Heavy Reads Cause Crash - SSD, RAID

[EXCellR8]

I've been using a pair of 80GB Intel X25-M's as my RAID-0 OS drive for awhile now but I am having some issues with crashing. It seems that whenever the disks are under heavy read (writing seems fine) like virus scan or even some large PDF's the computer crashes to a BSOD, usually a generic or unknown error. This sort of makes me wonder if it's driver related but the SSD array is fine otherwise (i.e. quick boot time, fast programs etc..)

I think I have the latest Intel RST installed as well as SSD Tweaker set to Auto. Like I said the drives are fine otherwise and never crash unless they are being read heavily. I'm not sure when this started but my guess is it wasn't always an issue. RST doesn't see any problems with the array and the only thing I see that's disabled is "write-back cache." I have no idea what that is and I'm pretty sure I never touched it. Also, defrag is disabled in Windows and I have found a few Superfetch errors in Events, if that has anything to do with this... any ideas?

Info:

Intel X25-M 80GB (x2)
ASUS P6T X58 motherboard
Windows 7 Ultimate 64-bit

 

[Spikesoldier]

i would try re-seating or replacing SATA cables.

 

[tweakboy]

I think that can be a sign of , death is soon to come to one of those drives. But like spike said make sure cables are nicely installed. gl

 

[EXCellR8]

the drives are relatively new and i would think that if one were corrupted i would have crashing all the time, but that's not the case. the cables are connected firmly but i'm wondering if heat could be an underlying factor. i am using the cooler master HAF chassis but where the SSD's are it's sort of a dead zone with little airflow. perhaps moving them down a few slots would resolve the issue...

 

[Emulex]

you overclocking or anything? i have raid-0 ssd's for a while - x25-m and sandforce and never had any crashes (crashes and raid are not good at all, if you crash - don't use raid) - hence why folks separate storage (SAN) from work - storage never crashes. you could have corrupted your raid-0 by crashing. might be a good idea to scan the drives to see if there is any corruption going on.

 

[sub.mesa]

you could have corrupted your raid-0 by crashing

I don't see how. That would mean the RAID metadata - stored at the last sector of each drive - needs to be corrupted. For that to happen, either the SSD has to corrupt itself upon losing power (a weakness of many SSDs without supercapacitors) or the RAID layer has to update the metadata during the power failure. In other words, I don't see this happening.

But when using Intel Volume Write-back cache option, a unique option in Intels RAID drivers that uses your RAM as write-back buffercache, you could corrupt your filesystem during a crash. But still that should not affect the RAID itself.

The problem of the OP sounds alot like cabling problems or general stability problems. Have you looked at SMART output? Have you ran a MemTest86+ test for 12+ hours?

 

[EXCellR8]

Yea I don't plan on enabling write-back cache but I'm wondering if I should swap SATA ports for the drives. I doubt that will change anything but it's worth a shot I suppose. I don't think my board is the greatest for RAID but it's been working great otherwise. I'm not really sure when the problem started but I can't run any heavy scans on the OS drive or it will crash, every time...

I guess if I had to abandon the RAID and just use a single SSD with AHCI enabled, it wouldn't be the end of the world but I'd rather not do that. I'm gonna do a little experimenting and see if I can stabilize the reads.

Thanks for the input!