Seagate ST3500071FC ユーザーズマニュアル

ページ / 88
NL35 Series FC Product Manual, Rev. D
37
7.3
FC-AL system errors
Information on the reporting of operational errors or faults across the interface is given in the Fibre Channel 
Interface Manual
. The FCP Response returns information to the host about numerous kinds of errors or faults. 
The Receive Diagnostic Results reports the results of diagnostic operations performed by the drive.
Status returned by the drive to the initiator is described in the Fibre Channel Interface Manual. Status reporting 
plays a role in systems error management and its use in that respect is described in sections where the various 
commands are discussed.
7.4
Background Media Scan
Background Media Scan (BGMS) is a self-initiated media scan. It performs sequential reads across the entire 
pack of the media while the drive is idle. In RAID arrays, BGMS allows hot spare drives to be scanned for 
defects prior to being put into service by the host system. On regualr duty drives, if the host system makes use 
of the BGMS Log Page, it can avoid placing data in suspect locations on the media. Unrecovered error sites 
found during BGMS will be logged and recovered error sites will be logged or reallocated per ARRE/AWRE 
settings.
With BGMS, the host system can consume less power and system overhead by only checking BGMS status/
results rather than tying up the bus and consuming power in the process of host-initiated media scanning activ-
ity.
Since the background scan functions are only done during idle periods, BGMS causes a negligible impact to 
system performance. Background media scan will begin after 500ms of idle time. Other features that normally 
use idle time to function will function normally because BGMS functions for bursts of 800ms and then sus-
pends activity to allow other background functions to operate. Additionally, BGMS operates in segments of 400 
block reads, checking for host commands upon completion of each of the block segments. 
The BGMS function will complete any BGMS-initiated error recovery prior to returning to service host-initiated 
commands. Overhead associated with a return to host-servicing activity from BGMS only impacts the first com-
mand that interrupted BGMS. This results in a typical delay of about 4 ms.
7.4.1
Media Pre-Scan
Media Pre-Scan is an optional feature that allows the drive to handle unrecovered media errors that would oth-
erwise have been found by the host system during critical data accesses early in the integration process at 
customer sites. This should reduce integration DPPM (defective parts per million) by identifying and repairing 
media errors that arose during shipping and handling. Media Pre-Scan replaces all host-initiated Write com-
mands with Write-Verify commands when the drive is accessing areas of the media yet untouched by BGMS-
initiated scanning activity.
The drive will attempt to run BGMS on the entire pack of the drive. This will result in increasingly large portion 
so fthe drive having been accesed, provided there is adequate idle time to allow BGMS to run. During the Pre-
Scan period, all Write commands that are in un-BGMS-scanned areas are converted to Write-Verify com-
mands. To expedite the scan of the full pack and the subsequenty exit from the Pre-Scan period, BGMS will 
begin immediately when the drive goes to idle during the Pre-Scan period, as opposed to waiting 500ms as is 
done under non-Pre-Scan conditions.