| * fungi swerves | 00:00 | |
| *** achanda_ has joined #openstack-infra | 00:00 | |
| *** dannywilson has quit IRC | 00:01 | |
| *** ashleighfarnham has quit IRC | 00:01 | |
| *** mahito has joined #openstack-infra | 00:01 | |
| fungi | the local farmers market and the various produce stands on the island are finally starting to be open again | 00:01 |
|---|---|---|
| fungi | excited | 00:01 |
| lifeless | clarkb: I think you're misreading it | 00:02 |
| lifeless | clarkb: thats all to do with pypy being incompatible with released cryptography | 00:02 |
| lifeless | clarkb: the subject of the commit line lies - the setuptools req stayed | 00:03 |
| *** achanda has quit IRC | 00:04 | |
| *** apmelton is now known as apmelton_away | 00:04 | |
| *** achanda_ has quit IRC | 00:05 | |
| anteaya | fungi: what would be available at this time of year | 00:06 |
| anteaya | in spring all we have are garlic scapes | 00:06 |
| anteaya | you can only handle so many of those | 00:06 |
| *** radez is now known as radez_g0n3 | 00:07 | |
| *** ujuc has quit IRC | 00:09 | |
| *** otter768 has joined #openstack-infra | 00:10 | |
| *** otter768 has quit IRC | 00:15 | |
| bknudson | I'm wondering what's going on with https://review.openstack.org/#/c/179321/ -- it's +W and +V but not gating | 00:17 |
| *** ociuhandu has quit IRC | 00:18 | |
| *** salv-orlando has joined #openstack-infra | 00:26 | |
| lifeless | I need a set of gnarly requirements to test 988 on | 00:28 |
| *** jamielennox|away is now known as jamielennox | 00:29 | |
| anteaya | bknudson: I think having it re-approved might get it going again | 00:29 |
| anteaya | bknudson: monday and tuesday weren't great days for zuul, it might have gotten missed in a gap | 00:29 |
| anteaya | bknudson: rechecking won't get it into the gate | 00:30 |
| *** salv-orlando has quit IRC | 00:30 | |
| anteaya | bknudson: since it already had a +1 from Jenkins and further Jenkins +1's don't register as new events | 00:31 |
| bknudson | anteaya: ok. Hopefully someone who can approve will come along and re-approve it. | 00:31 |
| anteaya | bknudson: hopefully | 00:31 |
| *** mtanino has quit IRC | 00:36 | |
| jamielennox | can keystone-release create feature branches on python-keystoneclient or do we need someone in infra to set it up? | 00:37 |
| *** freerunner has quit IRC | 00:38 | |
| anteaya | jamielennox: let's take a look: http://git.openstack.org/cgit/openstack-infra/project-config/tree/gerrit/acls/openstack/python-keystoneclient.config | 00:38 |
| anteaya | not right now | 00:39 |
| anteaya | jamielennox: I suggest having a chat with release managers to see if they want to create feature branches on keystone client | 00:39 |
| anteaya | jamielennox: if no, then you'll need a change to your acl file, but start with release managers first | 00:40 |
| jamielennox | anteaya: thanks i wasn't sure how to read that to read that from config | 00:40 |
| anteaya | sure | 00:40 |
| anteaya | under access refs/heads/* | 00:41 |
| jamielennox | anteaya: who is considered the relesae managers for ksc? | 00:41 |
| jamielennox | or in genreal | 00:41 |
| anteaya | you would need a create for -release | 00:41 |
| anteaya | jamielennox: ttx and dhellmann | 00:41 |
| mordred | clarkb: nice | 00:42 |
| anteaya | jamielennox: line 3: http://git.openstack.org/cgit/openstack-infra/project-config/tree/gerrit/acls/stackforge/fuel-astute.config | 00:42 |
| anteaya | jamielennox: as an example but you would give the create permission to keystone-release in your acl file | 00:42 |
| tchaypo | lifeless: last time i ran global-requirements through pip-compile, it found a conflict somewhere down the graph | 00:42 |
| anteaya | jamielennox: does that make sense? | 00:42 |
| tchaypo | lifeless: so that probably makes a starting point | 00:42 |
| anteaya | jamielennox: only if ttx and dhellmann say they would prefer not to manage feature branches | 00:43 |
| jamielennox | anteaya: yep - that's fine, i'll try and grap dhellmann when he's online | 00:43 |
| anteaya | jamielennox: great | 00:44 |
| jamielennox | anteaya: thank you | 00:44 |
| anteaya | you're welcome | 00:45 |
| anteaya | :) | 00:45 |
| *** zhiwei has joined #openstack-infra | 00:48 | |
| *** jtomasek has quit IRC | 00:48 | |
| *** ddieterly has joined #openstack-infra | 00:50 | |
| *** erlon has quit IRC | 00:51 | |
| *** sputnik13 has joined #openstack-infra | 00:54 | |
| mordred | jhesketh, lifeless: ok. pycon au talk submitted | 00:56 |
| jhesketh | mordred: winning :-) | 00:57 |
| lifeless | jhesketh: cutting it fine | 00:57 |
| lifeless | jhesketh: should I submit something do you think? or organising going to be enough .. | 00:57 |
| mordred | lifeless: I mean, there's a whole other day I could have waited in my timezone | 00:57 |
| jhesketh | lifeless: heh, I was going to submit something today, but was thinking of targeting it towards the amin conf | 00:57 |
| jhesketh | mordred: the advantages of living in the past I guess | 00:58 |
| openstackgerrit | Merged openstack-infra/zuul: Fix race in test_crd_check_duplicate more https://review.openstack.org/181183 | 00:59 |
| *** mahito has quit IRC | 01:00 | |
| *** yamahata has joined #openstack-infra | 01:00 | |
| openstackgerrit | Merged openstack-infra/zuul: Fix race condition relating to change updates https://review.openstack.org/181175 | 01:00 |
| *** mahito has joined #openstack-infra | 01:02 | |
| *** mahito_ has joined #openstack-infra | 01:05 | |
| *** mkoderer has quit IRC | 01:05 | |
| *** mpopow has joined #openstack-infra | 01:07 | |
| *** mahito has quit IRC | 01:08 | |
| *** alexsyip has quit IRC | 01:09 | |
| *** mpopow has quit IRC | 01:12 | |
| *** mkoderer has joined #openstack-infra | 01:13 | |
| *** tiswanso has joined #openstack-infra | 01:17 | |
| *** otter768 has joined #openstack-infra | 01:18 | |
| *** shashankhegde has quit IRC | 01:19 | |
| *** tiswanso has quit IRC | 01:20 | |
| *** tiswanso has joined #openstack-infra | 01:20 | |
| *** Krinkle|detached is now known as Krinkle | 01:22 | |
| kragniz | infra-folk, is it worth fighting the rebase battle with this? https://review.openstack.org/#/c/181000/ | 01:22 |
| *** rm_work|away is now known as rm_work | 01:23 | |
| *** tiswanso has quit IRC | 01:26 | |
| *** tiswanso has joined #openstack-infra | 01:26 | |
| *** sigmavirus24 is now known as sigmavirus24_awa | 01:26 | |
| *** crc32 has quit IRC | 01:39 | |
| *** crc32 has joined #openstack-infra | 01:40 | |
| *** kaisers has quit IRC | 01:44 | |
| *** patrickeast has quit IRC | 01:45 | |
| *** rmcall has quit IRC | 01:45 | |
| *** camunoz is now known as camunoz_afk | 01:47 | |
| *** unicell has quit IRC | 01:49 | |
| *** fedexo has joined #openstack-infra | 01:51 | |
| *** kaisers has joined #openstack-infra | 01:57 | |
| *** dims_ has quit IRC | 02:00 | |
| *** dims has joined #openstack-infra | 02:00 | |
| *** dims has quit IRC | 02:01 | |
| *** baoli has joined #openstack-infra | 02:05 | |
| *** baoli has quit IRC | 02:06 | |
| *** baoli has joined #openstack-infra | 02:07 | |
| *** david-lyle has joined #openstack-infra | 02:10 | |
| *** shashankhegde has joined #openstack-infra | 02:12 | |
| *** agentleexile has joined #openstack-infra | 02:15 | |
| *** annegentle has joined #openstack-infra | 02:15 | |
| *** patrickeast has joined #openstack-infra | 02:19 | |
| *** agentleexile has quit IRC | 02:20 | |
| *** annegentle has quit IRC | 02:20 | |
| *** fedexo has quit IRC | 02:21 | |
| *** craigbr has joined #openstack-infra | 02:22 | |
| *** stevemar has joined #openstack-infra | 02:23 | |
| *** craigbr has quit IRC | 02:24 | |
| *** tiswanso has quit IRC | 02:28 | |
| *** tiswanso has joined #openstack-infra | 02:29 | |
| *** achanda has joined #openstack-infra | 02:29 | |
| *** craigbr has joined #openstack-infra | 02:31 | |
| *** ddieterly has quit IRC | 02:32 | |
| *** craigbr has quit IRC | 02:37 | |
| *** mpopow has joined #openstack-infra | 02:38 | |
| *** r-daneel has quit IRC | 02:40 | |
| *** annegentle has joined #openstack-infra | 02:40 | |
| *** agentleexile has joined #openstack-infra | 02:41 | |
| *** stevemar has quit IRC | 02:49 | |
| *** agentleexile is now known as annegentle_ | 02:49 | |
| *** harlowja_at_home has joined #openstack-infra | 02:53 | |
| *** salv-orlando has joined #openstack-infra | 02:54 | |
| *** rkukura has joined #openstack-infra | 02:56 | |
| *** Swami_ has joined #openstack-infra | 02:57 | |
| *** amotoki has joined #openstack-infra | 02:58 | |
| *** Swami has quit IRC | 02:59 | |
| *** smallbig has joined #openstack-infra | 03:00 | |
| *** dims has joined #openstack-infra | 03:01 | |
| *** Swami__ has joined #openstack-infra | 03:02 | |
| *** salv-orl_ has joined #openstack-infra | 03:03 | |
| *** salv-orlando has quit IRC | 03:04 | |
| *** tiswanso has quit IRC | 03:05 | |
| *** Swami_ has quit IRC | 03:05 | |
| *** dims has quit IRC | 03:06 | |
| *** salv-orl_ has quit IRC | 03:08 | |
| *** mpopow has quit IRC | 03:13 | |
| *** akshai has joined #openstack-infra | 03:14 | |
| openstackgerrit | Rabi Mishra proposed openstack-infra/project-config: Enable orchestration service for gate-heat-templates-dsvm https://review.openstack.org/181286 | 03:15 |
| *** baoli has quit IRC | 03:16 | |
| *** salv-orlando has joined #openstack-infra | 03:18 | |
| *** akshai has quit IRC | 03:19 | |
| *** achanda has quit IRC | 03:22 | |
| *** salv-orlando has quit IRC | 03:22 | |
| *** r-daneel has joined #openstack-infra | 03:26 | |
| *** ivar-laz_ has joined #openstack-infra | 03:28 | |
| *** ivar-laz_ has quit IRC | 03:29 | |
| *** r-daneel has quit IRC | 03:31 | |
| *** ivar-lazzaro has quit IRC | 03:32 | |
| *** ddieterly has joined #openstack-infra | 03:33 | |
| *** marun has quit IRC | 03:37 | |
| *** ddieterly has quit IRC | 03:38 | |
| *** unicell has joined #openstack-infra | 03:42 | |
| *** rmcall has joined #openstack-infra | 03:42 | |
| *** annegentle has quit IRC | 03:43 | |
| *** annegentle_ has quit IRC | 03:43 | |
| *** annegentle has joined #openstack-infra | 03:44 | |
| *** Swami__ has quit IRC | 03:45 | |
| *** camunoz_afk is now known as camunoz | 03:45 | |
| *** r-daneel has joined #openstack-infra | 03:46 | |
| *** otter768 has quit IRC | 03:48 | |
| *** unicell has quit IRC | 03:48 | |
| *** unicell has joined #openstack-infra | 03:49 | |
| *** shashankhegde has quit IRC | 03:49 | |
| *** Swami has joined #openstack-infra | 03:50 | |
| *** patrickeast has quit IRC | 03:50 | |
| *** bhunter71 has quit IRC | 03:51 | |
| *** Swami has quit IRC | 03:55 | |
| *** craigbr has joined #openstack-infra | 04:00 | |
| *** annegentle has quit IRC | 04:02 | |
| *** craigbr has quit IRC | 04:05 | |
| *** craigbr has joined #openstack-infra | 04:05 | |
| *** craigbr has quit IRC | 04:06 | |
| *** xyang1 has quit IRC | 04:06 | |
| *** samueldmq has quit IRC | 04:11 | |
| *** pal has joined #openstack-infra | 04:11 | |
| *** camunoz is now known as camunoz_mtg | 04:12 | |
| *** sdake has quit IRC | 04:13 | |
| *** mmedvede has quit IRC | 04:14 | |
| *** baoli has joined #openstack-infra | 04:17 | |
| *** Alexandra_ has joined #openstack-infra | 04:17 | |
| *** changbl has joined #openstack-infra | 04:18 | |
| *** rm_work is now known as rm_work|away | 04:19 | |
| *** baoli has quit IRC | 04:22 | |
| *** achanda has joined #openstack-infra | 04:22 | |
| *** ivar-lazzaro has joined #openstack-infra | 04:27 | |
| *** achanda has quit IRC | 04:28 | |
| *** mmedvede has joined #openstack-infra | 04:28 | |
| *** ddieterly has joined #openstack-infra | 04:34 | |
| *** ddieterly has quit IRC | 04:38 | |
| *** achanda has joined #openstack-infra | 04:42 | |
| *** deepakcs has joined #openstack-infra | 04:43 | |
| *** rm_work|away is now known as rm_work | 04:44 | |
| *** julim has quit IRC | 04:49 | |
| *** mgedmin has quit IRC | 04:55 | |
| *** Sukhdev has joined #openstack-infra | 04:56 | |
| *** julim has joined #openstack-infra | 04:56 | |
| openstackgerrit | Ramakrishnan G proposed openstack-infra/project-config: Add python34 job for proliantutils https://review.openstack.org/181104 | 04:57 |
| *** mpopow has joined #openstack-infra | 04:57 | |
| *** mgedmin has joined #openstack-infra | 04:57 | |
| *** gokrokve has joined #openstack-infra | 04:58 | |
| deepakcs | AJaeger, morning, finally the patch to skip tempest test for glusterfs job is merged! :) thanks for all ur help in that | 04:59 |
| *** Longgeek has joined #openstack-infra | 05:00 | |
| *** annegentle has joined #openstack-infra | 05:03 | |
| *** annegentle_ has joined #openstack-infra | 05:03 | |
| *** ildikov has quit IRC | 05:04 | |
| *** emagana has joined #openstack-infra | 05:07 | |
| *** Krinkle is now known as Krinkle|detached | 05:07 | |
| *** mrmartin has joined #openstack-infra | 05:07 | |
| *** annegentle_ has quit IRC | 05:08 | |
| *** annegentle has quit IRC | 05:09 | |
| *** gokrokve has quit IRC | 05:09 | |
| *** gokrokve has joined #openstack-infra | 05:10 | |
| *** gokrokve has quit IRC | 05:13 | |
| *** gokrokve has joined #openstack-infra | 05:13 | |
| *** Somay has joined #openstack-infra | 05:15 | |
| *** ryanpetrello_ is now known as ryanpetrello | 05:15 | |
| *** ildikov has joined #openstack-infra | 05:17 | |
| *** baoli has joined #openstack-infra | 05:17 | |
| *** gokrokve has quit IRC | 05:18 | |
| *** salv-orlando has joined #openstack-infra | 05:19 | |
| *** mrmartin has quit IRC | 05:19 | |
| *** Somay has quit IRC | 05:20 | |
| *** camunoz_mtg is now known as camunoz | 05:21 | |
| *** teran has quit IRC | 05:21 | |
| *** teran has joined #openstack-infra | 05:22 | |
| *** baoli has quit IRC | 05:22 | |
| *** salv-orlando has quit IRC | 05:22 | |
| *** Longgeek has quit IRC | 05:29 | |
| *** Sukhdev has quit IRC | 05:30 | |
| *** ildikov has quit IRC | 05:30 | |
| *** mpopow has quit IRC | 05:31 | |
| *** jyuso1 has quit IRC | 05:32 | |
| *** macjack has joined #openstack-infra | 05:34 | |
| *** ddieterly has joined #openstack-infra | 05:35 | |
| *** sabeen1 has quit IRC | 05:36 | |
| *** ddieterly has quit IRC | 05:39 | |
| *** gokrokve has joined #openstack-infra | 05:40 | |
| *** gokrokve has quit IRC | 05:42 | |
| *** gokrokve has joined #openstack-infra | 05:42 | |
| *** emagana has quit IRC | 05:44 | |
| *** emagana has joined #openstack-infra | 05:45 | |
| *** pal has quit IRC | 05:48 | |
| *** pal has joined #openstack-infra | 05:48 | |
| *** otter768 has joined #openstack-infra | 05:49 | |
| *** emagana has quit IRC | 05:49 | |
| *** rossella_s has joined #openstack-infra | 05:52 | |
| *** otter768 has quit IRC | 05:53 | |
| StevenK | cinerama: Funny, seeing en-AU, en-Au, en-au and en-AU in Zanta. You didn't go far enough, eN-Au for example is missing. | 05:54 |
| *** abregman has joined #openstack-infra | 05:56 | |
| *** ibiris_away is now known as ibiris | 05:56 | |
| *** BharatK has joined #openstack-infra | 05:57 | |
| *** Alexandra_ has quit IRC | 05:58 | |
| AJaeger | StevenK, what about Ne-Au ? ;) | 05:59 |
| AJaeger | StevenK, Typo, I meant: En-uA | 05:59 |
| *** rossella_s has quit IRC | 06:00 | |
| StevenK | AJaeger: Filling out the rest of the missing ones was left as an exercise to the reader. | 06:00 |
| *** rossella_s has joined #openstack-infra | 06:01 | |
| AJaeger | :) | 06:01 |
| *** Kennan has quit IRC | 06:03 | |
| *** Kennan has joined #openstack-infra | 06:03 | |
| *** annegentle has joined #openstack-infra | 06:04 | |
| *** annegentle_ has joined #openstack-infra | 06:05 | |
| *** annegentle_ has quit IRC | 06:09 | |
| *** harlowja_at_home has quit IRC | 06:09 | |
| *** annegentle has quit IRC | 06:10 | |
| *** scheuran has joined #openstack-infra | 06:10 | |
| *** rossella_s has quit IRC | 06:11 | |
| *** salv-orlando has joined #openstack-infra | 06:13 | |
| *** Longgeek has joined #openstack-infra | 06:13 | |
| *** r-daneel has quit IRC | 06:15 | |
| *** mrunge has joined #openstack-infra | 06:16 | |
| openstackgerrit | Timo Tijhof proposed openstack-infra/zuul: Status: Expose $.zuul object in zuul_start to allow extension https://review.openstack.org/170337 | 06:20 |
| *** HeOS has quit IRC | 06:21 | |
| *** abregman is now known as abregman|afk | 06:23 | |
| openstackgerrit | Andreas Jaeger proposed openstack-infra/project-config: Fix django translation proposal https://review.openstack.org/181303 | 06:23 |
| *** tnovacik has joined #openstack-infra | 06:25 | |
| *** jyuso1 has joined #openstack-infra | 06:30 | |
| AJaeger | pleia2, babysitting my changes was needed - see https://review.openstack.org/181303 ;( | 06:31 |
| *** markvoelker has quit IRC | 06:33 | |
| *** ddieterly has joined #openstack-infra | 06:36 | |
| *** soren_ has joined #openstack-infra | 06:37 | |
| *** achanda has quit IRC | 06:38 | |
| *** ddieterly has quit IRC | 06:41 | |
| *** ildikov has joined #openstack-infra | 06:43 | |
| *** ibiris is now known as ibiris_away | 06:45 | |
| *** ujuc has joined #openstack-infra | 06:45 | |
| *** ibiris_away is now known as ibiris | 06:46 | |
| *** abregman|afk has quit IRC | 06:53 | |
| *** camunoz has quit IRC | 07:00 | |
| *** armax has quit IRC | 07:02 | |
| StevenK | cinerama: Bug 4, Zanata not logging people out after 10 minutes of inactivity would be AWESOME | 07:04 |
| *** notnownikki has joined #openstack-infra | 07:05 | |
| *** crc32 has quit IRC | 07:05 | |
| *** annegentle has joined #openstack-infra | 07:05 | |
| *** e0ne has joined #openstack-infra | 07:09 | |
| *** woodster_ has quit IRC | 07:10 | |
| *** annegentle has quit IRC | 07:10 | |
| *** markus_z has joined #openstack-infra | 07:14 | |
| *** mpopow has joined #openstack-infra | 07:16 | |
| *** sks has joined #openstack-infra | 07:16 | |
| *** jyuso1 has quit IRC | 07:17 | |
| StevenK | AJaeger, cinerama, pleia2: http://paste.openstack.org/show/216853/ | 07:21 |
| AJaeger | StevenK, btw. full list of languages is here: https://www.transifex.com/projects/p/horizon/ | 07:22 |
| AJaeger | You might want to download one project with all languages, run a tx pull -a -f... | 07:23 |
| StevenK | AJaeger: I was looking at http://www.transifex.com/languages/ , but I didn't want to add everything ever. | 07:23 |
| AJaeger | StevenK, just the languages for openstack | 07:24 |
| AJaeger | Better link: https://www.transifex.com/projects/p/openstack/ | 07:24 |
| AJaeger | I still don't udnerstand why we have pl and pl-pl | 07:24 |
| StevenK | They're different in Transifex too | 07:24 |
| StevenK | pl is Polish, and pl_PL is Polish as spoken in Poland. | 07:25 |
| StevenK | I don't know why, though. | 07:25 |
| *** luqas has quit IRC | 07:25 | |
| *** mmedvede has quit IRC | 07:25 | |
| *** jyuso1 has joined #openstack-infra | 07:27 | |
| AJaeger | https://www.transifex.com/projects/p/nova/language/pl_PL/ and https://www.transifex.com/projects/p/osloi18n/language/pl_PL/ | 07:27 |
| AJaeger | both pl_PL | 07:27 |
| StevenK | But the language in the nova code tree is pl | 07:27 |
| StevenK | Huh | 07:27 |
| AJaeger | where in nova code tree? http://git.openstack.org/cgit/openstack/nova/tree/nova/locale | 07:29 |
| *** HeOS has joined #openstack-infra | 07:29 | |
| *** e0ne has quit IRC | 07:29 | |
| AJaeger | I removed pl_PL a few months ago, check https://review.openstack.org/#/c/94685/ | 07:30 |
| AJaeger | Deleted nova/locale/pl_PL/LC_MESSAGES/nova.po in the change | 07:30 |
| AJaeger | So, where does your pl comes from? | 07:30 |
| AJaeger | Same might be true for sl. | 07:31 |
| * AJaeger didn't check | 07:31 | |
| *** luqas has joined #openstack-infra | 07:32 | |
| *** e0ne has joined #openstack-infra | 07:33 | |
| *** mmedvede has joined #openstack-infra | 07:35 | |
| *** rmcall has quit IRC | 07:35 | |
| ekarlso | heya guys, i'm trying to ./stack but I get an error saying that cffi can't be found http://paste.openstack.org/show/216858/ ? | 07:36 |
| *** jamielennox is now known as jamielennox|away | 07:36 | |
| *** ddieterly has joined #openstack-infra | 07:36 | |
| *** mpavone has joined #openstack-infra | 07:41 | |
| *** ddieterly has quit IRC | 07:41 | |
| *** tnovacik has quit IRC | 07:43 | |
| *** tnovacik has joined #openstack-infra | 07:44 | |
| *** jlanoux has joined #openstack-infra | 07:45 | |
| lifeless | ekarlso: yeah | 07:45 |
| lifeless | ekarlso: oh, no, I don't recognise anything in that snipper | 07:45 |
| lifeless | ekarlso: but this: grep: /opt/stack/requirements/global-requirements.txt: No such file or directory | 07:46 |
| lifeless | ekarlso: doesn't look good :) | 07:46 |
| *** mpopow has quit IRC | 07:50 | |
| *** otter768 has joined #openstack-infra | 07:50 | |
| *** otter768 has quit IRC | 07:54 | |
| *** Somay has joined #openstack-infra | 07:55 | |
| ekarlso | lifeless: http://review.openstack.org/181312 :p | 07:59 |
| ekarlso | was the fix :) | 07:59 |
| *** Somay has quit IRC | 07:59 | |
| *** ildikov has quit IRC | 08:01 | |
| lifeless | ekarlso: hah, indeed | 08:03 |
| *** annegentle has joined #openstack-infra | 08:06 | |
| *** annegent_ has joined #openstack-infra | 08:06 | |
| *** annegentle is now known as Guest59212 | 08:06 | |
| *** markvoelker has joined #openstack-infra | 08:07 | |
| openstackgerrit | Joseph Lanoux proposed openstack-infra/project-config: Enable SSL in devstack for experimental checks. https://review.openstack.org/180968 | 08:09 |
| *** Guest59212 has quit IRC | 08:10 | |
| *** annegent_ has quit IRC | 08:11 | |
| *** ildikov has joined #openstack-infra | 08:17 | |
| *** chlong has quit IRC | 08:20 | |
| *** sks has quit IRC | 08:21 | |
| *** e0ne has quit IRC | 08:23 | |
| *** freerunner has joined #openstack-infra | 08:24 | |
| *** freerunner has quit IRC | 08:26 | |
| *** dizquierdo has joined #openstack-infra | 08:26 | |
| *** mrmartin has joined #openstack-infra | 08:29 | |
| *** dizquierdo has quit IRC | 08:31 | |
| *** sergmelikyan has quit IRC | 08:35 | |
| *** derekh has joined #openstack-infra | 08:37 | |
| *** ddieterly has joined #openstack-infra | 08:37 | |
| *** rossella_s has joined #openstack-infra | 08:38 | |
| *** markvoelker has quit IRC | 08:38 | |
| *** dalgaaf has joined #openstack-infra | 08:41 | |
| *** mahito_ has quit IRC | 08:41 | |
| *** ddieterly has quit IRC | 08:41 | |
| *** ssam2 has joined #openstack-infra | 08:42 | |
| *** freerunner has joined #openstack-infra | 08:47 | |
| *** cdent has joined #openstack-infra | 08:56 | |
| *** dguitarbite has quit IRC | 08:56 | |
| *** dbite has quit IRC | 08:56 | |
| *** dguitarbite has joined #openstack-infra | 08:57 | |
| *** e0ne has joined #openstack-infra | 08:58 | |
| *** e0ne is now known as e0ne_ | 08:58 | |
| *** ildikov has quit IRC | 09:04 | |
| AJaeger | lifeless, pushed openstack-doc-tools 0.27, you should be able to use it now ;) | 09:05 |
| *** annegentle has joined #openstack-infra | 09:07 | |
| *** e0ne_ is now known as e0ne | 09:07 | |
| *** Somay has joined #openstack-infra | 09:09 | |
| lifeless | AJaeger: cool | 09:09 |
| *** annegentle has quit IRC | 09:12 | |
| AJaeger | lifeless, trove would have been hit by the problem if they would have updated requirements (but they are still old: https://review.openstack.org/#/c/179879) | 09:13 |
| *** Somay has quit IRC | 09:15 | |
| *** _nadya_ has joined #openstack-infra | 09:17 | |
| *** Somay has joined #openstack-infra | 09:17 | |
| *** ildikov has joined #openstack-infra | 09:20 | |
| *** e0ne is now known as e0ne_ | 09:21 | |
| *** freerunner has quit IRC | 09:21 | |
| *** Somay has quit IRC | 09:24 | |
| mgedmin | lifeless, you pinged me last night about some good news fungi ought to maybe share with me, but my znc's buffer overfloweth so I can't see the context | 09:25 |
| mgedmin | this is a high-volume channel | 09:25 |
| AJaeger | mgedmin, this channel is archived, see http://eavesdrop.openstack.org/irclogs/%23openstack-infra/ | 09:26 |
| mgedmin | oh, cool! | 09:26 |
| mgedmin | augh, raw logs | 09:27 |
| mgedmin | let me tell you about https://pypi.python.org/pypi/irclog2html | 09:27 |
| mgedmin | (and also apologize for being a bad maintainer for it who doesn't make proper docs) | 09:27 |
| *** _nadya_ has quit IRC | 09:27 | |
| AJaeger | mgedmin, I didn't set eavesdrop up ;) If you like to enhance eavesdrop with that tool, go for it! | 09:28 |
| mgedmin | cool, I'm gonna need ssh access ;) | 09:28 |
| AJaeger | mgedmin, nope! | 09:29 |
| mgedmin | haha | 09:29 |
| AJaeger | mgedmin, our infrastructure is completely administread like code - we use git repositories and scripts for deployment | 09:29 |
| mgedmin | but seriously, how would I go about enhancing eavesdrop? | 09:29 |
| AJaeger | So, anybody can change it | 09:29 |
| mgedmin | i.e. what's the infrastructure repo and what config management tool do you use? | 09:30 |
| AJaeger | puppet is used for configuration management | 09:30 |
| AJaeger | the openstack-infra/system-config repo has an eavesdrop file but I don't know all the details. | 09:31 |
| AJaeger | mgedmin, if you're interested, please come back later (US morning) and talk with the infra roots like jeblair, clarkb, fungi and pleia2 | 09:31 |
| mgedmin | thanks for the pointers! | 09:31 |
| AJaeger | you're welcome, mgedmin | 09:32 |
| lifeless | mgedmin: bindep on pypi | 09:33 |
| mgedmin | yay! | 09:34 |
| *** viktors|afk is now known as viktors | 09:34 | |
| *** mpopow has joined #openstack-infra | 09:34 | |
| *** abregman|afk has joined #openstack-infra | 09:36 | |
| *** ddieterly has joined #openstack-infra | 09:38 | |
| *** abregman|afk is now known as abregman | 09:40 | |
| *** abregman has quit IRC | 09:40 | |
| *** mrmartin has quit IRC | 09:40 | |
| *** abregman has joined #openstack-infra | 09:40 | |
| openstackgerrit | Derek Higgins proposed openstack-infra/tripleo-ci: temprevert : Remove passing Flavor's deploy_{kernel, ramdisk} https://review.openstack.org/181341 | 09:40 |
| *** ddieterly has quit IRC | 09:42 | |
| *** pelix has joined #openstack-infra | 09:44 | |
| openstackgerrit | Mikhail Dubov proposed openstack-infra/project-config: Add Rally job to Murano https://review.openstack.org/181343 | 09:44 |
| *** devvesa has joined #openstack-infra | 09:45 | |
| openstackgerrit | Mikhail Dubov proposed openstack-infra/project-config: Add Rally job to Murano https://review.openstack.org/181343 | 09:45 |
| *** alexpilotti has joined #openstack-infra | 09:48 | |
| *** hichihara has quit IRC | 09:48 | |
| *** teran has quit IRC | 09:48 | |
| *** ibiris is now known as ibiris_away | 09:51 | |
| *** otter768 has joined #openstack-infra | 09:51 | |
| *** ociuhandu has joined #openstack-infra | 09:54 | |
| *** e0ne_ is now known as e0ne | 09:55 | |
| *** otter768 has quit IRC | 09:55 | |
| openstackgerrit | Merged openstack-infra/project-config: Make tempest-dsvm-sahara job voting https://review.openstack.org/181210 | 09:58 |
| *** rameshg87 has joined #openstack-infra | 09:59 | |
| *** freerunner has joined #openstack-infra | 10:01 | |
| *** coolsvap|afk is now known as coolsvap | 10:03 | |
| *** ujuc has quit IRC | 10:06 | |
| openstackgerrit | Ramakrishnan G proposed openstack-infra/project-config: Add python34 job for proliantutils https://review.openstack.org/181104 | 10:06 |
| *** annegentle has joined #openstack-infra | 10:07 | |
| *** annegent_ has joined #openstack-infra | 10:08 | |
| *** annegentle is now known as Guest76235 | 10:08 | |
| *** mpopow has quit IRC | 10:08 | |
| *** sslypushenko has quit IRC | 10:11 | |
| *** woodster_ has joined #openstack-infra | 10:11 | |
| *** Guest76235 has quit IRC | 10:12 | |
| *** zhiwei has quit IRC | 10:12 | |
| *** annegent_ has quit IRC | 10:13 | |
| *** sslypush1nko has joined #openstack-infra | 10:16 | |
| *** sslypush1nko is now known as slypushenko_ | 10:17 | |
| *** pc_m has joined #openstack-infra | 10:20 | |
| *** yfried__ is now known as yfried|afk | 10:22 | |
| *** yfried|afk is now known as yfried__ | 10:26 | |
| *** ibiris_away is now known as ibiris | 10:27 | |
| *** freerunner has quit IRC | 10:30 | |
| *** rlandy has joined #openstack-infra | 10:31 | |
| *** freerunner has joined #openstack-infra | 10:33 | |
| *** hdd has joined #openstack-infra | 10:34 | |
| *** rameshg87 has left #openstack-infra | 10:34 | |
| *** markvoelker has joined #openstack-infra | 10:35 | |
| *** freerunner has quit IRC | 10:37 | |
| *** ddieterly has joined #openstack-infra | 10:39 | |
| *** markvoelker has quit IRC | 10:40 | |
| *** fawadkhaliq has joined #openstack-infra | 10:41 | |
| *** teran has joined #openstack-infra | 10:41 | |
| *** mrmartin has joined #openstack-infra | 10:43 | |
| *** ddieterly has quit IRC | 10:44 | |
| *** hdd has quit IRC | 10:45 | |
| *** baoli has joined #openstack-infra | 10:49 | |
| *** samueldmq has joined #openstack-infra | 10:49 | |
| samueldmq | morning | 10:49 |
| *** baoli has quit IRC | 10:53 | |
| *** weshay has joined #openstack-infra | 10:55 | |
| *** e0ne is now known as e0ne_ | 10:56 | |
| *** yfried__ is now known as yfried|afk | 10:56 | |
| *** koolhead17 has joined #openstack-infra | 10:59 | |
| *** moshele has joined #openstack-infra | 11:01 | |
| moshele | AJaeger: hi | 11:01 |
| *** ihrachyshka has joined #openstack-infra | 11:05 | |
| *** mwagner_lap has joined #openstack-infra | 11:06 | |
| *** freerunner has joined #openstack-infra | 11:07 | |
| openstackgerrit | Louis Taylor proposed openstack-infra/project-config: Add extra padding to some entries in layout.yaml https://review.openstack.org/181000 | 11:08 |
| *** annegentle has joined #openstack-infra | 11:08 | |
| *** annegent_ has joined #openstack-infra | 11:08 | |
| *** annegentle is now known as Guest46977 | 11:09 | |
| *** rfolco has joined #openstack-infra | 11:09 | |
| *** ihrachyshka has quit IRC | 11:11 | |
| openstackgerrit | Derek Higgins proposed openstack-infra/tripleo-ci: temprevert : Remove passing Flavor's deploy_{kernel, ramdisk} https://review.openstack.org/181341 | 11:12 |
| *** Guest46977 has quit IRC | 11:13 | |
| *** annegent_ has quit IRC | 11:13 | |
| *** dims has joined #openstack-infra | 11:13 | |
| *** e0ne_ is now known as e0ne | 11:14 | |
| *** moshele has quit IRC | 11:15 | |
| *** koolhead17 has quit IRC | 11:16 | |
| *** ihrachyshka has joined #openstack-infra | 11:16 | |
| *** koolhead17 has joined #openstack-infra | 11:16 | |
| *** yfried|afk is now known as yfried__ | 11:17 | |
| *** pal has quit IRC | 11:19 | |
| *** koolhead17 has quit IRC | 11:21 | |
| StevenK | AJaeger: Sorry, distracted by driving home, and other stuff. Huh, I see a lack of pl under nova, I shall have to debug my script it seems. | 11:22 |
| openstackgerrit | Merged openstack-infra/tripleo-ci: temprevert : Remove passing Flavor's deploy_{kernel, ramdisk} https://review.openstack.org/181341 | 11:22 |
| StevenK | Oh, from the .venv I have for sphnix. Nice. Not. | 11:22 |
| openstackgerrit | yolanda.robla proposed openstack-infra/jenkins-job-builder: Query all jobs once https://review.openstack.org/181364 | 11:24 |
| openstackgerrit | yolanda.robla proposed openstack-infra/jenkins-job-builder: Query all jobs once https://review.openstack.org/181364 | 11:26 |
| *** yfried__ is now known as yfried|afk | 11:27 | |
| *** dims has quit IRC | 11:28 | |
| *** e0ne is now known as e0ne_ | 11:28 | |
| *** dims has joined #openstack-infra | 11:29 | |
| *** e0ne_ is now known as e0ne | 11:30 | |
| *** fawadkhaliq has quit IRC | 11:31 | |
| *** _nadya_ has joined #openstack-infra | 11:31 | |
| *** ldnunes has joined #openstack-infra | 11:33 | |
| *** fawadkhaliq has joined #openstack-infra | 11:34 | |
| *** smccully has quit IRC | 11:36 | |
| *** sputnik13 has quit IRC | 11:36 | |
| *** markvoelker has joined #openstack-infra | 11:36 | |
| *** ddieterly has joined #openstack-infra | 11:39 | |
| *** hdd has joined #openstack-infra | 11:40 | |
| *** aswadr has joined #openstack-infra | 11:42 | |
| *** ddieterly has quit IRC | 11:44 | |
| *** r-daneel has joined #openstack-infra | 11:50 | |
| *** otter768 has joined #openstack-infra | 11:52 | |
| *** mpopow has joined #openstack-infra | 11:52 | |
| *** deepakcs has quit IRC | 11:53 | |
| mordred | mgedmin: hey! our logs are produced by meetbot, which is in openstack-infra/meetbot, and is configured with openstack-infra/puppet-meetbot | 11:55 |
| mordred | it looks like there is an HTML log config output parameter that we're not making use of | 11:56 |
| *** r-daneel has quit IRC | 11:56 | |
| *** otter768 has quit IRC | 11:56 | |
| mgedmin | I'm not familiar with meetbot | 11:56 |
| *** freerunner has quit IRC | 11:56 | |
| mgedmin | I've used supybot's channel logger with a cron script that runs irclog2html for my own logs | 11:56 |
| openstackgerrit | yolanda.robla proposed openstack-infra/jenkins-job-builder: Query all jobs once https://review.openstack.org/181364 | 11:57 |
| mordred | it also handles our irc meetings- I would imagine that it would not be hard to write an irclog2html output plugin for meetbot - I'm going to try a patch to enable the built in html output and we can see if we think the extra effort would be worth it | 11:58 |
| *** amotoki has quit IRC | 11:58 | |
| *** annegentle has joined #openstack-infra | 12:00 | |
| *** hdd has quit IRC | 12:01 | |
| *** fawadk has joined #openstack-infra | 12:01 | |
| mordred | hrm. I may be lying | 12:01 |
| mgedmin | the log files I saw should be usable with irclog2html | 12:02 |
| mgedmin | which is a standalone script that finds *.log and creates *.html, checking mtimes to avoid duplicate work | 12:02 |
| *** fawadkhaliq has quit IRC | 12:02 | |
| mordred | yah. I just meant that I was lying about it being meetbot that wrote the logs :) | 12:02 |
| *** freerunner has joined #openstack-infra | 12:02 | |
| mgedmin | meetbot is a supybot plugin, isn't it? | 12:02 |
| mordred | it is - but it turns out it's statusbot that does the channel logs | 12:03 |
| Daviey | meetbot is for doing agenda management and vote tallying, no? | 12:03 |
| *** yfried|afk is now known as yfried__ | 12:03 | |
| mordred | yah | 12:04 |
| mordred | I was thinking we'd combined them | 12:04 |
| mordred | but I was teh wrong | 12:04 |
| * mordred is still coffeeing | 12:04 | |
| mgedmin | yeah, supybot has a different plugin for logging (ChannelLogger) | 12:04 |
| mordred | mgedmin: how does the irclog2html script deal with logs being currently written? | 12:04 |
| mgedmin | you might get a truncated log line, I suppose, until the next cron run | 12:05 |
| mordred | ah - so it would re-convert it becase of the mtime | 12:05 |
| *** fawadk has quit IRC | 12:06 | |
| *** markvoelker has quit IRC | 12:07 | |
| *** sdake has joined #openstack-infra | 12:07 | |
| *** markvoelker has joined #openstack-infra | 12:07 | |
| *** dalgaaf has quit IRC | 12:07 | |
| mgedmin | sample output: https://mg.pov.lt/maemo-irclog/ | 12:08 |
| mgedmin | (it is terrible, I need to do something about grouping by month or something) | 12:08 |
| mgedmin | -ENOTIME | 12:08 |
| mgedmin | -ETOOMANYPROJECTS | 12:08 |
| *** koolhead17 has joined #openstack-infra | 12:08 | |
| *** koolhead17 has quit IRC | 12:08 | |
| *** koolhead17 has joined #openstack-infra | 12:08 | |
| mordred | k. I ran it on one file: http://eavesdrop.openstack.org/irclogs/%23openstack-infra/%23openstack-infra.2015-05-08.log.html | 12:08 |
| *** e0ne is now known as e0ne_ | 12:10 | |
| *** sdake_ has joined #openstack-infra | 12:10 | |
| *** derekh is now known as derekh_afk | 12:11 | |
| *** e0ne_ is now known as e0ne | 12:13 | |
| *** sdake has quit IRC | 12:14 | |
| openstackgerrit | Steven Dake proposed openstack-infra/project-config: Add CloudPulse https://review.openstack.org/181125 | 12:15 |
| *** claudiub has joined #openstack-infra | 12:15 | |
| *** nithyag_ has joined #openstack-infra | 12:16 | |
| *** NithyaG has quit IRC | 12:18 | |
| *** fawadkhaliq has joined #openstack-infra | 12:19 | |
| *** freerunner has quit IRC | 12:19 | |
| *** baoli has joined #openstack-infra | 12:19 | |
| *** mcdonagm has joined #openstack-infra | 12:20 | |
| *** mtanino has joined #openstack-infra | 12:20 | |
| *** macjack has left #openstack-infra | 12:22 | |
| *** dprince has joined #openstack-infra | 12:22 | |
| *** bhunter71 has joined #openstack-infra | 12:23 | |
| *** abregman has quit IRC | 12:24 | |
| *** baoli has quit IRC | 12:24 | |
| openstackgerrit | Sean Dague proposed openstack-infra/project-config: add stable branch reporting to qa channel https://review.openstack.org/181375 | 12:24 |
| *** mpopow has quit IRC | 12:26 | |
| *** gordc has joined #openstack-infra | 12:29 | |
| *** ihrachyshka has quit IRC | 12:32 | |
| *** craigbr has joined #openstack-infra | 12:33 | |
| openstackgerrit | Merged openstack-infra/irc-meetings: Update Rally meeting. https://review.openstack.org/181257 | 12:34 |
| *** ZZelle has joined #openstack-infra | 12:37 | |
| openstackgerrit | Monty Taylor proposed openstack-infra/puppet-meetbot: Produce html logs for our IRC logs https://review.openstack.org/181377 | 12:37 |
| mordred | mgedmin: ^^ there ya go | 12:37 |
| openstackgerrit | David Shrewsbury proposed openstack-infra/shade: Update server API for get/list/search interface https://review.openstack.org/180293 | 12:39 |
| *** craigbr has quit IRC | 12:40 | |
| *** aysyd has joined #openstack-infra | 12:40 | |
| *** dkranz has joined #openstack-infra | 12:40 | |
| *** ddieterly has joined #openstack-infra | 12:40 | |
| *** bknudson has quit IRC | 12:43 | |
| *** aswadr has quit IRC | 12:43 | |
| mgedmin | yay! | 12:43 |
| *** sdake has joined #openstack-infra | 12:44 | |
| openstackgerrit | Merged openstack-dev/hacking: Updated from global requirements https://review.openstack.org/181196 | 12:44 |
| mgedmin | this gives me an idea: incremental mode (limit amount of work done per run to, say, 30 logfiles? start with the most recently modified, so the backlog gets filled over time) | 12:44 |
| mgedmin | -ENOTIME | 12:44 |
| *** ddieterly has quit IRC | 12:45 | |
| openstackgerrit | David Shrewsbury proposed openstack-infra/shade: Update images API for get/list/search interface https://review.openstack.org/180600 | 12:45 |
| *** chlong has joined #openstack-infra | 12:47 | |
| *** sdake_ has quit IRC | 12:47 | |
| *** ddieterly has joined #openstack-infra | 12:48 | |
| *** _nadya_ has quit IRC | 12:50 | |
| openstackgerrit | Merged openstack-infra/system-config: Update rsyslog config after install https://review.openstack.org/181256 | 12:51 |
| *** baoli has joined #openstack-infra | 12:53 | |
| *** e0ne is now known as e0ne_ | 12:56 | |
| *** krtaylor has quit IRC | 12:56 | |
| fungi | mgedmin: or someone (me?) logs into eavesdrop.o.o and runs that at the command line for the initial catch-up, so that we don't need to overengineer it | 12:56 |
| *** e0ne_ is now known as e0ne | 12:57 | |
| *** baoli has quit IRC | 12:57 | |
| *** baoli has joined #openstack-infra | 12:58 | |
| *** derekh_afk is now known as derekh | 13:00 | |
| *** annegentle has quit IRC | 13:00 | |
| *** hdd has joined #openstack-infra | 13:01 | |
| *** freerunner has joined #openstack-infra | 13:01 | |
| *** dkranz has quit IRC | 13:01 | |
| *** annegentle has joined #openstack-infra | 13:02 | |
| *** annegentle_ has joined #openstack-infra | 13:02 | |
| *** AJaeger_ has joined #openstack-infra | 13:02 | |
| *** hdd has quit IRC | 13:03 | |
| AJaeger_ | fungi, mordred, pleia2: Could you review https://review.openstack.org/#/c/181303/ , please? django translation update failed this morning since the changes need one more initializion. I'm sorry for another problem ;( | 13:03 |
| *** ociuhandu has quit IRC | 13:05 | |
| *** krtaylor has joined #openstack-infra | 13:07 | |
| *** bknudson has joined #openstack-infra | 13:11 | |
| sdague | fungi: could you +A - https://review.openstack.org/#/c/181375/ - it's an irc event update for -qa | 13:11 |
| *** annegentle_ has quit IRC | 13:12 | |
| *** fawadk has joined #openstack-infra | 13:12 | |
| openstackgerrit | Merged openstack-infra/puppet-zuul: Correct the variable for the chain cert https://review.openstack.org/179935 | 13:12 |
| *** fawadkhaliq has quit IRC | 13:12 | |
| *** cdent has quit IRC | 13:12 | |
| *** tiswanso has joined #openstack-infra | 13:14 | |
| *** fawadk is now known as fawadkhaliq | 13:15 | |
| annegentle | jeblair: anteaya: thanks for the pointers last night for voting | 13:15 |
| annegentle | sdague: which launchpad project is the URL holder for cross-project specs? | 13:15 |
| *** EmilienM|afk is now known as EmilienM | 13:15 | |
| annegentle | sdague: also, the Service Catalog spec is a cross-project spec, right? | 13:15 |
| sdague | annegentle: I don't know on the launchpad front | 13:16 |
| sdague | I would think service catalog should be cross project spect | 13:16 |
| annegentle | sdague: ok, the template has this unspecified project URL: https://blueprints.launchpad.net//+spec/example | 13:16 |
| annegentle | (the cross project spec template) | 13:16 |
| annegentle | so I'm like "huh?" | 13:16 |
| sdague | yeh, I think that might just be copy/paste issues that never were cleaned up | 13:17 |
| annegentle | sdague: ok. I | 13:18 |
| annegentle | er | 13:18 |
| *** abregman has joined #openstack-infra | 13:18 | |
| annegentle | I'll ask if it could actually be sourced in keystone? does that seem reasonable? | 13:19 |
| mordred | AJaeger_: +2 | 13:19 |
| *** dims has quit IRC | 13:19 | |
| *** ddieterly has quit IRC | 13:20 | |
| *** dims has joined #openstack-infra | 13:20 | |
| *** dguitarbite has quit IRC | 13:21 | |
| sdague | annegentle: works for me | 13:21 |
| AJaeger_ | thanks, mordred | 13:21 |
| *** yfried__ is now known as yfried|afk | 13:22 | |
| *** dguitarbite has joined #openstack-infra | 13:23 | |
| *** pabelanger has joined #openstack-infra | 13:25 | |
| AJaeger_ | fungi, thanks! | 13:25 |
| *** yfried|afk is now known as yfried__ | 13:25 | |
| fungi | today's coffee-time fun involves figuring out why centos-6 yum install --downloadonly seems to error out on things depending on python-setuptools | 13:26 |
| *** peristeri has joined #openstack-infra | 13:27 | |
| annegentle | sdague: one more question, I didn't find Nova's *_host config variables in the nova.conf configs, are those somewhere else? | 13:28 |
| annegentle | sdague: you mention them in the cross-project proposal so I wanted to see what those are | 13:28 |
| openstackgerrit | yolanda.robla proposed openstack-infra/jenkins-job-builder: Query all jobs once https://review.openstack.org/181364 | 13:30 |
| mordred | fungi: I can run it if we decide we want to do it | 13:32 |
| mordred | fungi: (I've got the command loaded on the box since I used a version of it with "echo" to test the command :) ) | 13:32 |
| *** sabeen1 has joined #openstack-infra | 13:32 | |
| openstackgerrit | Merged openstack-infra/project-config: add stable branch reporting to qa channel https://review.openstack.org/181375 | 13:33 |
| *** dguitarbite has quit IRC | 13:33 | |
| *** dansmith is now known as superdan | 13:35 | |
| openstackgerrit | Merged openstack-infra/nodepool: Support spaces in Gearman functions names https://review.openstack.org/175791 | 13:35 |
| openstackgerrit | Merged openstack-infra/project-config: Fix django translation proposal https://review.openstack.org/181303 | 13:35 |
| openstackgerrit | Merged openstack-infra/project-config: Fix django translation proposal https://review.openstack.org/181303 | 13:35 |
| *** coolsvap is now known as coolsvap|afk | 13:37 | |
| *** mriedem_away is now known as mriedem | 13:37 | |
| sdague | annegentle: .... which things? | 13:37 |
| sdague | sorry, so many contexts, I lose some of them | 13:37 |
| annegentle | sdague: the full line is "Such a diverse Service Catalog means that services don't depend on it being correct (see Nova's *_host config variables)" | 13:37 |
| annegentle | sdague: sure understand! | 13:38 |
| annegentle | sdague: I do see cases where cinder's cinder.conf ensures the endpoints sanity | 13:38 |
| annegentle | sdague: there's endpoint_template nova.conf variable that will override service catalog lookup with template for cinder endpoint e.g. http://localhost:8776/v1/%(project_id)s | 13:38 |
| annegentle | sdague: but I wondered if there are some other configs going on | 13:40 |
| *** esker has joined #openstack-infra | 13:40 | |
| *** esker has quit IRC | 13:41 | |
| openstackgerrit | Merged openstack-infra/system-config: Avoid race on reload_gerrit_header https://review.openstack.org/180617 | 13:41 |
| *** freerunner has quit IRC | 13:41 | |
| *** esker has joined #openstack-infra | 13:41 | |
| openstackgerrit | Merged openstack-infra/system-config: Increase JJB timeout to 20min on Jenkins masters https://review.openstack.org/180753 | 13:41 |
| *** esker has quit IRC | 13:41 | |
| *** esker has joined #openstack-infra | 13:41 | |
| *** mtanino has quit IRC | 13:42 | |
| sdague | annegentle: oh, right | 13:43 |
| sdague | CONF.glance.host | 13:44 |
| *** erlon has joined #openstack-infra | 13:44 | |
| sdague | for instance | 13:44 |
| *** annegentle has quit IRC | 13:44 | |
| *** BharatK has quit IRC | 13:44 | |
| *** ociuhandu has joined #openstack-infra | 13:45 | |
| sdague | CONF.neutron.url | 13:45 |
| *** annegentle has joined #openstack-infra | 13:45 | |
| sdague | annegentle: like this is what we have to set on a gate run for it to work - http://logs.openstack.org/74/181374/1/check/check-tempest-dsvm-neutron-full/3289917/logs/etc/nova/nova.conf.txt.gz | 13:46 |
| annegentle | sdague: so that means "in nova.conf, in the [glance] section, host= | 13:46 |
| *** unicell1 has joined #openstack-infra | 13:46 | |
| sdague | you will notice a bunch of hard coded urls to services | 13:46 |
| *** cdent has joined #openstack-infra | 13:46 | |
| *** markus_z has quit IRC | 13:47 | |
| mordred | annegentle: where is service catalog spec? I'm not seeing it in openstack-specs? | 13:47 |
| sdague | I believe we should only need 1 keystone one | 13:47 |
| annegentle | an nova.conf, [glance], api_servers=<IP> | 13:47 |
| sdague | mordred: it's not up yet | 13:47 |
| annegentle | mordred: on my hard drive :) | 13:47 |
| annegentle | mordred: ready to submit to openstack-specs, just ironing out details now | 13:47 |
| annegentle | mordred: and I'll add you as a reviewer | 13:47 |
| *** coolsvap has joined #openstack-infra | 13:47 | |
| annegentle | sdague: so the ideal is, one setting that keystone gives you | 13:48 |
| *** unicell has quit IRC | 13:48 | |
| *** ddieterly has joined #openstack-infra | 13:48 | |
| sdague | so, it makes sense that you need someone to tell you where keystone is... but after that all service urls should be discoverable from there, so we don't keep hardcoding it all over the place. | 13:49 |
| *** eharney has joined #openstack-infra | 13:49 | |
| annegentle | sdague: yep | 13:49 |
| *** arielb has joined #openstack-infra | 13:50 | |
| *** arielb has joined #openstack-infra | 13:50 | |
| *** openstackgerrit has quit IRC | 13:51 | |
| *** openstackgerrit has joined #openstack-infra | 13:51 | |
| *** dims is now known as dimsum__ | 13:51 | |
| openstackgerrit | Sergey Skripnick proposed openstack-infra/project-config: Explicitly enable all required services in rally https://review.openstack.org/181391 | 13:52 |
| *** signed8bit has joined #openstack-infra | 13:52 | |
| *** AJaeger_ has quit IRC | 13:52 | |
| *** otter768 has joined #openstack-infra | 13:52 | |
| *** ociuhandu has quit IRC | 13:53 | |
| *** jpeeler has joined #openstack-infra | 13:53 | |
| mordred | annegentle: ah. ok. so it's fine that I haven't seen it yet then :) | 13:54 |
| mordred | sdague: ++ | 13:54 |
| mordred | sdague, annegentle: I 100% agree with this theory | 13:55 |
| *** samueldmq has quit IRC | 13:57 | |
| *** otter768 has quit IRC | 13:57 | |
| *** smccully has joined #openstack-infra | 13:59 | |
| annegentle | mordred: sdague: https://review.openstack.org/#/c/181393/ | 14:00 |
| *** soren_ has quit IRC | 14:00 | |
| *** smccully has quit IRC | 14:01 | |
| *** BharatK has joined #openstack-infra | 14:01 | |
| *** fawadkhaliq has quit IRC | 14:02 | |
| *** yfried__ is now known as yfried|afk | 14:03 | |
| *** smccully has joined #openstack-infra | 14:03 | |
| *** HeOS has quit IRC | 14:04 | |
| *** ildikov has quit IRC | 14:05 | |
| *** smccully has quit IRC | 14:05 | |
| *** smccully has joined #openstack-infra | 14:05 | |
| AJaeger | annegentle, this is needed, thanks! | 14:05 |
| annegentle | whew! AJaeger :) | 14:06 |
| *** smccully has quit IRC | 14:07 | |
| *** smccully has joined #openstack-infra | 14:07 | |
| mordred | annegentle: reviewed - let the quibbling begin | 14:07 |
| *** BharatK has quit IRC | 14:07 | |
| AJaeger | wow, mordred wants to marry a *spec* ;) | 14:08 |
| mordred | yup | 14:08 |
| *** baoli has quit IRC | 14:08 | |
| *** dizquierdo has joined #openstack-infra | 14:09 | |
| *** baoli has joined #openstack-infra | 14:09 | |
| mordred | our absolutely blantant disregard of supporting version discoverability through the catalog is one of my biggest beefs with openstack | 14:09 |
| fungi | hopefully nolafleur is on board with this plan | 14:09 |
| AJaeger | Let's get the papers together for the ceremony in Vancouver ;) | 14:09 |
| *** ramishra has joined #openstack-infra | 14:09 | |
| mordred | what's even worse is that the services all provide version discoverability | 14:09 |
| mordred | fungi: :) | 14:09 |
| *** fawadkhaliq has joined #openstack-infra | 14:09 | |
| mordred | but keystone does not provide any url pointers to the endpoint for that, and you can't back-infer it from the service url, since keystone returns versioned service urls | 14:10 |
| *** mpopow has joined #openstack-infra | 14:10 | |
| *** AJaeger has quit IRC | 14:11 | |
| ramishra | Hi! need some core reviews for https://review.openstack.org/#/c/181286/. We're having continuous gate failures due to this. | 14:11 |
| *** isviridov_away is now known as isviridov | 14:12 | |
| sdague | ramishra: +2 | 14:12 |
| sdague | mordred / fungi probably should have another +A on https://review.openstack.org/#/c/181286/ | 14:13 |
| ramishra | thanks sdague | 14:13 |
| sdague | oh, mordred just snagged it | 14:13 |
| sdague | great | 14:13 |
| *** sigmavirus24_awa is now known as sigmavirus24 | 14:13 | |
| *** bnemec has joined #openstack-infra | 14:13 | |
| *** isviridov is now known as isviridov_away | 14:13 | |
| ramishra | thanks all | 14:13 |
| *** radez_g0n3 is now known as radez | 14:14 | |
| *** dkranz has joined #openstack-infra | 14:14 | |
| *** annegentle has quit IRC | 14:16 | |
| *** smarcet has joined #openstack-infra | 14:18 | |
| *** BharatK has joined #openstack-infra | 14:19 | |
| *** ddieterly has quit IRC | 14:19 | |
| *** baoli has quit IRC | 14:20 | |
| *** ildikov has joined #openstack-infra | 14:21 | |
| *** gokrokve_ has joined #openstack-infra | 14:21 | |
| *** baoli has joined #openstack-infra | 14:21 | |
| *** ddieterly has joined #openstack-infra | 14:23 | |
| *** gokrokve has quit IRC | 14:25 | |
| *** baoli_ has joined #openstack-infra | 14:25 | |
| *** apmelton_away is now known as apmelton | 14:25 | |
| jroll | anyone know if this is a known issue? http://paste.openstack.org/show/217145/ | 14:25 |
| *** hdd has joined #openstack-infra | 14:25 | |
| openstackgerrit | Merged openstack-infra/project-config: Enable orchestration service for gate-heat-templates-dsvm https://review.openstack.org/181286 | 14:26 |
| *** armax has joined #openstack-infra | 14:26 | |
| fungi | jroll: there were recent fixes into devstack to preinstall/pin a specific cffi version, working around some setup_requires madness in xattr | 14:27 |
| fungi | jroll: could be a problem brought on by one of those fixes? | 14:27 |
| fungi | jroll: are you using a current openstack/requirements checkout there? | 14:28 |
| *** baoli has quit IRC | 14:28 | |
| jroll | fungi: maybe, no clue | 14:28 |
| openstackgerrit | Merged openstack-infra/zuul: Status: Expose $.zuul object in zuul_start to allow extension https://review.openstack.org/170337 | 14:28 |
| jroll | I assume so, this is a fresh vm | 14:28 |
| jroll | I provisioned it 20 minutes ago | 14:28 |
| fungi | grep: /opt/stack/requirements/global-requirements.txt: No such file or directory | 14:29 |
| fungi | that doesn't jive with http://git.openstack.org/cgit/openstack/requirements/tree/global-requirements.txt#n13 | 14:29 |
| mordred | grep: /opt/stack/requirements/global-requirements.txt: No such file or directory | 14:29 |
| mordred | yah | 14:29 |
| *** fawadkhaliq has quit IRC | 14:29 | |
| mordred | fungi: jinx | 14:29 |
| fungi | ahh, looks like maybe global requirements isn't being cloned into your opt/stack at all | 14:29 |
| jroll | weird | 14:29 |
| jroll | yeah, I just ran devstack with an ironic config | 14:29 |
| * jroll looks around | 14:29 | |
| fungi | er, openstack/requirements isn't | 14:30 |
| jroll | nothing's been cloned yet | 14:30 |
| fungi | i wonder if ironic's failing upstream on a similar error | 14:30 |
| sdague | oh, damn, do we have a sequencing problem? | 14:30 |
| jroll | aha, https://bugs.launchpad.net/devstack/+bug/1453012 | 14:30 |
| openstack | Launchpad bug 1453012 in devstack "Don't exists requirements before executing build_wheels.sh" [Undecided,In progress] - Assigned to Yuki Nishiwaki (uckey-1067) | 14:30 |
| * jroll applies that patch | 14:31 | |
| sdague | yep, ok +A on that | 14:31 |
| *** gokrokve_ has quit IRC | 14:31 | |
| openstackgerrit | Merged openstack-infra/project-config: Only import files if there are new translations https://review.openstack.org/179587 | 14:31 |
| jroll | sdague: nice, ty | 14:31 |
| *** gokrokve has joined #openstack-infra | 14:31 | |
| *** bnemec is now known as beekneemech | 14:32 | |
| *** emagana has joined #openstack-infra | 14:32 | |
| sdague | oh, hey, backport through gerrit web ui works ok | 14:33 |
| *** Krinkle|detached is now known as Krinkle | 14:33 | |
| *** koolhead17 has quit IRC | 14:35 | |
| *** fawadkhaliq has joined #openstack-infra | 14:35 | |
| *** koolhead17 has joined #openstack-infra | 14:36 | |
| mordred | sdague: yah - it's actually pretty great | 14:37 |
| peristeri | jeblair, Hi | 14:37 |
| openstackgerrit | Sergey Skripnick proposed openstack-infra/project-config: Explicitly enable all required services in rally https://review.openstack.org/181391 | 14:39 |
| *** wenlock has joined #openstack-infra | 14:39 | |
| *** samuelBartel has joined #openstack-infra | 14:40 | |
| *** koolhead17 has quit IRC | 14:40 | |
| *** wenlock1 has joined #openstack-infra | 14:41 | |
| *** Guest58709 has joined #openstack-infra | 14:41 | |
| samuelBartel | hi all | 14:41 |
| samuelBartel | fuel-plugin-tls-core group has been created : https://review.openstack.org/178995 | 14:42 |
| samuelBartel | can someone add me to the group please? | 14:42 |
| samuelBartel | my userbame is sbartel. thanks | 14:42 |
| *** freerunner has joined #openstack-infra | 14:43 | |
| fungi | samuelBartel: done | 14:44 |
| samuelBartel | great than you fungi | 14:44 |
| fungi | update: i have nova py26 passing and running the same number of tests on centos-6 now as bare-centos6 | 14:45 |
| *** mpopow has quit IRC | 14:45 | |
| fungi | and i think i'm one patch away from having dib creating working centos-6 images now too. just need to, er, write it ;) | 14:45 |
| *** packet has joined #openstack-infra | 14:45 | |
| *** Krinkle is now known as Krinkle|detached | 14:46 | |
| *** craigbr has joined #openstack-infra | 14:47 | |
| mordred | fungi, clarkb: ZOMG - I'm reviewing a patch from EmilienM and it just made me learn about ensure_resource - which apparently deals with our "if not defined blah { blah}" pattern in puppet | 14:49 |
| mordred | 178026 | 14:49 |
| fungi | full installation (this is with incomplete caching still) of the bindep fallback list and db prep for mysql and postgres on a fresh centos-6 worker is 1 minute 40 seconds wall clock time | 14:49 |
| EmilienM | mordred: ensure_resource is from stdlib, it helps to deal with duplicated resources | 14:50 |
| *** hemnafk is now known as hemna | 14:50 | |
| *** ashleighfarnham has joined #openstack-infra | 14:50 | |
| fungi | EmilienM: mordred: that's awesome--it could simplify a lot of our duplicate package resource definitions | 14:50 |
| EmilienM | https://github.com/puppetlabs/puppetlabs-stdlib#ensure_resource | 14:50 |
| EmilienM | fungi: I suggest to take care though, and not using it for everything all the time | 14:51 |
| mordred | fungi: ++ | 14:51 |
| EmilienM | there is an overhead | 14:51 |
| mordred | EmilienM: well, we have a ton of if !defined(Package['foo']) { package['foo'] } blocks | 14:51 |
| mordred | which are just hella ugly | 14:51 |
| fungi | EmilienM: yeah, it's mostly that there are a handful of packages which are required by more than one of our puppet classes, and so we end up with nasty conditionals around them to avoid duplicate resource definition errors when both classes need to be coinstalled | 14:52 |
| EmilienM | mordred: yeah, and also all the param = '' instead of using undef | 14:52 |
| *** craigbr has quit IRC | 14:52 | |
| EmilienM | mordred: do you think we can go ahead with https://review.openstack.org/#/c/175434/ too ? | 14:53 |
| mordred | elarson: I'll loko at that next | 14:53 |
| openstackgerrit | Merged openstack-infra/storyboard-webclient: Updated dependencies https://review.openstack.org/178312 | 14:54 |
| mordred | EmilienM: for the rubygems mirror - is there a good way to configure our build hosts to only use our mirror once we have it? | 14:54 |
| *** devvesa has quit IRC | 14:54 | |
| *** stevemar has joined #openstack-infra | 14:54 | |
| *** ajmiller has joined #openstack-infra | 14:55 | |
| clarkb | no undef does not work | 14:55 |
| EmilienM | clarkb: ? | 14:55 |
| *** e0ne is now known as e0ne_ | 14:55 | |
| clarkb | there is a reason you use '' | 14:55 |
| clarkb | undef doesnt override anything | 14:56 |
| clarkb | as a parameter its useless | 14:56 |
| fungi | except in latest puppet where '' now evaluates true ;) | 14:56 |
| EmilienM | fungi: ++ | 14:56 |
| mordred | yeah | 14:56 |
| *** yfried|afk is now known as yfried__ | 14:56 | |
| EmilienM | clarkb: '' works for you, I just think there is a puppet way to handle your case | 14:56 |
| clarkb | well undef doesnt do it | 14:56 |
| * fungi tries to remember, of the languages he knows which will recast strings too bools, how many of them cast a null string to a true value | 14:57 | |
| clarkb | you need a false value that overrides defaults | 14:57 |
| EmilienM | mordred: build hosts == ? | 14:57 |
| fungi | EmilienM: jenkins::jenkinsuser class | 14:57 |
| *** nelsnelson has joined #openstack-infra | 14:57 | |
| *** Krinkle|detached is now known as Krinkle | 14:58 | |
| mordred | EmilienM: the nodes in nodepool | 14:58 |
| *** Krinkle is now known as Krinkle|detached | 14:58 | |
| fungi | that is, we'd want our jenkins user accounts to default to hitting our rubygems mirror and not try to hit official rubygems | 14:58 |
| mordred | EmilienM: we configure pip on each of them to know about the pypi mirror - how do we do the same with gem? | 14:58 |
| *** ashleighfarnham has quit IRC | 14:59 | |
| * fungi wonders if this is going to involve baking /etc/hosts entries for the mirror into worker images | 14:59 | |
| clarkb | oh and if you mix types you cant assert type | 15:00 |
| mordred | well, if gem already has a mirror command, he imagines gem has a config setting to consume mirrors ... | 15:00 |
| clarkb | its all bad, I dont think there is a right way to do it just many wrong ways | 15:00 |
| sigmavirus24 | fungi: I think JS counts a null string as a true value, potentially ruby too but I forget | 15:00 |
| sigmavirus24 | (I think Ruby does because empty arrays are truthy too) | 15:00 |
| *** mpopow has joined #openstack-infra | 15:01 | |
| EmilienM | mordred: good question, let me look at it now | 15:01 |
| fungi | sigmavirus24: ahh. fun times! | 15:01 |
| rcarrillocruz | i think there's a setting in gemrc for that? | 15:01 |
| sigmavirus24 | fungi: yeah. Former lives as a consultant help sometimes =/ | 15:02 |
| EmilienM | mordred: it will be in our Gemfile | 15:02 |
| EmilienM | bundle config mirror.https://rubygems.org https://gem.openstack.org | 15:02 |
| fungi | sigmavirus24: i assume you mean "help you not want to write in those languages any more if you can help it" | 15:02 |
| mordred | EmilienM: we'd actually prefer not putting those into gemfiles | 15:02 |
| clarkb | mordred ++ | 15:02 |
| sigmavirus24 | fungi: I actually like rubby a bit | 15:02 |
| mordred | EmilienM: because we don't really want people hitting them as not part of our build infrastructure | 15:03 |
| mordred | however, it seems we can do: gem sources --add http://localhost:9292 | 15:03 |
| EmilienM | mordred: do you want to trick with DNS then? | 15:03 |
| rcarrillocruz | we have something similar , where we put specific server gem settings on gemrc, like proxy | 15:03 |
| *** mtanino has joined #openstack-infra | 15:03 | |
| EmilienM | mordred: on our build nodes? | 15:03 |
| rcarrillocruz | i assume you can put your own mirror in it | 15:03 |
| mordred | no, I think we can do gem sources and it should work on the build nodes | 15:03 |
| rcarrillocruz | so it doesn't hit upstream gems site | 15:03 |
| *** teran has quit IRC | 15:03 | |
| mordred | EmilienM: we also have to deal with each region having its own mirror - there is no "gem.openstack.org" - there will be "gem.DFW.openstack.org" and "gem.region-b.geo-1.openstack.org" | 15:04 |
| *** teran has joined #openstack-infra | 15:04 | |
| EmilienM | mordred: I see two options: DNS or Gem config | 15:04 |
| mordred | yah. I think gem config is similar to what we do with pip | 15:04 |
| mordred | gem sources --remove http://rubygems.org ; gem sources --add http://gem.$region.openstack.org | 15:05 |
| EmilienM | mordred: but Gemfile is in our modules | 15:05 |
| EmilienM | or maybe we can patch project-config to configure it *before* running bundle | 15:05 |
| mordred | yah - but those should still work if we've gone gem sources on the build hosts without having to do settings in Gemfiles | 15:05 |
| *** bcourt has quit IRC | 15:05 | |
| mordred | EmilienM: oh - do we need to configure bundle as well? | 15:06 |
| EmilienM | no | 15:06 |
| EmilienM | it's done by Gemfile | 15:06 |
| mordred | awesome | 15:06 |
| jeblair | mordred: catching up -- what config file does 'gem sources' write to? | 15:06 |
| *** tonytan4ever has joined #openstack-infra | 15:07 | |
| EmilienM | mordred: I'm in favor of not touching Gemfile so our mirrors would be used only by our CI | 15:07 |
| *** e0ne_ is now known as e0ne | 15:07 | |
| jeblair | EmilienM: agreed -- we don't want the actual published modules to have any dependency on our infrastructure | 15:07 |
| openstackgerrit | Merged openstack-infra/project-config: puppet: run beaker on both trusty & centos7 https://review.openstack.org/175434 | 15:08 |
| EmilienM | so I think we'll have to trick with DNS *or* project-config scripts | 15:08 |
| jeblair | mordred: oh ~/.gemrc | 15:08 |
| mordred | EmilienM: I do not think we need to do either | 15:08 |
| *** teran has quit IRC | 15:08 | |
| EmilienM | crinkle: we have centos job now, I'm working on it until it works | 15:08 |
| mordred | EmilienM: I think we can configure our build nodes to know about the gem mirrors | 15:08 |
| EmilienM | mordred: /etc/hosts? | 15:09 |
| mordred | EmilienM: no | 15:09 |
| mordred | EmilienM: .gemrc / gem sources | 15:09 |
| EmilienM | tell me how !!! | 15:09 |
| EmilienM | ah | 15:09 |
| mordred | gem sources --remove http://rubygems.org ; gem sources --add http://gem.$region.openstack.org | 15:09 |
| EmilienM | that's tricky but it works too | 15:09 |
| mordred | it's what we do with pypi/pip | 15:09 |
| crinkle | EmilienM: \o/ | 15:09 |
| EmilienM | ah, ok | 15:09 |
| mordred | so we can extend the same script | 15:09 |
| EmilienM | mordred: can I send a patch somewhere for that? | 15:09 |
| EmilienM | crinkle: well, maybe it's broken now :-P | 15:10 |
| jeblair | EmilienM: http://git.openstack.org/cgit/openstack-infra/project-config/tree/nodepool/scripts/configure_mirror.sh | 15:12 |
| openstackgerrit | Monty Taylor proposed openstack-infra/project-config: Configure to use gem mirrors https://review.openstack.org/181427 | 15:13 |
| mordred | EmilienM: ^^ there you go :) | 15:13 |
| EmilienM | lol | 15:13 |
| EmilienM | jeblair: mordred was too fast :( | 15:13 |
| *** ajmiller has quit IRC | 15:13 | |
| openstackgerrit | Mateusz Matuszkowiak proposed openstack-infra/project-config: Added owner permissions to -core group https://review.openstack.org/181428 | 15:13 |
| jeblair | EmilienM: he was writing that instead of answering my question ;P | 15:13 |
| EmilienM | ahah | 15:13 |
| *** ajmiller has joined #openstack-infra | 15:13 | |
| mordred | jeblair: so - this makes me think that maybe we should think about naming these servers "mirror.$region.openstack.org" | 15:13 |
| jeblair | gave me time to rtfm and catch up anyway :) | 15:13 |
| clarkb | do we have mirrors yet? | 15:13 |
| clarkb | last I checked we couldnt puppet them at all... | 15:14 |
| mordred | clarkb: no - we should not yet land that change | 15:14 |
| jeblair | mordred: yes, i think we may have discussed that offhand, and yes i think we should | 15:14 |
| jeblair | mordred: we should move the pypi mirror one dir level down | 15:14 |
| mordred | jeblair: yes | 15:14 |
| mordred | jeblair: I'd like to suggest that we add the dns names | 15:14 |
| *** yfried__ is now known as yfried|afk | 15:15 | |
| mordred | jeblair: and then add new vhosts that serve pypi one level down | 15:15 |
| jeblair | unless.... we actually want to vhost them to keep our options for having multiple hosts in the future... | 15:15 |
| mordred | jeblair: then land the ready script change to point to the new vhost and dir | 15:15 |
| mordred | jeblair: oh. hrm | 15:15 |
| jeblair | i think i favor all in one host so we don't have to futz with dns to add a new type of mirror. but i i think it's worth considering. | 15:15 |
| *** mpopow has quit IRC | 15:15 | |
| jeblair | at least until we hit a point where we _need_ multiple hosts to handle our CI mirror load :) | 15:15 |
| clarkb | cool last i chevked wasnt recent enough, asselins fix has gone in and fixed things | 15:15 |
| *** mpopow has joined #openstack-infra | 15:16 | |
| mordred | jeblair: I'm having a hard time imaginging the time when we exceed our ability to serve static files ... | 15:16 |
| jeblair | mordred: yeah. let's go with mirror.x for now and deal with it when it becomes a problem | 15:17 |
| mordred | speaking of that- we need to spin up a mirror server in HP West | 15:17 |
| mordred | maybe I'll add the mirror dns entries when I do that | 15:17 |
| jeblair | mordred: kk | 15:18 |
| *** dannywilson has joined #openstack-infra | 15:20 | |
| *** fawadk has joined #openstack-infra | 15:20 | |
| *** bcourt has joined #openstack-infra | 15:20 | |
| fungi | i'm in favor of the separate apache vhosts idea. mainly because it's possible we'll at some point run into something we want to mirror which doesn't have so flexible a url file path option | 15:20 |
| jeblair | fungi: okay... i guess we could wildcard dns to make it easier to spin them uP? | 15:21 |
| jeblair | mordred: ^ | 15:21 |
| Vivek | jeblair, mordred: I am comparing 2 openstack version docs, icehouse and juno regarding creating initial networks, the junp guide says neutron net-create ext-net --router:external True --provider:physical_network external --provider:network_type flat, the ice house guide days neutron net-create ext-net --shared --router:external=True | 15:22 |
| fungi | yeah, so long as we don't intent to need to pay for x.509 certs for them, something like mirror.dfw.openstack.org which can have pypi.mirror.dfw.openstack.org or something as a vhost if we want | 15:22 |
| Vivek | The URLs in question are http://docs.openstack.org/icehouse/install-guide/install/apt/content/neutron_initial-external-network.html and http://docs.openstack.org/juno/install-guide/install/apt/content/neutron_initial-external-network.html | 15:23 |
| fungi | eh, i can also see reasons to not bother with additional vhosts for now, as long as we don't paint ourselves into a corner of not being able to use them easily down the road | 15:23 |
| mordred | yeah - I mean, I don't think we need the specific vhost _now_ | 15:23 |
| mordred | because your example is something we can certainly add in the future if needed | 15:23 |
| fungi | yep. let's just do whatever's easiest | 15:23 |
| jeblair | okay, we back to putting everything on mirror.x for now then? | 15:23 |
| clarkb | and looking ahead, pypi, gems, apt, yum, maven, none should need special vhost stuff | 15:24 |
| *** dannywil_ has joined #openstack-infra | 15:24 | |
| *** fawadkhaliq has quit IRC | 15:24 | |
| jeblair | javascript? | 15:24 |
| Vivek | I am facing an issue that when I launch the vms the internal network which is in the 192.168.1.x series is not displayed in the dashboard. | 15:24 |
| clarkb | jeblair they use couchdb or something like that | 15:24 |
| *** dannywilson has quit IRC | 15:24 | |
| clarkb | and Go is just git.o.o/allofgithub | 15:24 |
| *** fawadkhaliq has joined #openstack-infra | 15:25 | |
| nibalizer | https://www.npmjs.com/package/npm-mirror makes it look like you can just do normal vhost on name/directory | 15:25 |
| nibalizer | might be a little more complicated | 15:25 |
| nibalizer | but not by much | 15:25 |
| fungi | speaking of mirrors and whatnot, apparently the problem i'm running into with centos-6 package caching is that python-setuptools is referenced as a dep in pylint and python-nose rpms, but isn't itself a thing yum can find to retrieve. i _think_ it just assumes the package will be preinstalled as a base component of the operating system and not removed? | 15:25 |
| Vivek | The internal network is created with the demo credentials. | 15:25 |
| jeblair | clarkb: okay, we have smart engineers. someone wants to add those they can figure them out :) | 15:25 |
| *** tjones1 has joined #openstack-infra | 15:25 | |
| nibalizer | point being that you don't need to run a linked couchdb any more :P | 15:26 |
| jeblair | nibalizer: woot | 15:26 |
| clarkb | nibalizer that is good news | 15:26 |
| *** hodos has joined #openstack-infra | 15:26 | |
| clarkb | Vivek I am not sure this is the best place for answers | 15:26 |
| clarkb | we run openstacks developer infrastructure but know little of the intricate details of deploying neutron | 15:27 |
| Vivek | I know but #openstack is silent for a while now.... | 15:27 |
| *** scheuran has quit IRC | 15:27 | |
| Vivek | clarkb: I understand. | 15:27 |
| *** fawadk has quit IRC | 15:27 | |
| fungi | Vivek: see the channel /topic (we don't for the most part actually run openstack api servers, nor did we write that documentation you linked) | 15:28 |
| fungi | Vivek: we run the wiki, the code review system, the test rig, and so on | 15:29 |
| *** btully has joined #openstack-infra | 15:29 | |
| *** sks has joined #openstack-infra | 15:29 | |
| *** yfried|afk is now known as yfried__ | 15:30 | |
| fungi | okay, confirmed. python-setuptools is in the base repo. i think if we uninstall it we can't actually get yum to pre-cache any rpms which depend on it. that's... very annoying | 15:31 |
| *** rlandy has quit IRC | 15:32 | |
| clarkb | uninstall after? | 15:32 |
| *** rm_work is now known as rm_work|away | 15:32 | |
| fungi | or somehow fake out the installed package set to think it's installed when it's not, so as to satisfy dependencies | 15:32 |
| fungi | maybe a dummy rpm which has no contents but claims to provide python-setuptools | 15:32 |
| *** samuelBartel has quit IRC | 15:33 | |
| *** craigbr has joined #openstack-infra | 15:33 | |
| fungi | i'm hesitant to change the order in which we uninstall distro-provided setuptools, since we want to do that before installing the one from pypi | 15:33 |
| fungi | which we do as early as possible in the image setup | 15:34 |
| *** craigbr has quit IRC | 15:34 | |
| openstackgerrit | Merged openstack-infra/system-config: Add statusbot to openstack-ansible https://review.openstack.org/179387 | 15:34 |
| nibalizer | jeblair: EmilienM so where did we get to on vhosting? | 15:34 |
| nibalizer | https://review.openstack.org/#/c/178026/9 right now is not set up for an all-in-one | 15:34 |
| *** mwagner_lap has quit IRC | 15:35 | |
| nibalizer | and changing it to be an all in one would be pretty significant because it would break the logical separation between the pypi/rubygems classes/vhost files | 15:35 |
| fungi | nibalizer: sounds like we're just going to do a common mirror.x.openstack.org in each provider/region and use the same vhost for pypi, rubygems, debs, et cetera | 15:35 |
| fungi | unless there's some complication with doing so | 15:35 |
| fungi | for example, make the vhost setup optional in those modules so that we can substitute our own server-specific vhost template? | 15:36 |
| *** panda is now known as panda|afk | 15:37 | |
| nibalizer | fungi: this is all in openstack_project, so no need to handle any use case that isn't ours | 15:38 |
| fungi | ahh, even better | 15:38 |
| nibalizer | Me or someone else may spin the bulk of EmilienM's patch out into something consumable but we don't have to deal with that now | 15:38 |
| jeblair | fungi: yeah, i think that's what we want. so in short: move pypi down one level, put gems mirror as a sibling, have one vhost file that points to the mirror root (which has pypi and gems direcotories)* | 15:39 |
| Vivek | Vivek> jeblair, mordred: I am comparing 2 openstack version docs, icehouse and juno regarding creating initial networks, the junp guide | 15:39 |
| Vivek | says neutron net-create ext-net --router:external True --provider:physical_network external --provider:network_type flat, the ice | 15:39 |
| Vivek | house guide days neutron net-create ext-net --shared --router:external=True | 15:39 |
| jeblair | fungi, nibalizer: * (we may have 2 vhost files for a short while while we perform the pypi move) | 15:39 |
| Vivek | 20:52 < fungi> yeah, so long as we don't intent to need to pay for x.509 certs for them, something like mirror.dfw.openstack.org which can have | 15:39 |
| EmilienM | jeblair: are you doing that a friday? | 15:39 |
| Vivek | pypi.mirror.dfw.openstack.org or something as a vhost if we want | 15:39 |
| Vivek | 20:53 < Vivek> The URLs in question are http://docs.openstack.org/icehouse/install-guide/install/apt/content/neutron_initial-external-network.html | 15:39 |
| Vivek | and http://docs.openstack.org/juno/install-guide/install/apt/content/neutron_initial-external-network.html | 15:39 |
| Vivek | 20:53 < fungi> eh, i can also see reasons to not bother with additional vhosts for now, as long as we don't paint ourselves into a corner of not | 15:39 |
| Vivek | being able to use them easily down the road | 15:39 |
| Vivek | 20:53 < mordred> yeah - I mean, I don't think we need the specific vhost _now_ | 15:39 |
| Vivek | 20:53 < mordred> because your example is something we can certainly add in the future if needed | 15:39 |
| nibalizer | I think we need an openstack_project::mirror class that include 1) the vhost file to sevre everything 2) include pypimirror(actually bandersnatch) 3) include rubygems mirror | 15:39 |
| Vivek | 20:53 < fungi> yep. let's just do whatever's easiest | 15:39 |
| Vivek | 20:53 < jeblair> okay, we back to putting everything on mirror.x for now then? | 15:39 |
| Vivek | 20:54 < clarkb> and looking ahead, pypi, gems, apt, yum, maven, none should need special vhost stuff | 15:40 |
| *** Guest58709 is now known as annegentle | 15:40 | |
| Vivek | 20:54 -!- dannywil_ [~dannywils@70.103.118.141] has joined #openstack-infra | 15:40 |
| Vivek | 20:54 -!- fawadkhaliq [~fawadkhal@mbl-109-48-176.dsl.net.pk] has quit [Read error: Connection reset by peer] | 15:40 |
| Vivek | 20:54 < jeblair> javascript? | 15:40 |
| Vivek | 20:54 < Vivek> I am facing an issue that when I launch the vms the internal network which is in the 192.168.1.x series is not displayed in the dashboard. | 15:40 |
| Vivek | 20:54 < clarkb> jeblair they use couchdb or something like that | 15:40 |
| Vivek | 20:54 -!- dannywilson [~dannywils@70.103.118.141] has quit [Ping timeout: 276 seconds] | 15:40 |
| Vivek | 20:54 < clarkb> and Go is just git.o.o/allofgithub | 15:40 |
| Vivek | 20:55 -!- fawadkhaliq [~fawadkhal@mbl-109-48-176.dsl.net.pk] has joined #openstack-infra | 15:40 |
| Vivek | 20:55 < nibalizer> https://www.npmjs.com/package/npm-mirror makes it look like you can just do normal vhost on name/directory | 15:40 |
| Vivek | 20:55 < nibalizer> might be a little more complicated | 15:40 |
| Vivek | My applogies for flooding. | 15:40 |
| Vivek | sorry did those lines get posted to the channel ? | 15:40 |
| EmilienM | yes | 15:40 |
| nibalizer | Vivek: yep | 15:40 |
| Vivek | Sorry again. | 15:40 |
| fungi | at least it stopped before i got finished adding you to the autokick list | 15:41 |
| *** yfried__ is now known as yfried|afk | 15:41 | |
| cinerama | StevenK: lol. yes, i had incidentally noticed the timeout issue but for some reason i hadn't filed it yet | 15:41 |
| nibalizer | jeblair: cool so then we'll tell clients to use mirror.<az>.o.o/<yourtypeofmirror> | 15:41 |
| jeblair | EmilienM: i think my schedule is booked today | 15:41 |
| Vivek | fungi: Thanks for clarifying :) | 15:41 |
| jeblair | nibalizer: yep | 15:41 |
| nibalizer | EmilienM: mind if I update your patchset with the above^? | 15:41 |
| clarkb | my schedule is, need to do gerrit prep | 15:41 |
| EmilienM | nibalizer: GO | 15:42 |
| nibalizer | clarkb: ++ | 15:42 |
| fungi | jeblair: also if we regularly change the paths/names of our pypi mirrors, we should be able to effectively dissuade others from using them | 15:42 |
| *** jaypipes is now known as leakypipes | 15:42 | |
| clarkb | fungi ha | 15:42 |
| nibalizer | its also worth noting that greghaynes has this https://review.openstack.org/#/c/165240/ which maybe wouldn't need its own hosts now? just run in another subdir under /srv/static/mirrors ? | 15:43 |
| jeblair | fungi: changing the apt repo to "test-only-do-not-use-in-production-it-will-break-go-away-beware-of-the-leopard" didn't work :/ | 15:44 |
| nibalizer | well maybe thats more about having a remote builder | 15:44 |
| clarkb | nibalizer ya that shouldnt affect this | 15:44 |
| clarkb | we would just have wheels/ as another root | 15:45 |
| clarkb | but the cffi stuff has me worried about it | 15:45 |
| jeblair | btw, i hope to get the afs stuff back on track so that all our mirroring can be "run once in a central location and publish to all mirrors via afs" | 15:45 |
| clarkb | I think we have to blacklist all cffi built wheels | 15:45 |
| jeblair | but that's not the highest priority thing around atm | 15:46 |
| fungi | jeblair: perhaps we should have hosted it from a disused lavatory | 15:46 |
| jeblair | fungi: i just had an idea for a cloud startup | 15:46 |
| fungi | bwahahahaha | 15:46 |
| *** sks has quit IRC | 15:46 | |
| clarkb | because if cffi updates, all our cffi built wheels will need to recompile then its gg | 15:47 |
| *** ildikov has quit IRC | 15:47 | |
| clarkb | at least until cffi 1.0 is in use and we build with that everywhere | 15:48 |
| *** shashankhegde has joined #openstack-infra | 15:48 | |
| *** ivar-lazzaro has quit IRC | 15:48 | |
| fungi | s/in use/even released at all/ | 15:49 |
| fungi | it's tough to make guarantees about what its actual behavior will be since it's still in beta | 15:50 |
| clarkb | ya | 15:50 |
| sigmavirus24 | fungi: so one solution would be for yet another foundation to give Armin a grant to work on it full time | 15:51 |
| sigmavirus24 | The PSF helped fund 1.0 beta | 15:51 |
| *** gyee has joined #openstack-infra | 15:52 | |
| *** mpavone has quit IRC | 15:52 | |
| *** bcourt has quit IRC | 15:53 | |
| Daviey | jeblair: just one afs cell, right? | 15:53 |
| *** otter768 has joined #openstack-infra | 15:53 | |
| clarkb | zaro: I can help shepherd things through today | 15:54 |
| fungi | sigmavirus24: i'm all for more funding to free software projects, though in our case the effort might be better invested in finding more pure python implementations of things so we can depend less on system libraries | 15:54 |
| derekh | If anybody has a chance could you take a look at the patch to add Fedora 21 to the tripleo cloud https://review.openstack.org/#/c/169778/ | 15:54 |
| clarkb | zaro: let me know when you are here and we can make sure that anything that can get in today is at least attempted | 15:54 |
| derekh | and the patch to start using it please https://review.openstack.org/#/c/181136/ | 15:54 |
| jeblair | derekh: yep | 15:54 |
| sigmavirus24 | fungi: yeah well I can think of at least one thing that we can't replace like that: cryptography | 15:54 |
| derekh | jeblair: thanks | 15:54 |
| jeblair | derekh: ah sorry tab fail | 15:55 |
| clarkb | sigmavirus24: well we sort of did, we pin to before they use cffi \o/ | 15:55 |
| jeblair | derekh: i have no idea what i just said yes to :) | 15:55 |
| jeblair | Daviey: yep | 15:55 |
| sigmavirus24 | clarkb: yuck | 15:55 |
| clarkb | sigmavirus24: sorry pyopenssl | 15:55 |
| jeblair | derekh: i will read what you said now :) | 15:55 |
| clarkb | sigmavirus24: no its much better | 15:55 |
| derekh | jeblair: ;-) , thanks | 15:55 |
| clarkb | sigmavirus24: it doesn't OOM your machines | 15:55 |
| fungi | sigmavirus24: or maybe that's top of the heap for something which should have a good pure python alternative after all | 15:55 |
| *** openstackstatus has quit IRC | 15:56 | |
| *** dkranz has quit IRC | 15:56 | |
| fungi | maybe not for ultra-high-performance situations, but there are plenty of places where you want good crypto running entirely within an interpreter | 15:56 |
| *** openstackstatus has joined #openstack-infra | 15:57 | |
| *** ChanServ sets mode: +v openstackstatus | 15:57 | |
| *** patrickeast has joined #openstack-infra | 15:57 | |
| jeblair | derekh: okay :) | 15:57 |
| dstufft | fungi: you can't do crypto in pure python | 15:58 |
| dstufft | it's totally insecure | 15:58 |
| clarkb | mordred: ssh to 15.126.143.112 thats your ubuntu-trusty node | 15:58 |
| dstufft | there are side channels all over the place | 15:58 |
| clarkb | mordred: I can get in so networking worked | 15:58 |
| *** otter768 has quit IRC | 15:58 | |
| dstufft | it's impossible to do constant time python | 15:58 |
| fungi | that also sounds like something worth improving then | 15:58 |
| dstufft | (there are some very limited situations where a side channel isn't a big deal, but it's subtle when it is or isn't and it's much safer to just assume not) | 15:59 |
| clarkb | mordred: one thing that simple-init does not appear to do is set the hostname (not a big deal, just noting my observations as I poke around) | 15:59 |
| derekh | jeblair: greghaynes clarkb thanks | 15:59 |
| *** hdd has quit IRC | 15:59 | |
| greghaynes | clarkb: did you see something didnt go well witho our recent dib zuul-cloner change? http://logs.openstack.org/61/181161/4/check/gate-dib-dsvm-functests-devstack-trusty/733d50b/console.html | 16:00 |
| dstufft | fungi: so it's possible to do that in Python if you added the needed primtitives to Python itself.. but you'd essentially being adding primitives which are only every good for cryptography and it'd only be 3.5 (or more likely 3.6+) if you started working on it *right now* | 16:01 |
| dstufft | only ever* | 16:01 |
| clarkb | greghaynes: ya I have no idea what caused that | 16:01 |
| openstackgerrit | Spencer Krum proposed openstack-infra/system-config: Create rubygems mirror from rubygems.org https://review.openstack.org/178026 | 16:01 |
| nibalizer | EmilienM: ^^ | 16:01 |
| clarkb | greghaynes: jeblair may recognize it? | 16:01 |
| *** freerunner has quit IRC | 16:01 | |
| fungi | or use a protocol to communicate with the crypto implementation rather than having it running in your interpreter's execution space (and then implement the crytpto service in a separate language). i've just had lots of headaches trying to get interpreted languages sanely loading precompiled libraries | 16:01 |
| sigmavirus24 | dstufft: 3.6 at best | 16:02 |
| greghaynes | clarkb: possible the removing gerrit-git-prep? | 16:02 |
| greghaynes | clarkb: since thats the only real change | 16:02 |
| clarkb | greghaynes: maybe? that would be really odd though | 16:02 |
| sigmavirus24 | fungi: I'm up to rewrite OpenStack in C | 16:02 |
| EmilienM | nibalizer: thanks | 16:02 |
| fungi | sigmavirus24: apparently we're rewriting it in go, shortly after google takes over | 16:02 |
| clarkb | greghaynes: its trying to clean up the repo it just cloned | 16:02 |
| clarkb | greghaynes: and that failed | 16:02 |
| *** annegentle has quit IRC | 16:02 | |
| sigmavirus24 | fungi: OpenStack Foundation is getting acquihired, no? | 16:02 |
| sigmavirus24 | =P | 16:03 |
| *** jlanoux has quit IRC | 16:03 | |
| *** e0ne is now known as e0ne_ | 16:04 | |
| clarkb | greghaynes: I wonder, that could be a bug due to the zuul-cloner change that was just merged | 16:04 |
| clarkb | jeblair: ^ that chnage did just merge ya? | 16:04 |
| dstufft | fungi: eh, loading the .so isn't the hard part really, CPython handles that pretty well. The problem with cffi < 1.0 is it uses a bunch of implicit magic all over the place | 16:04 |
| openstackgerrit | Emilien Macchi proposed openstack-infra/project-config: beaker: support ssh service name on centos https://review.openstack.org/181442 | 16:04 |
| fungi | dstufft: well, it's still a problem if you need to build c extensions to python packages. makes it a lot harder to keep them cross-platform-sane | 16:05 |
| jeblair | clarkb: no i don't think it's merged | 16:05 |
| greghaynes | clarkb: https://review.openstack.org/#/c/117626/ ? | 16:06 |
| clarkb | nevermind then | 16:06 |
| clarkb | hrm why would that fail? | 16:06 |
| dstufft | fungi: sure, compiled code isn't cross platform, but that's a reason to use C sparingly where it makes sense (like cryptography libraries ;)) not a reason to eschew the use of C all together | 16:06 |
| clarkb | greghaynes: is it consistent? | 16:06 |
| clarkb | mordred: also looks like simple-init unmounts the config drive, might be handy to leave that mounted? | 16:06 |
| openstackgerrit | Merged openstack-infra/system-config: Add a Fedora 21 node to tripleo cloud https://review.openstack.org/169778 | 16:07 |
| greghaynes | clarkb: it was late last night, I can recheck and see if it re-happens | 16:07 |
| clarkb | greghaynes: I wonder if it is related to the image it ran on | 16:07 |
| clarkb | greghaynes: maybe it didn't get HEAD because of the way we cached the repo? | 16:08 |
| *** ashleighfarnham has joined #openstack-infra | 16:08 | |
| greghaynes | hrm | 16:08 |
| clarkb | I can confirm there is no /opt/git/openstack/diskimage-builder/.git/refs/heads/HEAD on the ubuntu-trusty node I am on | 16:08 |
| jeblair | clarkb: what about other repos? | 16:08 |
| clarkb | jeblair: nova doesn't have one either. going to check system-config since we use zuul-cloner there | 16:09 |
| greghaynes | so, I wonder if its a bug in zuul cloner that we tickled by removing the gerrit-git-prep, maybe gerrit-git-prep was fetching it? | 16:09 |
| *** unicell1 has quit IRC | 16:09 | |
| fungi | clarkb: hpcloud or rax? did we skip a clone error in an image build? | 16:09 |
| mordred | clarkb: yes - I did not have it set the hostname because we do that in puppet thing - and the cloud-init setting of the hostname doesn't ever do what we expect anyway - happy to adapt that though if you think we should make it do the right thing | 16:09 |
| clarkb | ya none in system-config either | 16:09 |
| jeblair | greghaynes: i think we use zuul-cloner without ggp on system-config | 16:09 |
| *** bcourt has joined #openstack-infra | 16:09 | |
| mordred | clarkb: also - I'm _totally_ fine with not unmounting - I think I'd prefer that | 16:10 |
| clarkb | fungi: the fail is from hpcloud and I am on an hpcloud ubuntu-trusty node, all would've been dib built | 16:10 |
| openstackgerrit | Emilien Macchi proposed openstack-infra/project-config: beaker: support ssh service name on centos https://review.openstack.org/181442 | 16:10 |
| *** yfried|afk has quit IRC | 16:10 | |
| EmilienM | nibalizer, mordred ^ I had to fix ssh restart for beaker jobs on centos | 16:10 |
| *** shashankhegde has quit IRC | 16:11 | |
| nibalizer | EmilienM: lgtm | 16:12 |
| fungi | clarkb: yeah, for hpcloud a clone error in dib of that image would affect all regions there. seeing if the logs offer any insight | 16:12 |
| EmilienM | nibalizer: well, trusty is broken now | 16:12 |
| EmilienM | nibalizer: http://logs.openstack.org/39/181439/1/check/gate-puppet-keystone-puppet-beaker-rspec-dsvm-trusty/f216a4e/console.html#_2015-05-08_16_00_28_574 | 16:12 |
| EmilienM | is that related to the new image we're using? | 16:12 |
| EmilienM | actually it's http://logs.openstack.org/39/181439/1/check/gate-puppet-keystone-puppet-beaker-rspec-dsvm-trusty/f216a4e/console.html#_2015-05-08_16_00_28_592 | 16:13 |
| jeblair | clarkb, fungi, greghaynes: i just ran on a bare-trusty hpcloud-b3 node and it worked (that's a different image; the failure was on devstack-trusty in hpcloud-b3) | 16:13 |
| greghaynes | awesome | 16:14 |
| clarkb | jeblair: bare-trusty is snapshot built and devstack-trusty is dib built with the git cache | 16:14 |
| *** armax has quit IRC | 16:14 | |
| clarkb | I wonder if this is a bug in dib elements | 16:14 |
| fungi | looking in the dib log for that image, it looks like the clone didn't raise any errors | 16:15 |
| *** dan_ has joined #openstack-infra | 16:15 | |
| jeblair | oh, i forgot to mention, the node i tried on was 22 hours old | 16:15 |
| clarkb | nodepool:/opt/dib_cache/source-repositories/diskimage_builder_cd49334dd93562d7a6a7ff4d88b37662cb648227/.git/refs/heads has no HEAD either | 16:15 |
| fungi | HEAD is now at 9ac2a5b Merge "Package ldlinux.c32 along with isolinux.bin if it exists" | 16:15 |
| *** dan_ is now known as Guest64787 | 16:15 | |
| jeblair | i just tried on a 74 hour old devstack-trusty-2-node-rax-iad and it worked | 16:15 |
| clarkb | jeblair: thats also snapshot built so I think that does point to something odd with how we are making dib images | 16:16 |
| jeblair | brand new devstack-trusty-hpcloud-b2 failed | 16:16 |
| clarkb | jeblair: can you check the .git/refs/heads dir in a repo? | 16:16 |
| clarkb | jeblair: on a working node | 16:16 |
| *** gyee has quit IRC | 16:16 | |
| jeblair | clarkb: cache or clone? | 16:17 |
| *** freerunner has joined #openstack-infra | 16:17 | |
| clarkb | jeblair: cache /opt/git/openstack/diskimage-builder/.git/refs/heads | 16:17 |
| clarkb | though HEAD is a special ref and maybe the source side isn't the issue? | 16:18 |
| *** btully has quit IRC | 16:18 | |
| clarkb | this is one thing about git that I dislike quite a bit | 16:18 |
| jeblair | corvus@devstack-trusty-2-node-rax-iad-2416531:/opt/git/openstack/diskimage-builder$ ls /opt/git/openstack/diskimage-builder/.git/refs/heads | 16:18 |
| jeblair | master | 16:18 |
| jeblair | working ^ | 16:18 |
| jeblair | corvus@ubuntu:/opt/git/openstack/diskimage-builder$ ls /opt/git/openstack/diskimage-builder/.git/refs/heads | 16:18 |
| jeblair | fetch_master master | 16:18 |
| jeblair | failing ^ | 16:18 |
| fungi | basically last mention during the build is that it did a `sudo git reset --hard FETCH_HEAD` which reports that it set HEAD successfully | 16:18 |
| fungi | though maybe that's in reference to the work-tree HEAD and not the clone HEAD | 16:19 |
| clarkb | oh is .git/HEAD possibly pointing to something not in .git/refs/heads and that error is a bit confused? | 16:19 |
| jeblair | fetch_master is the ref that zuul-cloner is trying to clean up | 16:19 |
| jeblair | HEAD is "ref: refs/heads/master" in both places | 16:20 |
| clarkb | jeblair: what about in the cloned repo side for the host that fails? | 16:21 |
| * clarkb runs some zuul-cloner on this ubuntu-trusty node | 16:21 | |
| jeblair | ref: refs/remotes/origin/HEAD | 16:22 |
| *** gyee has joined #openstack-infra | 16:22 | |
| openstackgerrit | Emilien Macchi proposed openstack-infra/project-config: beaker: support ssh service name on centos https://review.openstack.org/181442 | 16:22 |
| clarkb | hrm zuul-cloner really wants me to set all of these zuul vars | 16:25 |
| fungi | yep, not entirely easy to test since they don't have defaults | 16:25 |
| jeblair | clarkb: http://paste.openstack.org/show/217233/ | 16:25 |
| jeblair | (we should have zuul-cloner emit that for easier debugging) | 16:26 |
| *** dboik has joined #openstack-infra | 16:27 | |
| *** dboik_ has quit IRC | 16:27 | |
| clarkb | ya so .git/HEAD -> .git/refs/remotes/origin/HEAD -> .git/refs/remotes/origin/master DNE | 16:28 |
| *** ildikov has joined #openstack-infra | 16:28 | |
| clarkb | I wonder if the prune is pruning too much? | 16:29 |
| jeblair | so what is the 'fetch_master' ref ? | 16:29 |
| jeblair | clarkb: it's trying to prune 'fetch_master' which only exists on the dib repos | 16:29 |
| *** dboik_ has joined #openstack-infra | 16:30 | |
| EmilienM | mordred: I had to change beaker job to add missing package due to a new image used - https://review.openstack.org/#/c/181442 | 16:30 |
| clarkb | jeblair: right but it appears to be failing because refs/remotes/origin/master DNE | 16:30 |
| clarkb | jeblair: that is what HEAD resolves to | 16:30 |
| *** dizquierdo has quit IRC | 16:30 | |
| fungi | so this points to a bug (lack of validation at least?) in the source-repositories element? | 16:32 |
| clarkb | fungi: no I don't think source-repositories is broken | 16:32 |
| mordred | o wow scrollback | 16:32 |
| * nibalizer bookmarks jeblair's link | 16:32 | |
| jeblair | the dib caching causes a ref called fetch_master to be created; the zuul cloner causes that to propogate to the clone and also sets HEAD to ref: refs/remotes/origin/HEAD; when zuul-cloner rewrites the origin it causes that ref to be invalid | 16:32 |
| clarkb | fungi: there is no fetch_master in that repo | 16:32 |
| jeblair | nibalizer: it's only valid for that job -- i got it by checking the jenkins parameters page for the job though -- all the vars are there | 16:32 |
| clarkb | fungi: its possibly in our clone from source-repositories cache into the image | 16:32 |
| fungi | 2015-05-08 15:17:24,102 INFO nodepool.image.build.devstack-trusty: * [new branch] master -> fetch_master | 16:33 |
| *** dboik has quit IRC | 16:33 | |
| fungi | seems to have created it | 16:33 |
| nibalizer | jeblair: but as a template its gold | 16:33 |
| jeblair | clarkb, fungi: so i think there's two ways to fix: cause fetch_master not to be created in the clone (this is a dib change); or have zuul cloner update HEAD to something else after cloning and before pruning | 16:33 |
| jeblair | nibalizer: ya | 16:33 |
| nibalizer | although yea zuul_ref is not guessable | 16:33 |
| *** Swami has joined #openstack-infra | 16:34 | |
| clarkb | fungi: what script is that a part of? if you go up in the log dib should tell ou what it ran | 16:34 |
| *** btran_ has joined #openstack-infra | 16:34 | |
| mordred | greghaynes: what is fetch_master used for? | 16:34 |
| nibalizer | jeblair: also 180782 180783 | 16:34 |
| *** Sukhdev has joined #openstack-infra | 16:34 | |
| greghaynes | mordred: that is a good question | 16:35 |
| mordred | git fetch $CACHE_PATH $REPOREF:fetch_$REPOREF | 16:35 |
| mordred | btw | 16:35 |
| clarkb | nibalizer: thats where the changes are /me updates their topics | 16:35 |
| mordred | is where it gets created | 16:35 |
| mordred | in elements/source-repositories/extra-data.d/98-source-repositories: | 16:35 |
| clarkb | huh the source-repository in the cache does not have that | 16:35 |
| clarkb | so that must only happen image side | 16:35 |
| mordred | # When we first clone we create a branch naming what we fetched | 16:35 |
| mordred | # that must match, or we are asking for two different references from the | 16:35 |
| mordred | # same repo, which is an error | 16:35 |
| jeblair | clarkb: i have a fetch_master ref on my nodes | 16:35 |
| greghaynes | mordred: yea, that is happening as it clones out of the cache | 16:36 |
| greghaynes | er | 16:36 |
| greghaynes | clarkb: ^ | 16:36 |
| mordred | I think it's in support of specifying a specific ref in the source-repositories file and then being able to have thigns consume "the ref that was fetched" without them also having to know the ref | 16:36 |
| clarkb | jeblair: ya me too | 16:36 |
| clarkb | jeblair: it does not exist in the dib cache on nodepool.o.o though | 16:36 |
| clarkb | mordred: we call that FETCH_HEAD | 16:37 |
| jeblair | clarkb: ah, gotcha, and i think mordred's paste just splains that | 16:37 |
| *** unicell has joined #openstack-infra | 16:37 | |
| fungi | clarkb: extra-data.d/98-source-repositories | 16:37 |
| mordred | clarkb: yah - I'm trying to dig in to why that wasn't juse used | 16:37 |
| fungi | is what's running that | 16:37 |
| jeblair | gah, i have to run do some errands, sorry | 16:37 |
| mordred | sudo git fetch $CACHE_PATH $REPOREF:fetch_$REPOREF | 16:38 |
| mordred | sudo git reset --hard FETCH_HEAD | 16:38 |
| *** dustins has joined #openstack-infra | 16:38 | |
| clarkb | ok, so we know where that ref is coming from, we know zuul-cloner wants to delete it, but why is refs/remotes/origin/HEAD pointing to refs/remotes/origin/master which does not exist? | 16:38 |
| greghaynes | https://review.openstack.org/#/c/135580/ | 16:38 |
| *** aduarte has joined #openstack-infra | 16:38 | |
| greghaynes | that is what added it | 16:38 |
| *** ajmiller has quit IRC | 16:38 | |
| fungi | clarkb: elements/source-repositories/extra-data.d/98-source-repositories in openstack/diskimage-builder | 16:39 |
| *** ajmiller has joined #openstack-infra | 16:39 | |
| clarkb | fungi: thanks | 16:39 |
| mordred | ok. I think that patch can get reworked | 16:39 |
| nibalizer | clarkb: ah i totally spaced on topic sorry | 16:39 |
| *** dboik_ has quit IRC | 16:39 | |
| mordred | but I agre, it does not explain clarkb's question | 16:39 |
| fungi | looks like line 162 | 16:39 |
| *** marcusvrn has quit IRC | 16:39 | |
| fungi | sudo git fetch $CACHE_PATH $REPOREF:fetch_$REPOREF | 16:39 |
| mordred | which is why is HEAD pointing to non-existant origin/master | 16:39 |
| clarkb | mordred: I think that if HEAD did not do that, this would work | 16:40 |
| *** gokrokve has quit IRC | 16:40 | |
| *** dboik has joined #openstack-infra | 16:40 | |
| mordred | yah | 16:40 |
| mordred | I think fetch_master is silly - but I'm better it's a red herring | 16:40 |
| clarkb | mordred: well it triggers this whole thing | 16:40 |
| *** yamahata has quit IRC | 16:41 | |
| clarkb | OH! | 16:41 |
| * mordred excites | 16:41 | |
| clarkb | I think its because we explode when changing our remote | 16:41 |
| *** radez is now known as radez_g0n3 | 16:41 | |
| clarkb | if you look at the zuul-cloner log what it is doing is changing the origin remote to point at our git mirror | 16:41 |
| clarkb | but then it also tries to clean up this thing which git won't do in this half broken state | 16:41 |
| *** marcusvrn has joined #openstack-infra | 16:41 | |
| clarkb | I wonder if this is just an order of operations thing inside of zuul-cloner | 16:42 |
| *** radez_g0n3 is now known as radez | 16:42 | |
| mordred | ah - nod | 16:43 |
| *** harlowja has quit IRC | 16:44 | |
| *** harlowja has joined #openstack-infra | 16:44 | |
| fungi | i'm sort of curious why it didn't start until today's image update though | 16:45 |
| clarkb | fungi: because this is the only job using zuul-cloner on dib built images | 16:46 |
| fungi | ohhh | 16:46 |
| zaro | morning | 16:46 |
| greghaynes | clarkb: why wasnt it failing before then? | 16:46 |
| clarkb | ok new update, if I use git to clone the repo I don't get a fetch_master | 16:46 |
| greghaynes | ggp cleaned up the repo? | 16:46 |
| *** tjones1 has quit IRC | 16:46 | |
| clarkb | greghaynes: you probably ran on rax exclusively | 16:46 |
| greghaynes | hrm, it ran a lot | 16:46 |
| zaro | clarkb: ready | 16:46 |
| mordred | clarkb: that's right- I would expect you would see a remote/origin/fetch_master in git branch -a though | 16:46 |
| *** packet has quit IRC | 16:46 | |
| clarkb | mordred: yes I do have that but not in refs/heads just under origin | 16:47 |
| mordred | right | 16:47 |
| clarkb | zaro: cool, so first question, do we have the version of gerrit that we want to deploy built in a war and running on review-dev? | 16:47 |
| mordred | I believe that is expected | 16:47 |
| clarkb | zaro: and is an up to date change to deploy that way on review.o.o pushed? | 16:48 |
| zaro | clarkb: did we upgrade all jenkins with your last change or was it only 07? | 16:48 |
| zaro | clarkb: checking | 16:48 |
| clarkb | zaro: all of them have my change but not he register functions change, just the work() loop update | 16:48 |
| *** armax has joined #openstack-infra | 16:49 | |
| clarkb | mordred: so when I do `git clone file:///opt/git/openstack/diskimage-builder && cd diskimage-builder && git branch -d -r origin/fetch_master` that works | 16:49 |
| *** freerunner has quit IRC | 16:49 | |
| pabelanger | Do we have any documentation about moving a project into the openstack namespace, from stackforge? Assuming governance approves it | 16:50 |
| zaro | clarkb: i expected 10.2.23 to show up at http://tarballs.openstack.org/ci/gerrit/ but don't see it. let me see if i can find build log in any of the jenkins masters | 16:50 |
| *** tjones1 has joined #openstack-infra | 16:50 | |
| *** fawadkhaliq has quit IRC | 16:50 | |
| *** achanda has joined #openstack-infra | 16:50 | |
| *** ddieterly has quit IRC | 16:51 | |
| mordred | pabelanger: yes, that is correct - governance approves such a thing | 16:51 |
| *** fawadkhaliq has joined #openstack-infra | 16:51 | |
| mordred | pabelanger: and then we try to batch up renames because they are a downtime | 16:51 |
| *** tjones1 has quit IRC | 16:51 | |
| pabelanger | mordred, Right, is that batching process documented any place? EG: how does a renamed project from governance get into -infra to be actually renamed? | 16:51 |
| *** tjones1 has joined #openstack-infra | 16:52 | |
| *** ddieterly has joined #openstack-infra | 16:52 | |
| mordred | pabelanger: one sec - I give you link | 16:52 |
| *** derekh has quit IRC | 16:52 | |
| pabelanger | ty | 16:53 |
| mordred | pabelanger: http://ci.openstack.org/gerrit.html#renaming-a-project | 16:53 |
| clarkb | also, if I clone from the /opt/git cache with git git then run zuul-cloner that works | 16:53 |
| clarkb | https://pypi.python.org/pypi/GitPython/1.0.1 released a week ago or so | 16:53 |
| *** ivar-lazzaro has joined #openstack-infra | 16:53 | |
| mordred | pabelanger: I also wrote this: https://review.openstack.org/#/c/175147/ | 16:54 |
| greghaynes | clarkb: wah? if you clone from it then it works? | 16:54 |
| mordred | pabelanger: which si as-of-yet untested | 16:54 |
| mordred | clarkb: so - maybe gitpython bug? | 16:54 |
| clarkb | greghaynes: if I use git to do the cloen first it work | 16:54 |
| mordred | clarkb: oh! was the fetch_master showing up in heads in the ... I get it | 16:54 |
| clarkb | greghaynes: instead of GitPython via zuul-cloner | 16:54 |
| clarkb | mordred: yes | 16:54 |
| mordred | clarkb: I understand what you were verifying now | 16:54 |
| clarkb | mordred: I think GitPython git clone is broken | 16:54 |
| greghaynes | huh | 16:54 |
| mordred | clarkb: that also makes sense that we maybe didnt' see this with greghaynes changes until we removed ggp - because ggp would have done the glone with git giti | 16:55 |
| greghaynes | yep | 16:55 |
| *** jogo is now known as fashgordon | 16:55 | |
| clarkb | mordred: no, different paths completely | 16:55 |
| mordred | okie | 16:55 |
| pabelanger | mordred, Yup, I've been reading that. Guess I'm confused on how the project gets from governance approved, to -infra actually doing the batch. In the past I have seen project renames on the meeting wiki, for weekly discussion. Guess that's the location to notify -infra? | 16:55 |
| clarkb | oh except that ggp did the clone in the past | 16:55 |
| clarkb | right so we replaced it with zuul-cloner yup got it | 16:55 |
| *** mwagner_lap has joined #openstack-infra | 16:55 | |
| *** fawadkhaliq has quit IRC | 16:55 | |
| mordred | pabelanger: yah - so - if someone gets the governance approal, they'll typically also write a change to project-config for the rename | 16:56 |
| *** sputnik13 has joined #openstack-infra | 16:56 | |
| mordred | and we'll notice | 16:56 |
| greghaynes | Also, the gitpython package site has videos of him live coding in case anyone was wondering | 16:56 |
| clarkb | greghaynes: so as a workaround we could add `git clone file:///opt/git/openstack/diskimage-builder openstack/diskimage-builder` and do the same for dib-utils prior to running zuul-cloner | 16:56 |
| mordred | we also may all watch the governance repo :) | 16:56 |
| *** fashgordon is now known as flashgordon | 16:56 | |
| *** alaski is now known as lascii | 16:56 | |
| clarkb | greghaynes: I am also going to try using older GitPython | 16:56 |
| *** freerunner has joined #openstack-infra | 16:57 | |
| pabelanger | mordred, okay, that's what I suspect. I offered up to do the project-config change in this case. Just trying to understand how to actively notify -infra. | 16:57 |
| fungi | yeah, openstack/governance is one of the only non-infra gerrit repos to which i subscribe | 16:57 |
| *** gokrokve has joined #openstack-infra | 16:57 | |
| clarkb | 0.3.7 and 1.0.0 and 1.0.1 all seem to exhibit this behavior | 16:58 |
| *** gokrokve has quit IRC | 16:58 | |
| *** gokrokve has joined #openstack-infra | 16:58 | |
| clarkb | greghaynes: so for immediate future lets make it work I think the do git clone first is a decent work around | 16:58 |
| greghaynes | clarkb: yep, patch inc. | 16:58 |
| *** ashleighfarnham has quit IRC | 16:58 | |
| *** annegentle has joined #openstack-infra | 16:59 | |
| fungi | pabelanger: also, since i think nobody mentioned it yet, generally those proposing the rename link it from https://wiki.openstack.org/wiki/Meetings/InfraTeamMeeting#Upcoming_Project_Renames and then we discuss scheduling in our weekly meetings | 16:59 |
| fungi | pabelanger: that's effectively the step for notifying the infra team that you've got a project to move/rename | 17:00 |
| openstackgerrit | greghaynes proposed openstack-infra/project-config: Clone DIB repos with git before zuul-cloner https://review.openstack.org/181463 | 17:01 |
| greghaynes | clarkb: ^ ty for the debug | 17:01 |
| zaro | clarkb: so the build to package and upload gerrit to tarballs failed downloading something from the internet. https://jenkins07.openstack.org/job/gerrit-package-buck/10/console | 17:02 |
| zaro | clarkb: can we run that again? | 17:02 |
| *** abregman has quit IRC | 17:02 | |
| fungi | zaro: yeah, i'll retrigger it now | 17:02 |
| zaro | fungi: thanks. | 17:02 |
| clarkb | fungi: zaro thanks | 17:03 |
| EmilienM | clarkb, fungi: when you got time, can you look at https://review.openstack.org/#/c/181442/ ? it would help to make progress on beaker work | 17:03 |
| zaro | clarkb: so i'm wondering how the jenkins master are holding up? | 17:03 |
| *** luqas has quit IRC | 17:03 | |
| clarkb | mordred: no screen on ubuntu-minimal, I am sad | 17:04 |
| zaro | clarkb: found an issue with the gearman-plugin but might not be exposed in how we are using. | 17:04 |
| clarkb | zaro: they seem to be doing fine | 17:04 |
| clarkb | zaro: what is the issue? | 17:04 |
| *** whoops has joined #openstack-infra | 17:05 | |
| *** yamahata has joined #openstack-infra | 17:06 | |
| *** Krinkle|detached is now known as Krinkle | 17:06 | |
| fungi | clarkb: https://github.com/larsks/fakeprovide looks like a potential solution to the centos-6 python-setuptools dependencies dilemma | 17:06 |
| mordred | clarkb: perhaps we should add it to template.pp ? | 17:07 |
| clarkb | EmilienM: is it ok if I use that patch as a "why I don't think we should use beaker" example on my spec :P | 17:07 |
| zaro | clarkb: scenario: send many jobs, while jobs are in flight, I send 1 job (offline slave), send more jobs, manually online the slave, gearman-plugin no longer registers functions so no more jobs run. | 17:07 |
| *** Hal has joined #openstack-infra | 17:07 | |
| clarkb | mordred: ya, probably | 17:07 |
| *** Hal is now known as Guest5218 | 17:08 | |
| clarkb | zaro: with which change? | 17:08 |
| EmilienM | clarkb: lol... why that? because of deps? | 17:08 |
| clarkb | EmilienM: because of the ridiculousness of making beaker "run puppet here damnit" | 17:08 |
| openstackgerrit | Merged openstack-infra/zuul: Adding doc reference of ZUUL_CHANGES https://review.openstack.org/176140 | 17:08 |
| mordred | EmilienM: because beaker tries to do to much | 17:08 |
| EmilienM | yeah... | 17:08 |
| zaro | clarkb: with https://review.openstack.org/#/c/179988/ | 17:09 |
| *** dkranz has joined #openstack-infra | 17:09 | |
| clarkb | zaro: its possible there is a bug with that behavior, but I think it existed before | 17:10 |
| clarkb | zaro: I didn't change how we run registerFunctions, I only changed the threading of it | 17:10 |
| clarkb | zaro: in any case we don't do that specific behavior so we should be ok until we fix it | 17:10 |
| fungi | wow, the mailing list archives at osdir.com have just become completely unusable these days. they're infested with "ads" that are social engineering/phishing exploits | 17:11 |
| zaro | clarkb: that's what i thought. i'll continue to debug. | 17:11 |
| pleia2 | fungi: yeah, they've been creeping up to unusuable levels in the past year or so | 17:12 |
| clarkb | zaro: if I had to guess at where that bug is, it has to do with not reseting the functions so we don't detect we need to reregister | 17:13 |
| clarkb | zaro: the child change of that change may fix it actually | 17:13 |
| clarkb | zaro: https://review.openstack.org/#/c/180371/ sorry it isn't parented anymore, but thats the one | 17:13 |
| zaro | clarkb: actually i noticed this with the child change first then check the parent to see if it was there as well. | 17:13 |
| clarkb | zaro: huh, then my hunch is probably wrong. The child is a lot more careful than the old code about registering if it and the server may have gotten out of sync | 17:14 |
| mordred | clarkb, fungi: I am activating US West on the openstackci2 account on hpcloud so that I Can spin up a mirror there | 17:14 |
| fungi | zaro: that retriggered post job has been in the gearman queue for a few minutes now, but that pipeline seems to be running an hour or so backlog so may be a while before it runs | 17:14 |
| clarkb | zaro: the "Ensure we start from scratch on reconnection comments" | 17:15 |
| fungi | mordred: thanks! can't wait to try extending our use to it | 17:15 |
| clarkb | zaro: I will try to reproduce locally now | 17:15 |
| zaro | clarkb: my plan was to switch back to the child change to debug some more. | 17:15 |
| clarkb | zaro: but its run job with offline request, queue jobs after offline, online node to reproduce? | 17:16 |
| *** crc32 has joined #openstack-infra | 17:17 | |
| fungi | zaro: nevermind. i guess bare-trusty workers aren't the holdup, so it's done now http://logs.openstack.org/03/039a1701aeeca922133ae55face73530a164cf5b/post/gerrit-package-buck/994350b,1/console.html | 17:17 |
| zaro | fungi: yup, file is there now. thanks. | 17:17 |
| fungi | np | 17:17 |
| zaro | clarkb: ^ so file is there now. | 17:17 |
| clarkb | zaro: cool, so now we need an up to date change to use that file and we need to ensure review-dev is running with that file? | 17:18 |
| *** teran has joined #openstack-infra | 17:18 | |
| zaro | clarkb: that should repro. let me check that exact scenario since i did have a step initially that runs a bunch of jobs. | 17:19 |
| *** crc32 has quit IRC | 17:20 | |
| *** Krinkle is now known as Krinkle|detached | 17:20 | |
| *** crc32 has joined #openstack-infra | 17:21 | |
| zaro | clarkb: ahh, already did that https://review.openstack.org/#/c/155463 | 17:21 |
| *** annegentle has quit IRC | 17:21 | |
| clarkb | zaro: ok, first attempt to reproduce the gearman-plugin bug failed. I ran one job with offline = 1, my single slave went offline, queued 10 jobs without offline requests, bought node back online and all 10 jobs ran | 17:22 |
| clarkb | zaro: and from wht I can tell it did a RESET_ABILITIES to reset all functions on offline, then CAN_DO'd them back after online so that looked correct to me | 17:23 |
| clarkb | zaro: probably an additional thing we need to reproduce | 17:23 |
| zaro | clarkb: ahh, i believe i did test that scenario before and it did work. try sending this scenario | 17:24 |
| clarkb | zaro: not to keep switching between conversations, but can we make sure we put the war in 155463 on review-dev today? | 17:25 |
| clarkb | zaro: let me know if I can help with ^ | 17:25 |
| zaro | send a bunch of jobs, send a job (offline) while previous jobs are in flight, send another bunch of jobs. bring node back online | 17:25 |
| clarkb | zaro: I can try that | 17:25 |
| greghaynes | fungi: Can I grab a review on the zuul-cloner workaround? https://review.openstack.org/#/c/181463/1 :) | 17:26 |
| zaro | clarkb: that change has already been prepped, https://review.openstack.org/#/c/181143 | 17:26 |
| pabelanger | fungi, perfect, thanks | 17:26 |
| fungi | mordred: clarkb: jeblair: opinions on doing something similar to https://github.com/larsks/fakeprovide to stub out the python-setuptools dependencies on centos-6 nodes? basically just a simple specfile which Provides:python-setuptools but installs no files, and then rpmbuild -bb it and install the result? | 17:26 |
| * zaro goes to run puppet on review-dev | 17:26 | |
| mordred | fungi: shudder what? | 17:27 |
| mordred | fungi: it's possible I haven't been following enough of the problem there | 17:27 |
| fungi | mordred: to deal with rpms which declare a depenency on python-setuptools | 17:27 |
| greghaynes | fungi: it is amazing that tool exists | 17:27 |
| openstackgerrit | Merged openstack-infra/project-config: beaker: support ssh service name on centos https://review.openstack.org/181442 | 17:27 |
| fungi | greghaynes: the "tool" is just a wrapper to make a simple specfile and rpmbuild on it | 17:27 |
| greghaynes | heh, yea | 17:28 |
| mordred | fungi: oh. gross | 17:28 |
| clarkb | ya but its existance implies that rpm/yum/something is broken | 17:28 |
| mordred | fungi: do we have a list of them? | 17:28 |
| mordred | let me guess - something in the dep chain for yum probably | 17:28 |
| fungi | mordred: if you uninstall python-setuptools on centos 6.x, you can't install (or even pre-cache) rpms which depend on it, because it's part of the base package set and so not resolvable through yum repository lists at all | 17:29 |
| mordred | fungi: oh wow | 17:29 |
| mordred | fungi: that's AMAZING | 17:29 |
| *** vponomaryov has quit IRC | 17:29 | |
| fungi | basically they never considered that someone might uninstall that package i guess | 17:29 |
| mordred | of course they didn't | 17:29 |
| mordred | fungi: so - we're treading enough here that it really makes me want to deal with 'infra needs package repo' more quickly | 17:30 |
| openstackgerrit | Khai Do proposed openstack-infra/system-config: Update Gerrit to newer 2.10 build on review-dev.o.o https://review.openstack.org/181143 | 17:30 |
| mordred | fungi: I guess I'm fine with that for the time being - but I really think we should put it in a list of hacks taht we need to resolve | 17:30 |
| clarkb | zaro: ya I am not able to reproduce that gearman-plugin bug. I am testing with the registration update change fwiw | 17:30 |
| fungi | right now devstack has pylint and python-nose in its rpms lists on at least some branches, so on centos-6 i can't pre-cache those because of the dependency yum is unable to resolve | 17:30 |
| mordred | fungi: ok. so the solution to that is simple - remove those | 17:31 |
| clarkb | mordred: I am not sure how infra needs a package repo solves this | 17:31 |
| clarkb | mordred: infra makes a new distro would solve it | 17:31 |
| clarkb | but then we aren't testing centos/rhel | 17:31 |
| mordred | clarkb: because then we could make rpms of modern setuptools for the hosts instead of running get-pip the whole tiem | 17:31 |
| fungi | mordred: i'll see if i can convince the devstack cores to take a patch which does that, possibly on old stable branches | 17:31 |
| *** nikita09 has joined #openstack-infra | 17:31 | |
| clarkb | mordred: this isn't setuptools its requests | 17:31 |
| mordred | fungi: arent' we on branchless devstack? | 17:31 |
| zaro | clarkb: hmm, might be some timing thing. i'm running in debug, hpi:run | 17:32 |
| mordred | clarkb: no it's not | 17:32 |
| fungi | mordred: branchless tempest | 17:32 |
| mordred | fungi: gotcha | 17:32 |
| fungi | devstack is branchybranchy | 17:32 |
| clarkb | ya devstack will likely be branched for ever | 17:32 |
| mordred | clarkb: we're talking about setuptools required by python-nose and pylint | 17:32 |
| *** e0ne_ is now known as e0ne | 17:32 | |
| mordred | clarkb: that's easy - both of those should be pip installed by devstack | 17:32 |
| clarkb | maye I crossed the streams on a couple things, I thought it was python requests | 17:32 |
| fungi | ideally yes | 17:32 |
| zaro | clarkb: i'll see if i can repro again and see how it works when installed on a real master | 17:32 |
| mordred | now - we may still need to do fungi's thing for somethign else | 17:32 |
| mordred | but for that one specific issue, there is a clearer and easier solution :) | 17:33 |
| fungi | mordred: well, i think setuptools is one of those rare cases where we want to uninstall a base package which the distro assumes will always be there | 17:33 |
| clarkb | fungi: and why can't we just uninstall it after we cache? | 17:33 |
| zaro | clarkb: ohh are you using multiple clients to submit the jobs? | 17:33 |
| clarkb | fungi: that seems much simpler to me | 17:33 |
| clarkb | zaro: yes I have 10 clients all submitting jobs to one worker in a loop, then I have an 11th send the offline job request | 17:34 |
| fungi | clarkb: we want to uninstall it before we install pip/setuptools in the instance. so you're talking about running cache-devstack well before our fundamental bootstrapping elements | 17:34 |
| mordred | fungi: yes. I agree - but I also think that if we have a generalized repo structure, we could make a replacement python-setuptools package and have to dance this dance way less | 17:34 |
| *** e0ne has quit IRC | 17:34 | |
| nikita09 | question related to oslo.... does it matter if we have multiple section" [DEFAULT]" declared in our conf file ?? like lets say i define some params under [DEFAULT] at start of file, and later after some other sections, i declare [DEFAULT] again and add more parameters ???? | 17:34 |
| *** koolhead17 has joined #openstack-infra | 17:34 | |
| mordred | nikita09: no idea - we don't really run openstack in here- you're better off asking in #openstack-dev | 17:34 |
| *** teran has quit IRC | 17:34 | |
| mordred | nikita09: although I do not believe iniparser would like if if you did what you suggest | 17:35 |
| fungi | or #openstack | 17:35 |
| mordred | fungi: ++ | 17:35 |
| clarkb | zaro: I have also done manual offlines via the web ui while those 10 clients are running | 17:35 |
| nikita09 | mordred: ok will hop there ..thanks ) | 17:35 |
| * zaro needs to afk, bb little later | 17:35 | |
| fungi | nikita09: see https://www.openstack.org/community#help for good places to get help with running openstack | 17:35 |
| clarkb | fungi: I would have to reread the elements but I don't think there is any reason we have to use the current order | 17:36 |
| *** shashankhegde has joined #openstack-infra | 17:36 | |
| clarkb | fungi: we need to run cache-devstack after we do the git repo stuff, but thats the only dep | 17:36 |
| nikita09 | fungi: thanks..i probably need to go one where doug hellmann can respond :-) thanks for the link again... | 17:36 |
| mordred | fungi: I do believe he's right, fwiw | 17:36 |
| fungi | clarkb: i guess as long as we don't use pip before we cache devstack rpms... | 17:37 |
| *** fawadkhaliq has joined #openstack-infra | 17:37 | |
| fungi | clarkb: does just changing the precedence on that element solve it? | 17:37 |
| clarkb | fungi: I don't think we do | 17:37 |
| *** teran has joined #openstack-infra | 17:37 | |
| clarkb | fungi: ya just chnage the numbers | 17:37 |
| fungi | er, rather on that script within the element | 17:37 |
| mordred | yup | 17:37 |
| *** teran has quit IRC | 17:37 | |
| mordred | it's run-parts | 17:37 |
| greghaynes | you could even move it to pre-install.d | 17:38 |
| greghaynes | since its kind of a pre-install thing | 17:38 |
| clarkb | doesn't that run outside of the chroot? | 17:39 |
| greghaynes | no | 17:39 |
| greghaynes | its the same as install.d, just before | 17:39 |
| clarkb | we do have a thing that runs out side the chroot to make a list of things to cache | 17:39 |
| greghaynes | yes | 17:39 |
| greghaynes | thats before pre-install.d, extra-data.d | 17:39 |
| *** HeOS has joined #openstack-infra | 17:40 | |
| fungi | hrm. we can't remove python-nose from devstack's rpms list because of swift functional testing apparently (according to the comments). i wonder if it can be adapted to use a virtualenv | 17:40 |
| * clarkb points out that unfortunately those words mean nothing about the usefulness of each step. Maybe we should alias "outside-chroot-init.d" to extra-data.d | 17:40 | |
| *** Sukhdev has quit IRC | 17:40 | |
| clarkb | fungi: they may actually be using virtualenv today in the gate at least let me look at that job | 17:41 |
| *** Sukhdev has joined #openstack-infra | 17:41 | |
| clarkb | yup they run tox | 17:41 |
| fungi | okay, so that's just cruft in devstack then | 17:41 |
| clarkb | I think so | 17:41 |
| greghaynes | The assumption was that extra-data.d was used to pull in extra data from the host environment | 17:42 |
| greghaynes | and so far I think thats mostly been true | 17:42 |
| *** annegentle has joined #openstack-infra | 17:43 | |
| *** sandywalsh has left #openstack-infra | 17:44 | |
| clarkb | sort of, its more "make sure the things we want to cache are cached on the host properly so we can pull them in at install time" | 17:44 |
| greghaynes | Is the issue just caching those packages? Maybe you could blacklist them from the list of packages we try to cache? | 17:45 |
| clarkb | greghaynes: we don't want to blacklist them if they are used | 17:45 |
| *** ildikov has quit IRC | 17:45 | |
| greghaynes | hrm | 17:45 |
| fungi | greghaynes: i considered that. at least if we only do it on the one obsolete platform where we're not likely to actually be using them anyway, that's an option but it means even more snowflakeness | 17:45 |
| *** jpeeler has quit IRC | 17:46 | |
| greghaynes | yea, sounds like super-early is the way to go | 17:46 |
| *** sarob has joined #openstack-infra | 17:46 | |
| fungi | though i *am* also proposing a change to devstack to drop nose and pylint from the distro package lists on all branches | 17:46 |
| *** abregman has joined #openstack-infra | 17:47 | |
| *** teran has joined #openstack-infra | 17:47 | |
| *** teran has quit IRC | 17:47 | |
| *** teran has joined #openstack-infra | 17:48 | |
| *** teran has quit IRC | 17:48 | |
| clarkb | ya pylint is an odd one to have in devstack | 17:48 |
| *** fitoduarte has joined #openstack-infra | 17:48 | |
| mordred | people want it to exist because they use devstack nodes as dev nodes | 17:49 |
| mordred | for local dev | 17:49 |
| clarkb | mordred: ya but all pylinting should be tox'd | 17:49 |
| mordred | clarkb: TOTALLY | 17:49 |
| mordred | I agree it sholdnt' be there | 17:49 |
| * greghaynes pokes mordred for +A on https://review.openstack.org/#/c/181463/1 | 17:50 | |
| greghaynes | I really want to get those *-minimal elements tested | 17:50 |
| greghaynes | (check out https://review.openstack.org/#/c/181162/ BTW) | 17:51 |
| fungi | oh wow, juno devstack actually has python-setuptools in distro package lists | 17:51 |
| *** aduarte has quit IRC | 17:52 | |
| mordred | greghaynes: ooh! | 17:52 |
| mordred | fungi: that's because bonghits | 17:52 |
| *** esker has quit IRC | 17:52 | |
| clarkb | mordred: you know, outside observers may think that is why we insist on the circular chair arrangment during summits | 17:53 |
| *** ildikov has joined #openstack-infra | 17:53 | |
| greghaynes | yes, passing formation | 17:53 |
| *** otter768 has joined #openstack-infra | 17:54 | |
| *** HeOS has quit IRC | 17:57 | |
| mordred | :) | 17:57 |
| mordred | clarkb, fungi: ok. pypi.region-a.geo-1.openstack.org has been created, and I've added mirror.$region.openstack.org entries for everything | 17:58 |
| mordred | I use A records and not CNAMES so that we could eventually switch the reverse as well and ditch the pypi records | 17:58 |
| clarkb | mordred: cool, have you sent a quota bump request for the openstackjenkins2 account in the west region? | 17:58 |
| clarkb | mordred: I can help do maths for that if it helps | 17:58 |
| openstackgerrit | Merged openstack-infra/elastic-recheck: Update query for bug 1403291 https://review.openstack.org/181150 | 17:58 |
| mordred | clarkb: not an official one - I sent one in an email - I'll follow up with a quota request | 17:58 |
| openstack | bug 1403291 in neutron "VM's fail to receive DHCPOFFER messages" [Critical,Fix released] https://launchpad.net/bugs/1403291 - Assigned to Kevin Benton (kevinbenton) | 17:58 |
| *** gokrokve has quit IRC | 17:58 | |
| mordred | clarkb: well, my request is "please make it match" | 17:59 |
| mordred | but maybe that's silly | 17:59 |
| clarkb | mordred: iirc they get really sad if you don't give them numbers for all values | 17:59 |
| mordred | clarkb: you don't happen to have a script/doc from when you did the neutron setup in east do you? | 17:59 |
| *** otter768 has quit IRC | 17:59 | |
| clarkb | mordred: but maybe that changed. IIRC we can't make the networks until that happens because default quota is too small | 17:59 |
| mordred | awesome | 17:59 |
| clarkb | mordred: I don't have a script because I clicked buttons in horizon :/ | 17:59 |
| mordred | awesome | 17:59 |
| clarkb | mordred: that is honestly the easiest way to use neutron | 17:59 |
| mordred | I will make a script this time | 17:59 |
| mordred | nah - I got it in python/ansible just fine these days :) | 18:00 |
| *** ildikov has quit IRC | 18:00 | |
| mordred | shade has made me learn many things I probably didn't want to know | 18:00 |
| clarkb | I guess, what horizon does that is nice is forces you to do all the things you have to do that are otherwise undocumented, and forces you to do them in the right order that is otherwise undocumented | 18:00 |
| *** alexpilotti has quit IRC | 18:01 | |
| clarkb | but its basically, make subnet, make network, make dhcp range, make router, connect router to netowkr and external net | 18:01 |
| clarkb | mordred: one thing to consider is using disjoint network ranges so that if we ever overlap the two we don't run into routing issues | 18:02 |
| *** alexpilotti has joined #openstack-infra | 18:02 | |
| clarkb | but maybe thats premature optimization | 18:03 |
| openstackgerrit | Russell Bryant proposed openstack-infra/project-config: Update jenkins jobs for networking-ovn https://review.openstack.org/181487 | 18:03 |
| *** claudiub has quit IRC | 18:04 | |
| dtroyer | clarkb: planning reasonable address usage is never premature… | 18:04 |
| * dtroyer one watched two companies merge networks that both used 10/8 | 18:04 | |
| openstackgerrit | Merged openstack-infra/project-config: Clone DIB repos with git before zuul-cloner https://review.openstack.org/181463 | 18:04 |
| *** viktors is now known as viktors|afk | 18:06 | |
| fungi | https://review.openstack.org/#/q/Ibd16ac5,n,z is the devstack patch and associated backports | 18:07 |
| fungi | but i'll also adjust out dib precedence on package caching to try to work around it as well | 18:07 |
| mordred | dtroyer: *shudder* | 18:07 |
| *** alexpilotti has quit IRC | 18:07 | |
| *** koolhead17 has quit IRC | 18:07 | |
| fungi | dtroyer: i once worked at a company which had undergone that. it had address-translating bridges (not routers, bridges) at interfaces between the networks | 18:08 |
| fungi | essentially we rewrote the source and destination addresses on the ip protocol headers of frames as they moved through those bridges | 18:09 |
| clarkb | mordred: maybe give east 10.0.0.0/16 and wast 10.1.0.0/16 ? did I do that math correctly? | 18:09 |
| clarkb | oh except that I think we use 10.1.0.0/24 in the gate? | 18:09 |
| clarkb | fun times | 18:09 |
| dtroyer | yeah, always get that second octet above 5 or so ;) | 18:09 |
| clarkb | dtroyer: well we stuck with 10.0.0.0/22 or so in east beacuse thats what hpcloud gave us before they used neutron | 18:10 |
| fungi | it taught me to assign rfc-1918 networks out of 172.<some random integer between 16 and 31> | 18:10 |
| clarkb | so we knew those ranges would work | 18:10 |
| mordred | how about 10.1{1,2,3,4,5}.0.0/24 ? | 18:10 |
| clarkb | mordred: that should work | 18:11 |
| *** jpeeler has joined #openstack-infra | 18:11 | |
| EmilienM | "Unable to locate package libxslt-dev" on Trusty now (beaker jobs) | 18:12 |
| EmilienM | is that because we have DIB now? | 18:12 |
| mordred | shouldn't be - that shoudl be a package | 18:12 |
| dtroyer | clarkb: right…that's default-land for sooo many things, including devstack and more than one openstack distro…so you can't run default devstack on default openstack in most cases :( | 18:12 |
| EmilienM | do | 18:12 |
| clarkb | EmilienM: you likely need an apt-get update prpr to installing anything | 18:12 |
| EmilienM | https://jenkins02.openstack.org/job/gate-puppet-keystone-puppet-beaker-rspec-dsvm-trusty/1/console | 18:12 |
| fungi | EmilienM: you'll need to apt-get install that in the job if it's devstack-trusty | 18:12 |
| mordred | yah. you need apt-get update | 18:13 |
| fungi | EmilienM: oh, right, you're getting that while trying to apt-get install | 18:13 |
| fungi | def need to update first | 18:13 |
| EmilienM | ah cool cool | 18:13 |
| EmilienM | I'm scare of clarkb saying it's because of beaker now. | 18:13 |
| clarkb | EmilienM: its not :) its due to apt caching this stuff | 18:13 |
| fungi | i've heard that all the world's troubles are because of beaker | 18:13 |
| clarkb | but I have no idea why puppet can't do this for itslef | 18:14 |
| clarkb | s/puppet/puppet or beaker/ | 18:14 |
| fungi | prof. honeydew's experiments would have gone so much more smoothly were it not for beaker | 18:14 |
| *** coolsvap has quit IRC | 18:15 | |
| dtroyer | some say beaker was a victim not an antagonist | 18:15 |
| *** freerunner has quit IRC | 18:15 | |
| EmilienM | clarkb: the beaker job on centos crashed: Unexpected termination of the channel | 18:15 |
| EmilienM | https://jenkins04.openstack.org/job/gate-puppet-keystone-puppet-beaker-rspec-dsvm-centos7/1/console | 18:15 |
| fungi | dtroyer: yes, but those people clearly deny science | 18:15 |
| *** ildikov has joined #openstack-infra | 18:15 | |
| clarkb | EmilienM: your puppet probably broke networking then | 18:16 |
| EmilienM | well, i'm installing a package. | 18:16 |
| clarkb | EmilienM: it may have gotten further than that when it crashed but the buffered data didn't get to jenkins | 18:16 |
| EmilienM | actually I'm deleting a package | 18:16 |
| clarkb | oh you are messing withssh right? | 18:16 |
| clarkb | thats what jenkins uses to talk to the slave and may affect stuff | 18:16 |
| fungi | yeah, jenkins will freak out if sshd stops working | 18:17 |
| EmilienM | no, I'm sudo yum -y remove facter puppet | 18:17 |
| *** mrunge has quit IRC | 18:17 | |
| clarkb | EmilienM: thats what the log is telling you but it ma be lying | 18:17 |
| clarkb | EmilienM: this is a buffered communication channel, if the channel dies before the buffer is flushed you won't see what it was actually doing | 18:17 |
| EmilienM | at least beaker don't lie | 18:17 |
| clarkb | well its sort of the only way to do this sort of communication otherwise you get a billion 1 byte packets | 18:18 |
| mordred | well, yeah - because you restart ssh to get that config change - I believe that's going to make the inbound connection unhappy | 18:18 |
| clarkb | mordred: at least on ubuntu that should be fine | 18:18 |
| clarkb | mordred: but I have no idea what crazy things centos might do to protect its users | 18:18 |
| EmilienM | well. We did all of that *before* | 18:18 |
| *** esker has joined #openstack-infra | 18:18 | |
| fungi | yeah, sshd restarts don't terminate open sockets | 18:18 |
| fungi | at least not usually | 18:19 |
| EmilienM | before https://review.openstack.org/#/c/175434/ - we didn't have any issue | 18:19 |
| clarkb | well you never ran on centos before | 18:19 |
| fungi | i just did `sudo service sshd restart` on a held centos6 node and it didn't disconnect my session, fwiw | 18:20 |
| EmilienM | fungi: we did that before on ubuntu | 18:20 |
| fungi | EmilienM: centos 6, centos 6, centos 6 | 18:21 |
| fungi | ubuntu is irrelevant in this case | 18:21 |
| EmilienM | ok | 18:21 |
| clarkb | fungi: I think this is centos7 | 18:21 |
| fungi | oh | 18:21 |
| fungi | i missed that. could be different there. i can fond out real quick | 18:21 |
| clarkb | and runs `sudo systemctl restart sshd` | 18:21 |
| openstackgerrit | Emilien Macchi proposed openstack-infra/project-config: puppet/beaker: run apt-get update before install deps https://review.openstack.org/181491 | 18:22 |
| *** ssam2 has quit IRC | 18:23 | |
| clarkb | EmilienM: I am +2 on ^ should fix trusty | 18:23 |
| clarkb | I need to pop out for a short bit, back to hopefully finish up the gerrit prep | 18:24 |
| fungi | yeah, restarting on centos7 doesn't faze it either | 18:24 |
| clarkb | fungi: huh | 18:24 |
| EmilienM | :( | 18:24 |
| fungi | even if i `sudo systemctl stop sshd;sleep 10;sudo systemctl start sshd` i just get a hung socket for 10 seconds and then my session resumes | 18:24 |
| fungi | openssh is amazingly resilient | 18:24 |
| *** rwsu_ has quit IRC | 18:24 | |
| clarkb | so its possible that its something to do with the other processes | 18:24 |
| clarkb | could we be OOMing? | 18:24 |
| clarkb | probably need to hold a centos7 node then go step by step so we avoid any buffering to find the exact thing that explodes | 18:25 |
| EmilienM | maybe just the VM crashing? | 18:25 |
| clarkb | EmilienM: ya could be | 18:25 |
| clarkb | kernel panic/OOM something like that | 18:25 |
| fungi | if we hold it before it crashes, we could probably see from the virtual console what's happened there | 18:25 |
| EmilienM | clarkb: is that happen often here? | 18:25 |
| clarkb | I can help with that in a bit if we need to do that | 18:25 |
| EmilienM | or am I alone today? | 18:25 |
| clarkb | EmilienM: yes centos7 has been a source of OOM trouble but only after installing and running openstack with latest pyopenssl | 18:26 |
| nibalizer | ya it looks like a vm crashing to me | 18:26 |
| clarkb | EmilienM: so this is somewhat new | 18:26 |
| EmilienM | I'll run a recheck once https://review.openstack.org/181491 is merged | 18:26 |
| fungi | afaik we have hardly any jobs gating on centos 7 | 18:26 |
| EmilienM | fungi: most of them on ubuntu I presume | 18:27 |
| fungi | yeah, or centos 6 | 18:27 |
| fungi | another possibility is iptables blocking ssh access | 18:27 |
| fungi | though it would need to terminate established connections as well if that's the cause | 18:28 |
| EmilienM | fungi: can you +A https://review.openstack.org/181491 please ? | 18:28 |
| EmilienM | I'll run recheck to see if it happens again | 18:28 |
| clarkb | greghaynes: have you rerun the dib tests since the workaround? | 18:28 |
| greghaynes | clarkb: nope, I was giving it a bit to puppet | 18:28 |
| greghaynes | clarkb: just kicked off a recheck | 18:29 |
| greghaynes | clarkb: https://jenkins01.openstack.org/job/gate-dib-dsvm-functests-devstack-trusty/5/console not sure how to tell what cloud thats running in | 18:30 |
| greghaynes | clarkb: oh, rax :( | 18:30 |
| clarkb | at the very beginning it says "Building remotely on..." | 18:30 |
| *** Krinkle|detached is now known as Krinkle | 18:30 | |
| pabelanger | So, for the people playing along at home, here is what a sample grafana dashboard looks like in yaml: https://github.com/pabelanger/grafyaml/blob/master/tests/schema/fixtures/dashboard-0001.yaml | 18:30 |
| pabelanger | going to be super simple to create dashboards | 18:31 |
| *** annegentle has quit IRC | 18:31 | |
| *** Guest5218 has quit IRC | 18:31 | |
| greghaynes | clarkb: https://jenkins01.openstack.org/job/gate-dib-dsvm-functests-devstack-trusty/6/console success | 18:32 |
| clarkb | greghaynes: awesome, now to figure out why it doesn't work in general :/ | 18:32 |
| *** Hal has joined #openstack-infra | 18:33 | |
| greghaynes | clarkb: yea, sounds like repro'ing should be easyish | 18:33 |
| *** Hal is now known as Guest62673 | 18:33 | |
| fungi | clarkb: we run install_puppet.sh at a precedence of 05, so should i drop cache-devstack's download-pkgs from 50 to 04, or raise the puppet element's puppet script to something higher? | 18:35 |
| fungi | or does it matter? | 18:36 |
| clarkb | fungi: there is also prepare-node which runs as 20 iirc | 18:36 |
| clarkb | fungi: also, I just realized we may be relying on install_puppet for git? | 18:36 |
| *** ddieterly has quit IRC | 18:36 | |
| clarkb | so I think we want install_git, cache-devstack, install-puppet, prepare-node | 18:37 |
| fungi | oh, possibly. we're relying on it for get-pip too | 18:37 |
| greghaynes | clarkb: source-repositories pulls in git | 18:37 |
| *** Longgeek has quit IRC | 18:37 | |
| clarkb | greghaynes: on the image side? | 18:37 |
| *** Longgeek has joined #openstack-infra | 18:37 | |
| greghaynes | clarkb: yes | 18:37 |
| clarkb | oh then I think if we shift to 04 that should work | 18:37 |
| greghaynes | it happens at install.d/00- | 18:38 |
| greghaynes | (package-installs.yamls) | 18:38 |
| clarkb | fungi: ^ that should at least be the minimal change to confirm it works | 18:38 |
| fungi | clarkb: i'm testing the theory now, just making sure there's nothing you know of i'm missing | 18:38 |
| clarkb | fungi: the only additional pieces were prepare node at 20 and potential git dep | 18:38 |
| greghaynes | how do youall get-pip? | 18:39 |
| greghaynes | in the dib images | 18:39 |
| *** Longgeek has quit IRC | 18:39 | |
| fungi | also concerned that when i add bindep into this stage, i'm going to need to make sure that the step which creates the bindep virtualenv happens before the step that caches packages | 18:39 |
| greghaynes | I have a todo item of moving the tripleo get-pip element into DIB proper | 18:39 |
| *** marun has joined #openstack-infra | 18:39 | |
| fungi | greghaynes: it happens as a part of install_puppet.sh | 18:39 |
| *** Longgeek has joined #openstack-infra | 18:39 | |
| greghaynes | ah, ok | 18:39 |
| *** fawadkhaliq has quit IRC | 18:39 | |
| mordred | if anybody else wants to go jump on voting on a defcore test thing about glance task-import: | 18:40 |
| mordred | https://review.openstack.org/#/c/181280/ | 18:40 |
| fungi | which, oh, if we rely on one script to both install pip and remove distro setuptools, i can't pip install bindep and then generate and pre-cache a list of packages between those | 18:40 |
| *** Krinkle is now known as Krinkle|detached | 18:40 | |
| greghaynes | mordred: ooo | 18:40 |
| fungi | so this may need some more serious refactoring/splitting up | 18:40 |
| greghaynes | fungi: yea, DIB + a ./install-all-the-stuff leads to that :/ | 18:41 |
| clarkb | fungi: ya sounds like we will have to split things up to have much better control over order | 18:41 |
| fungi | worse, we need to remove distro setuptools before installing pip, but i need pip to install bindep before removing setuptools. catch-22 | 18:41 |
| openstackgerrit | Merged openstack-infra/project-config: puppet/beaker: run apt-get update before install deps https://review.openstack.org/181491 | 18:41 |
| fungi | this is becoming un-fun | 18:41 |
| clarkb | wait | 18:41 |
| clarkb | why do we need to remove distro setuptools before installing pip? | 18:41 |
| clarkb | is it purely to avoid pip and yum from fighting? | 18:42 |
| mordred | yes | 18:42 |
| fungi | if we install setuptools after installing pip, the package manager (on at least some platforms) will remove our pip-installed setuptools | 18:42 |
| fungi | er, if we remove setuptools after installing pip, i mean | 18:42 |
| *** armax has quit IRC | 18:43 | |
| fungi | so anyway, we're back to needing other solutions for pre-caching rpms which depend on setuptools on centos-6 | 18:43 |
| *** Guest62673 has quit IRC | 18:43 | |
| clarkb | does yum have an ignore missing deps or just cache setuptools but don't install it flag? | 18:44 |
| clarkb | seems like when doing a download only it shouldn't care about what is installed | 18:45 |
| *** craigbr has joined #openstack-infra | 18:45 | |
| fungi | worth noting, centos/rhel 6.x is the only platform on which we uninstall distro setuptools before installing pip | 18:45 |
| fungi | clarkb: there are some fun mailing list threads on that topic | 18:45 |
| fungi | clarkb: yum's development "philosophy" is that it should be hard, if not impossible, for a user to override dependency relationships and get broken packages, even intentionally | 18:46 |
| clarkb | at install time that seems fine | 18:46 |
| clarkb | but if I just want you to download as many of the packages I think I need as possible why not just do it? yay yuim | 18:46 |
| clarkb | I wonder if we can just use zypper instead | 18:46 |
| greghaynes | haha | 18:47 |
| fungi | unfortunately yum install --downloadonly seems to exercise the entirely same codepath as without --downloadonly, and just skips the install step at the end | 18:47 |
| greghaynes | I wonder if we can use yumdownloader instead | 18:47 |
| greghaynes | and download the rpm's into the cache | 18:47 |
| greghaynes | it might end up with the same issue though... | 18:48 |
| greghaynes | maybe print a manifest and wget? | 18:48 |
| fungi | after some insistence from the community, there is apparently now a --skip-broken option which tells it to just not install things whose dependencies can't be resolved | 18:49 |
| greghaynes | oh hey | 18:49 |
| fungi | but that means we'll end up not caching things depending on python-setuptools (which is ultimately probably fine since nothing's going to be able to install them there anyway?) | 18:49 |
| *** armax has joined #openstack-infra | 18:49 | |
| clarkb | we should just use nix | 18:50 |
| clarkb | there I said it, made terrible idea to fix all the problems at once and introduce a whole host of new ones :) | 18:51 |
| fungi | my primary concern with doing --skip-broken is that it has the possibility to mask other problems too. but maybe we just don't care | 18:51 |
| clarkb | fungi: I certainly don't care much about centos6 as it is going away in the not distant future | 18:51 |
| *** ddieterly has joined #openstack-infra | 18:51 | |
| clarkb | its got what ~6 months left? | 18:51 |
| fungi | yep, give or take | 18:52 |
| clarkb | so ya lets do the simplest thing here, I thought it was simple to change order but it isn't so maybe we do larsks' thing | 18:52 |
| * larsks is guilty of horrible awful hacks. | 18:53 | |
| fungi | ooh, switching the package caching precedence earlier seems to work, though now we're at a new problem. grubby fails to find a suitable template to update the boot configuration | 18:53 |
| clarkb | fungi: the more we dig the more I want to say "no centos-6" :) | 18:53 |
| *** tnovacik has quit IRC | 18:54 | |
| fungi | larsks: well, i don't actually think that's a terrible hack. it's a standard solution in debianland (via the equivs package) to create dummy packages which provide some virtual package name satisfying dependencies you've worked around on your own | 18:54 |
| *** fawadkhaliq has joined #openstack-infra | 18:54 | |
| larsks | fungi: That is certainly an more pleasant phrasing :) | 18:54 |
| fungi | when mixing and matching between multiple package managers, it's a relatively sane way to manually bridge some dependencies between them | 18:55 |
| clarkb | fungi: ya for the general case I agree, I just can't get over that we are only asking to download and cache the package | 18:56 |
| clarkb | fungi: whcih should never care | 18:56 |
| fungi | clarkb: ultimately, i think the fault lies in the fact that you can't yum install packages from the base set if you ever yum erase them | 18:57 |
| fungi | though i guess it's also possible we're just missing a channel for them? | 18:57 |
| fungi | there is a "base" repo which we seem to already have enabled | 19:02 |
| EmilienM | clarkb, fungi : sshd restart works | 19:07 |
| EmilienM | https://jenkins05.openstack.org/job/gate-puppet-keystone-puppet-beaker-rspec-dsvm-centos7/1/console | 19:07 |
| EmilienM | now I have sudo error I'm debugging | 19:07 |
| EmilienM | the good news, it's consistent with trusty (got the same output) | 19:07 |
| EmilienM | the bad news, I did not have this error before (when using trusty too but another image) | 19:07 |
| clarkb | EmilienM: its removing your sudo privs | 19:08 |
| clarkb | EmilienM: then you try to sudo | 19:08 |
| clarkb | EmilienM: likely you just need to not remove sudo | 19:08 |
| EmilienM | oh | 19:08 |
| fungi | EmilienM: the revoke-sudo builder in your job definition | 19:09 |
| EmilienM | I can trop revoke-sudo ? | 19:09 |
| EmilienM | drop* | 19:09 |
| EmilienM | I prefer to move the gem install | 19:09 |
| EmilienM | to be consistent with what we had before | 19:09 |
| fungi | yeah, you can remove revoke-sudo or move it after whatever builder needs sudo | 19:10 |
| *** smarcet has quit IRC | 19:10 | |
| openstackgerrit | Emilien Macchi proposed openstack-infra/project-config: beaker/puppet: install bundle before revoking sudo https://review.openstack.org/181506 | 19:11 |
| dhellmann | jeblair, fungi, clarkb: The Oslo team is interested in setting up a blog. How do you feel about hosting a statically generated thing using git, gerrit, and sphinx and publishing it to http://docs.openstack.org/developer/ somewhere like we do with documentation? | 19:11 |
| clarkb | dhellmann: I am fine with it, annegentle may have other input though | 19:11 |
| fungi | dhellmann: arguably the vmt already does something similar with security.openstack.org and advisories | 19:11 |
| openstackgerrit | David Shrewsbury proposed openstack-infra/shade: Update images API for get/list/search interface https://review.openstack.org/180600 | 19:11 |
| dhellmann | clarkb: maybe a different hostname? how hard is that? | 19:11 |
| clarkb | dhellmann: basically you would just need a repo with the thing to build then add the doc publishing job to it | 19:11 |
| dhellmann | clarkb: yep, I have the toolchain all lined up, I was asking as a matter of policy | 19:12 |
| clarkb | dhellmann: would require setting that up, which isn't terrible just time | 19:12 |
| clarkb | DNS, vhost, file copying credentials etc | 19:12 |
| *** salv-orlando has quit IRC | 19:12 | |
| *** ajmiller has quit IRC | 19:13 | |
| dhellmann | ok, so I'll see what annegentle thinks and then come back and talk about details if she prefers a different hostname and figure out next steps | 19:13 |
| *** Swami has quit IRC | 19:13 | |
| *** soren has joined #openstack-infra | 19:13 | |
| *** ajmiller has joined #openstack-infra | 19:13 | |
| dtroyer | fungi: are those 4 devstack reviews order dependent? the nose/pylint ones? | 19:14 |
| fungi | dtroyer: i don't expect so, but am not sure yet since i haven't looked at test results | 19:14 |
| dtroyer | the master one failed the postgres job, the rest passed | 19:15 |
| fungi | yeah, my guess is it failed on something unrelated but i'll dig in and see | 19:15 |
| *** ashleighfarnham has joined #openstack-infra | 19:15 | |
| fungi | failed one test, tempest.api.compute.servers.test_delete_server.DeleteServersTestJSON.test_delete_active_server | 19:16 |
| dtroyer | yeah, looks unrelated | 19:16 |
| fungi | something tells me removing pylint and nose didn't trigger that | 19:16 |
| mtreinish | fungi: heh, I dunno | 19:18 |
| *** Somay has joined #openstack-infra | 19:19 | |
| zaro | clarkb: review-dev is down because i turned puppet on. | 19:19 |
| clarkb | zaro: ok? | 19:20 |
| *** freerunner has joined #openstack-infra | 19:20 | |
| zaro | clarkb: we ready to update the gerrit lib change? | 19:20 |
| clarkb | zaro: it needs reviewers | 19:20 |
| *** crc32 has quit IRC | 19:21 | |
| zaro | ok. hope it's ok for it to stay down until then | 19:21 |
| openstackgerrit | Davanum Srinivas (dims) proposed openstack/requirements: Allow Magnum to conform to global requirements process https://review.openstack.org/170830 | 19:21 |
| clarkb | zaro: for review-dev it should be fine, but we should get some reviewes | 19:22 |
| clarkb | reviewers can we get reviews for https://review.openstack.org/#/c/181173/ and its parent? | 19:22 |
| fungi | the gerritlib patch is only needed to support jeepyb's manage_projects.py which will only fire on updates to its projects list, right? | 19:22 |
| clarkb | hrm I thought I had approved that one | 19:23 |
| clarkb | or maybe it was the jeepyb update? | 19:23 |
| clarkb | fungi: what change are you referring to? | 19:24 |
| openstackgerrit | David Shrewsbury proposed openstack-infra/shade: Stop leaking server objects https://review.openstack.org/181513 | 19:24 |
| fungi | clarkb: zaro mentioned a change to gerritlib, and that was the only one i could remember | 19:25 |
| clarkb | fungi: oh, its the gerrit review_site/lib changes above I link to that he is referring to | 19:25 |
| fungi | oh, wait, he said "gerrit lib" not "gerritlib" | 19:25 |
| fungi | yeah, i'm reviewing those now | 19:25 |
| clarkb | a | 19:25 |
| fungi | okay, confusion averted | 19:25 |
| clarkb | I have not noop appliedthsoe changes but cando so if people are generally happy with them | 19:26 |
| clarkb | but first lunch | 19:26 |
| fungi | they look fine, but we should keep an eye on puppetboard | 19:26 |
| *** mwagner_lap has quit IRC | 19:26 | |
| clarkb | basic idea is if gerrit < 2.10 then do what we do today if greater do what we want to do tomorrow | 19:26 |
| fungi | right | 19:27 |
| fungi | and so should be safe to merge now | 19:27 |
| clarkb | review-dev should exercise the new side code and review.o.o the old side code when merged | 19:27 |
| clarkb | fungi: yup *should* be :) | 19:27 |
| fungi | i'm around in case i turn out to be wrong on that front ;) | 19:27 |
| *** HeOS has joined #openstack-infra | 19:27 | |
| zaro | yup, we gonna merged it? | 19:28 |
| fungi | wondering if we're doing some mucking around with grub.conf and blowing away the template stanza that crntos-6 grubby is expecting to find | 19:28 |
| zaro | ohh cool, looks like it's on its way. thanks fungi | 19:29 |
| fungi | though the only explicit modifications to grub.conf that i'm seeing in our custom elements come at precedence 99, long after yum update runs | 19:30 |
| fungi | oho! that seems to not be what's causing the failure though | 19:32 |
| fungi | `yum -y install --downloadonly foo` exits nonzero on centos-6 and that element is running under errexit | 19:32 |
| EmilienM | beaker: this one should be the last patch in -infra : https://review.openstack.org/#/c/181506/ | 19:33 |
| *** marcusvrn has quit IRC | 19:33 | |
| *** dustins has quit IRC | 19:33 | |
| fungi | i should clarify, exits nonzero once it successfully does the things you ask it to do. craaaaaazy | 19:33 |
| fungi | it's basically saying "i failed to install packages because you asked me not to, so exit 1" | 19:33 |
| *** hodos has quit IRC | 19:35 | |
| openstackgerrit | Merged openstack-infra/puppet-gerrit: Update bouncy castle libs for Gerrit 2.10+ https://review.openstack.org/172534 | 19:36 |
| *** salv-orlando has joined #openstack-infra | 19:38 | |
| openstackgerrit | Merged openstack-infra/puppet-gerrit: Get libs before running Gerrit https://review.openstack.org/181173 | 19:38 |
| *** Sukhdev has quit IRC | 19:41 | |
| clarkb | fungi: hahahaha | 19:43 |
| clarkb | fungi: so it just needs an || true and we are good? | 19:43 |
| *** pabelanger has quit IRC | 19:44 | |
| *** whoops has quit IRC | 19:44 | |
| fungi | clarkb: well, a release-specific ||true since i think centos7 and f12 probably behave more sanely | 19:45 |
| clarkb | fungi: gotcha | 19:45 |
| fungi | and we're not doing any release detection (yet) in that script | 19:45 |
| clarkb | zaro: so once we have review-dev running with those just merged changes, what is next? I think the remaining hcnages have to wait for tomorrow ya? so probably just reviews? | 19:45 |
| fungi | i'm testing other yum-using platforms now to confirm | 19:45 |
| *** Swami has joined #openstack-infra | 19:46 | |
| zaro | clarkb: yes, i concur. the only other change that can probably be approved now is https://review.openstack.org/#/c/173052/ | 19:48 |
| openstackgerrit | Dan Prince proposed openstack-infra/tripleo-ci: Enable pacemaker via the resource registry https://review.openstack.org/181522 | 19:48 |
| clarkb | zaro: will that work before we upgrade? | 19:49 |
| zaro | clarkb: it's got a depends-on so won't merge until the dependency goes thru? | 19:50 |
| clarkb | correct, but if we approve it now we just have to reapprove it later anyways | 19:50 |
| clarkb | so I think we are fine with the current state | 19:50 |
| openstackgerrit | Dan Prince proposed openstack-infra/tripleo-ci: Enable pacemaker via the resource registry https://review.openstack.org/181522 | 19:50 |
| clarkb | though maybe that should be cleaned up to use a release appropriate version? | 19:51 |
| clarkb | this is actualyl harder in puppet than I would like :/ would make sense to have packages here | 19:51 |
| openstackgerrit | Russell Bryant proposed openstack-infra/project-config: Update jenkins jobs for networking-ovn https://review.openstack.org/181487 | 19:54 |
| zaro | clarkb: ahh, forgot about that. agreed. | 19:55 |
| *** otter768 has joined #openstack-infra | 19:55 | |
| zaro | clarkb: i was able to repro that gearm-plugin issue. it only happens when using multiple executors on a node. | 19:57 |
| openstackgerrit | Merged openstack-infra/project-config: beaker/puppet: install bundle before revoking sudo https://review.openstack.org/181506 | 19:57 |
| zaro | clarkb: i assume you have your slave setup with only 1 executor? | 19:57 |
| *** dprince has quit IRC | 19:57 | |
| clarkb | zaro: yes only one executor | 20:00 |
| *** otter768 has quit IRC | 20:00 | |
| clarkb | zaro: I can add a second executor but gearman plugin only does one at a time iirc | 20:00 |
| *** notnownikki has quit IRC | 20:00 | |
| clarkb | zaro: with 2 executors and manual offline online that seems to work | 20:01 |
| zaro | clarkb: you mean we only use 1 exec per node but gearman should be able to handle more correct? | 20:02 |
| clarkb | zaro: no, I mean I am pretty sure that gearman only uses 1 executor | 20:02 |
| clarkb | zaro: I have enable concurrent builds set, but only one job at a time runs | 20:03 |
| zaro | clarkb: hmm, well when i submit multiple jobs (no offline) gearman run them all on multiple executors on jenkins and returns results jusft fine | 20:03 |
| clarkb | even if I have >1 executor | 20:03 |
| clarkb | and I am submitting jobs from 10 clients | 20:03 |
| zaro | ohh you need to restart the plugin after updating jenkins executors | 20:04 |
| clarkb | ok let me restart jenkins | 20:04 |
| zaro | there is no event for that in jenkins | 20:04 |
| zaro | not jenkins, just restart the plugin | 20:04 |
| clarkb | ah ok I see it running two at a time now | 20:05 |
| *** Somay has quit IRC | 20:06 | |
| clarkb | manual offline online is still working | 20:06 |
| *** packet has joined #openstack-infra | 20:06 | |
| *** ddieterly has quit IRC | 20:06 | |
| openstackgerrit | Davanum Srinivas (dims) proposed openstack-dev/pbr: Advertise support for Python3.4 https://review.openstack.org/181525 | 20:06 |
| zaro | clarkb: i think i see what's happening. the single job to offline runs on the 1st exec (brings down the slave) but the other job is already in the queue to run on the 2nd exec | 20:07 |
| clarkb | I have submitted the offline rquest job | 20:07 |
| clarkb | ok that reproduces, so you have to use the offline node request | 20:08 |
| clarkb | zaro: I am not sure this is a new bug | 20:08 |
| clarkb | definitely something to look into fixing, but we only do one executor per slave so we should be fine | 20:09 |
| zaro | clarkb: yeah, i think so i'll enter an issue into storyboard so i can remember. | 20:10 |
| *** Sukhdev has joined #openstack-infra | 20:11 | |
| EmilienM | clarkb: I'm working on Puppet design sessions scheduling, do you have a session about Puppet/Beaker/containers thing? maybe we should organize something together | 20:12 |
| zaro | clarkb: maybe this should go in now? https://review.openstack.org/#/c/181143/ | 20:14 |
| *** rmcall has joined #openstack-infra | 20:14 | |
| fungi | zaro: lgtm | 20:18 |
| *** signed8b_ has joined #openstack-infra | 20:19 | |
| *** ddieterly has joined #openstack-infra | 20:19 | |
| *** signed8bit has quit IRC | 20:20 | |
| *** tiswanso has quit IRC | 20:21 | |
| jeblair | clarkb, fungi, mordred: back | 20:22 |
| clarkb | EmilienM: I don't but sounds like a decnet idea | 20:22 |
| *** sdake_ has joined #openstack-infra | 20:22 | |
| jeblair | fungi: if i followed scrollback correctly, we're going the devstack route to resolve the python-setuptools thing, right? | 20:22 |
| clarkb | EmilienM: I am trying not to make too many things dependent on me just in case I don't end up going to the summit | 20:22 |
| *** banix has joined #openstack-infra | 20:23 | |
| jeblair | EmilienM: i'm getting ready to work on infra scheduling; let's keep in touch :) | 20:23 |
| EmilienM | jeblair: sorry | 20:23 |
| EmilienM | I'm confused who is PTL | 20:23 |
| jeblair | ouch | 20:23 |
| EmilienM | jeblair: yeah. Friday. | 20:24 |
| *** radez is now known as radez_g0n3 | 20:24 | |
| *** ddieterly has quit IRC | 20:24 | |
| openstackgerrit | Darragh Bailey proposed openstack-infra/jenkins-job-builder: Enable tests of builder object https://review.openstack.org/181541 | 20:24 |
| openstackgerrit | Darragh Bailey proposed openstack-infra/jenkins-job-builder: Test update command with --delete-old option https://review.openstack.org/181542 | 20:24 |
| jeblair | clarkb, greghaynes: was there a resolution to the zuul-cloner thing? | 20:24 |
| EmilienM | jeblair: so yeah, do you schedule a session about that already? | 20:24 |
| EmilienM | oh did not read your last comment | 20:25 |
| EmilienM | I'm available when you like to talk about that, cool | 20:25 |
| *** annegentle has joined #openstack-infra | 20:25 | |
| fungi | jeblair: i think so, for the moment (solution in devstack is already getting +2'd) | 20:25 |
| clarkb | jeblair: I found a workaround, if you use git to do the initial clone and have zuul-cloner update then it works fine | 20:25 |
| banix | anteaya: emagana At some point you were working on creating a multi-node setup for certain tests at the gate. Is that effort ongoing? Concluded? Essentially, I am wondering what the status of that work you were leading is. | 20:25 |
| *** annegentle has quit IRC | 20:26 | |
| *** sdake has quit IRC | 20:26 | |
| pelix | zaro: https://review.openstack.org/181542 adds tests to the area that the parallel threads code in JJB broke and shows the issue when you cherry-pick the patch onto it once you adjust the function names being mocked to align with the changes. | 20:26 |
| clarkb | jeblair: I think thee is a bug in GitPython but hasn't followed it much further than to check that GitPython 0.3.7 1.0.0 and 1.0.1 all do the same thing | 20:26 |
| jeblair | EmilienM: we're all pretty empowered around here, we operate a bit like a collective. I'm the PTL which means I don't get to do the fun stuff. | 20:26 |
| *** annegentle has joined #openstack-infra | 20:26 | |
| jeblair | clarkb: yeah, so using git to do the clone isn't a good solution for us | 20:26 |
| EmilienM | jeblair: heh, we try to have in collective too | 20:26 |
| clarkb | jeblair: ya its definitely subpar | 20:27 |
| jeblair | clarkb: i think we should do both things: i think dib elements should not leave the weird ref there; and i think zuul-cloner should protect itself from this. i think z-c may be able to do that by resetting HEAD after cloning. | 20:27 |
| clarkb | jeblair: but basically GitPython seems to copy refs/heads from the source rather than making those refs/remotes/$remote/things | 20:27 |
| EmilienM | jeblair: having a session together about Puppet/Beaker would make sense | 20:27 |
| dhellmann | fungi, clarkb : do we do anything specially when we set up nodes to run pypy jobs to ensure that pip is upgraded? This error looks suspicious, and I can't reproduce it locally. http://logs.openstack.org/76/179476/3/check/gate-oslo.middleware-pypy/9800dd6/console.html#_2015-05-08_13_58_31_918 | 20:28 |
| *** hodos|2 has joined #openstack-infra | 20:28 | |
| jeblair | EmilienM: i think we may have a fishbowl slot open for that | 20:30 |
| clarkb | dhellmann: no its basically just tox -epypy | 20:30 |
| clarkb | dhellmann: which won't be affected by the system wide pip upgrade | 20:30 |
| jeblair | clarkb, nibalizer: does an infra+puppet fishbowl to talk about testing (containers?), etc, seem usefil? | 20:30 |
| EmilienM | jeblair: my schedule is pretty open. Though we could have it on Tuesday morning before 11am that would be a good slot maybe | 20:31 |
| dhellmann | clarkb: ok, thanks | 20:31 |
| fungi | dhellmann: i think if something depends on pip and tox is set to do a pip install -U then pip gets upgraded, otherwise you run with the version bundled in virtualenv | 20:31 |
| clarkb | jeblair: I think if we go in with the plan to get some jobs/tests running that could be useful | 20:31 |
| EmilienM | jeblair: we have a room all Tuesday from 11am to 6pm so maybe we could stay in your before 11am to talk about that | 20:31 |
| greghaynes | jeblair: Agreed on fixing both things | 20:31 |
| clarkb | jeblair: fwiw z-c doesn't need to protect itself if gitpython worked | 20:32 |
| clarkb | jeblair: git is supposed to make this work for you by putting those refs in refs/remotes | 20:32 |
| clarkb | jeblair: but GitPython seems to mix refs/heads and refs/remotes when cloning locally | 20:32 |
| jeblair | clarkb: okay, maybe we fix three things? :) | 20:32 |
| greghaynes | heh | 20:32 |
| jeblair | clarkb: fishbowl is traditional design summit session, so more talking less working. | 20:33 |
| jeblair | clarkb: does that mean you think it would be better to have a work session and not a fishbowl? | 20:33 |
| *** e0ne has joined #openstack-infra | 20:33 | |
| clarkb | jeblair: oh then yes | 20:33 |
| dhellmann | fungi: I think you're right, and it is definitely pip 1.5.4. I guess I assumed we were doing something to force update virtualenv. | 20:33 |
| jeblair | clarkb: do we have a plan? | 20:34 |
| clarkb | jeblair: there are the two proposed specs, if we can get agreement on them soonish then we can treat those as the basis of some placns | 20:34 |
| clarkb | jeblair: biggest question for functional testing is what to use to test, openstack-puppet says beaker so maybe we should just more forward with that | 20:34 |
| asselin | greghaynes, hi....if you're not too busy, I'd like to go back and fix this: https://review.openstack.org/#/c/177043/1/bin/disk-image-create | 20:35 |
| clarkb | jeblair: and in the other case I didn't see much disagreement with my comments that we use ansible and containers for segregating services seem to be general happyness | 20:35 |
| nibalizer | jeblair: I would love to be in that session, but unfortunately I can't attend the summit | 20:35 |
| nibalizer | talking about testing would be good for the project | 20:35 |
| *** mriedem is now known as mriderman | 20:35 | |
| nibalizer | having crinkle and EmilienM and Hunner is probably enough that you won't really need me | 20:35 |
| dhellmann | hmm, I wonder if this works for me because I have a wheel of mcabe built | 20:35 |
| jeblair | nibalizer: if you're available remotely we may be able to call you | 20:35 |
| nibalizer | jeblair: ++ | 20:36 |
| nibalizer | I have the sip thing working with the asterisk server | 20:36 |
| pelix | clarkb: recall mention of a having a way to specify default values for params in JJB? https://review.openstack.org/180117 | 20:36 |
| greghaynes | asselin: ah, yep. I kind if wanted https://review.openstack.org/#/c/177043/ to merge which would fix it... | 20:36 |
| EmilienM | nibalizer: well, you could be more than useful | 20:36 |
| jeblair | clarkb: yeah, i'm just wondering if we think we can get agreement on everything before the summit, or should we use a fishbowl slot at the summit to get agreement | 20:36 |
| asselin | greghaynes, ok...I honestly, don't completely follow the discussion, but I will run it now anyway to see if it at least 'works'. | 20:37 |
| *** julim has quit IRC | 20:37 | |
| greghaynes | oh, that is the cahnge you linked :p I thought you were linking the nodepool change | 20:37 |
| *** esker has quit IRC | 20:38 | |
| clarkb | jeblair: I agreement is largely up to mordred | 20:38 |
| greghaynes | I might have to poke some people to get reviews on it since im guessing the -1 is making other cores overlook it | 20:38 |
| clarkb | jeblair: I am fine using beaker if that is what openstack-puppet wants to do | 20:38 |
| asselin | greghaynes, ok, so I'm clean. i'll pull that change, pip install it, then build an image using which version of nodepool? master, or before the nodepool change that hardcoded the value? | 20:38 |
| clarkb | and I think we largely have agreement on the other spec | 20:38 |
| asselin | s/clean/clear/ | 20:38 |
| greghaynes | asselin: master | 20:38 |
| *** kaisers has quit IRC | 20:38 | |
| *** freerunner has quit IRC | 20:40 | |
| emagana | banix: I just saw the note about multi-node environment | 20:40 |
| openstackgerrit | Doug Hellmann proposed openstack-infra/project-config: remove pypy job for oslo.middleware https://review.openstack.org/181547 | 20:40 |
| jeblair | EmilienM: tuesday is the cross-project day, and looking at the schedule, i think it will be very hard for some of us to join you in your room then. :( https://libertydesignsummit.sched.org/overview/type/design+summit/Cross+Project+workshops#.VU0gg-SVtph | 20:40 |
| emagana | banix: https://review.openstack.org/#/c/106495/ never went through | 20:40 |
| *** fawadkhaliq has quit IRC | 20:41 | |
| EmilienM | fungi, clarkb: thx for you rhelp on beaker today. The jobs are okay now, and I'm pushing patches on our modules for centos support. | 20:41 |
| clarkb | pelix: nice | 20:41 |
| *** kaisers has joined #openstack-infra | 20:41 | |
| jeblair | EmilienM: i think some of the more puppet-focused of us can join you, but some of us with a more project-overall focus may be locked in those rooms all day | 20:41 |
| EmilienM | jeblair: no problem, we can make it when you like | 20:41 |
| clarkb | banix: emagana: it was unnecessary | 20:41 |
| clarkb | banix: emagana we already had multinode testing for both neutron and nova net at that time iirc | 20:41 |
| clarkb | banix: emagana the jobs exist, they runagainst neutron experimental because they do not pass | 20:42 |
| asselin | greghaynes, pip install is failing...http://paste.openstack.org/show/217388/ | 20:42 |
| banix | emagana: clarkb thanks for the info. | 20:42 |
| emagana | clarkb: well, this is a very old proposal and anteaya was not aware of any multi-node and this is why she was supporting this work | 20:43 |
| clarkb | banix: emagana sure, I'm just saying it has happened | 20:43 |
| emagana | clarkb: The first patch is July 2014 | 20:43 |
| clarkb | we test DVR with multinode upstream | 20:43 |
| emagana | clarkb: got it! | 20:43 |
| anteaya | at the time the spec was written the solution didn't exist | 20:43 |
| emagana | clarkb: BTW. I was just replying to banix's question | 20:44 |
| asselin | greghaynes, nevermind...updated setuptools and it works now | 20:44 |
| banix | clarkb: So just for my understanding if I have a test that wnat to run on say 4 nodes (let’s assume it’s such a great test that there is no objection to what it does) say selectively for certain patches, I can do that now? | 20:44 |
| anteaya | I think we have 2 node ability | 20:45 |
| anteaya | one all in one and an extra compute node | 20:45 |
| clarkb | anteaya: correct | 20:46 |
| clarkb | https://git.openstack.org/cgit/openstack-infra/devstack-gate/tree/multinode_diagram.txt is what the network looks like | 20:46 |
| *** armax has quit IRC | 20:48 | |
| banix | anteaya: clarkb thanks very informative. So how the two nodes are setup are predetermined. For example I cannot bring the nodes up with neutron and ovs plugin (ml2 mech driver) using vlan. Is that correct? | 20:48 |
| clarkb | banix: its neutron with ovs, I don't know anything about VLANs but I think it should work | 20:49 |
| clarkb | banix: the only gotcha would be that the floating ip range would have to be untagged so that tempest can hit it | 20:49 |
| fungi | we should probably get https://review.openstack.org/181143 in so we can validate that war sooner rather than later | 20:50 |
| clarkb | fungi: approiving now | 20:50 |
| openstackgerrit | Steve Leon proposed openstack-infra/project-config: Add devstack gate job for Cue project https://review.openstack.org/180774 | 20:51 |
| fungi | i assume it will start gerrit back on review-dev when it gets applied? | 20:51 |
| jeblair | EmilienM, clarkb, nibalizer: i think we have room for a fishbowl session for this; if we have full agreement then, we can just use it to make sure everyone understands the plan; if not, we can discuss some more. | 20:51 |
| banix | clarkb: anteaya and emagana thanks again. I may bug you some more if I find you in Vancouver. | 20:51 |
| EmilienM | jeblair: TBH, I don't think someone will argue | 20:51 |
| clarkb | banix: is there something specific you are trying to accomplish? | 20:52 |
| nibalizer | jeblair: awesome | 20:52 |
| clarkb | banix: we are testing DVR multinode now | 20:52 |
| clarkb | banix: it would be great if neutron could get the job working though | 20:52 |
| jeblair | EmilienM, clarkb, nibalizer: asselin also proposed a 2-hour workroom block on puppet-openstackci and included testing in his description of what he would like to discuss/get accomplished; so i think that will touch on the subject as well and maybe get some things accomplished. | 20:52 |
| openstackgerrit | Doug Hellmann proposed openstack-infra/project-config: remove pypy job for oslo libraries https://review.openstack.org/181547 | 20:52 |
| openstackgerrit | Darragh Bailey proposed openstack-infra/jenkins-job-builder: Test update command with --delete-old option https://review.openstack.org/181542 | 20:52 |
| anteaya | banix: bug me when clarkb is around, we will make better progress | 20:53 |
| nibalizer | ya i worry about a workroom block on openstackci, since I don't think we have the patterns for that yet, so parallelising doesn't exactly makes sense, however getting the pattern set would be very valuable | 20:53 |
| openstackgerrit | Emilien Macchi proposed openstack-infra/project-config: beaker/puppet: use 2 different nodesets https://review.openstack.org/181555 | 20:53 |
| emagana | banix: anytime! | 20:53 |
| EmilienM | crinkle: like this ? ^^^ | 20:54 |
| jeblair | nibalizer: yeah, i don't think we'll come out of the openstackci workroom with patches, but maybe a plan for paralellization | 20:54 |
| anteaya | nibalizer: one of the purposes is to help onboard some people who want to help but don't have a clear sense how to begin | 20:54 |
| *** mriderman has quit IRC | 20:54 | |
| dhellmann | fungi, clarkb: it looks like the pypy job is failing for a bunch of oslo libs, so we're planning to just drop those jobs for now. Can you look at https://review.openstack.org/181547 to unblock us, when you have a minute, please? | 20:54 |
| anteaya | nibalizer: if they participate in the conversation it is possible they have a stronger motivation to test and write patches | 20:54 |
| asselin | and remain involved longer-term | 20:55 |
| anteaya | asselin: ++ | 20:55 |
| clarkb | banix: http://logs.openstack.org/79/181379/3/experimental/check-tempest-dsvm-neutron-multinode-full/9022ca3/console.html is a random job I pulled out of logstash | 20:55 |
| banix | clarkb: yes but I am not sure if it is something of value or not and whether my Neutron friends find it useful. So I want to discuss with them first. The specific thing I am interested in is whether having certain tests namely performance tests (beyond functional tests) can be integrated in our infrastructure and used selectively on certain patches which may have performance impact. WHether the gate is the right place for this you know bett | 20:56 |
| banix | just thinking along those lines. | 20:56 |
| asselin | nibalizer, jeblair but I do want to prioritize time on issues. Perhaps we can split that 2hour block into 2 1hour blocks? | 20:56 |
| clarkb | dhellmann: so are we not testing those because we don't care about pypy or because they fail? | 20:56 |
| asselin | not sure if that makes a difference or not | 20:56 |
| clarkb | dhellmann: if we are removing because of the failures we should consider working a fix first? | 20:56 |
| clarkb | dhellmann: or maybe going nonvoting? | 20:56 |
| dhellmann | clarkb: both | 20:56 |
| jeblair | asselin: do you think it will be useful to have EmilienM and the puppet folks in there to work through how to test it? | 20:56 |
| clarkb | banix: I am not sure how mutlinode and dvr is related to performance testing | 20:57 |
| dhellmann | clarkb: we had a handful of libs testing with pypy, but as many not (maybe more?) | 20:57 |
| asselin | jeblair, it seems testing is a bigger topic so would be good to get us all in sync on clarkb's proposal | 20:57 |
| dhellmann | clarkb: pypy support was a half-finished project from a while back, iirc, and the person driving it isn't around any more | 20:57 |
| clarkb | dhellmann: right I guess I am trying to figure out if we should make them nonvoting or if we should remove them | 20:58 |
| nibalizer | getting testing right makes big refactors like openstackci much safer | 20:58 |
| clarkb | dhellmann: if we generally don't care and no one will be working on a fix then lets remove them. But if we do care then we can make them nonvoting while others make them work? | 20:58 |
| dhellmann | clarkb: I'm not sure? I don't particularly care about pypy support, so I thought removing would free up those test nodes for things we do care about | 20:58 |
| banix | clarkb: I know the multinode was targeted for testing dvr but in the use case I have (not related to dvr) multinode is used to test certain aspects of network performance. | 20:59 |
| asselin | nibalizer, ++ and important for long-term stability | 20:59 |
| dhellmann | clarkb: but non-voting works if you know of someone who is interested in fixing them | 20:59 |
| *** mriedem has joined #openstack-infra | 20:59 | |
| clarkb | banix: gotcha, in generall I Would not use this setup for network performance testing. We have to create a logical l2 overlay with gre to make it work and performance is likely not great | 20:59 |
| anteaya | can I offer the proposal to move them to experiemental? | 20:59 |
| clarkb | dhellmann: I don't know of anyone beyond maybe lifeless? I assume lifeless is too busy to give that effort right now | 21:00 |
| jeblair | asselin, clarkb, EmilienM, nibalizer: i think we have two reasonable choices: A) fishbowl for puppet testing and 25% of puppet-openstackci workroom on that topic, or B) no fishbowl for testing (i'd give it over to kolla to talk about dockerhub) and 50% of the puppet-openstackci workroom focusing on puppet-testing | 21:00 |
| anteaya | would there be any way to evaluate after 2 weeks how many times soneone has bothered to trigger them? | 21:00 |
| *** rm_work|away is now known as rm_work | 21:00 | |
| jeblair | asselin, clarkb, EmilienM, nibalizer: have a preference, or alternate suggestion? | 21:00 |
| dhellmann | clarkb: assuming lifeless is busy seems like a reasonable assumption :-) | 21:00 |
| clarkb | anteaya: you mean the neutron dvr job? we can check logstash for runs, its already in experimental | 21:00 |
| anteaya | clarkb: sorry pypy | 21:00 |
| lifeless | clarkb: dhellmann: ECONTEXT | 21:00 |
| jeblair | asselin, clarkb, EmilienM, nibalizer: (also, i believe we may have some sprint room for puppet-openstackci) | 21:00 |
| banix | clarkb: I see. makes sense. thanks for the advice. | 21:00 |
| openstackgerrit | Merged openstack-infra/system-config: Update Gerrit to newer 2.10 build on review-dev.o.o https://review.openstack.org/181143 | 21:00 |
| EmilienM | jeblair: B looks cool | 21:01 |
| *** mrmartin has quit IRC | 21:01 | |
| dhellmann | lifeless: the pypy jobs on some oslo libs are broken and preventing non-code changes from landing; Alex isn't around; I don't care about pypy particularly so am not interested in fixing them myself; do you care enough for us to keep running the jobs non-voting or should we remove them? | 21:01 |
| clarkb | jeblair: I think it mostly comes down to how opinionated mordred wants to be. If we can get general agreement on the specs that are up then more time in workroom is my preference | 21:02 |
| lifeless | oh :( | 21:02 |
| clarkb | jeblair: if not then likely fishbowl to hash out the details is best | 21:02 |
| clarkb | mordred: ^ if yuo are around have ou started to settle on a process for puppet testing? | 21:02 |
| lifeless | so - I like the idea of pypy support. But it needs to be bottom up (as in community supported) | 21:02 |
| lifeless | one person caring is not a community :( | 21:02 |
| *** e0ne has quit IRC | 21:02 | |
| *** signed8bit has joined #openstack-infra | 21:02 | |
| dhellmann | lifeless: http://logs.openstack.org/76/179476/3/check/gate-oslo.middleware-pypy/9800dd6/console.html#_2015-05-08_13_58_31_918 | 21:02 |
| clarkb | lifeless: you had previously mentioned interest in pypy so I brought your name up :) | 21:03 |
| dstufft | Openstack should probably deploy onto PyPy as the recommended deployment tbh but D: | 21:03 |
| clarkb | dstufft: last time anyone really tested pypy with openstack they said it was slower | 21:03 |
| greghaynes | once nova supports it ;) | 21:03 |
| clarkb | dstufft: so, I dunno doesn't seem worth it | 21:03 |
| greghaynes | yea, eventlet | 21:03 |
| dhellmann | lifeless: right, I agree. I was willing to leave the jobs running while they were working, and I'd be happy for someone else to fix them, I'm just not interested enough to do it myself. I'm +0 on pypy support. | 21:03 |
| lifeless | why does that have pip 1.5.4 ? | 21:03 |
| dstufft | clarkb: how did they test it? and did they raise a bug with PyPy? | 21:03 |
| clarkb | dstufft: vishy wrote a blog on it | 21:03 |
| lifeless | yeah | 21:04 |
| lifeless | so | 21:04 |
| clarkb | dstufft: let me find it | 21:04 |
| greghaynes | dstufft: https://gist.github.com/vishvananda/89f0d6775ac335a824d5 | 21:04 |
| dhellmann | lifeless: no idea :-/ | 21:04 |
| *** bcourt has quit IRC | 21:04 | |
| greghaynes | dstufft: The key is at the bottom update | 21:04 |
| fungi | lifeless: it's the pip vendored in that release of virtualenv i think | 21:04 |
| clarkb | fungi: correct | 21:04 |
| lifeless | dstufft: http://logs.openstack.org/76/179476/3/check/gate-oslo.middleware-pypy/9800dd6/console.html#_2015-05-08_13_58_31_918 <- that rings a bell for me | 21:05 |
| *** abregman has quit IRC | 21:05 | |
| dhellmann | fungi: but why is an old virtualenv being used there? how do we get a new one on other nodes? | 21:05 |
| asselin | jeblair, when you mention 25%,50%, is that based on 1 hr or 2hrs? | 21:05 |
| lifeless | (the mETADATA file missing) | 21:05 |
| jeblair | asselin: 2 hours | 21:05 |
| clarkb | dhellmann: I want to say it is the version of virtualenv that tox uses? | 21:05 |
| openstackgerrit | Emilien Macchi proposed openstack-infra/project-config: beaker/puppet: use 2 different nodesets https://review.openstack.org/181555 | 21:05 |
| lifeless | dhellmann: so I think - ask on the -dev list, give a deadline of tuesday morning | 21:06 |
| dhellmann | lifeless: I found https://bugs.launchpad.net/openstack-ci/+bug/1315468 earlier | 21:06 |
| openstack | Launchpad bug 1315468 in OpenStack Core Infrastructure "*-pypy Failure: "No such file or directory: '/home/jenkins/workspace/gate-requirements-pypy/.tox/pypy/site-packages/setuptools-2.2.dist-info/METADATA"" [Undecided,Fix released] | 21:06 |
| *** signed8b_ has quit IRC | 21:06 | |
| dstufft | greghaynes: oof, yea if it was using a jit-less PyPy that's gonna be super slow | 21:06 |
| dhellmann | lifeless: well, the oslo team doesn't care about this and it's only our libs, so... | 21:06 |
| *** ddieterly has joined #openstack-infra | 21:06 | |
| clarkb | I take that back we pin virtualenv on our slaves | 21:06 |
| clarkb | probably because lol setuptools | 21:06 |
| dhellmann | clarkb: the version of tox is the newest, though? | 21:06 |
| lifeless | dhellmann: its not a matter of don't care | 21:06 |
| greghaynes | dstufft: yep | 21:06 |
| dhellmann | ah, yeah | 21:06 |
| lifeless | dhellmann: I do care, I can't context switch onto it on saturday morning | 21:07 |
| lifeless | this is family time | 21:07 |
| clarkb | dstufft: greghaynes it may be that things have changed but we are a long way away from even suggesting pypy should be the default | 21:07 |
| greghaynes | totally | 21:07 |
| dhellmann | lifeless: sure, I'm not asking you to -- if you care, I will make them non-voting for now instead of removing them | 21:07 |
| greghaynes | Theres definitely some things there that are less of an issue now | 21:07 |
| clarkb | virtualenv 1.11.4 is what we install | 21:07 |
| lifeless | dhellmann: and, is everyone that cares about oslo +pypy here right now? I'm saying we might well get more cares if we cast a wider net | 21:07 |
| greghaynes | I kind of want to replicate it now.. | 21:07 |
| dhellmann | lifeless: I don't need you to fix them today, just help me decide which approach to take to unblock the 3-4 libs that are blocked now | 21:07 |
| lifeless | dhellmann: ack, ok | 21:08 |
| clarkb | we can try lifting that | 21:08 |
| jeblair | clarkb, fungi, mordred, asselin, EmilienM: draft schedule at top of https://etherpad.openstack.org/p/infra-liberty-summit-planning | 21:08 |
| lifeless | dhellmann: so I think pypy is good to have | 21:08 |
| *** begal-sop__ has quit IRC | 21:08 | |
| lifeless | dhellmann: non-voting + a call for assistance IMO | 21:08 |
| dhellmann | lifeless: ok, I can do that | 21:08 |
| dstufft | greghaynes: I think it would be useful, I'm pretty sure that pypy for openstack would be a good thing if all the kinks are worked out. THat was using a beta version of PyPy 2.0, PyPy is up to uh, 2.5.1 or so now | 21:08 |
| *** lascii is now known as alaski | 21:08 | |
| *** eharney has quit IRC | 21:09 | |
| clarkb | jeblair: looks reasonable to me | 21:09 |
| clarkb | dstufft: 2.2.1 is what we would get on trusty | 21:09 |
| *** Longgeek has quit IRC | 21:10 | |
| *** mriedem has quit IRC | 21:11 | |
| fungi | jeblair: that proposed schedule seems fine to me | 21:11 |
| *** mriedem has joined #openstack-infra | 21:12 | |
| fungi | jeblair: as for docker registry vs puppet testing, i lean toward the latter | 21:12 |
| openstackgerrit | Doug Hellmann proposed openstack-infra/project-config: make pypy jobs non-voting for oslo libraries https://review.openstack.org/181547 | 21:13 |
| openstackgerrit | Clark Boylan proposed openstack-infra/system-config: Use latest current virtualenv https://review.openstack.org/181558 | 21:13 |
| dhellmann | lifeless, fungi, clarkb, dimsum__ : ^^ | 21:14 |
| EmilienM | jeblair: looks good to me, thanks (schedule) | 21:14 |
| clarkb | so ^ is totally not self testing, but we can give it a go | 21:14 |
| jeblair | yeah, i think the docker thing could be handled in an email or spec (probably "we would like to do this." "sure, go for it, here's what you need to know about the system"). | 21:14 |
| *** emagana has quit IRC | 21:14 | |
| fungi | right. doesn't need "an audience with the infra team" for that idea to be proposed | 21:15 |
| asselin | jeblair, I don't know how contentious puppet testing is going to be, but it'd be good to get everyone in agreenment and moving forward. | 21:15 |
| *** ldnunes has quit IRC | 21:16 | |
| fungi | asselin: i expect it to need a fair amount of bouncing design ideas back and forth, so suitable for fishbowling | 21:16 |
| clarkb | dhellmann: +2 | 21:16 |
| dhellmann | clarkb: ty | 21:16 |
| asselin | fungi, jeblair then we should do the fishbowl. now is the time | 21:16 |
| *** abregman has joined #openstack-infra | 21:17 | |
| *** rmcall has quit IRC | 21:17 | |
| jeblair | who wants to lead a puppet testing fishbowl? | 21:18 |
| clarkb | you can put me down with the caveat that I won't know if I am summiting until lsat minute | 21:19 |
| clarkb | appoitment is thursday | 21:19 |
| pleia2 | I'd like to voluntell nibalizer for that | 21:19 |
| clarkb | train saturday :) cutting it really close | 21:19 |
| jeblair | clarkb: okay, ill do that, and if you aren't there, i'll act as "moderator" :) | 21:19 |
| fungi | i'm happy to be backup if there are miniature people preventing clark's attendance | 21:19 |
| jeblair | pleia2: nibalizer can't attend :( | 21:19 |
| pleia2 | doh | 21:20 |
| jeblair | but we'll try to phone him in | 21:20 |
| pleia2 | yeah, definitely | 21:20 |
| anteaya | nibalizer: :( | 21:20 |
| clarkb | though I may know sooner as we have a different appointment on momday | 21:20 |
| clarkb | *monday | 21:20 |
| clarkb | day after momday | 21:20 |
| *** pfallenop has quit IRC | 21:21 | |
| pleia2 | hehe | 21:21 |
| *** ildikov has quit IRC | 21:21 | |
| *** pfallenop has joined #openstack-infra | 21:21 | |
| *** ddieterly has quit IRC | 21:22 | |
| EmilienM | clarkb, jeblair, crinkle and I will represent Puppet OpenStack community obviously | 21:24 |
| jeblair | ttx: i tried to add http://libertydesignsummit.sched.org/event/6dac7819d75ce432ecacca66010e3280#.VU0q9uSVtph to "Release Management" but that doesn't seem to be working | 21:25 |
| clarkb | mordred: greghaynes I have responded to comments on https://review.openstack.org/#/c/179840/ | 21:25 |
| *** sdake_ has quit IRC | 21:26 | |
| greghaynes | clarkb: ah, ok | 21:26 |
| *** pc_m has quit IRC | 21:26 | |
| *** peristeri has quit IRC | 21:28 | |
| mordred | clarkb: ok. gross | 21:28 |
| *** signed8bit has quit IRC | 21:29 | |
| pleia2 | my test server decided it had had enough with the dust in my office, but it's alive again now (woo canned air), running through nodepool instructions again | 21:29 |
| *** banix has quit IRC | 21:30 | |
| *** rwsu has joined #openstack-infra | 21:30 | |
| mordred | \o/ | 21:30 |
| jeblair | mordred: i think we decided to do this puppet-testing fishbowl http://libertydesignsummit.sched.org/event/42aaba40ddeda529fd027376da654cf2#.VU0sPOSVtpg does that sound good to you? | 21:30 |
| *** rwsu has quit IRC | 21:30 | |
| *** rwsu has joined #openstack-infra | 21:31 | |
| pleia2 | my office doesn't seem dusty, but I guess that's because it all gets sucked into my computers :D | 21:31 |
| mordred | jeblair: well... fwiw, I think we may be converging in clarkb's spec ... so it may be possible to use that slot for somethign else more contentious if you need the slot | 21:31 |
| *** armax has joined #openstack-infra | 21:31 | |
| openstackgerrit | Merged openstack-infra/project-config: make pypy jobs non-voting for oslo libraries https://review.openstack.org/181547 | 21:31 |
| clarkb | mordred: though https://review.openstack.org/#/c/181555/2 has me hating the world again | 21:32 |
| clarkb | mordred: we basically have one hack atop another atop another atop another to make beaker useable | 21:32 |
| jeblair | mordred: i don't think there is anything more contentious, so i think maybe we use it to make sure everyone is on the same page | 21:32 |
| mordred | jeblair: ok | 21:32 |
| mordred | clarkb: yes. I agree | 21:32 |
| mordred | clarkb: I think that what I currently agree with is "we and openstack-puppet should use the same thing" - especially since they have the hacky harness already | 21:33 |
| mordred | and then maybe someone who rubies, like crinkle, can go fix beaker | 21:33 |
| clarkb | mordred: yes I think we should use the same thing | 21:33 |
| mordred | so that we can get rid of the hacks | 21:33 |
| *** pfallenop has quit IRC | 21:34 | |
| *** ildikov has joined #openstack-infra | 21:34 | |
| *** pfallenop has joined #openstack-infra | 21:35 | |
| *** BobBall is now known as BobBall_AWOL | 21:36 | |
| *** gordc has quit IRC | 21:36 | |
| *** esker has joined #openstack-infra | 21:36 | |
| asselin | greghaynes, anything obvious to you? /usr/local/bin/../share/diskimage-builder/lib/common-functions: line 186: dib-run-parts: command not found http://paste.openstack.org/show/217440/ | 21:37 |
| greghaynes | asselin: you dont have dib-utils installed | 21:38 |
| greghaynes | asselin: or, its not in your path | 21:38 |
| fungi | clarkb: zaro: the tidy change isn't working the way we expect on review-dev | 21:38 |
| crinkle | mordred: clarkb my feeling on this is that puppet-openstack was without functional testing for too long and we converged on this thing called beaker, and at this point turning around to choose something else would be more costly than dealing with whatever dumbness beaker brings up | 21:39 |
| fungi | clarkb: zaro: we have 2x bcpkix, bcprov and mysql-connector-java | 21:39 |
| asselin | greghaynes, Requirement already up-to-date: dib-utils in /usr/local/lib/python2.7/dist-packages | 21:39 |
| crinkle | mordred: clarkb but if infra chooses to use something else i don't think anyone is really hurt by that | 21:39 |
| *** rwsu has quit IRC | 21:39 | |
| clarkb | fungi: wait is tidy wow | 21:39 |
| greghaynes | asselin: so it musnt be in your path | 21:39 |
| greghaynes | asselin: how are you running dib? | 21:39 |
| clarkb | fungi: its prepending them all with . and "backing" them up | 21:40 |
| asselin | greghaynes, nodepool image-build dpc | 21:40 |
| fungi | clarkb: oh, so it is | 21:40 |
| greghaynes | asselin: and disk-image-create is magically in your path? | 21:40 |
| fungi | that's rather silly | 21:41 |
| clarkb | fungi: incredibly silly | 21:41 |
| openstackgerrit | Merged openstack-infra/project-config: beaker/puppet: use 2 different nodesets https://review.openstack.org/181555 | 21:41 |
| asselin | greghaynes, how do you check? | 21:41 |
| *** erlon has quit IRC | 21:41 | |
| fungi | clarkb: but more to the point it also seems to not be cleaning out those now-unmanaged symlinks | 21:41 |
| asselin | greghaynes, http://paste.openstack.org/show/217446/ | 21:42 |
| fungi | clarkb: because in addition to the backed-up symlinks we have non-backed-up symlinks in there along with the downloaded libs | 21:42 |
| zaro | clarkb, fungi : ohh, ok that should work too. but very odd | 21:42 |
| greghaynes | asselin: can you which disk-image-create? | 21:43 |
| clarkb | fungi: a its definitely not doing what we want | 21:43 |
| zaro | fungi, clarkb : ohh wait, the .backup thing is my doing. | 21:43 |
| asselin | /usr/local/bin/disk-image-create | 21:43 |
| zaro | i did that for testing the new libs. so that's not from tidy. tidy just doens't seem to be working at all? | 21:44 |
| fungi | zaro: ahh, okay. well, regardless the tidy doesn't seem to have removed the symlinks we don't want | 21:44 |
| zaro | it didn't git rid of *-jdk15on-*.jar either | 21:44 |
| greghaynes | asselin: and is there a /usr/local/bin/dib-run-parts? | 21:44 |
| asselin | greghaynes, no | 21:45 |
| fungi | zaro: ohh, those aren't the ones you're downloading in the manifest? | 21:45 |
| greghaynes | asselin: wierd. I just did a pip install dib-utils and it put one there. | 21:45 |
| greghaynes | so, no idea whats different about your setup | 21:46 |
| asselin | greghaynes, ok...will uninstall and reinstall | 21:46 |
| zaro | fungi: the ones that get downloaded get renamed | 21:46 |
| greghaynes | maybe youre in a venv or something? | 21:46 |
| zaro | fungi: to bcpg.jar, bcpkix.jar, etc.. | 21:46 |
| fungi | zaro: okay, so perhaps those just wrote to the symlink targets rather than replacing the symlinks? | 21:46 |
| zaro | but actually the dates on those ae kinda old? | 21:46 |
| asselin | greghaynes, ok..uninstalled and reinstalled and it's there: -rwxr-xr-x 1 root root 3314 May 8 14:46 /usr/local/bin/dib-run-parts* | 21:47 |
| zaro | ohh wait there not even there. the links are there. | 21:47 |
| fungi | zaro: yeah, the links are linking to the ones provided by the distro packages still | 21:47 |
| greghaynes | asselin: cool, so there you go | 21:47 |
| asselin | thanks | 21:47 |
| zaro | ohh maybe need to delete those links before the download happens? | 21:48 |
| *** aysyd has quit IRC | 21:49 | |
| *** doug-fish has left #openstack-infra | 21:49 | |
| *** mriedem has quit IRC | 21:49 | |
| *** rfolco has quit IRC | 21:50 | |
| fungi | zaro: well, i would have expected the tidy on that dir to do that | 21:50 |
| zaro | fungi: the 'require' on the download of the libs is only set to 'require => File['/home/gerrit2/review_site/lib']' | 21:51 |
| zaro | fungi: probably needs to happen after the tidy? | 21:51 |
| zaro | ohh but tidy isn't even doing it's thing. | 21:51 |
| mordred | crinkle: I think I'd rather use the same thing since you've got it figured out - but I'm guessing that if there are things about beaker that are annoying in this context, they'll annoy you and EmilienM too ... so we might get some good upstream patches to make things shiner | 21:52 |
| mordred | shinier | 21:52 |
| mordred | not shiner | 21:52 |
| fungi | shriners? | 21:53 |
| clarkb | shiner bock? | 21:53 |
| crinkle | mordred: okay | 21:53 |
| fungi | schön | 21:53 |
| *** baoli_ has quit IRC | 21:53 | |
| mordred | jeblair: the swift hummingbird-in-go session is the same time as puppet testing - one of us may want to go lurk in that one | 21:55 |
| EmilienM | mordred: well, having beaker in the party for us is like using rspec-puppet. It just makes sense from a community & consistency perspective. Now if we want to talk about issues we run in rspec or beaker, we can, but they come with Puppet I would say | 21:55 |
| mordred | EmilienM: totally - I just think that we can probably make a patch or two to make them better, as we learn things | 21:55 |
| mordred | since we're a pretty different env, so we bring some more usecase I'm sure that the fine beaker folks would want to cover :) | 21:56 |
| *** otter768 has joined #openstack-infra | 21:56 | |
| zaro | fungi, clarkb : looking at the puppet again, tidy won't remove the link, it should only remove the '*-jdk*.jar' files. | 21:56 |
| zaro | and i bet the download failed because the links (with the same name) already exists. | 21:57 |
| *** rwsu has joined #openstack-infra | 21:58 | |
| fungi | zaro: that sounds like a good possibility, though i'm not sure what a downstream-friendly workaround is for that... perhaps we should use a slightly different filename than we used for the synlinks? | 21:58 |
| *** nelsnelson has quit IRC | 21:59 | |
| fungi | er, symlinks | 21:59 |
| zaro | fungi: i guess that would work. | 21:59 |
| zaro | fungi: but still tidy should have gotten rid of those *-jdk*.jar files but didn't do it | 22:01 |
| *** otter768 has quit IRC | 22:01 | |
| mordred | oslo.db wins best summit title | 22:02 |
| fungi | agreed, so i think it's either running before gerrit downloads them, or not at all | 22:02 |
| zaro | fungi: would tidy not run if steps before it failed? | 22:02 |
| fungi | zaro: possible | 22:02 |
| zaro | that could explain why it dien't work | 22:02 |
| fungi | zaro: if it is set to require some resource that couldn't be ensured successfully | 22:02 |
| jeblair | mordred: swift fishbowls have exactly the same timeslots as us, so that's not an easy conflict to resolve | 22:03 |
| * zaro needs to map the steps again, head is spinning from requires and befores | 22:03 | |
| *** jpeeler has quit IRC | 22:03 | |
| jeblair | mordred: i lean toward leaving the schedule as is and divide (as we always need to do anyway) | 22:04 |
| mordred | jeblair: yah. that's totally what I was thinking | 22:05 |
| *** stevemar has quit IRC | 22:05 | |
| clarkb | crinkle: I think we commented at the same time on that spec :) | 22:06 |
| jeblair | clarkb: did you say the same thing or opposite things? | 22:06 |
| clarkb | jeblair: I think very similar actually | 22:06 |
| jeblair | woot, we'll have a chorus in the fishbowl | 22:07 |
| clarkb | basically the big benefit is beaker lets devs run tests on laptop, if we can run those same tests in CI then win | 22:07 |
| clarkb | my concern is that we have hacked beaker so much for CI that we will end up with fragile tests that devs cannot reproduce though | 22:07 |
| clarkb | as evidenced by the churn we ran into today | 22:08 |
| *** EmilienM is now known as EmilienM|afk | 22:08 | |
| clarkb | but if we can get that out of the way and it just works then meh | 22:08 |
| mordred | clarkb: maybe we vargant the docker go cloudfoundry? | 22:08 |
| openstackgerrit | Jeremy Stanley proposed openstack-infra/project-config: Deal with yum --downloadonly CentOS 6.x quirks https://review.openstack.org/181574 | 22:10 |
| clarkb | mordred: I am pretty sure the words you just said made a large part of the world very excited | 22:10 |
| clarkb | mordred: but I have no idea what they mean :) | 22:11 |
| *** Sukhdev has quit IRC | 22:11 | |
| clarkb | zaro: fungi we should be able to do a puppet agent --test to see if and when tidy runs | 22:11 |
| fungi | i'm juiced to find out what a vargant is | 22:12 |
| mordred | fungi: it's like a vagrant, but you probably haven't heard of it | 22:12 |
| * fungi totally has not | 22:12 | |
| clarkb | mordred: also (and maybe I shouldn't even mention this at this point) but ansible would actually be a great substitute for beaker | 22:14 |
| clarkb | mordred: https://review.openstack.org/#/c/181555 made me think of that because with ansible you would just say localconnection | 22:14 |
| *** alexpilotti has joined #openstack-infra | 22:14 | |
| *** arielb has quit IRC | 22:14 | |
| clarkb | mordred: and it would happily do that, or if you openstack/dockered/lxced/etced a VM it could magically ssh into that too regardless of the distro | 22:14 |
| *** ddieterly has joined #openstack-infra | 22:15 | |
| clarkb | but I have no idea if ansible can run against OS X hypervisors which seems to be important | 22:15 |
| mordred | clarkb: I would be MASSIVELY surprised if it couldn't | 22:16 |
| *** beekneemech has quit IRC | 22:18 | |
| zaro | clarkb, fungi : so 2 things. we need the new libs to be in place before 'gerrit-init', not before 'gerrit-start' | 22:20 |
| openstackgerrit | Joe Gordon proposed openstack-infra/project-config: Add #defcore to the list of irc channels https://review.openstack.org/181576 | 22:20 |
| openstackgerrit | Joe Gordon proposed openstack-infra/system-config: Enable logging of defcore https://review.openstack.org/181577 | 22:20 |
| zaro | because init needs the new libs to do its thing | 22:20 |
| clarkb | zaro: except that init is what pulls them down right? | 22:20 |
| clarkb | zaro: so maybe we need two sets of things? | 22:20 |
| *** dimsum__ has quit IRC | 22:21 | |
| *** bknudson has quit IRC | 22:21 | |
| zaro | clarkb, fungi : ahh yes. | 22:21 |
| openstackgerrit | Jeremy Stanley proposed openstack-infra/project-config: Deal with yum --downloadonly CentOS 6.x quirks https://review.openstack.org/181574 | 22:21 |
| flashgordon | mordred: ^ | 22:22 |
| *** cdent has quit IRC | 22:23 | |
| *** annegentle has quit IRC | 22:23 | |
| clarkb | flashgordon: we need to get that channel registered first | 22:23 |
| flashgordon | clarkb: see top link | 22:24 |
| clarkb | flashgordon: I am looking at both changes? | 22:25 |
| clarkb | flashgordon: not sure what top link is | 22:25 |
| pleia2 | still needs to be registered, there's a check script for it (which it will fail) | 22:25 |
| flashgordon | clarkb: is https://review.openstack.org/181576 enough to register the channel? | 22:25 |
| pleia2 | flashgordon: nah, it's a chanserv command here on freenode | 22:25 |
| zaro | clarkb, fungi : actually i believe gerrit-init only uses the mysql lib to update the db schema. so i think that's actually fine. | 22:25 |
| flashgordon | ohh | 22:25 |
| flashgordon | gotcha | 22:25 |
| zaro | fungi, clarkb i mean actually fine that downloading of libs happens after gerrit-init | 22:26 |
| clarkb | zaro: cool | 22:26 |
| flashgordon | pleia2: anything I need to do to make that happen? | 22:27 |
| zaro | the 2nd thing i was going to point out is that tidy should execute before downloading new libs. which i don't think it is now. | 22:27 |
| pleia2 | flashgordon: does anyone in the channel have ops? | 22:27 |
| zaro | i think tidy executes after gerrit-start. | 22:28 |
| flashgordon | pleia2: not really sure actually I just went in the room moments ago myself | 22:28 |
| flashgordon | I bet mordred knows he is in there | 22:28 |
| pleia2 | flashgordon: ok, no ops, since it's not in the #openstack-* namespace, an appeal will likely need to go to freenode staff in #freenode to give someone ops to register it | 22:28 |
| zaro | gerrit-start fails so tidy doesn't run? | 22:28 |
| pleia2 | (also, why isn't it #openstack-defcore?) | 22:28 |
| jeblair | pleia2: i have magic powers in that regard | 22:28 |
| clarkb | pleia2: flashgordon or everyone leaves, then someone registers it | 22:28 |
| jeblair | pleia2: oh it's _not_ | 22:28 |
| pleia2 | jeblair: even outside of #openstack-* ? | 22:28 |
| jeblair | nm | 22:28 |
| jeblair | misread | 22:28 |
| clarkb | jeblair: for non #openstack- channels? | 22:28 |
| clarkb | kk | 22:29 |
| pleia2 | yeah | 22:29 |
| flashgordon | pleia2: great question | 22:29 |
| jeblair | i kind of think a new big tent thing should be that we register #openstack- channels for everything | 22:29 |
| jeblair | it's _really_ helpful to have that prefix | 22:29 |
| pleia2 | jeblair: ++ | 22:29 |
| flashgordon | jeblair: ++ | 22:29 |
| *** dalgaaf has joined #openstack-infra | 22:29 | |
| anteaya | the person who created it want to call it defcore, not openstack-defcore | 22:29 |
| anteaya | I didn't question why as I didn't see the possibility of making a case at the time | 22:30 |
| anteaya | wanted | 22:30 |
| openstackgerrit | Khai Do proposed openstack-infra/puppet-gerrit: Ensure that cleanup of libraries happens before gerrit starts https://review.openstack.org/181579 | 22:30 |
| zaro | fungi, clarkb ^ | 22:31 |
| asselin | greghaynes, still running into the issue: http://paste.openstack.org/show/217454/ | 22:31 |
| clarkb | zaro: we don't need that, see line 635 of the change | 22:32 |
| fungi | yeah, as i recall zaro explicitly requested that we use require relationships there rather than before relationships to make it easier to follow | 22:33 |
| zaro | clarkb: ahh i missed that. | 22:33 |
| hogepodge | pleia2: anteaya I don't see why we can't just move the room to conform. It makes sense in some ways. talk to eglute and zehicle about it, though | 22:34 |
| pleia2 | hogepodge: yeah, that would be good | 22:34 |
| zaro | yeah, still to one thing would help. but actually what i meant was to not use both require and before in one thing. | 22:35 |
| zaro | still/stick | 22:35 |
| anteaya | hogepodge: I have no argument but also no interest | 22:35 |
| asselin | greghaynes, the issue is that using 'compat' is not allowed.... | 22:35 |
| eglute | who created the room? | 22:36 |
| zaro | i meant block instead of thing. | 22:36 |
| *** annegentle has joined #openstack-infra | 22:36 | |
| clarkb | asselin: right qemu-img really screwed us here | 22:36 |
| anteaya | eglute: it doesn't really matter, there are no ops and it isn't registered | 22:37 |
| clarkb | asselin: they added a non backward compatible flag for backward compat | 22:37 |
| zaro | clarkb, fungi : anyways in it's curent state i can't see what would fail so i'll try to run puppet test to see what happens | 22:37 |
| clarkb | asselin: its super meta :) | 22:37 |
| asselin | clarkb, ha! | 22:37 |
| anteaya | eglute: so the minute the person who first entered the room leaves it, noone can get ops if it isnt' registered | 22:37 |
| clarkb | asselin: so if you want to be actually backward compatbile you have to do it yourself | 22:37 |
| eglute | i am not familiar with IRC operations- what would it take to get ops and re-name? | 22:37 |
| clarkb | asselin: I think greghaynes was working on pushing that into DIB so that nodepool could stop | 22:37 |
| clarkb | greghaynes: ^ any progress on that? | 22:37 |
| anteaya | eglute: well the rename is easier | 22:38 |
| asselin | clarkb, right...that's the patch I'm testing now...doesn't work | 22:38 |
| clarkb | eglute: if everyone left the room then it could be registered | 22:38 |
| asselin | clarkb, https://review.openstack.org/#/c/177043/1/bin/disk-image-create | 22:38 |
| clarkb | asselin: oh! I see | 22:38 |
| anteaya | eglute: type /join #openstack-defcore | 22:38 |
| anteaya | eglute: then you can register it | 22:38 |
| greghaynes | hes using the dib patch for that. theres a nodepool patch too I forgot about :( | 22:38 |
| asselin | greghaynes, which one? | 22:38 |
| greghaynes | but EONPHONE | 22:38 |
| greghaynes | so cant look it up | 22:39 |
| asselin | i'll try to find it | 22:39 |
| eglute | how do i register it | 22:39 |
| zaro | clarkb: should i disable the agent before running puppet agent --test? | 22:39 |
| clarkb | hrm I thought Iapproved the docs change to doc registration | 22:39 |
| clarkb | but I don't see it published | 22:39 |
| anteaya | eglute: working on that | 22:39 |
| clarkb | zaro: no shouldn't be necessary | 22:39 |
| asselin | greghaynes, https://review.openstack.org/#/c/177045/ i suppose | 22:39 |
| anteaya | jeblair: where did the page with your new patch end up? http://ci.openstack.org/irc.html#access | 22:39 |
| clarkb | eglute: anteaya https://review.openstack.org/#/c/181036/1/doc/source/irc.rst | 22:40 |
| clarkb | apparently it didn't publish, may have run into problems when jenkins/zuul were unhappy | 22:40 |
| anteaya | eglute: type /msg chanserv register #openstack-defcore | 22:41 |
| anteaya | eglute: then type /msg chanserv set #openstack-defcore guard on | 22:41 |
| anteaya | eglute: then type /msg chanserv access #openstack-defcore add openstackinfra +AFRefiorstv | 22:42 |
| anteaya | eglute: and you are done | 22:42 |
| eglute | anteaya thank you, done! | 22:42 |
| eglute | so now we need to tell everyone we are moving to the new channel? | 22:43 |
| anteaya | welcome and congratulations | 22:43 |
| anteaya | yes | 22:43 |
| eglute | thank you!! | 22:43 |
| clarkb | zaro: if you can paste the output when it is done that may be helpful | 22:43 |
| anteaya | eglute: please review https://review.openstack.org/#/c/181576/1 and comment that you have registered #openstack-defcore | 22:44 |
| anteaya | eglute: and you are most welcome | 22:44 |
| eglute | anteaya Jenkins is not happy with that PR | 22:44 |
| *** julim has joined #openstack-infra | 22:44 | |
| anteaya | no | 22:45 |
| anteaya | because #defcore isn't registered | 22:45 |
| anteaya | but if you tell flashgordon to change the patch to openstack-defcore | 22:45 |
| anteaya | jenkins will become happier, I predict | 22:45 |
| eglute | thanks, will do | 22:46 |
| anteaya | thank you | 22:46 |
| clarkb | zaro: is it running? | 22:46 |
| zaro | clarkb: http://paste.openstack.org/show/217463/ | 22:47 |
| clarkb | zaro: it definitely doesn't appear to have run tidy | 22:47 |
| zaro | clarkb: is that listed in order of execution? | 22:47 |
| *** abregman has quit IRC | 22:48 | |
| clarkb | zaro: yes should be | 22:48 |
| clarkb | zaro: though that paste looks incomplete | 22:48 |
| clarkb | zaro: also are you using --test? that looks more verbose than --test | 22:49 |
| zaro | here's the rest of it: http://paste.openstack.org/show/217464/ | 22:50 |
| zaro | clarkb: yes, i used --debug --test | 22:50 |
| clarkb | oh maybe thats extra debuggy | 22:50 |
| clarkb | zaro: maybe try it without the extra --debug? | 22:51 |
| zaro | well i don't see any references to tidy except for the beginning when i sets up the ordering? | 22:51 |
| clarkb | its kinda hard to follow what it is doing with --debug | 22:51 |
| zaro | clarkb: http://paste.openstack.org/show/217466/ | 22:52 |
| *** vipuls has joined #openstack-infra | 22:52 | |
| zaro | no info there | 22:52 |
| clarkb | no thats good info, thats the only thing puppet wanted to do | 22:52 |
| clarkb | so it isn't running tody | 22:52 |
| clarkb | *tidy | 22:52 |
| zaro | puppet bug? | 22:54 |
| clarkb | probably a bug in our puppet | 22:54 |
| *** hodos|2 has quit IRC | 22:55 | |
| clarkb | oh right the tidy doesn't match any of the symlinks | 22:55 |
| clarkb | so we don't want the tidy to run | 22:56 |
| clarkb | but we should instead get the modules installed | 22:56 |
| *** rwsu has quit IRC | 22:56 | |
| clarkb | er libs installed via maven | 22:56 |
| clarkb | I think I see the problem | 22:57 |
| clarkb | zaro: fungi: the execs to install stuff have creates clauses that match the symlinks | 22:57 |
| clarkb | if we remove the symlinks then it should work | 22:58 |
| clarkb | so we can either create another tidy step or make that a manual step or delete it in the exec if a symlink etc | 22:58 |
| *** Krinkle|detached is now known as Krinkle | 22:58 | |
| fungi | clarkb: yeah, that's why i suggested changing the names of them a little | 22:58 |
| fungi | then the tidy should blow away the old symlinks as unmanaged | 22:59 |
| fungi | that way, no manual step, so more downstream-friendly | 22:59 |
| clarkb | ya that should work too | 22:59 |
| clarkb | I have to run now, but I am pretty sure that is all that is happening | 22:59 |
| zaro | in the mean time i'll delete the .*.jar.backup links to remove confusion | 23:00 |
| clarkb | those did get deleted | 23:00 |
| openstackgerrit | Monty Taylor proposed openstack-infra/puppet-meetbot: Produce html logs for our IRC logs https://review.openstack.org/181377 | 23:01 |
| mordred | clarkb, fungi: ^^ now with puppet-lint happy | 23:01 |
| zaro | fungi, clarkb : i will put up a change. | 23:02 |
| *** dboik has quit IRC | 23:03 | |
| *** dboik has joined #openstack-infra | 23:04 | |
| *** zehicle has joined #openstack-infra | 23:04 | |
| asselin | greghaynes, with minor fix to your nodepool patch, image created successfullly. will test it now to make sure it really works :) | 23:07 |
| *** Krinkle is now known as Krinkle|detached | 23:07 | |
| *** annegentle has quit IRC | 23:09 | |
| greghaynes | asselin: w00t | 23:10 |
| *** btully has joined #openstack-infra | 23:12 | |
| *** patrickeast has quit IRC | 23:15 | |
| *** panda|afk has quit IRC | 23:15 | |
| *** mpopow has quit IRC | 23:15 | |
| *** packet has quit IRC | 23:15 | |
| *** panda|afk has joined #openstack-infra | 23:16 | |
| *** mpopow has joined #openstack-infra | 23:16 | |
| openstackgerrit | Joe Gordon proposed openstack-infra/system-config: Enable logging of #openstack-defcore https://review.openstack.org/181577 | 23:17 |
| openstackgerrit | Joe Gordon proposed openstack-infra/project-config: Add #openstack-defcore to the list of irc channels https://review.openstack.org/181576 | 23:18 |
| *** rwsu has joined #openstack-infra | 23:22 | |
| openstackgerrit | Khai Do proposed openstack-infra/puppet-gerrit: Ensure correct set of bouncy castle libs for Gerrit. https://review.openstack.org/181586 | 23:24 |
| zaro | fungi, clarkb ^ look right? | 23:25 |
| *** ddieterly has quit IRC | 23:27 | |
| openstackgerrit | Khai Do proposed openstack-infra/puppet-gerrit: Ensure only one set of bouncy castle libs are installed for Gerrit. https://review.openstack.org/181586 | 23:28 |
| *** marun has quit IRC | 23:32 | |
| openstackgerrit | Chris Hoge proposed openstack-infra/project-config: Add gerritbot to openstack-defcore channel https://review.openstack.org/181590 | 23:32 |
| * zaro needs to step away for dinner, back later. | 23:34 | |
| *** mestery has quit IRC | 23:34 | |
| openstackgerrit | Darragh Bailey proposed openstack-infra/jenkins-job-builder: Separate out custom yaml tags into separate classes https://review.openstack.org/180370 | 23:35 |
| *** pelix has quit IRC | 23:35 | |
| *** heyongli has quit IRC | 23:36 | |
| *** craigbr has quit IRC | 23:37 | |
| *** heyongli has joined #openstack-infra | 23:37 | |
| *** alexpilotti_ has joined #openstack-infra | 23:45 | |
| *** hemna is now known as hemnafk | 23:46 | |
| *** alexpilotti has quit IRC | 23:47 | |
| *** alexpilotti_ is now known as alexpilotti | 23:47 | |
| *** ajmiller has quit IRC | 23:48 | |
| *** nikita09 has quit IRC | 23:50 | |
| openstackgerrit | Jeremy Stanley proposed openstack-infra/puppet-gerrit: Ensure only one set of bouncy castle libs are installed for Gerrit. https://review.openstack.org/181586 | 23:52 |
| fungi | updated myself in the name of expediency | 23:52 |
| jeblair | fungi: i liked the old you, but i'll give the updated you a chance | 23:54 |
| anteaya | does he pass tests? | 23:54 |
| jeblair | we'll find out! | 23:55 |
| waynr | is there a place i can look for metrics about openstack's CI system? i'm interested in number of daily jobs triggered/run, number of projects served by the gerrit/zuul/gearman/jenkins stack, average CI response times, etc | 23:55 |
| jeblair | waynr: we have some graphs at the bottom of http://status.openstack.org/zuul/ | 23:56 |
| jeblair | waynr: they are generated from graphite | 23:56 |
| waynr | cool, i'm looking at that now | 23:56 |
| clarkb | fungi lgtm | 23:56 |
| jeblair | waynr: there's even more raw data available there: http://graphite.openstack.org/ | 23:56 |
| waynr | what about the number of jenkins masters/slaves/executors in use? | 23:56 |
| *** rwsu has quit IRC | 23:56 | |
| waynr | sweet! | 23:56 |
| *** otter768 has joined #openstack-infra | 23:57 | |
| jeblair | waynr: masters is constant at 8; dynamic slaves are a graph on the status page (nodes); there are a handful of static ones for special jobs | 23:57 |
| fungi | clarkb: if that version passes tests i'll approve it momentarily and we'll see if that does any better on review-dev | 23:57 |
| clarkb | fungi great | 23:57 |
| clarkb | I am finally getting lunch now so mostly not here but change makes sense | 23:57 |
| *** dannywil_ has quit IRC | 23:58 | |
| waynr | jeblair: is that the "Test Nodes" graph? | 23:58 |
| jeblair | waynr: you can take any of those images from the status page, open them in a new window, change 'render' to 'composer' in the url, and then further edit that graph in graphite | 23:58 |
| jeblair | waynr: ye | 23:58 |
| jeblair | yes | 23:58 |
| *** banix has joined #openstack-infra | 23:58 | |
| waynr | ha, that's pretty cool | 23:59 |
Generated by irclog2html.py 2.14.0 by Marius Gedminas - find it at mg.pov.lt!