narendly opened a new issue #327: Change the way we trigger rebalance in purgeExpiredJobs URL: https://github.com/apache/helix/issues/327 Currently, we use ResourceConfigs to trigger a rebalance. But this is causing Helix to write a ResourceConfig back to ZK, which is not desirable when a mixed-mode rebalance is happening (user may create a non-task ResourceConfig for custom preferenceLists). AC: 1. Use an alternative way to trigger a controller rebalance. See the log: ` if (expiredJobs.size() > 0) { // Update workflow context will be in main pipeline not here. Otherwise, it will cause // concurrent write issue. It is possible that jobs got purged but there is no event to // trigger the pipeline to clean context. HelixDataAccessor accessor = manager.getHelixDataAccessor(); List<String> resourceConfigs = accessor.getChildNames(accessor.keyBuilder().resourceConfigs()); if (resourceConfigs.size() > 0) { RebalanceScheduler.invokeRebalanceForResourceConfig(manager.getHelixDataAccessor(), resourceConfigs.get(0)); } else { LOG.warn( "No resource config to trigger rebalance for clean up contexts for" + expiredJobs); }`
---------------------------------------------------------------- This is an automated message from the Apache Git Service. To respond to the message, please log on to GitHub and use the URL above to go to the specific comment. For queries about this service, please contact Infrastructure at: [email protected] With regards, Apache Git Services
