I current use hadoop2.6.0 configured with fair scheduler. After trying Hive on Tez, it seems that fair scheduler get deadlock.
The output of jstack, jmap and resourcemanager log is here: https://drive.google.com/file/d/0B7zV3Ct1fa9dQWJvZW9hR0Q0dHc/view And deadlock occurred at 2015-07-27 16:10, so you guys can focus at this time when read resourcemanager log. >From jstack, it seemed hung at FairSchedulerUpdateThread Hope above information can help debug this issue. Below link is original posted at tez user email list. http://mail-archives.apache.org/mod_mbox/tez-user/201507.mbox/%3C2343C6D2-FD65-43B9-AE72-9E77E0A0E7CD%40apache.org%3E Thanks
