ICH10R RAID5 dropped one of my drives, folding@home to blame? (+How to align RAID10?)

Acejam

Weaksauce
Joined
Nov 4, 2009
Messages
65
Hey all,

Long story short, I picked up (4) HD204UI's about 2-3 weeks ago. I set them up in a RAID5, which yielded me ~6TB of usable space. Yes, I know RAID5 isn't the best for performance, but this is mostly for data storage.

Yesterday for the first time since adding the new array, I fired up my folding@home client. I'm using the official Windows client with the SMP flag on my i7 920. The folding client's files live on the RAID array. After getting home from work - one of my drives in the array was marked as Failed, according the Intel Matrix Storage Manager. In addtion, the entire array was marked as degraded. I did some quick Google searching, and decided to mark the failed drive as "Normal" in the IMSM utility.

The array then began to rebuild, which obviously will take a long time with 6TB. (but I'm only using 1TB right now, rest is free space)

My question is - is my folding client to blame? I run with hyperthreading on my i7 processor ,so I'm wondering if the 100% CPU usage on all 8 cores took away from the RAID array. In other words, it looks like the folding client took up so much CPU, that it didn't have enough power to calculate all of the parity bits being generated by the writes of the foldling client itself.

Thoughts? I ran a RAID1 array on this same board for almost a year without issue, including plenty of folding. I'm thinking about switching to a RAID10 now, and will try that out. Yes, I do have backups of my data. ;)

Also - once I get all 4 drives into a RAID10 array using the ICH10R on-boared controller, do I still need to align my partition? Do partitions on RAID10 volumes still need to be aligned? (or on RAID volumes in general) I'm using Windows 7 Ultimate x64, but I noticed that the partition I create using Win7 2 weeks ago starts at sector 63. If I'm going to align my partition, now is the time to do it. (I only have 1 giant NTFS partition)
 
Last edited:
High CPU usage can't be the cause. More likely you have a bad sector on one of your drives. Try gaining access to the SMART data. You can also just zero-write the dropped disk in another system and then add it to the raid again for it to be resynced.

Aligning partitions is fine with Windows 7 and Vista SP1; unless you use cloning/imaging programs which create the partition themselves. Do a fresh install instead and let Windows 7 create the partition, then it will be aligned. Check with AS SSD to check alignment.
 
What version of the Matrix Raid option ROM do you have? (It says when you boot the computer)

What version of the drivers?

You could have a bad drive, or some drives just don't work well in RAID.
 
High CPU usage can't be the cause. More likely you have a bad sector on one of your drives. Try gaining access to the SMART data. You can also just zero-write the dropped disk in another system and then add it to the raid again for it to be resynced.

Aligning partitions is fine with Windows 7 and Vista SP1; unless you use cloning/imaging programs which create the partition themselves. Do a fresh install instead and let Windows 7 create the partition, then it will be aligned. Check with AS SSD to check alignment.

I ran Samsung's ESTOOLS against the drive last night, and did a full diagnostic scan, which took ~6-7 hours. It reported back no errors.

Under IMSM it also reports that the drive is now missing. Apparently after rebooting, it's been kicked out of the array. I opened up CrystalDiskInfo and I'm able to view the SMART data. Anything I should look for in particular?

Screenshot:

drive.jpg





What version of the Matrix Raid option ROM do you have? (It says when you boot the computer)

What version of the drivers?

You could have a bad drive, or some drives just don't work well in RAID.

The storage console in Windows reports 8.9.0.1023 under the About menu. The ICH10R controller reports the same under Device Manager as well.
 
I would try modding 9.6.0 in with CBROM/MMTool... pretty sure it helped me by dropping drives less. Link me your BIOS and I can do it for you...
 
Last edited:
I would try modding 9.6.0 in with CBROM/MMTool... pretty sure it helped me by dropping drives less. Link me your BIOS and I can do it for you...

Yeah in all honesty I'm not sure if I want to go hacking at my BIOS....

FWIW, I have a vanilla Asus P6T board.
 
Cool, so it looks like you just took the latest BIOS ROM and dropped in a new version the Intel Matrix Storage ROM? Where did you find the newer Intel ROM from?

Now that I think of it, I probably have an *older* version of my BIOS. Perhaps I should try updating to the latest version first, rebuild my array, and see how it goes? I'm still weary of RAID5 now...

You don't think high CPU usage could do something like this? I'm talkin' 100% CPU on all 8 cores for 12 hours straight....
 
I found a zip file with most of the Intel option ROM versions. I modded my Gigabyte board with that 9.6.0.1014 one too.
 
I re-created my RAID5 array and it took 2.5 days for Windows to format the volume. It definitely didn't take this long before, so perhaps that was part of my issues? (then again, the drives were new at the time)

This time, I also tried aligning my partition. I opened up diskpart, and used the following command:

create partition primary align=1024

I then formatted the partition under Disk Management, and selected a allocation unit size of 4096.

The new results, with write-back caching turned on:

RAID5_1024.png



Old RAID5 volume (default Win7 alignment w/ write-back caching on)

brandnew_4TB_RAID5.png


Old RAID10 volume, used as a test:

brandnew_4TB_RAID10.png




I'm using (4) Samsung HD204UI drives, which have 4k sectors with 512 byte emulation. Am I aligning these drives properly?
 
I'm back!

My RAID5 array on my ICH10R broke again today! Same deal, computer was left on for a day or two folding, then I come home from work today and the Intel Matrix Storage Manager reports a RAID volume is degraded. I only have 1 RAID volume on this machine. I had even aligned this partition too, hoping it would help this time around.

Luckily, as always, I have a backup of my data. (fresh from this morning) I'm going to delete this RAID5 array, and move over to a RAID10.

ICH10R RAID5 = :mad:
 
This is a reason why I avoid on-board raid :(

Good luck, try ZFS and see if a certain drive degrades
 
After much reading, it looks like there have been a lot of issues with the Intel Matrix Storage Driver 8.9 driver causing "random" failed drives, which appears to be my exact issue. Each time, the drive that fails is random, and is on a different port. I've already tested my drives, both before and after failures, and they're 100% OK.

I have uninstalled the Intel Matrix Storage Manager and have now installed the Intel Rapid Storage Manager 9.6, which supposedly fixes this issue. I marked the failed drive as normal, and the RAID5 array is now rebuilding. If I can get away using the 9.6 driver and continue to keep my RAID5 array, I will be happy. Otherwise, I might switch to a RAID10.

Thoughts?

Also - should I update my BIOS and/or the Intel RAID ROM? (dandragonrage was nice enough to create a custom ROM file for me)
 
@AceJam,

How is the RAID5 setup working for you atm?

I bought 3x WD20EARS for my ICH10R raid 5 setup. Turns out a big mistake, and i guess it's the problem with the TLER not being enabled / available on the WD20EARS.
I'm also using the Intel Rapid Storage Tool 9.6.

I'm looking to trade em in for 3x HD204UI's.
I just can't seem to find any reports on problems with this drive, regarding to CCTL (supposedly the equivelant of WD's TLER).

I hope these drives (with the necessary Firmware upgrade ofc. http://www.samsung.com/global/business/hdd/faqView.do?b2b_bbs_msg_id=386) will result in a stable RAID5 via ICH10R onboard. I use the same mobo as you have (Asus P6T), so i'm really curious as to how it's working for you atm.

EDIT: PS: What's with this alignment? I understand that when you use Win7 x64, there is no need to manually change something to get em aligned proper, yet i see you did something to it. Could you explain as to how and where u make this change, and especially, why?
Thanks in advance.

Regards,
Kami.
 
Last edited:
Back
Top