Description of problem: StorageNodeManager.updateConfiguration(config) returns false when it should return true and the server goes into maintaineance mode. Version-Release number of selected component (if applicable): How reproducible: always Steps to Reproduce: 1. Connect to RHQ remote API. 2. Execute the following commands in order var criteria = new StorageNodeCriteria(); var nodes = StorageNodeManager.findStorageNodesByCriteria(criteria); var node = nodes.get(0); var config = StorageNodeManager.retrieveConfiguration(node) config.setHeapSize(<some string value say 700M or 1G etc.>) StorageNodeManager.updateConfiguration(config) Actual results: It returns false in command line and the server goes into maintaineance mode Expected results: It should return true and the server should not go into maintaineance mode. Additional info:
Please retest. This has been corrected by the recent changes done to the way cluster connectivity is reported.
CLI allows to set Heap New Size value >= Heap Size value which causes that cassandra is shutting down and server stays in maintenance mode after that. To start storage node it is necessary to set Heap New Size < Heap Size - found out experimentally setting up -Xmx -Xmn values in cassandra-jvm.properties file). CLI commands to create the situtation: var crit = new StorageNodeCriteria(); var nodes = StorageNodeManager.findStorageNodesByCriteria(crit); var node = nodes.get(0); var config = StorageNodeManager.retrieveConfiguration(node); var heapSizeValue = "700M"; var heapNewSizeValue = "700M"; // 700M or higher value config.setHeapSize(heapSizeValue); config.setHeapNewSize(heapNewSizeValue); StorageNodeManager.updateConfiguration(config);
Added SLSB validation for heap settings and jmx port settings. master commit: https://git.fedorahosted.org/cgit/rhq/rhq.git/commit/modules/enterprise/server/jar/src/main/java/org/rhq?id=c50b47c2ba50ddc9640ec6060313e36aaa86b44f
verified