| *** jamesmcarthur has quit IRC | 00:00 | |
| openstackgerrit | Ian Wienand proposed openstack/diskimage-builder master: simple-init: allow for NetworkManager support https://review.openstack.org/619120 | 00:07 |
|---|---|---|
| *** jamesmcarthur has joined #openstack-infra | 00:07 | |
| *** tosky has quit IRC | 00:12 | |
| *** armax has joined #openstack-infra | 00:13 | |
| *** mriedem has quit IRC | 00:17 | |
| openstackgerrit | Ian Wienand proposed openstack/diskimage-builder master: simple-init: allow for NetworkManager support https://review.openstack.org/619120 | 00:19 |
| *** vinaykns has quit IRC | 00:21 | |
| *** yamamoto has quit IRC | 00:34 | |
| *** gfidente has quit IRC | 00:34 | |
| *** longkb has joined #openstack-infra | 00:38 | |
| *** longkb has quit IRC | 00:38 | |
| openstackgerrit | Ian Wienand proposed openstack/diskimage-builder master: simple-init: allow for NetworkManager support https://review.openstack.org/619120 | 00:41 |
| *** scas has quit IRC | 00:44 | |
| *** hamzy has joined #openstack-infra | 00:44 | |
| openstackgerrit | Ian Wienand proposed openstack-infra/glean master: [wip] NetworkManager support https://review.openstack.org/618964 | 00:49 |
| openstackgerrit | Ian Wienand proposed openstack/diskimage-builder master: simple-init: allow for NetworkManager support https://review.openstack.org/619120 | 00:57 |
| *** jamesmcarthur has quit IRC | 01:05 | |
| *** jamesmcarthur has joined #openstack-infra | 01:08 | |
| openstackgerrit | Ian Wienand proposed openstack-infra/glean master: [wip] NetworkManager support https://review.openstack.org/618964 | 01:14 |
| *** jamesmcarthur has quit IRC | 01:17 | |
| *** jamesmcarthur has joined #openstack-infra | 01:18 | |
| *** eernst has joined #openstack-infra | 01:19 | |
| *** eernst has quit IRC | 01:21 | |
| *** jamesmcarthur has quit IRC | 01:22 | |
| *** roman_g has quit IRC | 01:26 | |
| *** eernst has joined #openstack-infra | 01:40 | |
| *** eernst has quit IRC | 01:45 | |
| *** jamesmcarthur has joined #openstack-infra | 01:56 | |
| *** mtreinish has quit IRC | 02:13 | |
| *** scas has joined #openstack-infra | 02:17 | |
| *** mrsoul has joined #openstack-infra | 02:30 | |
| *** ramishra has quit IRC | 02:41 | |
| openstackgerrit | Ian Wienand proposed openstack-infra/nodepool master: Add Fedora 29 testing https://review.openstack.org/618671 | 02:49 |
| *** anteaya has quit IRC | 02:51 | |
| *** fuentess has quit IRC | 03:40 | |
| *** bhavikdbavishi has joined #openstack-infra | 03:58 | |
| *** bhavikdbavishi1 has joined #openstack-infra | 04:01 | |
| *** bhavikdbavishi has quit IRC | 04:03 | |
| *** bhavikdbavishi1 is now known as bhavikdbavishi | 04:03 | |
| *** yamamoto has joined #openstack-infra | 04:05 | |
| openstackgerrit | Ian Wienand proposed openstack-infra/nodepool master: Add Fedora 29 testing https://review.openstack.org/618671 | 04:11 |
| openstackgerrit | Ian Wienand proposed openstack-infra/nodepool master: Add Fedora 29 testing https://review.openstack.org/618671 | 04:12 |
| *** psachin has joined #openstack-infra | 04:14 | |
| *** ykarel|away has joined #openstack-infra | 04:15 | |
| *** ykarel|away is now known as ykarel | 04:15 | |
| *** udesale has joined #openstack-infra | 04:17 | |
| *** rlandy has quit IRC | 04:31 | |
| *** janki has joined #openstack-infra | 04:42 | |
| *** yamamoto has quit IRC | 05:20 | |
| openstackgerrit | Merged openstack-infra/project-config master: Retire openstack-ansible-os_monasca-ui https://review.openstack.org/617321 | 05:21 |
| openstackgerrit | Ian Wienand proposed openstack/diskimage-builder master: simple-init: allow for NetworkManager support https://review.openstack.org/619120 | 05:21 |
| openstackgerrit | Ian Wienand proposed openstack-infra/nodepool master: Add Fedora 29 testing https://review.openstack.org/618671 | 05:23 |
| *** yamamoto has joined #openstack-infra | 05:24 | |
| *** chandankumar has joined #openstack-infra | 05:36 | |
| *** chandankumar is now known as chkumar|ruck | 05:37 | |
| *** yamamoto has quit IRC | 05:47 | |
| *** yamamoto has joined #openstack-infra | 05:51 | |
| *** d0ugal has quit IRC | 06:05 | |
| *** yamamoto has quit IRC | 06:07 | |
| *** annp has joined #openstack-infra | 06:17 | |
| openstackgerrit | Tobias Henkel proposed openstack-infra/zuul master: DNM: Test zuul gear keepalive https://review.openstack.org/619150 | 06:22 |
| *** noama has joined #openstack-infra | 06:22 | |
| *** apetrich has joined #openstack-infra | 06:40 | |
| *** yamamoto has joined #openstack-infra | 06:41 | |
| *** hjensas has joined #openstack-infra | 06:48 | |
| *** yboaron_ has joined #openstack-infra | 06:49 | |
| *** ifat_afek has joined #openstack-infra | 06:53 | |
| openstackgerrit | Tristan Cacqueray proposed openstack-infra/zuul master: executor: add support for generic build resource https://review.openstack.org/570668 | 07:05 |
| openstackgerrit | Ian Wienand proposed openstack-infra/glean master: [wip] NetworkManager support https://review.openstack.org/618964 | 07:06 |
| *** rfolco has quit IRC | 07:06 | |
| openstackgerrit | Tristan Cacqueray proposed openstack-infra/nodepool master: Implement zookeeper-auth https://review.openstack.org/619155 | 07:13 |
| openstackgerrit | Tristan Cacqueray proposed openstack-infra/zuul master: Implement zookeeper-auth https://review.openstack.org/619156 | 07:13 |
| *** ahosam has joined #openstack-infra | 07:14 | |
| *** rcernin has quit IRC | 07:26 | |
| *** quiquell|off is now known as quiquell | 07:28 | |
| openstackgerrit | Merged openstack-infra/nodepool master: Remove nodepool-k8s-functional and install-nodepool roles https://review.openstack.org/616409 | 07:43 |
| *** ykarel is now known as ykarel|lunch | 07:45 | |
| *** quiquell is now known as quiquell|brb | 07:53 | |
| *** bhavikdbavishi has quit IRC | 07:53 | |
| *** ccamacho has joined #openstack-infra | 07:59 | |
| *** jtomasek has joined #openstack-infra | 07:59 | |
| openstackgerrit | zhufl proposed openstack/diskimage-builder master: Add missing ws separator between words https://review.openstack.org/619169 | 08:02 |
| *** ginopc has joined #openstack-infra | 08:03 | |
| *** sean-k-mooney has quit IRC | 08:06 | |
| *** yboaron_ has quit IRC | 08:10 | |
| *** quiquell|brb is now known as quiquell | 08:15 | |
| *** rpittau has joined #openstack-infra | 08:19 | |
| *** yamamoto has quit IRC | 08:24 | |
| *** bhavikdbavishi has joined #openstack-infra | 08:32 | |
| *** shardy has joined #openstack-infra | 08:35 | |
| *** shardy has quit IRC | 08:39 | |
| *** dtantsur|afk is now known as dtantsur | 08:42 | |
| *** bobh has joined #openstack-infra | 08:43 | |
| *** eumel8 has joined #openstack-infra | 08:45 | |
| *** yamamoto has joined #openstack-infra | 08:47 | |
| *** ralonsoh has joined #openstack-infra | 08:50 | |
| *** ykarel|lunch is now known as ykarel | 08:52 | |
| *** jpena|off is now known as jpena | 08:52 | |
| openstackgerrit | Ian Wienand proposed openstack-infra/glean master: [wip] NetworkManager support https://review.openstack.org/618964 | 08:56 |
| *** tosky has joined #openstack-infra | 08:57 | |
| openstackgerrit | Tobias Henkel proposed openstack-infra/zuul master: DNM: Test zuul gear keepalive https://review.openstack.org/619150 | 09:02 |
| *** ccamacho has quit IRC | 09:02 | |
| *** e0ne has joined #openstack-infra | 09:03 | |
| *** ccamacho has joined #openstack-infra | 09:04 | |
| *** yboaron_ has joined #openstack-infra | 09:05 | |
| *** markmcd has joined #openstack-infra | 09:06 | |
| *** rfolco has joined #openstack-infra | 09:09 | |
| *** d0ugal has joined #openstack-infra | 09:11 | |
| *** chkumar|ruck has quit IRC | 09:18 | |
| *** chkumar|ruck has joined #openstack-infra | 09:18 | |
| *** rossella_s has joined #openstack-infra | 09:19 | |
| *** bhavikdbavishi has quit IRC | 09:22 | |
| *** d0ugal has quit IRC | 09:22 | |
| *** bhavikdbavishi1 has joined #openstack-infra | 09:22 | |
| *** roman_g has joined #openstack-infra | 09:23 | |
| *** bhavikdbavishi1 is now known as bhavikdbavishi | 09:24 | |
| openstackgerrit | Tristan Cacqueray proposed openstack-infra/zuul master: Implement zookeeper-auth https://review.openstack.org/619156 | 09:25 |
| *** jaosorior has quit IRC | 09:26 | |
| *** d0ugal has joined #openstack-infra | 09:26 | |
| *** derekh has joined #openstack-infra | 09:30 | |
| *** yboaron_ has quit IRC | 09:35 | |
| *** ahosam has quit IRC | 09:36 | |
| *** yboaron_ has joined #openstack-infra | 09:36 | |
| *** ahosam has joined #openstack-infra | 09:36 | |
| *** ahosam has quit IRC | 09:40 | |
| *** pcaruana has joined #openstack-infra | 09:48 | |
| *** kjackal has joined #openstack-infra | 09:49 | |
| *** shardy has joined #openstack-infra | 09:51 | |
| openstackgerrit | Kartikeya Jain proposed openstack/diskimage-builder master: Adding support for SLES 15 in element 'sles' https://review.openstack.org/619186 | 09:53 |
| *** yamamoto has quit IRC | 09:55 | |
| *** bobh has quit IRC | 10:01 | |
| *** sshnaidm|afk is now known as sshnaidm | 10:02 | |
| *** jaosorior has joined #openstack-infra | 10:03 | |
| *** ahosam has joined #openstack-infra | 10:03 | |
| openstackgerrit | Rafael Folco proposed openstack-infra/elastic-recheck master: Categorize error mounting image volumes due to libpod bug[1] https://review.openstack.org/619059 | 10:06 |
| *** Emine has joined #openstack-infra | 10:10 | |
| *** yboaron_ has quit IRC | 10:15 | |
| *** yboaron_ has joined #openstack-infra | 10:16 | |
| *** slivik has joined #openstack-infra | 10:17 | |
| *** yamamoto has joined #openstack-infra | 10:22 | |
| openstackgerrit | Rafael Folco proposed openstack-infra/elastic-recheck master: Categorize internal libpod error https://review.openstack.org/619197 | 10:24 |
| *** yamamoto has quit IRC | 10:28 | |
| openstackgerrit | Rafael Folco proposed openstack-infra/elastic-recheck master: Categorize internal libpod error https://review.openstack.org/619197 | 10:29 |
| *** yamamoto has joined #openstack-infra | 10:30 | |
| *** Emine has quit IRC | 10:31 | |
| *** priteau has joined #openstack-infra | 10:37 | |
| *** ahosam has quit IRC | 10:43 | |
| openstackgerrit | Chandan Kumar proposed openstack-infra/elastic-recheck master: Add query for ssh timeout in tempest scenario for bug 1802971 https://review.openstack.org/617579 | 10:48 |
| openstack | bug 1802971 in tripleo "tempest volume_boot_pattern and basic_ops running concurrently causing timeouts" [Critical,Triaged] https://launchpad.net/bugs/1802971 | 10:48 |
| *** ahosam has joined #openstack-infra | 10:51 | |
| *** shardy has quit IRC | 10:52 | |
| *** shardy has joined #openstack-infra | 10:53 | |
| openstackgerrit | Chandan Kumar proposed openstack-infra/elastic-recheck master: Add query for ssh timeout in tempest scenario for bug 1802971 https://review.openstack.org/617579 | 10:53 |
| openstack | bug 1802971 in tripleo "tempest volume_boot_pattern and basic_ops running concurrently causing timeouts" [Critical,Triaged] https://launchpad.net/bugs/1802971 | 10:53 |
| *** Emine has joined #openstack-infra | 10:54 | |
| *** ahosam has quit IRC | 10:55 | |
| *** ifat_afek has quit IRC | 10:58 | |
| *** kjackal has quit IRC | 10:59 | |
| *** kjackal_v2 has joined #openstack-infra | 10:59 | |
| *** sambetts_ has joined #openstack-infra | 11:19 | |
| *** sambetts_ is now known as sambetts|afk | 11:21 | |
| *** kjackal_v2 has quit IRC | 11:25 | |
| *** kjackal has joined #openstack-infra | 11:25 | |
| *** udesale has quit IRC | 11:28 | |
| *** kjackal has quit IRC | 11:30 | |
| *** kjackal_v2 has joined #openstack-infra | 11:30 | |
| *** aperevalov has quit IRC | 11:31 | |
| *** aojea has joined #openstack-infra | 11:39 | |
| *** xek has quit IRC | 11:39 | |
| *** BOKALDO_ has joined #openstack-infra | 11:40 | |
| *** yamamoto has quit IRC | 11:45 | |
| openstackgerrit | Jean-Philippe Evrard proposed openstack-infra/infra-manual master: Replace mailing list https://review.openstack.org/619216 | 11:46 |
| AJaeger | fungi, why do I get a thread that is directed to "openstack@lists.openstack.org" via openstack-discuss? See http://lists.openstack.org/pipermail/openstack-discuss/2018-November/000102.html and the rest in the thread. Was that a BCC? OR some miscommunication? | 11:46 |
| AJaeger | s/miscommunication/misconfiguration/ | 11:46 |
| *** sean-k-mooney has joined #openstack-infra | 11:47 | |
| ssbarnea|bkp2 | any logstash guru around? i need help on https://ask.openstack.org/en/question/117469/how-to-use-and-in-logstash-queries/ | 11:48 |
| *** mugsie has joined #openstack-infra | 11:49 | |
| *** njohnston has quit IRC | 11:49 | |
| *** njohnston_ has joined #openstack-infra | 11:49 | |
| *** janki has quit IRC | 11:50 | |
| *** janki has joined #openstack-infra | 11:56 | |
| frickler | AJaeger: iiuc mails to the old lists are getting automatically forwarded to -discuss for the transition now | 11:56 |
| frickler | AJaeger: the idea being that you can unsubscribe from the old lists now without losing anything | 11:56 |
| AJaeger | frickler: taht went to openstack - old list is openstack-dev! | 11:57 |
| frickler | AJaeger: there are three old lists that all move to discuss, openstack, openstack-dev and openstack-ops | 11:58 |
| * AJaeger reads fungi's announcement again now... | 11:58 | |
| frickler | AJaeger: oh, four in fact, openstack-sigs too | 11:59 |
| AJaeger | Oh indeed - frickler, you're right. I missed openstack, only had the others on the radar. | 11:59 |
| AJaeger | thanks, frickler ! | 11:59 |
| AJaeger | fungi, ignore me - frickler explained everything, all looks fine ;) | 11:59 |
| *** psachin has quit IRC | 12:01 | |
| *** evrardjp has joined #openstack-infra | 12:01 | |
| *** ifat_afek has joined #openstack-infra | 12:01 | |
| *** bhavikdbavishi1 has joined #openstack-infra | 12:02 | |
| *** bhavikdbavishi has quit IRC | 12:03 | |
| *** bhavikdbavishi1 is now known as bhavikdbavishi | 12:03 | |
| *** xek has joined #openstack-infra | 12:04 | |
| *** yamamoto has joined #openstack-infra | 12:07 | |
| *** janki has quit IRC | 12:12 | |
| *** sshnaidm is now known as sshnaidm|afk | 12:15 | |
| *** bhavikdbavishi has quit IRC | 12:25 | |
| *** yamamoto has quit IRC | 12:36 | |
| *** jpena is now known as jpena|lunch | 12:39 | |
| aspiers | hrm, is there some inconsistency regarding whether IRC meetings are prefixed with "openstack-"? http://eavesdrop.openstack.org/meetings/ | 12:45 |
| *** e0ne has quit IRC | 12:45 | |
| aspiers | e.g. there is nova and openstack-nova | 12:45 |
| *** aojea has quit IRC | 12:45 | |
| openstackgerrit | Jose Luis Franco proposed openstack-infra/project-config master: ansible-role-* projects switch to non-pypi jobs https://review.openstack.org/617974 | 12:46 |
| *** dtantsur is now known as dtantsur|brb | 12:47 | |
| *** cdent has joined #openstack-infra | 12:50 | |
| evrardjp | hey, do we have a best practice for .gitignore files? I have seen like many things in different projects, and I remember I got a series of patches to change details in OSA, I just wondering if at some point we decided on a convention | 12:50 |
| *** ramishra has joined #openstack-infra | 12:51 | |
| cdent | people have remarkably strong opinions on .gitignore files. which surprises me | 12:51 |
| evrardjp | I don't really care? | 12:52 |
| cdent | me too | 12:52 |
| evrardjp | ok will take yours then :p | 12:52 |
| evrardjp | you won the argument! | 12:52 |
| cdent | go me! | 12:52 |
| aspiers | if only there was a git command for figuring out how .gitignore rules apply to individual files ... ;-) | 12:54 |
| evrardjp | aspiers: show off | 12:54 |
| aspiers | haha | 12:54 |
| aspiers | SCNR | 12:54 |
| evrardjp | I had to urbandictionnary SCNR. Totally worth it. | 12:55 |
| aspiers | :) | 12:55 |
| evrardjp | aspiers: on that link is what people use to start the meeting | 12:58 |
| aspiers | which link? | 12:58 |
| evrardjp | what should be used in what's on eavesdrop | 12:58 |
| evrardjp | http://eavesdrop.openstack.org/meetings/ | 12:58 |
| aspiers | oh yeah, I know | 12:58 |
| evrardjp | that's why you have x things for OSA, because I cannot remember | 12:58 |
| aspiers | I started the "self-healing" meeting this morning, which is why I asked | 12:58 |
| aspiers | but openstack-self-healing already exists because of how I submitted the IRC meeting to Gerrit | 12:59 |
| evrardjp | ok | 12:59 |
| aspiers | haha | 12:59 |
| *** jcoufal has joined #openstack-infra | 12:59 | |
| aspiers | maybe there should be some validation done by the bot | 12:59 |
| evrardjp | seems #startmeeting self_healing | 12:59 |
| evrardjp | should be enough | 12:59 |
| aspiers | errr, careful! | 12:59 |
| evrardjp | yeah | 12:59 |
| aspiers | you almost started another meeting :) | 12:59 |
| aspiers | actually I used hyphen but it converted to _ | 12:59 |
| evrardjp | is that new? | 13:00 |
| aspiers | dunno | 13:00 |
| cdent | space, dash and underscore all mean the same thing to startmeeting | 13:01 |
| cdent | I just posted a making the most of zuul question to the ML: http://lists.openstack.org/pipermail/openstack-discuss/2018-November/000105.html | 13:01 |
| evrardjp | cdent: that's interesting, why do we have openstack-ansible and openstack_ansible in said meetings/ folder then? :) | 13:02 |
| cdent | well if they ended up in the same folder they must mean the same thing ? | 13:02 |
| *** e0ne has joined #openstack-infra | 13:02 | |
| cdent | or do you mean different folders within there? | 13:03 |
| *** e0ne has quit IRC | 13:03 | |
| evrardjp | different folders withing meetings/ | 13:03 |
| evrardjp | at least different folders within meetings exist | 13:03 |
| *** e0ne has joined #openstack-infra | 13:03 | |
| evrardjp | but hey I prefer your question | 13:03 |
| cdent | Maybe it changed some time after 2016 | 13:04 |
| evrardjp | I should maybe ask to move all the files in one location, then use said location in the bot | 13:04 |
| evrardjp | cdent: so your question is basically what are the base jobs definitions for your use case, right? | 13:05 |
| cdent | evrardjp: pretty much yeah. or anything else to make it little | 13:06 |
| openstackgerrit | Jakub Sliva proposed openstack-infra/project-config master: Create repository for ceilometer-dashboard plugin https://review.openstack.org/619235 | 13:06 |
| *** yamamoto has joined #openstack-infra | 13:06 | |
| evrardjp | cdent: well OSA jobs for example only inherit from base, then we have a top level one to checkout our basillion of repos and do minimal stuff (not running for documentation only changes for example) | 13:07 |
| evrardjp | then we inherit from that pseudo top level | 13:07 |
| evrardjp | so technically you can do whatever you want, but that's not the answer you were looking for, right? :p | 13:07 |
| cdent | well. I'm hoping for slightly more guidance on the "make sure the logs I want to capture are" part | 13:08 |
| evrardjp | maybe that part could help: https://github.com/openstack/openstack-ansible/blob/master/zuul.d/playbooks/post.yml#L30-L64 ? | 13:10 |
| evrardjp | basically https://github.com/openstack/openstack-ansible/blob/master/zuul.d/playbooks/post.yml#L57-L58 | 13:10 |
| *** rh-jelabarre has joined #openstack-infra | 13:10 | |
| evrardjp | in fact... I agree with you. I will check what's in base to see if there is a convention already | 13:11 |
| *** zul has joined #openstack-infra | 13:13 | |
| evrardjp | maybe a good question for #zuul | 13:16 |
| *** yamamoto has quit IRC | 13:18 | |
| * cdent nods | 13:18 | |
| *** kjackal_v2 has quit IRC | 13:24 | |
| *** kjackal has joined #openstack-infra | 13:25 | |
| *** ifat_afek has quit IRC | 13:30 | |
| *** rlandy has joined #openstack-infra | 13:37 | |
| fungi | evrardjp: cdent: probably closest thing we have to a project-wide consensus on .gitignore files is the result of the discussion starting at http://lists.openstack.org/pipermail/openstack-dev/2018-August/133393.html and really boils down to "don't add cruft from personal tools to .gitignore" because otherwise we're going to have $random_newcomer blasting 1k repos with a change to add a .gitignore line | 13:39 |
| fungi | for some random editor or ide only they're using. easier to just say that class of things is what you should configure your dev environment to ignore rather than trying to configure it in every project to which you might contribute | 13:39 |
| logan- | cdent: I think if you inherit base its on your job to ship the logs back to the executor from the test node (ie the OSA example evrardjp provided) | 13:39 |
| cdent | logan-: is there something between base and devstack that has the log magic built in? | 13:40 |
| fungi | evrardjp: cdent: aspiers: problem with relocating meeting logs is that the htmlified copies are pre-baked with urls embedding the original paths so would all need some careful stream editing to update for a new location | 13:42 |
| *** jpena|lunch is now known as jpena | 13:42 | |
| evrardjp | fungi: yeah I remember that conversation, thanks for the link :D | 13:42 |
| logan- | cdent: guessing its this https://github.com/openstack-dev/devstack/blob/master/roles/fetch-devstack-log-dir/tasks/main.yaml | 13:43 |
| evrardjp | fungi: for the meeting folder, it's fine like this then :p | 13:43 |
| *** boden has joined #openstack-infra | 13:43 | |
| fungi | even the text copies are going to make at least one mention of the old url in fact | 13:44 |
| evrardjp | logan-: hasn't the base jobs in openstack by default an upload? I'd say so, but I haven't managed to find exactly the same role name, and moved to somehting else :p | 13:44 |
| evrardjp | fungi: yeah, too much effort for not many results | 13:44 |
| logan- | evrardjp: i don't think so, because not every job runs on a test node requiring log uploads | 13:45 |
| logan- | like jobs running on the executor | 13:45 |
| logan- | you could put together a base-with-logs job and inherit that instead | 13:46 |
| evrardjp | yeah | 13:47 |
| evrardjp | because I have the impression everyone is doing that sync task, and because of the "onion" layer of pre/post jobs, it would be simpler to be consistent everywhere | 13:48 |
| *** mriedem has joined #openstack-infra | 13:49 | |
| logan- | hmm http://zuul.openstack.org/job/base says base does copy logs though | 13:49 |
| fungi | evrardjp: what the infra team did when they changed their meeting name was just add a url to the earlier meeting logs location in a note on the persistent agenda: https://wiki.openstack.org/wiki/Meetings/InfraTeamMeeting#Previous_meetings | 13:51 |
| fungi | though we had even earlier irc meetings still, there was no meetbot for them at the time | 13:52 |
| fungi | we didn't start logging the infra meeting with meetbot until august 2012 | 13:52 |
| evrardjp | fungi: that's what I did in OSA too | 13:53 |
| openstackgerrit | Markus Hosch proposed openstack-infra/zuul master: Add testcase to show inconsistency during merging https://review.openstack.org/619245 | 13:53 |
| evrardjp | aspiers: ^ | 13:53 |
| *** quiquell is now known as quiquell|brb | 13:53 | |
| logan- | ah the base job's log copying only means it copies logs from the executor to logs.openstack.org https://github.com/openstack-infra/project-config/blob/master/playbooks/base/post-logs.yaml -- the logs have to be shipped to the executor by child jobs currently | 13:54 |
| *** janki has joined #openstack-infra | 13:59 | |
| *** trown|outtypewww is now known as trown | 14:02 | |
| cdent | thanks for digging into that logan- , very helpful | 14:02 |
| cdent | I've got a wip going with a more minimal setup now. I expect it fail greatly, but we'll see | 14:02 |
| openstackgerrit | Tobias Henkel proposed openstack-infra/zuul master: Display ref instead of NA for time triggered items https://review.openstack.org/619252 | 14:03 |
| openstackgerrit | Tobias Henkel proposed openstack-infra/zuul master: Add testcase to show inconsistency during merging https://review.openstack.org/619245 | 14:06 |
| *** ykarel is now known as ykarel|away | 14:06 | |
| *** dtantsur|brb is now known as dtantsur | 14:07 | |
| *** pbourke has quit IRC | 14:07 | |
| *** fuentess has joined #openstack-infra | 14:07 | |
| *** pbourke has joined #openstack-infra | 14:09 | |
| *** ifat_afek has joined #openstack-infra | 14:09 | |
| *** kgiusti has joined #openstack-infra | 14:11 | |
| *** chkumar|ruck has quit IRC | 14:12 | |
| *** Douhet has quit IRC | 14:13 | |
| *** quiquell|brb is now known as quiquell | 14:19 | |
| *** bobh has joined #openstack-infra | 14:21 | |
| *** sthussey has joined #openstack-infra | 14:22 | |
| openstackgerrit | Lars Kellogg-Stedman proposed openstack-infra/glean master: changes for alpine compatibility https://review.openstack.org/619258 | 14:23 |
| fungi | ssbarnea|bkp2: can you point to an actual job log that you expected to match your logstash query? | 14:23 |
| openstackgerrit | Quique Llorente proposed openstack-infra/zuul master: Prevent exception at executor https://review.openstack.org/617984 | 14:24 |
| *** jaosorior has quit IRC | 14:25 | |
| *** Douhet has joined #openstack-infra | 14:26 | |
| *** njohnston_ has quit IRC | 14:29 | |
| *** njohnston has joined #openstack-infra | 14:29 | |
| *** quiquell is now known as quiquell|brb | 14:32 | |
| *** agopi|brb has quit IRC | 14:37 | |
| *** udesale has joined #openstack-infra | 14:39 | |
| *** quiquell|brb is now known as quiquell | 14:48 | |
| *** ykarel|away has quit IRC | 14:55 | |
| *** Douhet has quit IRC | 14:56 | |
| *** Douhet has joined #openstack-infra | 15:04 | |
| fungi | #status log rolled back garbled summit feedback pad using: wget -qO- 'http://localhost:9001/api/1.2.11/restoreRevision?apikey='$(cat /opt/etherpad-lite/etherpad-lite/APIKEY.txt)'&padID=BER-Feedback-Session&rev=5564' | 15:05 |
| openstackstatus | fungi: finished logging | 15:06 |
| *** agopi has joined #openstack-infra | 15:08 | |
| ssbarnea|bkp2 | fungi: in fact I am unable to find "newer" builds with the same error. See: https://review.openstack.org/#/c/608968/ | 15:09 |
| ssbarnea|bkp2 | this CR was created more than a month ago, and meanwhile the error no longer appears in logs. Even the specific occurences from the LP bug were recycled. | 15:10 |
| *** ykarel|away has joined #openstack-infra | 15:10 | |
| *** ykarel|away is now known as ykarel | 15:10 | |
| ssbarnea|bkp2 | it seems that time needed to review a change elastic-recheck may make it impossible to test it because log rotation, on both logstash and the fileserver. | 15:11 |
| ssbarnea|bkp2 | what should we do in this case? close both bug and CR with - unable to reproduce? | 15:11 |
| *** chandankumar has joined #openstack-infra | 15:11 | |
| ssbarnea|bkp2 | i think that we do need retention of >60 days on logs but probably we don't have the disk for it. | 15:12 |
| ssbarnea|bkp2 | that specific issue is not of concern right bow, I am more interested about how we should deal with these. | 15:14 |
| *** Emine has quit IRC | 15:17 | |
| *** Emine has joined #openstack-infra | 15:18 | |
| corvus | logan-, cdent: i think pabelanger picked up on a WIP effort started by mordred to update the base job to sync from ~/logs on all remote hosts automatically, so that jobs wouldn't need to do sync tasks themselves; i don't have links to emails/changes for that handy; pabelanger knows the latest status, but i could probably dig up more info if needed. | 15:19 |
| cdent | ✔ | 15:19 |
| *** chandankumar has quit IRC | 15:22 | |
| *** chandankumar has joined #openstack-infra | 15:24 | |
| *** jaosorior has joined #openstack-infra | 15:24 | |
| fungi | ssbarnea|bkp2: we probably could add extra volumes, but it's already something like 6tb of storage sharded across the cluster to get us ~10 days retention (i think) | 15:28 |
| fungi | performance might dictate that we add more cluster members if we're adding more data, and that's where things get tough because each cluster member is a 60gb ram flavor | 15:28 |
| fungi | that logstash backend is by far the most expensive (to our donors) service we operate besides nodepool | 15:29 |
| fungi | especially if you take the 20(?) logstash workers into account | 15:30 |
| ssbarnea|bkp2 | fungi: at least this explain why I had the impression that the query didn't work: i was looking at a log file which was like 10-12 old and logstash was failing to return anything when window size was 14 days. | 15:32 |
| ssbarnea|bkp2 | is not clear anywhere when did the cut-off happen. | 15:32 |
| ssbarnea|bkp2 | anyway at this moment I am more concerned about retension on logs.openstack.org than on logstash. | 15:33 |
| ssbarnea|bkp2 | for logstash I have low (retention) expectation. | 15:33 |
| corvus | we should be able to increase raw log storage retention when we switch to swift | 15:34 |
| ssbarnea|bkp2 | this also means that we are ok to close CRs made on elastic-search which do not identify anything, I can even try to document that. | 15:34 |
| openstackgerrit | Tobias Henkel proposed openstack-infra/zuul master: Fix manual dequeue of github items https://review.openstack.org/619272 | 15:34 |
| corvus | (which we could do at any point, but if we delay it a bit more, we might be able to have a better experience) | 15:34 |
| openstackgerrit | Tobias Henkel proposed openstack-infra/zuul master: Fix manual dequeue of github items https://review.openstack.org/619272 | 15:35 |
| ssbarnea|bkp2 | corvus: i have big personal interest in improving how we deal with log files, keep this in mind as I am more than willing to help there. | 15:35 |
| corvus | ssbarnea|bkp2: will do, thanks | 15:37 |
| *** chandankumar has quit IRC | 15:39 | |
| *** quiquell is now known as quiquell|off | 15:39 | |
| openstackgerrit | Tobias Henkel proposed openstack-infra/zuul master: Display ref instead of NA for time triggered items https://review.openstack.org/619252 | 15:43 |
| fungi | yeah, we certainly could benefit from much longer raw log storage retention, but that's not really searchable in a sane way. increasing the retention for logstash.o.o is probably not something we're going to find a way to do unless we can reduce the amount of data we're actually logging | 15:43 |
| *** ramishra has quit IRC | 15:45 | |
| ssbarnea|bkp2 | fungi: regarding log browsing please check https://docs.google.com/document/d/1TM2_USBVq55Xqk7WHbnsd-mrQthDTzl0rMq6-4MwbdY/edit# -- i started to write on it some time ago in order not to forget my idea. | 15:48 |
| clarkb | fungi: ya its 6TB total storage with about 5TB useable (so that we can handle the loss of a single node) and we tend to use a good chunk of it | 15:50 |
| clarkb | the other constraint is you need enough memory to load queries and we often run up against that with available memory on the nodes | 15:51 |
| *** ykarel is now known as ykarel|away | 15:52 | |
| clarkb | we already filter all debug logs out and still end up with a billion records a day | 15:52 |
| *** noama has quit IRC | 15:52 | |
| clarkb | fungi: how did you determine the revision to rollback to on the feedback etherpad? did you get the contents for each rev going back in time until the bad data went away? | 15:54 |
| clarkb | wonder if there is a smarter method | 15:54 |
| *** slivik has quit IRC | 15:57 | |
| *** udesale has quit IRC | 15:58 | |
| fungi | clarkb: someone else did that and i simply trusted them (after checking that revision looked sane) | 15:59 |
| clarkb | ah | 15:59 |
| fungi | with access to the api you can query pad contents by revision and also revision diffs i think, so that could be a future option in such cases | 15:59 |
| frickler | clarkb: fungi: you can use the timeslider (clock button on the top right) https://etherpad.openstack.org/p/BER-Feedback-Session/timeslider#5802 | 16:11 |
| clarkb | frickler: ah that shows the rev number got it | 16:11 |
| fungi | yeah, someone had provided me the direct timeslider url and i just fed the number from the url into the rev parameter | 16:14 |
| *** diablo_rojo has joined #openstack-infra | 16:16 | |
| *** agopi is now known as agopi|food | 16:17 | |
| *** jamesmcarthur has quit IRC | 16:17 | |
| *** jamesmcarthur_ has joined #openstack-infra | 16:18 | |
| *** bobh has quit IRC | 16:19 | |
| *** ykarel|away has quit IRC | 16:27 | |
| *** janki has quit IRC | 16:36 | |
| *** bhavikdbavishi has joined #openstack-infra | 16:40 | |
| *** yboaron_ has quit IRC | 16:43 | |
| openstackgerrit | Merged openstack-infra/system-config master: fedora mirror : drop -p from rsync https://review.openstack.org/618504 | 16:47 |
| *** pcaruana has quit IRC | 16:50 | |
| *** efried_back_mon has quit IRC | 16:51 | |
| *** agopi|food is now known as agopi | 16:51 | |
| *** xek_ has joined #openstack-infra | 16:54 | |
| clarkb | corvus: jamesmcarthur_ sent email with the info the registrar wants (and seems hopeful that providing this may make them happy) | 16:55 |
| jamesmcarthur_ | we shall see clarkb: | 16:57 |
| *** xek has quit IRC | 16:57 | |
| corvus | clarkb, jamesmcarthur_: great, thanks. i'll get that together in a jiffy | 16:57 |
| clarkb | corvus: thank you! | 16:57 |
| jamesmcarthur_ | corvus: danke | 16:58 |
| clarkb | ianw: it looks like you have a bunch of +2's on the bandersnatch retirement change. Let me know if I can help with the pypi afs volume cleanup | 16:58 |
| *** shardy has quit IRC | 17:01 | |
| *** e0ne has quit IRC | 17:02 | |
| *** Emine has quit IRC | 17:05 | |
| openstackgerrit | Clark Boylan proposed openstack-infra/glean master: [wip] NetworkManager support https://review.openstack.org/618964 | 17:06 |
| *** shardy has joined #openstack-infra | 17:07 | |
| clarkb | if anyone else is interested in future rhel/fedora glean support ^ and its parents are the changes for you | 17:08 |
| *** ifat_afek has quit IRC | 17:13 | |
| openstackgerrit | Merged openstack-infra/project-config master: Fix ansible-role-cloud-launcher docs https://review.openstack.org/616029 | 17:13 |
| *** sambetts|afk has quit IRC | 17:14 | |
| *** betherly has joined #openstack-infra | 17:15 | |
| *** yamamoto has joined #openstack-infra | 17:16 | |
| *** bhavikdbavishi has quit IRC | 17:16 | |
| *** xek_ is now known as xek | 17:18 | |
| *** bhavikdbavishi has joined #openstack-infra | 17:19 | |
| *** yamamoto has quit IRC | 17:20 | |
| *** betherly has quit IRC | 17:21 | |
| openstackgerrit | Clark Boylan proposed openstack/diskimage-builder master: package-installs: provide for skip from env var https://review.openstack.org/619119 | 17:21 |
| openstackgerrit | Clark Boylan proposed openstack/diskimage-builder master: simple-init: allow for NetworkManager support https://review.openstack.org/619120 | 17:21 |
| *** bhavikdbavishi has quit IRC | 17:24 | |
| rfolco | clarkb, quick question, is there any mechanism to indicate that a job is deprecated ? I was thinking in a custom message on emit-job-header or something... the issue is that I have deprecated jobs being referred in multiple projects and multiple branches. So the list of changes to remove is huge before actually removing the job definition. Any thoughts ? | 17:24 |
| clarkb | ianw: left a couple of thoughts on those dib changes (the new patchsets only fixed pep8) | 17:24 |
| clarkb | rfolco: emitting a warning message via ansible is probably the thing you can do today | 17:25 |
| clarkb | rfolco: then you will be able to query on logstash for that message to dig up where it is used | 17:25 |
| openstackgerrit | sebastian marcet proposed openstack-infra/openstackid-resources master: Added new endpoint to get my presentations per role and per summit: https://review.openstack.org/619321 | 17:25 |
| openstackgerrit | Merged openstack-infra/openstackid-resources master: Added new endpoint to get my presentations per role and per summit: https://review.openstack.org/619321 | 17:27 |
| rfolco | clarkb, cool, I was just thinking about emit-job-header, but that one I would need a good justification I think | 17:27 |
| clarkb | rfolco: also for the stable branhces you may not want to change the jobs there. At least typically openstack lets stable jobs sit as they were and then once the stable branch goes away or the testing itself no longer works the test is removed entirely | 17:28 |
| rfolco | clarkb, coz tripleo-ci (where my job stands) is not branched... many projects are branched and have in their zuul config these old branch jobs (ocata)... | 17:30 |
| rfolco | unless I am missing anything, zuul syntax check looks for all branches for untrusted-projects | 17:31 |
| clarkb | rfolco: it does, but if the jobs are already there they should be correct syntax right? Mostly just saying that changing jobs on stable branches is often more effort than it is worth | 17:32 |
| AJaeger | rfolco: also each job should have a description - If I deprecate jobs, I'll add that information to the description. | 17:32 |
| *** ginopc has quit IRC | 17:35 | |
| rfolco | I am blocked in removing the job definition from my zuul config because many of these branched projects refer to that old job name and don't pass ci for old branches like ocata | 17:35 |
| rfolco | I wish zuul could follow branches keyword for syntax check as well, not for where running the job | 17:36 |
| fungi | i'm not sure what you mean with that wish | 17:37 |
| corvus | jamesmcarthur_, clarkb, fungi: ohhhhh i think i get it. i think the registrar wants ds records and not the dnskey record. (gandi asks for dnskey and generates ds records for you). shouldn't be a problem, it's just a bit more work. | 17:39 |
| rfolco | fungi, like telling zuul which branches you want to ignore for syntax check | 17:39 |
| corvus | jamesmcarthur_, clarkb, fungi: the dnskey record is certainly "more characters" than the ds record :) | 17:39 |
| fungi | corvus: i concur | 17:39 |
| fungi | rfolco: oh, zuul doesn't ignore any branches. it loads its configuration from all branches of a project because it might at some point be asked to test a commit for one of them | 17:40 |
| clarkb | rfolco: the issue is you would braek those branches because the jobs are on an unbranched repo | 17:40 |
| clarkb | rfolco: it is an error to remove the jobs from the unbranched repo while still using the jobs in the branched repos | 17:41 |
| *** jaosorior has quit IRC | 17:41 | |
| clarkb | rfolco: if you converted to branched config for the job defs you could hide that debt away in older branches | 17:41 |
| clarkb | but potentially add more work when you sync between them | 17:41 |
| clarkb | corvus: aha | 17:41 |
| *** jamesmcarthur_ has quit IRC | 17:41 | |
| openstackgerrit | Kaio Kassiano Moura Oliveira proposed openstack-infra/project-config master: Retire openstack-ansible-os_monasca-ui https://review.openstack.org/617212 | 17:42 |
| tosky | not sure if I should ask here or on -tc or somewhere else: now that we are moving to openstack-discuss, can I update the author-email field for all sahara repositories to point to it instead of openstack-dev, | 17:42 |
| *** jamesmcarthur has joined #openstack-infra | 17:42 | |
| tosky | or should I wait for a set of coordinated patches? | 17:42 |
| tosky | (update in setup.cfg) | 17:43 |
| fungi | tosky: i don't expect any coordinated patches for that. feel free to make appropriate updates to your projects at any time | 17:43 |
| tosky | thanks | 17:43 |
| *** florianf has quit IRC | 17:45 | |
| rfolco | thx clarkb AJaeger fungi, will try to find a solution. For now I am moving these jobs to a separate file and parenting them to run a "deprecation warning" playbook. | 17:45 |
| corvus | fungi: should we send only 1 ds record with a sha256 digest (and omit the sha-1 digest record)? | 17:47 |
| corvus | fungi: it appears that's what gandi did for us with zuul-ci | 17:47 |
| corvus | i'm kind of assuming that we've aged out of using sha-1 for that now and it's not necessary. but i don't really know that fer shure. | 17:47 |
| fungi | that seems safe these days. the number of resolvers out there lacking sha-256 support should be few in number | 17:47 |
| corvus | bootstrapping is the best time to find out anyway :) | 17:48 |
| fungi | agreed | 17:48 |
| *** dtantsur is now known as dtantsur|afk | 17:49 | |
| *** derekh has quit IRC | 17:50 | |
| clarkb | it appears the rfc still has sha1 as set to mandatory but I agree gandi only has sha256 for my domain | 17:51 |
| openstackgerrit | James E. Blair proposed openstack-infra/system-config master: docs: add info on generating DS records https://review.openstack.org/619334 | 17:51 |
| corvus | clarkb, fungi: ^ | 17:51 |
| clarkb | I'm on a bit of a review streak. I think I've reivewed all teh things ianw needs for fedora 29 as well as pabelangers zuul executor zone work. At some point I need to write expense report(s) but let me know if there are other broad topics that could use review | 17:52 |
| clarkb | maybe I should double check on the docker work | 17:52 |
| *** jamesmcarthur has quit IRC | 17:54 | |
| *** jamesmcarthur has joined #openstack-infra | 17:54 | |
| corvus | jamesmcarthur, clarkb, fungi: i sent the info | 17:55 |
| *** rpittau has quit IRC | 17:57 | |
| jamesmcarthur | thanks corvus. Just sent and cc'd you. | 17:57 |
| *** boden has quit IRC | 17:59 | |
| clarkb | reviewing the docker stuff it is occuring to me that ipv6 is apparnetly not a thing that people using docker seem to really think about. But also we tend to front our services with apache anyway for ssl termination. I'm thinking it would probably be simplest for us if we terminated ipv6 in the webserver too then we don't have to rebuild a bunch of images to ipv6 too? On top of that do we want to run | 18:02 |
| clarkb | these webservers in a container too? Seems like it makes sense to do that since they are largely stateless (if you ignore connection state) | 18:02 |
| clarkb | Mostly throwing this out there so that people can think about it, I don't think there is a rush to have an answer | 18:02 |
| clarkb | also here hoping the NAT holds up :/ | 18:02 |
| fungi | corvus: lgtm, though i left an alternative in a comment based on how i did it for my domains... in short: dnssec-dsfromkey can take a dnskey rr as input | 18:03 |
| corvus | fungi: ah yep. maybe when we expand that doc out some more to encompass the whole process, we should use that instead. it's a good check. | 18:04 |
| *** jcoufal_ has joined #openstack-infra | 18:06 | |
| *** jamesmcarthur has quit IRC | 18:07 | |
| *** jamesmcarthur has joined #openstack-infra | 18:09 | |
| corvus | cdent: https://review.openstack.org/583346 is the work from pabelanger i was thinking of | 18:09 |
| AJaeger | any config-core to review some changes, please? https://review.openstack.org/616583 https://review.openstack.org/618633 https://review.openstack.org/618722 https://review.openstack.org/#/c/610888/ | 18:09 |
| *** jcoufal has quit IRC | 18:10 | |
| cdent | corvus: great thanks, I'll have a look. In the meantime I managed to get something basic going in a post job and almost have the guts of the job working | 18:10 |
| AJaeger | mnaser: do you want to +2A https://review.openstack.org/617212 - that's the openstack-ansible-os_monasca-ui retirement | 18:10 |
| corvus | cdent: cool. that's definitely the way to go for now since work on the other idea has stalled for the moment. hopefully we can push it over the line in the not-too-distant future. | 18:11 |
| clarkb | cdent: thinking about light weight I need a database and wsgi, I realized some projects (like zuul) do that with their tox invocations | 18:12 |
| clarkb | cdent: there is a tools/test_setup.sh (or similar name I forget what it is exactly) path that we'll run for you before running tox that you can use to setup the database | 18:12 |
| clarkb | pabelanger: is https://review.openstack.org/#/c/618633/1 used somewhere? | 18:12 |
| pabelanger | corvus: cdent: indeed, I haven't had much time to pick up and push on it. We are using some of it in ansible-network today, and it works well. I think we decided to hold off on it until openstack switched to swift for logs | 18:13 |
| pabelanger | clarkb: not yet, hopefull for zuul https://review.openstack.org/618634/ | 18:13 |
| *** jpena is now known as jpena|off | 18:13 | |
| pabelanger | hopefully* | 18:13 |
| corvus | pabelanger: oh... is there a dependency there, or was it just to reduce churn? | 18:13 |
| cdent | clarkb: thus far the difference in time consumed is only 3 minutes (down to 8 from 11) so less than I was expecting but still satisfying to not use extra stuff | 18:13 |
| pabelanger | corvus: reduce churn IIRC | 18:13 |
| *** jamesmcarthur has quit IRC | 18:14 | |
| *** jamesmcarthur has joined #openstack-infra | 18:15 | |
| fungi | clarkb: cdent: yes, there's even a role in zuul-jobs you can include: http://git.zuul-ci.org/cgit/zuul-jobs/tree/roles/test-setup/tasks/main.yaml | 18:15 |
| fungi | according to that it's tools/test-setup.sh | 18:16 |
| *** florianf has joined #openstack-infra | 18:16 | |
| openstackgerrit | Merged openstack-infra/project-config master: Move elastic-recheck zuul jobs definitions inside project repo https://review.openstack.org/616583 | 18:19 |
| *** jamesmcarthur has quit IRC | 18:20 | |
| clarkb | AJaeger: pabelanger double check my comments on https://review.openstack.org/#/c/618633/1 | 18:21 |
| AJaeger | thanks clarkb for reviewing | 18:22 |
| *** jamesmcarthur has joined #openstack-infra | 18:23 | |
| *** mriedem has quit IRC | 18:26 | |
| openstackgerrit | Merged openstack-infra/project-config master: Retire openstack-ansible-os_monasca-ui https://review.openstack.org/617212 | 18:26 |
| pabelanger | clarkb: ah yes, you are right. I'll fix up shortly | 18:27 |
| openstackgerrit | Paul Belanger proposed openstack-infra/project-config master: Create publish-zuul-python-branch-tarball job https://review.openstack.org/618633 | 18:29 |
| pabelanger | clarkb: AJaeger: ^should address the comments | 18:29 |
| openstackgerrit | Tobias Henkel proposed openstack-infra/zuul master: Display ref instead of NA for time triggered items https://review.openstack.org/619252 | 18:30 |
| *** trown is now known as trown|lunch | 18:33 | |
| *** shardy has quit IRC | 18:34 | |
| *** shardy has joined #openstack-infra | 18:36 | |
| *** mriedem has joined #openstack-infra | 18:37 | |
| *** ralonsoh has quit IRC | 18:40 | |
| *** ahosam has joined #openstack-infra | 18:53 | |
| dmsimard | I have a weird issue trying to use the synchronize module with zuul.project.src_dir | 18:55 |
| dmsimard | oh, wait, I think I know | 18:56 |
| *** shardy has quit IRC | 18:56 | |
| openstackgerrit | Merged openstack-infra/project-config master: Create publish-zuul-python-branch-tarball job https://review.openstack.org/618633 | 18:57 |
| prometheanfire | glean review please (for more networkd stuff https://review.openstack.org/610064 https://review.openstack.org/610107 | 18:58 |
| prometheanfire | ianw: clarkb you think glean should use a more generic install.py? | 18:59 |
| *** xek has quit IRC | 19:00 | |
| clarkb | prometheanfire: I need more context than that | 19:00 |
| *** apetrich has quit IRC | 19:00 | |
| prometheanfire | clarkb: k, I commented https://review.openstack.org/618964 but generally, should we have each network backend have a --use-foo option | 19:01 |
| clarkb | prometheanfire: ideally we wouldn't I think this is special beacuse its a transition point for fedora and rhel | 19:02 |
| *** jamesmcarthur has quit IRC | 19:02 | |
| clarkb | prometheanfire: an alternative could be to just assume nm if rhel version >=8 or fedora version >=29 ? | 19:03 |
| prometheanfire | clarkb: should the nm support also support other distros? | 19:03 |
| prometheanfire | or check /etc/os-release | 19:03 |
| clarkb | prometheanfire: the related dib changes seem to imply that it should. But I left a note about maybe not implying that since glean only does it for red hat | 19:03 |
| prometheanfire | ya, guess that should be answered first (rhel/fedora only or not) | 19:04 |
| clarkb | prometheanfire: looking at https://review.openstack.org/#/c/610064/1/glean/cmd.py the files it wouldn't check for are the slave interfaces and vlan logical interfaces? | 19:05 |
| clarkb | probably good enough for now to assume that if the main interface is configured so are the other bits (like vlan) | 19:05 |
| *** xek has joined #openstack-infra | 19:08 | |
| prometheanfire | I think it doesn't care what the name is, iirc, name should be eth4-vlan26 and find the netdev file | 19:09 |
| prometheanfire | see liberty.networkd.network.out for example | 19:10 |
| clarkb | right but that check you are modifying is checking if glean ahs already written things so that it can noop | 19:11 |
| clarkb | so it won't see that eth4-vlan26 is there but it will see eth4 and noop I think | 19:11 |
| clarkb | see comment on https://review.openstack.org/#/c/610107/2 that breaks an existing behavior I think | 19:12 |
| *** jamesmcarthur has joined #openstack-infra | 19:12 | |
| prometheanfire | clarkb: it just noops the single interface though | 19:12 |
| clarkb | prometheanfire: ah so it will write the others | 19:12 |
| clarkb | prometheanfire: we probably do wnat to fix that properly then | 19:12 |
| prometheanfire | ya, continue in the loop | 19:12 |
| prometheanfire | all that I was doing was removing 'gentoo' refrences | 19:12 |
| clarkb | ya its an improvement :) | 19:13 |
| *** jamesmcarthur has quit IRC | 19:13 | |
| prometheanfire | the only 'proper' fix would be to search ALL the netdev and network files for Name=iname | 19:14 |
| *** jamesmcarthur has joined #openstack-infra | 19:14 | |
| prometheanfire | I guess that's possible, but may also give false positives | 19:14 |
| *** apetrich has joined #openstack-infra | 19:15 | |
| clarkb | prometheanfire: thinking about nm as a general contruct it might be a good idea to get a general sense for what the configuration options are on the various currently supported distros | 19:17 |
| clarkb | prometheanfire: then maybe we can see where the venn diagram has overlap and reduce the total number of tools to support | 19:17 |
| prometheanfire | ya | 19:17 |
| clarkb | networkd should cover ubuntu and gentoo and I think suse for a future point (not sure where that line is) then it sounds like red hat is all nm for the future | 19:17 |
| clarkb | then maybe the sysconfig and /etc/network/interfaces stuff can be "legacy" and go away | 19:18 |
| prometheanfire | doesn't cover gentoo-openrc, but does gentoo-systemd | 19:18 |
| prometheanfire | kinda sad rh didn't go with networkd | 19:18 |
| clarkb | or maybe nm will work everwhere as another alternative | 19:18 |
| clarkb | its a bit heavyweight and definitely not my first choice but learning a single thing is probably less effort than paging in 6 tools | 19:19 |
| prometheanfire | ya, nm should work on gentoo (both) | 19:19 |
| prometheanfire | ya, nm isn't my first choice either | 19:19 |
| clarkb | whats weird is the rh nm config seems to be write out sysconfig then set use-nm to true | 19:20 |
| clarkb | and I guess it configures nm from that rather than direct configuration? | 19:20 |
| prometheanfire | hell if I know | 19:20 |
| prometheanfire | overcomplicated imo | 19:20 |
| clarkb | prometheanfire: before I go hunt down food, did you see my -1 about the dns thing? does it make sense? | 19:21 |
| fungi | nm is definitely a working option for ubuntu/debian, fwiw | 19:21 |
| clarkb | fungi: ya as well as suse | 19:21 |
| prometheanfire | ya | 19:21 |
| prometheanfire | clarkb: ya, I'll see about adding a check | 19:21 |
| cdent | clarkb: got it down to 4 minutes. how many cores does the vm used by the base job typically have? | 19:21 |
| clarkb | nm works everywhere because laptops | 19:21 |
| clarkb | cdent: 8 | 19:21 |
| cdent | great, thanks | 19:22 |
| fungi | well, the laptop argument is also valid for anything with hotplugged interfaces | 19:22 |
| fungi | which virtual machines tend to be able to make use of more than physical machines | 19:22 |
| fungi | i mean, not that ifupdown doesn't also support hotplugging network interfaces | 19:23 |
| clarkb | fungi: I guess? the only time I hot plug interfaces other than on my laptop is plugging in network devices so that I can configure them from desktop | 19:23 |
| clarkb | my VMs all come up with the interface they will live with until recycled | 19:23 |
| fungi | longer-lived pets in some virtual environments may benefit from users dynamically adding new networks/ports, but yeah it does seem like a bit of a corner case | 19:24 |
| fungi | and i agree, i'd personally rather _not_ have new network ports magically autoconfigured for me without some manual intervention | 19:24 |
| openstackgerrit | Merged openstack-infra/nodepool master: move 01-nodepool-setup to a later phase https://review.openstack.org/610188 | 19:29 |
| clarkb | ok gonna dig up food now. I'm likely to try and file expense reports after lunch but let me know if there are other stacks of changes that could use review | 19:29 |
| clarkb | ianw: ^ particularly happy to talk about your changes I've reivewed as I've left notes here and there | 19:29 |
| openstackgerrit | Matthew Thode proposed openstack-infra/glean master: write dns info to networkd https://review.openstack.org/610107 | 19:30 |
| prometheanfire | clarkb: ^ | 19:30 |
| *** e0ne has joined #openstack-infra | 19:33 | |
| *** bobh has joined #openstack-infra | 19:35 | |
| *** dklyle has quit IRC | 19:38 | |
| *** jamesmcarthur has quit IRC | 19:40 | |
| *** trown|lunch is now known as trown | 19:44 | |
| *** jamesmcarthur has joined #openstack-infra | 19:44 | |
| *** dklyle has joined #openstack-infra | 19:46 | |
| *** apetrich has quit IRC | 20:01 | |
| openstackgerrit | Doug Hellmann proposed openstack-infra/project-config master: add jobs to publish library from governance repo https://review.openstack.org/619347 | 20:05 |
| *** cdent has quit IRC | 20:18 | |
| ianw | clarkb / prometheanfire: i haven't caught up yet, but i was planning on switching centos and fedora to NM since it seems to work | 20:21 |
| fungi | (relative?) simplicity ftw! | 20:22 |
| ianw | and is the supported thing really | 20:22 |
| *** ahosam has quit IRC | 20:24 | |
| *** jamesmcarthur has quit IRC | 20:25 | |
| clarkb | ianw: I think the concern is whether we want a bunch oc switches for that --use-nm | 20:27 |
| clarkb | instead maybe we just use nm always | 20:27 |
| *** bobh has quit IRC | 20:29 | |
| ianw | this is a possibility; however my main concern was that we put it in the gate and something happens, and with this approach we just flip a flag and rebuild the image, or even create a duplicate image with it turned on for testing | 20:29 |
| *** jamesmcarthur has joined #openstack-infra | 20:29 | |
| ianw | it seems all the testing is working, just because i've added centos+f28+f29 to the same job it's too long and times out | 20:30 |
| clarkb | our testing is pretty good for all plattorms but rax | 20:31 |
| clarkb | maybe rely on that and make it the behavior after testing on rax? | 20:31 |
| ianw | i meant more that the jobs ontop don't like it, rather than not booting | 20:31 |
| ianw | it's in a odd place because it's using the ifcfg-rh plugin, so essentially the guts of the configuration are the same as the network-scripts path | 20:35 |
| ianw | rather than a completely new path | 20:36 |
| clarkb | ya | 20:37 |
| *** bobh has joined #openstack-infra | 20:39 | |
| ianw | does it work the same on debuntu? i think there's also a plugin there to read the existing files? | 20:41 |
| clarkb | plugin ifupdown | 20:42 |
| clarkb | looks like yes | 20:42 |
| clarkb | also ifcfg-suse | 20:42 |
| ianw | right, so that maybe suggests that (if someone cares to work on those) keeping one implementation per distro but with a flag for NM is a saneish thing | 20:45 |
| *** jamesmcarthur has quit IRC | 20:46 | |
| clarkb | ya I dont think it is the worst thing but do want to avoid a use wickd and use networkd and use whatver as a bunch of options | 20:48 |
| clarkb | I'm happy to say nm is default on rh and works everywhere else so is a bit of a special case | 20:48 |
| *** jamesmcarthur has joined #openstack-infra | 20:49 | |
| clarkb | ianw: fwiw I tried to fix pep8 in your stacks then lef comments on things that were more opinion | 20:49 |
| fungi | note that networkmangler has a number of dependencies which we may be avoiding installing so could bloat images and result in additional running daemons | 20:54 |
| clarkb | re nm maybe add a note to glean about its specialness? | 20:55 |
| clarkb | fungi: ya I'm coming around to the current stack where it is selectable | 20:55 |
| ianw | clarkb: yep, still wip, docs etc to add :) | 20:57 |
| fungi | if we don't need it on most platforms, i'm fine with the idea of continuing to run leaner there, but would be curious to try an `apt install network-manager` and finding out what it actually ends up pulling in | 20:57 |
| ianw | fungi: yes, for sure -- however it's a very realistic environment to run it; if we do have problems with the networkmanager daemons, then it doesn't bode well for actually useful code | 20:57 |
| fungi | like, are we already installing and starting dbus everywhere anyway? | 20:57 |
| ianw | fungi: yep, you can't even log in without dbus these days | 20:58 |
| clarkb | dbus is required for systemd things | 20:58 |
| fungi | oh, right, ugh | 20:58 |
| * fungi saves his disdain for more appropriate times and places | 20:58 | |
| * prometheanfire would rather just do networkd for most stuff (and something else for non-systemd) | 21:00 | |
| ianw | prometheanfire: unfortunately it seems others don't agree -- as i mentioned inline rhel8 (and centos8 eventually) won't have networkd and networkmanger is the thing | 21:01 |
| ianw | an eye on centos8 is why i've taken this on, really | 21:01 |
| fungi | unfortunate they weren't willing to choke down their own dogfood on this | 21:02 |
| prometheanfire | fungi: but their dns stuff is so stable and good | 21:03 |
| * fungi wouldn't know | 21:03 | |
| *** BOKALDO_ has quit IRC | 21:05 | |
| ianw | so yeah, i am sensitive that turning glean into cloud-init defies the entire point, but we can't really ignore the networkmanager side, especially that the scripts are split out into a now deprecated package on f29 | 21:05 |
| *** cloudnull has joined #openstack-infra | 21:06 | |
| *** jamesmcarthur has quit IRC | 21:09 | |
| *** jamesmcarthur has joined #openstack-infra | 21:13 | |
| mriedem | hmm, so i just posted my first email to openstack-discuss and am being told it's held for review since i'm not a member, although i confirmed membership earlier today | 21:13 |
| mriedem | is that just a cron or something that isn't showing my updated status? | 21:13 |
| *** betherly has joined #openstack-infra | 21:13 | |
| clarkb | mriedem: I think that would indicate an issue with your subscription. Maybe you subscribed with a different email address? | 21:14 |
| fungi | mriedem: i'll check the logs, but no it should have just worked | 21:14 |
| clarkb | fungi: ^ has been moderating and may be able to see details | 21:14 |
| clarkb | ianw: ya thinking a bit more I think probably the place to start is with --use-nm as you've done it. That gets us future rhel support. Then if anyone is interested in adding debuntu or suse or gentoo nm support that can come later. Separately if we decide that networkd is also a special universal like thing maybe we add a --use-networkd flag and do similar | 21:15 |
| mriedem | hmm, i sent as mriedem.os@gmail.com | 21:15 |
| fungi | mriedem: the moderation message definitely said "Reason: Post by non-member to a members-only list" so i'll check the subscribe log | 21:15 |
| clarkb | ianw: I imagine we'll want to further refactor things since a lot of that code is still messy (though hopefully the debuntu paths are a little bit happier after my refactorings) | 21:16 |
| fungi | also, that error makes me think of the members only jackets from the 80s | 21:16 |
| corvus | mriedem: the moderation queue says the message was from mriedemos@gmail.com (cc fungi) | 21:16 |
| mriedem | i believe members only jackets are back | 21:16 |
| corvus | note the lack of '.' | 21:16 |
| mriedem | corvus: hmm, ok | 21:17 |
| corvus | and that's what's in the 'From:' header | 21:17 |
| fungi | yep, i concur | 21:17 |
| fungi | the subscription confirmation includes the . in the user part of the address | 21:17 |
| clarkb | ianw: all that to say I think the current stack (ignoring bugs with implementation) is likely the path forward | 21:17 |
| mriedem | ok changing to mriedemos@gmail.com | 21:18 |
| fungi | i guess those are different aliases for the same account? | 21:18 |
| clarkb | prometheanfire: I think your test cases needed to be updated on https://review.openstack.org/#/c/610107/2..3 | 21:19 |
| clarkb | prometheanfire: the new cases always set the dns settings | 21:19 |
| fungi | mriedem: but yes, mailman has no way to know they're the same account since they're distinctly different addresses | 21:19 |
| mriedem | ok changed, | 21:19 |
| fungi | i'm happy to approve that one through the moderation queue for you anyway so you don't need to resend | 21:19 |
| mriedem | and i can confirm i was using mriedemos@gmail.com with the other lists... | 21:19 |
| mriedem | browser must have auto-filled | 21:19 |
| fungi | fun times! | 21:19 |
| mriedem | fungi: thanks | 21:20 |
| fungi | done! | 21:20 |
| prometheanfire | ya | 21:21 |
| prometheanfire | clarkb: I noticed that | 21:21 |
| prometheanfire | clarkb: not actually sure what to do there | 21:23 |
| clarkb | prometheanfire: you just need to update your new test cases to remove the dns stuff | 21:23 |
| clarkb | since we shouldn't set dns on those because skip dns is set | 21:23 |
| *** betherly has quit IRC | 21:23 | |
| prometheanfire | clarkb: shouldn't there be a test with dns enabled though (along with disabled)? | 21:24 |
| prometheanfire | clarkb: if I do that I fail the 'test_glean' test | 21:24 |
| clarkb | prometheanfire: maybe? I don't know if glean tests both paths for all platforms | 21:24 |
| prometheanfire | since that sets skip_dns to false | 21:24 |
| prometheanfire | so, since it's testing both true and false, and I only have one template file one or the other will always fail | 21:26 |
| clarkb | prometheanfire: you may need different fixtures in that case | 21:26 |
| clarkb | have a .dns fixture and check if that exists when skip_dns == True? | 21:26 |
| prometheanfire | ianw did the test_glean_systemd one | 21:26 |
| clarkb | the thing that makes this weird is the dns stuff going into the per interface files and not just resolv.conf | 21:26 |
| prometheanfire | ya, because dns is per interface (an interface property) | 21:27 |
| prometheanfire | that's how it's delivered via the cloud_init info | 21:27 |
| prometheanfire | and supported in networkd | 21:27 |
| *** bobh has quit IRC | 21:28 | |
| clarkb | its actually really weird to think of dns that way (beacuse dns is a layer 7 service for name lookups and interfaces are layer 3 (ish) constructs) | 21:28 |
| clarkb | but ya | 21:29 |
| prometheanfire | it is odd, but any service could be the same (just listening or sending requests on that interface) | 21:30 |
| prometheanfire | not normal for things like dns though, but it is how networkd expects it | 21:30 |
| prometheanfire | I think I have to make a 'if distro == networkd: set skip_dns to false' | 21:32 |
| *** e0ne has quit IRC | 21:32 | |
| prometheanfire | even on the skip dns test | 21:32 |
| clarkb | prometheanfire: no you can load a different fixture if skip_dns is set to false | 21:32 |
| prometheanfire | clarkb: as part of output_filename? | 21:33 |
| clarkb | prometheanfire: at line 195ish https://review.openstack.org/#/c/610107/3/glean/tests/test_glean.py check if there is a output_path + '.dns' file on disk and use that as output_path instead if it exists and skip_dns is false | 21:34 |
| clarkb | then you have the two versions of that file, with and without the dns data, in the fixtures dir | 21:34 |
| prometheanfire | clarkb: yep, that's where I was going with it (as you led me there) :D | 21:34 |
| prometheanfire | clarkb: means copying a ton of tests though, but sure | 21:38 |
| clarkb | just those for networkd/gentoo right? | 21:39 |
| prometheanfire | clarkb: with more if statements it could be limited to that (networkd is the 'distro' here, and rax-iad is the only scenario that it's hit in) | 21:39 |
| clarkb | prometheanfire: I think if the .dns file doesn't exist use the other file | 21:40 |
| clarkb | then you only have to copy them for gentoo/networkd as those are the only cases that will differe | 21:40 |
| clarkb | *differ | 21:40 |
| ianw | can you do something like i did with networkd and have a %DNS% entry in the file and patch it in or out? | 21:40 |
| prometheanfire | ok | 21:40 |
| *** xek has quit IRC | 21:40 | |
| prometheanfire | ianw: where is that? | 21:40 |
| ianw | prometheanfire: https://review.openstack.org/#/c/618964/7/glean/tests/test_glean.py line 204 | 21:41 |
| ianw | sorry i have to run out for a little, will come back and look properly in a bit | 21:42 |
| *** ahosam has joined #openstack-infra | 21:42 | |
| cmurphy | ianw: fungi if you have time and interest i have a few easy noop patches sitting in my queue that i'd like to clear out https://review.openstack.org/601464 https://review.openstack.org/581389 https://review.openstack.org/581697 https://review.openstack.org/582153 https://review.openstack.org/585692 https://review.openstack.org/577521 https://review.openstack.org/581335 | 21:43 |
| prometheanfire | will look, but I'd rather have a fixture for dns than have the test to special things (test the code, not the test) | 21:44 |
| *** agopi is now known as agopi|brb | 21:46 | |
| *** trown is now known as trown|outtypewww | 21:48 | |
| *** kjackal has quit IRC | 21:48 | |
| *** jcoufal_ has quit IRC | 21:50 | |
| openstackgerrit | Matthew Thode proposed openstack-infra/glean master: write dns info to networkd https://review.openstack.org/610107 | 21:50 |
| *** yamamoto has joined #openstack-infra | 21:50 | |
| prometheanfire | clarkb: let's try this ^ | 21:50 |
| prometheanfire | passes tests locally at least | 21:51 |
| *** agopi|brb has quit IRC | 21:51 | |
| *** betherly has joined #openstack-infra | 21:55 | |
| openstackgerrit | Colleen Murphy proposed openstack-infra/system-config master: Update puppetlabs-postgresql to 4.8.0 https://review.openstack.org/558995 | 21:57 |
| openstackgerrit | Colleen Murphy proposed openstack-infra/system-config master: Make ask/ask-staging numeric https://review.openstack.org/558992 | 21:57 |
| openstackgerrit | Colleen Murphy proposed openstack-infra/system-config master: Minor fork puppet-solr https://review.openstack.org/559178 | 21:57 |
| cmurphy | clarkb: fungi there are a few changes between ianw and i that depend on that stack ^ | 21:57 |
| *** eumel8 has quit IRC | 21:59 | |
| mnaser | any reported issues at rax? | 21:59 |
| *** mriedem has quit IRC | 21:59 | |
| mnaser | just had a job timeout there and zuul couldnt even connect to get logs | 21:59 |
| mnaser | and im trying to know if we broke a vm terribly :) or if its just an issue | 22:00 |
| clarkb | mnaser: could be the duplicate IP issue | 22:00 |
| *** betherly has quit IRC | 22:00 | |
| prometheanfire | mnaser: I haven't seen emails go out (incident management) | 22:00 |
| mnaser | i see | 22:00 |
| mnaser | http://zuul.openstack.org/stream/36278e5399614b8bb4c3e089da25aabf?logfile=console.log (warning really really big log) | 22:01 |
| mnaser | Timeout exception waiting for the logger. Please check connectivity to [104.130.216.193:19885] | 22:01 |
| *** betherly has joined #openstack-infra | 22:02 | |
| clarkb | prometheanfire: we've seen evidence of duplicate IP addrs in rax | 22:02 |
| clarkb | prometheanfire: cloudnull followed up on some of them week before summit but its possible whatever is causing it has caused more of them | 22:02 |
| clarkb | prometheanfire: if you are able checking if 104.130.216.193 is used in multiple places would be the check aiui | 22:02 |
| prometheanfire | ah, I can bug someone, but he's probably off right now | 22:04 |
| prometheanfire | msg'd him | 22:06 |
| *** kgiusti has left #openstack-infra | 22:06 | |
| *** betherly has quit IRC | 22:06 | |
| gmann | corvus: clarkb frickler do you know why it still pick bionic for py3 job - https://review.openstack.org/#/c/618169/11/.zuul.yaml@148 | 22:08 |
| gmann | it should pick xenial when py3 job running on stable/rocky | 22:08 |
| clarkb | gmann: I can take a look in a few minutes. The zuul info dir with the inventory file should show you the job inheritance path | 22:08 |
| *** rcernin has joined #openstack-infra | 22:09 | |
| gmann | clarkb: thanks, i checked that and job inheritance looks from stable/rocky all the way - http://logs.openstack.org/76/618176/1/check/tempest-full-py3/96b7329/zuul-info/inventory.yaml | 22:09 |
| *** priteau has quit IRC | 22:16 | |
| *** agopi|brb has joined #openstack-infra | 22:19 | |
| ianw | cmurphy: ok, that pre-existing comment in 581335 is now one of my favourites | 22:22 |
| cmurphy | heh | 22:24 |
| cmurphy | thanks for the reviews ianw | 22:33 |
| openstackgerrit | Merged openstack-infra/puppet-mailman master: Add beaker tests for custom providers https://review.openstack.org/601464 | 22:35 |
| *** rpioso is now known as rpioso|afk | 22:36 | |
| clarkb | gmann: ok this is weird | 22:38 |
| *** roman_g has quit IRC | 22:40 | |
| *** roman_g has joined #openstack-infra | 22:40 | |
| clarkb | gmann: that ran against 610977,2 but 610977,4 is current. Any chance the changes to make the nodeset explict there fix it? | 22:41 |
| clarkb | gmann: I'm wondering if openstack-single-node nodeset is not looked up in a branch specific context | 22:41 |
| *** imacdonn has quit IRC | 22:42 | |
| clarkb | hrm except that openstack-single-node nodeset appaers to have always been xenial not bionic | 22:42 |
| *** imacdonn has joined #openstack-infra | 22:42 | |
| gmann | yes, that is xenial always | 22:42 |
| gmann | i can check for explicit nodeset. but what difference we have between tempest-full and tempest-full-py3 is these extra branches setting - http://git.openstack.org/cgit/openstack-infra/openstack-zuul-jobs/tree/zuul.d/project-templates.yaml#n1401 | 22:44 |
| gmann | which does not exclude stable/rocky | 22:45 |
| *** roman_g has quit IRC | 22:45 | |
| clarkb | https://review.openstack.org/#/c/618169/4/.zuul.yaml is the version of tempest change that was tested there | 22:45 |
| clarkb | which is quite a bit older than current patchset. I'm trying to read through it and see if anything there may explain it | 22:45 |
| clarkb | gmann: https://review.openstack.org/#/c/611572/1/.zuul.yaml that is why | 22:47 |
| *** roman_g has joined #openstack-infra | 22:47 | |
| gmann | clarkb: ah i see, and that has not set the branch thing for py3 job | 22:47 |
| clarkb | gmann: the py3 job depends on devstack-tempest and since tempest is branchless there it takes that as is | 22:48 |
| gmann | but why it did not take latest version 618169/11 | 22:48 |
| clarkb | that is a different change | 22:48 |
| clarkb | https://review.openstack.org/#/c/611572/1 is the change | 22:49 |
| openstackgerrit | Merged openstack-infra/puppet-log_processor master: Add beaker tests https://review.openstack.org/581697 | 22:49 |
| openstackgerrit | Merged openstack-infra/puppet-statusbot master: Add beaker tests https://review.openstack.org/582153 | 22:49 |
| openstackgerrit | Merged openstack-infra/puppet-openstackci master: Update Gemfile for Zuulv3 https://review.openstack.org/577521 | 22:49 |
| openstackgerrit | Merged openstack-infra/puppet-openstackci master: Add beaker tests for nodepool https://review.openstack.org/577522 | 22:49 |
| openstackgerrit | Merged openstack-infra/puppet-subunit2sql master: Fix subunit2sql install guard for idempotency https://review.openstack.org/581335 | 22:49 |
| openstackgerrit | Merged openstack-infra/puppet-elastic_recheck master: Add beaker tests https://review.openstack.org/581389 | 22:49 |
| clarkb | Does that change need to exist? there should be the one chagne in tempest (618169) and the one change in devstack (610977) and the nova change that tests them | 22:49 |
| clarkb | gmann: I think order should be nova change depends on tempest change (618169) depends on devstack change (610977). I don't know where 611572 fits in | 22:50 |
| gmann | that was old one for old PS. | 22:50 |
| gmann | with current testing that is not in pic | 22:50 |
| gmann | that came there as it picked 618169/4 not 618169/11 | 22:51 |
| clarkb | ok then I think you just need to rerun your tests and check again | 22:51 |
| gmann | and in same run, tempest-full did pick the right one 618169/11 - http://logs.openstack.org/76/618176/1/check/tempest-full/9fe23a5/zuul-info/inventory.yaml | 22:51 |
| *** betherly has joined #openstack-infra | 22:51 | |
| gmann | clarkb: ok let me re run | 22:52 |
| clarkb | gmann: the patchset for 618169 in that inventory is 4 not 11 | 22:52 |
| clarkb | oh for tempest full | 22:52 |
| gmann | yeah | 22:52 |
| gmann | it was strange in same run different job picked different PS | 22:53 |
| clarkb | gmann: the py3 job ran a week ago | 22:53 |
| clarkb | I think the one with 11 ran after a recheck | 22:53 |
| clarkb | gmann: http://logs.openstack.org/76/618176/1/check/tempest-full-py3/45fac7f/ that is py3 that ran recently | 22:54 |
| clarkb | and it ran on xenial so I think it is working | 22:54 |
| gmann | clarkb: ahh, got it. | 22:57 |
| *** betherly has quit IRC | 22:57 | |
| gmann | my bad i picked up wrong run seems | 22:57 |
| prometheanfire | cloudnull: we've been talking about networkd on rhel8 as well | 23:00 |
| openstackgerrit | Clark Boylan proposed openstack-infra/glean master: write dns info to networkd https://review.openstack.org/610107 | 23:00 |
| clarkb | prometheanfire: ^ fyi I just fixed the pep8 issue | 23:00 |
| prometheanfire | clarkb: yep, thanks | 23:01 |
| prometheanfire | was just going to fix it, but less for me is better for me :D | 23:01 |
| gmann | clarkb: thanks, it's my early morning side effect seems. | 23:01 |
| * prometheanfire hates rhel more than normal right now | 23:02 | |
| * prometheanfire blames oracle | 23:02 | |
| *** diablo_rojo has quit IRC | 23:02 | |
| clarkb | prometheanfire: at the summit it came up that we are so afraid of changing other peoples patches that I figure fixing formatting issues like this is a good way to start pushing towards not being afraid ofthat :) | 23:02 |
| prometheanfire | meh, at this point I don't care so much, guess that means I should touch tonyb's patch then | 23:04 |
| clarkb | considering round trip time on CI is a cost being able to make CI green trivially so that when ianw or tonyb wake up its all ready to go is probably a good thing | 23:05 |
| *** betherly has joined #openstack-infra | 23:08 | |
| ianw | we could maybe do like the debian low-threshold nmu page, and somehow have people opt in to explicitly say "i'm happy for you to update my patch" | 23:11 |
| ianw | it doesn't really change anything, but people might feel better if the author has explicitly said "i won't get mad about it" :) | 23:12 |
| ianw | clarkb: you really got me thinking now on != and = behaviour on that skip_when patch ... | 23:13 |
| clarkb | ianw: it was just weird to me that the skip would match if the value was unset | 23:14 |
| clarkb | makes sense for != becuase unset can't equal a value | 23:14 |
| clarkb | but for = should be false I think | 23:14 |
| ianw | actually i think the opposite, see my comment :) | 23:15 |
| ianw | i think a unit test here would be most helpful to make sure the combinations are coming out as expected | 23:16 |
| *** rlandy is now known as rlandy|bbl | 23:17 | |
| ianw | which of course means a yak shaving adventure on unit testing elements, i'm not sure we do that | 23:18 |
| *** betherly has quit IRC | 23:18 | |
| *** ahosam has quit IRC | 23:19 | |
| clarkb | another option is for it to be an error if no value is set | 23:19 |
| clarkb | then it isn't ambiguous but slightly more work on element authors | 23:20 |
| clarkb | I think that would work for your existing use case without any changes to the nm stuff | 23:20 |
| ianw | yeah, i could buy that too | 23:24 |
| *** rm_work has quit IRC | 23:33 | |
| *** rm_work has joined #openstack-infra | 23:33 | |
| *** florianf has quit IRC | 23:38 | |
| *** slaweq has quit IRC | 23:45 | |
| *** lbragstad is now known as lbragstad_turkey | 23:46 | |
| *** betherly has joined #openstack-infra | 23:48 | |
| *** zul has quit IRC | 23:51 | |
| *** betherly has quit IRC | 23:53 | |
| clarkb | http://logs.openstack.org/85/605585/7/check/system-config-run-docker/62b9e11/job-output.txt.gz#_2018-10-31_20_34_58_500426 is what our iptables ruleset looks like with docker installed | 23:55 |
| *** lbragstad_turkey has quit IRC | 23:56 | |
| ianw | clarkb: oh, good -- yeah that was what i looked at initially and just threw up my hands :) | 23:57 |
Generated by irclog2html.py 2.15.3 by Marius Gedminas - find it at mg.pov.lt!