"Raid Failing" ?

[gorby]

"Raid Failing" ? SORTED NOW!

Does any one know what might be causeing windows to display a message saying my raid is failing?

I have two maxtor 80's SATA @ RAID 1 the volumes on it all seem fine its just error msg. I have defragged / scan disk'ed all volumes on the drives.

MSI 865pe neo2 Raid on the intel controller

 

[Epicurus]

whats the exact error message??

 

[gorby]

thats it m8, it sone of the balloons that comes up in the sys tray. Ive checked the manager at boot time and that also sais one drive of the raid volume is failing...

found a guy on msi.com with extaclly the same prob, but no solution

http://forum.msi.com.tw/thread.php?postid=479843#post479843

 

[Epicurus]

this may or may not help. While I was looking for some more info on this problem I saw a post that suggested that it might be a power shortage problem. Guess you have a good PSU???? What else do you have in your system?? have you tried taking out whats not needed while you trouble shoot?

 

[gorby]

not my pc but its a crappy psu that came with a chaintech case

Ill take my antec round tomorrow

cheers

 

[Epicurus]

fingers crossed is that simply. Let me know how you get on

 

[gorby]

No joy at my m8's house unfortunately.

Tried my antec tru 550 and still the same issue.

the windows balloon was actually a warning from the intel Application accelerator.

It still boots but the drive is still stated as failing. Tried updating the bios etc to no avail. Ive also noticed that on one of the partitions some 1 kb files are reported as being 10 megs aiieee we dont have much time!

Whilst moving around the net searching for a fix i saw some people mentioning that i you can upgrade the array to mirror as well.

So what i plan to do is buy a PATA 160 GB drive, add that to the array and set it up as raid 0 + 1. Do you think its possible to do that, remove the mirrored drive and then recreate the raid volume on the two maxtor disks?

hmmmmm

 

[Epicurus]

I take it you have broken the raid and reinstalled everything?? or not?

 

[gorby]

No i havent 'actioned' as yet

the system boots still so im waiting until i get the 160gb drive.

I will then attempt to migrate the raid volume to one disk (2X 80's onto the 160) (thru ghost or somethign like that) or makeing my current array (raid 0) raid 0 + 1, then pulling out the 2 drives and recreating on them.

 

[Epicurus]

humm.. fun fun fun...

I just updated my bios and thats messed my raid up. DOH!!!!!

on well, just as well there was nothing really on it...

 

[gorby]

No real solution here, but if anyone’s interested this is what I did.

This was the problem...

2 X maxtors @ raid 0, moderate overclock (220 fsb) WITH pci/agp locked.

Everything fine for a while then one day application accel starts whingeing about a drive failing. Did what low level checks i could on drive. (plugged individual drive into my drive repair / check board, confirmed there was no hardware fault.)

So how to backup? There was quite a lot critical data here so starting from scratch was not an option

went and bought a 160gb pata drive...

Plugged it in, and messed around with ghost. Ghost only worked for me if I did raw sector copying. However! my MFT and MBR had become corrupt. (many of the files on the drive were being misreported. (eg a 50 mb autoexec.bat file) so a complete restore was out of the question, something more extreme was needed. Step in linux

I already have a copy of debian* with file tools installed onto a cd, so i booted from that and used a lovely gnu program called partimage. No fancy gui but you can specify common raid failure patterns. It took a while to get my switches to work (you need to specify raid stripe size in bytes and things like that) but eventually it read / repaired / copied the data fine onto my 160 gb disk.**

That’s all the sweating over and done with now the easy part.

Just use what ever your comfortable with getting the image back onto the raid. I deleted and recreated my raid array (horray 2X green status ok's!) Booted of my linux disk again, got all the drives visible and sent the image back onto the new raid partition. Rebooted job done!

*afaik the only way to get support for all the latest bells and whistles on our uber tech boards is to either load about a million new modules on to a 2.4 kernel or use something a bit more bleeding edge 2.6.9 works a treat for me.

**nb: one word of warning!

Don’t try, (as I did) to boot of the ide disk with ur backup. It got to windows fine then the application accelerator started to 3h1t its pants because there was no raid . It settled down eventually but then after applying my image to the newly formatted array, it wouldn’t boot because it was trying to use the standard ata controller to access the raid and the BSOD'ing. Probably could have fixed this, but not worth the hassle. Just use something else to verify your written data after image has been made. (easy way to check for me was the block size on individual files. on the first image the files were still like 1kb but 10 megs on disk)

Hope this helps some one one day

 

[standin wave]

Useful post there gorby m8. I once cloned a Hewlett Packard 5.25" disc onto 3.5" format using a BBC Micro - the HP rep nearly had a thrombie - just shows a little lateral thinking can lead to a solution to seemingly impossible problems!

sw