CampusCluster Agent
Uses Volume Manager (VM) mirroring as a data mobility solution in a clustered environment for disaster recovery. The CampusCluster agent causes a fast mirror re-synch (FMR) to remote plexes that have experienced a temporary downtime and then re-connected.
 Caution
To use VM for a campus cluster, you must have expert knowledge of Volume Manager and VCS.
For more information on using this agent, see the VCS User's Manual.
Requirements
A soft requirement, from the clustered host's standpoint, is that you might want to distinguish the physical location of each disk either by controller number or enclosure name.
Several hard requirements exist for using this agent:
- You must have a single VCS cluster with at least one node in each of two sites, where the sites are separated by a physical distance of no more than 80 kilometers.
- All volumes that have application-required data must be mirrored across site boundaries, with at least one plex in each site.
- You must disable Volume Manager's Relocation Daemon. Otherwise in case of a temporary site outage, all plexes locate to the same site.
Limitations
Global Cluster Option is not supported.
Entry Point
Monitor---Parses output from vxnotify to determine when lost disks have returned. Upon determining that a site has been restored, initiates the steps to re-synch the lost disks if possible.
State Definitions
Online---CampusCluster resource type is always on.
Type Definition
type CampusCluster (
static int NumThreads = 1
static str ArgList[] = { DiskGroup, RemoteCtlr }
static str Operations = None
str DiskGroup
str RemoteCtlr
)
Required Attribute
| Required Attribute
|
Description, Type and Dimension, Default, and Example
|
|---|
DiskGroup
|
A string representing the name of the disk group to monitor.
|
Optional Attribute
| Optional Attribute
|
Description, Type and Dimension, Default, and Example
|
|---|
RemoteCtlr
|
Set this attribute if different controllers manage the disks at different sites from the standpoint of a host.
You might need to localize this attribute if the hosts in the cluster have different controllers that manage the remote disks. The value should be of the format "c#" where # is the controller number, e.g. c2. Setting this attribute minimizes the number of disks that need to be rescanned when a path returns and is for performance only. If a host has a single controller or if enclosure based naming is used do not set this attribute.
|
|
