The Controller Settings page appears when you choose Management >> Controller Settings from the menu bar.
Two Background Task settings allow you to control the balance of background tasks and I/O (reading and writing to disk) performed by the controller.
Background Task Rate lets you set the ratio at which background tasks are performed in comparison to I/O, on a scale from 5 to 1.
There are separate settings for Rebuild/Migrate Rate and Verify Rate. The Rebuild/Migrate Rate also applies to initialization. Although the same rate is used for rebuilding, migrating, and initializing, migrating has the highest priority.
Background Task Mode works with Background Task Rate, to address the issue of latency. Two settings are available:
Adaptive (the default) and
Low Latency.
Low Latency mode minimizes the latency (delay) in reading data from the RAID unit by slowing down the background task process. In contrast,
Adaptive mode allows the firmware to adjust the interaction of background tasks with host I/Os to maximize the speed of both host I/O and background tasks.
On the 3DM2 Controller Settings page you can enable or disable these policies:
Write Cache,
Auto-verify, Overwrite ECC, and Queuing You can also choose a
StorSave Profile and change Rapid RAID Recovery settings.
Write Cache. When write cache is enabled, data is stored locally in memory on the drive before it is written to the disk drive media, allowing the computer to continue with its next task. This improves performance. However, in the event of a power failure, the data in the controller's write cache will be lost if you do not have a battery backup unit (BBU) or an uninterruptable power supply (UPS).
Read Cache. The Read Cache unit policy lets you enable either
Basic read caching or
Intelligent read caching. You can also disable the Read Cache unit policy, if preferred. By default, Read Cache is set to
Intelligent.
Intelligent read caching includes a typical read-ahead caching method which is used to proactively retrieve data from media and store it locally on the controller with the anticipation that it may be requested by the host, and also includes the Intelligent Read Prefetch (IRP) feature, which can improve performance at higher queue depth in multiple read only or mixed read/write stream environments. For additional information, see
Working with Read Cache Settings.
Auto-Verify. Enabling Auto-Verify ensures that a verify takes place following either the Basic Auto-Verify schedule or the Advanced Auto-Verify schedule. For details, see
Enabling or Disabling Auto-Verify for a Unit.
When Auto-Verify is not enabled, verify tasks are only run if you manually request one on the 3DM2
Management page or in event of an unclean shutdown.
Overwrite ECC (Continue on Source Error When Rebuilding). This policy applies only to units which are redundant. (For units which are not redundant, a check box is not available.) When this policy is set, ECC errors are ignored when they are encountered during a rebuild. When this policy is not set, a rebuild will abort upon encountering an ECC error and the unit will be set back to Degraded.
Since this option could result in the loss of some source data in the event of source errors, select this option only if you want to ensure that a rebuild will complete successfully without manual intervention. If the rebuild fails and
Overwrite ECC is not selected, then you have the option to start a rebuild manually. It is recommended that you execute a file system check when the rebuild completes.
Under FreeBSD use the fsck utility command. For example,
fsck /dev/sda1.
Under Linux use the fsck utility command. For example,
fsck /dev/sda1.
On Mac OS, you can do this using the First Aid tab in the Disk Utility—select the disk on the left and then click
Verify Disk. If verification encounters problems, you can then use the
Repair Disk option on the same screen.
For OpenSolaris you should follow your operating system’s guide to perform a file system check.
Under Windows, you can do this by right-clicking on the Drive and choosing Properties; then on the Tools tab, click Check Now.
For VMware you should follow your operating system’s guide to perform a file system check.
|
|
Note: The policy Overwrite ECC is equivalent to the “Continue on Source Error” option in 3BM and the “IgnoreECC” option in the CLI.
|
Queuing. (Not applicable for SAS drives) This policy enables or disables Native Command Queuing (NCQ) for SATA drives in the unit. By default, queuing is disabled. You can enable it, if desired.
NCQ only operates when the feature is enabled at both the drive and the controller. If a drive does not support NCQ, the policy setting for the controller is ignored.
StorSave Profile. The StorSave feature includes an option that lets you change the StorSave Profile used for a unit. Three profiles are available: Protection, Balanced, and Performance. These profiles automatically adjust several different factors that affect protection and performance, including whether FUA (Forced Unit Access) is honored, whether Write Journaling is enabled, and whether Disable Cache on Degrade is enabled. For additional information, see
Setting the StorSave Profile for a Unit.
|
|
Note: If the write cache setting is disabled for a unit, the StorSave Profile capability does not apply and is automatically disabled.
|
Rapid RAID Recovery. Determines which Rapid RAID Recovery option to use,
All,
Rebuild or
Disable.
All will apply the policy to both rebuilds and unclean shutdowns.
Rebuild will apply only to rebuilds. You can switch between
All and
Rebuild settings. However, once Rapid RAID Recovery is disabled for the unit, it cannot be re-enabled. For additional information, see
Rapid RAID Recovery.
The Other Controller settings displays information about additional settings, some of which can only be changed in the BIOS (3BM) or CLI. 3BM is not available for Mac users.
Auto-Rebuild. The Auto-Rebuild policy determines how the controller firmware will attempt to rebuild degraded units.
When Auto-Rebuild is disabled, only spares will be used to automatically rebuild degraded units. When Auto-Rebuild is enabled, the firmware will automatically select drives to use for rebuilding a degraded unit using the following priority order.
Auto-Carving. Auto-carving can be enabled or disabled by selecting the appropriate radio button.
When this feature is enabled, any unit that is over a specified size (known as the
carve size) will be broken down into multiple volumes of that size, plus a remainder volume. The default carve size is 2048 GB (2 TB). For example, using the default carve size, if the unit is 2.5 TB then it will contain two volumes, with the first volume containing 2 TB and the second volume containing 0.5 TB. If the unit is 5.0 TB then it will contain 3 volumes, with the first two volumes containing 2 TB each and the last volume containing 1 TB.
Carve Size. Sets a size for dividing up units into volumes when Auto-Carving is enabled. This setting can be between 1024 and 2048 GB.
Number of Drives Per Spin-up. Number of drives that will spin up at the same time when the controller is powered up. (This setting only applies when the feature is supported by the disk drives.)
Delay between Spin-ups. The delay time (in seconds) between drive groups that spin up at one time on this particular controller.
Export Unconfigured Disks. This setting is not available for 9750 controllers.
Number of Controller Phys. The number of phy connections possible on the controller. For example the 9750-8i RAID controller has 8 phys and the
9750-24i4e controller has up to 28 phys.
The Update Firmware function allows you to update the firmware of your 3ware RAID controller to the latest version, which allows you to take advantage of new features and fixes. For additional information, see
Updating the Firmware Through 3DM2.
- Related Topics
-
-
-
-
-
-
-
-
