All times shown according to UTC.

Time Nick Message
01:04 KeiKun_ joined #buildbot
01:28 ckang buildbot web page question. We're about to change BB port to 80 from 8080. Is there a way to redirect 8080 to 80?
01:31 anish ckang:
01:32 that one redirects 80 to 8080, just reverse it
01:32 (hacky, use at your own risk)
01:33 ckang thanks, didn't think of using iptables :)
01:33 anish also, why would you need to redirect 8080 to 80 ? 80 is the default port anyways
01:33 ckang was thinking of redirecting within root.html or etc...
01:33 a long time ago, our BB was set to 8080
01:33 and its in everyones shortcuts now
01:34 anish I believe 8080 is the default anyways
01:35 ckang can BB listen to multiple ports?
01:35 errr not listen, the other
01:35 anish serve on multiple ports ?
01:35 ckang yeah
01:35 anish to my knowledge, you can only pass one port to webstatus
01:35 ckang cool deal
01:35 anish however my knowledge is weak and lacking
01:35 so please double check
01:37 also, at 80 you might see binding on priveleged port issues
01:38 ckang binding port issues as in "port in use" ?
01:39 anish no, this[…]ports-below-1024/
01:40 depending on how tomcat is starting up, you might run into that
01:40 ckang ah cool, got it. Thanks Anish
01:45 functorfun joined #buildbot
01:54 anish ckang: I forgot to mention, replace 'vnc18' by your actual hostname in the link I sent you
01:56 ckang anish: yup, thanks again
02:03 benallard__ joined #buildbot
02:32 bstack joined #buildbot
03:00 benallard_ joined #buildbot
03:27 jaredgrubb joined #buildbot
03:33 koobs joined #buildbot
03:50 booo joined #buildbot
04:27 KeiKun joined #buildbot
04:30 genpaku joined #buildbot
04:30 Kei__ joined #buildbot
06:08 bdbaddog joined #buildbot
06:10 bdbaddog I have some old buildbot code which uses IRenderable, but I see the docs no show a properties.renderer decorator. Are they equivalent?
06:44 fpotter joined #buildbot
07:19 mattiasr joined #buildbot
07:35 KeiKun joined #buildbot
07:53 benallard left #buildbot
08:01 kraiskil joined #buildbot
08:06 alicef left #buildbot
08:12 bahamas joined #buildbot
08:16 fpotter joined #buildbot
08:40 dirichlet joined #buildbot
08:49 mjw joined #buildbot
08:50 delanne morning
08:56 bb-github joined #buildbot
08:56 bb-github [13buildbot] 15delanne comment on issue #1363: I close this PR, as #1364 was opened. 02
08:56 bb-github left #buildbot
08:57 KeiKun hmmm
08:57 question
08:57 mastersrc what is the location?
08:59 rather confusing
09:00 so mastersrc = ./ <= of buildmaster?
09:10 bb-github joined #buildbot
09:10 bb-github [13buildbot] 15sa2ajj comment on issue #1364: @delanne what do you about this approach? 02
09:10 bb-github left #buildbot
09:17 bb-github joined #buildbot
09:17 bb-github [13buildbot] 15delanne comment on issue #1364: sounds good to me. And better than #1363. I use it right now ;) 02
09:17 bb-github left #buildbot
09:39 functorfun joined #buildbot
09:50 64MAA3PLE joined #buildbot
09:50 64MAA3PLE [13buildbot] 15sa2ajj closed pull request #1364: Use absolute import in the docker module (06master...06docker_absolute) 02
09:50 64MAA3PLE left #buildbot
09:50 7F1ABXHK4 joined #buildbot
09:50 7F1ABXHK4 [13buildbot] 15sa2ajj pushed 2 new commits to 06master: 02
09:50 13buildbot/06master 14c5e70f2 15Benoît Allard: Use absolute import in the docker module
09:50 13buildbot/06master 14d64079b 15Mikhail Sobolev: Merge pull request #1364 from benallard/docker_absolute...
09:50 7F1ABXHK4 left #buildbot
10:10 KeiKun O_O
10:10 what just happen
10:42 sa2ajj KeiKun: mastersrc != . of the buildmaster
10:42 or, more precisely, it might be different
11:00 KeiKun might?
11:05 ckang joined #buildbot
11:09 sa2ajj yes, might. it depends on the cwd when 'buildbot start /path/to/master' is executed
11:09 i've been looking at it about 2 months, but i soon realised it's a bit too complicated to "fix"
11:10 i wanted to make sure that mastersrc == . of the build master all the time
11:15 lissyx joined #buildbot
11:16 bb-trac [trac] #3023/defect (new) updated by Ben (Looks like the 'build started' event comes through, and that the waterfall makes an ...)
11:39 [trac] #3048/defect (v:master) created by Ben (ConnectionError when the Docker host is not available)
11:41 [trac] #3049/undecided (v:0.8.9) created by gauravmaraj (The Weight Loss Med Elite Green Coffee And Fat Loss Supplement)
11:41 lissyx joined #buildbot
11:43 dirichlet joined #buildbot
11:48 bb-trac [trac] #3049/undecided deleted (The Weight Loss Med Elite Green Coffee And Fat Loss Supplement)
11:51 delanne When using a AbstractLatentBuildSlave with build_wait_timeout >0, I have this tracback when the slave is disconnected when the build_wait_timer fired:[…]RtKGG9gtMg2MA9iM/
11:51 is the fix good ?
11:53 sa2ajj looks like it is
11:56 pepsiman joined #buildbot
12:00 bb-github joined #buildbot
12:00 bb-github [13buildbot] 15delanne opened pull request #1365: various minor fix for AbstractLatentBuildSlave and DockerLatentBuildSlave (06master...06DockerLatentFix) 02
12:00 bb-github left #buildbot
12:04 bb-github joined #buildbot
12:04 bb-github [13buildbot] 15sa2ajj comment on issue #1365: :+1: 02
12:04 bb-github left #buildbot
12:13 bahamas are builders stored anywhere in the database? I can't see any table named builders.
12:21 if I make changes to a builder, should I deleted the pickled files of that builder?
12:22 I have a weird situation where I received two emails when a builder fails. the URLs in the emails refer to different builds, but the changeset is the same
12:43 bb-github joined #buildbot
12:43 bb-github [13buildbot] 15sa2ajj pushed 6 new commits to 06master: 02
12:43 13buildbot/06master 14d738634 15Xavier Delannoy: DockerLatentBuildSlave:...
12:43 13buildbot/06master 14e08e431 15Xavier Delannoy: DockerLatentBuildSlave:...
12:43 13buildbot/06master 14182c6c5 15Xavier Delannoy: DockerLatentBuildSlave...
12:43 bb-github left #buildbot
12:50 mjw joined #buildbot
13:19 bb-trac [trac] #3050/undecided (v:0.8.9) created by axrdhwan88 (Garcinia Cambogia Supplements Online Bewertung)
13:23 lissyx joined #buildbot
13:29 bb-trac [trac] #3050/undecided deleted (Garcinia Cambogia Supplements Online Bewertung)
13:59 KeiKun_ joined #buildbot
14:00 functorfun joined #buildbot
14:04 verm__ i'm going to re-enable botscout and httpbl
14:05 KeiKun joined #buildbot
14:05 sa2ajj do you keep the config file changes under a vcs?
14:05 verm__ yes
14:05 not public
14:05 things need to be split, i don't want any of our spam busting information to be public
14:06 enabled.
14:07 sa2ajj ok :)
14:07 and thanks
14:07 verm__ i didn't know we were getting spam tickets :(
14:08 sa2ajj oh
14:08 not many. today's exceptional
14:08 verm__ for some reason i'm not getting the emails for those tickets it's strange
14:09 oh, i see hah
14:09 i'm not getting all the trac ticket mails to my inbot
14:09 inbox
14:09 doh
14:09 i filtered them to an mbox and forgot
14:09 sa2ajj :)
14:10 verm__ sa2ajj: anyway if any spam gets through please let me know
14:10 the settings need to be changed constantly to keep ahead of them.. technology gets better and better at guessing
14:10 we use 5 different systems to beat them now
14:14 sa2ajj sure
14:14 i'll create a ticket and reference it here
14:14 (or tickets :))
14:16 verm__ ok, well we're using botscout, httpbl (project honeypot), recaptcha, regex (login patterns), bayesian filtering (spam/ham), spamhaus (ip blacklist), ip throttling, and a custom plugin that lets us set users as SPAM/NOTSPAM (any post by a spam user results in it being learned as spam, a notspam users posts get learned as ham)
14:17 we can in the future enable: linksleeve (url name checker), stopforumspam (forum spam), akismet (comment/text), fspamlist (forum spam checker)
14:18 sa2ajj wow
14:28 bb-trac [trac] #3051/defect (v:master) created by Ben (master refuses to schedule new builds, doesn't really stop nor start)
14:35 [trac] #3051/defect (new) updated by Ben (And I have ~20 buildbot python process stuck.)
14:41 [trac] #3051/defect (new) updated by delanne ( - DockerSlave use threads.    - The issue seems to appears when stopping the master ...)
14:43 [trac] #3051/defect (new) updated by delanne (Or, an older container is still running and trying to connect to your master (I had ...)
14:44 KeiKun_ joined #buildbot
14:46 bb-trac [trac] #3051/defect (new) updated by Ben (Restarting my master doesn't help re-scheduling builds ... Nothing shows in my logs ...)
14:47 jklontz joined #buildbot
14:50 bb-trac [trac] #3051/defect (new) updated by Ben (@delanne: docker ps shows the container as running, and in my logs, I get the ...)
15:08 tromey joined #buildbot
15:12 bb-trac [trac] #3051/defect (new) updated by delanne (1. buildbot stop  2. ps aufx | grep python # check if buildbot stopped correctly as ...)
16:00 MiniCow joined #buildbot
16:02 jklontz joined #buildbot
16:13 bahamas can anyone give me any points on how I can debug this issue? basically, I'm receiving two emails when a build completes. sometimes the URLs of the builds are the same, sometimes they differ. sometimes the URL contains a build number that is higher than the number of the last build
16:13 s/points/pointers/
16:15 brson joined #buildbot
16:23 bahamas sendMessage is called only once for each particular mail, no? the strange part is that the 'sending mail' message appears only once in the logs
16:57 sa2ajj is it a single master system?
17:05 jklontz joined #buildbot
17:08 bahamas sa2ajj: yes
17:09 sa2ajj (i'm shooting in the dark)
17:10 do those mails refer to the same builder?
17:10 bahamas yes
17:10 they come exactly at the same time
17:10 sa2ajj the code that forms the mail notification is not a custom one?
17:11 so one sendMessage -> two mails? :/
17:11 damon__ joined #buildbot
17:12 skelly is there only one mailer set up too?
17:14 KeiKun joined #buildbot
17:14 bahamas sa2ajj: I'm just using MailNotifier.
17:15 skelly: yes. I'm assuming that if there were two, there would be two messages in the logs, no?
17:15 skelly presumably, but just trying to eliminate things
17:16 sa2ajj and it's 0.8.9?
17:17 bahamas yes
17:17 buildbot --version
17:17 Buildbot version: 0.8.9
17:17 Twisted version: 14.0.0
17:18 sa2ajj i would really expect a couple of sendMessage for a couple of messages
17:18 (iirc, each mail is sent separately)
17:19 bahamas from what I saw in status.mail I'm assuming the same
17:24 sa2ajj so just to summarise: log shows one outgoing mail, two are actually received, same builder, different build numbers, same changeset?
17:24 and it only happens for failed mails?
17:24 damon__ left #buildbot
17:40 bahamas sa2ajj: yes, regarding the first
17:40 bdbaddog bahama: did you check the raw source of the email to see if indeed both copies are sent to the same email address and that you are not getting one due to an alias or something like that?  Check your sendmail logs?
17:40 bahamas I don't know about the latter, because I only had failing packages recently
17:41 bdbaddog: I will check, but I don't have any aliases, as far as I know
17:42 thanks for your suggestions, guys. I have to go now
17:59 brson i have a buildstep with halteOnFailure=False, flunkOnFailure=false, warnOnFailure=True because it is flaky and non-essential. I expect that when this step fails the remaining build steps will proceed anyway, but that is not what happens
17:59 here is an example[…]64-opt/builds/817
17:59 any notions why?
18:00 hm, perhaps because it fails with an exception instead of a proper finished(FAILURE)
18:01 maybe not. the overall build does succeed with a warning as expected
18:02 er, maybe not. maybe the purple exception message means it failed
18:02 bb-meeting joined #buildbot
18:02 brson sorry, blathering
18:02 bb-meta joined #buildbot
18:03 brson could definitely use a pointer on how to make this exception not cause a failure though
18:03 skelly that's an exception in buildbot
18:03 jklontz joined #buildbot
18:04 skelly or, in twisted
18:04 the paths refer to linux but this is a Windows builders
18:13 sa2ajj brson: i think i mentioned this to you the other on irc (but you probably were not nearby)
18:13 brson skelly: i suspect it's the same bug as discussed[…]etup-for-windows/
18:13 for now i just want to work around it best i can
18:13 sa2ajj if a step errors (that's it, an exception is thrown during the execution), the build stop
18:13 brson skelly: ah, i did miss it i think
18:13 er sa2ajj
18:14 sa2ajj no worries :)
18:14 brson sa2ajj: makes sense, yeah. can i catch this exception in this instance and ignore it?
18:14 sa2ajj i also looked into ways to "disable" it, and it's not possible w/o patching the code (which i would not recommend)
18:14 brson sa2ajj: ah, even with e.g. subclassing DirectoryUpload/
18:15 sa2ajj it's on the other level
18:15 the level which checks the outcomes of buildsteps
18:15 lemme check again though
18:15 do your errors happen only because of the same cause?
18:16 brson sa2ajj: this error is consistently the same cause yes
18:16 sa2ajj and the "other process" is not another build?
18:17 brson sa2ajj: i don't know what the other process is :(
18:17 sa2ajj right
18:17 brson assuminng it is another build is there something i can do? limit slaves to 1 build each maybe?
18:17 like, if this is just the buildslave stepping on itself
18:18 sa2ajj you can use locks to prevent this
18:18 brson sa2ajj: you think putting a lock around just this step might help?
18:18 or maybe all the upload steps
18:19 fpotter joined #buildbot
18:19 brson sa2ajj: do you have an idea of what this temp file is?
18:19 sa2ajj one thing at a time :)
18:20 brson sure :)
18:20 sa2ajj the code that handles all those *OnFailure is here:[…]ess/
18:20 and it obviously handles exceptions and retries differently.
18:21 w/o looking at the build definition --> no
18:21 however looking at the log again, i got an idea: this temporary file/directory is created in the default path
18:22 so you might want to add a unique tmp dir for each build
18:23 and, in general, it's a good idea since it prevents "surprises"
18:24 brson hm, i'm not sure what that tmp dir even is, what process is creating it
18:25 sa2ajj regarding your suggestion that it might be same as discussed at mariadb site, it _might_ be the case. howevever, at work we run builds on windows 7 (not 8) and the programs under test do use temporary directories.
18:25 and i did not observe this problem once i set a unique tmp dir for each build
18:26 setting for each build = providing modified environment for each step
18:26 do you happen to have the builder definition somewhere?
18:26 (i checked, but the configs do not seem to be there)
18:27 brson sa2ajj: ah, yeah i could override TEMP env without understanding where the temp files are coming from
18:27 nice idea
18:29 sa2ajj since our programs under test generate a lot of intermediate files, i added a step to remove as the last step
18:29 brson sa2ajj: the builder definition is currently in a private repo but here's a gist, pointing to where the environment is configed[…]istfile1-txt-L550
18:30 it's ... a bit of a hairball. sry
18:31 this is where most of the builder is constructed[…]gistfile1-py-L825
18:33 sa2ajj it seems that most buildbot configs tend to be a bit messy :)
18:37 it looks that my suggestion will not work for you
18:38 unless you allow max_build>1 for your windows build slaves
18:38 (max_builds)
18:38 and, if you don't, then mariasdb site suggestion is what you'd need to go with.
18:39 (the parameter is read from a file, so only you would know)
18:39 brson sa2ajj: max_build is 3
18:40 if just setting it to 1 might help we could even try that
18:40 maybe with cheaper instances...
18:41 sa2ajj probably trying it will give an answer faster
18:42 brson i'll try that. thanks for the help sa2ajj
18:42 sa2ajj no worries :)
18:43 one question, if you are still there
18:43 you seem to upload directory called "doc", but i do not see "doc" anywhere in the error path :(
18:45 bstack joined #buildbot
18:46 brson sa2ajj: this DirectoryUpload build step should not involve any temporary directories as part of the executed build command. my suspicion is that this temp file is used by buildbot itself
18:46 (there is i tempfile created by SlaveDirectoryUploadCommand)
18:49 sa2ajj then we probably found a problem. thank you!
18:58 brson: i do not have a windows machine handy to check :(
19:04 Forkk joined #buildbot
19:05 sa2ajj brson: on my linux, my idea does not work
19:05 so your experiement with max_builds=1 might give an idea
19:10 sboudria_ joined #buildbot
19:34 brson sa2ajj: ok, i will try. thanks for looking into it
19:41 bstack_ joined #buildbot
19:42 bb-trac [trac] #3052/defect (v:0.8.9) created by sa2ajj ("File is being used by another process")
19:43 [trac] #3052/defect (new) updated by sa2ajj (The test coverage is collected under Linux, so I can't really say if this particular ...)
19:48 bstack joined #buildbot
19:51 mjw joined #buildbot
20:10 tromey` joined #buildbot
20:48 rtems-bot joined #buildbot
20:55 bb-trac [trac] #3048/defect (new) updated by Ben (I tried catching the {{{request.session.ConnectionError}}}, and raising a ...)
21:01 [trac] #3048/defect (new) updated by sa2ajj (`raise interfaces.LatentBuildSlaveFailedToSubstantiate`?)
21:01 [trac] #3048/defect (new) updated by sa2ajj (I should have read the first part :(    Maybe, you are raising it at the wrong ...)
21:02 bb-meeting` joined #buildbot
21:05 bb-github joined #buildbot
21:05 bb-github [13buildbot] 15sa2ajj comment on pull request #1344 14c88adb8: I should have noticed this earlier.  I believe validation of volumes (and generation of `self.volumes`) should be done in constructor, this way any problems would be reported at the config loading stage. 02
21:05 bb-github left #buildbot
21:08 bb-github joined #buildbot
21:08 bb-github [13buildbot] 15sa2ajj comment on pull request #1344 14c88adb8: There's no need to split the string twice (here and and at L101). 02
21:08 bb-github left #buildbot
21:09 bb-github joined #buildbot
21:09 bb-github [13buildbot] 15sa2ajj comment on pull request #1344 14c88adb8: Ditto for the generation of `binds`. 02
21:09 bb-github left #buildbot
21:09 sa2ajj benallard__: ayt?
21:13 bb-github joined #buildbot
21:13 bb-github [13buildbot-infra] 15sa2ajj comment on commit 143cadb5e: @djmitche, @verm, is it meta buildbot jail? 02
21:13 bb-github left #buildbot
21:15 bb-github joined #buildbot
21:15 bb-github [13buildbot] 15benallard comment on pull request #1344 14c88adb8: As I mentioned at that time, this code is taken verbatim from upstream, I did not checked it neither tested it, hence my original question about leaving it or re-implementing it when needed ? 02
21:15 bb-github left #buildbot
21:21 bb-github joined #buildbot
21:21 bb-github [13buildbot] 15sa2ajj comment on pull request #1344 14c88adb8: We might have this need now that you are working on TRAC:3048: minimising a number of error situations might help to deal w/ the problem as well (and, yes, I understand that the traceback happens later). 02
21:21 bb-github left #buildbot
21:22 tstellar joined #buildbot
21:23 bb-trac [trac] #3048/defect (new) updated by Ben (Copying the (new) traceback here to prevent ambiguity:    {{{  2014-11-18 ...)
21:25 bb-github joined #buildbot
21:25 bb-github [13buildbot] 15benallard comment on pull request #1344 14c88adb8: I'm on it. Also adding test and better doc ... (I don't believe that TRAC:3048 is linked to latent slaves ...) 02
21:25 bb-github left #buildbot
21:29 jaredgrubb joined #buildbot
21:37 bstack joined #buildbot
21:43 bb-trac [trac] #3048/defect (new) updated by sa2ajj (I '''think''' the traceback is a bit confusing: the 'while substantiating` part ...)
21:54 functorfun joined #buildbot
22:05 bstack joined #buildbot
22:08 bstack_ joined #buildbot
22:08 bb-trac [trac] #3048/defect (new) updated by Ben (While we're at it, just got that traceback (probably because the server require ...)
22:21 functorfun joined #buildbot
22:29 functorfun joined #buildbot
22:29 bb-trac [trac] #3048/defect (new) updated by Ben (As a final note for today, when the {{{stop_instance}}} goes wrong, the consequences ...)
22:35 functorfun joined #buildbot
22:51 bstack joined #buildbot
23:25 MiniCow joined #buildbot
23:41 bb-github joined #buildbot
23:41 bb-github [13buildbot] 15benallard opened pull request #1366: Docker better (06master...06docker_better) 02
23:41 bb-github left #buildbot
23:43 zhuang joined #buildbot

← Previous day | Index | Server Index | Channel Index | Today | Next day → | Atom Feed | Search | Google Search | Plain-Text | plain, newest first