Chad, I am aware of the magnitude of the problem *IF* corruption occurs in RAID 5. That is not even remotely relevant with a conversation about the probability of such corruption.
There appear to be two competing claims here: 1. RAID 5 itself is well known for having silent data corruption and should not be used. 2. Apple sold, until very recently a Mac OS X based, hardware RAID 5 solution, supposedly honoring a "do no evil, cause not harm" philosophy. RAID 5 is not a new thing. Silent data corruption is not a new thing. I won't buy an argument suggesting Apple just figured this out recently and that's one of the reasons why they killed the Xserve RAID. Chris Murphy On Mar 9, 2011, at 1:51 PM, objectwerks inc wrote: > > On Mar 9, 2011, at 1:46 PM, Chris Murphy wrote: > >> >> >> On Mar 9, 2011, at 1:35 PM, objectwerks inc wrote: >> >>>>> The integrity of the data with files after any issues such as these may >>>>> be suspect, especially if the fileystem was on a RAID5 which is very well >>>>> known for silent data corruption. This is why RAID5 should not be used. >>>> >>>> OK that's possibly a whole separate thread for qualifying such a >>>> statement. Apple has a supported product that uses RAID 5. I have clients >>>> with them and they've lost drives, and no data, and no data corruption. >>>> And an even larger sample size exists if filesystems other than jhfs+ are >>>> considered. RAID 5/6 are common with ext3, ext4, XFS and other file >>>> systems without anyone suggesting RAID 5 in particular is known for itself >>>> increasing the incidence of silent data corruption. >>>> >>> >>> >>> There is a reason why it is called silent data corruption. They may not >>> know they have it. It happens all the time and with HW raid 5 you may not >>> even know it for a long time. >>> >>> This is the whole reason why ZFS was made. >> >> ZFS was not made for combating this claim of RAID 5 specific silent data >> corruption, but rather silent data corruption in general. > > And the difference is? You corrupt one disk in a RAID 5 array and the whole > array is corrupted. > >> >> RAID 5 employs parity. RAID 1 does not, nor do conventional non-arrayed >> jhfs+ volumes. While RAID parity is not as sophisticated at ZFS >> checksumming, it is certainly better than nothing. So there is some error >> detection and correction possible, so I'm not understanding how RAID 5 is >> "well known for silent data corruption" and should not be used. I think this >> is a rather remarkable claim. >> >> >> Chris Murphy > > > If you have one disk that has silent data corruption (and parity may not help > you at all as you calculate your new parity on the corrupted data) you > compound it with RAID 5. > > http://www.raidinc.com/pdf/Silent%20Data%20Corruption%20Whitepaper.pdf > > > http://boink.superatomic.com/2009/04/25/the-raid-5-write-hole/ > > > http://www.google.com/#hl=en&sugexp=ldymls&xhr=t&q=raid+5+silent+data+corruption&cp=24&qe=cmFpZCA1IHNpbGVudCBjb3JydXB0aW9u&qesig=0VyEUZb1ePVXqTxPyJd8vg&pkc=AFgZ2tkTDI1HqM7fc6R1egXasC-1CNF3wT4BG5rzsZmDlr4IGBRuaMW3LWOqaaHzeI1IQlVZ2T5WTE3o3LlGRR6gLiYOy2lsHA&pf=p&sclient=psy&site=&source=hp&aq=0b&aqi=&aql=&oq=raid+5+silent+corruption&pbx=1&bav=on.2,or.&fp=ef34c9a9ed856910 > > > _______________________________________________ MacOSX-admin mailing list [email protected] http://www.omnigroup.com/mailman/listinfo/macosx-admin
