| *** abishop_ has quit IRC | 00:33 | |
| *** abishop has joined #oooq | 00:35 | |
| *** rnoriega has quit IRC | 01:29 | |
| *** rnoriega has joined #oooq | 01:30 | |
| *** skramaja has joined #oooq | 03:31 | |
| *** jaganathan has joined #oooq | 03:44 | |
| *** udesale has joined #oooq | 04:02 | |
| *** udesale has quit IRC | 04:06 | |
| *** udesale has joined #oooq | 04:06 | |
| *** udesale has quit IRC | 04:06 | |
| *** ykarel|away has joined #oooq | 04:07 | |
| *** udesale has joined #oooq | 04:08 | |
| *** jtomasek has joined #oooq | 04:44 | |
| *** saneax has joined #oooq | 04:51 | |
| *** saneax has quit IRC | 05:06 | |
| *** hamzy has quit IRC | 05:32 | |
| *** hamzy has joined #oooq | 05:32 | |
| *** quiquell has joined #oooq | 05:41 | |
| *** jfrancoa has joined #oooq | 05:50 | |
| *** chkumar|pto is now known as chandankumar | 06:05 | |
| *** ykarel|away is now known as ykarel | 06:08 | |
| *** kopecmartin has joined #oooq | 06:44 | |
| *** jaosorior has joined #oooq | 06:51 | |
| *** ykarel is now known as ykarel|lunch | 07:28 | |
| *** tosky has joined #oooq | 07:36 | |
| *** apetrich has joined #oooq | 07:43 | |
| quiquell | ssbarnea, marios: Good morning | 07:45 |
|---|---|---|
| marios|rover | quiquell: i/ | 08:07 |
| marios|rover | o/ | 08:07 |
| marios|rover | quiquell: talking on tripleo sec | 08:07 |
| marios|rover | quiquell: o/ | 08:08 |
| quiquell | marios|rover: ack | 08:09 |
| marios | so hows it going quiquell | 08:10 |
| marios | long time ! | 08:10 |
| marios | :) | 08:10 |
| marios | you then i was away | 08:10 |
| *** holser_ has joined #oooq | 08:10 | |
| quiquell | marios: Fine, just reading the pile of sh... :-) | 08:11 |
| marios | quiquell: my pc beeped | 08:11 |
| quiquell | marios: What do you mean ? | 08:12 |
| marios | quiquell: like when someone says swear work | 08:12 |
| marios | on tv | 08:12 |
| marios | joke | 08:12 |
| quiquell | Ahh ok... | 08:12 |
| quiquell | So haw is everything here ? | 08:12 |
| *** ykarel|lunch is now known as ykarel | 08:13 | |
| marios | quiquell: hmm 71 % pass on grafana | 08:13 |
| marios | quiquell: seems low | 08:14 |
| marios | has it been bad? | 08:14 |
| marios | quiquell: hmm but only 2 failed upstream gate jobs | 08:14 |
| marios | but quite a few check jobs | 08:14 |
| quiquell | timeouts again ? :-/ | 08:16 |
| quiquell | Is dashboard-ci working fine ? | 08:18 |
| quiquell | I see some updates for rocky but no the code change for it | 08:18 |
| quiquell | Did you update it directly over it ? | 08:18 |
| *** sshnaidm|off is now known as sshnaidm | 08:23 | |
| marios | quiquell: yeah might be i still didn't fully catch up gonna go do some reviews now and then cycle back | 08:27 |
| chandankumar | Good morning guys, /me is back | 08:27 |
| quiquell | chandankumar: o/ me2 | 08:29 |
| marios | o/ chandankumar | 08:30 |
| * marios 3 | 08:30 | |
| chandankumar | quiquell: marios \o/ | 08:30 |
| *** chem has joined #oooq | 08:59 | |
| *** skramaja has quit IRC | 09:01 | |
| *** skramaja has joined #oooq | 09:01 | |
| quiquell | sshnaidm: Hello sir | 09:12 |
| quiquell | sshnaidm: Are you there ? | 09:12 |
| sshnaidm | quiquell, hi! welcome back :) | 09:12 |
| quiquell | sshnaidm: Thanks man, how you doing | 09:13 |
| sshnaidm | quiquell, great) how was pto? | 09:13 |
| quiquell | sshnaidm: Sort as always :-) | 09:13 |
| quiquell | sshnaidm: Did we have change dashboard-ci without reviews ? | 09:14 |
| quiquell | sshnaidm: I see rocky all around | 09:14 |
| quiquell | sshnaidm: But there are no commits for it | 09:14 |
| sshnaidm | quiquell, yeah, changes of dashboard I didn't submit, waited for you, how do you exactly export them? | 09:14 |
| quiquell | sshnaidm: With the export script that sit's in the role | 09:15 |
| quiquell | sshnaidm: In case the change is too complex and cannot be done directly in the json | 09:15 |
| sshnaidm | quiquell, export-grafana.py? | 09:18 |
| quiquell | sshnaidm: yep | 09:18 |
| *** dtantsur|afk is now known as dtantsur | 09:19 | |
| quiquell | sshnaidm: In case is the first time you start it up you need to create an api key with create-api-key.py | 09:19 |
| sshnaidm | k | 09:20 |
| quiquell | ./create-api-key.py --key-name foo | 09:20 |
| quiquell | do the changes | 09:20 |
| quiquell | ./export-grafana.py | 09:20 |
| sshnaidm | quiquell, the only thing didn't work it's skipped rocky promotion table | 09:36 |
| sshnaidm | quiquell, I created all vars, but it didn't appear | 09:36 |
| quiquell | sshnaidm: Testing it with your commit | 09:38 |
| quiquell | sshnaidm: There is a space in the "releases" variable, maybe this is the issue | 09:42 |
| quiquell | sshnaidm: I mean missing space | 09:42 |
| sshnaidm | ok | 09:44 |
| quiquell | sshnaidm: It's not repeating the panels... weird | 09:48 |
| *** jbadiapa has quit IRC | 09:49 | |
| quiquell | sshnaidm: Found the issue, since it's a hidden variable, you need to "dehide it" select rocky and the hide it again, let me check if the change is somwhere in th ejson | 09:53 |
| *** jbadiapa has joined #oooq | 10:01 | |
| quiquell | sshnaidm: Commented on the change, that should fix the issue. | 10:03 |
| *** cibott has joined #oooq | 10:07 | |
| sshnaidm | @promotion status all | 10:07 |
| quiquell | sshnaidm: ^ what's that ? | 10:07 |
| cibott | sshnaidm: | 10:07 |
| cibott | Master: consistent - 4d tripleo-ci - 4d phase1 - 20d phase2 - 20d | 10:07 |
| cibott | Rocky: consistent - 0d tripleo-ci - 1d phase1 - 1d phase2 - N/A | 10:07 |
| cibott | Queens: consistent - 3d tripleo-ci - 13d phase1 - 13d phase2 - 13d | 10:07 |
| cibott | Pike: consistent - 0d tripleo-ci - 12d{:co | 10:07 |
| cibott | lor='red'} phase1 - 14d phase2 - 16d | 10:07 |
| sshnaidm | @patch status 567224 | 10:08 |
| cibott | sshnaidm: Patch is in CI pipeline "check". Jobs are in progress, but 2 failed :(. Overall progress: 7 jobs are running (95%), 4 passed, 2 failed (2 voting). | 10:08 |
| sshnaidm | quiquell, it's errbot :) | 10:08 |
| quiquell | sshnaidm: errbot ? it's super cool | 10:09 |
| sshnaidm | seems like colors don't work.. | 10:09 |
| quiquell | @help | 10:09 |
| cibott | quiquell: You're not allowed to access this command from a chatroom | 10:09 |
| quiquell | @help please | 10:09 |
| cibott | quiquell: You're not allowed to access this command from a chatroom | 10:09 |
| sshnaidm | quiquell, send him to private | 10:09 |
| sshnaidm | quiquell, it's pretty long | 10:09 |
| quiquell | sshnaidm: Did we implement the CI part or it was out of the box ? | 10:13 |
| quiquell | I am curious | 10:13 |
| sshnaidm | quiquell, no, I wrote it | 10:13 |
| sshnaidm | quiquell, it's pretty easy: https://github.com/sshnaidm/errbot-ci-tasks | 10:14 |
| marios | quiquell: please check https://review.openstack.org/#/c/595527/ for the ocata/queens timeouts like at https://review.openstack.org/#/c/564291/ | 10:14 |
| marios | quiquell: when you get a chance thanks | 10:14 |
| *** cibott has quit IRC | 10:15 | |
| quiquell | sshnaidm: Maybe we can put the dashboard-ci alerts there to ask for them | 10:17 |
| sshnaidm | quiquell, yeah, it has webhook too | 10:17 |
| quiquell | sshnaidm: Will send you a pull request with it | 10:18 |
| quiquell | sshnaidm: ok ? | 10:18 |
| sshnaidm | quiquell, I'd like bot will post some alerts to chat to ruck and rover.. | 10:18 |
| sshnaidm | quiquell, sure | 10:18 |
| quiquell | sshnaidm: I have that in the dashboard-ci bot, maybe we can merge both | 10:19 |
| sshnaidm | quiquell, cool | 10:19 |
| sshnaidm | quiquell, btw, it has a persistent db too.. | 10:20 |
| quiquell | sshnaidm: That's cool, grafana alerts API is not very consistent with the API from notifications, so we can store missing info in the db | 10:21 |
| quiquell | marios: tripleo-bootstrap role only exists at master and rocky ? | 10:23 |
| quiquell | marios: I think periodic jobs are missing in the change, they use different release file | 10:31 |
| *** ccamacho has joined #oooq | 10:41 | |
| *** ccamacho has quit IRC | 10:41 | |
| marios | quiquell: yeah | 10:43 |
| marios | quiquell: afaics (i wrote in the bug) | 10:43 |
| marios | quiquell: so it isn't easily backportable | 10:43 |
| marios | i checked with slagle and he agreed | 10:43 |
| marios | quiquell: so we merged the previous patch i pointed at in the commit message | 10:43 |
| quiquell | marios: But periodic jobs are not covered with this | 10:44 |
| quiquell | marios: I think they run the old workflow though... | 10:44 |
| marios | quiquell: well, it definitely fixed the pike job, and the logs on the failed queens job and ocata too have the same symptoms as https://bugs.launchpad.net/tripleo/+bug/1785067 | 10:44 |
| openstack | Launchpad bug 1785067 in tripleo "without bootstrap-subnodes queens/pike/ocata deployment fails for missing os-collect-config" [High,In progress] - Assigned to Marios Andreou (marios-b) | 10:44 |
| marios | quiquell: like http://logs.openstack.org/91/564291/20/check/tripleo-ci-centos-7-nonha-multinode-oooq/e113aba/logs/subnode-2/var/log/yum.log.txt.gz | 10:45 |
| marios | quiquell: i checked/reconfirmed again this morning | 10:45 |
| marios | quiquell: so there is no os-collect-config | 10:47 |
| marios | quiquell: Failed to restart os-collect-config.service: Unit not found. http://logs.openstack.org/91/564291/20/check/tripleo-ci-centos-7-nonha-multinode-oooq/e113aba/logs/undercloud/var/log/deployed-server-os-collect-config.log.txt.gz | 10:47 |
| quiquell | marios: Yep the commit is alright, maybe what we need to do is going back to this when periodic jobs are moved to the new workflow | 10:48 |
| quiquell | marios: Going to state that in the trello card | 10:48 |
| quiquell | marios: Going to remove my -1 as periodic jobs is different issue. | 10:50 |
| *** udesale has quit IRC | 10:56 | |
| *** skramaja has quit IRC | 10:57 | |
| marios | thanks for checking quiquell | 11:00 |
| ykarel | marios, quiquell queens jobs were switched to config-download, so ^^ would not affect queens i think, | 11:00 |
| quiquell | ykarel: What's the relation ? | 11:01 |
| marios | ykarel: well it *was* exibiting same symptoms on queens ~1 week ago see comment #11 at https://bugs.launchpad.net/tripleo/+bug/1785067 | 11:01 |
| openstack | Launchpad bug 1785067 in tripleo "without bootstrap-subnodes queens/pike/ocata deployment fails for missing os-collect-config" [High,In progress] - Assigned to Marios Andreou (marios-b) | 11:01 |
| ykarel | marios, yes 5 days ago it was switched: https://review.openstack.org/#/c/597141/ | 11:02 |
| ykarel | quiquell, os-collect-config not used with config-download | 11:02 |
| marios | ykarel: ah ok we can update the review then will you please vote at https://review.openstack.org/#/c/595527/ | 11:02 |
| marios | ykarel: so is the queens gate fine now ? me checks (gate check i mean) | 11:03 |
| ykarel | marios, still two scenarios failing | 11:03 |
| ykarel | 2 are passing after the patch i mentioned | 11:03 |
| marios | ykarel: hm yeah but is in better shape at least | 11:03 |
| ykarel | i think i got the issue with other two scenarios | 11:03 |
| marios | ykarel: before all the scenarios were at 3 hr timeouts | 11:03 |
| ykarel | marios, yes correct | 11:03 |
| marios | ykarel: ok cool please vote at https://review.openstack.org/#/c/595527/ | 11:04 |
| marios | and i will update immediately | 11:04 |
| marios | ! | 11:04 |
| ykarel | ack | 11:05 |
| marios | ykarel: ah | 11:05 |
| marios | ykarel: but | 11:05 |
| marios | ykarel: we will still need it for those featuresets | 11:05 |
| marios | that arent using config download | 11:05 |
| marios | like 7 i guess | 11:05 |
| marios | i mean looking at the logs now | 11:05 |
| marios | ykarel: yu | 11:05 |
| marios | yup | 11:05 |
| marios | Failed to start os-collect-config.service: Unit not found. | 11:06 |
| marios | at http://logs.openstack.org/24/567224/97/check/tripleo-ci-centos-7-scenario007-multinode-oooq-container/94715f8/logs/undercloud/var/log/deployed-server-os-collect-config.log.txt.gz | 11:06 |
| ykarel | marios, okk got it in hte patch only 4 scenarios were changed to config-download, so u are right | 11:06 |
| marios | ykarel: so might still be worth a discussion on the review and maybe i add a comment but we might just keep this one... it IS after all a bootstrap item imo | 11:06 |
| ykarel | we need that package in queens as well | 11:06 |
| *** ccamacho has joined #oooq | 11:06 | |
| ykarel | marios, +1 | 11:08 |
| ssbarnea | hi guys! I am now back now, please let me know what can I do to help. | 11:10 |
| quiquell | ssbarnea: Hello there, I am back too | 11:10 |
| ssbarnea | sadly the list of alert bugs is quite long and I am not sure in which order we can fix them. | 11:11 |
| *** jfrancoa has quit IRC | 11:12 | |
| quiquell | sshnaidm: Going to update your change for the dashboard-ci in rocky and give a +2 | 11:12 |
| quiquell | ok ? | 11:12 |
| *** jfrancoa has joined #oooq | 11:16 | |
| sshnaidm | quiquell, cool | 11:20 |
| *** jfrancoa has quit IRC | 11:23 | |
| ssbarnea | panda|rover: are you here? | 11:27 |
| *** skramaja has joined #oooq | 11:27 | |
| ssbarnea | i am looking at this failure and I have no idea from where did teh no_log came on that task, as master does not have it. http://logs.rdoproject.org/27/595527/2/openstack-check/legacy-tripleo-ci-centos-7-ovb-3ctlr_1comp-featureset001-master/201679b/job-output.txt.gz | 11:27 |
| ykarel | marios, good to address quiquell's comment about promotion job as well | 11:33 |
| ykarel | marios, are u planning a seperate patch for that? | 11:33 |
| quiquell | ykarel: We don't need that until we move promotion jobs to the new workflow, only the new workflow is using mario's work | 11:35 |
| *** jfrancoa has joined #oooq | 11:35 | |
| quiquell | marios: ^ | 11:35 |
| ykarel | quiquell, see https://logs.rdoproject.org/openstack-periodic-24hr/git.openstack.org/openstack-infra/tripleo-ci/master/legacy-periodic-tripleo-ci-centos-7-multinode-1ctlr-featureset016-queens/92eaadc/logs/undercloud/var/log/deployed-server-os-collect-config.log.txt.gz | 11:36 |
| ykarel | quiquell, after https://review.openstack.org/#/c/599088/1, will hit ^^ again | 11:36 |
| quiquell | ykarel: Humm following bad line of though... then we need the periodics | 11:37 |
| ssbarnea | Anyone aware about this: "/home/zuul/repo_setup.sh: line 212: NODEPOOL_CBS_CENTOS_PROXY: unbound variable" ? | 11:39 |
| quiquell | sshnaidm: Do you know what's the issue with https://review.openstack.org/#/c/589448 ? | 11:41 |
| ykarel | ssbarnea, i think u should not see ^^ after https://review.openstack.org/#/c/599280 | 11:45 |
| *** ccamacho has quit IRC | 11:50 | |
| *** ccamacho has joined #oooq | 11:50 | |
| *** jfrancoa has quit IRC | 11:54 | |
| *** jfrancoa has joined #oooq | 11:54 | |
| ssbarnea | ykarel: ok, thanks. btw, have a look at https://review.openstack.org/#/c/599358/ -- to save few clicks. | 11:58 |
| *** rfolco has joined #oooq | 11:58 | |
| quiquell | rfolco: Better here yep | 12:01 |
| rfolco | quiquell, there is one thing that I could not fix... | 12:02 |
| quiquell | rfolco: So looks like playbook executions are ok doing the comparagion https://paste.fedoraproject.org/paste/mX6E06gi2tjt2Cg80FOZ8g | 12:02 |
| rfolco | http://logs.openstack.org/48/589448/20/check/tripleo-ci-centos-7-3nodes-multinode/96834c4/job-output.txt.gz#_2018-09-01_09_53_52_160450 | 12:02 |
| rfolco | this --> --extra-vars' '@/home/zuul/src/git.openstack.org/openstack/tripleo-quickstart/config/nodes/2ctlr.yml | 12:02 |
| rfolco | it does not fail but this is annoying | 12:03 |
| quiquell | rfolco: Let me check, I remember fixing something similar before PTO | 12:05 |
| rfolco | quiquell, I tried changing <%- and -%> to eliminate whitespaces and line breaks... it is like nodes_file var had its own '' | 12:09 |
| quiquell | rfolco: It does not appear here http://logs.openstack.org/48/589448/20/check/tripleo-ci-centos-7-3nodes-multinode/96834c4/logs/playbook_executions.log | 12:11 |
| ykarel | ssbarnea, commented, it saves few clicks but am not sure about the policy for no_log, so good to standarize across all tasks | 12:11 |
| ykarel | it might be there to save disk on log server | 12:11 |
| ykarel | haven't digged in the history though | 12:11 |
| ssbarnea | the only policy that we can use is "common sense". if is too verbose we can do a tail. also the same task does not have no_log in other places. | 12:12 |
| rfolco | quiquell, because sed breaks --extra-vars in new lines | 12:14 |
| ykarel | yes that what i meant | 12:14 |
| rfolco | as I said no functional change, but this is annoying | 12:14 |
| rfolco | to my OCD level | 12:15 |
| rfolco | quiquell, :) | 12:15 |
| ssbarnea | marios: how can I find extra reason for stackcreation failure on http://logs.rdoproject.org/77/599077/1/openstack-check/legacy-tripleo-ci-centos-7-ovb-3ctlr_1comp-featureset001-ocata-branch/8cf4e60/logs/undercloud/home/zuul/overcloud_deploy.log.txt.gz ? | 12:15 |
| quiquell | rfolco: It does not appear before PTO http://logs.openstack.org/48/589448/5/check/tripleo-ci-centos-7-3nodes-multinode/33e5500/job-output.txt.gz | 12:15 |
| *** skramaja has quit IRC | 12:16 | |
| rfolco | quiquell, until PS9, yes, does not appear | 12:16 |
| rfolco | quiquell, I had to rebase on top of ronelle's patch, they were conflicting and blocking each other.... and blocking any other patch from sprint 18 | 12:17 |
| quiquell | rfolco: Ok, still landing from PTO :-) | 12:17 |
| quiquell | rfolco: I think we don't need ' " ' on it, that's why | 12:19 |
| sshnaidm | quiquell, tbh I'm not sure I understand what is going on in https://review.openstack.org/#/c/589448 | 12:20 |
| sshnaidm | rfolco, ^^ | 12:20 |
| quiquell | sshnaidm: no worries | 12:20 |
| quiquell | rfolco: Ok going to test one thing | 12:20 |
| sshnaidm | rfolco, there are changes from patchset 9 that I don't understand | 12:20 |
| rfolco | sshnaidm, I rebased on top of ronelle's patch | 12:21 |
| rfolco | it became much simpler and shorter | 12:21 |
| rfolco | with the reparenting and removal of environment_type files | 12:22 |
| quiquell | rfolco: But now we depends on a DNM patch ? | 12:22 |
| sshnaidm | rfolco, why do we still load ovb vars? | 12:23 |
| rfolco | sshnaidm, when env = ovb --> https://review.openstack.org/#/c/589448/20/playbooks/tripleo-ci/run-v3.yaml | 12:24 |
| quiquell | rfolco: Ok update the review, without '"' let's check the output | 12:24 |
| rfolco | I thought that way we didn't need all those if else on common.yaml | 12:24 |
| quiquell | rfolco: Ok... let help with the Ronelles review so we can deliver the two | 12:25 |
| quiquell | rfolco: But Ronelles change is DNM because could be not merged at the end ? | 12:26 |
| rfolco | quiquell, it was originally DNM since the beginning | 12:27 |
| sshnaidm | rfolco, so, shouldn't have it been parented from some ovb base job..? | 12:27 |
| rfolco | sshnaidm, hmm this is probably why we still cannot merge it | 12:29 |
| *** jaosorior has quit IRC | 12:29 | |
| sshnaidm | rfolco, why? | 12:29 |
| rfolco | oh no ignore me | 12:29 |
| rfolco | ovb base jobs are defined in tripleo-ci | 12:29 |
| *** ykarel is now known as ykarel|away | 12:30 | |
| rfolco | sshnaidm, well, featureset_conf changes with environment_type | 12:30 |
| quiquell | featurset_conf is now at common.yaml | 12:31 |
| marios | ssbarnea: call? | 12:31 |
| sshnaidm | rfolco, all this multinode settings in common - shouldn't they be in multinode base parent job? | 12:32 |
| quiquell | sshnaidm: They are used at multinode, singlenode and standalone | 12:32 |
| quiquell | sshnaidm: Name is missleading | 12:32 |
| *** dtantsur is now known as dtantsur|brb | 12:32 | |
| sshnaidm | quiquell, well, then in all of them | 12:33 |
| rfolco | sshnaidm, yes, all excpet ovb | 12:33 |
| quiquell | sshnaidm, rfolco: Yep per environment_type like before but in the new parent jobs | 12:33 |
| quiquell | rfolco: maybe the same for rdo_env_vars ? | 12:33 |
| sshnaidm | quiquell, I think we talked about having 4 parent job, each one for every environment. Not sure if something changed since then | 12:33 |
| quiquell | sshnaidm: I know, but there are variables that have been moving from the environment_type files to the common.yaml file | 12:34 |
| ssbarnea | #rhos-prod-chain | 12:34 |
| quiquell | sshnaidm, rfolco: We have to put now them at the new base jobs, they are featureset_conf, rdoi_env_vars and env_vars | 12:34 |
| quiquell | rfolco: Going to do it, ok ? | 12:35 |
| quiquell | So we have thos per environment_type like before | 12:36 |
| sshnaidm | quiquell, featureset is going to common or to environment base job? | 12:36 |
| quiquell | sshnaidm: environment base job | 12:36 |
| quiquell | sshnaidm: There will be some redundancy, but we can cope with that | 12:36 |
| rfolco | quiquell, ok, also need a QE on the parent patch (ronelle's) to verify all is good and we can remove DNM and merge it | 12:37 |
| rfolco | sshnaidm, ^ | 12:37 |
| sshnaidm | quiquell, but featureset is environment independent | 12:38 |
| quiquell | rfolco: Cool cool, thanks | 12:38 |
| quiquell | sshnaidm: not for OVB | 12:38 |
| sshnaidm | quiquell, why? | 12:38 |
| quiquell | sshnaidm: ovb -> https://review.openstack.org/#/c/589448/5/playbooks/tripleo-ci/vars/environment_type/ovb.yaml | 12:39 |
| quiquell | sshnaidm: the others -> https://review.openstack.org/#/c/589448/5/playbooks/tripleo-ci/vars/environment_type/multinode.yaml | 12:39 |
| *** udesale has joined #oooq | 12:39 | |
| quiquell | sshnaidm: ovb is not using featureset-multinode-common.yml | 12:40 |
| rfolco | well if featureset is only used to get the right featureset file and we move it to base job, there is no point in having featureset_conf | 12:40 |
| *** zul has joined #oooq | 12:40 | |
| sshnaidm | quiquell, yeah, so multinode-common is environment dependent, not featureset file | 12:40 |
| quiquell | rfolco: that's common_featureset_conf | 12:40 |
| rfolco | you mean it should be | 12:40 |
| quiquell | sshnaidm: Yep, don't know why this file has the featureset prefix... maybe that's the real issue | 12:41 |
| quiquell | rfolco: we cannot add conditionals in the zuul config ? | 12:49 |
| rfolco | quiquell, no, if you have conditional keep it in the code | 12:59 |
| quiquell | rfolco: Then I don't know how to resolve the ovb + multinode without the environment_type files | 13:00 |
| quiquell | env_vars are different between the two | 13:00 |
| rfolco | quiquell, whats wrong with my approach ? | 13:00 |
| rfolco | when: env = ovb include_vars: ovb.yml | 13:01 |
| marios | we have call? | 13:01 |
| rfolco | oops | 13:01 |
| quiquell | rfolco: Ahh silly me, I was missing that part :-/ sorry man | 13:02 |
| quiquell | Ok all good | 13:03 |
| ssbarnea | anyone aware about " Verify Sphinx build" tasking failing when it encounters and empty string in the loop? | 13:03 |
| ssbarnea | https://ci.centos.org/job/tripleo-quickstart-gate-master-delorean-quick-basic/6854/console | 13:05 |
| *** jaosorior has joined #oooq | 13:06 | |
| *** ykarel|away has quit IRC | 13:12 | |
| *** dtantsur|brb is now known as dtantsur | 13:30 | |
| quiquell | rfolco, sshnaidm: updated with the stuff we have talk about https://review.openstack.org/589448 | 13:41 |
| quiquell | humm going to remove included ovb wait | 13:42 |
| quiquell | sshnaidm: I think I am going to remove the github part fo dashboard-ci don't know why I added, ok ? | 13:44 |
| sshnaidm | quiquell, why, it's useful when github has problems | 13:44 |
| sshnaidm | quiquell, actually we added when it was down and job failed | 13:45 |
| quiquell | sshnaidm: Ups ok, didn't remember, ok. | 13:45 |
| rfolco | I think this won't work | 13:46 |
| rfolco | featureset_conf: "{{ common_featureset_conf}} {{ job_featureset_conf }}" | 13:46 |
| quiquell | we cannot do that at zuul config ? | 13:47 |
| rfolco | I think you should move literal part of it to zuul config and compose with the variable one at common | 13:47 |
| quiquell | rfolco: I see the jobs spinning, so the config have to be ok | 13:48 |
| quiquell | Damn you are right... this is evaluate before running the playbook :-) | 13:49 |
| quiquell | silly me | 13:50 |
| rfolco | featureset_common: "common-featureset.yml" | 13:50 |
| rfolco | then in common | 13:50 |
| quiquell | yep, the reverse | 13:51 |
| rfolco | featureset_conf: "{{ featureset_common|default('') }} {{ job_featureset_conf }}" | 13:51 |
| rfolco | so ovb job won't have common, defaults to "" | 13:51 |
| *** jaosorior has quit IRC | 13:53 | |
| quiquell | rfolco: Puff I don't want to put the whole path in the job config... will do a conditional | 13:55 |
| rfolco | quiquell, no don't do that | 13:56 |
| rfolco | quiquell, conditional in common.yml, good | 13:57 |
| quiquell | rfolco: Ahh ok | 13:57 |
| quiquell | going that path then | 13:57 |
| quiquell | thanks you | 13:57 |
| rfolco | well, conditional in common.yml we're back to where we were | 13:58 |
| rfolco | lolz | 13:59 |
| rfolco | quiquell, sshnaidm I still believe my ovb override approach was good | 13:59 |
| quiquell | ok updated | 14:01 |
| quiquell | Let's see now | 14:01 |
| rfolco | quiquell, symlink does not seem right... it will load "multinode.yml" string, not multinode.yml file contents | 14:07 |
| rfolco | quiquell, I hope I am wrong there | 14:07 |
| quiquell | rfolco: Really ? | 14:08 |
| quiquell | I need to use lookup ? | 14:08 |
| quiquell | or the like ? | 14:08 |
| quiquell | Arggg I have take over this to break it :-) | 14:09 |
| rfolco | quiquell, another option would be a symlink on tqe pointing to the same file... and a depends-on patch... | 14:10 |
| quiquell | rfolco: But those files are at tripleo-ci | 14:11 |
| rfolco | quiquell, oh, nevermind, crossed stations here | 14:11 |
| rfolco | multitasking w/ my other patches :) | 14:12 |
| quiquell | rfolco: no problem, going to run this locally to reduce noise | 14:12 |
| quiquell | rfolco: added some comments to them, small stuff | 14:12 |
| rfolco | quiquell, I saw, thanks for revewing | 14:12 |
| quiquell | yw | 14:12 |
| *** dtrainor has joined #oooq | 14:17 | |
| *** quiquell has quit IRC | 14:23 | |
| *** dtrainor_ has joined #oooq | 14:24 | |
| ssbarnea | sshnaidm: can you please give a +W on https://review.openstack.org/#/c/571176/ ? | 14:24 |
| sshnaidm | ssbarnea, only after I | 14:26 |
| sshnaidm | will review it | 14:26 |
| sshnaidm | :) | 14:26 |
| ssbarnea | you already did. | 14:26 |
| ssbarnea | it was one of the few waiting for a W, that was "lucky" to pass the checks :D | 14:26 |
| sshnaidm | ssbarnea, well, it should be tested on libvirt firstly, CI doesn't help here unfortunately | 14:27 |
| *** dtrainor has quit IRC | 14:27 | |
| *** dtrainor__ has joined #oooq | 14:37 | |
| *** dtrainor_ has quit IRC | 14:40 | |
| *** dtrainor_ has joined #oooq | 14:47 | |
| *** dtrainor__ has quit IRC | 14:49 | |
| sshnaidm | marios, fixed https://review.openstack.org/#/c/599206/\ | 15:22 |
| marios|rover | ack sshnaidm | 15:23 |
| marios|rover | ttyl | 15:24 |
| * marios|rover hometime | 15:24 | |
| *** udesale has quit IRC | 15:26 | |
| *** dtantsur is now known as dtantsur|afk | 15:32 | |
| sshnaidm | ssbarnea, can you update https://review.openstack.org/#/c/599201/ please in your time | 15:50 |
| *** jfrancoa has quit IRC | 16:05 | |
| *** ccamacho has quit IRC | 16:20 | |
| *** kopecmartin has quit IRC | 16:43 | |
| *** holser_ has quit IRC | 17:11 | |
| ssbarnea | sshnaidm: sorry, I am not changing my vote on that one because this is how I interpret the first two paragraphs from https://docs.openstack.org/tripleo-docs/latest/contributor/reproduce-ci.html | 17:20 |
| sshnaidm | ssbarnea, I don't understand how it's related. These paragraphs talk about parameters of the job, they are not related to reproducer as a tool itself | 17:21 |
| sshnaidm | ssbarnea, well, this patch is not for fun, it fixes reproducer that is broken due to your patch and can't reproduce jobs | 17:22 |
| ssbarnea | I do see the CRs are parameters of a job, not respecting them it means it will not reproduce the build, it would be another build. | 17:23 |
| ssbarnea | let me read it again, and test it. | 17:23 |
| sshnaidm | ssbarnea, we do respect them in the job | 17:23 |
| sshnaidm | ssbarnea, we should not apply them to reproducer script itself as in your patch | 17:23 |
| ssbarnea | maybe I didn't understand the reson for revert | 17:23 |
| ssbarnea | i am reading it again and *if* the reproducer was from a CR on top of tripleo-quickstart it makes sense to use old code, even if the repository may had been updated. | 17:26 |
| ssbarnea | downloading all CRs would be key to reproduce the behavior, otherwise it would not reprocue. Still, if you want to test if newer changes made to our repos changed behaviour you would have to run the gate, generate a new reproducer script and test this one. Running a reproducer that is 2 months old should "go back in time" and attempt to do the same thing. | 17:28 |
| ssbarnea | that is why is called "reproducer" and not "rebuilder". | 17:29 |
| ssbarnea | i can understand the need/deisred for both behaviours, but at the same time I am "opionated" about what I expect to happen. I would really be upset if each time I run the same script, I would get something else as a result. | 17:32 |
| *** dtrainor_ has quit IRC | 17:44 | |
| *** dtrainor has joined #oooq | 17:58 | |
| *** tosky has quit IRC | 18:04 | |
| *** dtrainor has quit IRC | 18:15 | |
| *** dtrainor has joined #oooq | 18:22 | |
| *** dtrainor has quit IRC | 18:33 | |
| sshnaidm | ssbarnea, your patch changes code of reproducer itself, without any relation to code of the job - this behavior I called undesirable. is it clear so far? | 18:54 |
| sshnaidm | ssbarnea, reproducer is a separate tool, we improve it or break it or whatever, it shouldn't affect any job or have any relation to jobs code | 18:55 |
| sshnaidm | ssbarnea, just an example I hit - we had reproducer broken in 30 Aug, I try to reproduce job from 30 Aug, download the reproducer from the logs and run it. It fails because your patch forces it to get extras from 30 Aug including breakage of reproducer | 18:58 |
| sshnaidm | ssbarnea, it's a nonsense! | 18:58 |
| *** agopi has joined #oooq | 21:17 | |
| *** tosky has joined #oooq | 23:03 | |
| *** d0ugal has quit IRC | 23:28 | |
| *** tosky has quit IRC | 23:52 | |
| *** d0ugal has joined #oooq | 23:56 | |
Generated by irclog2html.py 2.15.3 by Marius Gedminas - find it at mg.pov.lt!