12-02-2009 02:21 PM
I've inherited almost 200 2U servers at my current job, and as they are hitting 2-years of steady, rigourous use I am beginning to get a number of drives failing on our SAS Raid arrays. Most of the drives we use are Seagate and Hitachi - the Seagate drives are mostly model ST3300655SS.
After swapping in a new drive, I bring the failed drives back to the office and test them on a small 2-drive SAS/SATA enclosure, plugged into an LSI megaraid card on a spare desktop, which runs Win2003 Server 64bit. Once I can see the drive in the Megaraid management program I run SeaTools for Windows to test the drive. Some drives pass all the all tests, so I clear the foreign configuration, reformat them using the SeaTools advanced features, and then look for Raid array controller problems on the original server. These drives I've successfully put back into my other servers as replacements with no issues.
Some drives fail the short or long generic tests, or the short drive self test or long drive self test. When some drives are mounted I can't even launch the SeaTools program - it crashes with the error popup box stating "Argument 'Length' must be greater or equal to zero. System.string Mid(System.String, Int32, Int32).
So - I find it hard to believe I am seeing so many drives failing - on average of one per week. Only about 1 in 5 failed drives is actually a RAID array controller problem and actually a good drive.
My question - is there a firmware update and update tool I can download for these drives, to see if that reduces the failures? If not, is there another testing program I can download or purchase which will tell me exactly why these drives fail - for example, bad sectores, etc....?
12-11-2009 12:03 PM