Anyone deal with troubleshooting bad drives in a RAID set?

[Emulex]

sata drives don't report errors correctly. now that sas is only like 10% more (RE4 SAS) you should stick to that. IOEDC is very important.

Plus use a semi-modern raid controller like 9260-8i (2 gen old) rather than that 3rd/4th gen stuff.

M5014's go for like $65 without battery,double with LIPO. solid.

 

[imagoon]

sata drives don't report errors correctly. now that sas is only like 10% more (RE4 SAS) you should stick to that. IOEDC is very important.

Plus use a semi-modern raid controller like 9260-8i (2 gen old) rather than that 3rd/4th gen stuff.

M5014's go for like $65 without battery,double with LIPO. solid.

That is not blanket true. SATA drives report errors based on how the firmware is programed to do so. Also, while the PERC 6 series is "old" they work fine and are still getting firmware updates as of 2012.

 

[sornywrx]

Possibly true for software RAID, but Acronis TI handles my RAID perfectly any time I want it. The problem with software RAID is that cloning is usually done off line (command prompt reboot) and software RAID does not exist then. Hardware RAID - no problem!

Thank you for that, what you said makes perfect sense I just hadn't given it much thought before.

 

[sornywrx]

I thought I'd update the thread on what happened. I went back to his house and he was pretty anxious to get the drives tested so instead of running a 6 hour consistency check and coming back the next day I thought I'd at least run a quick test on the drives. I pulled them both and hooked them up to my laptop via a USB/SATA adapter and ran Seatools. I ran the short test and the first drive pass with no problem and when running the tests the progress bar went smoothly and quick. When I hooked up the second drive the progress bar would go smoothly for a second then pause then continue erratically. It failed. I tested both drives two more times just to be sure there wasn't a problem with my USB adapter and the results were consistent, the bad drive failed every test and the good drive passed no problem.

Here is where I may have made a mistake... since these were internal drives there was no way to set the PERC software to flash the lights on the drives to identify them. I didn't think to follow the cables and see if they were labeled on the PERC until later. So I didn't mark the bad drive as OFFLINE and just booted up with the good drive by itself. Of course I got an error saying the mirror was degraded but it booted Windows and you could see an immediate change. Instead of Explorer crashing every 30 seconds and freezing up it ran just fine. When he launched his photo editing program it did freeze up while loading thumbnails. It was getting late so I set chkdsk to run at boot and rebooted, hoping chkdsk would fix any file problems on the drive. I left it and called the next day and he said everything seems to be running fine and I haven't heard from him since so I guess that's a good thing.

He still had a month left on his extended warranty with Dell so I called them up and got a replacement coming tomorrow.

My question now is, since I didn't mark the bad drive offline, will I still be able to just throw another drive in and rebuild the mirror or will I have to image the good drive, delete and logical drive, and recreate a new mirror then reload the image? Or will I be able to install the new drive and add it back to the mirrored set? I've been reading the PERC 6i manual on Dell's site and it doesn't really say if I didn't mark the drive as offline.

 

[imagoon]

You can just throw the new disk in and let it rebuild. It may auto rebuild or you may need to use the MSM to add the new drive. It depends on how the card was configured.

It also shows in this case Emulex was right for this drive. It was not properly returning errors so the drive never failed out of the array.

 

[sornywrx]

You can just throw the new disk in and let it rebuild. It may auto rebuild or you may need to use the MSM to add the new drive. It depends on how the card was configured.

It also shows in this case Emulex was right for this drive. It was not properly returning errors so the drive never failed out of the array.

I'm glad to hear that. Although imagine and redoing the drive wouldn't be a big problem I would rather not spend the time doing that unless I need to.

The bad drive was making some unusual sounds like some light clicking and taking a long time to spin up when compared to the other drive. I was surprised that, as bad it was, that it wasn't flagged as failing by SMART or the PERC.

Is it better to use the MSM program in Windows instead of booting into the PERC's BIOS at boot?

 

[imagoon]

I'm glad to hear that. Although imagine and redoing the drive wouldn't be a big problem I would rather not spend the time doing that unless I need to.

The bad drive was making some unusual sounds like some light clicking and taking a long time to spin up when compared to the other drive. I was surprised that, as bad it was, that it wasn't flagged as failing by SMART or the PERC.

Is it better to use the MSM program in Windows instead of booting into the PERC's BIOS at boot?

Either will have the same effect. The BIOS may rebuild faster since there is no other disk IO but the machine is unusable during the rebuild. The MSM will let you use the computer while the array rebuilds.

 

[sornywrx]

Either will have the same effect. The BIOS may rebuild faster since there is no other disk IO but the machine is unusable during the rebuild. The MSM will let you use the computer while the array rebuilds.

Excellent, thank you (and everyone else) for the help. I will be going over there tomorrow to install the new disk and I'll update the thread in case when its finished in case this helps anyone else.