[jboss-jira] [JBoss JIRA] (WFLY-315) Avoid running out of threads when connecting to the DC from a slave to pull down missing data

[ https://issues.jboss.org/browse/WFLY-315?page=com.atlassian.jira.plugin.s... ] Emanuel Muckenhuber commented on WFLY-315: ------------------------------------------ Yeah, i mean most of this relies on unbounded thread pools, so that operations don't deadlock. We should really get rid of those at one point though. We do throttle (1) mainly because when i was running a lot of concurrent requests with an not limited thread pool the server pretty much ran out of memory pretty quickly (thanks to ModelNode.clone()). (2) are executed by single thread, because it's kinda pointless have multiple registration just use threads, when they cannot be executed concurrently. (3) we have a limited pool on the DC, however proxied request cannot be limited, because they are all blocking (awaiting the tx.complete from the DC). Also inbound request for fetching data and files are unlimited to avoid deadlocks. This all is a huge mess needs to be cleaned up at one point, however would require some fundamental changes. -- This message is automatically generated by JIRA. If you think it was sent incorrectly, please contact your JIRA administrators For more information on JIRA, see: http://www.atlassian.com/software/jira