What are the chances of getting a NIC with a duped MAC and how usefull would it be to you? I'd rather just use a BSD or Linux than consider having a duped NIC lucky. ; )
I know there are cards that can have their MAC address configured, but again, how usefull are 2 or more NICs with the same MAC address? Too much trouble to go to, making sure they're well enough away from each other, etc just to get some free XP : )
I personally hope that this harms M$ more than they thought it would help them.
more powerful OS that requires me to manually configure everything
Tom, I started out with Red Hat 5.0. I heard this type of gripe in the newgroups, etc back then, but when I installed RH5 I was shocked that getting it going was'nt as bad as I'd heard. My Matrox G200 did'nt work too well, I think 5.1 fixed that, but as far as setting everything up goes, it was OK. Even my ISA AWE32 was easy to set up, with Linus pronouncing Lee-Nooks Lee-Nooks.
But now!... (admitedly I don't use the AWE32 any more) if you have PCI and AGP gear, you should be fine. I never need to config any hardware besides video setup for X, which consists of choosing my card, monitor refresh rates, and desired resolutions. My PCI Sound (MaestroII), 10/100 NIC (Tulip), Mouse (Logitech optical PS/2+USB), Ultra SCSI (Adaptec), etc all get probed and configured at boot time and just work.
Choose the right distro and ALL of it (inc X+video card) will get probed and set up during install. Mandrake would be a nice place to start for a newbie (hell, I like the fact that Mandrake lets you set up and install to software RAID!). I have settled on Debian, though I think Mandrake is the go for newbies who want to get into some quick action to see what Linux has got to offer before deciding to put in the effort to learn what's under the hood. I'd advocate Progeny at the moment if it were'nt for the fact that I had problems with the slightly flakey installer and since it might be worth waiting for Progeny to melt back into Debian.
Enjoy!
PS, the PS/2 mouse port is/dev/psaux, COM1+2 are/dev/ttyS0 and/dev/ttyS1, LPT1 is/dev/par0. IDE drives are/dev/hd*# where * is a (Primary Master), b (Primary Slave), c (Secondary Master), d (Secondary Slave) and # is the partition number on that drive. So/dev/hdc8 would be the 8th partition of the Secondary Master IDE drive. If X dies, try Ctrl-Alt-BackSpace or Ctrl-Alt-F(1-6) and always make a habit of shutting down gracefully (shutdown now from a prompt or xdm should give you an option). wvdial is nice for getting on the net and if you have trouble otherwise with ppp, defaultroute might be something you need. ; )
My system shouldn't crash at all. I have brand name components
ROFL! Yeah, and you have a brand name OS, you say it should'nt and your point is? ; )
I've been using M$ OS since the late 80's, and up until about 1995 I thought that system crashes were just that, but now that I've been using Linux on those same systems for about 4 years, I now know that they were OS crashes. ; ) Linux has NEVER (or any of the BSD's for that matter) crashed on me in normal use in 4 years.
I have managed to make it crash once when I tried to force a hard drive into a UDMA mode and a CDROM on the same channel into a different mode, but besided that... rock solid stability. In contrast, the 4/8/16/24/32 {cough} MS OSes would crash once every day or two and NT every few weeks would chuck a bit of a fit. 2k seems better, but still nowhere near Linux, *BSD or some other decent OS.
Worse: this scheme is impossible to pull off in many commercial situation
Yeah, back when I first heard about the activation thing, I thought about all the medium to large size businesses and.edu environments that WON'T be upgrading to XP since they won't be able to employ quick large SOE rollouts or repairs with imaging through Ghost or whatever.
Surely M$ has a solution to that, and surely that solution will be H@x0r3D to death.
yeah, "no more rehearsing..." of course would be refering to Windows 1,2,3,NT,95,98,ME...
Somehow, I think they'll be singing that song again when Microsoft WinXP^2 comes out, in a feeble attempt to steal the desktop market back off OSX some stage in the future. ; )
I've been a Wintel user since the late 80's, but my next Arch/OS purchase will be PPC/OSX.
There is no way I am ever buying into M$ shit ever again.
I'm just curious how many security holes are going to exploited by hackers and crackers in Windows XP....they sure found enough in the earlier versions of windows to fill their time.
Maybe M$ theory is that if they keep churning out bugs and holes, the hackers will eventually just get bored and stop hacking, for a lack of challenge.
Please, please, please tell me Sting is NOT singing for M$! If he is, I will sell all my Sting CD's and keep my LAME 256k MP3's of them.
Fuck, first Metallica and now Sting, Van Halen said some nasty stuff about Matchbox 20(?) and now Linus says something along the lines of "I have'nt really looked at XP or FreeBSD, (duh) but I don't see anything good in them".
Amen brother! I grew up through the 70's and 80's, and heard plenty of my Mums 60's music, lots of really great stuff in that 30 years (I'm blocking most rap out, it's a sanity mechanism), then the 90's started to get crap and now we're well into the nothings it seems most music is really crap.
It would be OK though, if it were'nt for the fact that talentless morons insist on churning out the old music that I like with their fucking shit moans and grunts with their PC's, set to a gen-x feel that apparently is more enlightened than the uncool group that originally authored the tunes and lyrics without bloody sound sampling.
Give us a break! These new bullshit bands should receive the death penalty or XP or something.
I mean, patching 8 libs together to get one program to compile is fun and all, but I think I'll go with a defined environment....
You're going to Deb-E-An?
Or will you be going....
cd/usr/ports/wizbang/app
make && make install ???
Ahh, to hell with it! Just go with the company that makes you feel nice and comfortable with a pretty install and ease of use as long as you pay them big bucks, while their OS tick tick tick ticks... with it's security holes and legacy bugs.
Are you a RedHat user who thinks Linux is RedHat is Linux?
I see where you're coming from and would'nt argue with your rationale. Hell, I am using 2 Seagate UDMA 20Gb 7200rpm Barracuda's with Debian-current set up with/,/usr,/home and/tmp striped in RAID-0 and ReiserFS.
But, when I say "high I/O work", I'm not talking about a reliance on high transfer rates, I'm talking about huge quantities of perhaps small reads and writes, a'la database type loads, etc.
I personally guarantee, that under those "high I/O" conditions, the comparable (RAID level, drive access speed, drive transfer rate, # of drives) SCSI RAID will kill the IDE RAID severely.
I'll try to make an example.
SCSI: You and some friends are in your bedroom and want to get some stuff from the kitchen, where another friend is waiting to help. The whole bunch of you go to the kitchen and ask for (in this order): fork, plate, knife, coke, chips, OJ, fork, knife, plate, icecream, spoon, bowl. Instead of the person in the kitchen moving about getting each item as he is requested, he gets the first item and then before he has finished getting it, he re-orders the requests to make his movements more efficient. So he gets it all in this order: fork, fork, knife, knife, spoon, plate, plate, bowl, coke, OJ, icecream, chips. Then all the friends walk back to the bedroom with their stuff.
IDE: You and some friends are in your bedroom and want to get some stuff from the kitchen, where another friend is waiting to help. Unfortunately only one person is allowed to leave the bedroom to go to the kitchen and the next person cannot leave until the first person comes back. So the first person goes to the kitchen, asks the person in the kitchen for a fork, they get it and the fork is taken back to the bedroom. Then the next person leaves (while the person in the kitchen is doing nothing, waiting) and asks for a plate, etc. Next person asks for a knife, etc. This goes on for a few bits of cutlery, plates and food from the fridge and cupboard. All the while only ONE person is in transit at any one time and while they are in transit, the person in the kitchen waits, doing nothing. (Read ahead caching could help here, but still no match for SCSI smarts when there are huge numbers of small requests scattered all over the drive).
The problems with IDE could perhaps be helped a little with an IDE RAID controller with some smarts, but nothing can be done about the fact that an IDE *drive* is not capable of dealing with more than 1 request at any given time. This can be very limiting in some circumstances. Command re-ordering could be done with a controller but the IDE drive cannot work to it's best physical capability due to the fact that it can't queue requests.
IDE is of course great for home use, but keep it away from busy servers!
however listen to anything with acoustic instruments and even the CD standard itself isn't good enough to convey it without noticable loss.
I tend to beleive that the transducers involved (recording and playback) cause a lot more loss than the actual CD standard. With thermal noise and EMI noise getting amplified and phase reversals (the type of stuff BBE tries to get rid of), etc. Heard even on "flagship" equipment.
I don't doubt that noticable loss can occur with CD, I just don't think most of the fault falls upon the CD standard.
Years ago, I was hell bent on a Yamaha CDX-1060 CD player and a nice pair of Beyer Dynamic DT-911's. Since CD can restore a 20kHz sinewave, and that is about highest we can hear, and that player had fantastic SNR, low noise and high dynamic range and those phones sounded so good. But now, I make do with Lame 256k mp3's, hopefully soon through a decent soundcard (I have a Maestro II at the moment with unsatifactory noise levels, plus a GUS and AWE32, also with crap noise) and some good phones.
Sometimes, some audiophiles amaze me that they'll buy equipment based on reviews that contain no real tests. Some reviewers are certainly biased towards expensive equipment since they think it *must* be great. Take for example the $50,000 (au) Meridian CD player versus the $250 Marantz from a few years back. The Marantz sounds better, period. Test gear can prove it and so can ears, but the audiophile reviewer who can't see past all the gold, lead and teflon can't. : )
I usually notice pops, clicks, noise, etc where friends don't, but do I have golden ears, I doubt it, I'm happy with Lame 256k CBR, and in A-B tests absolutely cannot hear a difference. Sometimes I hear something whilst listening to an mp3 and think ah-ha! an artefact! and then listen to the CD only to find that it is there also. : )
If only more people would realise that paying a bit more gets them a longer lifespan on their products.
Hear hear! Look at the US Robotics Courier V.Everything.
Here is a MODEM that was brought out yonks ago, with a 20MHz Intel 80186 CPU (or 25MHz for the US model), 256k (from memory) flashable memory for firmware and a kick arse (for the intended application) TI DSP.
I bought it when it was just a "Courier Dual Standard V.34" 28.8k MODEM, then upgraded to V.34bis 33.6k with a small download from USR, then upgraded to X2 53.3k, then V.90 once it was ratified and then released by 3COM.
Sure, I paid $600 odd.au dollars for it, but I have NEVER regretted that purchase.
It comes with support for almost bloody everything on the DTE and modulation fronts:), is very customizable, has an excellent, huge, printed manual and excellent in-MODEM AT command summaries for checking stuff that is newer than the manual. It's the sort of product that you would brag about being involved in developing if you were. Those USR engineers should be very proud.
I set a pair up for the Aussie Stock Exchange for a broker connection through leased lines. I was able to set them up so that all you would have to do is plug them in (to DTE and leased line), switch them on and as soon as they could see each other whoosh, they'd connect in 3 or 4 seconds and then stay connected for YEARS.
It may seem out dated with broadband now, but it's as awesome initial design, given the interface limits (RS-232 and PSTN).They took the limits of RS-232 and PSTN into account when line rate technologies were at a much lower performance. And thats how these sorts of devices should be built.
For high I/O work, SCSI is best. Multiple requests can be issued to a drive, while it is working on another, and the drive can re-order those requests to reduce head movement and thus increase final throughput.
Add to that 15,000 rpm drives that could easily be limited by the highest ATA *burst* rates (especially with even a small (2 drive) SCSI RAID setup) and you'll see why SCSI has bus speeds up to 160MB/S.
I've had nothing but very impressive experiences with SCSI gear. My 16x SCSI CDROM is *much* faster than my 32x UDMA CDROM for example and does not choke on a simple dd! I've found most IDE gear to be of a low quality compared with most SCSI.
With the price of flat panel displays, RAM and Gigabit ethernet coming down to decent prices, I think the next homewide system for me will be home brewed thin client based.
1Gb RAM in each client to help releive reliance on network (hell in Sydney 1Gb of SDRAM is less than $au250 (4 * 256Mbyte DIMMS)!
With an AWARD BIOS equiped motherboard, I can easily flash an Etherboot ROM image to the mobo BIOS thanks to AWARDS modular design (so no need to burn ROMS for the NICS).
With Gigabit ethernet, when the network is relied upon for block device usage, ~120MBytes/S is very nice thank you very much...
from the server with 160MByte/S SCSI and also 1 or more gig of RAM. Urgh urgh urgh.
With Thunderbirds a plenty and each client with DVD-ROM. ; )
So I can have *silent*, diskless, flatpanel machines that are fast in both bedrooms and the lounge. With one central server with all our albums MP3'ed (145 so far done), central storage for video recorders like Tivo in the lounge and bedrooms also, and allowing quick easy backups of user ~ dirs.
Then add in some IP telephony to allow intercom and speakerphone transfers of incoming calls. All I need is more money!!!
I already have the US Robotics Courier V.Everything. I was sold on it YEARS ago with it was just a V.34, then upped it to V.34bis, then X2 and now V.90, talk about over engineering! What an awesome machine. Anyone know of any undocumented stuff or hacking info for the Courier V.Everything? I liked the old days when I had a MODEM dialing with a DTMF duration of 30mS. It sounded so cool, dialing up BBS' that quick (especially attack-dialing to get onto a BBS line once available), but my Courier won't allow that speed, it is limited to a longer length to appease Telstra/Austel.
Speaking of which, I just opened the pdf version of the FreeBSD Handbook in Acrobat 4.05 on my PII-300 256Mb RAM and striped swap (across 2 7200rpm UDMA drives) with kernel 2.4.10, then all of a sudden, while browsing through the docco, thrash thrash thrash, I had flashbacks to my Windows days (oh God the humanity).
My mouse cursor barely responded to movement, I managed to get my mouse over to close Acroreader but clicking seemed out of the question. I managed to get to a console prompt to see what top had to say when finally the thrashing stopped and my once responsive system was back to some sort of sane level.
WTF!?!? This is a first for me in my 4 years with Linux.
Sawfish died also and now won't restart. I wan't my AquaX buttons back! Damn you Rik!!
All I have said, is that not all overclocking causes the PCI bus to be overclocked and that I beleive that the thought of MOST hdd failures being caused by overclocking, is silly. Of course it IS possible to overclock your PCI bus in the pursuit of system overclocking.
I thought that I said this, but in case I didn't: You changed the FSB and the multiplier.
Yeah, what is your point? The PCI is still at 33MHz. I used a standard BIOS setting for the FSB and CPU multiplier. Choosing a 66FSB gives a PCI multiplier of 2 automatically, choosing a 100FSB gives a PCI multiplier of 3 automatically. They are standard increments that I am choosing that cause appropriate PCI speeds, not marginal increments like 112 that cause the PCI to be raised to 37MHz for example. It is preferable that I don't overclock the PCI or AGP bus as I am a beta tester for a card mfg that prefers results based on within-spec hardware.
Can you change your FSB from 100 now to 112 without changing the PCI bus speed? Or from 66 to 75?
No. But that does'nt disprove my point that overclocking can be done without changing the PCI speed, it merely proves that the PCI speed can be overclocked as a result of system overclocking, something I would never dispute, since I've been overclocking buses since before the existance of PCI or even VLB (My old SpeedStar 24X supported 16MHz ISA, but I could only go to 13MHz while being stable).
Thats fine, but do you have any facts or information
Hey, I'm not the one here making wild claims like "by far the most common reason [for hdd failure] was due to overclocking ", so perhaps you should provide some facts or information to support those claims.
I just searched on the internet and found no evidence to support what you were talking about
And just what was it that I was talking about? 1. Not all overclocking causes the PCI bus to run above spec? or 2. I highly doubt that "by far the most common reason" for hdd failure is due to overclocking.
I also love online debating, I've been doing so since 1991 starting with FidoNet and I avoid having words put into my mouth or arguing against things I do not dispute.
Woah, sorry to say but you are very misstaken on that.
No I'm not. Choosing say a standard 300MHz CPU option in the BIOS, will yeild a 33MHz PCI, 66MHz AGP, 66MHz FSB and a CPU multiplier of 4.5, or possibly 33MHz PCI, 66MHz AGP, 100MHz FSB and multiplier of 3, depending on whether the CPU is detected as a 66 or 100 FSB type. I had to open my PII 300 to doctor it a little for 100FSB.
Choose a standard option of 400 (PII), you should get 33MHz PCI, 66MHz AGP, 100MHz FSB and a multiplier of 4. These assume that the CPU's multiplier can be changed. And my point is, that overclocking CAN be done without increasing the PCI and AGP speeds. I have been doing it for many years.
Of course, some overclocking methods are achieved through increases that cause PCI, AGP and FSB to also increase. Are you trying to say that you can't overclock without increasing PCI? It is true that some hardware forces this method due to multiplier locks for example.
The thought that most hard drive failures occur due to overclocking, is absolutely ridiculous. The tech that said that, is either full of it or ignorant. I've been a hardware tech for 13 years and about 7 of that specifically with PC hardware (first with NEC then DEC).
Now the problem is that when you up the clockrate of your chip, you're really changing the BASE rate
That is NOT always the case. My CPU was a PII 300 with 66FSB. I opened it, cut a track, now it supports 100FSB. I choose in the BIOS 100 FSB, multiplier 3.5, and I get a PII-350 with 100FSB and the PCI runs at 33MHz and the AGP at 66MHz. So I overclocked my CPU WITHOUT increasing the the PCI or AGP speed. I can verify this with a CRO.
There are different methods of overclocking and there are multipliers in many places that allow for example 33PCI+100FSB or 33PCI+66FSB. It's too simplistic to say that all overclocking increases the PCI speed and downright silly to think that MOST hdd failures are due to overclocking! I have supported Australia's largest WAN (it's.edu) and the second largest (at the time), the ASX and a large legal firm. I assure you that NONE of the faulty hdds we got replaced were due to overclocking.
fires on the leading and falling edge of the signal
I've been dealing with positive and negative edge triggering in digital electronics since 1989 for the Royal Australian Navy, Electronic Weapons.
Thanks for elightening me on the IBM hdd head parking though, that is very neato.
Well, not so neccessarily... those IBM drives have their heads parked off the surface of the disk when powered off so you can shake'em around all that you like and won't damage the surface of the disk.
Can you provide a link to support this? I've been working in PC hardware repair since '94 (For NEC and DEC) and electronics since '89 before that. I've never seen any modern VC drive (non stepper) move the heads *off* the platters, they *always* move the heads to the inner most location of the platters, near the spindle, where head touch down is at the lowest surface speed (inner vs outer) and also required motor torque is the least when spinning up, since the heads are clamping down on the platters!
To move *off* the platters, would require some mechanism to transition the heads from off/on to the platters. I have'nt been watching hdd tech lately, so a link to this would be very interesting.
I like that and that was one of the reasons why I sprung the extra for the drive.
I hope you're right, for your sake. Not that I don't think IBM drives are not great.
I was told firsthand by a tech that by far the most common reason was due to overclocking.
You were told wrong. Loads of techs don't know shit. They come out, run a test from floppy, deem a drive OK or not, and if not replace it. They do the same kind of crap with printers, MODEM's, screens, etc. Most of them are good for what they do (replace hardware) and not much more.
Don't forget that running that wonderful 1.4GHz processor at 1.624GHz also overclocks your PCI bus by 4.5MHz.
It depends on how the overclocking is done. If you're just choosing standard CPU multiplier and FSB speeds from BIOS settings, or even jumpers, it should not change the PCI speed (but can). The small 1MHz tweaks that some BIOS support, can change the PCI speed also. You can also explicitely change the PCI speed too if you have the right mobo and read the manual. Just choosing different default CPU speeds should not though.
Doesn't sound like much? Well that's 14% above spec. After a while, and with the heat that these drives generate anyways, it's pretty easy to toast your electronics by doing that.
Possible, but definetely NOT the most common reason. I've been over clocking since it required soldering ; ), and I've found that hardware either does'nt work or becomes unreliable due to the overclocking. *Some* CPU's will die, though I've never killed an Intel. Having said that, some Intel PII's above 300MHz (and including some 300's), had faulty thermal diodes, which would allow destruction of those CPU's under overclocked settings. I've never killed a hdd due to overclocking.
Hell, I used to hot swap an old IDE WD Caviar (for years) to move big chunks of data before I got some 100base-T NICs. ; )
I like the old Caviars too. I had a WD Caviar 340Mb for about 7 years. Worked very nicely and was much faster than the crap Maxtor it replaced.
However, I upgraded (big time:) to two Seagate UDMA Barracuda 20Gb drives (Linux RAID-0+Reiser/,/usr,/home and/tmp. Urgh urgh urgh...). The *day* I plugged in one of the Seagates with the little Caviar as the slave.... boom, no more Caviar. Shit itself baddly, oddly somehow due to the Seagate? I don't know, I'm pretty happy with the Barra's though, going from about 1.2MB/s with the Caviar to ~28MB/s raw off the Barra's and 45MB/s Linux (2.2.x) RAID-0 was pretty sweet.
Of course now, with Linux 2.4.x, RAID-0 performance is SLOWER than the raw drive speed for some weird reason that I am still trying to track down. FreeBSD on the other hand, is consistently faster transfering data off these drives. I'll soon test FreeBSD's RAID-0 to see if I can get back my 40+ MB/s...
Slashdot Mirror
Yeah, it's not the wand, it's the wizard.
I guess that would explain Micro Soft?
These people would probably get the Corporate version which doesn't require Activation.
Do you know what mechanism is in place to prevent copying of that version? I am curious to know which will be most pirated.
Yeah, that's what I was getting at.
What are the chances of getting a NIC with a duped MAC and how usefull would it be to you? I'd rather just use a BSD or Linux than consider having a duped NIC lucky. ; )
I know there are cards that can have their MAC address configured, but again, how usefull are 2 or more NICs with the same MAC address? Too much trouble to go to, making sure they're well enough away from each other, etc just to get some free XP : )
I personally hope that this harms M$ more than they thought it would help them.
Definitely! I'm happy, you're happy, we're all happy! Except for maybe Osama:/bin/laden# mv / /dev/null.
.au uses a song that a once kinda liked. ; )
If only jingles on TV, etc did'nt rip off good 'ole songs. Target in
more powerful OS that requires me to manually configure everything
/dev/psaux, COM1+2 are /dev/ttyS0 and /dev/ttyS1, LPT1 is /dev/par0. IDE drives are /dev/hd*# where * is a (Primary Master), b (Primary Slave), c (Secondary Master), d (Secondary Slave) and # is the partition number on that drive. So /dev/hdc8 would be the 8th partition of the Secondary Master IDE drive. If X dies, try Ctrl-Alt-BackSpace or Ctrl-Alt-F(1-6) and always make a habit of shutting down gracefully (shutdown now from a prompt or xdm should give you an option). wvdial is nice for getting on the net and if you have trouble otherwise with ppp, defaultroute might be something you need. ; )
Tom, I started out with Red Hat 5.0. I heard this type of gripe in the newgroups, etc back then, but when I installed RH5 I was shocked that getting it going was'nt as bad as I'd heard. My Matrox G200 did'nt work too well, I think 5.1 fixed that, but as far as setting everything up goes, it was OK. Even my ISA AWE32 was easy to set up, with Linus pronouncing Lee-Nooks Lee-Nooks.
But now!... (admitedly I don't use the AWE32 any more) if you have PCI and AGP gear, you should be fine. I never need to config any hardware besides video setup for X, which consists of choosing my card, monitor refresh rates, and desired resolutions. My PCI Sound (MaestroII), 10/100 NIC (Tulip), Mouse (Logitech optical PS/2+USB), Ultra SCSI (Adaptec), etc all get probed and configured at boot time and just work.
Choose the right distro and ALL of it (inc X+video card) will get probed and set up during install. Mandrake would be a nice place to start for a newbie (hell, I like the fact that Mandrake lets you set up and install to software RAID!). I have settled on Debian, though I think Mandrake is the go for newbies who want to get into some quick action to see what Linux has got to offer before deciding to put in the effort to learn what's under the hood. I'd advocate Progeny at the moment if it were'nt for the fact that I had problems with the slightly flakey installer and since it might be worth waiting for Progeny to melt back into Debian.
Enjoy!
PS, the PS/2 mouse port is
If all else fails, get a Mac and OSX!
My system shouldn't crash at all. I have brand name components
ROFL! Yeah, and you have a brand name OS, you say it should'nt and your point is? ; )
I've been using M$ OS since the late 80's, and up until about 1995 I thought that system crashes were just that, but now that I've been using Linux on those same systems for about 4 years, I now know that they were OS crashes. ; ) Linux has NEVER (or any of the BSD's for that matter) crashed on me in normal use in 4 years.
I have managed to make it crash once when I tried to force a hard drive into a UDMA mode and a CDROM on the same channel into a different mode, but besided that... rock solid stability. In contrast, the 4/8/16/24/32 {cough} MS OSes would crash once every day or two and NT every few weeks would chuck a bit of a fit. 2k seems better, but still nowhere near Linux, *BSD or some other decent OS.
Worse: this scheme is impossible to pull off in many commercial situation
.edu environments that WON'T be upgrading to XP since they won't be able to employ quick large SOE rollouts or repairs with imaging through Ghost or whatever.
Yeah, back when I first heard about the activation thing, I thought about all the medium to large size businesses and
Surely M$ has a solution to that, and surely that solution will be H@x0r3D to death.
yeah, "no more rehearsing..." of course would be refering to Windows 1,2,3,NT,95,98,ME...
Somehow, I think they'll be singing that song again when Microsoft WinXP^2 comes out, in a feeble attempt to steal the desktop market back off OSX some stage in the future. ; )
I've been a Wintel user since the late 80's, but my next Arch/OS purchase will be PPC/OSX.
There is no way I am ever buying into M$ shit ever again.
I'm just curious how many security holes are going to exploited by hackers and crackers in Windows XP....they sure found enough in the earlier versions of windows to fill their time.
Maybe M$ theory is that if they keep churning out bugs and holes, the hackers will eventually just get bored and stop hacking, for a lack of challenge.
and Sting will be singing!!!!!
Please, please, please tell me Sting is NOT singing for M$! If he is, I will sell all my Sting CD's and keep my LAME 256k MP3's of them.
Fuck, first Metallica and now Sting, Van Halen said some nasty stuff about Matchbox 20(?) and now Linus says something along the lines of "I have'nt really looked at XP or FreeBSD, (duh) but I don't see anything good in them".
Is there anyone who is genuine out there?
No, I think that's just cause music sucks now.
Amen brother! I grew up through the 70's and 80's, and heard plenty of my Mums 60's music, lots of really great stuff in that 30 years (I'm blocking most rap out, it's a sanity mechanism), then the 90's started to get crap and now we're well into the nothings it seems most music is really crap.
It would be OK though, if it were'nt for the fact that talentless morons insist on churning out the old music that I like with their fucking shit moans and grunts with their PC's, set to a gen-x feel that apparently is more enlightened than the uncool group that originally authored the tunes and lyrics without bloody sound sampling.
Give us a break! These new bullshit bands should receive the death penalty or XP or something.
I mean, patching 8 libs together to get one program to compile is fun and all, but I think I'll go with a defined environment....
/usr/ports/wizbang/app
You're going to Deb-E-An?
Or will you be going....
cd
make && make install ???
Ahh, to hell with it! Just go with the company that makes you feel nice and comfortable with a pretty install and ease of use as long as you pay them big bucks, while their OS tick tick tick ticks... with it's security holes and legacy bugs.
Are you a RedHat user who thinks Linux is RedHat is Linux?
though I can't see why you couldn't install XP on several machines with the exact same hardware
/dev/hda for an example.
The MAC address in NICs are unique (theoretically and practically. Absolutely if not for a very few cheapo no-name crap cards).
Hard drives have serial numbers too, hdparm -i
Motherboards probably have unique serial numbers too, though I'm not completely sure.
Intel's famed id number?
SCSI controllers and devices, perhaps also?
RAM EEPROM data?
Video card serial number? I'm sure my Matrox G400 has one.
Surely they would just grab the hdd serial and generate from that, forcing the installed OS to follow that drive only.
I could'nt give a shit anyway, after far too many years with MS (starting with Windows 3.0), I am very glad I have discovered Debian GNU/Linux.
Hi James,
/, /usr, /home and /tmp striped in RAID-0 and ReiserFS.
I see where you're coming from and would'nt argue with your rationale. Hell, I am using 2 Seagate UDMA 20Gb 7200rpm Barracuda's with Debian-current set up with
But, when I say "high I/O work", I'm not talking about a reliance on high transfer rates, I'm talking about huge quantities of perhaps small reads and writes, a'la database type loads, etc.
I personally guarantee, that under those "high I/O" conditions, the comparable (RAID level, drive access speed, drive transfer rate, # of drives) SCSI RAID will kill the IDE RAID severely.
I'll try to make an example.
SCSI: You and some friends are in your bedroom and want to get some stuff from the kitchen, where another friend is waiting to help. The whole bunch of you go to the kitchen and ask for (in this order): fork, plate, knife, coke, chips, OJ, fork, knife, plate, icecream, spoon, bowl. Instead of the person in the kitchen moving about getting each item as he is requested, he gets the first item and then before he has finished getting it, he re-orders the requests to make his movements more efficient. So he gets it all in this order: fork, fork, knife, knife, spoon, plate, plate, bowl, coke, OJ, icecream, chips. Then all the friends walk back to the bedroom with their stuff.
IDE: You and some friends are in your bedroom and want to get some stuff from the kitchen, where another friend is waiting to help. Unfortunately only one person is allowed to leave the bedroom to go to the kitchen and the next person cannot leave until the first person comes back. So the first person goes to the kitchen, asks the person in the kitchen for a fork, they get it and the fork is taken back to the bedroom. Then the next person leaves (while the person in the kitchen is doing nothing, waiting) and asks for a plate, etc. Next person asks for a knife, etc. This goes on for a few bits of cutlery, plates and food from the fridge and cupboard. All the while only ONE person is in transit at any one time and while they are in transit, the person in the kitchen waits, doing nothing. (Read ahead caching could help here, but still no match for SCSI smarts when there are huge numbers of small requests scattered all over the drive).
The problems with IDE could perhaps be helped a little with an IDE RAID controller with some smarts, but nothing can be done about the fact that an IDE *drive* is not capable of dealing with more than 1 request at any given time. This can be very limiting in some circumstances. Command re-ordering could be done with a controller but the IDE drive cannot work to it's best physical capability due to the fact that it can't queue requests.
IDE is of course great for home use, but keep it away from busy servers!
however listen to anything with acoustic instruments and even the CD standard itself isn't good enough to convey it without noticable loss.
I tend to beleive that the transducers involved (recording and playback) cause a lot more loss than the actual CD standard. With thermal noise and EMI noise getting amplified and phase reversals (the type of stuff BBE tries to get rid of), etc. Heard even on "flagship" equipment.
I don't doubt that noticable loss can occur with CD, I just don't think most of the fault falls upon the CD standard.
Years ago, I was hell bent on a Yamaha CDX-1060 CD player and a nice pair of Beyer Dynamic DT-911's. Since CD can restore a 20kHz sinewave, and that is about highest we can hear, and that player had fantastic SNR, low noise and high dynamic range and those phones sounded so good. But now, I make do with Lame 256k mp3's, hopefully soon through a decent soundcard (I have a Maestro II at the moment with unsatifactory noise levels, plus a GUS and AWE32, also with crap noise) and some good phones.
Sometimes, some audiophiles amaze me that they'll buy equipment based on reviews that contain no real tests. Some reviewers are certainly biased towards expensive equipment since they think it *must* be great. Take for example the $50,000 (au) Meridian CD player versus the $250 Marantz from a few years back. The Marantz sounds better, period. Test gear can prove it and so can ears, but the audiophile reviewer who can't see past all the gold, lead and teflon can't. : )
I usually notice pops, clicks, noise, etc where friends don't, but do I have golden ears, I doubt it, I'm happy with Lame 256k CBR, and in A-B tests absolutely cannot hear a difference. Sometimes I hear something whilst listening to an mp3 and think ah-ha! an artefact! and then listen to the CD only to find that it is there also. : )
If only more people would realise that paying a bit more gets them a longer lifespan on their products.
.au dollars for it, but I have NEVER regretted that purchase.
:), is very customizable, has an excellent, huge, printed manual and excellent in-MODEM AT command summaries for checking stuff that is newer than the manual. It's the sort of product that you would brag about being involved in developing if you were. Those USR engineers should be very proud.
Hear hear! Look at the US Robotics Courier V.Everything.
Here is a MODEM that was brought out yonks ago, with a 20MHz Intel 80186 CPU (or 25MHz for the US model), 256k (from memory) flashable memory for firmware and a kick arse (for the intended application) TI DSP.
I bought it when it was just a "Courier Dual Standard V.34" 28.8k MODEM, then upgraded to V.34bis 33.6k with a small download from USR, then upgraded to X2 53.3k, then V.90 once it was ratified and then released by 3COM.
Sure, I paid $600 odd
It comes with support for almost bloody everything on the DTE and modulation fronts
I set a pair up for the Aussie Stock Exchange for a broker connection through leased lines. I was able to set them up so that all you would have to do is plug them in (to DTE and leased line), switch them on and as soon as they could see each other whoosh, they'd connect in 3 or 4 seconds and then stay connected for YEARS.
It may seem out dated with broadband now, but it's as awesome initial design, given the interface limits (RS-232 and PSTN).They took the limits of RS-232 and PSTN into account when line rate technologies were at a much lower performance. And thats how these sorts of devices should be built.
For high I/O work, SCSI is best. Multiple requests can be issued to a drive, while it is working on another, and the drive can re-order those requests to reduce head movement and thus increase final throughput.
Add to that 15,000 rpm drives that could easily be limited by the highest ATA *burst* rates (especially with even a small (2 drive) SCSI RAID setup) and you'll see why SCSI has bus speeds up to 160MB/S.
I've had nothing but very impressive experiences with SCSI gear. My 16x SCSI CDROM is *much* faster than my 32x UDMA CDROM for example and does not choke on a simple dd! I've found most IDE gear to be of a low quality compared with most SCSI.
Of course, you pay for good SCSI, big time.
and has (or had:) net access, I guess this must be Osama Bin Laden!
With the price of flat panel displays, RAM and Gigabit ethernet coming down to decent prices, I think the next homewide system for me will be home brewed thin client based.
1Gb RAM in each client to help releive reliance on network (hell in Sydney 1Gb of SDRAM is less than $au250 (4 * 256Mbyte DIMMS)!
With an AWARD BIOS equiped motherboard, I can easily flash an Etherboot ROM image to the mobo BIOS thanks to AWARDS modular design (so no need to burn ROMS for the NICS).
With Gigabit ethernet, when the network is relied upon for block device usage, ~120MBytes/S is very nice thank you very much...
from the server with 160MByte/S SCSI and also 1 or more gig of RAM. Urgh urgh urgh.
With Thunderbirds a plenty and each client with DVD-ROM. ; )
So I can have *silent*, diskless, flatpanel machines that are fast in both bedrooms and the lounge. With one central server with all our albums MP3'ed (145 so far done), central storage for video recorders like Tivo in the lounge and bedrooms also, and allowing quick easy backups of user ~ dirs.
Then add in some IP telephony to allow intercom and speakerphone transfers of incoming calls. All I need is more money!!!
I already have the US Robotics Courier V.Everything. I was sold on it YEARS ago with it was just a V.34, then upped it to V.34bis, then X2 and now V.90, talk about over engineering! What an awesome machine. Anyone know of any undocumented stuff or hacking info for the Courier V.Everything? I liked the old days when I had a MODEM dialing with a DTMF duration of 30mS. It sounded so cool, dialing up BBS' that quick (especially attack-dialing to get onto a BBS line once available), but my Courier won't allow that speed, it is limited to a longer length to appease Telstra/Austel.
Speaking of which, I just opened the pdf version of the FreeBSD Handbook in Acrobat 4.05 on my PII-300 256Mb RAM and striped swap (across 2 7200rpm UDMA drives) with kernel 2.4.10, then all of a sudden, while browsing through the docco, thrash thrash thrash, I had flashbacks to my Windows days (oh God the humanity).
My mouse cursor barely responded to movement, I managed to get my mouse over to close Acroreader but clicking seemed out of the question. I managed to get to a console prompt to see what top had to say when finally the thrashing stopped and my once responsive system was back to some sort of sane level.
WTF!?!? This is a first for me in my 4 years with Linux.
Sawfish died also and now won't restart. I wan't my AquaX buttons back! Damn you Rik!!
Hi Telek,
All I have said, is that not all overclocking causes the PCI bus to be overclocked and that I beleive that the thought of MOST hdd failures being caused by overclocking, is silly. Of course it IS possible to overclock your PCI bus in the pursuit of system overclocking.
I thought that I said this, but in case I didn't: You changed the FSB and the multiplier.
Yeah, what is your point? The PCI is still at 33MHz. I used a standard BIOS setting for the FSB and CPU multiplier. Choosing a 66FSB gives a PCI multiplier of 2 automatically, choosing a 100FSB gives a PCI multiplier of 3 automatically. They are standard increments that I am choosing that cause appropriate PCI speeds, not marginal increments like 112 that cause the PCI to be raised to 37MHz for example. It is preferable that I don't overclock the PCI or AGP bus as I am a beta tester for a card mfg that prefers results based on within-spec hardware.
Can you change your FSB from 100 now to 112 without changing the PCI bus speed? Or from 66 to 75?
No. But that does'nt disprove my point that overclocking can be done without changing the PCI speed, it merely proves that the PCI speed can be overclocked as a result of system overclocking, something I would never dispute, since I've been overclocking buses since before the existance of PCI or even VLB (My old SpeedStar 24X supported 16MHz ISA, but I could only go to 13MHz while being stable).
Thats fine, but do you have any facts or information
Hey, I'm not the one here making wild claims like " by far the most common reason [for hdd failure] was due to overclocking ", so perhaps you should provide some facts or information to support those claims.
I just searched on the internet and found no evidence to support what you were talking about
And just what was it that I was talking about? 1. Not all overclocking causes the PCI bus to run above spec? or 2. I highly doubt that "by far the most common reason" for hdd failure is due to overclocking.
I also love online debating, I've been doing so since 1991 starting with FidoNet and I avoid having words put into my mouth or arguing against things I do not dispute.
I stand by what I have said in this thread.
Absolutely!
.edu) and the second largest (at the time), the ASX and a large legal firm. I assure you that NONE of the faulty hdds we got replaced were due to overclocking.
Thanks, that is really cool.
Woah, sorry to say but you are very misstaken on that.
No I'm not. Choosing say a standard 300MHz CPU option in the BIOS, will yeild a 33MHz PCI, 66MHz AGP, 66MHz FSB and a CPU multiplier of 4.5, or possibly 33MHz PCI, 66MHz AGP, 100MHz FSB and multiplier of 3, depending on whether the CPU is detected as a 66 or 100 FSB type. I had to open my PII 300 to doctor it a little for 100FSB.
Choose a standard option of 400 (PII), you should get 33MHz PCI, 66MHz AGP, 100MHz FSB and a multiplier of 4. These assume that the CPU's multiplier can be changed. And my point is, that overclocking CAN be done without increasing the PCI and AGP speeds. I have been doing it for many years.
Of course, some overclocking methods are achieved through increases that cause PCI, AGP and FSB to also increase. Are you trying to say that you can't overclock without increasing PCI? It is true that some hardware forces this method due to multiplier locks for example.
The thought that most hard drive failures occur due to overclocking, is absolutely ridiculous. The tech that said that, is either full of it or ignorant. I've been a hardware tech for 13 years and about 7 of that specifically with PC hardware (first with NEC then DEC).
Now the problem is that when you up the clockrate of your chip, you're really changing the BASE rate
That is NOT always the case. My CPU was a PII 300 with 66FSB. I opened it, cut a track, now it supports 100FSB. I choose in the BIOS 100 FSB, multiplier 3.5, and I get a PII-350 with 100FSB and the PCI runs at 33MHz and the AGP at 66MHz. So I overclocked my CPU WITHOUT increasing the the PCI or AGP speed. I can verify this with a CRO.
There are different methods of overclocking and there are multipliers in many places that allow for example 33PCI+100FSB or 33PCI+66FSB. It's too simplistic to say that all overclocking increases the PCI speed and downright silly to think that MOST hdd failures are due to overclocking! I have supported Australia's largest WAN (it's
fires on the leading and falling edge of the signal
I've been dealing with positive and negative edge triggering in digital electronics since 1989 for the Royal Australian Navy, Electronic Weapons.
Thanks for elightening me on the IBM hdd head parking though, that is very neato.
Well, not so neccessarily... those IBM drives have their heads parked off the surface of the disk when powered off so you can shake'em around all that you like and won't damage the surface of the disk.
Can you provide a link to support this? I've been working in PC hardware repair since '94 (For NEC and DEC) and electronics since '89 before that. I've never seen any modern VC drive (non stepper) move the heads *off* the platters, they *always* move the heads to the inner most location of the platters, near the spindle, where head touch down is at the lowest surface speed (inner vs outer) and also required motor torque is the least when spinning up, since the heads are clamping down on the platters!
To move *off* the platters, would require some mechanism to transition the heads from off/on to the platters. I have'nt been watching hdd tech lately, so a link to this would be very interesting.
I like that and that was one of the reasons why I sprung the extra for the drive.
I hope you're right, for your sake. Not that I don't think IBM drives are not great.
I was told firsthand by a tech that by far the most common reason was due to overclocking.
You were told wrong. Loads of techs don't know shit. They come out, run a test from floppy, deem a drive OK or not, and if not replace it. They do the same kind of crap with printers, MODEM's, screens, etc. Most of them are good for what they do (replace hardware) and not much more.
Don't forget that running that wonderful 1.4GHz processor at 1.624GHz also overclocks your PCI bus by 4.5MHz.
It depends on how the overclocking is done. If you're just choosing standard CPU multiplier and FSB speeds from BIOS settings, or even jumpers, it should not change the PCI speed (but can). The small 1MHz tweaks that some BIOS support, can change the PCI speed also. You can also explicitely change the PCI speed too if you have the right mobo and read the manual. Just choosing different default CPU speeds should not though.
Doesn't sound like much? Well that's 14% above spec. After a while, and with the heat that these drives generate anyways, it's pretty easy to toast your electronics by doing that.
Possible, but definetely NOT the most common reason. I've been over clocking since it required soldering ; ), and I've found that hardware either does'nt work or becomes unreliable due to the overclocking. *Some* CPU's will die, though I've never killed an Intel. Having said that, some Intel PII's above 300MHz (and including some 300's), had faulty thermal diodes, which would allow destruction of those CPU's under overclocked settings. I've never killed a hdd due to overclocking.
Hell, I used to hot swap an old IDE WD Caviar (for years) to move big chunks of data before I got some 100base-T NICs. ; )
I like the old Caviars too. I had a WD Caviar 340Mb for about 7 years. Worked very nicely and was much faster than the crap Maxtor it replaced.
/, /usr, /home and /tmp. Urgh urgh urgh...). The *day* I plugged in one of the Seagates with the little Caviar as the slave.... boom, no more Caviar. Shit itself baddly, oddly somehow due to the Seagate? I don't know, I'm pretty happy with the Barra's though, going from about 1.2MB/s with the Caviar to ~28MB/s raw off the Barra's and 45MB/s Linux (2.2.x) RAID-0 was pretty sweet.
However, I upgraded (big time:) to two Seagate UDMA Barracuda 20Gb drives (Linux RAID-0+Reiser
Of course now, with Linux 2.4.x, RAID-0 performance is SLOWER than the raw drive speed for some weird reason that I am still trying to track down. FreeBSD on the other hand, is consistently faster transfering data off these drives. I'll soon test FreeBSD's RAID-0 to see if I can get back my 40+ MB/s...