This should be fixed now. Some conclusions can be seen in at: https://issues.apache.org/jira/browse/HIVE-18263?focusedCommentId=16292354&page=com.atlassian.jira.plugin.system.issuetabpanels:comment-tabpanel#comment-16292354
On 11 December 2017 at 18:48, Adam Szita <sz...@cloudera.com> wrote: > Hi all, > > There are two things related to this: > 1, We have recently applied a fix on the ptest component so we did a > restart this morning. We also re-triggered all pending jobs at that time. > More details: https://issues.apache.org/jira/browse/HIVE-18212 > 2, It seems like that for some time (even before Yetus integration was > added) ptest was suffering from very slow executions from time to time. > Today we looked in this as well in depth. I have found that some executor > slaves are getting shut down in the middle of the testing. I already have a > fix ready and will shortly upload it at: > https://issues.apache.org/jira/browse/HIVE-18263 > > Thanks, > Adam > > > On 11 December 2017 at 10:06, Anishek Agarwal <anis...@gmail.com> wrote: > >> Hello >> >> The apache builds seem to be taking very long and most of them just being >> either manually shut or getting killed >> Builds after https://builds.apache.org/job/PreCommit-HIVE-Build/8160/ >> seem >> to be showing the problem. Problems are surfacing after TestExecDriver is >> being run. Is there someone here with better infrastructure understanding >> to help figure out what is happening. >> >> Thanks >> Anishek >> > >