Proware Fibre to SAS/SATA RAID Subsystem, Руководство пользователя

Страница: 88 / 122

Fibre to SAS/SATA RAID Subsystem
88 User Manual
TLER Setting
TLER (time-limited error recovery) functions provide support for WD Caviar RE (RAID)
series disks. This is a new option from WD to support RAID features that were
traditionally missing from standard desktop drives. TLER is a method of signaling the
system RAID controller in the event that an error recovery process is taking longer
than time-out specifications allow. This prevents the RAID controller from dropping
the drive from the array during this period. Default value is manufacture setting. You
can select between 5, 6 and 7 second. This feature is to setup the HDD internal
timeout value.
Timeout Setting
Disk time-out is a registry setting that defines the time that RAID controller will wait
for a hard disk to respond to a command. You can modify the retry value by entering
a new value in the edit box beside this button and then selecting the button.
Normally you should not need to modify this value. Default value is 12 seconds: You
can select between 8~32 second.
Number of Retries
This setting determines the number of access that will be attempted before the
current command from the RAID controller to the disk drive is aborted. You can
modify the retry value by entering a new value in the edit box beside this button and
then selecting the button. Normally you should not need to modify this value. There
are two selections, either 2 retry or 3 retry.
Buffer Threshold
This new feature there are 4 options; 5%, 25%, 50%, 75%. The percentage
represents how much data should be kept in resident cache memory (how full cache
should get) before controller starts to flush data onto the hard drives. If the buffer is
set for 25%, then all 25% will be cached and is used for writing data. The remaining
cache memory will be used for reading and other system overhead. Write buffer
threshold for 5% is added for video recording. This option will push data to disk early.
This feature gives controller extra buffer time in case of slow response from the hard
drives within a given time. Consequently, this can prevent a pause in data flow and
there will be continues data access and stream. This feature is very useful for the
video streaming applications where there is high demand for constant non-stop data
flow with no interruption due to lower performance of specific hardware.
Amount of Read Ahead
Read-Ahead data is buffered in the RAID controller cache, however, thereby cutting
down on the amount of I/O traffic to the disk. The Amount of Read Ahead defines
how many data of reading at a time, making more efficient use of the RAID
subsystem. This makes it possible to locate and re-issue the data without repetitive
hard parsing activities.
The Amount of Read Ahead parameter is used to allocate an amount of memory in
the cache memory the frequently executed queries and return the result set back to
the host without real disk read execution. Default value is Auto: Controller will base
on the HDD number to set the amount of Read Ahead value. You can select between
512KB ~ 16MB.