Internal RAID 5 often drops out during data-intensive work

P

peppergomez

2[H]4U
Joined
Sep 15, 2011
Messages
2,151
So this is an intermittent but complete pain in the ass--

My D drive, which is an internal RAID 5 array of 4 HDD drives, will sometimes drop out/disappear in Windows 10 when I am doing data/drive-intensive tasks like exporting large images. It's necessary to reboot in order for the D drive to be recognized again by Windows.

How can I troubleshoot this? It's seriously annoying.

Thanks.
 
On a discrete RAID card? - try taking the side of the case off and place a fan blowing on it - it might be overheating and dropping the array. Also try reseating the data and power cables to the drives, as well as reseating the RAID card (do this with the power OFF and power cable UNPLUGGED).

On-motherboard hardware RAID 5? GOD HELP YOU. I lost about 6TB of data to an on-board RAID 5 when the motherboard died back when 1TB drives just came out. An exact same model replacement from RMA wouldn't even see the array. Back up that data elsewhere, and fast! You could also try the case / fan trick, and also try reseating the power and data cables to the drives.

Software RAID 5? Reseat the power and data cables.
 
You can also load up Event Viewer and take a look at the System logs, you might see a bunch of errors/warnings about the time the array drops out. It does sound suspiciously like a hardware RAID card having heat issues, as some RAID cards can generate a ridiculous amount of heat, and are designed for the high airflow server cases. I'd imagine a software/Windows RAID problem would have a different failure mode than the drive just disappearing, as well.

Our old NAS here at work has a motherboard problem that causes the PCI bus to "glitch" every few months, which resets any cards that are plugged in. When that happens, the RAID card resets but Windows doesn't recognize it as being there, so the drives just disappear. They show back up again after a reboot, just like they do for the OP.
 
arestavo said:
On a discrete RAID card? - try taking the side of the case off and place a fan blowing on it - it might be overheating and dropping the array. Also try reseating the data and power cables to the drives, as well as reseating the RAID card (do this with the power OFF and power cable UNPLUGGED).

On-motherboard hardware RAID 5? GOD HELP YOU. I lost about 6TB of data to an on-board RAID 5 when the motherboard died back when 1TB drives just came out. An exact same model replacement from RMA wouldn't even see the array. Back up that data elsewhere, and fast! You could also try the case / fan trick, and also try reseating the power and data cables to the drives.

Software RAID 5? Reseat the power and data cables.
Click to expand...

The SATA cables are connected to my mobo, yes, if that makes it on-motherboard RAID. I'm a bit of a noob about this...what is so terrible about on-mobo RAID?
I do back up my data, yes, and have just begun backing it up to the cloud (which is going to take awhile).
 
peppergomez said:
The SATA cables are connected to my mobo, yes, if that makes it on-motherboard RAID. I'm a bit of a noob about this...what is so terrible about on-mobo RAID?
I do back up my data, yes, and have just begun backing it up to the cloud (which is going to take awhile).
Click to expand...

Maybe I didn't spell it out very well...

I lost a SHIT TON of data, back when 1TB drives were new, when my on-board RAID 5 wasn't even detected with a replacement motherboard of the same model, manufacturer, and BIOS revision.

The motherboard had died and needed to be RMAd. Theoretically, as with discrete RAID cards, all I needed to do was plug the drives back in - nominally in the same SATA ports, which is what I did.

The array was not detected. Up and vanished like a fart in the wind, never to return with the ~6TB of data that took years and years to accumulate.

After that I bought discrete RAID cards. I went with the Adaptec 6805, then 7805, then 8805. Each card accepted the array created from the old cards, and I can move them from PC to PC with no issue - I LOVE IT!




Hopefully it's nothing serious for you, but at least you're doing the smart thing by backing up all of your data.
 
OK thanks for clarifying. Christ. That's terrible. I had a RAID 5 9TB drive go - 2 of the four drives died (they were the shit Seagate 3TB drives that I have since joined a class action lawsuit against) and I lost all my data. I am a photographer and lost 4 or 5 shoots that I hadn't gotten around to backing up. No fun at all.
 
drescherjm said:
Could be a bad drive. See what CrystalDiskInfo says about your disks.
Click to expand...
Thanks for the CDI suggestion. Just ran it. The disks in the RAID all read as good.

Another drive, an external one that I have my movies on, is listed as Yellow/Caution for status. "C5 Current Pending Sector Count" has a yellow circle next to it. What is this telling me--that a sector is likely to die?
 
Is this new behavior for the array, or has it always done this?

Did the trouble start under Windows 10?

Your mobo BIOS updates also contain on board RAID controller updates.
Make sure mobo BIOS is updated and you are using the newest drivers.

Mobo RAID controllers are generally very low end as far as performance
and reliability. Although I've had good luck with Intel mobo RAID controllers.

You may want to get a "real" RAID card and setup your array on that.

I'd also recommend a good cloud backup for work related critical data.
I use iDrive.com on my own server and for customers.

.
 
peppergomez said:
Thanks for the CDI suggestion. Just ran it. The disks in the RAID all read as good.

Another drive, an external one that I have my movies on, is listed as Yellow/Caution for status. "C5 Current Pending Sector Count" has a yellow circle next to it. What is this telling me--that a sector is likely to die?
Click to expand...

Its telling you that you wrote sectors to the drive (look at the raw count) and the last read the drive could not read the sector and it has not been able to recover the data that was lost.
 
arestavo said:
On a discrete RAID card? - try taking the side of the case off and place a fan blowing on it - it might be overheating and dropping the array. Also try reseating the data and power cables to the drives, as well as reseating the RAID card (do this with the power OFF and power cable UNPLUGGED).
Click to expand...

I was thinking the exact same thing with the above statement. Had a coworker with the exact issue with the raid controller over heating under load and dropping out. Hopefully you get it resolved.
 
Update---in case it helps, here's my mobo info:


asus z97-pro motherboard version Rev 1.xx

Going to google for updated drivers and install if needed.
Curious if anyone has other suggestions. Thanks.
 
Did you hook a fan up to it? Nothing else you do will matter if the controller is overheating.
 
There's already a side fan that's blowing directly on the mobo. There's no room for an additional fan, though I can try by opening the side of the case to let air in, though it will also let dust in as well.

I googled the issue and added my mobo to the search term, and found this post, which though 2 years old might be relevant:
http://www.coyoteblog.com/coyote_bl...-windows-7-randomly-appear-and-disappear.html


  • Check your power plan in Windows control panel. Even if the computer is set never to sleep, your hard drives may be set to sleep (this is in fact the default in windows 7). Go to the power plan advanced settings, look for hard drives, and set the time to sleep to 0 which causes them never to sleep. I am not sure this is necessary but others report some success with this. I may go back later to see if I can change it back -- I don't necessarily need my hard drives spinning all day.
  • It turns out that installing the Intel chipset driver is not enough. I had thought that since the SATA controller is part of the chipset, the chipset driver would cover it. However, once I installed the Intel chipset driver, when I checked the SATA / AHCI controller in device manager, it still showed the driver to still be a Microsoft driver. Turns out this is the problem. You need this to be an Intel driver
  • The drivers I wanted for the Z97 were right there on my motherboard support site and were called:
    • Intel AHCI/RAID Driver Path for Windows Win7 32bit & Win7 64bit & Win8 32bit & Win8 64bit & Win8.1 32bit & Win8.1 64bit.
    • Intel Rapid Storage Technology Driver software V13.1.0.1058 for Windows Win7 64bit & Win8 64bit & Win8.1 64bit---(WHQL).
  • I had originally thought these were some sort of utility (and a utility is included) but these are essentially the eSATA drivers I needed. Once installed, checking device manager now showed an Intel rather than a Microsoft driver.
And that fixed it.


I followed his first step, setting the power plan settings for hard drive (they were set to turn off after 20 minutes).

I've all the updated drivers on the ASUS site, but there are no installation instructions included in the ZIP files, which I think is very lame of ASUS. So I'm googling to find out the best way to install mobo updates first.
 
Sounds like you are on the right track.

It must have the correct drivers installed, and as you noted, no power save spin down enabled on the drives.

.
 
You must log in or register to reply here.
Back
Top