughts about hard drives and you believe it as fact. It's just not. Read up on this stuff and don't rely on Internet people so much for the facts.
Ohh that's utter bullshit, and even more so with your original FUD "open-to-corruption-and-complexity" bullshit line.
With a $400 SATA RAID card I can achieve 250MB/s with six 750GB SATA disks (RAID5) - copying to a 5-disk 500GB SATA array (RAID5) on the same controller. My CPU time is nearly zero. The SAME CONFIGURATION with software RAID and JBOD (the original setup of this machine) were 1/2 the performance and 40% CPU time on the quad-core Opteron.
Software RAID uses a lot of resources that are better offloaded to something else if you can. So with your little home server you can get decent speeds, awesome.
I don't think I'll be recommending that all our clients switch to software RAID because of some anecdotal evidence.
The most common new SAS disks are 300GB. 400GB FIber Channel disks are also gaining a lot of popularity, and I can't tell you how many people are using 500GB - 1TB SATA disks for more "bulkish" type storage.
The only reason why the SAS/FC disks are smaller in capacity is because of the rotational speed of the spindle. It's a lot more difficult to make a reliable 10k or 15k disk at high capacities. The tolerances are unbelievable. As the processes get better, as the machines get better, they are able to produce higher capacity, high speed disks.
IOPS are more important than capacity of an individual disk in many cases, and the faster the rotation the faster the disk can find and retrieve your data.
If Seagate or Hitachi made a 1TB 15K disk you better believe people would be buying the hell out of them.
NewEgg has some 8 port SATA controllers (non RAID) for pretty cheap. I recently purchased one and it was a PCI-X card that also worked fine in a normal PCI slot. The card is nice and fast even over PCI (because PCI isn't THAT slow.)
I originally went software RAID too. I did this because what happens if the RAID controller dies? That's a big problem if it's 5 years from now and I can't get the same controller anymore.
The big reason I went with HW was not only the performance (which wasn't excellent under Windows) but the fact that if the machine doesn't shut down cleanly the machine will have to do an entire rescan of the array. With a 7 disk 750GB SATA array, this was taking about 20 hours. This happens on Linux too. It's normal for software RAID but can be avoided on HW RAID.
I don't have a complete battery and generator power solution, so I got a HW raid card. The Accusys cards actually have a little bit of flash on them that can keep track of things so if you power off it won't have to rebuild, even without a battery unit.
Software RAID can be a lot more flexible than HW RAID but HW RAID has a lot of benefits so I went that way.
A lot of RAID controllers do this on their own, too. I have a couple SATA RAID cards at home (Accusys) that will scrub periodically and make sure the parity is accurate. If not, it will find the offending drive, mark it as bad and alert you.
Most RAID controllers do that type of thing now. Modern controllers aren't just dumb parity calculators anymore.
Slashdot Mirror
I read it too, and it's true - they DO say all of the machines need access to the source which they do not.
Maybe there's some special cases, but I've never had to have a shared source repository in order to use distcc.
They also say the machines need to be exactly the same configuration, and they do elaborate on that a little bit, but it's not strictly true. Depending on the source you're compiling, you might only need to just have the same major version of GCC.
Yea. I used to use distcc a lot about five years ago. It doesn't help with all compile functions but it can help if you're compiling something big like X11 or KDE.
Blah blah blah Gentoo blah blah compiling forever blah blah punchline.
When I first used Gentoo several years ago it was with a 950Mhz Athlon CPU and it wasn't too bad.
Now, with quad-core CPU's becomming the norm even in Desktop machines, the compiling thing is even less of an issue.
I was able to run software on Gentoo that I could never get to run well together on any other distribution. You can almost always get the latest and greatest versions of everything with Gentoo. With Kernels taking almost no time to compile these days, the source distribution excuse just keeps getting weaker and weaker.
The main problem with Gentoo was it's moving target when it came to core system tools. They'd upgrade the ebuild system and it could sometimes hose your system. I haven't used Gentoo in a couple years because of it, but I'd be willing to give it another shot eventually.
No, you sarcastic prick.
I was a lego boy when I grew up. I had a lot of legos.
I also had some imitation blocked made by Tyco. These immitation blocks never fit together well. You'd build something and it would fall apart. Although the bricks looked almost identical, the Tyco bricks just sucked.
So, I do worry about imitation blocks. Lego blocks are the best because they have impossibly high standards during manufacturing in order to avoid the frustration I experienced with the Tyco blocks.
If someone else is going to start really making sets to compete with Lego, let's hope they go the distance and implement quality control like Lego does.
Wait, you said "obviously, they did something wrong if their economy collapsed" - but who's they? The government?
I thought the solution was no government involvement at all?
There's a big problem with no government oversight, and that's accountability. With no controls, no oversight, and no laws - nobody goes to jail when these big companies screw people into the ground.
Even a retard is going to be right SOMETIMES.
Predicting a big failure in a system is easy. I predict that the Golden Gate bridge will someday collapse or be town down. It might not happen for 200 years but it will collapse I promise you!
It would be like another attack on the USA two months after Obama is elected; one that was planned for five years. Would that give McCain the right to say "TOLD YOU SO!!"
Coming up with a reasonable solution is entirely different to predicting failure, and it doesn't prove anything. Saying "This economic crisis is BECAUSE of the government!" can't be proven, and indeed this particular issue could actually prove otherwise. Because there wasn't ENOUGH regulation, this happened.
I really hate this type of "pop economics" crap that comes up whenever there's any problems. Free markets require regulation because individuals need to be protected. This isn't a strip mall where I can choose to buy a loaf of bread from three different stores, where it makes them set prices to compete. This is big banks, big economics, and we as individuals do not have the ability to force companies to compete for our business simply by choosing one over another. It's much too complicated and big, and these companies have too much power over our economy and therefor country to be left to their own devices.
That was a well worded, well thought out post and I agree with you completely.
You shouldn't post Anonymous!
Yea because we need an entire data center up and running to service all 20 million people in Australia. What is it with the Australians anyways? They're always bitching because they're not considered some huge superpower that everyone caters to. Get over it, you don't have the population. Australia is great but it's very small (in population.)
Australia has only slightly double the population New York City.
The idea was that you don't need to keep a huge web farm running outside of US/UK timezone areas because usually, an English web site will be hit during those times and mostly idle at other times.
I figured it would be plainly obvious that by "UK Region" it would include Spain.
Douche.
Sure, but that's just trying to rationalize a dumb argument that every web site is used all over the world, and that's simply not true. English speaking sites usually only get hit from the North America and the UK regions.
I guess now we're both feeding the troll.
PS. The funny thing is that I actually agreed with your original post, but not about the failure rate of "enterprise" disks versus desktop disks.
You're a very ignorant person. I'm glad I don't know you in real life.
Are you kidding me?
Software RAID is a lot more flexible. You can plug the disks into ANY machine running the right OS and get your data up and running.
You can also RAID just PARTS of your disk, like with Linux LVM.
You can get new features with a software patch.
There's a lot of software out there to recover data from failed Windows software RAID sets.
Versus hardware, where you have a single vendor that you need to deal with to help you, and that you might need to acquire the same exact model controller if yours fails. This could be a problem if your controller is several years old.
Hardware RAID is a lot faster (in most cases) but that's really the ONLY benefit of it.
Well, because it's still too expensive for a lot of businesses to go full RAID 10 on their main storage system.
The disks you buy at NewEgg are cheap, but the disks you buy for your SAN are not as cheap. They might be the same disks, but that's just the way it is. And, the big costs come in the form of cost per slot, not necessarily the disk that plugs into it.
RAID-5 doesn't suffer from any real performance issues. Not for the last 10 years, anyway. Read speed is as fast as a stripe set, and write performance hits are easily mitigated by on-board cache. I kinda thought this was common knowledge..
Replication can be done to a much cheaper unit or DAS, and/or can be sent to an off-site location for better recoverability in the case of a real disaster.
Daily KOS doesn't report anything that CNN doesn't. Daily KOS isn't exactly a Blog. It's an entire web site, not just some guy or group of people ranting away like most of them.
Ohh, wait. Do you actually think that all web sites are blogs?
I'm not saying that Blogs aren't interesting to read sometimes or that they are no good at all, but they aren't a reliable news source and they WON'T replace main media.
There's no revolution here. Web news simply adds to the overall picture, just like TV did - except that Blogs are notoriously biased and/or extremely poorly researched (or not researched at all.)
I know it's popular around the Intarwebs to accuse the so-called "main stream" media of being nothing but crap, but it's just not true.
You have to get your news from multiple sources, but I'll trust the so-called mainstream media sources a lot more than "Bob's Obama Blog" or "Julie's McCain Blog."
There's some good blogs, yes. But if you notice, a lot of those people ALSO work with the larger media outlets or newspapers also.
It's better than some random blog. Blogs are - nearly all of them - just random musings from every day people. Usually, they will have a severe bias one way or another, and never (practically speaking) check facts.
No single news source should be taken as 100% accurate or unbiased, but I do trust that what they're telling me on CNN is accurate to the best of their knowledge, or that they big newspapers (sans editorials) will try their best to make sure that they are printing is accurate.
But hey, if you want want to get all your news from angry bloggers spewing nonsense with absolutely zero credibility, go for it! I'm sure you can get your conspiracy theory fix really easy there.
That's such crap. Weblogs don't equal news. There's almost NO reliable news sources in "blog" format.
It's all anecdotal, partisan, and un-researched crap.
You can't get any reliable news from blogs. It's a waste of time to try.
Blogs will never replace the normal news outlets and these "bloggers" are deluding themselves if they think they will.
Wow, you really don't know what you're talking about!
MTBF is mostly marketing. It's not a guarantee, and drives will go above or below the MTBF. There's no possibly way for a manufacturer to know when these drives will fail. It's an estimate. They give "high end" disks a higher number but the desktop drives are built with exactly the same quality control.
"Server" drives are absolutely "bleeding" edge. The reason they don't have as big of capacity (usually) is because of the spindle speeds. Your average server disk is a 10k or 15k disk. It's more difficult to make higher capacity disks with spindle speeds that high. So, you end up with faster spindles, and lower capacity. As the manufacturing process improves, you see larger capacities in higher spindle speeds.
SAS and SATA use the same connector and are electrically compatible. I guess you didn't know that. So there goes that theory.
I don't love the fragility of the SATA/SAS connector but I love the interface. By doing what they did for standard connector between disks, standard positioning on the disk itself, and electrically compatible means that almost all SAS controllers and SAS drive enclosures can also use SATA disks. You can have one array with SAS, one array with larger SATA disks, all in one enclosure with one controller.
The warranties aren't always better. Seagate, for instance, warranties all of their desktop drives for 5 years, and they have an excellent returns system.
I have no idea what you're talking about for data recovery. No hard drive manufacturer I know of offers data recovery for warranty work. They warranty the part, not the data - it's up to YOU to provide yourself with recoverability of failed disks.
Why don't you take a look at Google's hard drive study and see for yourself. http://research.google.com/archive/disk_failures.pdf
I wouldn't use ExcelStore drives, but I would, and DO, use Seagate, Samsung, and Hitachi SATA drives in production.
It seems as though you've picked up on the "common" thoughts about hard drives and you believe it as fact. It's just not. Read up on this stuff and don't rely on Internet people so much for the facts.
I'm actually quite familiar with RAID-like systems. I've been using Usenet for a long time =)
I've actually read the second link before, but I'll read the first one just because I love this stuff.
Ohh that's utter bullshit, and even more so with your original FUD "open-to-corruption-and-complexity" bullshit line.
With a $400 SATA RAID card I can achieve 250MB/s with six 750GB SATA disks (RAID5) - copying to a 5-disk 500GB SATA array (RAID5) on the same controller. My CPU time is nearly zero. The SAME CONFIGURATION with software RAID and JBOD (the original setup of this machine) were 1/2 the performance and 40% CPU time on the quad-core Opteron.
Software RAID uses a lot of resources that are better offloaded to something else if you can. So with your little home server you can get decent speeds, awesome.
I don't think I'll be recommending that all our clients switch to software RAID because of some anecdotal evidence.
Hey whatever. Go on with your bad self.
NTFS shares it's lineage with HPFS, which was part of OS/2 - originally a joint IBM and Microsoft venture.
That's quite untrue.
The most common new SAS disks are 300GB. 400GB FIber Channel disks are also gaining a lot of popularity, and I can't tell you how many people are using 500GB - 1TB SATA disks for more "bulkish" type storage.
The only reason why the SAS/FC disks are smaller in capacity is because of the rotational speed of the spindle. It's a lot more difficult to make a reliable 10k or 15k disk at high capacities. The tolerances are unbelievable. As the processes get better, as the machines get better, they are able to produce higher capacity, high speed disks.
IOPS are more important than capacity of an individual disk in many cases, and the faster the rotation the faster the disk can find and retrieve your data.
If Seagate or Hitachi made a 1TB 15K disk you better believe people would be buying the hell out of them.
NewEgg has some 8 port SATA controllers (non RAID) for pretty cheap. I recently purchased one and it was a PCI-X card that also worked fine in a normal PCI slot. The card is nice and fast even over PCI (because PCI isn't THAT slow.)
I originally went software RAID too. I did this because what happens if the RAID controller dies? That's a big problem if it's 5 years from now and I can't get the same controller anymore.
The big reason I went with HW was not only the performance (which wasn't excellent under Windows) but the fact that if the machine doesn't shut down cleanly the machine will have to do an entire rescan of the array. With a 7 disk 750GB SATA array, this was taking about 20 hours. This happens on Linux too. It's normal for software RAID but can be avoided on HW RAID.
I don't have a complete battery and generator power solution, so I got a HW raid card. The Accusys cards actually have a little bit of flash on them that can keep track of things so if you power off it won't have to rebuild, even without a battery unit.
Software RAID can be a lot more flexible than HW RAID but HW RAID has a lot of benefits so I went that way.
A lot of RAID controllers do this on their own, too. I have a couple SATA RAID cards at home (Accusys) that will scrub periodically and make sure the parity is accurate. If not, it will find the offending drive, mark it as bad and alert you.
Most RAID controllers do that type of thing now. Modern controllers aren't just dumb parity calculators anymore.