Disk Arrays
cancel
Showing results for 
Search instead for 
Did you mean: 

Smart Array 641 disk failure

Simon Pooler
Occasional Visitor

Smart Array 641 disk failure

I have a SmartArray 641 array running RAID 5 across 4 disks. One of the disks (slot 0) reported a predictive failure and (subsequently turned into DISK I/O errors) and while we were waiting for a replacement disk a second drive (slot 3) actually failed. I have had the contents of the slot 0 drive imaged onto an equivalent new disk.

Does anyone know whether the 641 is clever enough to recognise the contents of the new disk as being the data for the old slot 0 disk, or whether it will just flag the new disk as being a replacement and fail the logical drive based on the fact it will have a different serial number?

I think I read somewhere that SMART array controllers store the details of the physical drive as part of the data held on the disk. Can anyone confirm this?
2 REPLIES
Steven Clementi
Honored Contributor

Re: Smart Array 641 disk failure

Simon:

The smart Array DOES store array configuration on the drive, but it is not likely to recognize the new disk as usable.

1st, the data must be different now on the rest of the disks so that the parity information is not accurate. Even if the controller picked up the drive and placed the array into a reduced state, your data would most likely be corrupted anyway.

2nd, it is more likely that the array will see a "new" drive, but at this point... there is nothing it can do since the array lost 2 drives.

Hope you have ag ood backup of the data.


Steven
Steven Clementi
HP Master ASE, Storage and Clustering
MCSE (NT 4.0, W2K, W2K3)
VCP (ESX2, Vi3, vSphere4, vSphere5)
RHCE
NPP3 (Nutanix Platform Professional)
Simon Pooler
Occasional Visitor

Re: Smart Array 641 disk failure

Steve,

As the disk in slot 0 was still functioning when the disk in slot 3 failed, I believe the most uptodate set of data will be in disks 0,1, and 2. hence there shouldn't be a parity problem.

The machine was able to boot successfully (for a while) after the disk in slot 3 failed completely. My problem was that the rebuild of slot 3 onto a new drive failed persumable because the controller took exception to the drive IO errors during the rebuilt from slot 0,1, and 2 onto slot 3.

If I plug disk 0 into the controller it still shows as OK. so in teory I only have 1 failed drive. It's just the drive I/O errors mean that the rebuilt of slot 3 starts then stops.