-
Title
Additional Information on High availability on Job Servers, Load Balancing and Failover -
Description
Use case scenario involving Two Job Servers for the High Availability issues on Job Servers.
One Identity Manager support processes to be run on second server if the first server is unresponsive. To be able to have this option both servers should have the same job function and the process step should be configured to run on that function. So both servers can run the job if the other one is down. The scheduled jobs or events that triggered in database level (which is created by QBMDBQueueProcess), the firegenevents only runs on the main server.
Is there a way to change this or any way to achieve the HA function? when the main job server is down, the events(created by QBMDBQueueProcess - for example custom target system synchronizations defined on dialogschedule table) tries to run on main server, not the second one.
-
Cause
This is a mix up in the expectation of load balancing vs failover of the job server. Load balancing is support by the One Identity Job Server however the failover is not supported yet. This is related to Product Enhancement Request CR ID 26097.
-
Resolution
Additional Info on a Job Server
At the time of the generation of a job, One Identity Manager decides which Jobserver is responsible for the particular job (that's the QueueName for the job in the JobQueue). The QueueName cannot be changed later which means, two Jobservers are required which are configured for the same QueueName.Note that as per the definition, there can be only one running Jobserver for the same queue.If there are two running Jobservers for the same queue, one of them will get the jobs and the other one will fave this error:---[810143] Database error 50000: Session ID for queue \xyz does not match./---The only possibility is to stop the secondary Jobserver as long as the primary is running. As soon as the primary job server stops, the secondary can be started. However, this is not an automatic process as this requires an external failover service which monitors the Jobservers and once one is stopped it starts another one automatically. -
Change Request
26097