| *** nicolasbock has joined #openstack-ansible | 00:26 | |
| *** nicolasbock1 has joined #openstack-ansible | 00:26 | |
| *** nicolasbock has quit IRC | 00:30 | |
| *** nicolasbock1 is now known as nicolasbock | 00:30 | |
| *** born2bake has quit IRC | 00:31 | |
| *** cshen has joined #openstack-ansible | 00:36 | |
| *** cshen has quit IRC | 00:40 | |
| *** d34dh0r53 has quit IRC | 00:51 | |
| *** cloudnull has quit IRC | 00:52 | |
| snadge | is there a way to override the download location for the EPEL repository? | 00:58 |
|---|---|---|
| snadge | if it uses the default download.fedoraproject.org .. it could end up using one of very many mirrors, which is difficult to whitelist | 00:59 |
| *** nurdie has quit IRC | 01:05 | |
| *** hamzy_ has quit IRC | 01:33 | |
| *** nurdie has joined #openstack-ansible | 01:39 | |
| *** nurdie has quit IRC | 01:44 | |
| *** nurdie has joined #openstack-ansible | 02:13 | |
| *** nurdie has quit IRC | 02:18 | |
| *** nurdie has joined #openstack-ansible | 02:22 | |
| *** grumbert has quit IRC | 02:23 | |
| *** spsurya has joined #openstack-ansible | 02:32 | |
| *** nurdie has quit IRC | 03:13 | |
| *** nurdie has joined #openstack-ansible | 03:13 | |
| *** nurdie has quit IRC | 03:18 | |
| *** radeks has joined #openstack-ansible | 03:19 | |
| *** nicolasbock has quit IRC | 03:23 | |
| *** radeks has quit IRC | 03:59 | |
| snadge | im getting lxc containers failed to start.. but it doesn't appear to be overwriting the log files in /var/log/lxc | 04:18 |
| snadge | is there a way to reset the lxc container state.. without blowing everything away and starting again | 04:18 |
| snadge | lxc-destroy -n seems to do the job | 04:21 |
| *** cshen has joined #openstack-ansible | 04:40 | |
| *** cshen has quit IRC | 04:45 | |
| *** cshen has joined #openstack-ansible | 04:56 | |
| *** cshen has quit IRC | 05:01 | |
| *** dave-mccowan has quit IRC | 05:08 | |
| *** jhesketh has quit IRC | 05:11 | |
| *** markvoelker has joined #openstack-ansible | 05:21 | |
| *** markvoelker has quit IRC | 05:26 | |
| jrosser | snadge: there is a playbook for destroying containers and their associated data | 06:27 |
| jrosser | https://github.com/openstack/openstack-ansible/blob/master/playbooks/lxc-containers-destroy.yml | 06:28 |
| snadge | i'll check that out.. my next problem is.. the bootstrap aio takes care of creating all the network interfaces etc.. in particular br-mgmt | 06:30 |
| snadge | im just trying to go one step beyond that very basic configuration, and have 5 physical vms that I want to do an install on, but only really the management or primary interface needs to work | 06:30 |
| jrosser | for the yum repo thing, the internet suggests “In the files in /etc/yum.repo.d/*.repo , use baseurl instead of mirrorlist.” | 06:31 |
| snadge | yeah the problem is the EPEL repo is added by ansible during the setup hosts playbook | 06:33 |
| snadge | that was something i was going to look at, as well.. to see if a config can override that, for the moment i just edited the file immediately after it gets deployed, and before it installs from it | 06:34 |
| jrosser | the ansible module appears to understand mirrorlist/baseurl https://docs.ansible.com/ansible/latest/modules/yum_repository_module.html | 06:35 |
| jrosser | tbh my systems are pretty much entirely Ubuntu, if you have chance to try out some of those yum_repository options we could make a patch | 06:36 |
| jrosser | for going beyond an AIO to multiple VM it’s now up to you how you want the networking to be | 06:37 |
| jrosser | the only requirement is that the necessary br-* bridges exist when the config is done | 06:37 |
| jrosser | could be done with multiple interfaces / vlans / whatever you like | 06:38 |
| jrosser | snadge: look here https://github.com/openstack/openstack-ansible-openstack_hosts/blob/master/tasks/openstack_hosts_configure_yum.yml#L73 | 06:40 |
| jrosser | if you find where openstack_hosts_package_repos is defined, and you want to get rid of any mirrorlist setup, override that variable with just the fields that you want | 06:42 |
| jrosser | well that’s not epel - but you get the idea | 06:44 |
| snadge | excellent.. yeah.. my AIO example uses ubuntu 18.04, but production was RHEL 7, but will likely be Centos 7, unless there's a reason to go newer | 06:49 |
| *** ianychoi has quit IRC | 06:51 | |
| *** jbadiapa has joined #openstack-ansible | 06:52 | |
| *** ianychoi has joined #openstack-ansible | 06:54 | |
| *** gokhani has joined #openstack-ansible | 07:01 | |
| snadge | i've made a note of that, and will sort it out tomorrow. thanks :) (yay AU timezone) | 07:02 |
| *** jbadiapa has quit IRC | 07:08 | |
| *** tosky has joined #openstack-ansible | 07:14 | |
| *** pcaruana has joined #openstack-ansible | 07:16 | |
| *** jbadiapa has joined #openstack-ansible | 07:21 | |
| *** miloa has joined #openstack-ansible | 07:29 | |
| *** cshen has joined #openstack-ansible | 07:29 | |
| miloa | Morning | 07:29 |
| openstackgerrit | Jonathan Rosser proposed openstack/openstack-ansible master: [WIP] Centos-8 support https://review.opendev.org/689629 | 07:42 |
| openstackgerrit | Bernd Mueller proposed openstack/ansible-hardening master: changed disable dccp conf for preventing kernel messages https://review.opendev.org/689682 | 07:43 |
| openstackgerrit | Bernd Mueller proposed openstack/ansible-hardening master: changed disable dccp conf for preventing kernel messages https://review.opendev.org/689682 | 07:48 |
| *** rpittau|afk is now known as rpittau | 07:51 | |
| *** lkoranda has joined #openstack-ansible | 07:56 | |
| openstackgerrit | Jonathan Rosser proposed openstack/openstack-ansible master: [WIP] Centos-8 support https://review.opendev.org/689629 | 07:59 |
| openstackgerrit | OpenStack Proposal Bot proposed openstack/openstack-ansible master: Imported Translations from Zanata https://review.opendev.org/689684 | 08:02 |
| *** lkoranda_ has joined #openstack-ansible | 08:04 | |
| *** lkoranda has quit IRC | 08:06 | |
| *** DanyC has joined #openstack-ansible | 08:08 | |
| *** DanyC has quit IRC | 08:10 | |
| *** tosky_ has joined #openstack-ansible | 08:15 | |
| *** tosky_ has quit IRC | 08:15 | |
| *** tosky_ has joined #openstack-ansible | 08:15 | |
| *** tosky is now known as Guest39640 | 08:16 | |
| *** tosky_ is now known as tosky | 08:16 | |
| *** DanyC has joined #openstack-ansible | 08:23 | |
| *** electrofelix has joined #openstack-ansible | 08:35 | |
| ebbex | anyone else running into problems building networking-calico on the repo_containers? something about a mismatch between versions 3.8.4 and 3.9.2 | 08:36 |
| *** vesper has joined #openstack-ansible | 08:40 | |
| *** vesper11 has quit IRC | 08:40 | |
| *** hamzy_ has joined #openstack-ansible | 08:41 | |
| CeeMac | morning | 08:42 |
| jrosser | ebbex: sort of - i see a load of our CI jobs are broken with exactly the same thing | 08:53 |
| *** kopecmartin has joined #openstack-ansible | 09:01 | |
| jrosser | ebbex: which branch are you deploying? | 09:01 |
| tacco | short question regarding the secret_uuid. Do i have to UUID decode the key from the keyring or can i just generate "something" random ? | 09:05 |
| tacco | but first of all good morning everyone | 09:05 |
| tacco | s/decode/encode/ | 09:11 |
| *** lkoranda has joined #openstack-ansible | 09:16 | |
| *** lkoranda_ has quit IRC | 09:16 | |
| ebbex | jrosser: rocky. I've worked around it by going into networking-calico and running "git tag --annotate -m "Calico v3.9.2" --force 3.9.2" | 09:35 |
| ebbex | if you run "git describe" in the dir it reports back 3.8.4 but "git tag" reports a bunch of versions/tags after that release. | 09:36 |
| *** openstackgerrit has quit IRC | 09:37 | |
| noonedeadpunk | mornings | 09:44 |
| noonedeadpunk | oh, that doesn't look good... https://zuul.opendev.org/t/openstack/build/95c59b998d2947ebac4fd08cafe1400b/log/logs/openstack/aio1_repo_container-9cf59a00/repo/wheel_build.log.txt.gz#162 | 09:48 |
| noonedeadpunk | so looks like we have broken rocky and queens | 09:52 |
| jrosser | noonedeadpunk: i think we use master version of networking-calico everywhere | 09:57 |
| noonedeadpunk | yep, that's true. but there's no change since last release in terms of sha... | 09:58 |
| noonedeadpunk | But I guess new tags were added | 09:59 |
| noonedeadpunk | as 3.8.4 and 3.9.2 are made on the same commit | 09:59 |
| noonedeadpunk | (as well as all tags in between) | 10:01 |
| miloa | noonedeadpunk: Hi, I saw your comment on Change 671762, I will check Logan comments asap, but i am actually very busy, sorry. :) | 10:04 |
| noonedeadpunk | miloa: yep, sure:) thanks:) | 10:04 |
| noonedeadpunk | we may backport this into train later | 10:05 |
| jrosser | noonedeadpunk: is this the one which needs mnaser vote https://review.opendev.org/#/c/686573/ ? | 10:06 |
| noonedeadpunk | actually, that one will require it https://review.opendev.org/#/c/686581/ But if he's not agree with 686573, than updating 686581 has not so much sense | 10:08 |
| noonedeadpunk | ebbex: yeah, our gates are also complaining about it... The things is that 3.8.4 and 3.9.2 are made against the same commit... | 10:10 |
| jrosser | noonedeadpunk: i wonder if that is just a mistake - those tags all being at the same point | 10:16 |
| jrosser | it looks very very strange indeed | 10:16 |
| *** lkoranda_ has joined #openstack-ansible | 10:17 | |
| *** lkoranda has quit IRC | 10:19 | |
| ebbex | Well, even if you check out master, you'll get an error on a version mismatch between 3.9.2 and 3.8.4, because it's looking at git to verify some version. Did they do something wrong in the tagging process? | 10:24 |
| ebbex | (setup.py that is.) | 10:24 |
| *** lkoranda_ has quit IRC | 10:24 | |
| jrosser | from a git point of view nothing stops you attaching as many tags to a particular revision as you like | 10:25 |
| ebbex | why does git describe" | 10:25 |
| ebbex | why does "git describe" return 3.8.4 on master though? | 10:25 |
| jrosser | well, thats the thing, i guess a tag is just that, a tag, it doesnt convey any meaning as far as git is concerned | 10:26 |
| jrosser | and i expect that pbr is making an assumption that there is only one tag | 10:27 |
| ebbex | i force updated the tag to 3.9.2 and everything worked as normal. (I don't know if 'annotate' is/was important in that step.) | 10:28 |
| jrosser | ebbex: by a bit of random experimenting here whichever tag was last modified seems to be the one returned by git describe when there are several at the same position in the history | 10:31 |
| noonedeadpunk | maybe we should point to tag at this point instead of sha? | 10:32 |
| noonedeadpunk | at least for the calico | 10:32 |
| jrosser | i'm not sure if that will help, becasue the stack trace comes from inside pbr? | 10:32 |
| ebbex | Is this the bug we're experiencing? https://bugs.launchpad.net/pbr/+bug/1453996 | 10:38 |
| openstack | Launchpad bug 1453996 in PBR "multiple tags of the same commit have non-deterministic results" [High,Triaged] | 10:38 |
| jrosser | it looks very much like it | 10:40 |
| ebbex | It's been a known issue for ~4 years, still not solved. Yikes. | 10:42 |
| jrosser | you can sort of show when the tags were updated | 10:43 |
| jrosser | git log --tags --simplify-by-decoration --pretty="format:%ai %d" | 10:43 |
| jrosser | but there still seems to be an inner ordering of all of the ones at HEAD | 10:44 |
| *** lkoranda has joined #openstack-ansible | 10:51 | |
| noonedeadpunk | hm, interestingly how they had created tags https://opendev.org/openstack/releases/src/branch/master/deliverables/_independent/networking-calico.yaml | 10:57 |
| ebbex | I updated the pbr bug-report, any of you wanna chime in on it? | 11:00 |
| *** openstackgerrit has joined #openstack-ansible | 11:01 | |
| openstackgerrit | Merged openstack/openstack-ansible master: Imported Translations from Zanata https://review.opendev.org/689684 | 11:01 |
| *** lkoranda has quit IRC | 11:07 | |
| *** markvoelker has joined #openstack-ansible | 11:10 | |
| *** trident has quit IRC | 11:11 | |
| *** trident has joined #openstack-ansible | 11:15 | |
| *** markvoelker has quit IRC | 11:16 | |
| noonedeadpunk | I guess we should run update-alternatives to set default python now... | 11:17 |
| openstackgerrit | Bernd Mueller proposed openstack/ansible-hardening master: changed disable dccp conf for preventing kernel messages https://review.opendev.org/689682 | 11:18 |
| noonedeadpunk | In terms of "/bin/sh: /usr/bin/python: No such file or directory" https://openstack.fortnebula.com:13808/v1/AUTH_e8fd161dc34c421a979a9e6421f823e9/zuul_opendev_logs_ad3/689629/10/check/openstack-ansible-deploy-aio_metal-centos-8/ad37079/job-output.txt | 11:18 |
| *** goldyfruit has quit IRC | 11:26 | |
| *** lkoranda has joined #openstack-ansible | 11:33 | |
| *** strobelight has joined #openstack-ansible | 11:48 | |
| *** mgariepy has joined #openstack-ansible | 11:51 | |
| *** markvoelker has joined #openstack-ansible | 11:59 | |
| jrosser | noonedeadpunk: heres what update-alternatives says on centos-8 http://paste.openstack.org/show/785356/ | 12:00 |
| noonedeadpunk | lol /usr/libexec/no-python | 12:00 |
| jrosser | which is "echo "Run: man unversioned-python"" | 12:01 |
| noonedeadpunk | maybe we can do smth like "update-alternatives --install /usr/bin/python python /usr/bin/python3.6"? | 12:02 |
| jrosser | i did locally just hack that to python3 where that job failed | 12:02 |
| jrosser | there was a fairly long discussion about this on the ML recently | 12:02 |
| noonedeadpunk | oh, I've missed it | 12:03 |
| jrosser | here http://lists.openstack.org/pipermail/openstack-discuss/2019-October/010146.html | 12:04 |
| *** kopecmartin has quit IRC | 12:05 | |
| *** kopecmartin has joined #openstack-ansible | 12:06 | |
| *** nicolasbock has joined #openstack-ansible | 12:13 | |
| selvakumarnms | Hi All , can you please help me in how to configure HA proxy & keepalived for the openstack controller nodes? Is there any sample inventory I can refer? do let me know. | 12:15 |
| jrosser | selvakumarnms: there are a few reference examples in the openstack-ansible repo, like this for the config https://github.com/openstack/openstack-ansible/blob/master/etc/openstack_deploy/openstack_user_config.yml.prod.example | 12:19 |
| jrosser | and this for the user_variables.yml https://github.com/openstack/openstack-ansible/blob/master/etc/openstack_deploy/user_variables.yml.prod.example | 12:20 |
| selvakumarnms | can we have two haproxy_hosts? also can it be different than shared-infra_hosts or os-infra-hosts? | 12:20 |
| jrosser | but it is almost certain that you need to customise these for your environment | 12:20 |
| jrosser | yes it is, in fact I have dedicated haproxy nodes in my environment | 12:20 |
| selvakumarnms | thanks #jrosser | 12:20 |
| jrosser | selvakumarnms: in the example config the haproxy nodes are here https://github.com/openstack/openstack-ansible/blob/master/etc/openstack_deploy/openstack_user_config.yml.prod.example#L105-L111 | 12:23 |
| jrosser | but you can define them to be a different set of hosts if you need to | 12:23 |
| selvakumarnms | thanks jrosser I will refer and configure. 172.29.236.12 is the container ip I have to give or host physical ips? | 12:25 |
| jrosser | everything you see in openstack_user_config like that is physical host IP | 12:26 |
| jrosser | the container IP are allocated dynamically by the openstack-ansible dynamic inventory script | 12:26 |
| jrosser | to keep things simple the reference configs also assmune that the openstack management network is the one that use also use for your deploy host and to ssh to the physical servers | 12:27 |
| selvakumarnms | okay cool. | 12:27 |
| jrosser | but again openstack-ansible doesnt require that, and if you want a specific network for the bare metal that is also possible | 12:28 |
| selvakumarnms | I am using physical host for deploying openstack. what is vip IP do I need to give? unused IP from the physical hosts? | 12:29 |
| selvakumarnms | I mean internal vip or external vip | 12:30 |
| jrosser | the internal VIP should be one of the addresses out of the 'used_ips' range for the openstack management network | 12:30 |
| jrosser | the external one is up to you as you need to provide your own external network | 12:30 |
| selvakumarnms | so basically unused ip address? | 12:33 |
| jrosser | in fact exactly like this https://github.com/openstack/openstack-ansible/blob/master/etc/openstack_deploy/openstack_user_config.yml.prod.example#L14 | 12:33 |
| jrosser | selvakumarnms: which one, the internal or external? | 12:33 |
| selvakumarnms | internal | 12:34 |
| selvakumarnms | thanks again for the rerference link | 12:34 |
| jrosser | it should be one from this range (or equivalent in your environment) https://github.com/openstack/openstack-ansible/blob/master/etc/openstack_deploy/openstack_user_config.yml.prod.example#L8 | 12:34 |
| jrosser | those ip are reserved for the bare metal hosts mgmt interface IP | 12:34 |
| jrosser | and also for whatever routers/gateways physical networking equipment you may need | 12:35 |
| jrosser | the rest of the IP outside the reserved range are allocated randomly to containers by openstack-ansible | 12:35 |
| selvakumarnms | ok got it. How to specify other used IPs(physical hosts) from the same network in the inventory? can I also mention in used_ips? | 12:40 |
| jrosser | do you have a dedicated subnet for this? | 12:41 |
| noonedeadpunk | jrosser: ebbex logan- seems like calico is generaly broken now http://paste.openstack.org/show/785357/ | 12:42 |
| selvakumarnms | yes for example I have another hosts also in the same subnet. so I dont want to take away that ip by openstack-ansible | 12:42 |
| jrosser | so you should specify everything that you don't want OSA to touch in used_ips | 12:42 |
| arxcruz | odyssey4me: https://github.com/ansible/ansible-lint/issues/595 | 12:43 |
| noonedeadpunk | I'm really wondering how did they made these tags... | 12:44 |
| selvakumarnms | ok thanks jrosser for your help | 12:44 |
| jrosser | selvakumarnms: i would recommend having a good block of addresses available, even for a modest lab /24 felt a bit small for mgmt network | 12:46 |
| selvakumarnms | yes I do agree | 12:46 |
| jrosser | noonedeadpunk: i am wondering if networking-calico is still an official openstack deliverable | 12:47 |
| noonedeadpunk | hm, I guess that everything under https://opendev.org/openstack/ should use deliverables? | 12:47 |
| jrosser | it doesnt seem to be here https://opendev.org/openstack/governance/src/branch/master/reference/projects.yaml | 12:48 |
| noonedeadpunk | that's true. I just don't know the way of managing repos under openstack without releases | 12:49 |
| jrosser | noonedeadpunk: https://review.opendev.org/#/c/659337/4/RELEASING.md | 12:50 |
| *** DanyC has quit IRC | 12:52 | |
| jrosser | looks like they have some kind of independant release process there | 12:55 |
| noonedeadpunk | Yep, but still don't undertand it... As I'd dropped some tags from it... | 13:01 |
| noonedeadpunk | Submitted bug report https://bugs.launchpad.net/networking-calico/+bug/1849110. | 13:01 |
| openstack | Launchpad bug 1849110 in networking-calico "Tagging several versions on a single commit breaks installation" [Undecided,New] | 13:01 |
| noonedeadpunk | logan-: I can recall that you had some slack channel of calico folks? Can you reach them regarding ^? | 13:01 |
| noonedeadpunk | As I guess master is also broken | 13:02 |
| jrosser | becasue the repo build works differently i think we will only see that on the neutron role? | 13:02 |
| noonedeadpunk | I guess yes | 13:03 |
| noonedeadpunk | I mean that I didn't check neutron role, but yes - we wouldn't see it anywhere else | 13:04 |
| noonedeadpunk | mnaser: can you kindly check https://review.opendev.org/#/c/689145/ as it should do db migration for placement? | 13:14 |
| *** DanyC has joined #openstack-ansible | 13:14 | |
| *** strobelight_ has joined #openstack-ansible | 13:15 | |
| *** strobelight is now known as Guest76784 | 13:15 | |
| *** strobelight_ is now known as strobelight | 13:16 | |
| *** lkoranda has quit IRC | 13:17 | |
| *** Guest76784 has quit IRC | 13:18 | |
| *** lkoranda has joined #openstack-ansible | 13:19 | |
| *** lkoranda has quit IRC | 13:23 | |
| *** michielp has quit IRC | 13:25 | |
| *** dave-mccowan has joined #openstack-ansible | 13:28 | |
| *** lkoranda_ has joined #openstack-ansible | 13:28 | |
| *** nurdie has joined #openstack-ansible | 13:28 | |
| *** goldyfruit has joined #openstack-ansible | 13:32 | |
| *** lkoranda has joined #openstack-ansible | 13:32 | |
| *** nurdie has quit IRC | 13:33 | |
| *** lkoranda_ has quit IRC | 13:34 | |
| *** hamzaachi has joined #openstack-ansible | 13:36 | |
| noonedeadpunk | folks, can someone review https://review.opendev.org/#/c/679245/ one more time? | 13:50 |
| noonedeadpunk | oh, for the first time:) | 13:51 |
| jrosser | hmm 6 weeks ago | 13:52 |
| jrosser | the logs are gone :) | 13:52 |
| openstackgerrit | kourosh vivan proposed openstack/openstack-ansible-lxc_hosts stable/stein: Add Centos GPG key custom url https://review.opendev.org/689103 | 13:53 |
| *** cshen has quit IRC | 14:11 | |
| *** ianychoi has quit IRC | 14:21 | |
| *** dmsimard1 is now known as dmsimard | 14:26 | |
| *** spatel has joined #openstack-ansible | 14:28 | |
| *** ianychoi has joined #openstack-ansible | 14:29 | |
| *** nurdie has joined #openstack-ansible | 14:35 | |
| *** nurdie_ has joined #openstack-ansible | 14:38 | |
| *** nurdie has quit IRC | 14:40 | |
| *** macz has joined #openstack-ansible | 14:54 | |
| *** lkoranda has quit IRC | 14:56 | |
| *** goldyfruit has quit IRC | 15:08 | |
| *** goldyfruit has joined #openstack-ansible | 15:08 | |
| *** goldyfruit has quit IRC | 15:09 | |
| *** nurdie_ is now known as nurdie | 15:09 | |
| openstackgerrit | Dmitriy Rabotyagov (noonedeadpunk) proposed openstack/openstack-ansible-os_manila master: PDF Documentation Build tox target https://review.opendev.org/689820 | 15:25 |
| *** trident has quit IRC | 15:29 | |
| mgariepy | the logs from the last run are gone, but the logs from aug 29 are not.. that's kinda weird (https://review.opendev.org/#/c/679245/) | 15:32 |
| *** trident has joined #openstack-ansible | 15:32 | |
| openstackgerrit | Dmitriy Rabotyagov (noonedeadpunk) proposed openstack/openstack-ansible-os_manila master: service_setup: refactor service setup to a single file https://review.opendev.org/689824 | 15:35 |
| noonedeadpunk | yep, let's recheck | 15:35 |
| openstackgerrit | Dmitriy Rabotyagov (noonedeadpunk) proposed openstack/openstack-ansible master: Add cloudkitty and zun to a-r-r https://review.opendev.org/689828 | 15:39 |
| *** ianychoi has quit IRC | 15:43 | |
| *** ianychoi has joined #openstack-ansible | 15:44 | |
| *** markvoelker has quit IRC | 15:46 | |
| *** markvoelker has joined #openstack-ansible | 15:49 | |
| *** nurdie has quit IRC | 15:54 | |
| *** nurdie has joined #openstack-ansible | 15:58 | |
| noonedeadpunk | rgh, cirros again fails... | 16:03 |
| *** mgariepy has quit IRC | 16:06 | |
| logan- | noonedeadpunk: yep, i'll ping on their slack, i did test out making a new commit and tag and it seemed to make pbr happy again, so it seems like it'll be pretty easy to fix that up | 16:08 |
| jrosser | noonedeadpunk: i was thinking about that patch again for cirros | 16:08 |
| jrosser | and it might be nicer if we use stat in the os_tempest role to check if it is already downloaded | 16:08 |
| jrosser | i wasnt too happy that my patch overrode the os_tempest defaults in a slightly not obvious way for CI tests | 16:09 |
| *** hamzaachi has quit IRC | 16:09 | |
| *** mgariepy has joined #openstack-ansible | 16:10 | |
| noonedeadpunk | yep, that's exactly what I've tried to do, but faced with some probem in it. | 16:11 |
| noonedeadpunk | but more porblem is that your method will copy image on local filesystem when it can be omited. | 16:11 |
| jrosser | and i think the uri module will check the SHA too which is loads of IO for no reason | 16:12 |
| *** hamzaachi has joined #openstack-ansible | 16:13 | |
| *** hamzaachi has quit IRC | 16:18 | |
| *** markvoelker has quit IRC | 16:23 | |
| *** nurdie has quit IRC | 16:24 | |
| *** nicolasbock has quit IRC | 16:26 | |
| noonedeadpunk | jrosser: so I wanted to do smth like this http://paste.openstack.org/show/785372/ in https://opendev.org/openstack/openstack-ansible-os_tempest/src/branch/master/tasks/tempest_resources.yml#L32 but I finished up that I don't know how to map the stat result to tempest_imagesall in all... | 16:30 |
| noonedeadpunk | and I'm not sure if list order is smth I can rely on | 16:30 |
| noonedeadpunk | to map results by their index | 16:31 |
| *** macz has quit IRC | 16:31 | |
| *** macz has joined #openstack-ansible | 16:31 | |
| jrosser | noonedeadpunk: probably with_together loop? | 16:32 |
| noonedeadpunk | and another option to use block/recover there | 16:32 |
| noonedeadpunk | yeah, maybe, but I'm not sure how good this idea is... It.s actually mapping by id inside list | 16:34 |
| noonedeadpunk | and another option was to use another block with rescue | 16:34 |
| noonedeadpunk | but this will require 2 using 2 blocks and 2 delegates instead of 1 | 16:35 |
| noonedeadpunk | just to do smth like try/except | 16:36 |
| noonedeadpunk | oh, there's always which is kinda finally I guess | 16:36 |
| *** spsurya has quit IRC | 16:40 | |
| *** nurdie has joined #openstack-ansible | 16:45 | |
| noonedeadpunk | lol | 16:45 |
| noonedeadpunk | https://opendev.org/openstack/openstack-ansible/src/branch/master/tests/roles/bootstrap-host/templates/user_variables.aio.yml.j2#L216-L227 | 16:46 |
| *** nurdie has quit IRC | 16:49 | |
| noonedeadpunk | so, generaly, real file download shouldn't occur... | 16:51 |
| noonedeadpunk | "if a checksum is passed to this parameter, and the file exist under the dest location, the destination_checksum would be calculated, and if checksum equals destination_checksum, the file download would be skipped" | 16:51 |
| *** electrofelix has quit IRC | 16:56 | |
| *** miloa has quit IRC | 16:56 | |
| openstackgerrit | Dmitriy Rabotyagov (noonedeadpunk) proposed openstack/openstack-ansible-os_tempest master: Don't download image when it's already there https://review.opendev.org/689843 | 16:59 |
| noonedeadpunk | wondering if it will work... ^ | 16:59 |
| *** DanyC has quit IRC | 17:02 | |
| *** markvoelker has joined #openstack-ansible | 17:04 | |
| *** KeithMnemonic has joined #openstack-ansible | 17:05 | |
| *** nicolasbock has joined #openstack-ansible | 17:17 | |
| *** hamzaachi has joined #openstack-ansible | 17:28 | |
| *** hamzaachi has quit IRC | 17:29 | |
| bjoernt | Hey it seems OSA is not really suitable for 100% baremetal deployments as per https://bugs.launchpad.net/openstack-ansible/+bug/1847395. What is the feeling around this, should this be a option or we better to remove the support for baremetal entirely | 17:32 |
| openstack | Launchpad bug 1847395 in openstack-ansible "Nova/Neutron Playbook run on baremetal disables keystone service" [Undecided,New] | 17:32 |
| *** weifan has joined #openstack-ansible | 17:41 | |
| *** weifan has quit IRC | 17:41 | |
| *** weifan has joined #openstack-ansible | 17:41 | |
| *** weifan has quit IRC | 17:46 | |
| *** nurdie has joined #openstack-ansible | 17:46 | |
| jrosser | bjoernt: there was a ton of effort put into bare metal co-installability during the last cycle | 17:46 |
| bjoernt | train ? | 17:46 |
| jrosser | and there are real clouds running OSA on bare metal today, perhaps needing a large bunch of config overrides to achieve that | 17:47 |
| jrosser | bjoernt: yes but not addressing the specific thing you make a bug about though | 17:47 |
| jrosser | you will see that the CI jobs for bare metal don’t include all components because this work is not yet complete | 17:48 |
| bjoernt | yes I have one engineer looking at that problem because there are cross roles issues between apache2, nginx-full and light | 17:49 |
| bjoernt | one role does turnoff /remove the package from the other | 17:49 |
| jrosser | federation makes that complicated currently afaik | 17:50 |
| *** tosky has quit IRC | 17:50 | |
| ThiagoCMC | I can't wait to try OSA Train! Sorry about the buzz... I'm excited... lol | 17:50 |
| *** weifan has joined #openstack-ansible | 17:53 | |
| cjloader | saw this similar thing in stein, had to manually rerun playbooks to keep services from going down | 17:55 |
| cjloader | bjoernt: | 17:55 |
| bjoernt | yeah its an aspect of not all roles using the same wsgi service due to many reasons | 17:55 |
| *** prometheanfire has quit IRC | 17:56 | |
| jrosser | well like I say the aio_metal scenario currently avoids these issues by removing components | 17:57 |
| jrosser | so extra work is needed to make that more complete | 17:57 |
| *** prometheanfire has joined #openstack-ansible | 17:57 | |
| cjloader | so what you're saying is its not fully tested? | 17:58 |
| bjoernt | I check train and if there are still similar issues then I add a warning that those problems are known before making deployment choices | 17:59 |
| jrosser | train is no different | 17:59 |
| jrosser | there are a bunch of extra knobs to twiddle for what ports things bind to | 18:00 |
| jrosser | which is part of the story | 18:00 |
| jrosser | but imho if you want to do a metal deployment with the same slickness that you can do lxc today, then a bunch of extra work needs contributing | 18:01 |
| *** shananigans has joined #openstack-ansible | 18:06 | |
| jrosser | bjoernt: noonedeadpunk and I have been slowly chipping away at the bare metal use case last cycle | 18:10 |
| *** weifan has quit IRC | 18:10 | |
| jrosser | but there is still work to do so if you have effort to help complete that it would be great to accelerate it | 18:10 |
| bjoernt | Yeah thats fine and appreciated perhaps we start thinking about a some docs what feature or what configuration is considered stable/tested to make those things more obvious | 18:12 |
| *** canori01 has quit IRC | 18:13 | |
| *** weifan has joined #openstack-ansible | 18:14 | |
| *** weifan has quit IRC | 18:23 | |
| *** int-0x21 has joined #openstack-ansible | 18:27 | |
| *** weifan has joined #openstack-ansible | 18:31 | |
| *** goldyfruit has joined #openstack-ansible | 18:43 | |
| *** kopecmartin is now known as kopecmartin|off | 18:50 | |
| *** goldenfri has joined #openstack-ansible | 18:53 | |
| *** weifan has quit IRC | 18:55 | |
| *** DanyC has joined #openstack-ansible | 18:59 | |
| *** weifan has joined #openstack-ansible | 19:02 | |
| *** spatel has quit IRC | 19:04 | |
| *** weifan has quit IRC | 19:12 | |
| *** CeeMac has quit IRC | 19:15 | |
| *** weifan has joined #openstack-ansible | 19:17 | |
| *** DanyC_ has joined #openstack-ansible | 19:18 | |
| *** starborn has joined #openstack-ansible | 19:18 | |
| *** DanyC has quit IRC | 19:18 | |
| *** DanyC has joined #openstack-ansible | 19:19 | |
| *** weifan has quit IRC | 19:22 | |
| *** DanyC_ has quit IRC | 19:23 | |
| *** aedc has quit IRC | 19:26 | |
| *** gyee has joined #openstack-ansible | 19:29 | |
| *** DanyC has quit IRC | 19:30 | |
| *** weifan has joined #openstack-ansible | 19:33 | |
| *** schwicht has joined #openstack-ansible | 19:39 | |
| *** aedc has joined #openstack-ansible | 19:43 | |
| openstackgerrit | Merged openstack/openstack-ansible-os_manila master: PDF Documentation Build tox target https://review.opendev.org/689820 | 19:47 |
| *** weifan has quit IRC | 19:59 | |
| openstackgerrit | Merged openstack/ansible-hardening master: changed disable dccp conf for preventing kernel messages https://review.opendev.org/689682 | 20:07 |
| *** spatel has joined #openstack-ansible | 20:08 | |
| *** markvoelker has quit IRC | 20:11 | |
| *** markvoelker has joined #openstack-ansible | 20:12 | |
| *** weifan has joined #openstack-ansible | 20:14 | |
| *** weifan has quit IRC | 20:15 | |
| *** pcaruana has quit IRC | 20:16 | |
| *** weifan has joined #openstack-ansible | 20:29 | |
| *** shananigans has quit IRC | 20:32 | |
| *** dave-mccowan has quit IRC | 20:47 | |
| *** spatel has quit IRC | 20:51 | |
| *** nurdie_ has joined #openstack-ansible | 20:57 | |
| *** starborn has quit IRC | 20:58 | |
| openstackgerrit | Merged openstack/openstack-ansible-os_placement master: Add nova to placement DB migration https://review.opendev.org/689145 | 20:59 |
| *** nurdie_ has quit IRC | 20:59 | |
| *** nurdie has quit IRC | 21:00 | |
| *** goldyfruit has quit IRC | 21:01 | |
| *** jralbert has joined #openstack-ansible | 21:10 | |
| *** tosky has joined #openstack-ansible | 21:10 | |
| jralbert | I think this change breaks repo usage for CentOS/RedHat folks: https://review.opendev.org/#/c/608636/ | 21:11 |
| *** schwicht has quit IRC | 21:11 | |
| *** dave-mccowan has joined #openstack-ansible | 21:11 | |
| jralbert | The change to the lineinfile invocation against yum.conf specifies quotes around the url which the preceding release did not create, so the directive to remove the line "succeeds" but removes nothing, and the proxy line remains in place | 21:12 |
| jralbert | As a result, when following the upgrade procedure, the apt-cacher-ng service is removed but the containers are all still configured to use it | 21:12 |
| jralbert | The result is play failures in the galera upgrade step, and the first affected galera node being brought down and not brought back up | 21:13 |
| jralbert | The regexp for the lineinfile invocation either needs to omit the quotes, or allow them to be an optional requirement if we really do believe some installations would have had them | 21:14 |
| jralbert | Do folks generally agree? If so I can propose a change | 21:14 |
| *** DanyC has joined #openstack-ansible | 21:18 | |
| jrosser | jralbert: looks like the ' got changed for " in the lineinfile | 21:24 |
| *** markvoelker has quit IRC | 21:25 | |
| jrosser | if you could test a fix and make a patch that would be great | 21:25 |
| jralbert | @jrosser the line as created by the queens installer includes no quotes, single or double | 21:25 |
| jralbert | And it's even worse than that: the decision to remove the caching service also misses that the percona yum repos also specify the proxy line | 21:26 |
| jralbert | so the galera upgrade is completely hosed, and it turns out the upgrade play goes ahead and actually removes the package before it establishes that a new one can be installed | 21:26 |
| jrosser | i really meant the ' vs " in the lineinfile tast | 21:26 |
| jrosser | *task | 21:26 |
| jralbert | The single quotes in the pre-change version were for Ansible's benefit; the changed version moves to the heredoc syntax, and therefore doesn't need them, but the quotes are also moved to surround just the url | 21:29 |
| jralbert | It's possible (likely?) that this issue will also apply to DNF-using hosts, though I don't have an installation to test that on | 21:30 |
| jrosser | well please do make a patch, particularly if you are in a position to have somewhere to test it out | 21:33 |
| *** goldyfruit has joined #openstack-ansible | 21:38 | |
| *** nurdie has joined #openstack-ansible | 21:40 | |
| *** DanyC has quit IRC | 21:43 | |
| *** DanyC has joined #openstack-ansible | 21:44 | |
| *** nurdie has quit IRC | 21:45 | |
| *** schwicht has joined #openstack-ansible | 22:00 | |
| *** goldyfruit has quit IRC | 22:01 | |
| *** dave-mccowan has quit IRC | 22:03 | |
| *** schwicht has quit IRC | 22:05 | |
| *** schwicht has joined #openstack-ansible | 22:06 | |
| *** openstackgerrit has quit IRC | 22:07 | |
| *** openstackgerrit has joined #openstack-ansible | 22:11 | |
| openstackgerrit | Merged openstack/openstack-ansible stable/stein: Bump SHAs for stable/stein https://review.opendev.org/689653 | 22:11 |
| *** jhesketh has joined #openstack-ansible | 22:12 | |
| *** dave-mccowan has joined #openstack-ansible | 22:14 | |
| *** hwoarang has quit IRC | 22:19 | |
| *** hwoarang has joined #openstack-ansible | 22:20 | |
| *** schwicht has quit IRC | 22:37 | |
| *** nicolasbock has quit IRC | 22:45 | |
| *** weifan has quit IRC | 22:50 | |
| *** weifan has joined #openstack-ansible | 22:51 | |
| *** weifan has quit IRC | 22:54 | |
| *** macz has quit IRC | 22:54 | |
| *** macz has joined #openstack-ansible | 22:55 | |
| *** ianychoi has quit IRC | 23:00 | |
| *** ianychoi has joined #openstack-ansible | 23:00 | |
| *** weifan has joined #openstack-ansible | 23:03 | |
| *** DanyC has quit IRC | 23:05 | |
| redkrieg | Has anyone seen a case where a single hypervisor gets a "BadStatusLine" error when reaching out to the placement api? I've done a complete reinstall and my playbooks have not changed from my other regions. This is the only hypervisor in the region with the issue: http://paste.openstack.org/show/785456/ | 23:08 |
| *** ianychoi has quit IRC | 23:24 | |
| *** ianychoi has joined #openstack-ansible | 23:26 | |
| *** markvoelker has joined #openstack-ansible | 23:26 | |
| *** markvoelker has quit IRC | 23:30 | |
| *** tosky has quit IRC | 23:32 | |
| *** weifan has quit IRC | 23:57 | |
| *** weifan has joined #openstack-ansible | 23:57 | |
| *** weifan has quit IRC | 23:58 | |
| *** weifan has joined #openstack-ansible | 23:58 | |
Generated by irclog2html.py 2.15.3 by Marius Gedminas - find it at mg.pov.lt!