Branch: refs/heads/master Home: https://github.com/rhq-project/rhq Commit: 2fa7c545cb0225916ae27965a143de2a6f85748f https://github.com/rhq-project/rhq/commit/2fa7c545cb0225916ae27965a143de2a6f... Author: Jay Shaughnessy jshaughn@redhat.com Date: 2014-06-13 (Fri, 13 Jun 2014)
Changed paths: M modules/core/client-api/src/main/java/org/rhq/core/clientapi/agent/discovery/DiscoveryAgentService.java M modules/enterprise/server/jar/src/main/java/org/rhq/enterprise/server/measurement/AvailabilityManagerBean.java M modules/enterprise/server/jar/src/main/java/org/rhq/enterprise/server/measurement/AvailabilityManagerLocal.java M modules/enterprise/server/jar/src/main/java/org/rhq/enterprise/server/resource/ResourceManagerBean.java
Log Message: ----------- [1094540] Agents get permanently backfilled if backend database goes offline for a moment Clean up some of the live availability doco/workflow: - Fix the jdoc for DiscoveryAgentService.getCurrentAvailability() to reflect reality - Fix ResourceManagerBean.getLiveResourceAvailability() to mark its avail report as a "ServerSideReport". This prevents it from interfering with server-agent backfill coordination. Also, correct some confusing inline doco and doe some cleanup. - Fix AvailabilityManagerBean.mergeAvailabilityReport() such that logic checking for the backfill flag does not execute *after* the backfill flag has been reset. And clean up the overall logic a bit more. - Improve jdoc for AvailabilityManagerLocal.updateLastAvailabilityReportInNewTransaction to clearly indicate its side-effect of clearing the backfill flag.
rhq-commits@lists.fedorahosted.org