[infinispan-issues] [JBoss JIRA] Commented: (ISPN-669) Impose Hot Rod topology cache configuration restrictions

[ https://jira.jboss.org/browse/ISPN-669?page=com.atlassian.jira.plugin.sys... ] Galder Zamarreño commented on ISPN-669: --------------------------------------- Actually, I've looked at this again and we do indeed impose restrictions on the topology cache in HotRodServer.scala since April/May, way before 4.1.0.FINAL: defineTopologyCacheConfig(cacheManager) topologyCache = cacheManager.getCache(TopologyCacheName) And: protected def defineTopologyCacheConfig(cacheManager: EmbeddedCacheManager) { val topologyCacheConfig = new Configuration topologyCacheConfig.setCacheMode(CacheMode.REPL_SYNC) topologyCacheConfig.setSyncReplTimeout(10000) // Milliseconds topologyCacheConfig.setFetchInMemoryState(true) // State transfer required cacheManager.defineConfiguration(TopologyCacheName, topologyCacheConfig) } By default Configuration is not configured with eviction, so this is not precisely the problem. The only way I could see HotRodEncoder's topologyCache using the default cache instead of the predefined cache is if the HotRodEncoder had retrieved the topologyCache before the configuration had been defined by HotRodServer.start. This could maybe have happened cos the topology config is set after the Netty server has started, so there could be small window of opportunity. However, if that had happened, "view" key would have been null from the start cos it means that the topologyCache would be queried before "view" had been updated by HotRod server.start(). I'm moving topologyCacheConfig definition to before the Netty server is started to make sure there's no such chance and will ask for the config used in the forum cos I'm slightly intrigued by this. -- This message is automatically generated by JIRA. - If you think it was sent incorrectly contact one of the administrators: https://jira.jboss.org/secure/Administrators.jspa - For more information on JIRA, see: http://www.atlassian.com/software/jira