Product specifications

Redistributing Data
83
Symmetric Configuration Redistribution
Full redistributions happen faster on workspaces that have had symmetric configuration changes. For example, the
same number of ISBs that were removed from a storage group are also added in a single configuration change. For
more examples of symmetric configuration changes search the Avid Knowledge Base for the Avid ISIS Performance
and Redistribution Guide. As with all full redistributions, all workspaces in the storage group should be selected for
the full redistribution.
A full redistribution is still recommended at 20 configuration changes. At 30 configuration changes the Avid ISIS
software automatically performs a full redistribution.
Automatic Redistribution on Disk Failure
Avid ISIS performs an automatic redistribution on Disk Failure notification. Storage Managers continuously
monitor disk status and sends a “Disk Failed” notification to the System Director upon determination that a disk is
not usable. The System Director then removes the Storage Manager from its associated Storage Group. The removal
of the Storage Manager from the Storage Group initiates redistributions on all workspaces associated with that
Storage Group. The System Director then prevents the Storage Manager that reported the disk failure from being
added to a Storage Group.
The Automatic Removal feature is controlled by a system preference that is configured using the ISIS Management
Console Preferences tool. The preference is called “Auto Remove Redistribution on Disk Failure” The default
setting is “enabled.” The Storage Manager is only removed from the Storage Group. It is not unbound from the ISIS
Systems for the following reasons:
• The Storage Manager participates in the removal redistribution. This is required to prevent data loss when full
redundancy has not been attained for all data blocks.
• The Storage Manager continues to report status, such as physically replacing the blade.
Auto removal status is reported by way of the System Event Log. A sequence of event log entries are generated for
the initial report of the failure, the decision to remove the Storage Manager, and subsequent success or failure.
Events are also logged for the start and stop of all the workspace redistributions.
When using Automatic Redistribution:
• Keep a certain percentage of your Storage Group unused at all times. Failure to do so can cause the system to
run out of space after an automatic redistribution has initiated and cause a client outage. The recommended
amount of free space varies by ISIS platform; for more information, see
“Minimum Free Storage Space
Requirement” on page 69
.
• If an automatic redistribution of a Storage Group is started during a critical time the overall system performance
can degrade significantly and be disruptive. For more information on understanding the performance
characteristics during remove redistribution, see the Avid ISIS Performance and Redistribution Guide.
• An automatic redistribution removes the Storage Manager from the Storage Group but not from the file system.
Once an automatic redistribution has occurred and completed, the removed Storage Manager should be
removed from the file system. You must first remove the Storage Manager with the software and then physically
replace it. A replacement ISB must then be added to the file system and then the Storage Group, triggering
another redistribution.
• If unmirrored workspaces are in use, the data will be damaged on those workspaces in that Storage Group.