Mageia forum

In the past I have always run hddtemp. So I just tonight installed it on my Cauldren Mageia 3 beta 4? system. All of a sudden I noticed I was taking errors on one of my 3ware RAID controllers.

Code: Select all


Mar 31 20:35:58 localhost.localdomain kernel: 3w-xxxx: scsi8: Unknown scsi opcode: 0x41
Mar 31 20:35:58 localhost.localdomain kernel: sd 8:0:0:0: [sdf] Unhandled error code
Mar 31 20:35:58 localhost.localdomain kernel: sd 8:0:0:0: [sdf] 
Mar 31 20:35:58 localhost.localdomain kernel: Result: hostbyte=DID_BAD_TARGET driverbyte=DRIVER_OK
Mar 31 20:35:58 localhost.localdomain kernel: sd 8:0:0:0: [sdf] CDB:
Mar 31 20:35:58 localhost.localdomain kernel: Write Same(10): 41 00 11 a1 95 48 00 00 08 00
Mar 31 20:35:58 localhost.localdomain kernel: end_request: I/O error, dev sdf, sector 295802184
Mar 31 20:35:58 localhost.localdomain kernel: sdf1: WRITE SAME failed. Manually zeroing.



Eventually I discover that /etc/sysconfig/hddtemp includes the following by default:

Code: Select all


#
# hddtemp(8) daemon options.  Add at least the disk(s) you want to monitor here.
#
HDDTEMP_OPTIONS="-l 127.0.0.1 /dev/hd? /dev/sd?"



That would seem to force hddtemp to probe all drives including the 3ware pseudo drives. That is not a good idea. I understand that if you don't include this, the user will complain that hddtemp is not working. I don't know what the solution is. But in the meantime, be forewarned. If you are running hardware RAID, don't enable hddtemp without fixing this config file. AND installing hddtemp causes it to start running upon installation.

- George

Well, apparently hddtemp is not the problem. It was quiet for a while and then started again. It seems to be happening when I have mcc running. Could mcc be probing my drives in some way to cause this. I just unmounted it and ran a full fsck on it and it is clean. I have changed the 3-ware card, so I know it is not the hardware. It has got to be related to something I installed that was previously installed on my Mageia 2 system. Strange.

Well, depending what you run within MCC, yes. Enable the logs in MCC via the menu, and also have a look in /var/log/explanations (IIRC)