Recently I rented a new server with an LSI MegaRAID SAS 9260-4i controller and - of course - installed VMware ESXi 6.0 on it. I will soon migrate my "production workloads" (the V-Front Online Depot, the ESXi Patch Tracker site and my Zimbra E-Mail server) onto this box to improve their availability. A disk failure is one of the most likely disaster scenarios, so a hardware RAID controller with two mirrored disks will certainly help.
After I had the host installed and configured I did a routine check of the log files and found a lot of SCSI errors like these in /var/log/vmkernel.log:
cpu1:34460)NMP: nmp_ThrottleLogForDevice:3178: Cmd 0x1a (0x439d80668b00, 0) to dev "naa.600605b0058b25f01c995a4f0b03da18" on path "vmhba1:C2:T0:L0" Failed: H:0x0 D:0x2 P:0x0 Valid sense data: 0x5 0x24 0x0. Act:NONE cpu1:32779)ScsiDeviceIO: 2646: Cmd(0x439d80716d00) 0x4d, CmdSN 0x1 from world 34425 to dev "naa.600605b0058b25f01c995a4f0b03da18" failed H:0x0 D:0x2 P:0x0 Valid sense data: 0x5 0x20 0x0.
These errors were referring to the RAID 1 disk, and first I was worried about them - according to KB289902 the D:0x2 translates to "Device: Check condition" and the sense data codes 0x5 0x24 0x0 and 0x5 0x20 0x0 mean "Illegal request", "Invalid Field in CDB", resp. "Invalid command operation code". I double checked that the RAID controller is on the VMware HCL, I searched the Internet for these errors, found few peoples' posts describing this issue, but no solution. I even updated the firmware of the controller, all to no avail ... and then, finally, I stumbled over KB1031221 which - basically - describes that these errors are "normal" and "can be safely ignored".
Well, that calmed me down... But I'm a log purist: I wanted to get rid of these false error messages in vmkernel.log!