The google survey appears to show that drives are happiest arround 35-40 celcius with failure rates increasing both sides of that band.
Of course there are a couple of issues with that data
Firstly the data comes from the drives built in sensors so if a particular brand of drive has both an abnormal failure rate and an abnormal reported running temperature (either due to producing a different ammount of heat or due to a bad sensor) it would skew the results.
The second problem is they simply don't have much data outside the range 25-45 celcius.
Still I don't think it's a huge issue as long as you don't try to run your datacenter insanely hot.
Antartica would indeed not be a good choice but afaict there are places with temperatures low enough that you could use outside air to cool stuff year round while not being so low as to cause major logistical problems.
Well... if that's the case, it shouldn't get too far in court since just about every ethernet implementation from the past ~5 years has been crap integrated ethernet based on host signal processing Do you have a source for that outlandish claim? PCI isn't even really fast enough to handle the data from gigabit ethernet, let alone handle raw samples off the line. Even at 100 megabit it would be putting a HUGE load on the bus.
Winmodems were only practical because the data rates for a POTs modem are so low.
Uh, no. Range is not a function of signal power. Range is a measurement of how far you can go before path loss (note: many people think path loss is inverse square, this is only true in free space, there are a few situations where it is better and a load where it is much worse) reduces the signal to a level where the SNR is too bad to receive it. Crank up the transmit power and you increase the recieved signal power and hence improve the SNR at any given distance.
An antenna system with gain can both concentrate and attenuate signals, depending on the directivity and where its pointed. Indeed it can, what this means is that antenna gain only helps you if you know something about the relative location of transmitter and reciever and can reasonablly arrange for them to be oriented correctly.
Afaict what tends to happen is that they start with early retirements, that means the oldest slice of your workforce, possiblly the most experianced too. OTOH they were people you would probablly lose in a few years time anyway due to normal retirement.
Then they tend to go for "voluntary redundancies", basically anyone who leaves gets paid extra for doing so. This means all your most employable people go find another job. Even if there is no payment for leaving people who can leave are likely to do so because they can get a more secure job elsewhere.
Only if both of those fail to cull enough people do they go for compulsary redundancies.
Unfortunately most of these mutuals are converting into banks here in the UK. You are a bit behind the times, the wave of conversions happened about a decade ago.
XP X64 sucks Do you have a source for that claim, i've only run it briefly but the only real issue I found was driver availibility.
and it is most assuredly not the same thing on any level as Win2K3 server. The server specific functionality has of course been stripped out and the crippling adjusted but afaict the version of the major components is the same and it even uses the same hotfixes and service packs as the x64 version of server 2003.
Now, I have no idea what this gets you as the marketing stuff is vague -- it doesn't expressly list shared calendaring It seems to indeed be difficult to find good information on small buisness server but my understanding is it's a slightly crippled (mainly in terms of allowed user counts) package of microsofts various server products. There is a version of exchange in there so i'd be very surprised if it didn't support shared calendering.
At Amazon "Microsoft Small Business Server Standard 2003 R2 32-bit for System Builders" (whatever that means) costs $460 for a five user license. Additional users cost $50 or $60 each, or $150 in bundles of five. Now, I have no idea what this gets you as the marketing stuff is vague -- it doesn't expressly list shared calendaring (killer business app.). Throw in a crappy $540 computer which is noisy, bulky, and power hungry, and you almost have the equivalent of the Mini Server. OTOH that cheap and cheerful PC will almost certainly take desktop hard drives (which are available in higher capacity and afaict generally give better performance than the laptop drives in the mini) and will probablly have room for more than two of them.
I guess apple is hoping to get people hooked with the mini server and then get them to upgrade to the mac pro or xserve later.
None of this changes the fact that apple does not have any machines in the price range commodore64_love was asking about.
P.S. when the shit hits the fan on OS-X server can you sort things out using the GUI or do you need unix skills?
True but the rarer a phenomenon is the bigger the sample size you need to quantify it.
Suppose we could tell whether or not there is life in a hundred thousand star systems (I don't think we are anywhere near that yet) and the chance of a star system developing life is one in 10 million. We would be far from alone in the universe and yet we would also be unlikely to spot one of the other star systems with life.
Question - anyone with the 1 hdd version of the new Mac Mini know if it is possible to mount a second drive in there? On the previous model it could be done with the right adaptors (you had to sacrifice the optical drive just like the server mini does) but it's a bit messy. Google "ifixit 1TB mini" for details. I suspect the new refresh of the desktop model will be the same (but we will have to wait until someone tears it down to know for sure).
A $599 Mac Mini + $100 drive seems like a better deal than a $999 Mac Mini. The $599 mini only comes with a 160GB hard drive though, it also comes with a slower processor and less ram. If you compare the server model to the desktop model with the 500GB drive and the same processor and ram it's only $100 more.
Why not 640GB of 750GB drives? Two 750GB drives can be had for a little over $300. I'm pretty sure the mini needs 2.5 inch 9.5mm high drives (though I haven't seen anyone tear down this new model yet). If i'm right that rules out the 750GB and 1TB drives. 640GB drives a possibility though afaict. I've noticed in general apple doesn't like to ship the biggest drives on the market, maybe they can't get them in sufficiant quantity or somthing.
And it also provides apple with a nice way to get small buisnesses hooked on OS-X server, then when they outgrow the server mini they will hopefully spend the big bucks on a mac pro or xserve.
Even a raid without hotswap has a few big advantages over non-raid
1: it saves the data created between the last backup run and when the drive fails. 2: it lets you schedule the downtime for off hours 3: it probablly means less downtime than a restore from backup
I agree raid isn't perfect since it only protects from drive failures and as you say there can be problems with the second drive failing during rebuild (though i'd like to see some stats on this, how often does the second drive in a mirror really fail during rebuild?) but I still think even without hotswap it's much better than a single drive.
Every other small form factor desktop i've seen has been either far less powerful than the mac mini (e.g. the asus EEEbox, that dell you just linked), far larger (e.g. the shuttle series) or both.
The difference is even more pronounced now apple has made the option to have a second hard drive instead of the optical drive official (it was previously available only as a third party hack). I personally don't fancy setting up a server without raid, especially in an environment where backups are likely to be infrequent.
No, mac's aren't price competitive with ordinary PCs of similar spec, you pay a large premium for apples style and miniaturisation. These new releases don't change anything in that regard.
Depends on your requirements, afaict the eeebox only supports a single internal hard drive (and I don't think it has esata or firewire either so you are left with shitty USB if you want a second drive for raid). This new server mini supports two hard drives (the previous gen mini could also be hacked to support this but it's nice that apple have made it official). The mini also has a much better processor (which you say is not important to you, fair enough doesn't mean it isn't important to anyone)
In terms of bang per cubic centimeter the mac mini is pretty hard to beat.
As always there are trade-offs, the eeebox is small and cheap but not powerful. The mini is small and reasonably powerful but not particularly cheap. A bottom of the range dell vostro has a price comparable to the eeebox and specs comparable to the base model mini but isn't small.
Part of the value is that the Mac mini Server is only $100 more than the standard mini equipped with a single 500 GB drive, when OS X Server costs $500 on its own. I think the server model is the only way to get the mini with two hard drives instead of an optical drive. I consider raid to be an important feature for servers (yes I know raid isn't perfect and shouldn't be used as a substitute for backup but realisitically there is a limit to how often backups will actually get done so it's good to have something to cover the time since the last one) so it may be worth it even if you don't plan to use the supplied OS.
I wonder how apple has done the second hard drive, unfortunately I can't seem to find a teardown of this machine.
Star trek voyager was a series about a starfleet ship that due to a dying godlike being called the "caretaker" had ended up stuck an extremely long way from home.
Re:The web server can finally serve large files
on
OpenBSD 4.6 Released
·
· Score: 1
Torrent and FTP are much more efficient FTP wastes server resources and complicates interactions with firewalling and NAT by using seperate control and data connections. FTP and HTTP both have resume functionality nowadays.
Torrent is designed for peer to peer distribution, of peices this can save the server a lot of bandwith but also adds a lot of checking overhead and is somewhat controversial.
Unreliable networks can make a net-based installation process drag on and on or even freeze. That is more likely a case of poorly chosen timeouts and retry logic than any fundamental problem with the http protocol.
While I don't know for sure I suspect slashdot story urls are generated automatically from the article title by running it through some kind of special character striper routine which is being a bit more aggressive than it needs to be.
In this case we're all hoping the problem gets fixed before Malware writers get a hold of the idea of stealth installing plugins! I was under the impression that they already had got that idea and that was a large part of the reason the blacklist system was present in the first place.
Slashdot Mirror
The google survey appears to show that drives are happiest arround 35-40 celcius with failure rates increasing both sides of that band.
Of course there are a couple of issues with that data
Firstly the data comes from the drives built in sensors so if a particular brand of drive has both an abnormal failure rate and an abnormal reported running temperature (either due to producing a different ammount of heat or due to a bad sensor) it would skew the results.
The second problem is they simply don't have much data outside the range 25-45 celcius.
Still I don't think it's a huge issue as long as you don't try to run your datacenter insanely hot.
Antartica would indeed not be a good choice but afaict there are places with temperatures low enough that you could use outside air to cool stuff year round while not being so low as to cause major logistical problems.
Just make the building well insulated and then have controlled fans to bring in just enough outside air to keep the temperature where you want it.
Out of interest what brand was the computer in question?
Have you tried taking ownership of the files on the d: drive via the GUI?
Well... if that's the case, it shouldn't get too far in court since just about every ethernet implementation from the past ~5 years has been crap integrated ethernet based on host signal processing
Do you have a source for that outlandish claim? PCI isn't even really fast enough to handle the data from gigabit ethernet, let alone handle raw samples off the line. Even at 100 megabit it would be putting a HUGE load on the bus.
Winmodems were only practical because the data rates for a POTs modem are so low.
In my experiance outdoor antennas are usually a lot larger with a lot more elements than set top ones and presumablly this means they are higher gain.
Uh, no. Range is not a function of signal power.
Range is a measurement of how far you can go before path loss (note: many people think path loss is inverse square, this is only true in free space, there are a few situations where it is better and a load where it is much worse) reduces the signal to a level where the SNR is too bad to receive it. Crank up the transmit power and you increase the recieved signal power and hence improve the SNR at any given distance.
An antenna system with gain can both concentrate and attenuate signals, depending on the directivity and where its pointed.
Indeed it can, what this means is that antenna gain only helps you if you know something about the relative location of transmitter and reciever and can reasonablly arrange for them to be oriented correctly.
Afaict what tends to happen is that they start with early retirements, that means the oldest slice of your workforce, possiblly the most experianced too. OTOH they were people you would probablly lose in a few years time anyway due to normal retirement.
Then they tend to go for "voluntary redundancies", basically anyone who leaves gets paid extra for doing so. This means all your most employable people go find another job. Even if there is no payment for leaving people who can leave are likely to do so because they can get a more secure job elsewhere.
Only if both of those fail to cull enough people do they go for compulsary redundancies.
Unfortunately most of these mutuals are converting into banks here in the UK.
You are a bit behind the times, the wave of conversions happened about a decade ago.
XP X64 sucks
Do you have a source for that claim, i've only run it briefly but the only real issue I found was driver availibility.
and it is most assuredly not the same thing on any level as Win2K3 server.
The server specific functionality has of course been stripped out and the crippling adjusted but afaict the version of the major components is the same and it even uses the same hotfixes and service packs as the x64 version of server 2003.
Now, I have no idea what this gets you as the marketing stuff is vague -- it doesn't expressly list shared calendaring
It seems to indeed be difficult to find good information on small buisness server but my understanding is it's a slightly crippled (mainly in terms of allowed user counts) package of microsofts various server products. There is a version of exchange in there so i'd be very surprised if it didn't support shared calendering.
At Amazon "Microsoft Small Business Server Standard 2003 R2 32-bit for System Builders" (whatever that means) costs $460 for a five user license. Additional users cost $50 or $60 each, or $150 in bundles of five. Now, I have no idea what this gets you as the marketing stuff is vague -- it doesn't expressly list shared calendaring (killer business app.). Throw in a crappy $540 computer which is noisy, bulky, and power hungry, and you almost have the equivalent of the Mini Server.
OTOH that cheap and cheerful PC will almost certainly take desktop hard drives (which are available in higher capacity and afaict generally give better performance than the laptop drives in the mini) and will probablly have room for more than two of them.
I guess apple is hoping to get people hooked with the mini server and then get them to upgrade to the mac pro or xserve later.
None of this changes the fact that apple does not have any machines in the price range commodore64_love was asking about.
P.S. when the shit hits the fan on OS-X server can you sort things out using the GUI or do you need unix skills?
True but the rarer a phenomenon is the bigger the sample size you need to quantify it.
Suppose we could tell whether or not there is life in a hundred thousand star systems (I don't think we are anywhere near that yet) and the chance of a star system developing life is one in 10 million. We would be far from alone in the universe and yet we would also be unlikely to spot one of the other star systems with life.
Question - anyone with the 1 hdd version of the new Mac Mini know if it is possible to mount a second drive in there?
On the previous model it could be done with the right adaptors (you had to sacrifice the optical drive just like the server mini does) but it's a bit messy. Google "ifixit 1TB mini" for details. I suspect the new refresh of the desktop model will be the same (but we will have to wait until someone tears it down to know for sure).
A $599 Mac Mini + $100 drive seems like a better deal than a $999 Mac Mini.
The $599 mini only comes with a 160GB hard drive though, it also comes with a slower processor and less ram. If you compare the server model to the desktop model with the 500GB drive and the same processor and ram it's only $100 more.
Why not 640GB of 750GB drives? Two 750GB drives can be had for a little over $300.
I'm pretty sure the mini needs 2.5 inch 9.5mm high drives (though I haven't seen anyone tear down this new model yet). If i'm right that rules out the 750GB and 1TB drives. 640GB drives a possibility though afaict. I've noticed in general apple doesn't like to ship the biggest drives on the market, maybe they can't get them in sufficiant quantity or somthing.
And it also provides apple with a nice way to get small buisnesses hooked on OS-X server, then when they outgrow the server mini they will hopefully spend the big bucks on a mac pro or xserve.
Even a raid without hotswap has a few big advantages over non-raid
1: it saves the data created between the last backup run and when the drive fails.
2: it lets you schedule the downtime for off hours
3: it probablly means less downtime than a restore from backup
I agree raid isn't perfect since it only protects from drive failures and as you say there can be problems with the second drive failing during rebuild (though i'd like to see some stats on this, how often does the second drive in a mirror really fail during rebuild?) but I still think even without hotswap it's much better than a single drive.
Got any pics of this setup?
Every other small form factor desktop i've seen has been either far less powerful than the mac mini (e.g. the asus EEEbox, that dell you just linked), far larger (e.g. the shuttle series) or both.
The difference is even more pronounced now apple has made the option to have a second hard drive instead of the optical drive official (it was previously available only as a third party hack). I personally don't fancy setting up a server without raid, especially in an environment where backups are likely to be infrequent.
No, mac's aren't price competitive with ordinary PCs of similar spec, you pay a large premium for apples style and miniaturisation. These new releases don't change anything in that regard.
Depends on your requirements, afaict the eeebox only supports a single internal hard drive (and I don't think it has esata or firewire either so you are left with shitty USB if you want a second drive for raid). This new server mini supports two hard drives (the previous gen mini could also be hacked to support this but it's nice that apple have made it official). The mini also has a much better processor (which you say is not important to you, fair enough doesn't mean it isn't important to anyone)
In terms of bang per cubic centimeter the mac mini is pretty hard to beat.
As always there are trade-offs, the eeebox is small and cheap but not powerful. The mini is small and reasonably powerful but not particularly cheap. A bottom of the range dell vostro has a price comparable to the eeebox and specs comparable to the base model mini but isn't small.
Part of the value is that the Mac mini Server is only $100 more than the standard mini equipped with a single 500 GB drive, when OS X Server costs $500 on its own.
I think the server model is the only way to get the mini with two hard drives instead of an optical drive. I consider raid to be an important feature for servers (yes I know raid isn't perfect and shouldn't be used as a substitute for backup but realisitically there is a limit to how often backups will actually get done so it's good to have something to cover the time since the last one) so it may be worth it even if you don't plan to use the supplied OS.
I wonder how apple has done the second hard drive, unfortunately I can't seem to find a teardown of this machine.
Star trek voyager was a series about a starfleet ship that due to a dying godlike being called the "caretaker" had ended up stuck an extremely long way from home.
Torrent and FTP are much more efficient
FTP wastes server resources and complicates interactions with firewalling and NAT by using seperate control and data connections. FTP and HTTP both have resume functionality nowadays.
Torrent is designed for peer to peer distribution, of peices this can save the server a lot of bandwith but also adds a lot of checking overhead and is somewhat controversial.
Unreliable networks can make a net-based installation process drag on and on or even freeze.
That is more likely a case of poorly chosen timeouts and retry logic than any fundamental problem with the http protocol.
While I don't know for sure I suspect slashdot story urls are generated automatically from the article title by running it through some kind of special character striper routine which is being a bit more aggressive than it needs to be.
In this case we're all hoping the problem gets fixed before Malware writers get a hold of the idea of stealth installing plugins!
I was under the impression that they already had got that idea and that was a large part of the reason the blacklist system was present in the first place.