A server switchover can be triggered either manually from the user, or automatically with the Auto-Recovery function.
Regardless of what triggered a switchover, the user must decide whether to switch back applications to their original server (failback), or let the spare server indefinitely take over those applications (takeover). Choosing takeover will result in the spare server becoming the new active server.
For details, refer to "4.3 Post-Switchover Operations".
Different switchover methods are available according to each managed server's hardware configuration.
For details, refer to the "Functions Available for Agents" in "6.2.1 All Editions" in the "Overview".
Note
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.
During switchover, server restarts and configuration changes may trigger SNMP Traps (which are shown in the Event Log).
For details, refer to "Part 1 Messages Displayed by Virtual Edition and Cloud Edition" in "Messages".
When using the backup and restore switchover method, do not start up the original primary server during or after the switchover operation.
The primary server and spare server both run the same system image. Having the two servers running together will cause conflicts of IP addresses and other information. This can adversely affect the applications recovered on the spare server.
If it becomes necessary to start the primary server, for maintenance or other tasks, ensure that it does not start up from the same system image as that of the spare server. This can be done by turning off the spare server first, or by stopping the primary server at its BIOS screen (before startup of the OS).
When using PRIMERGY BX servers, the maintenance LED of a switched over server is automatically activated.
If switchover takes place when a spare server is operating, the spare server is turned off.
After shutting down a spare server, if the power is not turned off within 15 minutes, the spare server will be forcibly powered off.
When performing server switchover using the backup and restore method in a SAN data server environment that uses local boot, configure a target disk for image operations on both the primary server and the spare server.
If the switchover is performed without configuring the target disk for image operations, data may be overwritten on an unintended disk.
For details, refer to "9.1.13 Changing Target Disks of Image Operations" in the "User's Guide VE".
[VMware] [Hyper-V]
When switching over to the spare server of a VM host, the VM host of the spare server will be placed into VM maintenance mode.
[OVM for SPARC]
In OVM for SPARC environments, the XML files of the OVM configuration information used during server switchover are saved in the following location. Delete them if necessary.
[Control domain]
/etc/opt/FJSVrcvat/config
ovm_configyear_month_minute_second.xml
The year_month_minute_second at the end of the file name is UTC.
Before performing server switchover, if the XML file of the domain configuration information has not been saved in the control domain, the domain configuration cannot be recovered after server switchover. In this case, it will be necessary to reconfigure the domain configuration.
This is also true for the spare server that is being used as OVM for SPARC.
For details on how to save the XML files of the OVM configuration information, refer to "18.5.4 XML File of OVM Configuration Information" in the "User's Guide VE".
Auto-Recovery
For PRIMERGY BX servers
By enabling Auto-Recovery in the spare server settings of a server, it will be automatically switched over to a spare server when its status changes to either "error" or "fatal", and no response is obtained from its OS.
For rack mount servers, tower servers, SPARC Enterprise, and FUJITSU M10
By enabling Auto-Recovery in the spare server settings of a server, it will be automatically switched over to a spare server when it receives an "Error" level SNMP trap, and no response is obtained from its OS.
For the decision criteria for whether an OS has stopped, refer to "9.4 Conditions Required for Auto-Recovery" in the "Setup Guide VE".
Manual Switchover
Use the following procedure to manually switch over applications from a primary server to a spare server.
Manual switchover can be performed either at will when necessary, or to verify that the switchover process operates properly.
For details on the conditions for a server switchover, refer to "Conditions for Server Switchover" in "9.3 Server Switchover Conditions" in the "Setup Guide VE".
In the ROR console server resource tree, right-click the physical OS or VM host to be switched, and select [Spare Server]-[Switchover] from the popup menu.
The [Execute Switchover Operation] dialog is displayed.
Select the switchover destination server.
If "Automatic allocation" is selected, the switchover destination server is automatically selected.
Click the [OK] button.
The configuration for server switchover is started. The process status can be checked in the Recent Operations area of the ROR console. The primary server stops, and the physical OS or VM host starts on the spare server. For the backup and restore method, clicking the [Cancel] button in the Recent Operations area displays the confirmation dialog.
Note
If the server switchover/failback operation is canceled, the original server will be powered off. To continue server operations, power on the server. If the server switchover/failback operation is canceled, OS images that have the same information (such as IP addresses) as that of the original server may remain on the internal disk of the destination server.
Information
Manual switchover can be performed regardless of the status of the primary server.
If "Automatic allocation" is selected for Auto-Recovery or manual switchover, the spare server to be allocated will be selected based on the following order of priority:
Physical servers with no settings configured
Physical servers for which HBA address rename information or a VIOM server profile is configured
Physical servers with no agents registered
Spare servers are selected based on the names of their physical servers, in the following order of priority.
"-"
Numbers
Uppercase letters
Lowercase letters
