Wednesday, 2013-09-11

jeblairdtroyer: ping00:00
jeblairdtroyer, clarkb, notmyname: that just passed the swift functional tests (the tempest tests are still running)00:00
jeblairclarkb, notmyname: i'm deleting the held node00:02
*** jhesketh has joined #openstack-infra00:08
*** dkliban has quit IRC00:10
*** Ryan_Lane has quit IRC00:11
*** dkliban has joined #openstack-infra00:11
clarkbjeblair: are we waiting on dtroyer?00:18
*** reed has quit IRC00:19
*** KLevenstein1 has quit IRC00:24
*** dkliban has quit IRC00:27
*** jhesketh__ has joined #openstack-infra00:31
*** jhesketh has quit IRC00:32
*** jhesketh has joined #openstack-infra00:32
*** nosnos has joined #openstack-infra00:36
*** fifieldt has quit IRC00:36
*** rev has left #openstack-infra00:37
*** dcramer_ has quit IRC00:41
jeblairclarkb: i went ahead and approved it under the circumstances; i wanted to rope him in if he were around00:42
*** weshay has joined #openstack-infra00:44
*** MIDENN_ has joined #openstack-infra00:51
*** branen_ has joined #openstack-infra00:51
*** mdenny has quit IRC00:54
*** dkliban has joined #openstack-infra00:54
*** branen has quit IRC00:55
*** vipul is now known as vipul-away00:57
clarkbjeblair: awesome, I will poke the swift changes once that merges00:57
openstackgerritJames E. Blair proposed a change to openstack-infra/config: Gate more projects on swift-functional
jeblairsdague: you should take a look at
jeblairsdague: the fact that the change it reverts slipped through the gate suggests a test gap00:59
*** vipul-away is now known as vipul00:59
pleia2so, we went from 178 bugs to 166 today, not as dramatic as other bug days but I think this just means we have more work + actually are doing a decent job keeping on top of things lately00:59
jeblairsdague: (it was caught by the swift functional tests; the way they are run, as it turns out, is more like an integration test; they actually did exercise the thing that added)01:00
anteayanice job on the curation pleia201:00
clarkbpleia2: yeah, I think the drastic backlogs of updates aren't happening as often01:00
jeblairpleia2: thanks; i dropped off talking about them, but i did keep going through the list; i also managed to triage some new ones01:00
clarkbI also got completely distracted by six and swift01:00
clarkbbut changes are moving through jenkins again so \o/01:01
clarkbjeblair: I will probably take it easy tomorrow. Larissa has the day off, weather is supposed to be really nice (for perhaps one of the last times this summer), but I intend on getting backups going on review.o.o01:03
clarkbjeblair: I was wondering if the jenkins servers need to be backed up01:03
lifelessjeblair: I don't understand why 1223073 is won't fix: I can understand hard-to-do-and-not-a-priority; but I don't understand why you'd reject patches if someone put them forward.01:04
*** jhesketh has quit IRC01:04
clarkbwe aren't quite able to spin up new jenkins from scratch independent of anything but puppet, but we have 3 and if we lose one the other two serve as templates. So redundancy and not backups01:04
clarkbnot sure how safe we want/need to play it01:04
*** jhesketh has joined #openstack-infra01:05
jeblairclarkb: i don't think i'd bother with jenkins backups at the moment01:08
jeblairlifeless: that has never been a problem, except for the other day, when a core member of a related project gave is incorrect information01:09
jeblairlifeless: and the person proposing the patch should have known better01:09
*** pcrews has quit IRC01:09
jeblairlifeless: it's a people problem, not a technical problem01:09
jeblairlifeless: also, i just _intentionally_ broke the gate for devstack.  that happens far more often than accidentally breaking a gate.01:11
jeblairlifeless: there are ample ways to test new jobs that don't involve breaking a gate for a project01:11
jeblairlifeless: a job can be added as non-voting, or added to the silent or experimental pipelines, for instance01:11
lifelessjeblair: ah, this is my lack of operational experience here01:12
lifelessjeblair: thanks01:12
morganfainbergclarkb, jeblair, is the weird pep8 issue (six) resolved?01:13
jeblairlifeless: no worries01:13
clarkbmorganfainberg: not completely, but for most projects yes01:13
lifelessjeblair: in the context of reducing human load, I still think some automation around this would be valuable01:14
morganfainbergclarkb, so, revierfy on a patchset or two for keystone should be safe?01:14
clarkbmorganfainberg: I expect only ceilometer may have trouble01:14
clarkbmorganfainberg: I think so01:14
lifelessjeblair: it sounds perhaps that could be very simple - clone variation into a silent job etc; but will leave for future discussion01:14
morganfainbergclarkb, is there a bug id i should reference with these, or just "no bug" is fine?01:15
clarkbmorganfainberg: there is a bug that dolphm had earlier01:17
uvirtbotLaunchpad bug 1219694 in ceilometer/grizzly "Jenkins broken due to incompatibility between six 1.4.0 and wsme 0.5b2" [Critical,Confirmed]01:18
morganfainbergah thanks.01:18
morganfainbergi'll reference that one01:18
morganfainbergthank you01:18
clarkbI have rechecked/reverified swift changes hit by the ceilometer middleware01:19
*** dkliban has quit IRC01:19
torgomaticclarkb: fwiw, potential fix to the ceilometer middleware is here:
torgomaticapparently some code linter hates my commit message (seriously?), but the rest of the tests passed01:26
clarkbtorgomatic: cool. and we will test it when added to devstack again so it shouldn't sneak through as broken01:26
clarkbjeblair closed that hole earlier01:26
torgomaticah, good to know. thanks, jeblair!01:27
clarkbtorgomatic: that console.html has logged the paths used by the functional swift tests01:33
torgomaticclarkb: that'll be handy in the future; I see that change in Gerrit, and it looks like it just needs a little PEP8 help01:34
torgomaticcode looks good to me01:34
clarkbyeah I will fix pep8 tomorrow01:35
*** dkliban has joined #openstack-infra01:36
*** lcestari has quit IRC01:45
*** melwitt has quit IRC01:49
*** lcestari has joined #openstack-infra01:53
*** yaguang has joined #openstack-infra01:58
*** gyee has quit IRC02:03
*** gyee has joined #openstack-infra02:06
Alex_GaynorDo any infra team members live in SF? I feel like I ask for reviews so often I should at least be bribing you with cookies / coffee cake / something.02:08
*** SergeyLukjanov has joined #openstack-infra02:13
*** dkliban has quit IRC02:16
*** xchu has joined #openstack-infra02:26
anteayaAlex_Gaynor: I think the closest you have would be jeblair02:26
anteayanot sure what he takes as bribes, but a good candidate02:27
*** anteaya_ has joined #openstack-infra02:31
*** lnxnut has joined #openstack-infra02:31
anteaya_Alex_Gaynor: and pleia202:31
* anteaya_ completely forgot she will be pleia2 house guest in SF next weekend02:32
*** anteaya has quit IRC02:32
anteaya_good thing I am headed for sleep02:32
anteaya_night again02:32
*** xchu has quit IRC02:34
*** lcestari has quit IRC02:35
*** anteaya_ has quit IRC02:37
*** marun has quit IRC02:39
*** salv-orlando has quit IRC02:41
*** salv-orlando_ has joined #openstack-infra02:41
*** xchu has joined #openstack-infra02:42
*** xchu has quit IRC02:47
*** gyee has quit IRC02:51
*** salv-orlando_ has quit IRC02:52
*** salv-orlando has joined #openstack-infra02:53
*** dims has quit IRC02:54
*** lcestari has joined #openstack-infra02:55
*** lcestari has quit IRC03:04
*** lcestari has joined #openstack-infra03:05
*** salv-orlando has quit IRC03:05
*** xchu has joined #openstack-infra03:06
*** salv-orlando has joined #openstack-infra03:06
*** dkliban has joined #openstack-infra03:10
*** SergeyLukjanov has quit IRC03:10
*** salv-orlando_ has joined #openstack-infra03:11
*** salv-orlando has quit IRC03:11
*** salv-orlando_ is now known as salv-orlando03:11
*** vipul has quit IRC03:11
*** vipul has joined #openstack-infra03:12
*** lnxnut has quit IRC03:17
*** salv-orlando_ has joined #openstack-infra03:18
*** salv-orlando has quit IRC03:18
*** salv-orlando_ is now known as salv-orlando03:18
*** Guest11685 has quit IRC03:18
*** lillie has joined #openstack-infra03:19
*** lillie is now known as Guest3178303:19
*** weshay has quit IRC03:19
*** salv-orlando has quit IRC03:38
*** salv-orlando has joined #openstack-infra03:39
*** salv-orlando has quit IRC03:46
*** salv-orlando has joined #openstack-infra03:46
*** sarob has joined #openstack-infra03:47
*** HenryG has quit IRC03:49
mordredAlex_Gaynor: jeblair lives in berkeley - and I believe does like cookies03:50
*** sarob has quit IRC03:53
*** sarob has joined #openstack-infra03:54
*** xchu has quit IRC03:54
*** sridevi has joined #openstack-infra03:55
clarkbI will accept cookies in the mail03:57
clarkbmordred: did you see I have no idea why the different venvs are different03:57
*** sarob has quit IRC03:58
mordredsee - this is an indication to me that something we're doing is too complex03:59
*** salv-orlando has quit IRC03:59
clarkboh I agree03:59
clarkbthe way we do config options is just wrong03:59
*** salv-orlando has joined #openstack-infra03:59
clarkbbut I haven't managed to have that argument with markmc yet03:59
clarkband ...04:00
clarkbit is doing something wrong. I think it isn't feeding pip a -r filename argument and instead trying to list each one individually04:00
*** salv-orlando has quit IRC04:02
*** salv-orlando_ has joined #openstack-infra04:02
mordredclarkb:         self.pip_install('-r', self.requirements)04:03
*** salv-orlando has joined #openstack-infra04:06
*** salv-orlando_ has quit IRC04:06
clarkbhmm I don't know then04:07
*** vogxn has joined #openstack-infra04:10
*** xchu has joined #openstack-infra04:12
mordredclarkb: install_venv does install -U04:12
mordredclarkb: although so does nova04:12
*** salv-orlando_ has joined #openstack-infra04:14
*** salv-orlando has quit IRC04:14
*** salv-orlando_ is now known as salv-orlando04:14
*** KLevenstein has joined #openstack-infra04:18
*** DennyZhang has joined #openstack-infra04:19
openstackgerritMonty Taylor proposed a change to openstack-infra/config: Update projects when requirements change
mordredclarkb: jeblair says "This should not be a regex ^ (Zuul's behavior changes if the name is a regex)."04:24
mordredclarkb: what behavior change?04:24
clarkblet me look, it does do somethign slightly idfferent04:24
mordredclarkb: (because, actually, with the new change to make it a job template, I actually need it to be a regex)04:24
openstackgerritMonty Taylor proposed a change to openstack-infra/config: Add job to propose tag merges
openstackgerritMonty Taylor proposed a change to openstack-infra/config: Move MySQL creation out of gerrit class
*** CaptTofu_ has joined #openstack-infra04:28
*** salv-orlando_ has joined #openstack-infra04:29
*** salv-orlando has quit IRC04:29
*** salv-orlando_ is now known as salv-orlando04:29
clarkbgah not finding it in the code, the name stuff isn't handled where the otherfiltering is done04:29
*** CaptTofu has quit IRC04:30
*** salv-orlando_ has joined #openstack-infra04:33
*** salv-orlando has quit IRC04:33
*** salv-orlando_ is now known as salv-orlando04:33
clarkbmordred: it has to do with how it maps those things across all of the projects04:36
clarkbmordred: basically if it is a regex then it has to check everything, if not a regex then it is 1:104:36
mordredclarkb: ah - ok. so, in this case, regex is probably appropriate04:36
mordred(in the new patch) - but not in the old one04:36
*** dcramer_ has joined #openstack-infra04:37
openstackgerritA change was merged to openstack-infra/config: Run the troveclient tests under PyPy
openstackgerritA change was merged to openstack-infra/config: Merge duplicate bugs in recheckwatch
clarkbmordred: found it
clarkbthe check on 928 does the 1:1 if that falls through then you ahve to do more work04:40
openstackgerritA change was merged to openstack-infra/config: Clarify bup server setup directions.
mordredclarkb: nod04:44
*** salv-orlando_ has joined #openstack-infra04:50
*** salv-orlando has quit IRC04:50
*** salv-orlando_ is now known as salv-orlando04:50
*** boris-42 has joined #openstack-infra04:51
jog0mordred clarkb I take it the recheck server doesn't restart automatically04:53
jog0if the code is changed04:53
mordredjog0: hrm. dunno04:54
jog0either my patch didn't work or it doesn't04:55
jog0auto restart04:55
mordredjog0: nope. no auto-restart04:57
mordredit autorestarts if the conf file changes04:57
mordredjog0: restarted04:59
*** Ryan_Lane has joined #openstack-infra05:01
openstackgerritA change was merged to openstack-infra/devstack-gate: add c-bak to MY_ENABLED_SERVICES
*** sridevi has quit IRC05:12
openstackgerritHaomeng,Wang proposed a change to openstack/requirements: Update Ceilometer alembic requirement to 0.6.0 version
pleia2Alex_Gaynor: no need for bribes, but if you want to join me for an Ubuntu Hour in downtown SF tomorrow evening you're welcome:
*** dzyu has joined #openstack-infra05:18
*** nicedice has quit IRC05:18
*** dmakogon_ has joined #openstack-infra05:22
*** KLevenstein1 has joined #openstack-infra05:29
*** KLevenstein has quit IRC05:29
dzyuhi all, I made one code change #45605, but jenkins failed, the root cause is Jenkins use no-recommended arguments. e.g. when create one volume instance use "--display_name", "--display_description", based on 'cinder help create', we should use '--display-name', "--display-description" to replace these no-use arguments05:38
dzyuSo want I should to do? still save these no-use argument in cinderclient? or open one bug to change Jenkins test case?05:39
clarkbdzyu: whether or not we need to maintain backward compatiblity probably depends on an answer from jgriffith05:40
clarkbdzyu: if backward compatibility is not necessary you can update grenade yourself or submit a bug to fix it or both05:40
clarkbbut grenade will need updating before 45605 can merge05:40
*** pblaho has joined #openstack-infra05:41
*** sridevi has joined #openstack-infra05:48
*** DennyZhang has quit IRC05:54
*** DennyZhang has joined #openstack-infra05:55
*** sarob has joined #openstack-infra05:56
*** jhesketh has quit IRC06:06
*** jhesketh has joined #openstack-infra06:07
*** amotoki has quit IRC06:07
*** salv-orlando has quit IRC06:11
*** salv-orlando has joined #openstack-infra06:11
*** afazekas has joined #openstack-infra06:15
*** salv-orlando has quit IRC06:17
*** salv-orlando has joined #openstack-infra06:18
*** DennyZhang has quit IRC06:18
*** amotoki has joined #openstack-infra06:23
*** sarob has quit IRC06:26
*** sarob has joined #openstack-infra06:26
*** sridevi has quit IRC06:27
*** vogxn has quit IRC06:28
jog0mordred: hmm there are still duplicate files on that page06:29
jog0mordred: (such as bug 1220386 on )06:30
uvirtbotLaunchpad bug 1220386 in swift "gate-tempest-devstack-vm-full fails with bunch of tests "failed to build and is in ERROR status" (dup-of: 1221987)" [Undecided,New]
uvirtbotLaunchpad bug 1221987 in nova "compute node heartbeat out of sync causing scheduler to fail in devstack:  VMs fail to spawn" [Critical,Confirmed]
*** sarob has quit IRC06:31
*** vogxn has joined #openstack-infra06:36
openstackgerritAndreas Jaeger proposed a change to openstack-infra/git-review: Prefer .gitconfig username
*** yolanda has joined #openstack-infra07:03
*** giulivo has quit IRC07:07
*** vogxn has quit IRC07:15
GheRiveroanyone from infra awake? There are several repos duplicated in stackforge and openstack (os-*-config, tripleo-heat-templates, tripleo-image-elements and diskimage-builder). Last changes in stackforge are over a month ago, so transitional period should be over07:16
*** salv-orlando has quit IRC07:18
*** bswrchrd has quit IRC07:20
*** vishy has quit IRC07:26
*** vishy has joined #openstack-infra07:28
*** flaper87|afk is now known as flaper8707:29
*** KLevenstein has joined #openstack-infra07:32
*** KLevenstein1 has quit IRC07:32
*** dmakogon_ has quit IRC07:34
*** boris-42 has quit IRC07:35
*** sarob has joined #openstack-infra07:37
*** sarob has quit IRC07:41
*** KLevenstein1 has joined #openstack-infra07:43
*** KLevenstein has quit IRC07:43
*** jhesketh has quit IRC07:48
*** fbo_away is now known as fbo07:50
*** Ryan_Lane has quit IRC08:05
*** DennyZhang has joined #openstack-infra08:14
*** salv-orlando has joined #openstack-infra08:15
*** dzyu_ has joined #openstack-infra08:17
*** dzyu has quit IRC08:18
*** dzyu_ is now known as dzyu08:19
*** DennyZhang has quit IRC08:19
*** salv-orlando has quit IRC08:24
*** sridevi has joined #openstack-infra08:28
*** jpich has joined #openstack-infra08:30
*** Ryan_Lane has joined #openstack-infra08:35
*** alexpilotti has joined #openstack-infra08:36
*** salv-orlando has joined #openstack-infra08:36
*** Ryan_Lane has quit IRC08:42
*** alexpilotti has quit IRC08:43
*** boris-42 has joined #openstack-infra08:49
*** xchu has quit IRC08:50
*** giulivo has joined #openstack-infra09:02
*** xchu has joined #openstack-infra09:07
*** Ryan_Lane has joined #openstack-infra09:09
*** Ryan_Lane has quit IRC09:14
*** dzyu_ has joined #openstack-infra09:16
*** sridevi has quit IRC09:17
*** dzyu has quit IRC09:19
*** dzyu_ is now known as dzyu09:19
*** sridevi has joined #openstack-infra09:29
*** vogxn has joined #openstack-infra09:39
*** ruhe has joined #openstack-infra09:46
*** KLevenstein1 has quit IRC09:53
*** KLevenstein has joined #openstack-infra09:54
*** dguitarbite has joined #openstack-infra09:56
*** vogxn has quit IRC10:02
*** ruhe has quit IRC10:03
*** sridevi has quit IRC10:05
*** HenryG has joined #openstack-infra10:11
*** vogxn has joined #openstack-infra10:13
*** xchu has quit IRC10:13
*** swifterdarrell has quit IRC10:14
*** DennyZhang has joined #openstack-infra10:14
*** cdevienne has joined #openstack-infra10:17
*** ruhe has joined #openstack-infra10:17
*** dzyu has quit IRC10:17
cdevienneHello, I need to make a release of wsme. I did my signed tag, but do not know where to push it... can someone help me ? Thanks10:18
*** DennyZhang has quit IRC10:19
*** sridevi has joined #openstack-infra10:19
*** swifterdarrell has joined #openstack-infra10:20
*** MIDENN_ has quit IRC10:20
*** weshay has joined #openstack-infra10:24
*** sridevi has quit IRC10:31
*** pcm_ has joined #openstack-infra10:32
*** sridevi has joined #openstack-infra10:33
*** pcm_ has quit IRC10:33
*** cdevienne has quit IRC10:33
*** pcm_ has joined #openstack-infra10:34
*** vogxn has quit IRC10:34
*** dizquierdo has joined #openstack-infra10:35
*** CaptTofu_ has quit IRC10:39
*** CaptTofu has joined #openstack-infra10:39
*** sridevi has quit IRC10:49
*** sridevi has joined #openstack-infra10:50
*** KLevenstein has quit IRC10:52
*** yaguang has quit IRC10:52
*** sridevi_ has joined #openstack-infra10:53
*** sridevi has quit IRC10:55
*** sridevi_ is now known as sridevi10:55
*** dims has joined #openstack-infra11:04
*** vogxn has joined #openstack-infra11:07
*** Ryan_Lane has joined #openstack-infra11:10
*** Ryan_Lane has quit IRC11:14
*** alexpilotti has joined #openstack-infra11:14
*** vogxn has left #openstack-infra11:16
*** SergeyLukjanov has joined #openstack-infra11:20
*** jcoufal has joined #openstack-infra11:38
*** dkranz has joined #openstack-infra11:46
*** markmcclain has joined #openstack-infra11:50
*** ArxCruz has joined #openstack-infra11:51
*** yongli is now known as yongli_away11:54
*** ruhe has quit IRC12:01
*** ruhe has joined #openstack-infra12:04
*** ruhe has quit IRC12:06
*** afazekas has quit IRC12:06
*** dkliban has quit IRC12:06
*** sridevi has quit IRC12:09
*** afazekas has joined #openstack-infra12:10
*** amotoki has quit IRC12:10
*** vogxn has joined #openstack-infra12:11
*** vogxn has quit IRC12:13
*** ruhe has joined #openstack-infra12:17
*** rfolco has joined #openstack-infra12:17
*** thomasm has joined #openstack-infra12:18
*** DennyZhang has joined #openstack-infra12:21
*** dprince has joined #openstack-infra12:31
*** afazekas has quit IRC12:32
*** markmcclain has quit IRC12:32
*** sandywalsh has quit IRC12:32
*** DennyZhang has quit IRC12:40
*** DennyZhang has joined #openstack-infra12:40
*** sandywalsh has joined #openstack-infra12:44
*** afazekas has joined #openstack-infra12:53
*** afazekas has quit IRC12:54
*** afazekas has joined #openstack-infra12:54
*** weshay has quit IRC12:57
*** ruhe has quit IRC13:01
*** ruhe has joined #openstack-infra13:03
*** afazekas has quit IRC13:05
*** anteaya_ has joined #openstack-infra13:07
*** anteaya_ is now known as anteaya13:13
*** mriedem has joined #openstack-infra13:13
*** KLevenstein has joined #openstack-infra13:15
*** mriedem has joined #openstack-infra13:15
*** dkliban has joined #openstack-infra13:21
*** afazekas has joined #openstack-infra13:28
*** weshay has joined #openstack-infra13:30
*** ruhe has quit IRC13:30
*** dcramer_ has quit IRC13:33
*** matty_dubs|gone is now known as matty_dubs13:33
*** ruhe has joined #openstack-infra13:33
*** KLevenstein has quit IRC13:35
*** yaguang has joined #openstack-infra13:36
*** DennyZhang has quit IRC13:37
*** nosnos has quit IRC13:38
dhellmannmordred, clarkb, Alex_Gaynor: interesting requirements issue again this morning: -- looks like hacking is listed twice, but I'm not sure why we have dev- and global- files so I'm reluctant to "fix" it13:39
*** burt has joined #openstack-infra13:41
*** adalbas has joined #openstack-infra13:41
*** pcrews has joined #openstack-infra13:52
*** zeus has joined #openstack-infra14:01
*** dkranz has quit IRC14:01
*** zeus has quit IRC14:01
openstackgerritThierry Carrez proposed a change to openstack-infra/odsreg: Fix links in comments
openstackgerritA change was merged to openstack-infra/odsreg: Fix links in comments
*** markmcclain has joined #openstack-infra14:08
*** pblaho has quit IRC14:10
*** mrodden has joined #openstack-infra14:10
*** ruhe has quit IRC14:17
*** alexpilotti_ has joined #openstack-infra14:17
*** SergeyLukjanov has quit IRC14:17
*** alexpilotti has quit IRC14:18
*** alexpilotti_ is now known as alexpilotti14:18
*** zehicle_at_dell has joined #openstack-infra14:21
*** ruhe has joined #openstack-infra14:22
*** pentameter has joined #openstack-infra14:22
dhellmannmordred, clarkb, Alex_Gaynor: I think the wiki page of instructions for editing requirements may be out of date with the actual file names in that repo (mentions "pip-requires", doesn't describe global-requirements.txt, etc.)14:25
Alex_Gaynordhellmann: indeed. Sounds wrong14:25
*** ruhe has quit IRC14:26
dhellmannI'd fix it, but with the hacking thing mentioned above now I'm confused, too.14:26
*** afazekas_ has joined #openstack-infra14:26
*** afazekas_ has quit IRC14:26
openstackgerritA change was merged to openstack-infra/config: Gate more projects on swift-functional
*** SergeyLukjanov has joined #openstack-infra14:28
*** DennyZhang has joined #openstack-infra14:31
*** ruhe has joined #openstack-infra14:32
*** ruhe has quit IRC14:34
*** jgriffith has quit IRC14:37
fungiGheRivero: i'll take a look at the duplicate repos if nobody has beaten me to it14:37
*** UtahDave has joined #openstack-infra14:39
fungicdevienne seems to have disappeared before i could wake up and link for him14:41
dhellmannfungi: we got the tag taken care of, thanks14:41
*** cppcabrera has joined #openstack-infra14:41
fungidhellmann: awesome--thanks!14:41
dhellmannfungi: are the related jobs on a schedule, or should they have run right after the tag was pushed? IIRC, they're scheduled14:42
*** sandywalsh has quit IRC14:42
fungiwhich related jobs? the release jobs to generate tarballs, upload to pypi, et cetera?14:42
fungithose are triggered in the pre-release or release pipeline as soon as the tag is pushed14:43
* fungi looks to see if they failed14:43
dhellmannfungi: ok, something's not set up right then14:43
dhellmannlooking at I don't see it as ever having run14:43
fungidhellmann: entirely possible. if this is the first time, it can be a little finicky (for example pypi refusing malformed/invalid trove classifiers)14:44
dhellmannI think this is the first time we have released wsme this way14:44
fungiif pypi-upload didn't run then tarball may have failed. looking there first anyway14:44
dhellmannI know we've done releases before14:44
dhellmannmakes sense14:44
jeblairneeds a 'venv' section in the tox.ini14:45
fungiyeah, so tox.ini needs patching14:45
fungiwhat jeblair said14:45
dhellmanncan I use oslo.config as an example of what to put there?14:45
dhellmannand which job failed? I couldn't find anything that seemed to have been triggered by the tag14:46
fungidhellmann: the wsme-tarball job failed14:46
*** yaguang has quit IRC14:46
jeblairdhellmann: failed14:46
dhellmannI looked at and found no instances of that job14:46
jeblairdhellmann: zuul runs the upload job only after the tarball succeeds14:46
jeblairdhellmann: yeah, it ran on a slave attached to the 'jenkins01' master14:47
jeblairdhellmann: (we have 3 masters now)14:47
dhellmannok. I'm trying to understand better how I would have known to figure that out. Check all of them?14:47
fungidhellmann: yeah, the tox.ini in oslo.config looks like it has the right testenv:venv (it's just a passthrough)14:47
dhellmannfungi: ok, that's what I thought, thanks14:47
jeblairdhellmann: so actually, the best way to find job results for a post job is...14:48
jeblairdhellmann: grab the git sha of the commit in question14:48
jeblairdhellmann: then go to
jeblairdhellmann: (that's the git_sha[:2]/git_sha)14:48
fungiand the git sha for the tag is identified with 'git show-ref x.y'14:48
jeblairdhellmann: all the jobs related to that commit land there14:49
dhellmannwow, ok, I'm going to have to make some notes about that :-)14:49
jeblairdhellmann: (that is not what i actually did though; i was sort of lazy and just looked at jenkins01 manually since the url was mostly already in my browser; looking at each master in turn does work)14:49
jeblairdhellmann: (at least until we delete the build records from jenkins, which is after about 24 hours)14:50
fungialso if you catch them while still running, you can query things like the appropriate jenkins url through zuul's http api14:50
jeblairdhellmann: but the useful thing about logs.o.o is this page:
jeblairdhellmann: where you can see all of the jobs tha ran for that change in the pre-release pipeline14:51
jeblairdhellmann: (if tarball had worked, pypi-upload would be there too)14:51
* dhellmann ponders creating a git plugin "git show-os-job" to open that url automatically14:52
jeblairdhellmann: dude!14:53
*** flaper87 is now known as flaper87|afk14:58
*** yamahata has joined #openstack-infra14:59
*** sandywalsh has joined #openstack-infra14:59
jeblairBobBall: will you be at linuxcon in new orleans?14:59
*** pblaho has joined #openstack-infra15:00
*** SergeyLukjanov has quit IRC15:01
openstackgerritJulien Danjou proposed a change to openstack/requirements: Update WSME to allow six to be unpined
*** SergeyLukjanov has joined #openstack-infra15:03
*** davidhadas has quit IRC15:03
*** ruhe has joined #openstack-infra15:04
mordredmorning humans15:05
*** sandywalsh has quit IRC15:05
*** w_ is now known as olaph15:06
*** flaper87|afk is now known as flaper8715:08
Alex_Gaynormorning mordred15:08
Mithrandirjeblair: going to lca, eh? :-)15:09
*** mriedem1 has joined #openstack-infra15:11
*** mrodden has quit IRC15:12
*** mriedem has quit IRC15:12
jeblairMithrandir: yes!15:12
MithrandirI'm trying to make up my mind if I'm going too.15:13
Mithrandirspeaker lineup looks pretty good15:13
*** mdenny has joined #openstack-infra15:13
* ttx just decided to skip Pycon US(NA) next year15:13
Mithrandirbut it's a 24 hour flight for me, so ugh.15:14
jeblairMithrandir: it's going to be fun!  mordred will be there, clarkb's giving a talk on logstash, i'm sure lifeless will be there as well15:15
zulttx: nooooooo15:16
Mithrandirjeblair: yes, and lots of folks I haven't seen in a while.  Would also be great to meet some of you guys in RL.15:16
*** dcramer_ has joined #openstack-infra15:16
anteayattx it is in Montreal next year15:17
mordredttx: you're not coming? I'm going to talk about pbr (I hope)15:18
ttxmordred: falls like at releaseweek-115:18
mordredttx: oy15:18
ttxmordred: and at ODS-4weeks15:18
mordredttx: so - in the world of shared calendar things15:19
mordredttx: perhaps we should make one with conferences overlayed with release schedule overlayed with things we know exist that more than one people do like burningman15:19
mordredjust so we can be aware15:19
ttxmordred: we could definitely use a "things we may care about" calendar15:19
ttxto help us pick dates for stuff15:20
ttxmordred: not sure what's the best tool for that. Wiki?15:20
fungimordred: you never know, there may eventually be more than one person attending burning man too15:20
*** sandywalsh has joined #openstack-infra15:21
*** markmcclain has quit IRC15:21
jeblairfungi: what ever happened with the zuul HEAD thing?15:21
mordredfungi: :)15:21
mordredttx: me either. I _think_ there is a shared calendar thing in owncloud15:21
fungijeblair: i could not find any commonality between the three affected projects, nor was there any sign of distress in zuul's debug log when it cloned them15:21
mordredmaybe if we ever stand one of those up we can use it15:21
anteayaowncloud is on my radar, after puppet-dashboard15:22
mordredfungi, jeblair: just had a thought about those15:22
ttxmordred: I'd definitely welcome some way of submitting meetigns changes that didn't involve me watching a wiki page changes and syncing them to gcal15:22
fungijeblair: merely echoing the appropriate string for the master branch to the HEAD file in each of them was sufficient to get them back into working order15:22
mordredfungi, jeblair: I _think_ at least 2 of them were ones where there was an initial problem with the import and where we force pushed the right thing in15:22
mordredas in, I _know_ that I screwed up the mirantis one and had to force-push a fix. and I think I remember doing that for one of the others too15:23
fungimordred: ahh, so perhaps the repo in gerrit was lacking a default branch when zuul first tried to run a job for them? and then they got updated later (zuul doesn't update the remote HEAD)15:23
*** mrodden has joined #openstack-infra15:23
SergeyLukjanovjust remembered, are there any thoughts about
ttxthe unbelievably early CFP deadline for PyCON made me think too much about it, I guess15:24
fungimordred: also fuel-main was affected though. and it looked like it was added simultaneously with several other fuel-* projects which weren't15:24
*** rnirmal has joined #openstack-infra15:25
fungiwhich is also odd because it looks like fuel-main has never had any changes submitted to gerrit (none open, merged, abandoned), so why would zuul have cloned it already?15:27
fungioh, wait, searched in gerrit wrong15:27
sdaguemordred: I've done a lot of calendaring things in drupal as well, if something along those lines seemed interesting15:27
*** dcramer_ has quit IRC15:27
fungibut yeah, still not finding any when searching correctly either15:28
* fungi looks back in the debug log to see what triggered cloning that15:29
*** yaguang has joined #openstack-infra15:33
ttxmordred: should I go ahead and merge tags back for H3 manually ? I'll use a human-usable version of that .sh in your patch15:34
*** prad has joined #openstack-infra15:38
*** DennyZhang has quit IRC15:48
*** dcramer_ has joined #openstack-infra15:50
*** erfanian has quit IRC15:50
*** thomasm has quit IRC15:52
*** sridevi has joined #openstack-infra15:54
*** thomasm has joined #openstack-infra15:54
dhellmannfungi, jeblair : trying again for a wsme release. makes me think the tarball job worked, but I don't see any other jobs under
*** pblaho has quit IRC15:56
fungidhellmann: looks like (from zuul's layout.yaml) your pre-releases are set to only generate a tarball and not upload to pypi15:56
fungithat's a fairly common configuration15:57
dhellmannfungi: ah, damn it15:57
dhellmannI thought we got that all sorted15:57
* dhellmann sees the light15:58
fungiso you should be able to find the tarball created with the correct filename/contents sitting on (the shoutycaps there worry me a little)15:58
dhellmannfungi: so this is because we're doing a pre-release (b4) and not a real release?15:58
*** pycabrera has joined #openstack-infra15:58
*** dcramer_ has quit IRC15:58
fungidhellmann: yes, i the release pipeline you have it set to also upload to pypi after a successful tarball creation15:59
dhellmanngot it15:59
* dhellmann hates beta release tags anyway15:59
fungibut for pre-release tags you only make tarballs, based on the current zuul config for wsme15:59
*** cppcabrera has quit IRC16:00
*** pycabrera is now known as cppcabrera16:00
dhellmannfungi: I'll talk to cdvienne about whether we can just tag this 0.5 or if we should add pre-release jobs, thanks16:00
fungianyway, i'd take the opportunity to at least download the generated tarball and make sure it's versioned the way you intended16:00
dhellmannthe date at the end looks odd16:00
*** sridevi has quit IRC16:00
* fungi agrees. i think pbr may be doing something you're not anticipating there (is this pbr-based?)16:01
UtahDavemordred or fungi: mordred commented on the Salt review that he wanted to create a dedicated Salt Jenkins slave.  Is there someone in particular I can discuss that with?16:01
dhellmannfungi: the version comes out as 0.5b4dev-2013091116:02
fungiUtahDave: i can spin one up. i'll get a review up in a sec to add a special-purpose slave for it and make sure everyone else is on the same page16:02
dhellmannthat doesn't look like anything I'd want16:02
dhellmannaha, there's a version in the setup.cfg file16:02
UtahDavefungi: ok, cool. thanks!16:02
mordredttx: yes. although I think we might need to give you ability to propose merge commits16:03
fungiUtahDave: i need to do a related review to revert the "disable salt" change as well16:03
mordreddhellmann: sorry, was looking away for a sec - reading scrollback16:03
*** jcoufal is now known as jcoufal_mtg16:04
ttxmordred: feel free to merge them back yourself :)16:04
mordredttx: I'll do that16:04
mordreddhellmann: wsme?16:04
ttxmordred: just let me know when done so that I delete MP branches16:04
mordredttx: you can delete them now - I can merge the tags16:04
mordredwithout the branches16:04
dhellmannmordred: yes, I'm trying to cut a release to fix the six pinning problem16:05
* mordred looks at what you've got16:05
ttxmordred: [keystone, glance, nova, horizon, neutron, cinder, ceilometer, heat, trove]16:05
dhellmannI have a version in my setup.cfg, which I think is wrong16:05
mordreddhellmann: (also, fwiw, soon we'll publish pre-release versions to pypi too)16:05
mordredyes. you should remove the version from your setup.cfg16:05
ttxwill remove branches tomorrow. eod here16:06
mordredthat's only useful if you want to do versions like nova and glance do with intra-versions being marked as versions before things16:06
mordreddhellmann: if you leave a version out of there, all versions will come from signed tags in the repo16:06
dhellmannmordred: that's what we want16:06
mordred(you have to make sure to sign the tags, unsigned tags are ignored)16:06
mordreddhellmann: also, btw - you don't need to list subpackages - pbr recurses for you16:07
*** reed has joined #openstack-infra16:07
*** pblaho has joined #openstack-infra16:07
dhellmannmordred: ok16:07
mordredbut it doesn't hurt - just happen to be looking at your file right now :)16:07
dhellmannmordred: where does the version come from for an sdist? is the metadata updated when the sdist is created?16:07
mordredyes. the version is put into the PKG_INFO file16:08
mordredand is read from there if it exists16:08
*** hemnafk is now known as hemna16:08
mordredso, the only thing that doesn't work is a github tarball snapshot, because it has neither the files created sdist nor the git metadata16:08
mordredbut those are a pile of crap anyway :)16:08
*** erfanian has joined #openstack-infra16:09
mordreddhellmann: are you doing python release somewhere?16:09
mordredoh - nevermind16:10
mordredI misread it16:10
*** mriedem has joined #openstack-infra16:10
mordreddhellmann: [egg_info]16:10
mordredtag_build = dev16:10
mordredtag_date = true16:10
mordredI've never used those options with pbr before16:10
mordredbut I think we might be doing double-duty here16:10
dhellmannwonder where that stuff came from16:10
mordredhysterical raisins probably16:11
mordredit's a way to do some of what we do with pbr differently16:11
*** mriedem1 has quit IRC16:11
*** gyee has joined #openstack-infra16:13
mordreddhellmann: you beat me16:13
dhellmannmordred: I'll merge yours into mine16:13
mordreddhellmann: great16:13
mordredjeblair, clarkb: I believe I have addressed your concerns here:
*** rev has joined #openstack-infra16:15
jeblairmordred: ttx has a question in there16:15
mordredjeblair: ah - thanks.16:16
mordredttx: ^^ the answer is yes, it is ok to merge back tags from the same branch multiple times16:16
*** jpich has quit IRC16:19
*** yaguang has quit IRC16:19
*** reed has quit IRC16:20
*** sarob has joined #openstack-infra16:21
*** pblaho has quit IRC16:23
clarkbI love that pycon and burning man potentially cause problems for our release schedule16:26
*** sarob has quit IRC16:26
*** matty_dubs is now known as matty_dubs|lunch16:27
*** sarob has joined #openstack-infra16:27
*** afazekas has quit IRC16:27
jeblairi'm not sure burning man really put a dent in project velocity16:27
mordrednope. I think you guys did great without me (possibly better without me distracting you with useless things)16:29
mordredbut, in general, there's at least a few openstack people who go, so being aware that it's happening is reasonable16:29
zulshiny squirrel!16:29
*** boris-42 has quit IRC16:29
mordredmmm. squirrel16:29
mordredspeaking of - I'm eating lunch at a place called Bacon today16:30
jeblairpretty sure you eat lunch there every day16:30
clarkbI am going to do manual review.o.o backup setup, unWIP the change that does the puppet bits today16:30
clarkbthen go to discovery park (or do something to enjoy the nice weather before fall happens)16:31
clarkbmordred: but now I really want bacon16:32
jeblairok, i sumbitted the overview talk to pycon.  that wasn't easy.16:34
*** dguitarbite has quit IRC16:35
mordredjeblair: no. it wasn't16:36
mordredjeblair: they want more info than a normal conference :)16:36
jeblairand they want it 10 months in advance.16:36
mordredAlex_Gaynor, dhellmann: ^^ re: jeblair submitting OpenStack Infra overview talk16:36
mordredjeblair: I know!16:36
mordredI mean, jeez16:37
jeblairsubmitting code to openstack probably feels like this.16:37
*** dguitarbite has joined #openstack-infra16:37
clarkbjeblair: it is a plateau though, (I pushed changes to swift, devstack, devstack-gate, and config yesterday and didn't have to change anything)16:38
*** reed has joined #openstack-infra16:41
*** marun has joined #openstack-infra16:46
*** dachary has quit IRC16:46
*** dkranz has joined #openstack-infra16:46
UtahDavefungi: I'm going to be afk for a couple hours. I'll check back to see if you've had a chance to set up the salt jenkins slave16:47
*** dachary has joined #openstack-infra16:47
fungiUtahDave: sure. should get to posting those reviews here shortly16:49
*** matty_dubs|lunch is now known as matty_dubs16:50
jeblairclarkb, fungi: ^16:51
ekarlso-jeblair: care for a question ?16:51
ekarlso-could zuul be used to split up a java test run into parallells ?16:52
clarkb is ready for review16:52
jeblairekarlso-: zuul runs jobs, if you can split up a test run into multiple jobs, it can trigger them at the same time.16:52
*** fbo is now known as fbo_away16:53
mordredjeblair: what's A and B - just two different runs?16:54
fungijeblair: i'm having trouble wrapping context around that paste. i get the region abbrevs but not sure what a and b correspond to16:54
jeblairmordred: yes16:54
*** UtahDave has quit IRC16:54
mordredthose dfw numbers seem pretty highly variable16:54
jeblairthey were triggered at the same time on two different days, so should be very comparable16:54
fungisuccess and failure of... a devstack-tempest job in the gate pipeline?16:55
jeblairi lean toward thinking that iad is faster because it is not loaded.16:55
clarkbjeblair: that seems to be the initial implication16:55
mordredjeblair: I'd lean towards thinking that as well16:55
jeblairfungi: no, a trial run of the full tempest run16:55
fungigot it16:55
jeblairfungi: gate-tempest-devstack-vm-{iad,ord,dfw}-trial16:56
fungiand yeah, agreed... we're probably the only ones on our compute nodes in that region16:56
jeblairespecially since we see a 36 min run in both iad and dfw16:56
fungiso no contention with neighbors16:56
jeblair(got lucky once in dfw, not so lucky in ord)16:56
*** nati_ueno has joined #openstack-infra16:57
fungihow do those compare to hp?16:57
jeblairfungi: 25-30 seems to be the range for hp based on gate runs16:57
mordredhp norm is 25 minutes right now16:57
fungiso no appreciable improvement from the new region then, other than being a ghost town at the moment16:58
*** dcramer__ has joined #openstack-infra16:59
*** markmcclain has joined #openstack-infra16:59
clarkbjog0: seems like a common issue, is that a load problem where heartbeats don't happen or do you expect something more broken than that?17:00
uvirtbotLaunchpad bug 1221987 in nova "compute node heartbeat out of sync causing scheduler to fail in devstack:  VMs fail to spawn" [Critical,Confirmed]17:00
*** krtaylor has quit IRC17:03
openstackgerritKhai Do proposed a change to openstack-infra/config: new maven plugin jobs to build and deploy maven plugins to maven nexus
*** adalbas has quit IRC17:04
*** harlowja has joined #openstack-infra17:05
*** zehicle_at_dell has quit IRC17:05
zaroclarkb: ^ check this to sync up with creation of new user for maven central repo.17:06
zaroclarkb: sonotype doesn't allow you to upload directly to their release repo.  it requires a 2 stage release, first upload to staging repo then promote to release repo.17:08
Alex_Gaynormordred, jeblair: Yeah, it's crazy to have a CfP this far out, but with hundreds of proposals to be read and rev iewed individually this is what you gotta do to have a program by january so people can tell their company to buy tickets17:08
*** SergeyLukjanov has quit IRC17:09
zaroclarkb: i think we'll need to test that after you create the account.17:09
Alex_Gaynor(Don't ask me when conference management becamse a thing I had an opinion on, it just... happened)17:09
jeblairAlex_Gaynor: yeah, I understand.  it seems that like openstack, the process is optimizing for the reviewers since there is no shortage of potential contributors.17:09
*** zehicle_at_dell has joined #openstack-infra17:12
mordredI also forget that there are people who don't book tickets or travel until they see the program17:13
Alex_Gaynormordred: yup, a lot of people who need to justify a thing to their boss17:13
*** dizquierdo has left #openstack-infra17:14
*** ruhe has quit IRC17:14
*** hashar has joined #openstack-infra17:15
*** hashar has quit IRC17:16
*** adalbas has joined #openstack-infra17:18
clarkbswift doesn't have a base test case class apparently...17:18
*** nicedice has joined #openstack-infra17:18
clarkbwell the tests use unittest.TestCase /me chainsaws safe_repr into a common location17:18
jog0clarkb: I have no idea actaully17:22
portanteclarkb: that would be a good change to make for swift17:22
jog0I have a few patches up to test some things though17:22
portantebase test case class17:22
clarkbportante: yeah it would make finding a home for safe_repr easier. It is something we will probably end up doing to testr17:23
clarkbportante: so that we can properly capture logging, stderr, and stdout into the subunit stream when using subunit17:23
portanteclarkb: there is another set of tests that have/use a safe_repr in the tree (I forgot where)17:24
portanteif you want to make that change, i'd support it17:24
clarkbportante: yeah I am ripping it out of that change and putting it in test/ so that it is available to unit and functional17:25
*** Ryan_Lane has joined #openstack-infra17:25
clarkbportante: get_config seems to be done the same way17:25
*** ruhe has joined #openstack-infra17:25
*** adalbas has quit IRC17:27
*** Ryan_Lane1 has joined #openstack-infra17:27
*** Ryan_Lane has quit IRC17:27
clarkbportante: should that be two changes or one?17:27
dhellmannfungi: I'm not finding any other projects with a pre-release job to upload to pypi. Is that "against policy" or just not something anyone other than WSME wants to do?17:28
portantelet's see it as one first, might not be so bad17:29
clarkbportante: ok, pushed17:29
fungidhellmann: i think it's just not something other projects have wanted to do in the past (particularly since up until recently, i think, pip didn't really distinguish between them)17:30
openstackgerritMonty Taylor proposed a change to openstack-dev/pbr: Rework run_shell_command
*** KennethWilke has joined #openstack-infra17:30
fungidhellmann: for official projects, the usual solution has been to depend on specific tarball urls in the requirements files when another official project's beta releases are needed17:31
openstackgerritKhai Do proposed a change to openstack-infra/config: new maven plugin jobs to build and deploy maven plugins to maven nexus
fungiwhich prevents people who use older pip from accidentally upgrading to a beta release on pypi17:32
dhellmannfungi: yeah, we're trying to maintain coordination with non-openstack projects. I'm working on a patch to add the pypi publishing jobs to our pre-release job list. I'll add you as a reviewer, if that's OK?17:32
fungidhellmann: sure17:32
jog0clarkb: ping17:32
openstackgerritDoug Hellmann proposed a change to openstack-infra/config: Add pre-release PyPI jobs for WSME
clarkbjog0: pong17:33
jog0clarkb: either isn't running or my patch is broken17:33
Alex_Gaynordhellmann: so it looks like the double hacking thing is blocking
jog0I was wondering if you could confirm which it is17:34
clarkbAlex_Gaynor: that was the other thing I wanted to fix before AFKing17:34
clarkbjog0: looking17:34
jog0clarkb: thanks17:34
mordreddhellmann: we haven't gotten to it yet17:34
mordreddhellmann: it'll be a thing we do for all of our projects once we have creation and uploading of wheels sorted17:35
clarkbAlex_Gaynor: I think the way to solve the problem is an update to that job so that it doesn't try installing test-requirements with globale-requirements because they are different now17:35
mordreddhellmann: the problem is, pip pre-1.4 doesn't know how to ignore prereleases on pypi17:35
clarkbmordred: can yo utake a look at the thing Alex_Gaynor points out?17:35
mordredso "pip install wsme" will get you latest pre-release, not latest release if you are running pre-1.4 pip17:35
Alex_Gaynormordred: "upgrade all teh pips" :P17:35
mordredAlex_Gaynor: yup. in the mean time though...17:35
mordreddhellmann: our solution is going to be to upload pre-release wheels to pypi17:36
mordreddhellmann: only 1.4 knows how to find them17:36
*** Ryan_Lane1 has quit IRC17:36
mordredso we'll get the 1.4 behavior even if you're using pre-1.417:36
*** Ryan_Lane has joined #openstack-infra17:36
mordredbut we only upload pure-python wheels anyway, so the problem with platform specific binary wheels should not hit us17:36
mordredclarkb: looking17:36
clarkbmordred: global-requiremenst and test-requirements both have hackign in them17:37
mordredclarkb: AH17:37
clarkbmordred: then we try installing both lists together. But test-requirements are now self gating so we should only need to focus on global-requiurements17:37
*** pmoosh has joined #openstack-infra17:38
dhellmannmordred: sounds like a good plan17:38
mordredtest-requirements in the requirements repo is just for requirements17:38
* dhellmann just wants to get a release uploaded today17:38
mordredclarkb: I thikn we should remove it from the list of things in the gate-requirements-install job17:38
mordredclarkb: patch coming17:38
*** adalbas has joined #openstack-infra17:39
mordredclarkb: actually - I think adding global-requirements.txt to that list was a mistake17:39
mordredwait. hrm17:40
clarkbmordred: that test is there to make sure that the list of requirements can install together17:40
mordredclarkb: gate-requirements-install only runs on changes to requirements17:40
*** kiall has quit IRC17:41
clarkbmordred: right, it is a simple test to make sure that the list isn't uninstallable17:41
mordredright. so, if that's what it is, it's completely wrong now17:41
clarkbmordred: yah, it was missing global-requirements so that got added, but pruning needed to happen too17:41
clarkbzaro: ^17:42
mordredI'm now wondering if we need this job at all anymore17:42
mordredgiven that we're gating requirements on devstack now17:42
mordredwhich is updating the actual project lists17:42
clarkbjog0: the recheckwatch daemon isn't running so no updates are happening17:42
mordredand installing things in a sensible order17:42
clarkbmordred: I think we do beacuse devstack doesn't check the whole set17:42
clarkbit checks whatever subset is installed by your enabled services17:43
*** hashar has joined #openstack-infra17:43
jog0clarkb: ahh17:43
jog0thats not good17:43
*** krtaylor has joined #openstack-infra17:44
openstackgerritMonty Taylor proposed a change to openstack-infra/config: Only process global-requirements.txt
mordredclarkb: ^^17:45
*** HenryG has quit IRC17:45
fungimordred: jeblair: clarkb: it was pointed out earlier (by GheRivero) that we have some cruft repos on github, mostly from project org moves. here's what i get when diffing them programmatically...
dhellmannAlex_Gaynor: we have to get a good WSME release out before we update anything anyway17:45
clarkbjog0: best guess is new bug in new code was tickled causing it to die?17:46
clarkbjog0: did any bugs merge?17:46
Alex_Gaynordhellmann: that review request seems to indicate that there's a beta release that's good?17:46
fungimordred: jeblair: clarkb: i can delete the stackforge copies from github if there's consensus. i gather the melange project deletions probably merit further discussion however17:46
dhellmannAlex_Gaynor: there will be when I can publish it. :-)17:47
Alex_Gaynordhellmann: ah!17:47
jog0clarkb: no :(17:47
jog0mordred: said he restarted the service last night so it must have been down since then17:47
* dhellmann needs lunch17:47
*** dhellmann is now known as dhellmann_17:47
*** blamar has joined #openstack-infra17:47
* mordred checked and it was running after restart17:47
clarkbmordred: probably ran until an event came it that required it to do work17:47
mordredclarkb: that's what I'm thinking17:48
*** dcramer__ has quit IRC17:48
mordredfungi: I think removing the first set is a no-brainer17:48
mordredfungi: for melange, I kinda think we should either add them to the cgit replication, or remove them17:49
mordredfungi: but, I don't like the idea of removing from gerrit, because it did exis17:49
mordredso perhaps we should just add it to the cgit replication and live with the fact that we have an abandoned project17:50
jog0think I found my bug17:50
mordredwe could rename it to stackforge/melange17:50
jog0if the duplicate bug isn't in the list of rehecks already it fails17:51
clarkbjog0: ah17:51
clarkbmordred: fungi: what is wrong with leaving it only on gerrit17:51
*** kiall has joined #openstack-infra17:51
clarkbgerrit is canonical, if we choose not to mirror a dead project I Think that is fine17:51
mordredclarkb: it's kinda weird17:51
jeblairi think we should mirror it or delete it.  we could ask the tc for their opinion.17:52
mordredclarkb: it means that now we've created a de-facto project categorization called "dead" which has different rules17:52
* mordred doesn't like deleting because it feels like deleting history17:52
openstackgerritA change was merged to openstack-infra/devstack-gate: Log pre and post test hooks to file and stdout.
mordredalso, I told some peope who wanted to "remove their repo from stackforge" that we don't do that17:53
*** salv-orlando has quit IRC17:54
mordredso I think deleting the melange repo and review history would set a bad precedent17:54
hub_capping dtroyer got a devstack+trove question17:54
*** HenryG has joined #openstack-infra17:54
jeblairmordred: i don't feel compelled to do something for a stackforge project because we do it for an official project, but nonetheless, i'm okay with keeping the repo around for the change history, etc.17:55
zaroclarkb: pruning global-requirements.txt?  there was no mention of that in review.  does this need a bug?17:55
clarkbzaro: right, we missed it. mordred has a proposed fix up17:55
clarkbzaro: just pinging you as you were involved17:55
jeblairmordred: (and if we keep it around, i'm in favor of mirroring it)17:56
mordredjeblair: I agree17:56
mordredjeblair: remaining question - should we move it to stackforge since it is no longer an active core project?17:56
mordredwhich would preserve the history, but also not have it sitting in the list of 'real' thing17:57
hub_caphorray for moving melange! (sry /me is lurking)17:58
zaroclarkb, mordred : ahh ok.  i see the patch.  it's going to conflict with
*** david-lyle has quit IRC17:58
mordredzaro: I'm fine with just removing the extra things from
mordredzaro: in your patch, REQ_FILES should just become "global-requirements.txt"17:59
zarook, want to abandon yours then?17:59
mordredzaro, clarkb- or we can land my patch to get the other things unstuck if we want to wait longer on zaro's patch18:00
*** boris-42 has joined #openstack-infra18:00
zaromordred, clarkb : yeah, i think i prefer to land yours first.  then i'll just rebase.  good to keep to seperate changes with messages.18:01
jog0that should fix it18:01
openstackgerritJoe Gordon proposed a change to openstack-infra/config: if duplicate bug not in recheckwatch, add it in
*** alexpilotti has quit IRC18:02
jog0about to test it itoo18:02
openstackgerritSergey Kolekonov proposed a change to openstack-infra/jenkins-job-builder: Publish Over SSH Plugin option added
*** melwitt has joined #openstack-infra18:03
*** mrmartin has joined #openstack-infra18:04
fungiclarkb: i have no problem with leaving it on gerrit, just pointing out the current situation, and wondering whether to leave it on github if we're not going to mirror it to git.o.o18:04
*** ruhe has quit IRC18:05
*** lnxnut has joined #openstack-infra18:06
fungii'll go ahead and remove the stale duplicate github cruft from the stackforge->whatever org moves for now18:07
fungialso, something worth considering for this weekend's marconi move... i assume we need to perform similar directory mv operations on each of the current gitXX.o.o servers?18:08
zarofungi: i submitted 2 patches for the same bug with 'fixes bug' but only one of the patches reported to launchpad.  shouldn't both report back to the lp bug?18:09
*** yolanda has quit IRC18:10
zarofungi: this is the one that didn't report back
jog0clarkb: just tested the fix and it works18:10
fungizaro: which reviews? note that the comment only gets added if the fixes-bug header is in the commit message in the first patchset. if it gets added to a later patchset's commit message it won't add a url comment in the bug18:10
fungizaro: ahh, having a look18:11
clarkbjog0: cool change has my +218:11
jog0clarkb: thanks18:11
fungizaro: bug 1082812 is opened against the openstack-ci project on lp while review 46103 is for the clouddocs-maven-plugin project on gerrit (which is not mapped to the openstack-ci project on lp via jeepyb's
uvirtbotLaunchpad bug 1082812 in openstack-ci "Move rackspace/clouddocs-maven-plugin to openstack-dev/clouddocs-maven-plugin" [High,In progress]
*** boris-42 has quit IRC18:15
fungizaro: you would have needed to add a bugtask to that bug for the clouddocs-maven-plugin project (assuming it exists on lp at all) for that to work18:15
*** dcramer__ has joined #openstack-infra18:15
zarofungi: exist where? on lp?18:17
*** boris-42 has joined #openstack-infra18:17
fungizaro: yes. is there an lp project for it?18:17
*** vipul is now known as vipul-away18:17
zarofungi: nope.18:18
*** KennethWilke has quit IRC18:18
fungizaro: for a gerrit review to update a bug on lp, the review and the bug have to be for the same project. there is a mapping in jeepyb for when the project name on lp is different than in gerrit, and a bug in lp can correspond to multiple lp projects by using bugtasks18:18
zarofungi: ok. should that mapping be changed to include clouddocs-maven-plugin project?18:19
fungizaro: also your bug title says you're moving the project into openstack-dev but review 39337 seems to have imported it into stackforge instead. not sure which is correct but i suspect you want to update the bug title18:20
fungizaro: what lp project do the clouddocs-maven-plugin devs expect bugs for it to be opened in?18:20
fungizaro: could be as simple as creating a clouddocs-maven-plugin project on lp for it18:21
fungibut it's really up to the core devs on that project as to how they want to handle it18:22
*** dprince has quit IRC18:22
zarofungi: ok. creating lp project for it would work.  but seems like not necessary.  looks like all bugs are in openstack-infra project for this plugin. so can we map this project to openstack-infra?18:22
fungizaro: is clouddocs-maven-plugin going to be maintained by the infra team? if so, why was it imported into stackforge?18:22
fungi(rather than imported into openstack-infra)18:23
zarofungi: hmm. good question.  probably a question for someone with more background.18:23
clarkbit is not going to be maintained by us18:23
zaroclarkb: so should we create lp clouddocs project?18:24
clarkbzaro: no, dwcramer and annegentle can do that if they like18:24
*** hartsocks has joined #openstack-infra18:24
fungizaro: yes. the fundamental questions are "who will triage the bugs opened against clouddocs-maven-plugin, and who will be committing patches/making overall decisions on its development direction"18:24
annegentlefungi: we already triage bugs against clouddocs-maven-plugin in openstack-manauls, tagged with doc-builds18:24
annegentleDavid Cramer will continue to maintain it18:25
clarkbstackforge was the safe bet18:25
clarkbopenstack-dev felt wrong18:25
openstackgerritA change was merged to openstack-dev/hacking: Add check for deprecated method assert_
annegentlewe have interest from other OpenStack companies in helping with it18:25
annegentleclarkb: yeah stackforge is good18:25
fungiare we treading on any questionable project boundaries by tying a stackforge project's bugs to the openstack-docs project in lp then?18:25
clarkbannegentle: yeah it seemed to most closely mimic the old setup18:25
*** vipul-away is now known as vipul18:26
fungiand if so, then it probably deserves a separate project on lp18:26
clarkbfungi: not if clouddocs is already using it (it is/was hosted under rackerlabs github org)18:26
fungik, just making sure it's sane to keep doing that18:26
*** dachary has quit IRC18:26
annegentlefungi: I don't think so, do each of the tempest/tox/other QA-y things track separately?18:26
annegentlefungi: yeah good questions18:27
fungiwell, not tox. we don't write tox18:27
*** dachary has joined #openstack-infra18:27
fungibut for example
clarkbfungi: fwiw I do think it would be better to split the bugs out, but if this is the way it has been with an even bigger gulf between github and lp...18:30
*** SergeyLukjanov has joined #openstack-infra18:30
fungii'm mostly just coming at it from the angle of how we support it with our gerrit/lp bug updater. either that project needs to exist on lp or we need to map it to one which does (and if so, which one?)18:32
fungior just ignore it and deal with the fact that reviews for that project mentioning an lp bug number won't auto-update (maybe its community tracks bugs elsewhere anyway?)18:33
fungianyway, point being, i'm not saying it necessarily needs to be addressed at all, just explaining why the review in question didn't update the lp bug it referenced18:35
*** hartsocks has left #openstack-infra18:35
fungithere are several ways to potentially fix it if there is a desire to do so, but it really depends on that particular community's expected workflow18:35
*** dmakogon_ has joined #openstack-infra18:45
boris-42fungi hi18:45
fungihi boris-4218:45
boris-42fungi how are you?18:45
fungiboris-42: doing quite well! heavily caffeinated and getting work done ;)18:46
*** yolanda has joined #openstack-infra18:46
fungii hope all is well with you18:46
boris-42yeah after OpenStack CI=)18:46
*** bswrchrd has joined #openstack-infra18:47
boris-42much better then plain github_)18:47
boris-42fungi but have some question18:47
boris-42fungi do you know how to remove old branches from stackforge ?18:47
boris-42fungi I mean these 2
fungiboris-42: yes, you ask me (or clarkb or jeblair or mordred) and we delete them for you18:48
clarkbfungi: maybe I should add a bigger warning that importing a project will grab as much stuff as it can?18:48
fungiboris-42: unfortunately the permissions acl granularity on gerrit is insufficient to be able to safely delegate branch deletion for a project, or we would make that more self-service too18:48
*** svarnau has joined #openstack-infra18:49
reedselfservice rules18:49
boris-42fungi I mean is somebody able to remove these 2 branches?)18:49
fungiclarkb: yeah, wouldn't be a bad idea... "remember to prune your repository tags and branches before proposing the change to import that repo)18:49
fungiboris-42: sure, i'm happy to delete those for you right now. only takes me a moment18:49
boris-42fungi thank you=)18:49
*** thedodd has joined #openstack-infra18:50
boris-42fungi joke of the day
*** dcramer__ has quit IRC18:50
fungithat's a classic photo too. very amusing ;)18:50
fungiboris-42: so actually, it seems we imported three additional branches for rally besides master... those are named DevstackDeployment, init and test_scenarios. i'm not sure why the "init" branch didn't wind up on github, but i'm guessing you want it deleted too18:52
boris-42fungi no pls don't delete init branch=)18:53
fungiahh, it did. it's showing up as a branch which has been merged into master:
*** david-lyle has joined #openstack-infra18:53
fungiokay, so you want two published branches for rally... "master" and "init"18:53
fungiand you want DevstackDeployment and test_scenarios deleted18:53
fungidone. the branch deletion may take a few minutes to show up in the github web interface though18:54
fungiahh, it's already replicated there now18:55
fungishould be all set18:55
*** dcramer_ has joined #openstack-infra18:55
boris-42fungi thank you = )18:55
*** dcramer_ has quit IRC18:55
fungiboris-42: you're welcome!18:55
boris-42fungi btw one more joke
fungiheh. git humor. very nice18:56
jog0fungi: if you have a minute have  infra/config patch to unbreak something I broke
fungijog0: making a minute, sure18:56
fungimy dyslexia makes bug.addHit look like bongHit for some reason18:58
clarkbboris-42: fungi you may enjoy that for slightly more subtle git humor18:58
jog0fungi: :)18:59
boris-42clarkb ahaha =)18:59
*** dhellmann_ is now known as dhellmann19:00
*** vipul is now known as vipul-away19:00
*** boris-42 has quit IRC19:01
clarkbthe hobgoblin is my favorite19:02
clarkbbecause I struggle with that all the time. thankfully we have man pages19:02
pleia2clarkb: haha19:02
openstackgerritA change was merged to openstack-infra/config: if duplicate bug not in recheckwatch, add it in
clarkbI don't use aliases so silence was really funny19:03
clarkbI am sure that if I had suffered through that I wouldn't be laughing19:04
jog0clarkb: can you rekick recheckwatch19:04
clarkbjog0: puppet should do it automagically19:04
jog0ohh neato19:04
clarkbyup there is a notify => service['recheckwatch'] so the service will get restarted when the script updates19:05
*** dcramer_ has joined #openstack-infra19:05
*** dcramer_ has quit IRC19:06
jog0it hasn't restarted yet it looks like, will check up on it after lunch19:06
fungiit will take upwards of 30 minutes from merging to implementation currently19:07
fungiso lunch is warranted19:07
SpamapShmm... transient error downloading from, or more confusing?
fungijog0: actually somewhere between 0 and 25 minutes, but i like to provide conservative estimates19:09
SpamapS2013-09-11 18:33:51.117 | Downloading/unpacking websockify>=0.5.1,<0.6 (from nova==2013.2.a58.gce76e4e)19:09
SpamapS2013-09-11 18:33:51.117 |   Cannot fetch index base URL
*** david-lyle has quit IRC19:10
clarkbSpamapS: I am going to guess transient error as there aren't a ton of failures in the zuul pipelines19:12
fungiSpamapS: possible it was a network issue local to that slave, or to static.o.o or something more broken19:12
clarkbSpamapS: we have noticed occasional trouble talking through the hpcloud NAT to rackspace19:12
fungibut yeah, more likely some transparent network voodoo at the hpcloud end19:13
SpamapScool will do a recheck no bug then19:14
fungiSpamapS: of course, any time you report problems, i like to blame it on hpcloud just out of principle ;)19:14
SpamapSor is there an "occasional flakiness through the NAT..." bug? :p19:14
clarkbjeblair: fungi: mordred: I am considering a summit session along the lines of "Adding new tests to the infrastructure" to do an initial howto of the process and get feedback on making it better. But also to sort out things like when do we gate on a thing, how can we avoid/fix the asymmetric issues we had yesterday and so on19:14
SpamapSfungi: so hpcloud is the new cosmic rays... ;)19:14
clarkbdoes that seem useful? we seem to do a lot of this informally today and I want to make it less tribal knowledge19:14
SpamapSclarkb: the tribe has spoken19:15
* SpamapS snuffs out clarkb's torch19:15
clarkbqueue survivor music19:15
fungiclarkb: as long as it heavily tows the party line for getting tests into existing jobs whenever possible and only creating new jobs when that doesn't make sense, i think it would be great to cover19:16
*** atiwari has joined #openstack-infra19:16
*** SergeyLukjanov has quit IRC19:16
*** boris-42 has joined #openstack-infra19:17
fungijust don't want to wind up at the other end of the spectrum, where everyone knows how to add new test jobs and we wind up with an unnecessary proliferation19:17
clarkbmaybe a better way of asking my question would be "is something like that a valuable summit session that we want to solicit feedback on, or do we just document and formalize the rules that have grown organically?"19:17
atiwariunable to push keystone code, need some help19:17
clarkbI do see it as being partially educational19:17
atiwariatiwari@arvind-laptop:~/cloudDev/keystone-os/ks-bug-1186059 (master) $ git push19:17
atiwariUsername for '': arvind-tiwari19:17
atiwariPassword for '':19:17
atiwarifatal: Authentication failed19:17
atiwarican some one check what is going on19:18
clarkbatiwari: we don't use github for development, it is there as a handy mirror19:18
fungiclarkb: eh, good point. if it's just presenting a howto, and not soliciting discussion on how to improve the job submission workflow, then maybe not a design summit session so much as just some formalized docs19:19
clarkbfungi: I think it can be both, but you seem wary of the feedback bits :)19:19
*** lnxnut has quit IRC19:19
funginot at all wary. but if we don't expect feedback, then it might not be as well suited to the design summit19:20
fungiif we expect feedback and involved discussion (or if we can tailor it to that) then it's probably a great fit19:20
*** dcramer_ has joined #openstack-infra19:21
atiwariclarkb, thanks19:21
*** zehicle_at_dell has quit IRC19:21
clarkbfungi: I expect it. The idea has grown out of the Xen thread on the list, the last infra meeting, work with experimental queue + heat + large ops jobs, and the recent asymmetric failures19:21
clarkbthere are definitely large gaps that exist19:22
*** SergeyLukjanov has joined #openstack-infra19:22
clarkbatiwari: no problem see as well if you haven't read that before19:22
*** zehicle_at_dell has joined #openstack-infra19:22
fungiclarkb: sounds great then. i will readily admit i've been behind on some of the ml discussions since vacation19:22
clarkbfungi: don't worry I am way behind on mail too (I blame trying to get off of gmail) and that thread was started a while back by BobBall iirc19:23
fungiatiwari: and in general, we try to link new contributors to most of that information from the CONTRIBUTING.rst at the root of each of our projects19:24
clarkbfungi: gmail is a ridiculously good mail client, still fiddling with alternatives. I have decided that going back to thunderbird probably isn't going to happen19:24
zaromgagne: question about
clarkbtempted to try mutt and notmuch19:24
mgagnezaro: I have 5 minutes =)19:24
openstackgerritJames E. Blair proposed a change to openstack-infra/gearman-plugin: Set a node offline even if there is an exception
openstackgerritJames E. Blair proposed a change to openstack-infra/gearman-plugin: Always return WORK_COMPLETE when a build finishes regardless of the result.
pleia2I just caught up on hp mail this morning, openstack mail seems to be more busy than usual this week, I thought I got caught yesterday morning :)19:24
pleia2s/caught/caught up19:24
zaromgagne: did you mean capitalization should be for params on left side of = ?19:25
fungii use mutt heavily and have for over a decade. tried notmuch but it didn't mesh well with my current habits of doing everything over imap though19:25
zaromgagne: i mean only for params on left side of '=' , right?19:25
fungineed to reengineer things to use a local message cache with mutt before notmuch is relevant19:25
mgagnezaro: checking context19:25
pleia2I miss hooks in mutt (particularly for changing your email address based on folder), but otherwise gmail is ok19:25
zaromgagne: i think code would be better if pierre kept variables capitalized.19:26
clarkbfungi: ya offlineimap is common19:26
zaromgagne: that way he doesn't have to make those vars dictionaries, just lists.19:26
mgagnezaro: this is not the style elsewhere19:26
*** zehicle has quit IRC19:26
*** yolanda has quit IRC19:26
mgagnezaro: and the style used in JJB19:27
*** zehicle has joined #openstack-infra19:27
zaromgagne: ok,  just making sure you mean for all vars.19:27
mgagnezero: yes, all, left and right side19:28
*** dcramer_ has quit IRC19:28
zaromgagne: i would make an exception in this case, but i'm ok if we want to enforce caps here as well.19:28
fungipleia2: does gmail respect the mail-followup-to header?19:28
mgagnezaro: caps?19:28
jeblairinclusion_dict = {'include-by-string': 'INCLUDE_BY_STRING'} is the sort of thing i'd expect19:29
zaromgagne: sorry i mean enforce lower case.19:29
pleia2fungi: good question, I don't know (it does fine with -reply-to, but anything that calls itself a mail client should)19:29
mgagnezaro: I'll review again later today19:30
fungipleia2: i'm also way too used to adding arbitrary rfc822 headers by hand in my mail editor19:30
* zaro won't dispute author of jjb19:30
jeblairX-Arbitrary-Header: true19:30
fungipleia2: so pretty sure it would be a disappointment even if i wasn't concerned about google spying on me19:30
pleia2fungi: hah, yeah, I actually do miss that - I have mutt configured on my gmail account for when I need to do stuff like that (or things like "bounce" an email)19:31
pleia2fungi: my husband works for google and they pay my health insurance, pretty sure I already lost on the google spying on me game :)19:31
fungioh, yeah, i use mutt's bounce many times a day to resend messages. would definitely miss that19:31
pleia2I love bounce19:32
fungipleia2: oh, yeah, google probably even had you sign over your discarded toothbrushes in the clickwrap eula you signed to get health benefits19:32
clarkbreading threaded mail and mute are things I miss19:33
clarkbthough mute didn't seem to work correctly as often as I would like19:33
pleia2I do miss threaded mail19:33
fungimutt does great with threading (for me anyway)19:33
pleia2yeah it does19:33
clarkbfungi: gmail is very context aware. I haven't found anyhting that manages the top post bottom post mess that happens as well as gmail19:34
fungithough it doesn't do views threaded between different mailboxes or anything fancy like that19:34
fungiclarkb: i also delete messages which have top-posting in them. very easy to manage ;)19:34
clarkbjeblair: thoughts on the summit session discussion that began at 12:15 PDT?19:35
portanteclarkb: we had a set of 3 tempest jobs fail because of oslo-config error in module19:36
portanteis this a known problem?19:36
clarkbfungi: the way gmail formats it is if top posting or bottom posting (so not inline) you just see the content that was added and messages are displayed first to last effectively normalizing to bottom posting19:36
clarkbportante: I would need to look at the failure, do you have a link?19:37
portanteI'll post a bug19:37
*** ruhe has joined #openstack-infra19:37
clarkbportante: were they for the same chnage?19:37
fungiclarkb: i take it that totally doesn't deal with inline context quoting then (or does it just hide it so you can expand when needed)?19:38
clarkbfungi: it deal with inline by giving you context sort of like diff would19:38
uvirtbotLaunchpad bug 1224077 in oslo "oslo.config failes to install in gate job: TypeError: dist must be a Distribution instance" [Undecided,New]19:38
fungireading and thinking i would totally never be satisfied with gmails interface19:38
*** fbo_away is now known as fbo19:38
clarkbso you still don't see the entire message, just the context near the inlining19:38
*** ruhe has quit IRC19:41
clarkbportante: that is a knownish issue19:41
clarkbmordred: ^ there is the setuptools bug that clayg was working19:41
clarkbfungi: gmail is definitely not a good client if you want total control. it does a lot of things based on context. Its "threading" is based on context (and is better imo), you don't get reply to list because reply replies to list if mail came from a list and so on19:42
fungiahh. but manually curating threads by controlling in-reply-to and references headers seem to be lost in the process19:44
fungithe whole "gmail makes a new thread any time the subject header changes" would definitely screw with me19:45
*** boris-42 has quit IRC19:47
*** david-lyle has joined #openstack-infra19:50
*** sarob has quit IRC19:52
*** vipul-away is now known as vipul19:56
jeblairclarkb: i think it's worth proposing, but i'm not sure there would be a big enough audience for a tutorial type session, and as it is largely 'designed' at this point, i'm not sure we need to have a design session about it.19:59
jeblairclarkb: a howto doc, definitely.19:59
clarkbjeblair: ok20:02
clarkband now time to find lunch then do a walk around discovery park on the possibly last hot day of the year20:02
openstackgerritJames E. Blair proposed a change to openstack-infra/config: Move running-your-own to the sidebar
*** sarob has joined #openstack-infra20:06
jeblairwow, gerrit did an atrocious job on the diff for
jeblairit's just an indentation change20:07
mordredclarkb: 2013-09-11 18:36:36.828 | 2013-09-11 18:33:46 Download error on [Errno 101] Network is unreachable -- Some packages may not be found!20:07
mordredportante: ^^20:07
jeblairzaro: do you have any ideas on how to make the maven output cleaner here?
jeblairzaro: it's filled with download progress reports which aren't interesting20:08
mordredclarkb, portante: also, that oslo.config needs to be updated20:09
fungimordred: SpamapS reported an grenade run which errored trying to retrieve the package index from pypi.o.o a little while ago20:10
portantemordred: so every job downloads directly from pypi?20:11
portanteoh good20:11
mordredno jobs download directly from pypi20:11
mordredjobs download from our mirror20:11
mordredhowever, that seems to have something unhappymaking20:12
portanterats, missed that, cool20:12
mordredportante: the real issue is that there is still some latent issues with the old d2to1 codebase lurking around20:12
fungimordred: was that job also running on a devstack slave or on a static/unit test slave?20:12
mordredthat we need to stomp out20:12
portanteis there some process that audits the mirror periodically?20:12
mordredportante: audit?20:12
fungioh, hey, also a grenade job even20:12
mordredportante: we have processes that update the mirror20:13
mordredbut I'm not sure I'd call those audit jobs20:13
mordred2013-09-11 18:36:37.113 | 2013-09-11 18:33:58 Best match: oslo.config 1.2.0a320:14
mordredoh - right. we do that20:14
reedguys, what else do we need to do to have this approved?
fungimordred: clarkb and i were blaming hpcloud nat being flaky and dropping connections. your download error came within 2.5 minutes of the one in SpamapS's grenade job log20:15
mordredfungi: lovely20:16
fungimordred: for comparison, his...
jeblairreed: i think it's ready; i didn't approve it just to give other infra folks a chance to review it.  I believe since some of the names changed we do need to adjust the database and hiera configuration.20:17
fungia briefly overloaded network device somewhere (ddos burst or what have you) could certainly account for a period of random connection failures20:17
jeblairreed: since i both did those originally, and suggested the changes, it probably means i get to do it.  should be able to today.20:17
reedjeblair, wonderful20:18
mordredfungi: yah. now, I'm still not sure that's the actual reason for that failure20:19
mordredbut it defintely is an error in that run20:19
mordredI'd like to see a run with the reported issue without problems downloading pbr so we can diagnose20:19
portantemordred: regarding the mirror update, perhaps it would be worth making sure there is no bit rot in the mirror20:20
*** anteaya has quit IRC20:21
portantenow if the mirror used swift ... ;)20:21
mordredportante: I think that would be neat - but it would require someone writing some web interfaces :)20:21
openstackgerritPierre Rognant proposed a change to openstack-infra/jenkins-job-builder: Fix plot plugin support
mordredportante: right now, we just rsync files into a tree and have apache mod_autoindex them20:21
*** giulivo has quit IRC20:22
portanteperhaps rsync can us cksum for the comparison periodically?20:24
*** MarkAtwood has joined #openstack-infra20:26
*** sarob has quit IRC20:28
openstackgerritA change was merged to openstack-infra/zuul: Separate reporters from triggers
openstackgerritA change was merged to openstack-infra/zuul: Add support for emailing results via SMTP
mroddenso... looking through the jenkins-job-builder publishers i dont see any publisher that maps to the Git Publisher stuff? (to push a git ref back somewhere)20:30
mroddenam i looking in the wrong spot for that?20:31
*** jcoufal_mtg is now known as jcoufal20:31
mgagnejeblair: In JJB, why are some plugins using ", ".join(a_dict.keys()) Isn't ", ".join(a_dict) enough? Is it for backward compatibility?20:31
*** adalbas has quit IRC20:31
jeblairmgagne: if I wrote it, it's because i always forget that dictionaries are iterators for their keys.  that's still a new idea for me.  :)20:33
jeblairmgagne: i don't think that we are intending to maintain that level of backwards compat.20:33
jeblair(though i wouldn't object to that code if someone wanted to write it that way -- it is very explicit)20:35
*** SergeyLukjanov has quit IRC20:35
mgagnejeblair: it's a matter of consistency. I should rename myself to Mr. Consistency20:36
* mordred welcomes Mr. Consistency20:40
*** sarob has joined #openstack-infra20:40
zarojeblair: about your maven question.. you can use '-B' to supress the download progress.  i think that's what you are looking for right?20:41
jeblairzaro: yeah, sounds like it'd be a good idea for those jobs20:43
zaromgagne: funny, i was thinking it should be a_dict.keys() to be more explicit.20:43
jeblairzaro: also, do you have any idea why is failing?20:44
*** dizquierdo has joined #openstack-infra20:44
jeblairzaro: oh!  something's not showing up there20:44
mgagnezaro: I think .keys() would force more code to be wrapped (and made less readable) to account for the 80 characters limit20:44
*** jpeeler1 has joined #openstack-infra20:44
zarojeblair: looking..20:46
jeblairzaro: the "[ERROR] symbol  : variable currBuild20:46
jeblairlines don't seem to be showing up20:46
* zaro agrees with my consistency.20:46
*** jpeeler has quit IRC20:46
*** jpeeler1 is now known as jpeeler20:46
* fungi hides the other straws20:48
mordredapparently the problem with google maps in firefox on linux is not a problem with linux or google20:48
mordredbut with UBUNTU20:48
mordredapparently Ubuntu has managed to break Firefox20:49
funginow i'm curious... have a link?20:49
zaromordred: tell more.. is this why my firefox keeps hanging?20:49
mordredLinuxJedi just told me it works fine in firefox on Fedora20:49
fungii've used google maps in iceweasel on debian, though i try to make myself use openstreetmap instead these days20:49
zaromordred: when using g maps?20:49
mordredzaro: yah20:49
mordredsomething having to do with webgl20:50
LinuxJediyep, I didn't even know there was supposed to be a problem until mordred said and I googled for it :)20:50
LinuxJediI'm even using it hooked up to a 30" and 23" screens20:50
LinuxJediand it is pretty smooth20:50
mgagnemordred: not the first time I heard problem about google maps + firefox + ubuntu20:51
fungioh, then could also be i don't see a problem because i'm not using opengl-accelerated video hardware or something20:51
jeblairmgagne: i'm not sure if it's a code review prompting this or not, but i'll go ahead and say this just in case.  i'm not a fan of nit-picky code reviews -- i try to focus on errors, user experience, and things that will make maintenance harder.  i wouldn't leave a negative review for someone who wrote something in a slightly different way than i would.20:51
jeblairmgagne: anyway, that may or may not have anything to do with what you're working on20:51
mordredjeblair: ++20:51
jeblairjust wanted to get it out there :)20:51
mgagnejeblair: alright, I'll put my OCD to rest20:51
* mordred thinks jeblair needs that as a tattoo20:51
pleia2mordred: eww, that's no good20:52
jeblairmordred: we'd never get the kerning right to my satisfaction.  :)20:52
mordredjeblair: excellent point20:52
fungiand in agreement with jeblair, i definitely try not to make negative reviews purely on non-enforced style issues unless the code in question could be substantially simplified and the author obviously didn't know there was an easier way20:52
mordredpleia2: yeah. grumble grumble20:53
portanteclarkb, mordred: this job seems to take a long time to fail ...
mordredbtw - because I know everyone wants to keep up with my personal desktop choices - I'm now running an android emulator on my desktop so that I can use the android calendar application20:54
mordredportante: looking20:54
mordredportante: it certainly doesn't look happy20:55
pleia2mordred: wow, haha20:55
mordredpleia2: it's working surprisingly well20:55
*** ArxCruz has quit IRC20:55
pleia2I never really liked the android calendar app, preferred the browser one20:55
mordredpleia2: the android app can also sync with my terrible-corporate-borg outlook calendar20:56
pleia2mordred: aah, that could be useful20:56
mordredpleia2: but that's great - because it means that I can actually view all of my schedule in one place20:56
mordredand sadly, android is the only place I've found that can do this20:56
mordred(I agree, the web-based google calendar is a better experience)20:57
*** pmoosh has quit IRC20:58
mgagnejeblair: I still have concerns about how some plugins implements stuff differently from other plugins: docs or error reporting. My comment about using hypen-separated-words still stands and I hope you don't mind me enforcing that one =)20:58
zaromordred: ohh that's an interesting solution.20:58
* zaro likes it mgagne reviews20:59
mordredmgagne: I'd say that things that are actual operational consistency like how to do error reporting are worthwhile, because they would impact user experience21:00
mgagnemordred: ok, won't nitpick variable declaration =)21:00
zaromgagne: i'm still on the fence on the capitalization one though.21:00
mgagnezaro: I'm on the side of jeblair for that one. I would hate having BOTH styles in JJB.21:01
zarojeblair: i'll patch maven builds with the '-B'21:01
jeblairzaro: thanks21:02
zaroyes sir mr consistency sir.21:02
mgagnedamn, now I have that hat =(21:02
jeblairzaro: that one is about user experience; jenkins does all kinds of things (ALL_CAPS, CamelCase, lower_case, etc), and by being consistent in the YAML, we make it easier for users21:03
*** rfolco has quit IRC21:04
*** pcm_ has quit IRC21:04
fungimordred: to support trove cloud databases in our puppet, were you proposing moving all db connection params into hiera, or is keeping stuff like the database hostname in modules/openstack_project/manifests still going to happen?21:04
jeblairwho otherwise would need to have a copy of the documentation nearby and look up everything they type to make sure they did the right kind of capitalization/word separation21:04
mordredfungi: I think that we should make all db connection info be input params to things that consume it21:05
jeblairmgagne: yes, please do enforce the lowercase-hypen thing in yaml!  i think it was a mistake for us to not start mapping jenkins values earlier21:06
fungimordred: so put the mysql db hostname in hiera, or in site.pp, or in modules/openstack_project/manifests?21:06
mordredfungi: so, like, rather than having ::gerrit create a mysql, we should pass that info in - and in the sortterm, we can have modules/openstack_project/manifests/gerrit be the place where we make a mysql, and then pass the hostname/user/pass to ::gerrit21:06
*** rnirmal has quit IRC21:06
* jeblair reads backwards21:06
* mgagne feels better21:06
*** changbl has joined #openstack-infra21:06
jeblairmordred: have you reviewed the groups-dev change?21:06
fungijeblair: yeah, that's why i was asking about it21:07
jeblairfungi, mordred: you may want to look at it, as it has my suggestions for how to implement it21:07
mordredjeblair: I have looked over it a bit, but then there was a bunch of stuff I didn't grok21:07
mordredjeblair: lemme look again real quick21:07
fungijeblair: okay, so putting the db hostname in hiera is actually intentional in that change? if so i'll delete the comments i had about it21:07
jeblairfungi: yes.  the hostname is gobbledygook anyway.  and mrmartin has know way of knowing it.21:08
jeblair(it's appears to have a hash as part of the name)21:08
*** krtaylor has quit IRC21:09
mordredfungi, jeblair: ah, neat - are we using a trove db for the db for that one?21:10
*** sarob has quit IRC21:11
jeblairmordred: yep21:11
mordredI'm excited - and although think that's a great thing to try that with21:11
jeblairmordred: it's the way of the future, and clarkb's backup stuff seems about ready.21:11
mordredjeblair: ++21:11
mordredhub_cap: ^^ we're starting to use your stuff21:11
jeblairfigured we'd get it in at the start instead of having to rework it later21:11
fungidefinitely seems like a low-risk trial21:12
mordredhub_cap: if you break things, we'll blame you personally21:12
hub_capplz do21:12
hub_capi feed on blame and suffering21:12
hub_capcall me mordred-in-training21:12
hub_capbut srsly thats pretty freakin cool21:12
mordredjeblair: I need to run - feel free to approve groups-dev without me21:13
openstackgerritJames E. Blair proposed a change to openstack-infra/gearman-plugin: Always return WORK_COMPLETE when a build finishes regardless of the result.
openstackgerritJames E. Blair proposed a change to openstack-infra/gearman-plugin: Set a node offline even if there is an exception
*** pcrews has quit IRC21:16
*** dkliban has quit IRC21:17
stevebakerhey, how can gate-requirements-install be made to stop failing on ?21:19
hub_capjeblair: hey im a bit lost on context, what is the use case for trove? /me is curious21:20
hub_capalso, feel free to ping me anytime to help if you have questions / concerns about trove21:20
fungihub_cap: so we can have something more portable between servers than local mysql databases21:21
hub_capahh cool21:21
fungihub_cap: we're not talking about running trove, just talking about using trove as served from our providers21:21
*** changbl has quit IRC21:22
hub_capoh cool! like trove hp and trove rax21:22
* fungi nods21:22
hub_capthats even cooler!@21:22
fungiin theory, more available and more flexible than running local databases on every vm which needs one (there are quite a few things we manage which need databases)21:23
fungiwe could do most of the same things by setting up a cluster of virtual machines to serve highly-available mysql, but that's just more things which then need care and feeding21:24
hub_capand trove can do this for you, with backups sent to swift21:25
hub_capin the event of booboo's21:25
*** dkranz has quit IRC21:26
fungiall very exciting21:26
hub_capyes well you are a client now (or will be)21:27
hub_capplz file bugs and feature requests when applicable ;)21:27
fungihub_cap: we are notorious for that--don't worry ;)21:27
hub_capyes yes21:27
hub_capyou will be my favorite client21:27
hub_capcuz i dont run the infra21:27
*** blamar has quit IRC21:27
*** Ryan_Lane has quit IRC21:28
*** Ryan_Lane has joined #openstack-infra21:28
*** salv-orlando has joined #openstack-infra21:29
*** dizquierdo has left #openstack-infra21:29
*** Ryan_Lane has quit IRC21:30
*** Ryan_Lane1 has joined #openstack-infra21:30
*** mrmartin has quit IRC21:30
*** sarob has joined #openstack-infra21:36
*** thomasm has quit IRC21:37
openstackgerritPierre Rognant proposed a change to openstack-infra/jenkins-job-builder: Fix plot plugin support
*** hashar has quit IRC21:40
*** anteaya has joined #openstack-infra21:41
openstackgerritKhai Do proposed a change to openstack-infra/config: supress maven download progress messages
openstackgerritPierre Rognant proposed a change to openstack-infra/jenkins-job-builder: Fix plot plugin support
*** cppcabrera has quit IRC21:44
*** zehicle_at_dell has quit IRC21:47
*** sarob has quit IRC21:48
*** sarob has joined #openstack-infra21:48
zarojeblair: could you please tell me what this failure is about?
jeblairzaro: sorry, i'm halfway through a breaking config change; gimme a few mins; you can ignore it.21:50
*** alagalah has joined #openstack-infra21:53
alagalahNeed help on a failing build:
alagalahIt seems to be some font error that's been popping up, locally builds clean, looks good.21:54
jeblairalagalah: it appears to be a bug in the clouddocs-maven-plugin; i believe annegentle was going to talk to david cramer about it21:55
alagalahAh, many thanks!21:55
openstackgerritA change was merged to openstack-infra/jenkins-job-builder: Fix plot plugin support
alagalahjeblair: David is aware, he was at a boot camp with Colin McNamara and Sean Roberts. No solution was found21:58
*** rev has left #openstack-infra21:59
*** UtahDave has joined #openstack-infra22:00
*** pmoosh has joined #openstack-infra22:01
jeblairreed: fungi caught a couple of potential errors that clarkb and i missed; it looks like will need another patchset22:02
reedsaw the comments :(22:02
jeblairreed: on the plus side, i've set up the db and configured hiera, so when it's updated, it's ready to merge and be launched.22:02
jeblairthough, i mean, 'caught potential errors' is still on the plus side, if not as exciting.22:03
alagalahHey there, I made a whitespace change and ran it and it worked22:03
alagalahPrecise2 was the slave that it failed on22:03
reedwhitespaces are annoying22:03
alagalahPrecise31 worked22:03
* fungi hopes his reviews are a potential plus22:03
*** alagalah has left #openstack-infra22:07
openstackgerritJames E. Blair proposed a change to openstack-infra/config: Update Zuul layout to support multiple reporters
*** burt has quit IRC22:08
jeblairhub_cap: !!!!  the database i just created on rackspace is in 'error' status!22:10
*** fbo is now known as fbo_away22:10
* jeblair deletes it, and tries again22:11
*** alagalah has joined #openstack-infra22:12
hub_capjeblair: did you say the proper incantation?22:12
hub_capwhat DC are you in?22:12
hub_capalso, how big a DB you creating? /me is curious22:12
jeblairhub_cap: i said "this better work, hub_cap!" before i started22:12
jeblairhub_cap: dfw, 2gb ram, 2gb disk22:13
hub_capshiiiiii that shoudl be fine22:13
* hub_cap pops some popcorn and waits22:13
* hub_cap pictures jeblair saying "this better work, hub_cap" to the sky22:14
hub_capjeblair: you can come hit me in person, im on jefferson in berkeley22:14
jeblairfungi: if you're around, can you look at but not approve?  i'll need to manually stop zuul and restart to pick that up22:14
*** Ryan_Lane1 is now known as Ryan_Lane22:14
*** Ryan_Lane has joined #openstack-infra22:14
jeblairhub_cap: wow so many of us in berkeley now.  you know you're supposed to tell everyone it sucks here, right?  :)22:16
pleia2Ryan_Lane: how big are those devstack images? (looking at size of usb sticks I need to scour my home office for)22:16
jeblairno one else move to berkeley!  it's full!22:17
*** shardy is now known as shardy_afk22:17
Ryan_Lanepleia2: a little over 3GB22:17
hub_capjeblair: its terrible22:17
Ryan_LaneI need to make a final version22:17
pleia2Ryan_Lane: ok cool, so my 4G sticks will do it22:17
Ryan_Lanedevstack eats up a shitload of space22:17
*** lifeless has quit IRC22:17
pleia2Ryan_Lane: I figure I'll take instructions from you and build them all the Saturday prior to (anteaya will be in town that day too)22:17
*** mrodden has quit IRC22:18
hub_capworst place on earth jeblair... nothign but crotchety old people who think cell towers will give them brain cancer22:18
pleia2we might go to the zoo :) but otherwise openstack prep!22:18
*** krtaylor has joined #openstack-infra22:18
Ryan_Lanepleia2: well, the image itself should just work22:18
anteayaI was wondering when I would get to test drive that image, Saturday works well22:18
jeblairhub_cap: this one is green.  thanks for logging in and creating a db so it looked like it worked! :)22:18
Ryan_Laneit installs devstack, then unconfigures it22:18
anteayaopenstack at the zoo?22:18
pleia2Ryan_Lane: ah ok22:18
Ryan_Laneso everything is there, but re-running devstack will go through the process and start everything22:18
hub_capjeblair: shhhhh, its under my desk at home actually22:19
Ryan_Laneif you want, I could have one that just has devstack installed, but will require apt22:19
Ryan_LaneI'm assuming a bad network, though22:19
pleia2yeah, that's what I'm thinking too22:19
anteayagood assumption22:19
pleia2even on a good network, lots of people22:19
hub_capfwiw jeblair every users' first instance fails. its a good way to lower expectations22:20
jeblairhub_cap: mission accomplished!22:20
openstackgerritA change was merged to openstack-infra/config: Update Zuul layout to support multiple reporters
jeblairi'm gracefully shutting down zuul to restart it with that change ^22:23
* anteaya admires the lovely Queue only mode announcement bar in the zuul status page22:25
*** alagalah has quit IRC22:26
jeblairbeen a while since we've seen it22:26
*** tstevenson has quit IRC22:28
*** pentameter has quit IRC22:29
*** prad has quit IRC22:29
*** hashar has joined #openstack-infra22:30
*** thedodd has quit IRC22:33
*** changbl has joined #openstack-infra22:38
*** flaper87 is now known as flaper87|afk22:42
*** hashar has quit IRC22:43
jog0so I think I still broke recheck even with my second patch :(22:44
jog0is it possible to get a copy of rechecks pickle file22:44
jog0to test locally22:44
*** sandywalsh has quit IRC22:44
jeblairjog0: 1 sec22:45
jeblair-rw-rw-r--  1 recheckwatch recheckwatch 107510 Sep 11 02:57 scoreboard.pickle22:46
jeblairjog0: it is not running22:46
jeblair(and hasn't touched the pickle for a while)22:46
jog0if you try to run it, and I broke it, it will crash without touching the pickle22:47
jog0that sounded dirty22:47
*** hashar has joined #openstack-infra22:47
jeblairjog0: pickle emailed to you22:48
jog0btw I assume you tried restarting it and it didn't work22:49
*** rockyg has joined #openstack-infra22:50
*** dmakogon_ has quit IRC22:53
jeblairdhellmann: that's cool :)22:53
dhellmannjeblair: feedback on the use of "ref" in the UI would be appreciated, I'm not certain it's 100% accurate22:54
jog0jeblair: thanks I broke things :(22:54
jog0now I can debug it22:55
*** matty_dubs is now known as matty_dubs|gone22:56
*** sandywalsh has joined #openstack-infra22:56
jeblairdhellmann: that seems right to me.  i don't consider myself an expert on git terminology, but it seems like it works on more than a commit-ish (which is a tag or commit), and seems to match up with the git show-ref command.22:57
dhellmannjeblair: cool, that's what I thought22:57
dhellmannneeds more readme before a release, so maybe tomorrow22:58
dhellmannjeblair: have any thoughts on the pre-release pypi jobs for wsme?
jeblairdhellmann: my only potential concern would be whether the versioning (it's using pbr i believe) is constructed in such a way that openstack projects that consume it might start using it before they expect to.23:02
*** dhellmann is now known as dhellmann_23:02
*** dhellmann_ is now known as dhellmann23:02
jeblairdhellmann: i don't know enough about the details to answer that myself at the moment23:02
dhellmannit is using pbr23:02
dhellmannfair point on the early consumption23:02
dhellmannwe're on a beta release now, need to make another beta release with a fix for the six pinning then, and will make a real 0.5 after a couple of other patches land23:03
*** UtahDave has quit IRC23:03
jeblairdhellmann: in general though there's no reason why we shouldn't support that from an infra pov.  so i'm likely to +2 it on the basis that it looks correct and you probably know what you're doing with versions.  :)23:04
dhellmannwe'd like to be able to use the same process for beta and non-beta releases -- isn't all of openstack using the version of pip that handles betas properly?23:04
jeblairdhellmann: i don't believe so (and i don't think we assume that all of openstacks users can use it)23:04
dhellmannah, ok23:05
jeblairdhellmann: i believe that's pip 1.4, and mordred was engineering some wheel-based stuff to get the same behavior without pip 1.423:05
dhellmannyeah, he mentioned that earlier, but I wasn't clear why it was needed23:05
clarkbjeblair: it sucks in the northwest too. my walk around discovery park was aweful23:05
dhellmannI guess, for now, everyone actually does need this new beta release, so it doesn't matter this time around23:05
dhellmannand as I don't like beta tagging at all, I'm inclined to argue against it for future releases, but I'm trying to be helpful and get this out for the six issue clarkb is working on23:06
clarkbthank you23:06
dhellmannclarkb: for reference:
* dhellmann will be less accommodating about letting things get back into this state, once they are out of this state23:07
*** prad has joined #openstack-infra23:08
*** yidclare has joined #openstack-infra23:08
jeblairdhellmann: oops, actually the change isn't right; left a comment.23:08
dhellmanntime for dinner, I'll check in with you guys tomorrow morning -- have a good night23:08
dhellmannoh, ok23:08
jeblairrestarting zuul23:09
jeblairer, rather, running puppet to install new config first, then restarting zuul23:09
clarkbthis is just reporter work?23:10
clarkbno nnfi?23:10
jeblairclarkb: correct23:10
openstackgerritDoug Hellmann proposed a change to openstack-infra/config: Add pre-release PyPI jobs for WSME
dhellmannand now dinner for real23:11
*** dhellmann is now known as dhellmann_23:11
jeblairdhellmann_: good night! :)23:11
dhellmann_jeblair: have a good one :-)23:11
*** lifeless has joined #openstack-infra23:12
openstackgerritJames E. Blair proposed a change to openstack-infra/config: Lower devstack-gate timeouts
clarkbfungi: there is a gerrit mail thread from today indicating that installing bouncy castle with gerrit will fix the ssh key hash mismatch errors23:17
clarkbfungi: I am going to try adding that to the gerrit install of the git-review tests to see if it helps23:17
fungiclarkb: on the gerrit ml or one of ours?23:20
jeblairzuul has reported some check jobs and reported and merged some gate jobs23:20
clarkbfungi: the gerrit ml23:20
clarkbjeblair: jhesketh__ woot23:20
jeblairjhesketh__: ^ your reporters change is in production! :)23:20
* fungi just got back from a walk. apparently there's scrollback23:20
jhesketh__oh shiny :-)23:20
jhesketh__thanks guys23:20
jeblairjhesketh__: thank you!23:21
clarkbfungi: same here23:21
*** jcoufal has quit IRC23:21
*** salv-orlando has quit IRC23:21
jhesketh__jeblair: when you say in production, has infra's zuul deployed it?23:21
jeblairjhesketh__: yep23:21
jhesketh__sweet, nice work :-)23:22
fungiclarkb: AHA! right--the issue in the integration tests for git-review23:27
fungi(context successfully restored in brane)23:27
*** yidclare has left #openstack-infra23:28
fungiand worth noting, we already install bc on our production servers because the encryption for its contactstore feature uses it23:28
*** prad has quit IRC23:29
*** rockyg has quit IRC23:34
hub_capsdague: i think the trove devstack review is finally at a place where you might approve it. can you take a look when you have a chance? <323:35
*** anteaya_ has joined #openstack-infra23:35
*** anteaya has quit IRC23:38
clarkbfungi: I think there are two bouncy castle libs23:38
clarkbfungi: the bcpg and bcprov23:38
*** dcramer_ has joined #openstack-infra23:40
*** sarob has quit IRC23:41
anteaya_I think I am in the market for a new router23:42
fungiclarkb: yeah, we install both. i think gerrit provides the bcprov jar and we install the libbcpg-java deb to get the bcpg jar23:42
clarkbfungi: any idea where we are telling gerrit to provide bcprov?23:43
clarkbI wonder if that is a thing we did by hand and never puppetted23:43
fungiooh, maybe23:43
clarkbbecause I don't think any config is necessary beyond put .jar in place and generate keys23:44
fungii think gerrit's installer script might add it noemally23:44
clarkbbut not when you --batch by default23:44
clarkbwhich git-review is doing23:44
clarkb(and it should --batch makes more sense than using expect or something)23:44
*** sarob has joined #openstack-infra23:45
clarkbfungi: I am going to try doing what we do for libbcpg and only create the symlink if the remote end is available23:45
clarkbthat way tests should work locally without the lib, but we hopefully get more reliable tests on the slaves23:45
fungisounds good23:45
mordreddhellmann_: all of openstack is using 1.4 - but we can't assume that all openstack _users_ are yet23:46
mordredso, our infra is using 1.4 and for the most part if you're using pip and doing openstack dev at the point and not using 1.4 you're in a world of hurt23:46
*** sarob has quit IRC23:46
mordredbut as they said above, I've been working on the wheel-based thing to provide the same logic to non-1.4 people23:46
clarkbprotip just use a newish virtualenv then your global pip can be ancient23:47
mordredhub_cap: we can't really rely on your backups yet though23:48
fungii have joined the modern era as of a couple months ago and no longer use pip to do anything global23:48
mordredhub_cap: because they're all per-account and per-provider23:48
fungivirtualenv is the only pip for me23:48
mordredhub_cap: and we currently back up things to the other cloud too - for DR purposes23:48
hub_capyes very good idea mordred23:49
hub_capif only trove was multi provider aware23:49
* hub_cap wants that for the trove future23:49
*** jcoufal has joined #openstack-infra23:49
fungiclarkb: it looks like maybe we should add libbcprov-java to the packages list for ::gerrit and then symlink it similarly23:50
hub_capmulti provider replicas :)23:50
clarkbfungi: I need global pip for tox and virtualenv23:50
clarkbmordred: no we don't currently backup to the other cloud yet23:50
clarkbmordred: but we should, doing this one step at a time23:51
fungiclarkb: technically not. you could use pip to locally install virtualenv in your homedir, add it to your path, install tox into a venv and then run tox from that venv to get other venvs ;)23:51
hub_capfeel free to run backups on the trove instance clarkb, its just putting a xtraback'dup output in your swift acct23:51
hub_capeasily accessible / movable23:51
mordredclarkb: ssssh. I'm telling him theory23:52
fungithough i'm also coming around to the preferences of some of my other more dev-heavy friends who just compile and install whichever python versions they're writing for in their homedirs now23:52
mordredhub_cap: oh - I thought there was an encryption thing23:53
mordredhub_cap: maybe it's just hp that's doing that23:53
hub_capmordred: rot1323:53
hub_capthe nsa can get it anyway23:53
mordredhub_cap: you're saying I can retreive and move my db backup out of my swift and re-use it?23:53
clarkbmordred: whoops I will keep my mouth shut23:53
* fungi uses rot26. the nsa will never crack that23:54
clarkbmordred: I actually like bup for the most part23:54
hub_caposnap fungi23:54
hub_capmordred: yes iirc23:54
hub_capi can triple check for you23:54
clarkbmordred: I don't like that the non bulk backups are experimental23:54
clarkbbut for easy backups using minimal space it seems to do a good job23:54
mordredclarkb: sure. I'm just saying, one of our original issues with cloud dbs was that we could not access the magical backups23:54
mordredbut that was apparently just an hp-ism23:54
mordredhp cloud, continually subverting the actual production desires of its users since 201123:55
hub_capya we dont use a shadow acct for that mordred at rax23:55
fungimordred: if only you knew people at hp23:55
*** anteaya__ has joined #openstack-infra23:55
mordredfungi: I'm going to add this to the list of 'features' that keep me from using hp cloud for real work23:55
hub_capyall sound like me @rax23:55
mordredhub_cap: thank you, btw, for backing up my database to a place from where I can retreive it23:55
hub_capmordred: right???23:55
mordredhub_cap: and, also, thank other people at rax for not  blocking port 25 and not randomly deleting instances23:56
fungimordred: i assume it currently fills more than one page of greenbar when echoed to /dev/lp023:56
clarkbhowever all is not spectacular and rosy23:56
hub_capHAHA mordred im sure we can do all thats possible to underwhelm you in other ways23:56
mordredhub_cap: oh, you do :)23:56
hub_capps, i think the rax people dont view me as one of them anymore23:56
hub_caplucky me!23:56
mordredhub_cap: w00t! job well done23:56
fungihub_cap: you're irreversibly contaminated23:57
mordredhub_cap: well, you don't work at the castle, so you can't possibly be one of them23:57
clarkbrax could be even more awesomer if we didn't need to check ipv6 on each host we build and if the performance was better >_>23:57
hub_capclarkb: its comin, performance wise23:57
hub_capfrom the little i hear thru the grapevine23:57
mordredat least hp doesn't pretend to have ipv623:57
* mordred has heard the same thing, btw23:57
clarkbis that different than what you were testing?23:58
mordredI'm inferring from what I read that new hardware is coming or something23:58
fungii need to check the ipv6 bug again. it's been a month since i looked to see if they'd added a new excuse23:58
*** anteaya_ has quit IRC23:58
hub_capfungi: excuse == bug buffer23:59
fungiyah. the problem is that the tier 3 buffer has a built-in tier 1 expiration. it's like a dead-man swith23:59

Generated by 2.14.0 by Marius Gedminas - find it at!