Depending on the type of deployment, the Qlik Sense Scheduler Service (QSS) runs as master, slave, or both on a node.
Master
There is only one master Qlik Sense Scheduler Service (QSS) within a site and it is always located on the central node, where the master Qlik Sense Repository Service (QRS) runs. This means that the central node must have the Qlik Sense Scheduler Service (QSS) installed even if more QSS nodes are added. This is because the QSS on the central node coordinates all QSS activities within the site.
The master QSS handles all task administration (for example, which tasks to execute and when to execute a specific task). When the time comes to execute a task, the master QSS sends the task ID to a slave QSS within the site. Which slave QSS to distribute the task ID to is determined by a load balancing operation performed by the master QSS.
When a slave QSS completes a task, it returns the task state (successful or fail) to the master QSS. The master QSS uses the task state to perform task chaining, which means that it uses the task state to determine if other events are affected by the state of the completed task and need to be executed. This can be configured in the Qlik Management Console (QMC).
If the slave QSS fails to perform the task, the master QSS repeatedly requests the same or another slave QSS to perform the task until it has been completed or until the maximum number of attempts has been reached.
Slave
If a Qlik Sense Scheduler Service (QSS) runs on a rim node, the QSS is considered to be a slave QSS.
When receiving a task ID from the master QSS, the slave QSS reads the task from the local repository database and executes the task.
When a slave QSS completes a task, it returns the task state (successful or fail) to the master QSS.
Master and slave
Within a single node site, the master Qlik Sense Scheduler Service (QSS) also acts as a slave QSS.