Clustered Drive corruption

[Guest Jake]

I have approx 50 Win2k AS 2-node File and Print clusters connected to IBM

FastT200 Fibre Arrays, configured with a Q: quorum drive, D: data drive and 2

spooler drives for Windows and Unix printing.

The D data drives continually get MFT corruption, eventually causing drive

failure, however there is no corruption on any other drive. The d: Data

drive contains user profiles, homedirs, shareddata and shared folders for

application data and DHCP and WINS db's. Although a the majority of I/O is

on this drive it is well within the limits of the hardware. After running a

chkdsk d: /f on the drive a read only chkdsk shows no corruption. However a

day later the MFT corruption is back even if there has been no failover.

 

I have experimented with different block and stripe sizes and with just

about every setting available on the Fastt200 FC array but the corruption

remains.

 

Any ideas or anyone had similar issues?

 

Thanks

 

Jake

[Guest Paul]

Re: Clustered Drive corruption

 

I am having similar experiences. I think it is hardware related, probably a

flaky hdd ribbon cable, since that was the only thing I changed recently.

The error seems to come and go randomly. Am scheduling a shutdown over the

weekend to swap out the cables so can't tell you now if it is the cause. But

if you have been swapping h/w recently, would bet something with h/w is the

source of your problem.

 

"Jake" <Jake@discussions.microsoft.com> wrote in message

news:F95780D3-8B7A-4792-B846-BEF65691EC43@microsoft.com...

> I have approx 50 Win2k AS 2-node File and Print clusters connected to IBM

> FastT200 Fibre Arrays, configured with a Q: quorum drive, D: data drive

and 2

> spooler drives for Windows and Unix printing.

> The D data drives continually get MFT corruption, eventually causing drive

> failure, however there is no corruption on any other drive. The d: Data

> drive contains user profiles, homedirs, shareddata and shared folders for

> application data and DHCP and WINS db's. Although a the majority of I/O

is

> on this drive it is well within the limits of the hardware. After running

a

> chkdsk d: /f on the drive a read only chkdsk shows no corruption. However

a

> day later the MFT corruption is back even if there has been no failover.

>

> I have experimented with different block and stripe sizes and with just

> about every setting available on the Fastt200 FC array but the corruption

> remains.

>

> Any ideas or anyone had similar issues?

>

> Thanks

>

> Jake