In case you didn't see the latest comment on the JIRA, here is the response I got from the builds team:
Unfortunately we've been running into some bugs with Jenkins which is > causing us to restart the server. We're in the early stages of migrating to > a new host which we're confident will solve most of the issues we're seeing > now. Watch the bui...@apache.org list > https://lists.apache.org/list.html?bui...@apache.org for more information > about the process going forward. We'll also try and tweet about when we > have problems and are investigating them: https://twitter.com/infrabot > On Thu, May 3, 2018 at 2:41 PM, Vihang Karajgaonkar <vih...@cloudera.com> wrote: > I had sent an email to bui...@apache.org (not sure if that is the right > list). Also created this JIRA https://issues.apache. > org/jira/browse/INFRA-16478 to see if there is some response. > > On Thu, May 3, 2018 at 2:21 PM, Vihang Karajgaonkar <vih...@cloudera.com> > wrote: > >> I saw patches getting disappeared also. I know the patch for HIVE-17824 >> was not run since last 1 week. When we looked into this yesterday I found >> that for some reason the jenkins server was restarted overnight and we lose >> the pending queue. Does anyone know why this might be happening? Does >> jenkins restart happen if the pending queue is long enough or they really >> had some scheduled maintenance? If its a maintenance thing we should be >> able to find out the mailing list which informs users of this. >> >> Even if you submit the job manually, the current ptest is only able to >> serve one job at a time. We will still end up with long pending queue and >> any jenkins restart will drain the queue again. I think we need general >> improvements to Ptest to speed things up. The testing infrastructure is not >> able to keep up with the number of patches. Any suggestions to improve this >> situation are welcome but I don't know if submitting manually is really >> going to help with the problem. >> >> >> On Thu, May 3, 2018 at 12:53 PM, Кривенко Ігор <kryvenko7i...@gmail.com> >> wrote: >> >>> My JIRAs also got disappeared. >>> But, what other developers which have no access to submit manually on >>> builds.apache.org will be doing? >>> >>> Thanks, Igor. >>> >>> 2018-05-03 22:47 GMT+03:00 Prasanth Jayachandran < >>> pjayachand...@hortonworks.com>: >>> >>> > My JIRAs also got disappeared mysteriously and had to manually submit >>> the >>> > patch. >>> > +1 on making it manual (atleast this will make the patch not disappear >>> and >>> > is more predictable/no babysitting). >>> > >>> > Thanks >>> > Prasanth >>> > >>> > > On May 3, 2018, at 12:40 PM, Sergey Shelukhin < >>> ser...@hortonworks.com> >>> > wrote: >>> > > >>> > > Ping? I just had 3-4 JIRAs again disappear from the queue >>> mysteriously, >>> > for the 2nd or 3rd time in one week. >>> > > >>> > > Given that there doesn’t appear to be major interest in improving the >>> > script in the community (other than some minor changes), should we just >>> > nuke it and have manual submission w/2 fields, a patch file and jira >>> > number? That way the tracking at least can be improved. And the patch >>> > choice to run or not run. >>> > > This would take less time than trying to appease the fickle spirits >>> of >>> > HiveQA as it is now. And also probably result in somewhat shorter >>> queues. >>> > > >>> > > From: Sergey Shelukhin <ser...@hortonworks.com<mailto: >>> > ser...@hortonworks.com>> >>> > > Date: Thursday, April 26, 2018 at 17:57 >>> > > To: "dev@hive.apache.org<mailto:dev@hive.apache.org>" < >>> > dev@hive.apache.org<mailto:dev@hive.apache.org>> >>> > > Subject: switching HiveQA to manual submission? >>> > > >>> > > Given the constant cluster.... with HiveQA, namely queue constantly >>> > losing or not picking up patches for whatever reason (not limited to >>> dedup >>> > - I uploaded some patches today where there’s only one attached to the >>> JIRA >>> > and still it didn’t get in the queue or got removed) I wonder if we >>> should >>> > add manual submission option to HiveQA. Or make it the only option >>> actually. >>> > > That way one can target a specific file and get a specific job link >>> back >>> > that won’t got anywhere (one can hope)... >>> > >>> > >>> >> >> >
