Re: [CGUYS] RAID Risks Rising
Not really cause it won't make any difference in your opinion of RAID. Z for Zealot! You probably think I pal around with terrorists too. * ** List info, subscription management, list rules, archives, privacy ** ** policy, calmness, a member map, and more at http://www.cguys.org/ ** *
Re: [CGUYS] RAID Risks Rising
Never said RAID was best, just that it has it's place. Tom's argument previously was that because of HD's high MTBF RAID was useless, this new article that came up said RAID was useless because of low MTBF. I think you should reread the 2009 article more carefully. That article was not about MTBFs. It was about the probability of errors vs the number of bytes on the drive. As the number of bytes increases the probability of an error increases. The author said that that probability of 1 error would soon approach 100%. Under those condtions RAID would fail to rebuild a bad drive. Hence a RAID set would be no more secure than a single drive and it would be less reliable because the probability of something going wrong on a RAID set is higher than the probability of a single drive failing. Probability lesson (simplified but accurate): 1) The non-RAID bet: I flip a coin. Heads you lose. 2) The RAID bet: I flip 7 coins. If any one is heads, you lose. Do you want to take bet 1 or 2? * ** List info, subscription management, list rules, archives, privacy ** ** policy, calmness, a member map, and more at http://www.cguys.org/ ** *
Re: [CGUYS] RAID Risks Rising
Single drives still aren't fast enough for enterprise work. We've been over this...the horse is dead, buried, he's been exhumed and is now being beaten again within an inch of his death. RAID with backup on/offsite is still cheaper then putting together a cluster for some mid range customers. I've never said RAID should be used for storage...and of course always backup. This from my RAID 0 fully backed up pc. On Sat, Oct 25, 2008 at 10:16 AM, Tom Piwowar [EMAIL PROTECTED] wrote: Never said RAID was best, just that it has it's place. Tom's argument previously was that because of HD's high MTBF RAID was useless, this new article that came up said RAID was useless because of low MTBF. I think you should reread the 2009 article more carefully. That article was not about MTBFs. It was about the probability of errors vs the number of bytes on the drive. As the number of bytes increases the probability of an error increases. The author said that that probability of 1 error would soon approach 100%. Under those condtions RAID would fail to rebuild a bad drive. Hence a RAID set would be no more secure than a single drive and it would be less reliable because the probability of something going wrong on a RAID set is higher than the probability of a single drive failing. Probability lesson (simplified but accurate): 1) The non-RAID bet: I flip a coin. Heads you lose. 2) The RAID bet: I flip 7 coins. If any one is heads, you lose. Do you want to take bet 1 or 2? * ** List info, subscription management, list rules, archives, privacy ** ** policy, calmness, a member map, and more at http://www.cguys.org/ ** * -- Make sure you support your local CarbonONset programs! * ** List info, subscription management, list rules, archives, privacy ** ** policy, calmness, a member map, and more at http://www.cguys.org/ ** *
Re: [CGUYS] RAID Risks Rising
This from my RAID 0 fully backed up pc. Ever see Stoppard's Rosencrantz and Guildenstern? * ** List info, subscription management, list rules, archives, privacy ** ** policy, calmness, a member map, and more at http://www.cguys.org/ ** *
Re: [CGUYS] RAID Risks Rising
I can recall Robert Metcalfe, the co-inventor of Ethernet and founder of 3Com, writing in 1996 or so about the coming Internet crash. He called it a gigalapse. He publicly and very enthusiastically predicted that very soon the Internet would not be able to handle the data load and would fail. I must have overslept that day and missed it. Oh yeah, I remember something about Y2K destroying the world too. IOW, I'll believe this latest apocalyptic prediction when I (or someone else) see(s) it. -Original Message- From Zdnet storage blogs - titled Why RAID 5 stops working in 2009 Robin Harris discusses why RAID risks are starting to overwhelm its protections. A must read for anyone serious about storage (and especially retrieval) - not a be-all, but these risks are real and not considering them invites disastrous data loss. http://blogs.zdnet.com/storage/?p=162 * ** List info, subscription management, list rules, archives, privacy ** ** policy, calmness, a member map, and more at http://www.cguys.org/ ** *
Re: [CGUYS] RAID Risks Rising
On Thu, Oct 23, 2008 at 1:44 PM, Larry Sacks [EMAIL PROTECTED] wrote: I've done that with computers. But with a car? Wow... you must really be one of the Glitterati to be able to send a car back because you made a mistake... Haven't you heard of a lemon law? Many jurisdictions implemented them for the very purpose of making sure that a poc could be returned. And it was specifically because of cars. I would think that if you documented a computer blue-screened on you several times a week, needed to be rebooted every day or two to get decent performance, etc.; you should qualify under the lemon law to return it. -- John DeCarlo, My Views Are My Own * ** List info, subscription management, list rules, archives, privacy ** ** policy, calmness, a member map, and more at http://www.cguys.org/ ** *
Re: [CGUYS] RAID Risks Rising
Well Tom's implication is that if the car got a scratch it'd be all over...new car time! On Fri, Oct 24, 2008 at 12:45 PM, John DeCarlo [EMAIL PROTECTED]wrote: On Thu, Oct 23, 2008 at 1:44 PM, Larry Sacks [EMAIL PROTECTED] wrote: I've done that with computers. But with a car? Wow... you must really be one of the Glitterati to be able to send a car back because you made a mistake... Haven't you heard of a lemon law? Many jurisdictions implemented them for the very purpose of making sure that a poc could be returned. And it was specifically because of cars. I would think that if you documented a computer blue-screened on you several times a week, needed to be rebooted every day or two to get decent performance, etc.; you should qualify under the lemon law to return it. -- John DeCarlo, My Views Are My Own * ** List info, subscription management, list rules, archives, privacy ** ** policy, calmness, a member map, and more at http://www.cguys.org/ ** * -- Make sure you support your local CarbonONset programs! * ** List info, subscription management, list rules, archives, privacy ** ** policy, calmness, a member map, and more at http://www.cguys.org/ ** *
Re: [CGUYS] RAID Risks Rising
Well Tom's implication is that if the car got a scratch it'd be all over...new car time! You want to provide the quote and explain your tortuous logic to get from there to here. * ** List info, subscription management, list rules, archives, privacy ** ** policy, calmness, a member map, and more at http://www.cguys.org/ ** *
Re: [CGUYS] RAID Risks Rising
Not really cause it won't make any difference in your opinion of RAID. Z for Zealot! On Fri, Oct 24, 2008 at 5:17 PM, Tom Piwowar [EMAIL PROTECTED] wrote: Well Tom's implication is that if the car got a scratch it'd be all over...new car time! You want to provide the quote and explain your tortuous logic to get from there to here. * ** List info, subscription management, list rules, archives, privacy ** ** policy, calmness, a member map, and more at http://www.cguys.org/ ** * -- Make sure you support your local CarbonONset programs! * ** List info, subscription management, list rules, archives, privacy ** ** policy, calmness, a member map, and more at http://www.cguys.org/ ** *
Re: [CGUYS] RAID Risks Rising
On Fri, Oct 24, 2008 at 8:37 PM, mike [EMAIL PROTECTED] wrote: Not really cause it won't make any difference in your opinion of RAID. Z Well, the more evidence people post on here that RAID is less and less useful, riskier and riskier, the more you hew to your RAID is best approach. Perhaps you just need to stop reading the mounting evidence, since it won't change your opinion. -- John DeCarlo, My Views Are My Own * ** List info, subscription management, list rules, archives, privacy ** ** policy, calmness, a member map, and more at http://www.cguys.org/ ** *
Re: [CGUYS] RAID Risks Rising
Never said RAID was best, just that it has it's place. Tom's argument previously was that because of HD's high MTBF RAID was useless, this new article that came up said RAID was useless because of low MTBF. He seemed to agree with both sides because of his conclusion that IT companies use it to overcharge their clients. JBOD's have their place, RAID has it's place...entire mirrored systems have their place. I'm not an idealogue, I'll use whatever works best and is priced best for the need. Mike On Fri, Oct 24, 2008 at 5:47 PM, John DeCarlo [EMAIL PROTECTED] wrote: On Fri, Oct 24, 2008 at 8:37 PM, mike [EMAIL PROTECTED] wrote: Not really cause it won't make any difference in your opinion of RAID. Z Well, the more evidence people post on here that RAID is less and less useful, riskier and riskier, the more you hew to your RAID is best approach. Perhaps you just need to stop reading the mounting evidence, since it won't change your opinion. -- John DeCarlo, My Views Are My Own * ** List info, subscription management, list rules, archives, privacy ** ** policy, calmness, a member map, and more at http://www.cguys.org/ ** * -- Make sure you support your local CarbonONset programs! * ** List info, subscription management, list rules, archives, privacy ** ** policy, calmness, a member map, and more at http://www.cguys.org/ ** *
Re: [CGUYS] RAID Risks Rising
So the first time your new car breaks down, you buy a new one? Depends on why it failed. If I look at it and realize that I made a very bad mistake then I would send it back. I have done that with PCs on more than one occasion. I've done that with computers. But with a car? Wow... you must really be one of the Glitterati to be able to send a car back because you made a mistake... Larry * ** List info, subscription management, list rules, archives, privacy ** ** policy, calmness, a member map, and more at http://www.cguys.org/ ** *
Re: [CGUYS] RAID Risks Rising
Yeah especially when it's just say a scratch on the paint or maybe a dent in a fender. Those mac glitterati can really throw money around I guess. On Thu, Oct 23, 2008 at 10:44 AM, Larry Sacks [EMAIL PROTECTED] wrote: So the first time your new car breaks down, you buy a new one? Depends on why it failed. If I look at it and realize that I made a very bad mistake then I would send it back. I have done that with PCs on more than one occasion. I've done that with computers. But with a car? Wow... you must really be one of the Glitterati to be able to send a car back because you made a mistake... Larry * ** List info, subscription management, list rules, archives, privacy ** ** policy, calmness, a member map, and more at http://www.cguys.org/ ** * -- Make sure you support your local CarbonONset programs! * ** List info, subscription management, list rules, archives, privacy ** ** policy, calmness, a member map, and more at http://www.cguys.org/ ** *
Re: [CGUYS] RAID Risks Rising
So the first time your new car breaks down, you buy a new one? What about when you just change the oil? Or do you buy a new car then too? You must buy a LOT of computers, even macs if you toss them if they get a kernel panic. Interestingly the article actual reaches the same conclusion as you, but from completely opposite directions. Yet you agree. If you can have the same conclusion with completely opposite facts, then you can't possibly let facts get in the way of what you believe. Z for zealot for you! On Wed, Oct 22, 2008 at 10:38 AM, Tom Piwowar [EMAIL PROTECTED] wrote: The main story has a link to Desktop RAID is a bad idea. http://blogs.zdnet.com/storage/?p=116 The reasons are much the same as we covered here last month. I'm not sure why in the case of RAID, the idea is that if it fails at all, the entire idea of RAID is a failure. Statements like this make me think that the R in RAID stands for Religion. * ** List info, subscription management, list rules, archives, privacy ** ** policy, calmness, a member map, and more at http://www.cguys.org/ ** *
Re: [CGUYS] RAID Risks Rising
We missed it like global freezing, and that other catastrophe the inventor of the internet keeps droning about. Funny how all these RAID setups just keep humming along in face of the 'facts'. On Wed, Oct 22, 2008 at 12:14 PM, Jeff Wright [EMAIL PROTECTED] wrote: I can recall Robert Metcalfe, the co-inventor of Ethernet and founder of 3Com, writing in 1996 or so about the coming Internet crash. He called it a gigalapse. He publicly and very enthusiastically predicted that very soon the Internet would not be able to handle the data load and would fail. I must have overslept that day and missed it. Oh yeah, I remember something about Y2K destroying the world too. IOW, I'll believe this latest apocalyptic prediction when I (or someone else) see(s) it. -Original Message- From Zdnet storage blogs - titled Why RAID 5 stops working in 2009 Robin Harris discusses why RAID risks are starting to overwhelm its protections. A must read for anyone serious about storage (and especially retrieval) - not a be-all, but these risks are real and not considering them invites disastrous data loss. http://blogs.zdnet.com/storage/?p=162 * ** List info, subscription management, list rules, archives, privacy ** ** policy, calmness, a member map, and more at http://www.cguys.org/ ** * -- Make sure you support your local CarbonONset programs! * ** List info, subscription management, list rules, archives, privacy ** ** policy, calmness, a member map, and more at http://www.cguys.org/ ** *
Re: [CGUYS] RAID Risks Rising
So the first time your new car breaks down, you buy a new one? Depends on why it failed. If I look at it and realize that I made a very bad mistake then I would send it back. I have done that with PCs on more than one occasion. * ** List info, subscription management, list rules, archives, privacy ** ** policy, calmness, a member map, and more at http://www.cguys.org/ ** *