I had a drive going bad in a RAID-1 set. Replaced it, rebuilt the RAID, everything running fine. When I look in the warn.log file it’s full of end_request: i/o errors. The problem is those errors are on 2 drives that are not even in the system. I mounted the bad hard drive in a hot swap bay to wipe it before doing an RMA on it. It came up as sdf, I tried formatting it but that failed and the partition table was wiped. Drive was never mounted so I removed it from the hot swap bay. I then figured it was a good time to do a backup so I put 3 drives into the hot swap bays. Fired up partitioner to format them and got the following:
“The partitioning on disk /dev/sdf is not readable by the partitioning tool parted, which is used to change the partition table. You can use the partitions on /dev/sdf as they are, format them and assign mount points, but you cannot add, edit, resize or remove partitions from the disk with this tool.”
Of course the drive that was /def/sdf was sitting on the desk. Figured maybe it did mount so I popped it back in the remaining hot swap bay (other 3 drives were still in there) and it came up as sde. The other drives were listed as sdg, sdh, and sdi. This was unusual since they are unusual since they normally come up as sdc sdd and sde. I had a usb stick in which was taking sdd so I didn’t think too much about it. Removed the bad drive again. The error message when loading partitioner still shows up for sdf. The warn.log is full of messages about sdf and a few about sde.
How the heck do I fix this so the system knows that sde and sdf don’t exist so the warnings and partitioner error go away?