Yep sure. Still it's not clear when it happens, under which circumstances....
Also for the error, can I just make the corresponding column "bigger"? Or maybe change its type? Not the that very same config works just fine with sqlite. On Fri, Feb 19, 2016 at 7:23 PM, Pierre Tardy <[email protected]> wrote: > Thanks for the rootcausing! > > We would really appreciate if you can contribute this as a documentation > patch > > Le ven. 19 févr. 2016 à 15:14, Francesco Di Mizio < > [email protected]> a écrit : > >> Searching around it seems most people only get that error on Win. Kind of >> strange I am also getting it as I am using Docker 5.7 images which is >> Debian\Jesse. >> >> Anyway adding sql_mode='MYSQL40' to my custom .cnf file. It now looks >> like: >> >> [mysqld] >> bind-address=0.0.0.0 >> default-storage-engine=MYISAM >> sql_mode='MYSQL40' >> >> Rather then being an error, it's now a warnign like you said it should be >> Dustin. >> >> I then bumped into the very same error I got before (it's up in this >> thread) using postgres. Here's the mysql variant of the error: >> >> xceptions.RuntimeError: value for column objects.name is greater than >> max of 128 characters: >> MailNotifier_builders_try_waf_win_x64_msvc_shipping+try_waf_win_x64_msvc_development+try_waf_linux_x64_gcc_shipping+try_waf_linux_x64_clang_shipping >> >> Attached is the entire call stack. >> >> As a test, by removing my mailnotifiers to work the above around I was >> finally able to start the master ;) >> >> >> >> >> >> >> >> On Fri, Feb 19, 2016 at 5:15 AM, Dustin J. Mitchell <[email protected]> >> wrote: >> >>> Hm, in our testing 5.7 worked, although it *warned* about that. Is >>> there, by chance, a MySQL setting to control whether that's a warning >>> or not? >>> >>> Dustin >>> >>> On Thu, Feb 18, 2016 at 1:30 PM, Francesco Di Mizio >>> <[email protected]> wrote: >>> > Did some more testing with b07 and Mysql 5.7. I am failing to buildbot >>> > upgrade-master /buildbot/master. >>> > >>> > >>> > 2016-02-18 18:17:07+0000 [-] sqlalchemy.exc.OperationalError: >>> > (_mysql_exceptions.OperationalError) (1101, "BLOB, TEXT, GEOMETRY or >>> JSON >>> > column 'state_string' can't have a default value") [SQL: u"\nCREATE >>> TABLE >>> > builds (\n\tid INTEGER NOT NULL AUTO_INCREMENT, \n\tnumber INTEGER NOT >>> NULL, >>> > \n\tbuilderid INTEGER, \n\tbuildrequestid INTEGER NOT NULL, >>> \n\tbuildslaveid >>> > INTEGER, \n\tmasterid INTEGER NOT NULL, \n\tstarted_at INTEGER NOT >>> NULL, >>> > \n\tcomplete_at INTEGER, \n\tstate_string TEXT NOT NULL DEFAULT '', >>> > \n\tresults INTEGER, \n\tPRIMARY KEY (id), \n\tFOREIGN KEY(builderid) >>> > REFERENCES builders (id), \n\tFOREIGN KEY(masterid) REFERENCES masters >>> > (id)\n)CHARACTER SET utf8\n\n"] >>> > >>> > 5.7 does not seem to like >>> > tate_string TEXT NOT NULL DEFAULT '' >>> > >>> > On Tue, Feb 9, 2016 at 9:14 AM, Francesco Di Mizio >>> > <[email protected]> wrote: >>> >> >>> >> Sure thing. >>> >> >>> >> I am using a config that works just well with sqlite. >>> >> >>> >> I then spun up a container whose image is provided directly by >>> postgres. >>> >> The setup look alright as I can connect from other containers and >>> open a >>> >> postgree console and run my commands against the DB. >>> >> >>> >> When I set bbot up to use that DB (bbot master in just an other >>> container >>> >> linking the postgres one) I get: >>> >> >>> >> sqlalchemy.exc.DataError: (DataError) value too long for type >>> character >>> >> varying(128) >>> >> 'INSERT INTO objects (name, class_name) VALUES (%(name)s, >>> %(class_name)s) >>> >> RETURNING objects.id' {'class_name': >>> 'buildbot.reporters.mail.MailNotifier', >>> >> 'name': >>> >> >>> 'MailNotifier_builders_try_waf_win_x64_msvc_shipping+try_waf_win_x64_msvc_development+try_waf_linux_x64_gcc_shipping+try_waf_linux_x64_clang_shipping+try_server_x64_release'} >>> >> >>> >> those above are what my builders are named. >>> >> >>> >> On Sun, Feb 7, 2016 at 6:15 PM, Dustin J. Mitchell <[email protected] >>> > >>> >> wrote: >>> >>> >>> >>> Can you give some more context for that error? >>> >>> >>> >>> On Fri, Feb 5, 2016 at 12:42 PM, Francesco Di Mizio >>> >>> <[email protected]> wrote: >>> >>>> >>> >>>> After spending too much with MYSQL official docker images I gave e >>> a try >>> >>>> to postgres. This was much smoother to setup, I could at least >>> connect from >>> >>>> the cmd line in 5 mins. >>> >>>> >>> >>>> Sadlysqlalchemy.exc.DataError: (DataError) value too long for type >>> >>>> character varying(128) >>> >>>> >>> >>>> All in Nine. >>> >>>> >>> >>>> >>> >>>> >>> >>>> >>> >>>> On Fri, Feb 5, 2016 at 12:18 AM, Pierre Tardy <[email protected]> >>> wrote: >>> >>>>> >>> >>>>> There are some legacy files that are actually unused (like the >>> template >>> >>>>> dir), and we did not have time yet to cleanup (help wanted!) >>> >>>>> >>> >>>>> Unless you are using sqlite, I swear there is no data in the >>> basedir. >>> >>>>> Everything is in the db.. >>> >>>>> >>> >>>>> Pierre >>> >>>>> >>> >>>>> Le jeu. 4 févr. 2016 à 23:01, Francesco Di Mizio >>> >>>>> <[email protected]> a écrit : >>> >>>>>> >>> >>>>>> you mean Nine does not generate anything in the Master folder when >>> >>>>>> running? I swear to have seen something, I'll double check >>> tomorrow! >>> >>>>>> >>> >>>>>> On Thu, Feb 4, 2016 at 9:15 PM, Pierre Tardy <[email protected]> >>> wrote: >>> >>>>>>> >>> >>>>>>> What I do with my eight instance is just to make a symlink of >>> the >>> >>>>>>> master.cfg to your pickle dir. >>> >>>>>>> >>> >>>>>>> What I would really do is switch to nine :-p >>> >>>>>>> bye bye pickle! >>> >>>>>>> >>> >>>>>>> Pierre >>> >>>>>>> >>> >>>>>>> Le jeu. 4 févr. 2016 à 21:03, Francesco Di Mizio >>> >>>>>>> <[email protected]> a écrit : >>> >>>>>>>> >>> >>>>>>>> Hello Pierre, >>> >>>>>>>> >>> >>>>>>>> what I meant was more about having a folder with the config >>> (i.e. >>> >>>>>>>> master.cfg) and one with the stuff generated at runtime by the >>> master >>> >>>>>>>> itself. >>> >>>>>>>> >>> >>>>>>>> Chhers, >>> >>>>>>>> Francesco >>> >>>>>>>> >>> >>>>>>>> On Thu, Feb 4, 2016 at 7:44 PM, Pierre Tardy <[email protected]> >>> >>>>>>>> wrote: >>> >>>>>>>>> >>> >>>>>>>>> Hi Francesco. >>> >>>>>>>>> You can specify the master's basedir in the buildbot.tac file. >>> >>>>>>>>> >>> >>>>>>>>> HTH >>> >>>>>>>>> Pierre >>> >>>>>>>>> >>> >>>>>>>>> Le jeu. 4 févr. 2016 à 14:59, Francesco Di Mizio >>> >>>>>>>>> <[email protected]> a écrit : >>> >>>>>>>>>> >>> >>>>>>>>>> Is there any way to do this? >>> >>>>>>>>>> >>> >>>>>>>>>> In a context where the master is dockerized, being able to >>> tell >>> >>>>>>>>>> buildbot to use a specific folder (either mounted or a docker >>> volume) would >>> >>>>>>>>>> allow to persist logs/pickles when the container hosting the >>> master gets >>> >>>>>>>>>> killed and recreated. >>> >>>>>>>>>> >>> >>>>>>>>>> The db side is not a problem, plenty of solutions there. >>> >>>>>>>>>> >>> >>>>>>>>>> Any thoughts? >>> >>>>>>>>>> _______________________________________________ >>> >>>>>>>>>> users mailing list >>> >>>>>>>>>> [email protected] >>> >>>>>>>>>> https://lists.buildbot.net/mailman/listinfo/users >>> >>>>>>>> >>> >>>>>>>> >>> >>>>>> >>> >>>> >>> >>>> >>> >>>> _______________________________________________ >>> >>>> users mailing list >>> >>>> [email protected] >>> >>>> https://lists.buildbot.net/mailman/listinfo/users >>> >>> >>> >>> >>> >> >>> > >>> >> >>
_______________________________________________ users mailing list [email protected] https://lists.buildbot.net/mailman/listinfo/users
