Bad Raid 50 Stripe!

[Apostle 83]

've been working on a new server and getting it in. I'm about to go crazy becaus emy $600 adaptec raid card is reporting a bad stripe and it won't tell me how to fix it. The server was functional today. I replaced two of the six drives to see if that would fix it (They are all Western Digitals). The server is fine for everyone on the network, but it is finnicky when logged directly onto it (W2k3 R2 SB) or when logged on using terminal services or remote desktop.

I'm almost through running a utility from the RAID card (just post boot) on the first drive which verifies and repairs what it can.

That said, if this doesn't fix it... do I have to wipe and start from scratch?

[Midknyte]

Makes and models of your parts would help.

Western Digitals in a RAID 50? So you're running SATA RAID? I hope they are Raptor drives.

How did you know which 2 drives to change? Did you run diagnostics on them?

[Apostle 83]

Thanks for your reply. I'm logged onto the server through remote desktop now. It's sluggish. Like I said, it's fulfilling it's role as domain server for all the local stuff as well as file storage ok. I am making a LOT of data backups throughout the day.

Sorry, I should have started with the info that you needed. The controller is an Adaptec 2820SA, the hard drives are Western Digital WD1600SB's. I'm a rather small company and my vendor told me they are very reliable... (barf). He's been very good so far but this is driving me up a wall!

I changed one drive because I got an error that I needed to. Initially (I'm new to raid) I opened the Adaptec utility, and I disabled said drive Things didn't get faster.

Right now, I have it running a "Build/Verify" on the entire array overnight. It's 27% complete.

What to do?

[Midknyte]

disabling a drive wouldn't help with speed. you wouldn't get back to normal until you replaced the drive and rebuilt the array.

most SATA drives are not made for 24x7 operation in an enterprise environment. Raptors would have been the best bet. I'm guessing that cost was the reason you went with SATA drives.

You're already doing a build, so there's nothing to do but sit and wait.

[Apostle 83]

Yeah. It's been running almost 3 hours and it's nearly 50% done. I'm gonna hit the hay and I guess I'll find out in the morning.

The weird thing is I'm getting a cray crypkey error that I think is from the previous write errors. I think crypkey is a piracy plugin... my CA etrust anti-virus realtime protection seems to have been a victim of crypkey going down. It always fails on load now. :(

[Apostle 83]

Well, it gone done and gave the error:

One or more logical devices contain a bad stripe: Controller 1.

[Midknyte]

stripe could refer to either of the RAID5 arrays or the RAID0 stripe of the arrays (50).

at this point it could be a bad drive. like I said, SATA drives aren't usually meant for enterprise use. Since it's a production server, you'll need to find some down time to run diags on each drive.

why did your company choose RAID50? Wouldn't a RAID5 array have been sufficient? You lose 2 drive's worth of space in RAID50 versus just 1 in RAID5.

[Apostle 83]

We just wanted the maximum redundancy possibly.

So I need to go in, pre boot, and check each drive? Is that the sum of things?

[Midknyte]

RAID6 would have been a better choice then. If 2 drives die in the same RAID5 array on a RAID50, you lose everything.

[Midknyte]

did you check the event viewer? it may say which disk has problems.

[Apostle 83]

Yeah, I checked it. It does not have that logged, that I can see.

[Apostle 83]

I have two open spots on it and it says I can create "hot swap" drives for it.

[Midknyte]

http://support.wdc.com/download/index.asp?cxml=n&pid=999&swid=3
run WD datalifeguard tools on each drive first. no sense in trying to rebuild the array on bad drives.

[Apostle 83]

Hmmm, doesn't look like that supports server 2003.

[Midknyte]

ah forgot about that. maybe seatools. it says it works on non-seagate drives. not sure if it'll run on 2003 or not, but it's free.

http://www.seagate.com/www/en-us/support/downloads/seatools/

otherwise, you'll have to down the server or remove the hd and run diags on another system.