Github user jiny2 commented on the issue: https://github.com/apache/incubator-hawq/pull/1029 @zhangh43 YARN mode, 2 segments, each segment has 75gb mem, 15core resource, and the default queue has 25% capacity, 80% maximum capacity. Then allocated and return 12 x 256mb vsegs, then wait more than 5 min. RM will try to return yarn containers but due to precision problem, there are always some yarn containers cannot be returned. because 0.9999999999... is less than 1. As to EPSILON, I adopted previous definition introduce by Wen. I think it should work appropriately. (gdb) p *containerset $6 = {Allocated = {MemoryMB = 5120, Core = 1, Ratio = 0}, Available = {MemoryMB = 5120, Core = 0.99999999999999956, Ratio = 0}, Containers = 0x7ff07a93c870} (gdb) quit A debugging session is active.
--- If your project is set up for it, you can reply to this email and have your reply appear on GitHub as well. If your project does not have this feature enabled and wishes so, or if the feature is enabled but not working, please contact infrastructure at infrastruct...@apache.org or file a JIRA ticket with INFRA. ---