Perc 5 driving me crazy

Steve Thompson smt at
Fri Apr 13 05:29:20 CDT 2012

On Fri, 13 Apr 2012, Chris Good wrote:

> We've seen exactly the same thing.  We noticed that it would always be 
> sequentially numbered drives that it marked failed so we took the machine
> down and juggled the drive order such that each half of the mirror was on
> opposite halves of the array.  That way when it failed it just got marked
> as degraded so we could just power the whole thing down at a convenient
> time and it rebuild it with no data loss.

Unfortunately, it appears to be completely random which drives are going 
to be marked as failed; they're not sequential. We've had failures on 
drives with Dell firmware as well as the non-Dell drives, and the drives 
themselves seem to work fine in other systems (although these are the only 
MD1000's that we have). I've even replaced the "failed" drives with brand 
new drives, and have failures on the new ones within days. There just 
doesn't seem to be a pattern emerging :(


More information about the Linux-PowerEdge mailing list