Article Original Creation Date: 2014-10-21
Problem
This issue is replicated by the scenario below:
After the introducing of the configuration change add JMS workflow manager: allow 128 requestors instead of 16. (To solve the issue with inform handling queued for a long time.) We now notice a rise in the hazelcast map sgDeviceWorkflowStates. It is building u in 24 hours to 65K.
We will perform cleaning of this map.
See screenshot:
Environment
ServiceGateway 4.4.1
Root Cause
After reviewing the application code Engineering has identified a scenario whereby Service Gateway will orphan device entries in the SPRT_SG_POLICY_DEVICE_HOLD table. The build-up of entries in the sgDeviceWorkflowStates Hazelcast map is a direct result of this.
Solution
Engineering has released hotfix ServiceGateway_HotFix_Pivotal10168720 to address this issue.