This appendix provides important reminders for the operation of Resource Coordinator VE.
Server Switchover
In configurations where a server OS is operating on a spare server, if the boot methods of primary servers and spare servers are different, server switchover may fail and damage the primary server OS. Ensure that the boot methods are the same.
It is recommended that system images of primary servers are regularly backed up in order to prepare for unexpected situations.
Servers can be set as spare servers even if it may not be possible to switch over with them.
Ensure that server switchover is possible after setting up spare servers.
For details on server switchover conditions, refer to "9.3 Server Switchover Conditions" of the "ServerView Resource Coordinator VE Setup Guide".
Auto-Recovery cannot be performed if maintenance mode has not been released.
Ensure maintenance mode is released once maintenance operations are complete.
Note the following points regarding the occurrence of a fault in the switchover destination spare server during server switchover.
The server switchover state returns to pre-switchover state. Additionally, maintenance mode is set for the switchover source server, and server switchover processing ends. If the switchover source server has been stopped, the switchover source server does not start.
Even when more than one spare server is set for the primary server, there is no automatic switch to another spare server. Specify another spare server, then perform server switchover manually.
To perform operations on the switchover source server, start the server and then release maintenance mode. However, when Auto-Recovery is enabled and the switchover source server is faulty, Auto-Recovery occurs when maintenance mode is released. Restore the switchover source server to its pre-fault status and then release maintenance mode.
Redundancy Configurations for the Admin LAN
If communication issues occur on the admin LAN, or one of the network interfaces used by a managed server on the admin LAN fails, the following operations may result in errors. In such cases, restore the admin LAN network as quickly as possible.
Backup and restore operations
Collection and deployment of cloning images
Server switchover and failback
HBA Address Rename
With Resource Coordinator VE, the factory-set WWN of a managed server's HBA is overridden when the HBA address rename function is used. The WWN is reset to its factory-set value when the server is deleted from Resource Coordinator VE.
Before using HBAs in an environment that is not managed by Resource Coordinator VE, first delete the server in which it is mounted using the RC console.
For information on deleting servers, refer to "6.4.2 Deleting Managed Servers" of the "ServerView Resource Coordinator VE Setup Guide".
The WWN of a managed server is set during startup, using a network boot session to connect to the admin server. Once set up with a proper WWN, the managed server reboots into its own Operating System.
Therefore, a managed server may reboot during its startup.
Do not move HBAs whose HBA address rename settings have been set up to different managed servers. If operating HBAs without resetting their WWNs, when the same WWN is configured on multiple servers data may be damaged by same volume access.
ETERNUS SF Storage Cruiser Coordination
When using ETERNUS SF Storage Cruiser integration for Resource Coordinator VE, zoning of Fibre Channel switches connecting to the HBAs of managed servers, and host affinity configurations for storage units will be changed by the WWN information settings. When deleting a server, the relevant zoning and host affinity settings are also deleted.
For information on deleting servers, refer to "6.4.2 Deleting Managed Servers" of the "ServerView Resource Coordinator VE Setup Guide".
When configuring WWN information during OS operation, do not delete existing configurations. If a server is deleted accidentally, users may be unable to access the disk, the OS may hang, or the contained data may be damaged.
Changing the Manager's System Time
When the admin server's system time is reset to a time in the past, the resource monitoring by the manager stops for this period. To reset the system time to more than just a few minutes in the past, return the time and then restart the manager.
To restart the manager, refer to "5.1 Manager" in the "ServerView Resource Coordinator VE Setup Guide".
Restarting Managers
By default, the manager services restart at 3:45 am everyday for stable system operation.
The settings for restarting can be changed depending on the authority level. To change the configuration, perform the following:
Configuration File
[Windows]
Installation_folder\Manager\rails\config\rcx\rcx_manager_params.rb
[Linux]
/opt/FJSVrcvmr/rails/config/rcx/rcx_manager_params.rb
Configuration Parameters
Parameter | Meaning | Default Value |
|---|---|---|
RESTART_ENABLE | Select the restart operation status. | true |
RESTART_HOUR | Specify the restart time (hour) from 0 to 23. | 3 |
RESTART_MIN | Specify the restart time (minutes) from 0 to 59. | 45 |
RESTART_CYCLE | Specify the restart interval (days) from 1 to 5. | 1 |
Parameter change procedure
Stop the manager.
Use an editor and change the parameters of the rcx_manager_params.rb file.
Restart the manager.
For details on how to stop and restart the manager, refer to "5.1 Manager" in the "ServerView Resource Coordinator VE Setup Guide".
Note
The conditions for restarting are, that more than RESTART_CYCLE * 24 hours have passed since manager was started and it is the time specified for RESTART_HOUR and RESTART_MIN.
