How to reset a single drive?

How to reset a single drive?

[joystick]

Hello,
it just happened that there was an unresponsive HDD in the system (a WD 
RE3 1TB SATA drive).
Multiple reboots of the machine (not poweroffs) did not revive it, and 
the server did not create a /dev device for it on subsequent reboots. So 
I thought it had failed.
But when I tried removing and inserting it in another hot-swap slot it 
started working again. After reinserting it in the original slot also it 
was working.
Probably a poweroff of the server would also have revived it.
FWIW: smartctl -x tells no error is logged by the drive.
It could also have been a controller glitch maybe?

For a future time, I know there are SCSI reset messages which can be 
sent to various things in the path to the drive. Could someone teach me 
how to send a reset message to the controller card, to a slot and/or to 
a drive, from bash?
Or any other suggestion for reviving a slot / drive which do not imply 
powering off the machine?

Our controller is LSI 6201-16i. No SAS expanders. I am able to determine 
the phy port number of the unresponsive drive.

Thanks for suggestions
J.

Re: How to reset a single drive?

[Emmanuel Florac]

Le Tue, 13 Mar 2012 17:44:35 +0100
joystick <joystick@shiftmail.org> écrivait:

> It could also have been a controller glitch maybe?
> 
> For a future time, I know there are SCSI reset messages which can be 
> sent to various things in the path to the drive. Could someone teach
> me how to send a reset message to the controller card, to a slot
> and/or to a drive, from bash?
> Or any other suggestion for reviving a slot / drive which do not
> imply powering off the machine?
> 
> Our controller is LSI 6201-16i. No SAS expanders. I am able to
> determine the phy port number of the unresponsive drive.

You're out of luck here. Usually removing the driver (rmmod mpt_sas) is
the best you can do, it almost always force a complete scsi bus reset
and rescan, but that obviously won't do anything more than a reboot. Of
course you could try to force the drive offline, then online with
MegaCli:

MegaCli -PDOffline -PhysDrv [E:S] -aN
MegaCli -PDOnline -PhysDrv [E:S] -aN

Or you could try a controller scan:

MegaCli -CfgForeign -Scan -aALL 

You're obviously either got an LSI firmware bug or a drive
firmware bug. The fact that several reboots and therefore controller
firmware restarts occured makes me thing that's the drive.  WD SATA
drives are notoriously complete sh*te in professional environment and
love living their life by themselves, out of RAID arrays.

-- 
------------------------------------------------------------------------
Emmanuel Florac     |   Direction technique
                    |   Intellique
                    |	<eflorac@intellique.com>
                    |   +33 1 78 94 84 02
------------------------------------------------------------------------
--
To unsubscribe from this list: send the line "unsubscribe linux-scsi" in
the body of a message to majordomo@vger.kernel.org
More majordomo info at  http://vger.kernel.org/majordomo-info.html