| openstackgerrit | Merged openstack-infra/zuul-jobs master: Expand docs for fetch-sphinx-output role https://review.openstack.org/507999 | 00:41 |
|---|---|---|
| *** lennyb has quit IRC | 01:11 | |
| *** lennyb has joined #zuul | 01:11 | |
| *** huangtianhua has joined #zuul | 02:44 | |
| *** zhuli has joined #zuul | 02:45 | |
| zhuli | hi folks, how to integrate zuul with github? after install zuul from feature/zuulv3 branch and follow the quick start guide of official doc, I encounter a problem when starting the zuul-scheduler service that it raised an AttributeError said 'Github' object has no attribute 'session' | 03:06 |
| zhuli | # cat /etc/zuul/zuul.conf | 03:11 |
| zhuli | [scheduler] | 03:11 |
| zhuli | tenant_config=/etc/zuul/main.yaml | 03:11 |
| zhuli | [gearman_server] | 03:11 |
| zhuli | start=true | 03:11 |
| zhuli | [gearman] | 03:11 |
| zhuli | server=127.0.0.1 | 03:11 |
| zhuli | [connection github] | 03:11 |
| zhuli | driver=github | 03:11 |
| zhuli | server=github.com | 03:11 |
| zhuli | api_token=xxxxxxxxxxxxxxxxxxxxxxxxxx | 03:11 |
| zhuli | sshkey=/home/zuul/.ssh/id_rsa | 03:11 |
| zhuli | and this is my configuration | 03:12 |
| mordred | jlk: ^^ | 03:30 |
| mordred | zhuli: as soon as the rollout is done this week we'll be working on more documentation to help folks get up and going | 03:32 |
| mordred | zhuli: I think I remember someone else having the session isue and it turned out to be a too-old version of Github3.py ... | 03:32 |
| mordred | zhuli: we unfortunately depend on an unreleased version of the library for the moment (waiting on a release) | 03:32 |
| * mordred isn't really here - goes back to being AFK | 03:33 | |
| zhuli | mordred thanks, I will update the Github3 and try again to see if this error go away | 03:41 |
| *** pbrobinson has quit IRC | 03:54 | |
| *** pbrobinson has joined #zuul | 03:59 | |
| openstackgerrit | Merged openstack-infra/zuul-jobs master: Emit a message about the final tox invocation https://review.openstack.org/507990 | 04:39 |
| *** bhavik1 has joined #zuul | 04:57 | |
| *** bhavik1 has quit IRC | 05:23 | |
| *** xinliang has quit IRC | 07:51 | |
| *** xinliang has joined #zuul | 08:05 | |
| *** xinliang has quit IRC | 08:05 | |
| *** xinliang has joined #zuul | 08:05 | |
| *** electrofelix has joined #zuul | 08:05 | |
| *** hashar has joined #zuul | 08:20 | |
| kklimonda | what's the high availability story with zuul-scheduler? | 08:24 |
| kklimonda | I can probably run two gerrit instances, and I can run a ZK cluster, zuul executors are dumb anyway, and nodepool-launchers seem to be pretty self contained too | 08:25 |
| *** jesusaur has quit IRC | 08:40 | |
| *** jesusaur has joined #zuul | 08:48 | |
| tobiash | kklimonda: zuul-scheduler is currently a single point of failure | 08:53 |
| tobiash | kklimonda: there are thoughts to eventually change zuul-scheduler in v4 into a scale out component synced via zookeeper | 08:54 |
| kklimonda | tobiash: you just brightened my day with a mention of zuulv4 ;) | 08:55 |
| tobiash | kklimonda: but that's probably a long term roadmap entry ;) | 08:56 |
| *** smyers has quit IRC | 09:03 | |
| *** smyers has joined #zuul | 09:14 | |
| *** _ari_ has quit IRC | 10:08 | |
| *** _ari_ has joined #zuul | 10:13 | |
| *** jkilpatr has quit IRC | 10:35 | |
| *** jkilpatr has joined #zuul | 10:52 | |
| *** yolanda has quit IRC | 10:54 | |
| *** yolanda has joined #zuul | 11:08 | |
| pabelanger | zhuli: you have the wrong version of github.py installed. You need to install git+https://github.com/sigmavirus24/github3.py.git@develop#egg=Github3.py | 12:00 |
| pabelanger | for some reason pip install doesn't install it correctly | 12:00 |
| zhuli | pabelanger: thanks, it works now :) | 12:40 |
| *** dkranz has quit IRC | 13:25 | |
| *** yolanda has quit IRC | 14:08 | |
| *** yolanda has joined #zuul | 14:08 | |
| *** isaacb has joined #zuul | 14:15 | |
| *** dkranz has joined #zuul | 14:41 | |
| *** SpamapS is now known as Zuulv2 | 14:43 | |
| *** Zuulv2 is now known as SpamapS | 14:43 | |
| *** nhicher has joined #zuul | 14:46 | |
| SpamapS | When the dust settles on the migration... | 14:55 |
| SpamapS | http://paste.openstack.org/show/622163/ | 14:55 |
| SpamapS | I would really like to see that ansible problem sent to the user. | 14:55 |
| SpamapS | Pretty simple "oops I forgot to git add roles/k8s-certs" since I could read the debug log | 14:55 |
| SpamapS | but from the outside, it looked like the playbook was invisible. | 14:55 |
| *** isaacb has quit IRC | 15:10 | |
| openstackgerrit | James E. Blair proposed openstack-infra/zuul feature/zuulv3: Remove "Applying template" log line https://review.openstack.org/508203 | 15:32 |
| *** jkilpatr has quit IRC | 15:34 | |
| rbergeron | I'll just share this for team spirit while y'all get things done: | 15:44 |
| rbergeron | https://www.infoworld.com/article/3228171/devops/why-ansible-is-the-future-of-red-hatand-automated-devops.html | 15:44 |
| *** hashar is now known as hasharAway | 15:51 | |
| openstackgerrit | Merged openstack-infra/zuul feature/zuulv3: Remove "Applying template" log line https://review.openstack.org/508203 | 16:06 |
| openstackgerrit | David Shrewsbury proposed openstack-infra/nodepool feature/zuulv3: Add debug line for deleting unlocked nodes https://review.openstack.org/508235 | 16:56 |
| openstackgerrit | James E. Blair proposed openstack-infra/zuul feature/zuulv3: Improve debug logging around dynamic layout loading https://review.openstack.org/508237 | 16:57 |
| openstackgerrit | James E. Blair proposed openstack-infra/zuul feature/zuulv3: Improve debug logging around dynamic layout loading https://review.openstack.org/508237 | 17:02 |
| openstackgerrit | James E. Blair proposed openstack-infra/zuul feature/zuulv3: Add ZK session timeout option https://review.openstack.org/508251 | 17:35 |
| harlowja | timeouts i hear :-P | 17:44 |
| *** jkilpatr has joined #zuul | 18:14 | |
| openstackgerrit | James E. Blair proposed openstack-infra/zuul feature/zuulv3: Wait up to 30s to lock a node https://review.openstack.org/508260 | 18:19 |
| openstackgerrit | Merged openstack-infra/zuul feature/zuulv3: Add ZK session timeout option https://review.openstack.org/508251 | 18:25 |
| openstackgerrit | Merged openstack-infra/zuul feature/zuulv3: Improve debug logging around dynamic layout loading https://review.openstack.org/508237 | 18:25 |
| openstackgerrit | Merged openstack-infra/zuul feature/zuulv3: Wait up to 30s to lock a node https://review.openstack.org/508260 | 18:25 |
| openstackgerrit | Monty Taylor proposed openstack-infra/zuul feature/zuulv3: Parent legacy jobs with artifacts legacy-publish-openstack-artifacts https://review.openstack.org/508265 | 18:26 |
| openstackgerrit | Monty Taylor proposed openstack-infra/zuul feature/zuulv3: Set BUILD_TIMEOUT environment variable in zuul_legacy_vars https://review.openstack.org/508266 | 18:26 |
| openstackgerrit | Monty Taylor proposed openstack-infra/zuul feature/zuulv3: Set BUILD_TIMEOUT environment variable in zuul_legacy_vars https://review.openstack.org/508266 | 18:26 |
| *** hasharAway is now known as hashar | 18:27 | |
| openstackgerrit | Merged openstack-infra/zuul feature/zuulv3: Set BUILD_TIMEOUT environment variable in zuul_legacy_vars https://review.openstack.org/508266 | 18:32 |
| openstackgerrit | Merged openstack-infra/nodepool feature/zuulv3: Add debug line for deleting unlocked nodes https://review.openstack.org/508235 | 18:44 |
| openstackgerrit | Monty Taylor proposed openstack-infra/zuul-jobs master: Remove the right file in revoke-sudo https://review.openstack.org/508274 | 19:11 |
| openstackgerrit | Jeremy Stanley proposed openstack-infra/zuul-jobs master: Remove the right file in revoke-sudo https://review.openstack.org/508274 | 19:25 |
| openstackgerrit | Jeremy Stanley proposed openstack-infra/zuul-jobs master: Remove the right file in revoke-sudo https://review.openstack.org/508274 | 19:36 |
| openstackgerrit | Jeremy Stanley proposed openstack-infra/zuul-jobs master: Remove the right file in revoke-sudo https://review.openstack.org/508274 | 19:53 |
| openstackgerrit | Jeremy Stanley proposed openstack-infra/zuul-jobs master: Stop running ansible-lint on playbooks https://review.openstack.org/508285 | 19:53 |
| *** hashar has quit IRC | 19:53 | |
| *** weshay has quit IRC | 20:10 | |
| openstackgerrit | Monty Taylor proposed openstack-infra/zuul feature/zuulv3: Only add BUILD_TIMEOUT if there is a timeout https://review.openstack.org/508290 | 20:11 |
| *** weshay has joined #zuul | 20:11 | |
| openstackgerrit | Jeremy Stanley proposed openstack-infra/zuul-jobs master: Stop running ansible-lint on playbooks https://review.openstack.org/508285 | 20:15 |
| openstackgerrit | Jeremy Stanley proposed openstack-infra/zuul-jobs master: Remove the right file in revoke-sudo https://review.openstack.org/508274 | 20:15 |
| openstackgerrit | Monty Taylor proposed openstack-infra/zuul feature/zuulv3: Only add BUILD_TIMEOUT if there is a timeout https://review.openstack.org/508290 | 20:22 |
| openstackgerrit | Merged openstack-infra/zuul feature/zuulv3: Parent legacy jobs with artifacts legacy-publish-openstack-artifacts https://review.openstack.org/508265 | 20:35 |
| openstackgerrit | Monty Taylor proposed openstack-infra/zuul feature/zuulv3: Only add BUILD_TIMEOUT if there is a timeout https://review.openstack.org/508290 | 20:35 |
| openstackgerrit | Merged openstack-infra/zuul feature/zuulv3: Only add BUILD_TIMEOUT if there is a timeout https://review.openstack.org/508290 | 21:08 |
| kklimonda | do you see any issue with making sure that we add user aux groups when we drop root privileges for zuul-executor? | 21:08 |
| kklimonda | I'm thinking of using puppet CA and certs to authenticate gearman connections, adding zuul user to the puppet group so it can read certs. | 21:09 |
| dmsimard | jeblair, mordred: I was pointing out to mnaser that his attempt at uploading logs should probably not be working.. but then it worked and I'm not sure I understand what's going on | 21:10 |
| dmsimard | the play is here http://logs.openstack.org/96/508296/2/check/puppet-openstack-integration-scenario004-ubuntu-xenial-puppet-4/c30f8bf/ara/file/02f6cfdf-e9a7-4487-800c-ec6bdabbcf53/#line-3 | 21:10 |
| dmsimard | and the result is here: http://logs.openstack.org/96/508296/2/check/puppet-openstack-integration-scenario004-ubuntu-xenial-puppet-4/c30f8bf/ara/result/3c5a2558-f993-482d-ba70-48986d380892/ | 21:10 |
| jeblair | kklimonda: nothing jumps out at me right now, but i'm distracted; SpamapS would be a good person to think about that | 21:11 |
| dmsimard | it looks like it ran against 'ubuntu-xenial', not localhost or anything.. | 21:11 |
| jeblair | dmsimard: that task is slightly misnamed | 21:11 |
| jeblair | dmsimard: that's "pull log files from worker onto executor" | 21:11 |
| jeblair | dmsimard: then the base job handles uploading | 21:11 |
| mordred | yah | 21:11 |
| dmsimard | that play is not from the base job though I think | 21:12 |
| jeblair | dmsimard: correct; it's auto-converted | 21:12 |
| jeblair | dmsimard: any job can pull log files onto the executor | 21:12 |
| jeblair | then whatever they pull, base job uploads | 21:12 |
| dmsimard | okay, I thought that was limited to the base job | 21:13 |
| dmsimard | i.e, that an untrusted playbook could not run https://review.openstack.org/#/c/508296/4/playbooks/upload-logs.yaml | 21:13 |
| jeblair | dmsimard: all jobs get a build-specific workspace on the executor in which they can do (almost) whatever they want. so logs all go to a scratch space. | 21:13 |
| dmsimard | is it a DoS vector to pull a series of huge files on the executor or something ? | 21:14 |
| jeblair | dmsimard: yep. if you exceed a certain size the job will be killed. | 21:14 |
| dmsimard | ok | 21:14 |
| dmsimard | thanks for clarifying | 21:14 |
| jeblair | i only hope i did :) | 21:14 |
| dmsimard | yeah I was mistakenly thinking we did not authorize pulls :) | 21:15 |
| dmsimard | mnaser: ^ fwiw you can "manually" upload things beyond what ends up being in the "logs" directory | 21:15 |
| mnaser | nice :> | 21:16 |
| *** dkranz has quit IRC | 21:45 | |
| openstackgerrit | Merged openstack-infra/zuul-jobs master: Stop running ansible-lint on playbooks https://review.openstack.org/508285 | 22:23 |
| SpamapS | kklimonda: You should be fine. The executor is just a plain old python daemon at the gearman level. | 22:42 |
| SpamapS | 2017-09-28 22:33:30.287774 | control | failed: [localhost] (item=control) => {"changed": false, "checksum": "82e1fe58c157d7df7d63b5f9cbdf2c91b9e0451e", "failed": true, "item": "control", "msg": "Aborting, target uses selinux but python bindings (libselinux-python) aren't installed!"} | 22:43 |
| SpamapS | well that's a new one on me :-P | 22:43 |
| openstackgerrit | James E. Blair proposed openstack-infra/zuul-jobs master: Fix post log location https://review.openstack.org/508346 | 22:49 |
| SpamapS | dmsimard: https://dmsimard.com/2016/01/08/selinux-python-virtualenv-chroot-and-ansible-dont-play-nice/ <-- ARGH | 22:52 |
| SpamapS | W T F | 22:52 |
| dmsimard | SpamapS: that again? :) | 22:52 |
| SpamapS | yep | 22:52 |
| dmsimard | SpamapS: I should probably edit that post | 22:52 |
| * SpamapS decides it's not worth it and just installs ansible in system python land | 22:52 | |
| dmsimard | SpamapS: it holds true for anything that can't be pip installed basically | 22:53 |
| dmsimard | SpamapS: libselinux-python, rpm-python are two examples I know offhand | 22:53 |
| * SpamapS grumbles something about maybe somebody should make those pip installable | 22:53 | |
| dmsimard | SpamapS: depending on the situation, you can also selectively set the ansible_python_interpreter per task | 22:53 |
| SpamapS | yeah no I want this to be simple | 22:53 |
| SpamapS | why does it even care? No selinux, don't set selinux. Done. | 22:53 |
| dmsimard | because it tries to import selinux to check if selinux is there | 22:54 |
| dmsimard | :D | 22:54 |
| SpamapS | we have this amazing thing | 22:54 |
| SpamapS | try | 22:54 |
| dmsimard | maybe it's worth revisiting upstream | 22:55 |
| dmsimard | if I'm not mistaken bcoca said it wasn't a bug :p | 22:55 |
| SpamapS | I mean, if I have to fix my PR's to work with python2.6, they should have to fix their modules to work with virtualenv'd pythons | 22:55 |
| dmsimard | maybe some core Ansible folks would have a better idea | 22:56 |
| dmsimard | Shrews, mordred ^ | 22:56 |
| openstackgerrit | Merged openstack-infra/zuul-jobs master: Remove the right file in revoke-sudo https://review.openstack.org/508274 | 22:59 |
| SpamapS | dmsimard: your bug report and bcoca's response were accurate. The bug, however, is trying to use selinux when it's unnecessary, IMO. | 23:00 |
| * SpamapS is reading the file action now | 23:00 | |
| dmsimard | SpamapS: it doesn't know if it has to use selinux until it's able to check if selinux is activated and it does that with the libselinux-python lib | 23:00 |
| SpamapS | dmsimard: right... but if libselinux-python isn't available......... that's a sign. ;) | 23:02 |
| dmsimard | SpamapS: I'm not sure libselinux-python is a requirement of selinux | 23:06 |
| SpamapS | hm | 23:08 |
| SpamapS | it looks like it is already conditional | 23:08 |
| SpamapS | yeah... it's as I expected | 23:09 |
| SpamapS | dmsimard: selinux _is_ enabled on the box I'm testing on. Doh. | 23:09 |
| dmsimard | ಠ_ಠ | 23:10 |
| dmsimard | Also, it's localhost, not remote host | 23:11 |
| openstackgerrit | Merged openstack-infra/zuul-jobs master: Fix post log location https://review.openstack.org/508346 | 23:14 |
| SpamapS | dmsimard: I guess if you build a centos-minimal and don't disable selinux, you get selinux ;) | 23:16 |
| SpamapS | and I'm fine with that. It doesn't seem to be breaking stuff yet | 23:16 |
| dmsimard | there's a dib element to disable selinux | 23:16 |
| SpamapS | CI should run with selinux enabled if you can swing it ;) | 23:16 |
| SpamapS | have always felt that the only reason everybody disables it is that they start with it turned off | 23:17 |
| SpamapS | oh also nobody knows how to use the tools | 23:17 |
| dmsimard | running selinux in prod is awesome and not that hard, you just need to understand how it works | 23:17 |
| dmsimard | there's even a tool that is able to write policies automatically if you're not able to do it yourself | 23:17 |
| SpamapS | it's more the bleary eyed 3am 'somfinkbroke' that gets it turned off | 23:18 |
| dmsimard | Stuff isn't supposed to break at 3am suddenly cause of selinux :) | 23:22 |
| dmsimard | It'll break in broad daylight long before that | 23:22 |
| Shrews | SpamapS: the last time i was bleary eyed at 3am, i was with you... in Atl... in some dark corner of the city | 23:30 |
Generated by irclog2html.py 2.15.3 by Marius Gedminas - find it at mg.pov.lt!