| *** EvilienM is now known as EmilienM | 00:50 | |
| *** dhellmann has quit IRC | 02:55 | |
| *** dhellmann has joined #openstack-tc | 03:01 | |
| *** jamesmcarthur has quit IRC | 05:43 | |
| *** Bhujay has joined #openstack-tc | 05:46 | |
| *** Bhujay has quit IRC | 06:15 | |
| *** chenyb4 has joined #openstack-tc | 06:33 | |
| *** diablo_rojo has joined #openstack-tc | 07:03 | |
| *** diablo_rojo has quit IRC | 07:04 | |
| openstackgerrit | Omer Anson proposed openstack/governance master: Remove Dragonflow from the official projects list https://review.openstack.org/613856 | 08:19 |
|---|---|---|
| *** e0ne has joined #openstack-tc | 08:45 | |
| *** cdent has joined #openstack-tc | 09:57 | |
| * cdent brings an endless meeting of endless speculating about Red Hat and IBM to order | 10:03 | |
| cdent | Actually, sorry for being flip, this is probably challenging on many dimensions. | 10:04 |
| *** dtantsur|afk is now known as dtantsur | 10:35 | |
| *** chenyb4 has quit IRC | 10:51 | |
| *** wxy-xiyuan has quit IRC | 11:15 | |
| *** wxy-xiyuan has joined #openstack-tc | 11:15 | |
| *** spsurya has quit IRC | 11:16 | |
| *** spsurya has joined #openstack-tc | 11:18 | |
| dhellmann | o/ | 11:21 |
| cdent | welcome back dhellmann | 11:22 |
| * dhellmann looks at 3500+ email messages and considers going back to bed | 11:22 | |
| dhellmann | this may also distract me a bit over the next couple of days: https://www.redhat.com/en/about/press-releases/ibm-acquire-red-hat-completely-changing-cloud-landscape-and-becoming-world’s-1-hybrid-cloud-provider?intcmp=701f2000000RWK2AAO | 11:22 |
| cdent | I imagine so | 11:22 |
| dhellmann | but I am here | 11:22 |
| cdent | back to bed sounds like a good life choice | 11:23 |
| dhellmann | it so often is | 11:23 |
| *** e0ne has quit IRC | 11:25 | |
| *** e0ne has joined #openstack-tc | 11:34 | |
| openstackgerrit | Samuel de Medeiros Queiroz proposed openstack/governance master: Reword open source definition in the Four Opens https://review.openstack.org/613894 | 11:49 |
| *** jamesmcarthur has joined #openstack-tc | 11:54 | |
| *** jamesmcarthur has quit IRC | 11:59 | |
| dims_ | dhellmann : ++ | 12:21 |
| *** edmondsw has quit IRC | 12:29 | |
| *** edmondsw has joined #openstack-tc | 12:36 | |
| *** e0ne has quit IRC | 12:58 | |
| *** cdent has quit IRC | 13:03 | |
| *** jamesmcarthur has joined #openstack-tc | 13:07 | |
| *** e0ne has joined #openstack-tc | 13:19 | |
| *** mriedem has joined #openstack-tc | 13:19 | |
| *** elbragstad is now known as lbragstad | 13:22 | |
| *** e0ne has quit IRC | 13:24 | |
| *** jaosorior has quit IRC | 13:30 | |
| *** e0ne has joined #openstack-tc | 13:30 | |
| smcginnis | It would be great if we could get moving on the T release naming. dhellmann, I think that is just pending the approval of allowing Train on the list - https://review.openstack.org/#/c/611511/ | 13:32 |
| * mtreinish still thinks it should be the 'Treinish' release :p | 13:39 | |
| smcginnis | Hah | 13:39 |
| *** tellesnobrega has left #openstack-tc | 13:50 | |
| *** jamesmcarthur has quit IRC | 13:55 | |
| *** jpich has joined #openstack-tc | 13:58 | |
| *** jaosorior has joined #openstack-tc | 14:05 | |
| *** jamesmcarthur has joined #openstack-tc | 14:11 | |
| *** jamesmcarthur has joined #openstack-tc | 14:11 | |
| fungi | or the tonyb release | 14:15 |
| *** SteelyDan is now known as dansmith | 14:16 | |
| *** jaypipes_ is now known as jaypipes | 14:17 | |
| *** cdent has joined #openstack-tc | 14:20 | |
| *** dklyle has joined #openstack-tc | 14:47 | |
| *** irclogbot_3 has joined #openstack-tc | 15:08 | |
| *** jpich has quit IRC | 15:09 | |
| *** dtantsur is now known as dtantsur|brb | 15:10 | |
| *** cdent has quit IRC | 15:22 | |
| *** jaosorior has quit IRC | 15:38 | |
| *** cdent has joined #openstack-tc | 15:48 | |
| openstackgerrit | Merged openstack/governance master: charms: Add repository for Octavia charm https://review.openstack.org/608283 | 16:12 |
| openstackgerrit | Merged openstack/governance master: Don't quote {posargs} in tox.ini https://review.openstack.org/609132 | 16:13 |
| openstackgerrit | Merged openstack/governance master: Import puppet-crane https://review.openstack.org/610015 | 16:14 |
| openstackgerrit | Merged openstack/governance master: Retire project Anchor - Step 5 https://review.openstack.org/611187 | 16:14 |
| *** dtantsur|brb is now known as dtantsur | 16:16 | |
| mtreinish | fungi: heh, although tonyb has a potential conflict being the adminstrator for the name selection :) | 16:20 |
| openstackgerrit | Lance Bragstad proposed openstack/project-team-guide master: Add section about collecting feedback to PTL guide https://review.openstack.org/613617 | 16:26 |
| smcginnis | fungi: Do you know what the plan is for (timeline, etc.) for merging the mailing lists? | 16:29 |
| openstackgerrit | Merged openstack/governance master: New Repo - OpenStack-Helm Images https://review.openstack.org/611895 | 16:29 |
| fungi | smcginnis: http://lists.openstack.org/pipermail/openstack-dev/2018-September/134911.html | 16:32 |
| fungi | i plan to send out a reminder later today | 16:32 |
| fungi | was going to send it last week but last week got away from me | 16:32 |
| smcginnis | Oh, I forgot that actually did mention some dates. Great - thanks! | 16:33 |
| openstackgerrit | Merged openstack/governance master: Add blazar-specs repository https://review.openstack.org/612431 | 16:34 |
| *** e0ne has quit IRC | 16:36 | |
| fungi | i've sent the reminder just now | 16:54 |
| *** markvoelker has joined #openstack-tc | 17:13 | |
| *** jamesmcarthur has quit IRC | 17:41 | |
| *** jamesmcarthur has joined #openstack-tc | 17:42 | |
| clarkb | https://review.openstack.org/#/c/613674/2 generated http://paste.openstack.org/show/733572/ which is a bit more human friendly listing of zuul resource usage | 17:43 |
| *** jamesmcarthur has quit IRC | 17:44 | |
| *** jamesmcarthur has joined #openstack-tc | 17:44 | |
| cdent | wow | 17:46 |
| clarkb | the data span there is all jobs completed since Thursday ~1500UTC when re restarted zuul | 17:46 |
| cdent | I wonder if IBM will have anythning left over after the 34bn to buy some triplelo specific CI resources ;) | 17:47 |
| clarkb | script is written such that it will loo kat all available zuul logs (we have thirty days retention I think) but we only started logging this extra data at that point | 17:47 |
| clarkb | going forward we'll be able to do a thirty day snapshot though | 17:47 |
| cdent | that will be a handy report to help avoid speculation | 17:49 |
| cdent | this data gathering has certainly altered my perceptions | 17:50 |
| *** diablo_rojo has joined #openstack-tc | 17:51 | |
| dhellmann | clarkb : "resource usage" there is defined as "node time", right? | 17:54 |
| dhellmann | no log storage, etc. | 17:54 |
| clarkb | dhellmann: yes its number of nodes * job runtime | 17:54 |
| clarkb | correct | 17:54 |
| dhellmann | k | 17:54 |
| clarkb | ajaeger points out I should print the time span that data was collected for so I'll work on adding that in after I eat some food | 17:55 |
| *** jamesmcarthur has quit IRC | 17:55 | |
| dhellmann | good idea | 17:56 |
| fungi | as i said last week, i'm less interested in using that for blamethrowing and witch hunts, more as a tool to be able to see how effective efforts to improve utilization really wind up being | 18:01 |
| cdent | fungi: I was thinking of it more in terms of correcting incorrect blamethrowing. For example for a long time I've assumed that nova was much more of a hog than it is | 18:03 |
| fungi | cdent: that too. or the unfounded assertions that unofficial projects or new top-level projects under the osf are significantly eating into our available pool of test resources | 18:04 |
| notmyname | this isn't normalized for activity is it? absolute numbers are good, but so are per-patch usage, right? | 18:05 |
| fungi | notmyname: correct | 18:06 |
| clarkb | notmyname: but anyone can normalize it now because activity data is public in gerrit (this absolute data was not available | 18:07 |
| clarkb | I need to give you time frames for that though | 18:07 |
| clarkb | which is my next step | 18:07 |
| notmyname | clarkb: are there docs on how to pull it? | 18:07 |
| clarkb | notmyname: it being activity? the gerrit api docs | 18:07 |
| notmyname | ok | 18:08 |
| *** jamesmcarthur has joined #openstack-tc | 18:08 | |
| clarkb | my concern as infra ptl isn't to blame or witchhunt but is to be clearer to people that have worries that airship and starlingx etc are the problem that they are such a small amount of usage and that if we are serious about fixing nova's time to merge in the gate there are other better targets for change | 18:10 |
| fungi | useful to note that the top 20 repositories there are all official openstack deliverables, and account for 84% of overall node usage. would be interesting to find out what the total for all official deliverables together is | 18:10 |
| clarkb | I also have the selfish need/want that people will help run the infrastructure and maybe showing them that they are a major consumer of the infrastructure is a great way to enoucrage that :) (absolute numbers work fine for this) | 18:11 |
| clarkb | fungi: I can have this generate that number too | 18:12 |
| fungi | i figured, since you're already mapping to the deliverable repos | 18:12 |
| clarkb | fungi: Total OpenStack usage: 154614005.00s, 98.76% | 18:19 |
| notmyname | (fwiw, that's 4.9 years) | 18:20 |
| * dtroyer ponders a new unit of measurement, the zuul-year: how long it takes Zuul to run a year's worth of test jobs | 18:27 | |
| clarkb | ok improvements added http://paste.openstack.org/show/733577/ you now get the time range and total of openstack usage | 18:34 |
| clarkb | change for script updated to generate ^ too | 18:34 |
| *** jamesmcarthur has quit IRC | 18:41 | |
| *** jamesmcarthur has joined #openstack-tc | 18:41 | |
| fungi | thanks clarkb! so unofficial repos and top-level projects used 1.24% of the total amount. definitely miniscule given they account for over half the repositories we host | 18:42 |
| *** jamesmcarthur_ has joined #openstack-tc | 18:42 | |
| TheJulia | dtroyer: worth in time, or worth in sheer job numbers | 18:44 |
| dtroyer | TheJulia: in time, ie "the Berlin Summit will be just over 3 zuul-years long next month" | 18:45 |
| clarkb | one caveat there is I'm not doing anything to ensure those float values don't get all crazy (but you can review the script https://review.openstack.org/#/c/613674/ and help me fix that if we suspect that to be an issue) | 18:45 |
| *** jamesmcarthur has quit IRC | 18:46 | |
| openstackgerrit | Doug Hellmann proposed openstack/governance master: report the number of votes needed https://review.openstack.org/614007 | 18:49 |
| dhellmann | tc-members: we need 1 more RC+1 on https://review.openstack.org/#/c/608751/ in order to approve the meeting schedule | 18:50 |
| dhellmann | I don't see votes from mugsie, evrardjp, or fungi there | 18:51 |
| *** jamesmcarthur_ has quit IRC | 18:51 | |
| *** jamesmcarthur has joined #openstack-tc | 18:52 | |
| *** jaosorior has joined #openstack-tc | 18:52 | |
| fungi | i've voted now | 18:53 |
| zaneb | didn't realise we needed 2/3 to change that one | 18:56 |
| dhellmann | zaneb : yeah, because it's part of the charter | 18:57 |
| dhellmann | fungi : thank you | 18:57 |
| clarkb | using the gerrit query: 'after:"2018-10-25 15:16:32.079" before:"2018-10-29 18:32:23.997" project:openstack/tripleo-heat-templates' and changing the project to openstack/nova those two repos have about the same activity based on changes being updated within that time period | 19:07 |
| clarkb | (this doesn't 100% correlate to what triggers jobs in zuul, but should be fairly close?) | 19:07 |
| clarkb | notmyname: ^ thta was for you | 19:07 |
| clarkb | neutron has less activity | 19:08 |
| clarkb | (slightly) | 19:08 |
| clarkb | I can probably scrape the zuul logs to get more accurate activity levels | 19:10 |
| clarkb | since zuul alread yhas that | 19:10 |
| *** dtantsur is now known as dtantsur|afk | 19:11 | |
| *** scas has quit IRC | 19:11 | |
| *** diablo_rojo has quit IRC | 19:16 | |
| openstackgerrit | Merged openstack/governance master: T Release Name https://review.openstack.org/611511 | 19:19 |
| openstackgerrit | Merged openstack/governance master: Reword open source definition in the Four Opens https://review.openstack.org/613894 | 19:19 |
| *** diablo_rojo has joined #openstack-tc | 19:20 | |
| *** diablo_rojo has quit IRC | 19:22 | |
| *** diablo_rojo has joined #openstack-tc | 19:23 | |
| * tonyb could vote for Treinish! | 19:26 | |
| * tonyb sees that https://review.openstack.org/#/c/611511 merged so I'll get on that today | 19:27 | |
| clarkb | hrm the zuul side logs actually don't make that easier either because zuul will reenqueue things in a flaky gate type situation | 19:29 |
| clarkb | so one human activity event can be many zuul events | 19:30 |
| smcginnis | Thanks tonyb! | 19:31 |
| fungi | clarkb: also the variety of human interaction events includes not only patchset upload and change approval but also restoring abandoned changes and leaving recheck and check experimental comments | 19:49 |
| clarkb | fungi: ya | 19:50 |
| fungi | so simply counting the number of patchsets isn't going to get a terribly accurate picture | 19:50 |
| fungi | depending on what you're wanting to measure there | 19:50 |
| clarkb | on the patchset front for the most recently logfile triple-heat-elements has 31 patchsets created to neutrons 26 | 19:50 |
| clarkb | fungi: its going to be fuzzy I think becuase a check experimental in one repo will be a noop but actually be expensive in another repo | 19:51 |
| clarkb | notmyname: ^ may have a better idea of what he would expect from an activity report for scaling usage numbers? | 19:51 |
| fungi | one useful gauge of repository activity would be to simply count how many changes merged | 19:52 |
| fungi | resource utilization scaled against merged changes could be a relatively useful gauge of efficient use of the ci system | 19:52 |
| *** mrhillsman is now known as mrhillsman_lunch | 19:53 | |
| clarkb | fungi: I like that | 19:53 |
| fungi | though that still fails to take into account the relative value/complexity of different changes | 19:53 |
| fungi | and how those might vary between projects | 19:53 |
| notmyname | clarkb: cool, thanks (the info on "per-capita" cost) | 19:55 |
| fungi | the value/complexity of a spec change measured against a typo correction, for example | 19:56 |
| clarkb | fungi: using my above gerrit query neutron merged 20 changes, nova 16, and tripleo heat templates 14 | 19:56 |
| clarkb | after:"2018-10-25 15:16:32.079" before:"2018-10-29 18:32:23.997" project:openstack/tripleo-heat-templates status:merged is query, adjust project name as necessary | 19:56 |
| clarkb | notmyname: ^ | 19:56 |
| clarkb | tripleo-quickstart-extras is the other repo in top four repos by usage and it merged one change in that period | 19:57 |
| fungi | one other thing to keep in mind, amount of ci resources consumed per change merged _could_ also be a measure of how thorough your testing is. this cuts both ways | 19:59 |
| fungi | projects with effectively no testing would have a really great utilization-per-change-merged | 20:00 |
| clarkb | fungi: to understand that you probably want to look at gate resets per time | 20:00 |
| clarkb | high number of resets would indicate poor coverage for keeping bugs out | 20:00 |
| fungi | unless, again, your testing was so feeble/nonexistent that what jobs you did run simply didn't fail (and maybe also didn't exercise anything of value in doing so) | 20:01 |
| notmyname | yeah, per-patch CI resources consumed (not wall time) is interesting because a project can track resource usage over time (are our tests running faster or not). the wall time is dominated by integrated queue resets | 20:02 |
| clarkb | fair enough. We likely can't do this entirely statistically then. Would need to have objective review by humans | 20:02 |
| *** lbragstad has quit IRC | 20:03 | |
| notmyname | absolutely | 20:04 |
| notmyname | a long time ago, I tracked gate status via graphite. I had a graph showing the chance a patch would make it through the gate without a recheck and stenciled it to my desktop. we never really got above about 80% success rate. (that was a while back) | 20:05 |
| fungi | this data is probably a bit more useful for the tripleo team because they have their own gate queue separate from the main integrated queue | 20:05 |
| notmyname | with this new information, I'd anticipate being able to do something similar on a per-project basis to measure and track a project's usage of CI resources | 20:06 |
| fungi | so negative interactions related to gate resets for them are being caused by other deliverables controlled by their team | 20:06 |
| fungi | (well, or caused by bugs in other software used in those jobs) | 20:06 |
| fungi | but regardless, gate resets for the tripleo queue are happening because tripleo repos are failing jobs, not because a change ahead of theirs for some other team failed | 20:07 |
| clarkb | fungi: ya they consume the integrated gate software (and other software) so the integrated queue instability can feed into the tripleo queue | 20:08 |
| clarkb | not directly as within a queue, but a bug causing resets in nova will also likely case resets in tripleo | 20:08 |
| fungi | definitely | 20:08 |
| notmyname | what's the status of graphite in -infra these days? is it still used? I found my old stuff (much older that I realized), and at least the values are all different because I'm getting errors | 20:11 |
| clarkb | notmyname: the service is still there but the values have moved because zuul is multitenant now and the keys have tenant prefixes | 20:12 |
| clarkb | let me find examples | 20:12 |
| notmyname | ah ok. so `stats.zuul.pipeline.gate.job.gate-swift-pep8.{SUCCESS,FAILURE}` has a prefix now? (for example) | 20:12 |
| clarkb | notmyname: https://git.openstack.org/cgit/openstack-infra/project-config/tree/grafana/neutron.yaml shows you examples of the new keys | 20:12 |
| clarkb | ya would be stats.zuul.tenant.openstack.pipline.... I think | 20:13 |
| notmyname | thanks! | 20:13 |
| *** lbragstad has joined #openstack-tc | 20:18 | |
| *** jamesmcarthur has quit IRC | 20:25 | |
| *** jamesmcarthur has joined #openstack-tc | 20:30 | |
| *** jamesmcarthur has quit IRC | 20:55 | |
| *** jamesmcarthur has joined #openstack-tc | 20:56 | |
| *** mrhillsman_lunch is now known as mrhillsman | 21:04 | |
| *** weshay has joined #openstack-tc | 21:08 | |
| openstackgerrit | Doug Hellmann proposed openstack/governance master: New Repo: OpenStack-Helm Docs https://review.openstack.org/611896 | 21:16 |
| *** e0ne has joined #openstack-tc | 21:29 | |
| *** jamesmcarthur has quit IRC | 21:32 | |
| *** e0ne has quit IRC | 21:36 | |
| *** mriedem is now known as mriedem_away | 21:42 | |
| *** cdent has quit IRC | 22:13 | |
| *** eandersson has joined #openstack-tc | 22:36 | |
Generated by irclog2html.py 2.15.3 by Marius Gedminas - find it at mg.pov.lt!