mozilla :: #releng

18 May 2017
00:33glandiumnthomas: is it deployed now?
00:34nthomasshould be
00:34glandiumlet's see :)
00:41glandiumnthomas: it worked
01:18nthomaswoo
01:55tjrTwo questions. Fore buildbot, I see a log line like 17:01:06 INFO - Uploading to S3: filename=c:\builds\moz2_slave\try-w64-d-00000000000000000000\logs\log_info.log mimetype=text/plain length=9352565 - how would I find that file?
01:55tjrI don't see it in (e.g.) https://archive.mozilla.org/pub/firefox/try-builds/tritter@mozilla.com-5bb23907928c62b28b282889c5e1c8b5d8c1cd8f/try-win64-debug/
01:56tjrSecondly, I was trying to produce another output file here: https://hg.mozilla.org/try/rev/4977e787e3eb30dbb31fd0bc99408a03509d0a62 it didn't seem to work (the file tom_test is not in the raw log as a file uploaded to S3 or otherwise mentioned)
02:03travis-cibuild-tools#1786 (master - e1b8841 : ffxbld): The build passed. (https://travis-ci.org/mozilla/build-tools/builds/233465288)
02:15nthomasfor 1, thats buildbot uploading things into the taskcluster index (so both buildbot and tc are available there)
02:16nthomasyou can grab the taskId (ZjR5ixaCRuGdBlGSdykpCg) and go to https://tools.taskcluster.net/task-inspector/#ZjR5ixaCRuGdBlGSdykpCg/
02:16nthomasThen Runs > Run 0 > Artifacts
02:16nthomashowever, log_info.txt all appears on treeherder anyway
02:16nthomasits the output from mozharness
02:19nthomasFor the 2nd,
02:19nthomas17:03:38 INFO - In the test function
02:19nthomas17:03:38 INFO - Copying logs to upload dir
02:19nthomasis in the treeherder log
02:20nthomaslooks like youd just get an empty file anyway ?
02:23nthomasanother option to upload a file is blobber
02:24nthomasnot sure if we use that in builds though, may just be tests
02:54tjrnthomas: yea, just an empty file; just trying to get it to produce it. Don't see it in TC though...
02:55nthomasI suspect itd exist on the worker, but the function ran after the s3 upload happened
10:05travis-cibuild-tools#1787 (master - 38526c0 : ffxbld): The build passed. (https://travis-ci.org/mozilla/build-tools/builds/233569204)
13:10travis-cibuild-buildbot-configs#2623 (master - a719349 : Jordan Lund): The build was broken. (https://travis-ci.org/mozilla-releng/build-buildbot-configs/builds/233626173)
13:11travis-cibuild-buildbotcustom#1046 (master - d7558a5 : Jordan Lund): The build passed. (https://travis-ci.org/mozilla-releng/build-buildbotcustom/builds/233626255)
13:22travis-cibuild-buildbot-configs#2624 (master - f4feed1 : Rail Aliiev): The build was fixed. (https://travis-ci.org/mozilla-releng/build-buildbot-configs/builds/233630284)
13:34travis-cibuild-buildbotcustom#1047 (production-0.8 - a5253eb : Rail Aliiev): The build passed. (https://travis-ci.org/mozilla-releng/build-buildbotcustom/builds/233635151)
13:35travis-cibuild-buildbot-configs#2625 (production - c7c9c30 : Rail Aliiev): The build passed. (https://travis-ci.org/mozilla-releng/build-buildbot-configs/builds/233635012)
13:46bhearsumthis is a heck of a task name: beetmover-checksumssigning-beetmover-signing-android-x86-nightly/opt
13:47Pikelol
13:48bhearsumi'm a little confused by so many of these tasks having "signing" in their name, eg: the one above as well as balrog-beetmover-signing-android-x86-nightly/opt and beetmover-signing-android-api-15-nightly/opt. none of them appear to actually do signing? the beetmover ones put stuff on s3, and the balrog ones submit to balrog (unless i'm missing something)
13:53mtabarabhearsum: heh, I couldn't agree more.
13:54mtabarathing is, the way we initially build the tasks, we ended up prepending the name of the task to the task dependency's name. that said, you can stop after first word to understand what that job really does
13:54mtabaras,build,created
13:54bhearsumah
13:55bhearsumso each task represents a downstream task i guess
13:55bhearsumor upstream, depending on which way you're looking at it
13:55wcostaI have a BBB task pending for quite some time now, is there something wrong or the number of pending jobs is just large? https://treeherder.mozilla.org/#/jobs?repo=try&revision=b786deee005a&selectedJob=100064033
13:55mtabaraI wanted to change that too when I did the beetmover part but it was too late, hah. there's a bug somewhere to dive into these and rename the entire schema but since we've gotten so many on our plate, we've postponed
13:56mtabarawcosta: we had the dot release running for the past few hours. could have been eating logs of BBB bandwitdh maybe
13:56mtabaras,logs,lots
13:57wcostamtabara: ah, thanks
13:57mtabarait's just a thought, not sure though. normally, it should be able to handle more than one release at a time so could be something else
13:58mtabararail is BBB expert from our side :)
13:58railshhhh
13:58mtabarahhe
13:58mtabarahehe
13:58bhearsummtabara: thanks for the background, it's very helpful in understanding!
14:00railwcosta: I'd check https://secure.pub.build.mozilla.org/buildapi/pending
14:01mtabarabhearsum: yeah, precisely. in this case "beetmover-checksumssigning-beetmover-signing-android-x86-nightly/opt" is actually a beetmover job that uploads just the checksums, depending on "checksumssigning-beetmover-signing-android-x86-nightly/opt".
14:01mtabaraThe latter is the checksums signing job (what you're looking for) that at its turn depend on beetmover job which is "beetmover-signing-android-x86-nightly/opt".
14:01mtabaraThe latter at its turn depends on the signing job to download all the artifacts and upload them to S3, which is "signing-android-x86-nightly/opt". This latter task is the actual build-signing job you're most likely looking for.
14:01railwcosta: I see Rev7 MacOSX Yosemite 10.10.5 try opt test cppunit scheduled, but has been waiting for 6+h
14:01mtabarait's a mess, I know. we'll fix this when our plates get emptier a bit
14:01bhearsummtabara: i'm having horrible memories of buildbot when you say that last part
14:02mtabarabhearsum: the one where i promise that we'll fix this or which one? :P
14:02wcostarail: thanks, I was just concerned I did something wrong
14:02railwp, np, it's just very low priority I guess
14:04bhearsummtabara: "we'll fix this when our plates get emptier"
14:04mtabarahaha
14:04mtabarafair enough :)
14:04mtabaraI thought you meant the naming part reminding you of "misc.py" :P
14:05bhearsumhehe
14:12mtabarawell, first step is done at least, I've filed bug 1365941 for it :P
14:12firebothttps://bugzil.la/1365941 ASSIGNED, mtabara@mozilla.com rename all nightly graph TC tasks to decent labels
15:11travis-cibuild-tools#1788 (master - 0c29810 : Connor Sheehan): The build passed. (https://travis-ci.org/mozilla/build-tools/builds/233672394)
15:26tjrIf I've stuck 150-200 MB of garbage onto tooltool that will never be used, should I care about cleaning it up? And if so, how?
16:03RyanVMtjr: i don't think you can
16:22travis-cibuild-buildbot-configs#2626 (master - b8c0927 : Kim Moir): The build was broken. (https://travis-ci.org/mozilla-releng/build-buildbot-configs/builds/233701788)
16:23kmoirlooking into it^^
16:38travis-cibuild-buildbot-configs#2627 (master - 4c10ed2 : Kim Moir): The build is still failing. (https://travis-ci.org/mozilla-releng/build-buildbot-configs/builds/233707193)
16:50travis-cibuild-buildbot-configs#2628 (master - 86d1d69 : Kim Moir): The build was fixed. (https://travis-ci.org/mozilla-releng/build-buildbot-configs/builds/233710992)
16:53tjrIt looks like we use gcc 4.9.4 on the linux x32 builds. I'd like to get some preliminary output from a GCC 5 warning - is there a build target that uses gcc 5? Or an easy way to bump it up? Or plans to move to 5?
16:54RyanVMtjr: I believe we have some newer gcc packages on tooltool from glandium playing around on Try
16:57travis-cibuild-buildbot-configs#2629 (production - 538cdfc : Kim Moir): The build passed. (https://travis-ci.org/mozilla-releng/build-buildbot-configs/builds/233713564)
17:19travis-cibuild-buildbot-configs#2630 (master - 0e52c2b : Kim Moir): The build passed. (https://travis-ci.org/mozilla-releng/build-buildbot-configs/builds/233720437)
17:21travis-cibuild-tools#1789 (master - 5d65657 : Kim Moir): The build passed. (https://travis-ci.org/mozilla/build-tools/builds/233720533)
17:31travis-cibuild-buildbot-configs#2631 (production - b078c2a : Kim Moir): The build passed. (https://travis-ci.org/mozilla-releng/build-buildbot-configs/builds/233723221)
17:36akidhouse: maybe i need access granted to my vpn acct?
18:01dhouseaki: I was just talking with pdawda about the flows and asked re: your ssh. Pranav said everything incoming is allowed right now in the firewall. so it may be a vpn thing. what is your ip? (or pm me)
18:02bhearsumkmoir: do you know how the devedition desktop tests are coming?
18:03kmoirbhearsum: I talked to sebastian about it this morning and discussed the path forward to fix his issues but I don't see updated patches.
18:03kmoirlet me me go look, might be easy to fix
18:04bhearsumnp
18:04bhearsumnot trying to push, was just wondering
18:05kmoirno worries, it has been taking a while
18:06bhearsumlike much of this project!
18:06JoeSRyanVM: Thanks for the uplift. Do we need to wait for a merge to production before building
18:06RyanVMJoeS: there's no production branch on mozilla-bea
18:06RyanVMbeta*
18:07JoeSRyanVM: OK thanks again
18:07RyanVMassuming you create relbranches off default, you should be fine whenever
18:07JoeSYes we have a relbranch
18:07RyanVMyou'll be creating a new one I assume?
18:26fox2mikebhearsum: https://bugzilla.mozilla.org/show_bug.cgi?id=1179339 <- that&#39;s the bug we got the thawte aus5 cert for (the one expiring in 89 days)
18:26firebotBug 1179339 FIXED, rwatson@mozilla.com please set-up aus5.m.o w/ sha2 (or higher) SSL cert
18:33bhearsumfox2mike: thanks
18:34fox2mikebhearsum: so should we go ahead and get a new one with thawte? or? file a bug and discuss?
18:34bhearsumi&#39;m still really confused by thawte being in the loop for aus5, i need to dig into it more first
18:34fox2mikebhearsum: it&#39;s our backup, it&#39;s not live
18:34bhearsumoh
18:35fox2mikewe always have a backup for aus
18:35fox2mikeand the live one is digicert
18:35bhearsumhuh, TIL thawte is still our backup
18:35fox2mikethe backup is expiring in 90 days
18:35fox2mikewell :) we always wanted different SSL providers
18:35bhearsumfor some reason i thought it was someone else now
18:35fox2mikeand that bug I linked you to has more info
18:35bhearsumin any case! as long as the issuer will match the pins from https://wiki.mozilla.org/Balrog/Client_Domains we can just renew
18:36travis-cibuild-buildbot-configs#2633 (master - cbd16cf : ffxbld): The build passed. (https://travis-ci.org/mozilla-releng/build-buildbot-configs/builds/233744244)
18:36travis-cibuild-buildbot-configs#2632 (production - e86f33f : ffxbld): The build passed. (https://travis-ci.org/mozilla-releng/build-buildbot-configs/builds/233744236)
18:36travis-cibuild-buildbotcustom#1048 (production-0.8 - 88d1f09 : ffxbld): The build passed. (https://travis-ci.org/mozilla-releng/build-buildbotcustom/builds/233744420)
18:39fox2mikebhearsum: who does aus5 ops these days?
18:40bhearsumfox2mike: relud and miles
18:40fox2mikecool
18:40fox2mikeshould I just ask them to renew it?
18:40bhearsumyeah, sounds good
18:41travis-cibuild-tools#1790 (master - f466461 : ffxbld): The build passed. (https://travis-ci.org/mozilla/build-tools/builds/233744487)
19:11bhearsumwho knows about balrogworker? do i need to login to them and pull after landing https://bugzilla.mozilla.org/show_bug.cgi?id=1366034 ?
19:11firebotBug 1366034 NEW, bhearsum@mozilla.com make balrog submission tools support old id builds
19:12akithat&#39;s likely. i can check
19:13akibhearsum: if you mean balrog scriptworker, i believe https://hg.mozilla.org/build/puppet/file/tip/modules/balrog_scriptworker/manifests/init.pp#l106 will not re-pull after the initial pull, so you&#39;ll need to ssh in
19:13bhearsumi think that&#39;s what i mean
19:13bhearsumbalrogworker-v1
19:14mtabara|focusthere&#39;s two of them
19:14akiyeah
19:14bhearsumok, thanks!
19:14bhearsumeasy
19:15mtabara|focusoh, wow. dates back to Jan :)
19:15bhearsumi was afraid of that
19:15bhearsumnot risky at all!
19:15mtabara|focusheh
19:15bhearsumi better check the diff before i pull though
19:16mtabara|focus++
19:18mtabara|focusbhearsum: should be fairly easy to check actually. balrogscript uses just https://github.com/mozilla-releng/balrogscript/blob/master/balrogscript/script.py#L10 so if stuff under lib/python/balrog hasn&#39;t drastically change since then, it should just work
19:21bhearsumoh, nice
19:21bhearsumi think you mean line 19 though?
19:22mtabara|focuserr, right - https://github.com/mozilla-releng/balrogscript/blob/master/balrogscript/script.py#L19
19:28bhearsumhere&#39;s a dumb question: how do i find the hostnames for the balrogworker machines?
19:28mtabara|focushttps://hg.mozilla.org/build/puppet/file/default/manifests/moco-nodes.pp#l1028
19:29bhearsumyeah, but what&#39;s in the wildcard section?
19:29akihttps://papertrailapp.com/groups/4560112/events
19:29mtabara|focusbalrogworker-1 and balrogworker-2
19:29bhearsumhmmm, i should get a papertrail account
19:30bhearsumah, i guessed the datacenter wrong, didn&#39;t realize they are use1
19:30mtabara|focusthey aren&#39;y both use1
19:30akibhearsum: you have access to the papertrail secret
19:30bhearsumthanks
19:31mtabara|focusbalrogworker-1 is in use1 while balrogworker-2 is in usw2
19:31bhearsumgot it
19:35bhearsumok, looks like there&#39;s nothing risky changing with my patch \o/
19:35mtabara|focus\o/
19:35bhearsumthanks for the knowledge transfer
19:35mtabara|focusnp
19:37arrjmaher|afk: catlee: sphilp: we&#39;re running 1511
19:44arrah, apparently 1607 in AWS
19:44arrso we did at least one upgrade there
20:36travis-cibuild-tools#1791 (master - 0ab5968 : Ben Hearsum): The build passed. (https://travis-ci.org/mozilla/build-tools/builds/233783395)
20:55railoremj: whenever you have time, may I ask you to take a look at 1365991
21:48travis-cibuild-tools#1792 (master - b22d07f : tbirdbld): The build passed. (https://travis-ci.org/mozilla/build-tools/builds/233808992)
23:12travis-cibuild-tools#1793 (master - b299726 : ffxbld): The build passed. (https://travis-ci.org/mozilla/build-tools/builds/233831872)
19 May 2017
No messages
   
Last message: 41 days and 5 hours ago