P410i Array Controllers Handling Predictive Failures

[mikegw]

I've had the same problem on various servers using P410i array controllers. Replacing drives in predictive failure with known good drives will not clear the predictive failure. The drives are replaced and go through the rebuild process only to come back online and starting flashing the predictive failure again. As far as I can tell there are no problems. I've had to go so far as to replace all the drives with exactly the same part number/model number drives and completely redo the array and reload the OS. It's acting like the errors aren't cleared out of the controller when the drive is replaced although the ADU shows the new serial numbers. Firmware is up to date and the problem has been occuring through a couple of releases. Any ideas as I currently have one incedent outstanding and I don't want to have them start from scratch just to clear the led.

[Robert_Jewell]

Surely not normal.  Check the ADU for each drive to see if you can tell why its in PF state.  Read/write errors, etc.  Its not impossible to have a bad batch of disks; though unlikely.

-Bob

[mikegw]

Thanks for the reply. I've talked to a few people and the closest thing I can come up with is a "Punctured Array". A perscribed fix is to delete the array, move all the drives one slot up with the last drive going back to the first slot in the array and recreate the array.