1

I have a workstation that uses an adaptec RAID 71605 controller with 4x8TB WD drives in a RAID5 config. One of these drives failed, and after replacing it and rebuilding the RAID, a lot of the data stored on the RAID is corrupted. The Hyper-V machines do not start any more, and a lot if graphics are just broken or have a lots of artefacts.
I have but the controller in a different workstation, with the same result. I then swapped the controller, but the data is still corrupted. Checkdisk and rebuilding the array did not help
Now I am very curious as to why that happened?
The controllers only job is to assure that a drive failure is non-critical, and that was not the case here. Is that something "normal" that can happen? Is there a way to protect for such errors? And most importantly: Is there a way of recovering the files?
smiet
  • 11
  • 1
  • 3
    See the section on RAID 5 in this Q&A https://serverfault.com/questions/339128/what-are-the-different-widely-used-raid-levels-and-when-should-i-consider-them and continue with what a URE is here: https://serverfault.com/q/812891/546643 - additionally, no RAID level is a substitute for regular backups – Bob Feb 12 '21 at 08:41
  • 1
    Every few weeks we get someone come to this site with exactly the same story - RAID 5 is dangerous, personally I think it shouldn't be allowed to be sold/provided by controller manufacturers. Sorry it's you this time, oh and only R1/10 and R6/60 are reliable, oh and ZRAID/ZFS if you're into that. – Chopper3 Feb 12 '21 at 10:55
  • 1
    Hi, and thanks for your comments. Of course we have a backup strategy, and its restoring at the moment, but restoring 24TB of data takes it's sweet time. I guess I am manly interested as to what exactly happend? Like what caused it to fail? Bad thing is that we are scheduled to phase out our RAID5 workstations and the end of the year and replacing them with new controllers and a raid 6 + hotspare config. Guess it was just bad luck – smiet Feb 13 '21 at 09:06
  • @Chopper3 *RAID 5 is dangerous* That's a bit strong. It wouldn't surprise me in this case that more than one drive was going bad, but a lot of the bad sectors were never read for a while so they were never noticed. Once one drive failed, the rebuild exposed the latent issues. A RAID6 array is *less* susceptible to such failures, but they're still possible especially if you do something like creating a RAID 6 array from 19 8TB drives (with a 1 MB segment size "because bigger is faster"...) But a RAID 5 array of 5 400GB SSDs would be fine (yeah, two 1.6TB RAID 1 SSDs would be better, but...) – Andrew Henle Feb 13 '21 at 13:53
  • Google 'raid 5 bad' or similar - this is the first thing that came up for me, and it's over 11 years old - so yeah, dangerous; https://www.zdnet.com/article/why-raid-5-stops-working-in-2009/ – Chopper3 Feb 15 '21 at 12:47
  • @Chopper3 RAID5 is a tool. A poor mechanic blames his tools.. I guess I'll just have to go tell my customer that their, oh, probably tens of thousands of RAID5 arrays that have been fine for years all need to be trashed and we have to go out and buy more disks because the internet said so. Of course, we don't conflate data availability with data security so if a RAID array of any type goes bad we just restore from backup and try not to apply broad-brushed and trite aphorisms to real-world situations. – Andrew Henle Feb 15 '21 at 23:08
  • I'm sorry but you're wrong - if you'd spent the best part of the last decade repeating the same thing to literally hundred of R5 users who've lost their data you'd agree too. And while I'll admit that for arrays using very small disks or disks with high UREs that R5 *can* work just fine what we find on this site is that inexperienced users think it's a good idea to use it with large consumer-grade disks and come to pay the price. There's overwhelming evidence to this effect, sorry if it annoys you but you won't find anyone on this site who will agree with you. – Chopper3 Feb 16 '21 at 09:04
  • @Chopper3 You're just saying this site is full of poor mechanics who misuse tools. I bet if you put diesel fuel into a gasoline-powered car it runs poorly, too. – Andrew Henle Feb 21 '21 at 17:36
  • Yep, us 10,000 or so, totally wrong, you must be right, all those people who turn up asking for help must be wrong too - good luck! – Chopper3 Feb 22 '21 at 16:51

0 Answers0