Date - 07/2007
Affected Product & Version - NetVault: Backup ALL
Affected Module & Version - N/A
OS Version - Linux 2.4 and some 2.6 kernels (Red Hat and variations)
Application Information - N/A
Description:
This article describes how to troubleshoot and possibly resolve cases where a SCSI connected standalone drive to linux servers, fails to function with NetVault and logs the following error message:
"Unemulated command" leading to a "Command failed, re-initializing driver"
Symptoms:
The scenario is as follows: A linux Netvault Server (2.4 and 2.6 kernels) has a SCSI attached stand-alone drive that no longer functions and fills the NetVault logs with the following messages:
Example:
Background 2007/10/18 05:00:13 0 Media NVPE Starting up device LTO3:NVPE
Background 2007/10/18 05:00:13 0 Media NVPE Device 'LTO3:NVPE' check media
Error 2007/10/18 05:00:13 0 Media NVPE Command failed, re-initializing driver
The additonal log context to the error shows the following trace information:
TAPESCSI 48 0 40013 Check Media (BEGIN)
CAM 159 5 40013 (LINUX) FAILED TO FLAGS: Input/output error
SCSI 14 0 40013 CMD INQUIRY [ 12 00 00 00 32 00 ]
CAM 102 0 40013 Unemulated command 0x12
CAM 103 5 40013 Scsi command failed [12]: Input/output error
SCSI 15 0 40013 CMD failed: errno 5
SCSI 38 0 40013 Command failed, re-initializing driver
SCSI 14 0 40013 CMD PREVENT/ALLOW MEDIA REMOVAL [ 1e 00 00 00 00 00 ]
CAM 101 5 40013 Failed to ALLOW/PREVENT drive: Input/output error
Although the error "Command failed, re-initializing driver" means that the Netvault Device Manager process has not received any response to its SCSI command sent to the device, we should be careful in quickly concluding that there is a fault in the SCSI subsystem and flag this as a hardware error.
If you take a closer look at the trace, you will notice that the SCSI comm
1) run 'modprobe sg' to make sure the sg driver is loaded
2) remove and readd your device
3) make sure this module will always be autoloaded at reboot. There are many ways of achieving this and you should contact your OS support for this.
As an example, the following is a way of making sure this sg module is loaded at boot in Red hat.
“mkinitrd --with=sg /boot/initrd-2.6.9-42.ELsmp.img 2.6.9-42.ELsmp”
where 2.6.9-42.EL.smp is the exact name of your kernel. This will create an initrd file that will automatically include the autoloading of sg.
Please contact your OS support for more detailed procedures.
© 2024 Quest Software Inc. ALL RIGHTS RESERVED. Terms of Use Privacy Cookie Preference Center