Re: [BuildBot] Notifications disabled for Debian-s390x-* and Fedora-ppc64*-* builders

[Sergio Durigan Junior <sergiodj at redhat dot com>]

On Friday, December 15 2017, David Edelsohn wrote:

> On Fri, Dec 15, 2017 at 9:20 AM, Sergio Durigan Junior
> <sergiodj@redhat.com> wrote:
>> On Friday, December 15 2017, David Edelsohn wrote:
>>
>>> On Fri, Dec 15, 2017 at 8:45 AM, Sergio Durigan Junior
>>> <sergiodj@redhat.com> wrote:
>>>> Hi there,
>>>>
>>>> This is a small announcemente to let you know that I have disabled
>>>> e-mail notifications for a few builders.
>>>>
>>>> All of the Fedora-ppc*-* builders:
>>>>
>>>>   Fedora-ppc64be-m64
>>>>   Fedora-ppc64be-native-gdbserver-m64
>>>>   Fedora-ppc64be-native-extended-gdbserver-m64
>>>>   Fedora-ppc64be-cc-with-index
>>>>   Fedora-ppc64le-m64
>>>>   Fedora-ppc64le-native-gdbserver-m64
>>>>   Fedora-ppc64le-native-extended-gdbserver-m64
>>>>   Fedora-ppc64le-cc-with-index
>>>>
>>>> All of the Debian-s390x-* builders:
>>>>
>>>>   Debian-s390x-m64
>>>>   Debian-s390x-native-gdbserver-m64
>>>>   Debian-s390x-native-extended-gdbserver-m64
>>>>
>>>> I have also removed them from the list of Try Builders, so if you use
>>>> the try build, please update your config file.  An updated list can be
>>>> found here:
>>>>
>>>>   <https://sourceware.org/gdb/wiki/BuildBot#Available_Builders>
>>>>
>>>> I took this decision after noticing that these builders have been
>>>> lagging behind for quite some time.  Some of them have more than 900
>>>> queued builds!
>>>>
>>>> I've been talking with Edjunior Machado about the PPC64* machines.
>>>> Unfortunately it appears that Minicloud, the project that provided the
>>>> machines, is going through some difficulties right now.  We're now
>>>> investigating the possibility to move these resources to the GCC Compile
>>>> Farm.
>>>>
>>>> As for the s390x machine, I haven't had the chance to investigate
>>>> further what is happening.
>>>>
>>>> I would like to take this opportunity to invite more people/companies to
>>>> join our BuildBot effort!  There's always opportunity for improvement,
>>>> and we're always looking for more machines and architectures to add to
>>>> our infrastructure.  If you have interest, please get in touch (in
>>>> private) with me and we'll get the ball rolling.
>>>
>>> This decision is beyond stupid and completely negates the purpose of
>>> the buildbots.  GDB testsuite now requieres 9 hours to complete on
>>> s390x.  Something is fundamentally broken.  Instead of the GDB
>>> community noticing the issue, investigating, and fixing, it chooses to
>>> ignore the problem.
>>
>> Hi David,
>>
>> Perhaps I should have explained better.  The only thing that I did was
>> to disable the email notifications, that's all.  The builders are still
>> going to keep building the commits, and the results are still going to
>> be displayed in the website.  This is hopefully a temporary action to
>> avoid the report of false positives (like we're seeing with PPC64BE, for
>> example), and to give us more time to investigate what is going on with
>> the machines/GDB and take necessary actions.
>>
>>> And yet you ask for more companies to participate and contribute build
>>> bots. Why should any company participate when the response is pathetic
>>> and issues that the buildbots raise are ignored?  Companies should not
>>> participate until the GDB community shows some interest in utilizing
>>> and showing appreciation for the valuable resources contributed to the
>>> program.  Otherwise this is just a vanity project for Sergio  with no
>>> real benefit to the stability of GDB on the architectures, nor to the
>>> companies.
>>
>> I find this part very offensive.  Even though I run the BuildBot with my
>> own resources, I cannot keep track of all the failures that happen all
>> the time, and I also cannot monitor the machines attached to the
>> BuildBot master.  What I try to do is to look, from time to time, at the
>> status of each buildslave and get in touch with the owner of the machine
>> if there's something strange.  This is not the first time I have had to
>> disable e-mail notifications for specific builders, for example.  So
>> far, we managed to solve all the problems that appeared with the
>> buildslaves.  I hope we can keep working together and doing that.
>>
>> Our BuildBot sends the testsuite reports to the gdb-testers mailing
>> list.  It is somewhat hard to keep track of the messages there, but it
>> would be good if we had more eyes looking at the list.  Aside from that,
>> I consider that the Try Builds and the breakage emails are nice features
>> that have helped us.
>>
>> So yeah, it would be good to see more people interested in the project.
>> I personally don't consider this a vanity project, but by all means, if
>> the community doesn't see value in this then please tell me, because I
>> certainly could find other uses for the resources I dedicate to this.
>
> Sergio,
>
> I believe that the GDB buildbots should be useful to the GDB community
> and I appreciate your efforts to maintain it.
>
> But the irony in your message should not be ignored.  You implicitly
> express that the buildbots are ignored, except possibly x86 and Try
> Builds on x86, while you proceed to ask for additional participation.
> This is fundamentally inconsistent.  Actions speak louder than words
> and this demonstrates the priorities of the GDB community.

David,

There is absolutely no irony in my message, and please do not take my
words as being the official GDB community position.

Perhaps I should have explained it better in my e-mail, but again, I am
absolutely not disabling anything other than e-mail notifications, and
again, this is a temporary action until we figure things out with these
machines.

The only reason I also asked for more participation in my e-mail is
because I do not want to send two e-mails regarding BuildBot things to
gdb-patches.  I try to keep the "noise" about the project to a minimum
here (perhaps I should be more vocal and report whenever there's a
change, but that's another question).

Anyway, I do hope that things are more clear right now.  This has
nothing to do with "not caring about these architectures" or "blindly
ignoring the failures" or "the world doesn't like my contributions".  It
is just a temporary thing.

-- 
Sergio
GPG key ID: 237A 54B1 0287 28BF 00EF  31F4 D0EB 7628 65FC 5E36
Please send encrypted e-mail if possible
http://sergiodj.net/