| *** jbadiapa has quit IRC | 00:10 | |
| hubbot1 | FAILING CHECK JOBS on master: tripleo-ci-centos-7-scenario000-multinode-oooq-container-upgrades, tripleo-ci-centos-7-scenario003-standalone @ https://review.openstack.org/604298, stable/pike: tripleo-ci-centos-7-scenario001-multinode-oooq-container @ https://review.openstack.org/602248, master: tripleo-ci-centos-7-scenario000-multinode-oooq-container-upgrades, tripleo-ci-centos-7-scenario003-standalone @ (1 more message) | 00:25 |
|---|---|---|
| *** tosky has quit IRC | 00:32 | |
| hubbot1 | FAILING CHECK JOBS on master: tripleo-ci-centos-7-scenario000-multinode-oooq-container-upgrades, tripleo-ci-centos-7-scenario003-standalone @ https://review.openstack.org/604298, stable/pike: tripleo-ci-centos-7-scenario001-multinode-oooq-container @ https://review.openstack.org/602248, master: tripleo-ci-centos-7-scenario000-multinode-oooq-container-upgrades, tripleo-ci-centos-7-scenario003-standalone @ (1 more message) | 02:25 |
| *** saneax has joined #oooq | 02:31 | |
| *** sanjayu_ has joined #oooq | 03:14 | |
| *** saneax has quit IRC | 03:16 | |
| *** apetrich has quit IRC | 03:16 | |
| *** ykarel|away has joined #oooq | 03:42 | |
| *** udesale has joined #oooq | 04:05 | |
| *** udesale has quit IRC | 04:15 | |
| *** udesale has joined #oooq | 04:16 | |
| hubbot1 | FAILING CHECK JOBS on master: tripleo-ci-centos-7-scenario000-multinode-oooq-container-upgrades, tripleo-ci-centos-7-scenario003-standalone @ https://review.openstack.org/604298, stable/pike: tripleo-ci-centos-7-scenario001-multinode-oooq-container @ https://review.openstack.org/602248, master: tripleo-ci-centos-7-scenario000-multinode-oooq-container-upgrades, tripleo-ci-centos-7-scenario003-standalone @ (1 more message) | 04:25 |
| *** udesale has quit IRC | 04:26 | |
| *** chkumar|out is now known as chandankumar | 05:38 | |
| *** ykarel|away is now known as ykarel | 05:38 | |
| *** ykarel has quit IRC | 05:43 | |
| *** ykarel has joined #oooq | 06:00 | |
| *** udesale has joined #oooq | 06:05 | |
| *** udesale has quit IRC | 06:12 | |
| *** udesale has joined #oooq | 06:13 | |
| *** ratailor has joined #oooq | 06:13 | |
| hubbot1 | FAILING CHECK JOBS on master: tripleo-ci-centos-7-scenario000-multinode-oooq-container-upgrades, tripleo-ci-centos-7-scenario003-standalone @ https://review.openstack.org/604298, stable/pike: tripleo-ci-centos-7-scenario001-multinode-oooq-container @ https://review.openstack.org/602248, master: tripleo-ci-centos-7-scenario000-multinode-oooq-container-upgrades, tripleo-ci-centos-7-scenario003-standalone @ (1 more message) | 06:25 |
| *** udesale has quit IRC | 06:33 | |
| *** udesale has joined #oooq | 06:33 | |
| *** brault has quit IRC | 06:33 | |
| chandankumar | marios: quiquell \o/ | 06:51 |
| chandankumar | marios: quiquell please have a look at the spec review https://review.openstack.org/630654 | 06:51 |
| chandankumar | thanks :-) | 06:51 |
| quiquell | Will do | 06:52 |
| quiquell | H | 06:54 |
| marios | o/ chandankumar ack | 06:55 |
| *** jfrancoa has joined #oooq | 07:10 | |
| *** jbadiapa has joined #oooq | 07:15 | |
| *** kopecmartin|off is now known as kopecmartin | 07:35 | |
| *** rascasoft has joined #oooq | 07:44 | |
| *** apetrich has joined #oooq | 07:48 | |
| *** ccamacho has joined #oooq | 07:56 | |
| marios | chandankumar: with batteries included :) | 08:11 |
| *** panda|off is now known as panda | 08:19 | |
| *** ccamacho has quit IRC | 08:25 | |
| hubbot1 | FAILING CHECK JOBS on master: tripleo-ci-centos-7-scenario000-multinode-oooq-container-upgrades, tripleo-ci-centos-7-scenario003-standalone @ https://review.openstack.org/604298, stable/pike: tripleo-ci-centos-7-scenario001-multinode-oooq-container @ https://review.openstack.org/602248, master: tripleo-ci-centos-7-scenario000-multinode-oooq-container-upgrades, tripleo-ci-centos-7-scenario003-standalone @ (1 more message) | 08:25 |
| *** jpena|off is now known as jpena | 08:33 | |
| chandankumar | marios: thanks, updated :-) | 08:34 |
| *** holser_ has joined #oooq | 08:35 | |
| marios | chandankumar: ack | 08:36 |
| chandankumar | marios: quiquell need one more helo | 08:39 |
| chandankumar | marios: quiquell in this review https://review.openstack.org/#/c/628421/ I am looking for a nicer way to handle rolepath https://review.openstack.org/#/c/628421/5/ansible.cfg@16 instead each time putting soemthing there if it is not discovered by TQ | 08:41 |
| quiquell | chandankumar: you role need to have a setup.py that install at correct paths | 08:41 |
| chandankumar | quiquell: it has alreadt | 08:41 |
| quiquell | so just having the requirements is enough | 08:41 |
| quiquell | then you don't need to change ansible.cfg | 08:42 |
| chandankumar | quiquell: https://github.com/openstack/openstack-ansible-os_tempest/blob/master/setup.cfg#L20 | 08:42 |
| quiquell | pip instdall -r requirements will install at correct place | 08:42 |
| quiquell | chandankumar: the you are good you don't need to hack ansible.cfg | 08:42 |
| chandankumar | quiquell: TQ venv looks for rolepath at virtualenv/usr/share | 08:43 |
| chandankumar | quiquell: but if we simply create venv localy it already prefix with /usr | 08:43 |
| *** tosky has joined #oooq | 08:44 | |
| chandankumar | quiquell: let me find the logs | 08:44 |
| quiquell | chandankumar: If we do the pip install in the venv the place should be correct | 08:45 |
| quiquell | chandankumar: If not you have to fix the role | 08:45 |
| chandankumar | quiquell: http://logs.openstack.org/00/627500/39/check/tripleo-ci-centos-7-standalone-os-tempest/53d1b64/job-output.txt.gz#_2019-01-14_11_30_20_712927 | 08:45 |
| chandankumar | quiquell: let me check another roles | 08:46 |
| quiquell | chandankumar: what's the repository of the role ? | 08:50 |
| chandankumar | quiquell: https://github.com/openstack/openstack-ansible-os_tempest | 08:54 |
| chandankumar | quiquell: I checked tripleo upgrade repo they are using /usr/share | 08:55 |
| *** ykarel is now known as ykarel|lunch | 08:55 | |
| chandankumar | quiquell: https://github.com/openstack/tripleo-upgrade/blob/master/setup.cfg#L22 | 08:55 |
| *** chem has joined #oooq | 08:55 | |
| chandankumar | quiquell: for https://github.com/openstack/ansible-role-tripleo-modify-image/blob/master/setup.cfg#L23 is share/ | 08:55 |
| chandankumar | things are incosistent here | 08:56 |
| *** dtantsur|afk is now known as dtantsur | 08:56 | |
| quiquell | chandankumar: executing pip at venv with the os_tempest role instdall at .os_tempest/share/ansible/roles/os_tempest/ | 08:57 |
| quiquell | .os_tempest is the venv dir | 08:57 |
| quiquell | chandankumar: you have to change stuff so it install at /usr/share/... | 08:58 |
| chandankumar | quiquell: yes, sure | 08:58 |
| quiquell | chandankumar: But I think you already know this | 08:58 |
| quiquell | chandankumar: I mean $VIRTUAL_ENV/usr/local/share/ansible/roles is the place the role has to install in | 08:59 |
| chandankumar | quiquell: my question is when installing in venv it already Using base prefix '/usr' | 08:59 |
| chandankumar | so i thought it is not needed but for system wide installation I need to fix to install in /usr/share | 09:00 |
| quiquell | chandankumar: but why not adding usr/ prefix to data_files ? | 09:01 |
| quiquell | What the issue with that ? | 09:01 |
| quiquell | q! | 09:02 |
| chandankumar | quiquell: no issue I am fixing that | 09:02 |
| quiquell | chandankumar: ack, so you only need this and quickstart should see it with its default ansible.cfg | 09:02 |
| quiquell | chandankumar: let me know | 09:02 |
| quiquell | chandankumar: btw, maybe you want to test the role with molecule | 09:02 |
| chandankumar | quiquell: on the list | 09:02 |
| quiquell | chandankumar: ack | 09:02 |
| *** ccamacho has joined #oooq | 09:08 | |
| quiquell | weshay: dashboard stuff workflowed, but maybe we need just a green/red box for noop instead of a guage | 09:08 |
| arxcruz|ruck | quiquell: but the noop should always be green right? | 09:09 |
| quiquell | arxcruz|ruck: yep, that's why a gauge des not make sense | 09:10 |
| quiquell | arxcruz|ruck: its kind of binary stuff | 09:10 |
| quiquell | arxcruz|ruck: green or red | 09:10 |
| quiquell | arxcruz|ruck: gauge is to show intermediate states | 09:10 |
| quiquell | arxcruz|ruck: maybe you want to try to change that | 09:10 |
| *** bogdando has joined #oooq | 09:26 | |
| *** derekh has joined #oooq | 09:33 | |
| chandankumar | quiquell: marios https://review.openstack.org/#/c/630917/ | 09:49 |
| quiquell | chandankumar: Do you have test patches ? | 09:50 |
| quiquell | chandankumar: btw after merge this maybe is good that you put there some standalone job | 09:50 |
| *** ratailor_ has joined #oooq | 09:52 | |
| *** ratailor has quit IRC | 09:55 | |
| panda | snape, snape, severus snape. | 10:06 |
| *** jaosorior has joined #oooq | 10:09 | |
| quiquell | sshnaidm: you there we have to talk about ssh keys and containers | 10:18 |
| *** udesale has quit IRC | 10:25 | |
| hubbot1 | FAILING CHECK JOBS on master: tripleo-ci-centos-7-scenario000-multinode-oooq-container-upgrades, tripleo-ci-centos-7-scenario003-standalone @ https://review.openstack.org/604298, stable/pike: tripleo-ci-centos-7-scenario001-multinode-oooq-container @ https://review.openstack.org/602248, master: tripleo-ci-centos-7-scenario000-multinode-oooq-container-upgrades, tripleo-ci-centos-7-scenario007-multinode-oooq- (2 more messages) | 10:25 |
| sshnaidm | quiquell, yep | 10:25 |
| quiquell | sshnaidm: so, looks like bind mounting with :z those kind of directories is not a good idea | 10:27 |
| quiquell | sshnaidm: it does a relabeling to share them, and removing :z does not work on centos7 (Maybe it should) fedora28 works fine without :z at ssh dir | 10:27 |
| quiquell | sshnaidm: so maybe we need to have Dockerfiles to inject ssh keys | 10:28 |
| quiquell | sshnaidm: genrated playbook that runs before | 10:28 |
| sshnaidm | quiquell, so you mean to build containers | 10:29 |
| quiquell | sshnaidm: yep at least th eones that use ssh keys | 10:29 |
| quiquell | sshnaidm: or to generate playbook with ssh embedded and run inside the container | 10:29 |
| quiquell | sshnaidm: also good thing abot building the containers is it can detect changes at config | 10:29 |
| quiquell | sshnaidm: and restart only needed services (but that's just an improvement) | 10:30 |
| *** chandankumar has quit IRC | 10:31 | |
| sshnaidm | quiquell, can we try w/o :z maybe, and with :Z, just for tests | 10:32 |
| quiquell | sshnaidm: without :z fedora28 passes but centos7 don't :-/ | 10:32 |
| sshnaidm | quiquell, I'd like to avoid building, but if no choice.. | 10:32 |
| quiquell | sshnaidm: lets try with :Z | 10:32 |
| *** arxcruz|ruck has quit IRC | 10:34 | |
| quiquell | ok, launched let's see | 10:34 |
| quiquell | sshnaidm: I will try to generate playbook with ssh keys in it and run it inside containers | 10:34 |
| *** udesale has joined #oooq | 10:35 | |
| *** chandan_kumar has joined #oooq | 10:35 | |
| sshnaidm | quiquell, you mean to change containers locally? | 10:38 |
| quiquell | sshnaidm: even issier run a playbook before zuul commands that injects keys there | 10:39 |
| quiquell | sshnaidm: we have to be sure that we don't show the playbook at CI :-) | 10:39 |
| quiquell | sshnaidm: keys are inside the playbook embedded | 10:39 |
| *** jtomasek has quit IRC | 10:40 | |
| sshnaidm | quiquell, ok | 10:40 |
| *** jtomasek has joined #oooq | 10:40 | |
| quiquell | sshnaidm: we have to be sure that keeps people's dirs intact after repro is running | 10:40 |
| quiquell | sshnaidm: so bind mounting not repo dirs are not a good idea | 10:41 |
| quiquell | sshnaidm: in general | 10:41 |
| *** ykarel|lunch is now known as ykarel | 10:42 | |
| panda | jfrancoa: question on 630228 | 10:45 |
| jfrancoa | panda: sure | 10:46 |
| jfrancoa | panda: I've just read your question in the patch. So, in an overcloud upgrade, you would first perform the undercloud upgrade to retrieve the last tht version you're going to upgrade your overcloud to | 10:47 |
| jfrancoa | panda: but as the standalone is a one-node-for-everything environment, you need to at least update python-tripleoclient and tripleo-heat-templates before running the upgrade | 10:48 |
| jfrancoa | panda: otherwise you're running the upgrade with the last release packages | 10:48 |
| sshnaidm | quiquell, we can mount readonly | 10:49 |
| quiquell | sshnaidm: Can you try it in the patch ? | 10:49 |
| jfrancoa | panda: in the end, the standalone upgrade right now is almost the same as an undercloud-upgrade, where you need to do the same...update the CLI packages and tht before running the uc upgrade | 10:49 |
| sshnaidm | quiquell, I mean for having dirs intact | 10:50 |
| jfrancoa | panda: anyway, I'll reply in the patch | 10:50 |
| panda | jfrancoa: so since the standalone is a hack, and is not consideres something that the customer sees, we cant' add part of the upgrade process for it, we resort to manually handling the part that is not considered official, then proceed with the rest ? | 10:50 |
| panda | oh | 10:50 |
| sshnaidm | quiquell, I don't think it will help in selinux case.. | 10:50 |
| quiquell | sshnaidm: no ? | 10:50 |
| sshnaidm | quiquell, no, but if we already try, will submit a patch :) | 10:51 |
| panda | jfrancoa: that is what I don't understand, why don't you take care of the package upgrade in the uprgrade command itself even for the undercloud ? | 10:51 |
| panda | jfrancoa: that is, why something that upgrades stuff is not included in the "upgrade" command :) | 10:53 |
| quiquell | sshnaidm: test it at your centos7 patch is the place of my experiments | 10:53 |
| quiquell | sshnaidm: only working solution a have there is doing a permissive selinux after repro :-( | 10:53 |
| panda | jfrancoa: honest question, I'm not criticiing the choice | 10:53 |
| jfrancoa | panda: so perform the needed packages update from the cli command? | 10:53 |
| sshnaidm | quiquell, I'll duplicate it | 10:53 |
| quiquell | sshnaidm: ack | 10:53 |
| *** brault has joined #oooq | 10:54 | |
| *** apetrich has quit IRC | 10:54 | |
| panda | jfrancoa: yes, this feels to me you're modifying the CI to get it to work, but the customers will not get this, and will have to perform another manual step (hoping it's written in the documentation) | 10:54 |
| *** udesale has quit IRC | 11:01 | |
| jfrancoa | panda: even though I'd add it in python-tripleoclient now, we would need to perform the package update by hand before the upgrade process step start (we would get the gain in next release from Train to R, because the packages updating would be already in the command itself. Anyway, the reason for this change is not to modify the CI to make it work, but the CI job wasn't testing the patch. As we create a repo with the build-test-pack | 11:02 |
| jfrancoa | le and install-built-repo doesn't update that package (https://github.com/openstack/tripleo-quickstart-extras/commit/05b5b272f38217ce09bc0e9a4a484f97bffcd2e2#diff-3214758135904ff969021b2459a0781c ) we are never using the tripleo-heat-templates patch in the upgrade | 11:02 |
| jfrancoa | panda: anyway, it's a fair point, including the packages update steps inside the upgrade process. I'll discuss it with the DFG | 11:02 |
| *** arxcruz has joined #oooq | 11:03 | |
| panda | jfrancoa: ok, is this safe to merge at this point ? | 11:03 |
| panda | noop for me, I can +W | 11:03 |
| jfrancoa | panda: but the modification in 630228 is needed so far, otherwise standalone-upgrade is all the time testing what's in master, not the gerrit patch | 11:03 |
| panda | uh ? | 11:04 |
| chandan_kumar | arxcruz: please have a look at last comment https://tree.taiga.io/project/tripleo-ci-board/us/441 we need to finalize the design for this user story | 11:05 |
| panda | jfrancoa: when a job is triggered by a gerrit patch, a new package is built and installed automatically, why you need to update ? | 11:05 |
| *** ChanServ sets mode: +o arxcruz | 11:06 | |
| arxcruz | chandan_kumar: ack | 11:06 |
| arxcruz | chandan_kumar: did you check my last comment? | 11:07 |
| *** apetrich has joined #oooq | 11:07 | |
| jfrancoa | panda: but in our case we are deploying a release less version (so we don't need the patch instealled at that point) and upgrading to the version which includes the patch (and that's the moment when we need the package to be installed) | 11:07 |
| chandan_kumar | arxcruz: sorry seen right now thanks! | 11:08 |
| panda | jfrancoa: you and your n-1 installations!! :) | 11:08 |
| panda | jfrancoa: alright | 11:08 |
| jfrancoa | panda: haha how to do an upgrade if not? | 11:09 |
| *** holser_ is now known as holser|lunch | 11:09 | |
| *** ratailor__ has joined #oooq | 11:10 | |
| panda | jfrancoa: +W | 11:12 |
| *** ratailor_ has quit IRC | 11:12 | |
| jfrancoa | panda: thanks sir | 11:12 |
| quiquell | sshnaidm: created a taiga task for the selinux labeling issue https://tree.taiga.io/project/tripleo-ci-board/taskboard/unified-sprint-4?kanban-status=1447275 | 11:15 |
| sshnaidm | quiquell, maybe we can run restorecon in containers before command.. | 11:17 |
| quiquell | sshnaidm: Humm if something fails and command is not run labels are wrong | 11:18 |
| quiquell | sshnaidm: Better ensure that we do nothing with those directories | 11:18 |
| panda | chandan_kumar: commented https://review.openstack.org/628415 | 11:24 |
| chandan_kumar | panda: sure, will update it :-) | 11:27 |
| quiquell | marios: Do you know if we can resuse "files:" block with some yaml stuff ? | 11:34 |
| panda | ssbarnea|rover: sorry you were so close ttps://review.openstack.org/627633 | 11:35 |
| *** udesale has joined #oooq | 11:36 | |
| ssbarnea|rover | panda: haha, risks of the trade. i will rebase. | 11:36 |
| ssbarnea|rover | panda: in fact, i think is wrong, the script is running with set -e, so it will stop if rebase fails. | 11:38 |
| ssbarnea|rover | if you look up a little bit, you will find a set -e | 11:39 |
| panda | ssbarnea|rover: is there also a set -x ? | 11:39 |
| ssbarnea|rover | panda: set -x is verbosity, not added by me. arguably if we want it or not but clearly outside the scope of this change. | 11:40 |
| ssbarnea|rover | panda: unless you know about a place in the source that calls "set +e" the rebase will fail and stop execution. | 11:40 |
| panda | ssbarnea|rover: I mean, the rebase fails and the run stops. How do I understand why ? | 11:40 |
| ssbarnea|rover | panda: imho the git gives a enough information why. | 11:41 |
| *** apetrich has quit IRC | 11:44 | |
| panda | ssbarnea|rover: some comments in https://review.openstack.org/628409 too | 11:47 |
| marios | quiquell: o/ not sure what you mean have smoething to point at/? | 11:48 |
| marios | jaosorior: o/ hey man can you check this please and do you know someone else with core we can ask to vote? https://review.openstack.org/#/c/628244/ i also pinged openstack-barbican a little while ago | 11:49 |
| quiquell | marios: https://markusholtermann.eu/2016/04/reference-variables-in-ansible-yaml/ | 11:49 |
| marios | rfolco: (you'll see i just sent another round of harrass-mail :) still telemetry/barbican/heat @ https://review.openstack.org/#/q/topic:replace-scen2+(status:open+OR+status:merged) | 11:50 |
| marios | quiquell: didn't read all but maybe set_fact and use that with_items that what you mean? | 11:51 |
| ssbarnea|rover | panda: can you suggest me a good message for the cleanup part? no need to worry about template, it will override the file (is default behavior) | 11:51 |
| marios | quiquell: but not something i know off hand though maybe ssbarnea|rover might know | 11:51 |
| quiquell | marios: we cannot do that with "zuul" file config | 11:51 |
| marios | quiquell: why/how you mean you cant use set_fact or isn't read properly/as you want? | 11:52 |
| quiquell | marios: The idea is to find something to reduce duplicity at those "files: " sections | 11:52 |
| marios | quiquell: OH i see what you mean | 11:52 |
| quiquell | marios: zuul config is not an ansible task | 11:52 |
| marios | quiquell: you mean in the layout.yaml | 11:52 |
| quiquell | marios: yep | 11:52 |
| quiquell | marios: maybe yaml references can help there | 11:52 |
| ssbarnea|rover | marios: quiquell: i asked on zuul channel but didn't get a rely yesterday. we should try again. | 11:53 |
| marios | quiquell: yeah would be nice if we could 'inherit' there somehow :) but for some of them is different | 11:53 |
| quiquell | ssbarnea|rover: yaml references can help ? | 11:53 |
| marios | quiquell: i think for first pass though lets get it merged and we can tidy up as next step? ssbarnea|rover | 11:53 |
| quiquell | marios: Yep, just commented it but added a +2 is just an idea | 11:53 |
| marios | quiquell: anyway is still early in sprint | 11:53 |
| marios | quiquell: ack | 11:53 |
| ssbarnea|rover | quiquell: marios: yaml references may work but we need to test, depends on the parser and how is used. as a rule when they work, they work only inside the same file. | 11:54 |
| marios | ssbarnea|rover: thanks | 11:54 |
| quiquell | ssbarnea|rover: Do you know if you can like concantenate partial list with them ?= | 11:55 |
| quiquell | ssbarnea|rover: or is all or nothing ? | 11:55 |
| ssbarnea|rover | i used that a lot with jenkins-job definitions, pretty ugly. nope I do not know. | 11:55 |
| marios | ssbarnea|rover: yeah thats a good idea (yaml refs) we use those in the upgrade/update_tasks in fact in tht already for prevention of duplicate blocks | 11:55 |
| quiquell | ssbarnea|rover: would be nice to have a zuul mechanism | 11:55 |
| marios | ssbarnea|rover: but unkown how zuul will deal with them right | 11:55 |
| ssbarnea|rover | see what parser is used by zuul and make a local test with it. we may get lucky. | 11:56 |
| ssbarnea|rover | i would not even look at zuul itself, probably is inside pyyaml | 11:56 |
| panda | ssbarnea|rover: the message can stay, I think we have to revisit that option | 11:56 |
| *** holser|lunch is now known as holser_ | 11:57 | |
| ssbarnea|rover | i am more curious if we can merge dictionaries, like: gate to include all that check contains and something more. | 11:57 |
| panda | ssbarnea|rover: and anyway -x and -e in bash for me are lazy writing :) | 11:57 |
| marios | review request please folks adding the periodic standalone-scenarios 1-4 so they start running. we can defer the addition into the master promote criteria till later but would be nice to have them running please thanks https://review.rdoproject.org/r/#/q/topic:standalone-scenario-promotion | 11:58 |
| panda | ssbarnea|rover: oh, you are changing the default in 627633 | 12:01 |
| rfolco | marios, quick question: https://review.rdoproject.org/r/#/c/18094/2/ci-scripts/dlrnapi_promoter/config/master.ini - is it agreed to add to promotion criteria now ? I thought we should add jobs to pipeline but not prevent promotion if fails. | 12:01 |
| *** apetrich has joined #oooq | 12:01 | |
| marios | rfolco: ack yeah i just -1 both of those and added a comment to that effect | 12:01 |
| panda | ssbarnea|rover: mmhh no you're not | 12:01 |
| marios | rfolco: i.e. lets land the rdo-jobs ones but not the ci-config ones | 12:01 |
| marios | rfolco: thanks | 12:01 |
| panda | I hate perlisms. | 12:02 |
| rfolco | marios, ok thank you for confirming :) | 12:02 |
| rfolco | did i hear PERL ? | 12:02 |
| panda | rfolco: no, PERLisms | 12:02 |
| panda | rfolco: like do_something || die | 12:02 |
| *** jpena is now known as jpena|lunch | 12:03 | |
| rfolco | panda, perl_you_like || don't | 12:03 |
| ssbarnea|rover | panda: i agree with you that is confusing, i made the condition wrong initially. Well, i do have ~4years of perl but these were not my best ones. I will try to use if/then/fi in the future. | 12:03 |
| panda | ssbarnea|rover: +W that one, for the other, not sure how to proceed | 12:05 |
| panda | ssbarnea|rover: the option doesn't do much, unless you skip tripleo-inventory, and I'm not sure where we are doing that anymore | 12:06 |
| quiquell | panda: commented https://review.openstack.org/#/c/617617 | 12:11 |
| quiquell | ssbarnea|rover: libvirt update images takes sooo long, maybe we can update it's image too if we put in place the images mechanism | 12:12 |
| quiquell | I mean sshnaidm sorry | 12:13 |
| sshnaidm | quiquell, which update? in nodepool-setup? | 12:13 |
| quiquell | sshnaidm: Yep I am reaching it at CI and takes sooo long | 12:13 |
| sshnaidm | quiquell, which image do you use? | 12:14 |
| quiquell | sshnaidm: Default one at libvirt... humm maybe I can take the last one from cloud ? | 12:14 |
| sshnaidm | quiquell, yeah, maybe it's old one | 12:14 |
| sshnaidm | quiquell, what is the default now there? | 12:14 |
| quiquell | sshnaidm: libvirt step passed at CI !!!! :-) | 12:15 |
| quiquell | https://softwarefactory-project.io/zuul/t/rdoproject.org/stream/d1e9ccae73f24164b80a3a9c847cc040?logfile=console.log | 12:15 |
| panda | ssbarnea|rover: also for https://review.openstack.org/620146 | 12:15 |
| quiquell | sshnaidm: https://cloud.centos.org/centos/7/images/CentOS-7-x86_64-GenericCloud-1802.qcow2 | 12:15 |
| sshnaidm | quiquell, \o/ | 12:15 |
| panda | sshnaidm: I'm always terrified by the the lack of documentation on static includes. | 12:15 |
| panda | sshnaidm: sorry for ssbarnea|rover | 12:16 |
| quiquell | panda: hahah that sound so bad :-) | 12:16 |
| sshnaidm | quiquell, let's use https://cloud.centos.org/centos/7/images/CentOS-7-x86_64-GenericCloud-1809.qcow2 then | 12:16 |
| panda | ssbarnea|rover: in the past when you wanted something to include dynamically you had to specify static: no. So I'm assuming all those roles: role inclusions are static | 12:17 |
| quiquell | sshnaidm: Will try to introspect and take the latest | 12:17 |
| panda | ssbarnea|rover: on the other hand, include_role is dynamic by default, so we are effectively changing ALL the inclusion types in ALL our includes | 12:17 |
| panda | ssbarnea|rover: I'll commmend on the change | 12:17 |
| panda | what sounds bad | 12:20 |
| panda | ? | 12:20 |
| panda | quiquell: https://review.openstack.org/629872 please, a +145, -258 change deserves some brief explanation, otherwise I'll stay here and tempest you with questions, then ask all the other cores to do the same :) | 12:23 |
| quiquell | panda: Agree didn't want to throw it at people until libvirt job passes | 12:24 |
| quiquell | panda: let me link third party to this | 12:24 |
| quiquell | panda: now that it passes so we can review the changes in the review itself and will add comments | 12:24 |
| marios | sshnaidm: o/ hey can you please revote https://review.openstack.org/#/c/630239/ | 12:25 |
| panda | quiquell: replied https://review.openstack.org/617617 | 12:25 |
| panda | quiquell: ok, thanks. | 12:25 |
| hubbot1 | FAILING CHECK JOBS on master: tripleo-ci-centos-7-scenario000-multinode-oooq-container-upgrades, tripleo-ci-centos-7-scenario007-multinode-oooq-container, tripleo-ci-centos-7-scenario003-standalone @ https://review.openstack.org/604298, stable/pike: tripleo-ci-centos-7-scenario001-multinode-oooq-container @ https://review.openstack.org/602248, master: tripleo-ci-centos-7-scenario000-multinode-oooq-container- (2 more messages) | 12:25 |
| marios | sshnaidm: we don't want it in gate as it isn't voting (alex replied thanks for reviewing) | 12:25 |
| chandan_kumar | quiquell: ostempest in action http://logs.openstack.org/00/627500/43/check/tripleo-ci-centos-7-standalone-os-tempest/04c119d/job-output.txt.gz#_2019-01-15_10_48_39_794179 | 12:26 |
| quiquell | panda: commented | 12:28 |
| quiquell | chandan_kumar: \o/ !! | 12:28 |
| marios | sshnaidm: thanks | 12:29 |
| quiquell | sshnaidm: success libvirt ! :-) 30 minutes :-( | 12:29 |
| sshnaidm | quiquell, great! | 12:30 |
| sshnaidm | quiquell, maybe with newer image will take less time | 12:30 |
| quiquell | panda, sshnaidm: How do I hock third party CI to a upstream repo ? | 12:30 |
| quiquell | want to run the reproducer CI at tq/tqe changes for libvirt/nodepool | 12:30 |
| panda | quiquell: hock ? | 12:31 |
| quiquell | panda: Run RDO jobs at upstream change | 12:31 |
| quiquell | panda: Third Party CI | 12:31 |
| panda | quiquell: you'll need to change the trigger in config | 12:32 |
| quiquell | panda: Argg config project :-( | 12:32 |
| panda | quiquell: it's time you face the dismay and terror we face every day | 12:32 |
| quiquell | panda: thanks will look | 12:32 |
| panda | quiquell: the blindness that takes you while changing those repositories | 12:33 |
| panda | quiquell: what's the job that needs to trigger ? | 12:33 |
| quiquell | panda: <advertisment> with the zuul repro you can play with trusted projects before put in place a review </advertisment> | 12:33 |
| panda | quiquell: maybe it's not in config, but in rdo-job | 12:33 |
| panda | quiquell: I'll be happy when I see it. It was in the "future plans" wasn't it ? | 12:34 |
| panda | quiquell: if the job is in rdo-job you just need to change the pipeline | 12:34 |
| quiquell | panda: differnt project | 12:35 |
| panda | quiquell: to openstack-check-rdo | 12:35 |
| panda | quiquell: subject ? | 12:35 |
| quiquell | panda: let me do some investigations | 12:37 |
| panda | quiquell: don't forget the raincoat and the black fedora. | 12:38 |
| quiquell | panda: dthanks | 12:38 |
| panda | quiquell: 617617 updated thanks | 12:39 |
| quiquell | np | 12:40 |
| *** ratailor__ has quit IRC | 12:40 | |
| panda | chandan_kumar: question about tempest in standalone | 12:41 |
| chandan_kumar | panda: yes | 12:41 |
| sshnaidm | quiquell, finally succeeded with nodepool and zuul user.. | 12:41 |
| panda | chandan_kumar: ok, thanks. | 12:41 |
| sshnaidm | quiquell, good day for reproducer :D | 12:41 |
| quiquell | sshnaidm: with the userdata ? | 12:42 |
| sshnaidm | quiquell, yeah | 12:42 |
| quiquell | sshnaidm: so we can use key-name there ? | 12:42 |
| chandan_kumar | panda: I will be updating the patch soon, currnetly updating tempestconf one | 12:42 |
| sshnaidm | quiquell, no need, we can put ssh-key in userdata itself | 12:42 |
| sshnaidm | quiquell, just need to merge this patch to nodepool | 12:43 |
| quiquell | sshnaidm: it means pre cooking the images ? | 12:43 |
| quiquell | sshnaidm: pass me the review | 12:43 |
| *** zul has joined #oooq | 12:43 | |
| sshnaidm | quiquell, yeah, it requires a cloud-init installed | 12:43 |
| panda | chandan_kumar: no, so my question was | 12:43 |
| quiquell | Let's use marios technics on harrasing reviews | 12:43 |
| panda | chandan_kumar: what are we testing with tempest in standalone ? | 12:43 |
| sshnaidm | quiquell, well, I need to prepare tests for it before | 12:43 |
| quiquell | sshnaidm: so the placen with that is install there our tripleo ci pub keys for zuul user but with upstream image is that it ? | 12:44 |
| panda | chandan_kumar: how do you understand what's installed and what to test ? undercloud has a defined set of services installed, standalone is a free for all entropic simulation | 12:44 |
| chandan_kumar | panda: that thing, we need to figure out | 12:45 |
| sshnaidm | quiquell, the possible workflow is - download image, install cloud-init, upload images, share image, then you just put userdata in nodepool config on client side | 12:45 |
| chandan_kumar | panda: on osa side, they use group_var | 12:45 |
| quiquell | sshnaidm: ahh shit understood, that super cool | 12:45 |
| sshnaidm | quiquell, ssh-key is taken like lookup('file', 'id_rsa.pub') and added to nodepool config | 12:46 |
| quiquell | sshnaidm: how big where those images ? | 12:46 |
| quiquell | sshnaidm: libvirt one is 800 gb | 12:46 |
| sshnaidm | quiquell, 6GB | 12:46 |
| panda | chandan_kumar: ok | 12:46 |
| sshnaidm | quiquell, they're pretty big.. | 12:46 |
| quiquell | I mean 800mb | 12:47 |
| *** panda is now known as panda|lunch | 12:47 | |
| sshnaidm | quiquell, yeah, I understood :D | 12:47 |
| quiquell | sshnaidm: Maybe we can give the option at libvirt of setip things up or just download it | 12:47 |
| sshnaidm | quiquell, will look into image-manager, maybe will be easier to use it.. | 12:48 |
| quiquell | sshnaidm: because seting thins up takes a lot of time too | 12:48 |
| quiquell | sshnaidm: ack, not bad to have a generator for images | 12:48 |
| quiquell | sshnaidm: even out of reproducer | 12:48 |
| chandan_kumar | panda: if we are talking about this one https://review.openstack.org/#/c/628415/19/playbooks/multinode-standalone.yml@61 we might use OS::TripleO::Services::CinderApi for setting the tempest_service_available vars or we need to find something other thing | 12:48 |
| chandan_kumar | panda|lunch: I am thinking of first goal is to run os_tempest in ci | 12:48 |
| sshnaidm | quiquell, I think.. if we'll use nodepool-setup with libvirt, why not to use it with cloud images too | 12:48 |
| chandan_kumar | panda|lunch: next step is to go with cleanup | 12:48 |
| quiquell | sshnaidm: yep | 12:49 |
| quiquell | sshnaidm: well let's go step by step | 12:49 |
| sshnaidm | quiquell, or maybe to make it optional.. by default - cloud image with our modifications, and by request - nodepool image | 12:49 |
| sshnaidm | quiquell, yeah, need to check other options now | 12:50 |
| quiquell | sshnaidm: "nodepool iamge" == role setup images ? | 12:50 |
| sshnaidm | quiquell, what is "role setup images"? | 12:50 |
| quiquell | sshnaidm: vanilla centos + libvirt role + nodepool setup role | 12:50 |
| sshnaidm | quiquell, well, we mimic nodepool image, but it's still not equivalent | 12:51 |
| quiquell | sshnaidm: Also libvirt + fedora28 will neeed a lot of adaptations... | 12:51 |
| sshnaidm | quiquell, for python3? | 12:52 |
| quiquell | sshnaidm: python3 + os things | 12:52 |
| quiquell | sshnaidm: You don't know until you don't run :-/ | 12:52 |
| quiquell | s/don't// | 12:52 |
| sshnaidm | quiquell, maybe we can postpone it, not in mvp | 12:53 |
| quiquell | sshnaidm: sure is not in mvp it will take time to adapt libvirt to fedora28 | 12:53 |
| *** chandan_kumar has quit IRC | 12:58 | |
| *** chandan_kumar has joined #oooq | 12:59 | |
| *** jpena|lunch is now known as jpena | 12:59 | |
| chandan_kumar | weshay: Hello | 13:01 |
| chandan_kumar | weshay: add new role to tq meeting | 13:01 |
| weshay | chandan_kumar, hey.. joining.. had to make some coffee | 13:01 |
| chandan_kumar | weshay: sure | 13:01 |
| weshay | chandan_kumar, https://github.com/openstack/tripleo-quickstart/blob/master/quickstart-extras-requirements.txt | 13:04 |
| quiquell | weshay: libvirt CI pass :-) | 13:07 |
| weshay | woot | 13:07 |
| weshay | rrcockpit :) | 13:07 |
| weshay | love it | 13:07 |
| quiquell | panda|lunch: added comments to the libvirt patches they are ready for review now | 13:08 |
| weshay | chandan_kumar, https://docs.openstack.org/tripleo-quickstart/latest/working-with-extras.html | 13:13 |
| weshay | https://docs.openstack.org/tripleo-quickstart/latest/working-with-extras.html#developing-new-roles | 13:14 |
| chandan_kumar | weshay: https://specs.openstack.org/openstack/tripleo-specs/specs/stein/ostempest-tripleo.html | 13:15 |
| arxcruz | weshay: why sometimes the jobs on periodic are skipped? | 13:18 |
| arxcruz | me and ssbarnea|rover was wondering that last week | 13:18 |
| *** rlandy has joined #oooq | 13:19 | |
| chandan_kumar | weshay: https://tree.taiga.io/project/tripleo-ci-board/us/441?milestone=215191 | 13:21 |
| *** arxcruz is now known as arxcruz|ruck | 13:21 | |
| *** trown|outtypewww is now known as trown | 13:29 | |
| *** udesale has quit IRC | 13:36 | |
| *** udesale has joined #oooq | 13:36 | |
| weshay | marios, good morning sir | 13:40 |
| weshay | let me know if you have any interest in chatting about the upgrade job and those few reviews | 13:41 |
| weshay | panda|lunch, we should chat too when ur back | 13:41 |
| marios | weshay: ack sure, wana do it now? | 13:41 |
| weshay | now is fine by me | 13:41 |
| weshay | if I'm awak enough.. my blue | 13:41 |
| marios | weshay: or we can talk on community call too? whatever | 13:42 |
| * marios unpackign headphones | 13:42 | |
| weshay | marios, ya.. better idea | 13:42 |
| weshay | ur right | 13:42 |
| weshay | see you then | 13:42 |
| weshay | TONIGHT WE DINE IN HELL | 13:42 |
| marios | weshay: ack wfm (want to revisit that bug anyway since bandini updated earlier but didnt context switch yet) | 13:42 |
| *** panda|lunch is now known as panda | 13:43 | |
| panda | weshay: marios community meeting conflicts with UA meeting for me. | 13:44 |
| weshay | panda, that's fine.. we need to chat about containers | 13:44 |
| * weshay hangs out in my blue | 13:45 | |
| weshay | so ronrie | 13:45 |
| weshay | so sad and roooonrie | 13:45 |
| weshay | marios, ah.. so the multinode and standalone are probably hitting the same issue | 13:46 |
| marios | weshay: i saw your note about the multinode fail but didn't dig there . we are still missing something ... there must be some default switched on the 8th cos that job was green for a while | 13:47 |
| marios | weshay: i mean, if it isn't the rabbit thing (even after bandini fix hits the same again)( | 13:48 |
| weshay | panda, https://trunk.rdoproject.org/fedora/current-tripleo/ | 13:49 |
| weshay | marios, I guess we could diff against scen02 | 13:53 |
| weshay | the only diff should be... | 13:53 |
| weshay | sahara | 13:53 |
| weshay | mistral | 13:53 |
| weshay | and designate | 13:54 |
| weshay | or 4 | 13:54 |
| ssbarnea|rover | arxcruz|ruck: i think skip is done when there is a job already running (last one didn't finish), right weshay? | 14:02 |
| arxcruz|ruck | ssbarnea|rover: some periodic jobs today were skipped | 14:03 |
| *** apetrich has quit IRC | 14:07 | |
| lyarwood | panda: sorry, didn't want to spam in the channel during the meeting | 14:11 |
| lyarwood | panda: so I've spent my morning trying to work the container build flow into an existing reproduer I had to hand without any luck | 14:11 |
| lyarwood | panda: there's a build job on tripleo-common that is successfully building the image but it doesn't have an associated reproducer script | 14:12 |
| ssbarnea|rover | arxcruz|ruck: I do not see skipped for today on https://review.rdoproject.org/zuul/builds?pipeline=openstack-periodic | 14:12 |
| *** apetrich has joined #oooq | 14:12 | |
| arxcruz|ruck | ssbarnea|rover: https://review.rdoproject.org/zuul/status check the periodic running | 14:14 |
| arxcruz|ruck | a bunch of jobs marked as skipped | 14:14 |
| arxcruz|ruck | ssbarnea|rover: plus on the link you point, it's not sort by date | 14:14 |
| panda | lyarwood: mmhhh, if you can't build the image, I can't imagine how long it will take me :) | 14:15 |
| ssbarnea|rover | arxcruz|ruck: oops, i was sure it was sorted by date. | 14:15 |
| panda | lyarwood: I'll start the build, reproducing the promotion job, I'll ping you if I encounter problems. | 14:16 |
| ssbarnea|rover | arxcruz|ruck: me clueless, and the table is not helping at all, all links on the skipped lines are broken and there is no time recorded. so clearly not helping much. | 14:17 |
| arxcruz|ruck | ssbarnea|rover: that's why i would like to know what's happening why/how | 14:18 |
| ssbarnea|rover | arxcruz|ruck: lets ask in #rdo channel | 14:18 |
| ssbarnea|rover | unrelated: just discovered that ansible-galaxy is being replace with mazer : https://github.com/ansible/galaxy/issues/1455#issuecomment-454401566 -- made a note to play with it over the weekend. | 14:20 |
| panda | weshay: reported results for fedora mixed job to fedora and centos, but the promoter script is not resilient enough to promote a fedora hash without having any image available. It will just fail saying " ther is no overcloud image for this hashm and WHERE ARE MY CONTAINERS???" then explode. So manual promotion it is. | 14:22 |
| hubbot1 | FAILING CHECK JOBS on master: tripleo-ci-centos-7-scenario000-multinode-oooq-container-upgrades, tripleo-ci-centos-7-scenario007-multinode-oooq-container, tripleo-ci-centos-7-scenario003-standalone @ https://review.openstack.org/604298, stable/pike: tripleo-ci-centos-7-scenario001-multinode-oooq-container @ https://review.openstack.org/602248, master: tripleo-ci-centos-7-scenario000-multinode-oooq-container- (2 more messages) | 14:25 |
| weshay | brb.. folks my son's bday | 14:30 |
| *** vinaykns has joined #oooq | 14:33 | |
| ykarel | Happy b'day to junior weshay | 14:35 |
| rfolco | CI community meeting (office hours) starts now at https://bluejeans.com/4113567798 | 14:40 |
| weshay | http://logs.openstack.org/63/630363/4/check/tripleo-ci-centos-7-scenario000-multinode-oooq-container-upgrades/2495ab4/job-output.txt.gz | 15:02 |
| *** ykarel has quit IRC | 15:04 | |
| *** quiquell is now known as quiquell|off | 15:07 | |
| weshay | arxcruz|ruck, hey you around | 15:15 |
| arxcruz|ruck | weshay: i'm always around | 15:15 |
| arxcruz|ruck | :) | 15:15 |
| weshay | arxcruz|ruck, join the https://bluejeans.com/4113567798 | 15:16 |
| *** ykarel has joined #oooq | 15:23 | |
| panda | weshay: https://trunk.rdoproject.org/fedora28-master/?C=M;O=D current-tripleo/ 2019-01-14 09:13 | 16:13 |
| *** dougbtv has joined #oooq | 16:15 | |
| *** ccamacho has quit IRC | 16:16 | |
| weshay | panda, thank you sir | 16:17 |
| *** ykarel is now known as ykarel|away | 16:23 | |
| hubbot1 | FAILING CHECK JOBS on master: tripleo-ci-centos-7-scenario000-multinode-oooq-container-upgrades, tripleo-ci-centos-7-scenario007-multinode-oooq-container, tripleo-ci-centos-7-scenario003-standalone @ https://review.openstack.org/604298, stable/pike: tripleo-ci-centos-7-scenario001-multinode-oooq-container @ https://review.openstack.org/602248, master: tripleo-ci-centos-7-scenario000-multinode-oooq-container- (2 more messages) | 16:26 |
| weshay | ykarel|away, we should chat | 16:26 |
| panda | arxcruz|ruck: ssbarnea|rover heads up I uploaded a new container which didn't pass through promotion https://hub.docker.com/r/tripleomaster/centos-binary-placement-api/tags | 16:34 |
| panda | arxcruz|ruck: ssbarnea|rover but I had to tag it with current-tripleo and the latest hash, otherwise it would not work | 16:34 |
| panda | arxcruz|ruck: ssbarnea|rover if it causes any problem, ping me | 16:35 |
| panda | arxcruz|ruck: ssbarnea|rover or lyarwood | 16:35 |
| arxcruz|ruck | panda: ack | 16:35 |
| lyarwood | yup, I'll be AFK for a few hours but back again this evening if there are any issues with this being there | 16:39 |
| ykarel|away | weshay, can we do tomorrow? it's late here | 16:41 |
| weshay | ykarel|away, oh ya.. no rush man :) | 16:43 |
| weshay | sorry to ping you | 16:43 |
| ykarel|away | np :) | 16:43 |
| *** hamzy_ has quit IRC | 16:43 | |
| *** udesale has quit IRC | 17:04 | |
| rlandy | panda: re: https://review.openstack.org/#/c/622575/7/roles/run-test/tasks/main.yaml | 17:07 |
| rlandy | is there any other way to pass vars to the inner playbooks other than what I have commented on? | 17:08 |
| rlandy | am I missing something here? | 17:08 |
| rlandy | reproducer case only | 17:08 |
| *** derekh has quit IRC | 17:22 | |
| weshay | kopecmartin++ | 17:26 |
| hubbot1 | weshay: kopecmartin's karma is now 5 | 17:26 |
| weshay | thanks for chasing the doc issues | 17:26 |
| *** dtantsur is now known as dtantsur|afk | 17:28 | |
| panda | rlandy: the quickest way would be to rename featureset_override to just "overrides" in all our jobs. | 17:33 |
| rlandy | panda: I have defined it in allowed_overrides | 17:33 |
| rlandy | will raise this in our next meeting | 17:33 |
| panda | rlandy: the less intrusive way would be to add a generic extra_vars variable that will be handled like the featureset_override | 17:34 |
| panda | rlandy: yes, but that variable needs to be set in the featureset_override dict in the job variable | 17:34 |
| panda | and we need to put the dlrn_hash_tag, featureset override doesn't make sense anymore, beause we are not overriding featureset anymore | 17:35 |
| rlandy | panda: what worries me about extra_vars is that it's already over used | 17:35 |
| rlandy | look at how it is defined | 17:35 |
| rlandy | in diff cases | 17:35 |
| *** dtantsur|afk has quit IRC | 17:36 | |
| *** ykarel has joined #oooq | 17:37 | |
| *** ykarel|away has quit IRC | 17:38 | |
| panda | rlandy: if you're passing it as vars: to the zuul job directly, it will end up in @{{ workspace }}/logs/zuul-variables.yaml | 17:38 |
| rlandy | that won;t help | 17:38 |
| panda | rlandy: which is already the highest extra vars in the inner ansible | 17:38 |
| rlandy | so then we could just pass the vars directly? | 17:39 |
| * rlandy is never sure about this | 17:39 | |
| panda | rlandy: yes everything that you pass in zuul as vars for the job is dumped in @{{ workspace }}/logs/zuul-variables.yaml and passed to the inner ansible-playbook as last configuration, that is it will override everything | 17:40 |
| rlandy | ok - then we can pass directly | 17:40 |
| panda | rlandy: the problems is that everything will need to be addressed as zuul.dlrn_hash | 17:40 |
| panda | zuul.dlrn_hash* | 17:41 |
| rlandy | hmmm | 17:41 |
| rlandy | that won;t work | 17:41 |
| panda | which again is on a different namespace ... | 17:41 |
| rlandy | the override is better | 17:41 |
| panda | I need a redesign on the featureset at this point ... I'm not sure where we're heading. The override list is growing, we are going to add override for variables that are not in the featuresets, and we may end up having too many places where we can specify a variable and not being sure which one wins in the end | 17:44 |
| *** sanjayu_ has quit IRC | 17:50 | |
| *** trown is now known as trown|lunch | 17:52 | |
| *** jpena is now known as jpena|off | 17:59 | |
| rlandy | panda: understood - for the moment I am designing the reproducer with what is available - allowed_overrides | 18:00 |
| weshay | panda, rlandy what var are you guuys talking about? | 18:02 |
| weshay | that needs to be overridden in ALL jobs | 18:02 |
| panda | weshay: not all jobs | 18:03 |
| panda | weshay: https://review.openstack.org/#/c/622575/7/roles/run-test/tasks/main.yaml@19 | 18:03 |
| panda | the problem is how to inject a variable from zuul configuration to inner ansible call in a way taht can be used by quickstart roles | 18:04 |
| weshay | oh.. this is cool | 18:04 |
| panda | a bit less cool if you picture how this variable is going to be passed | 18:07 |
| rlandy | weshay: quiquell|off and I today went through the design of the reproducer scipt | 18:07 |
| *** holser_ is now known as holser|eod | 18:08 | |
| rlandy | script | 18:08 |
| rlandy | what gets created etc. | 18:08 |
| weshay | cool | 18:08 |
| rlandy | I am working on the review now | 18:08 |
| rlandy | what goes into playbooks vs. bash etc | 18:08 |
| weshay | so the script that folks download w/ a dlrn hash will be setting the override var | 18:08 |
| rlandy | for the reproducer we need to pass a couple things | 18:08 |
| weshay | for dlrn | 18:08 |
| weshay | ya | 18:08 |
| weshay | this is nice | 18:08 |
| weshay | +2 | 18:08 |
| rlandy | getting there :) | 18:08 |
| * rlandy is still explaining | 18:09 | |
| rlandy | dlrn hash OR releases file | 18:09 |
| rlandy | depending on fs | 18:09 |
| weshay | k | 18:09 |
| rlandy | and depends-on for the job itself | 18:09 |
| weshay | someone remind me.. check-rdo to recheck 3rd party? | 18:09 |
| weshay | what's the keyword | 18:09 |
| * rlandy thinks so - checking | 18:10 | |
| panda | check-rdo | 18:10 |
| panda | weshay: ^ | 18:10 |
| weshay | thanks | 18:10 |
| weshay | lol | 18:10 |
| panda | or verify-rdo | 18:10 |
| rlandy | weshay: https://github.com/rdo-infra/review.rdoproject.org-config/blob/2dcdfd80d0a30e8c11142dfeb858982cd8ea8b46/zuul.d/upstream.yaml#L21 | 18:10 |
| rlandy | recheck|check-rdo|verify-rdo | 18:10 |
| panda | or recheck if you don't mind checking everything else too | 18:10 |
| weshay | ya | 18:11 |
| weshay | for some reason rdo is not trigging on everything in the last 24hrs | 18:11 |
| weshay | I think things may be reset w/o us knowing | 18:11 |
| weshay | so it goes | 18:12 |
| *** holser|eod has quit IRC | 18:12 | |
| *** brault has quit IRC | 18:14 | |
| *** brault has joined #oooq | 18:15 | |
| panda | and I thought the day was going well. | 18:19 |
| *** panda is now known as panda|off | 18:19 | |
| *** brault has quit IRC | 18:20 | |
| *** ykarel has quit IRC | 18:25 | |
| *** bogdando has quit IRC | 18:26 | |
| hubbot1 | FAILING CHECK JOBS on master: tripleo-ci-centos-7-scenario000-multinode-oooq-container-upgrades, tripleo-ci-centos-7-scenario007-multinode-oooq-container, tripleo-ci-centos-7-scenario003-standalone, tripleo-ci-centos-7-ovb-3ctlr_1comp-featureset053 @ https://review.openstack.org/604298, stable/pike: tripleo-ci-centos-7-scenario001-multinode-oooq-container @ https://review.openstack.org/602248, master: tripleo- (2 more messages) | 18:26 |
| *** chem has quit IRC | 18:43 | |
| *** hamzy_ has joined #oooq | 18:44 | |
| weshay | rlandy, when you have a sec.. https://review.openstack.org/#/c/628024/ sshnaidm | 18:51 |
| *** kopecmartin is now known as kopecmartin|off | 18:52 | |
| *** hamzy_ has quit IRC | 18:59 | |
| *** hamzy_ has joined #oooq | 19:00 | |
| *** hamzy_ has quit IRC | 19:13 | |
| *** rlandy is now known as rlandy|brb | 19:14 | |
| *** trown|lunch is now known as trown | 19:19 | |
| *** hamzy_ has joined #oooq | 19:31 | |
| *** jfrancoa has quit IRC | 19:31 | |
| *** rlandy|brb is now known as rlandy | 19:33 | |
| *** dougbtv has quit IRC | 19:44 | |
| *** brault has joined #oooq | 20:06 | |
| *** brault has quit IRC | 20:11 | |
| *** hamzy_ has quit IRC | 20:12 | |
| ssbarnea|rover | weshay: can you run this for me and tell if it worked? https://gist.github.com/ssbarnea/c6100ec231055a5c861acbfebe089108 | 20:23 |
| weshay | k | 20:25 |
| * weshay tries | 20:25 | |
| hubbot1 | FAILING CHECK JOBS on master: tripleo-ci-centos-7-scenario000-multinode-oooq-container-upgrades, tripleo-ci-centos-7-scenario007-multinode-oooq-container, tripleo-ci-centos-7-scenario003-standalone, tripleo-ci-centos-7-ovb-3ctlr_1comp-featureset053 @ https://review.openstack.org/604298, stable/pike: tripleo-ci-centos-7-scenario001-multinode-oooq-container @ https://review.openstack.org/602248, master: tripleo- (2 more messages) | 20:26 |
| weshay | ssbarnea|rover, fatal: [localhost]: FAILED! => {"changed": false, "msg": "Error creating container: 400 Client Error: Bad Request (\"{\"message\":\"No command specified\"}\")"} | 20:28 |
| *** rfolco has quit IRC | 20:28 | |
| ssbarnea|rover | weshay: and if you replace fedora:28 with centos:7 ? ... it works, right? | 20:28 |
| weshay | will do so now | 20:28 |
| weshay | ssbarnea|rover, yup | 20:29 |
| weshay | centos-7 works | 20:29 |
| weshay | I wonder if there is another name | 20:29 |
| weshay | https://developer.fedoraproject.org/tools/docker/docker-usage.html | 20:30 |
| ssbarnea|rover | weshay: what can I say. Now try to add "command: ls" | 20:30 |
| ssbarnea|rover | and the fedora:28 will work. | 20:30 |
| ssbarnea|rover | it happens only with fedora* when command is null (default) | 20:30 |
| ssbarnea|rover | not sure who to blame for this.... unexpected surprise | 20:31 |
| weshay | odd | 20:32 |
| weshay | fedora should equal fedora:latest | 20:32 |
| weshay | probably docker is a like f u red hat | 20:32 |
| weshay | w/ your kubernettes and podman | 20:33 |
| ssbarnea|rover | weshay: is more of a [f]edora you, because centos works :D | 20:33 |
| weshay | heh | 20:33 |
| weshay | rlandy, https://review.openstack.org/#/c/622575/ looks good re: 3rd party | 20:34 |
| ssbarnea|rover | molecule uses this module, now we need to narrow down who is the culprit: ansible module, python docker, docker itself or the fedora image having somethign weird in it. | 20:34 |
| weshay | oh ya | 20:34 |
| weshay | ssbarnea|rover, so why not use the centos7-base and fedora-base image we create | 20:35 |
| weshay | in docker.io | 20:35 |
| * weshay wonders if thhose are busted | 20:35 | |
| rlandy | weshay: thanks - still part of the whole reproducer design we will be bringing to Thurs meeting | 20:35 |
| ssbarnea|rover | weshay: I doubt. it has somethign to do with command. if you override command it works. | 20:36 |
| weshay | rock | 20:36 |
| ssbarnea|rover | i coult put "command: echo" and avoid the bug | 20:36 |
| weshay | meh.. weird | 20:37 |
| ssbarnea|rover | ended up creating https://github.com/ansible/ansible/issues/50951 | 20:44 |
| weshay | arxcruz|ruck, tomorrow.. saw this fail in tempest https://ci.centos.org/view/rdo/view/promotion-pipeline/job/rdo_trunk-promote-master-current-tripleo/428/ | 21:05 |
| * weshay wonders if there is a no-op tempest test.. you need one pass to generate results | 21:06 | |
| weshay | oh there was a pass. | 21:07 |
| weshay | arxcruz|ruck, def. look into this tomorrow https://ci.centos.org/artifacts/rdo/jenkins-tripleo-quickstart-promote-master-current-tripleo-delorean-minimal-446/undercloud/home/stack/tempest.log.gz | 21:07 |
| ssbarnea|rover | weshay: got progress, proved to be broken fedora image, bug raised at https://github.com/fedora-cloud/docker-brew-fedora/issues/62 -- wondering when/if it will be fixed. | 21:08 |
| arxcruz|ruck | weshay: i'm on pto tomorrow... | 21:09 |
| weshay | bah k | 21:09 |
| weshay | ssbarnea|rover, heh.. no need to test or ci that stuff.. geez | 21:10 |
| weshay | arxcruz|ruck, I'll take care of it, have a good night | 21:10 |
| weshay | rlandy, can I close https://bugs.launchpad.net/tripleo/+bug/1810269 | 21:11 |
| openstack | Launchpad bug 1810269 in tripleo "tempest.api.compute.admin.test_agents.AgentsAdminTestJSON tests failing with Internal Server Error" [Critical,Triaged] - Assigned to Ronelle Landy (rlandy) | 21:11 |
| rlandy | weshay: has phase 1 run yet? | 21:12 |
| rlandy | or do you see tempest running on fs020 | 21:12 |
| rlandy | if som yes | 21:12 |
| weshay | I just rekicked it.. but not w/ new tripleo | 21:12 |
| rlandy | so | 21:12 |
| rlandy | if we haven't see tempest run, we don;t know it's fixed | 21:12 |
| weshay | ah. .we need a new nova | 21:13 |
| weshay | rfc3986-1.2.0 is available in Repo[1], now needs promotion of nova containers(to current-tripleo) to get RDO phase1 clear. | 21:13 |
| rlandy | weshay: very preliminary - but you can see the design starting here ... https://review.openstack.org/#/c/631067/ | 21:19 |
| * rlandy has lots to fill in | 21:19 | |
| weshay | k | 21:20 |
| *** hamzy_ has joined #oooq | 21:21 | |
| *** hamzy_ has quit IRC | 21:29 | |
| weshay | holy cow we had a ton of old prmotion blockers | 21:30 |
| *** trown is now known as trown|outtypewww | 21:41 | |
| Tengu | raaaaahhhhh | 22:24 |
| Tengu | podman ;_;. such a nasty boy | 22:24 |
| hubbot1 | FAILING CHECK JOBS on master: tripleo-ci-centos-7-scenario000-multinode-oooq-container-upgrades, tripleo-ci-centos-7-scenario003-standalone, tripleo-ci-centos-7-ovb-3ctlr_1comp-featureset053 @ https://review.openstack.org/604298, stable/pike: tripleo-ci-centos-7-scenario001-multinode-oooq-container @ https://review.openstack.org/602248, master: tripleo-ci-centos-7-scenario000-multinode-oooq-container-upgrades, (2 more messages) | 22:26 |
| *** brault has joined #oooq | 22:45 | |
| *** brault has quit IRC | 22:45 | |
| *** vinaykns has quit IRC | 22:50 | |
| *** jtomasek has quit IRC | 23:01 | |
| *** rascasoft has quit IRC | 23:06 | |
| *** rlandy has quit IRC | 23:33 | |
| *** rascasoft has joined #oooq | 23:34 | |
| *** rascasoft has quit IRC | 23:39 | |
Generated by irclog2html.py 2.15.3 by Marius Gedminas - find it at mg.pov.lt!