[ https://jira.jboss.org/jira/browse/JBPM-2360?page=com.atlassian.jira.plug... ] Tom Baeyens updated JBPM-2360: ------------------------------ Fix Version/s: jBPM 4.1 (was: jBPM 4.0) -- This message is automatically generated by JIRA. - If you think it was sent incorrectly contact one of the administrators: https://jira.jboss.org/jira/secure/Administrators.jspa - For more information on JIRA, see: http://www.atlassian.com/software/jira

[ https://jira.jboss.org/jira/browse/JBPM-2360?page=com.atlassian.jira.plug... ] Joram Barrez resolved JBPM-2360. -------------------------------- Resolution: Done There are two options to delete a deployment: * repositoryService.deleteDeployment(): this method will do a check if there are still process instances running. If there are, then the deployment is not deleted. If not, then only history is kept. * repositoryService.deleteDeploymentCascade(): this method will not do a check and delete everything associated with the deployment (including history). The first operation is a subset of the second one, so the following is valid for both operations: * After deployment deletion * Starting a new process instance will throw a JbpmException. Als on other nodes, since before starting, a DB check is done * Signalling a process instance that was active at the time of the deletion, will throw a JbpmException on all nodes * Finding a process instance associated with the deployment on both nodes will return a null value on all nodes -> Added a testcase to verify this behaviour. There was however another issue with the repositoryCache. If a deployment deletion occurs on one node, then all the use case above will work correctly, but the caches on the other nodes are not aware of the deletion. Only on the node where the deletion happens, the cache is updated (as part of the command). This could potentially lead to a memory leak since the repositoryCache on the other nodes only grows. -> Added a check and cache cleanup to the repositorySession when a process definition isn't found. This removes any entry in the cache that is not valid (ie deployed) anymore. -- This message is automatically generated by JIRA. - If you think it was sent incorrectly contact one of the administrators: https://jira.jboss.org/jira/secure/Administrators.jspa - For more information on JIRA, see: http://www.atlassian.com/software/jira