6.1.5 Processing performed when failover occurs on SLM - Manager
When failover occurs on SLM - Manager, the service performance information that has been written onto the shared disk by SLM - Manager on the active server up to that point is inherited by SLM - Manager on the standby server. The standby server resumes business operations on the basis of the inherited information. However, monitored-service detection processing is not resumed.
The following figure shows the processing that is performed when failover occurs on SLM - Manager.
|
![[Figure]](GRAPHICS/F0000144.GIF)
When failover occurs, the service performance is inherited, but the icons indicating error and warning statuses displayed on the window are not inherited. When the service is restarted after failover processing, the service performance is re-analyzed and the appropriate icons are displayed based on the analysis results.
The following subsections explain the failover processing flows depending on the timing of the failover.
- Organization of this subsection
(1) While neither monitoring nor detection of monitored services is being performed
When failover occurs on SLM - Manager while neither monitoring nor detection of monitored services is being performed, only the connection to SLM - UR is restored after SLM - Manager is started on the standby server. Neither monitoring nor detection processing is started.
(2) While monitored services are being monitored
When failover occurs on SLM - Manager while monitored services are being monitored, the monitoring status is restored to what it was immediately before the failure, on the basis of the service performance information in the database on the shared disk after SLM - Manager has started on the standby server.
The following figure shows the processing flow when failover occurs on SLM - Manager while monitored services are being monitored.
|
![[Figure]](GRAPHICS/F0000216.GIF)
The following explains the processing flow shown in the figure, where the numbers correspond to the numbers in the figure:
The cluster software (active server) starts SLM - Manager (active server). The cluster software (active server) also starts periodic monitoring of server status.
Note that you must start SLM - UR manually. For details about how to start SLM - UR manually, see 2.1.2 Starting SLM - UR.
SLM - Manager (active server) starts monitoring the monitored services.
Because a failure has occurred on SLM - Manager (active server), the Windows services stop.
All Windows services for SLM - Manager that have been registered into SLM - Manager (active server) are stopped by the cluster software (active server), after which failover processing starts.
The cluster software (standby server) starts SLM - Manager (standby server). The cluster software (standby server) also starts periodic monitoring of server status.
The cluster software (standby server) restarts monitoring of the monitored services by SLM - Manager (standby server). The service performance collected by SLM - UR is sent to SLM - Manager (standby server).
(3) While detection of monitored services is being performed
When failover occurs on SLM - Manager while detection of monitored services is being performed, the detection processing is canceled.
The following figure shows the processing flow when failover occurs on SLM - Manager while monitored services are being detected.
|
![[Figure]](GRAPHICS/F0000188.GIF)
The following explains the processing flow shown in the figure, where the numbers correspond to the numbers in the figure:
The cluster software (active server) starts SLM - Manager (active server). The cluster software (active server) also starts periodic monitoring of server status.
Note that you must start SLM - UR manually. For details about how to start SLM - UR manually, see 2.1.2 Starting SLM - UR.
SLM - Manager (active server) starts detecting monitored services.
Because a failure has occurred on SLM - Manager (active server), the Windows services stop.
All Windows services for SLM - Manager that have been registered into SLM - Manager (active server) are stopped by the cluster software (active server), after which failover processing starts.
The cluster software (standby server) starts SLM - Manager (standby server). The cluster software (standby server) also starts periodic monitoring of server status.
The cluster software (standby server) stops detection of monitored services by SLM - Manager (standby server). It also sends a notification to SLM - UR indicating that detection of monitored services is stopped.