After upgrading to version 11.x of Aurea Monitor you notice that provisioning takes several minutes compared to several seconds on 10.x.
This can happen if huge Export profiles are used with many Service Groups.
The config manager on the Aurea Monitor Intermediary (AI) side is recreated after each change. To avoid this, there is a parameter that can be used to tell the engine to wait x milliseconds after each change before dropping and recreating the configuration manager. If during this time a new change is made the timer is reset as well.
This way many changes can be done in a batch without the recreation of the config manager. The new config will be live only after the given timeout expires.
This timeout can be set by adding the following JVM argument to the Additional JVM Options of the provisioned profile:
- Stop the AI you want to provision. Ensure it is really stopped (Launcher + AI).
Delete the deploy and working folder of this AI (NOT the source AI)
eg. on Linux this can be done like this:
rm -Rf /aurea_test/profiles/CX_AI_profile02/working rm -Rf /aurea_test/profiles/CX_AI_profile02/deploy
Edit the launcher profile on AMS and add this JVM argument:
- Start the AI launcher
This will wipe the local AI config, but that shouldn't be a problem because it should be stateless anyway. The new config should be provisioned again from AMS (Aurea Monitor Management Server), this time the com.actional.config.RetainRuntime JVM option should be applied.