Maintenance work on the database manager system hosting the UIM database needs to be performed and is estimated to take 5 hours.
Which UIM components need to be shutdown during this period and in what order should they be shut down? What order would you start up the UIM components?
The following provides suggested guidelines for shutting down UIM components when the database manager is going to be down for an extended period of time.
1. On all robots where the UMP, Unified Reporter, and/or CABI components are deployed, stop the UIM service.
2. If an HA backup hub exists in the UIM domain, stop the service on this hub
3 Stop the UIM service on the primary hub
All UIM messages (QoS, alarm, probeDiscovery, etc) will back up on the robots and all other hubs in the UIM domain. These messages will be stored in local queue files on these systems until the primary hub is back up and running. If there is a risk of running out of disk space on any of these devices, then it is recommended that the UIM service be stopped on these systems as well. This will result in a gap in the QoS metrics collected from the probes on these devices.
If there are a number of "collector" hubs in the UIM domain (secondary hubs that are collecting messages from robots managed by this hub or other downstream hubs to be funneled to the primary hub), it would be a good idea to set up a means of blocking the traffic from these hubs to the primary hub so that when the primary hub comes back on-line, it is not flooded with the build up of messages from all of these hubs all at once.
Once the database server is back up and on-line again, activate the UIM service in reverse order:
1. Start the UIM service on the primary hub. It is recommended that you verify that there are no issues with connecting to the database manager but testing the connection from the data_engine probe's Configure GUI.
2. Start the UIM serfvice on the HA backup hub where applicable.
3. Start the UIM service on all of the UMP, Unified Reported, and CABI robots.
If you have multiple collector hubs in the UIM domain, unblock the traffic from these hubs to the primary hub one at a time. When the "attach" queues to the primary hub are drained from one collector hub, then do the same on the next until all messages are flowing from all collector hubs to the primary hub.
On Windows system, you Stop and Start the Nimsoft Robot Watcher service from the Windows services.msc utility to Stop/Start the UIM service.
On Linux systems you Stop and Start the UIM service by executing the following commands from a root user login shell:
To Stop the UIM service:
service nimbus stop
To Start the UIM service:
service nimbus start