spark-dev mailing list archives

Site index · List index
Message view « Date » · « Thread »
Top « Date » · « Thread »
From shane knapp <>
Subject Re: PR permission to kick Jenkins?
Date Fri, 05 May 2017 20:41:13 GMT
alright.  this kinda sucks.

after looking at that PR, i found a bunch of comments by tom and
marcelo on the morning of may 2nd.  i grepped for that PR# in the
jenkins logs and found things that made me REALLY SAD (as well as
making me ask myself how this even works to begin with):

INFO: Pull request #17,658 was updated on apache/spark at 5/2/17 7:38
AM by User:redsanket
INFO: Pull request #17,658 was updated on apache/spark at 5/2/17 8:08
AM by User:redsanket
INFO: Pull request #17,658 was updated on apache/spark at 5/2/17 9:13
AM by User:redsanket

these three comments were actually authored by tgravescs and vanzin...
including the 9:13am one which triggered the build.

so, the logs are useless, and at this point i really don't know if
there's anything that can be done short of upgrading everything...

...which badly needs to happen, but the "upgrade jenkins" project i've
been sitting on all spring is still blocked by a bug here @ UC
Berkeley.  some of the GIDs of the students that will be accessing the
new systems are low (<200), and have GID collisions w/things like
"libvirtd" and other potentially dangerous systems.  that is getting
fixed really soon, which will let me get back on track w/rolling out
jenkins 2.0+ and a much more modern, and hopefully less flaky ghprb

wish i had better news.  :\


On Fri, May 5, 2017 at 1:14 PM, shane knapp <> wrote:
> ok cool, thanks.  i'll check the jenkins logs from then and see if i
> can find anything.  no promises.  :\
> On Fri, May 5, 2017 at 1:10 PM, Marcelo Vanzin <> wrote:
>> I noticed it on this one (before I said "ok to test"):
>> My requests went through, Tom's seemed to be ignored.
>> On Fri, May 5, 2017 at 1:06 PM, shane knapp <> wrote:
>>> also, do we have any recent PRs where i can see this happening?  it
>>> will make my log-diving a bit easier.
>>> On Fri, May 5, 2017 at 12:48 PM, shane knapp <> wrote:
>>>> tom:  i checked and your username is still on the list.
>>>> anyways, it's a pretty big list...  i pulled a couple of names out of
>>>> it (including mine -- i can add myself back if needed), but i'd be
>>>> down to have someone (rxin, maybe?) audit the list and pare it down.
>>>> also, after checking out the code for the ghprb plugin, it reads in
>>>> and passes the adminlist around as a string(!?), but there are only
>>>> 322 characters total in the entire admin list...  while using a string
>>>> seems a bit strange, at this small size (322 chars) i can't see what
>>>> might be causing these things to drop.
>>>> and finally, and unsurprisingly, there's nothing in the logs that
>>>> point to anything like this happening at all (admin/whitelist failures
>>>> or hitting the github rate limit).
>>>> i'll poke around and see if i can get a better idea of what's happening.
>>>> shane
>>>> On Fri, May 5, 2017 at 12:23 PM, Reynold Xin <>
>>>>> I suspect the list is getting too big for Jenkins to function well. It
>>>>> stopped working for me a while ago.
>>>>> On Fri, May 5, 2017 at 12:06 PM, Tom Graves <>
>>>>> wrote:
>>>>>> Does  anyone know how to configure Jenkins to allow committers to
tell it
>>>>>> to test prs?  I used to have this access but lately it is either
not working
>>>>>> or only intermittently working.
>>>>>> The commands like "ok to test", "test this please", etc..
>>>>>> Thanks,
>>>>>> Tom
>>> ---------------------------------------------------------------------
>>> To unsubscribe e-mail:
>> --
>> Marcelo

To unsubscribe e-mail:

View raw message