Thunderstone Search Appliance Manual

RAID Array Management

Note: This area only applies to larger models (such as 3000) that include multiple hot-swap disks in a RAID-5 configuration.

Note: The pages in this area may load somewhat slowly as they collect information from the RAID controller.

Overview

Little to no maintenance of the RAID array is required. In the event of a disk failure the hot-spare will automatically take over and the array will be automatically rebuilt. The rebuild process takes several hours. After the array is rebuilt the failed disk will have to be manually removed (no shutdown required) and replaced with a same size or larger disk of the same type. After the disk is replaced it needs to be Added to the array.

Details

The RAID Status page displays a summary of the RAID's state. It's an abbreviated form of the information on the RAID Management page to provide a quick Good/Bad check.

The RAID Management page lists information about the overall RAID array as well as each of the hard disks in the system. Each item starts with a Status and is color-coded to indicate its state. Green is good, red is problem, blue is hot-spare disk, light blue is unused disk, yellow is verifying/testing.

The first line of the Storage table contains information about the overall array with the Use column set to Array. The remaining lines are individual disks, either Member, Spare, or None. Member disks are part of the RAID array. Spare disks are hot-spares that will take over for a failed member disk.

Each item in the Storage table has associated actions that may be taken.

Rebuild

An array that is in a non-optimal state may be forced into a rebuild to become optimal again.

Verify

Verifies the integrity of the parity information for the array. This is not generally needed as the array is automatically verified periodically as controlled by the hardware BIOS.

Fail

Forces an individual disk into a failed state so that it may be replaced. This is not generally needed as failures will be automatically detected.

Remove

This removes a hot-spare disk from the array. It then becomes an unassociated disk with a Use of None. All arrays should have a hot spare.

Add

This adds an unassociated disk with a Use of None to the array as a hot-spare.

The first number in the Disk:Addr column (everything up to the :) is the disk number which corresponds to labels on the front panel of the Parametric Search Appliance.

Rebuild/Verify Rate

The Rebuild/Verify Rate is how aggressively the RAID will rebuild. A higher rate will rebuild a partially failed array more quickly so that it's in a non-fault-tolerant state for the shortest possible time. The downside of the higher rate is that operations that use the disk such as walks and searches will see slower performance.

Controller

This table shows various model and version information about the RAID controller.

Command

The Command input box should not be used except at the request of Thunderstone technical support. It is for issuing arbitrary commands to the RAID controller. Putting the wrong thing in this field could irrevocably damage the RAID array and render your machine completely unusable! If the "Ok to run this command" checkbox is not also checked anything in the command input will be ignore.

Perform

The Perform button at the bottom will perform all of the actions selected on the form. You must also set "Are you sure you want to perform these actions?" to "Yes" or the actions will not be performed.

Front panel

This provides a rough approximation to the physical front panel of the Parametric Search Appliance. It shows the drive arrangement to aid in locating the proper disk when performing maintenance.