Backbone shows the wrong synchronization status in a Data Center environment
Description
In a Data Center environment, there are some situations when Backbone does not show the correct status of a synchronization.
Steps to reproduce
Start a Jira DC with three nodes
Increase the log level to TRACE
Create a synchronization with two different syncusers A and B
Start the synchronization
Delete the syncuser B
Shutdown the active Backbone node (you can recognize this in the trace logs)
Wait some time until another Backbone node becomes the active Backbone node and tries to start the synchronizations
The startup of the synchronization fails (expectedly due to the missing sync user)
Expected behaviour The synchronization should show as STOPPED on all nodes.
Current behaviour The active Backbone node shows the correct status STOPPED, all other nodes show the state from before, RUNNING although the synchronization couldn't be started.
Workaround
Fix the errors why the synchronization with the wrong state can't be started.
On the active node: Start the synchronization
Now all nodes should show the correct status again.
In a Data Center environment, there are some situations when Backbone does not show the correct status of a synchronization.
Steps to reproduce
Start a Jira DC with three nodes
Increase the log level to TRACE
Create a synchronization with two different syncusers A and B
Start the synchronization
Delete the syncuser B
Shutdown the active Backbone node (you can recognize this in the trace logs)
Wait some time until another Backbone node becomes the active Backbone node and tries to start the synchronizations
The startup of the synchronization fails (expectedly due to the missing sync user)
Expected behaviour
The synchronization should show as STOPPED on all nodes.
Current behaviour
The active Backbone node shows the correct status STOPPED, all other nodes show the state from before, RUNNING although the synchronization couldn't be started.
Workaround
Fix the errors why the synchronization with the wrong state can't be started.
On the active node: Start the synchronization
Now all nodes should show the correct status again.