*** gyee has quit IRC | 00:07 | |
*** spatel has joined #openstack-ansible | 01:01 | |
*** cshen has joined #openstack-ansible | 01:34 | |
*** cshen has quit IRC | 01:39 | |
*** markvoelker has joined #openstack-ansible | 01:39 | |
*** markvoelker has quit IRC | 01:44 | |
*** NewJorg has quit IRC | 02:16 | |
*** cshen has joined #openstack-ansible | 03:35 | |
*** spatel has quit IRC | 03:36 | |
*** cshen has quit IRC | 03:39 | |
*** markvoelker has joined #openstack-ansible | 03:40 | |
*** markvoelker has quit IRC | 03:50 | |
*** udesale has joined #openstack-ansible | 03:52 | |
*** pmacdonnell has quit IRC | 03:56 | |
*** spikysnow_ has joined #openstack-ansible | 04:03 | |
spikysnow_ | Hello I want to ask about cinder_service_internaluri_insecureTo ignore validate_certs cinder_service_internaluri_insecure is set to false, but based on the name like keystone's keystone_service_internaluri_insecure is set to true to disable validate_certs. https://usercontent.irccloud-cdn.com/file/ssjfuXum/image.png | 04:06 |
---|---|---|
*** rh-jelabarre has quit IRC | 04:23 | |
*** evrardjp has quit IRC | 04:34 | |
*** evrardjp has joined #openstack-ansible | 04:34 | |
*** cshen has joined #openstack-ansible | 04:45 | |
*** cshen has quit IRC | 04:49 | |
*** NewJorg has joined #openstack-ansible | 05:48 | |
*** jbadiapa has quit IRC | 06:18 | |
jrosser | spikysnow_: the internal endpoint is normally http rather than https so there would not be a cert to validate either way | 06:20 |
*** rpittau|afk is now known as rpittau | 06:20 | |
openstackgerrit | Jonathan Rosser proposed openstack/openstack-ansible-tests stable/ussuri: Rewrite CI provided apt sources to be http instead of https on focal https://review.opendev.org/737907 | 06:32 |
*** jbadiapa has joined #openstack-ansible | 06:36 | |
*** miloa has joined #openstack-ansible | 06:38 | |
*** grantza has joined #openstack-ansible | 06:39 | |
openstackgerrit | Jonathan Rosser proposed openstack/openstack-ansible-lxc_hosts stable/ussuri: Stop installing python2.7 on debian/focal https://review.opendev.org/737769 | 06:47 |
openstackgerrit | Jonathan Rosser proposed openstack/openstack-ansible-lxc_hosts stable/ussuri: Add functional tests for ubuntu focal https://review.opendev.org/737770 | 06:49 |
*** spikysnow_ has quit IRC | 07:04 | |
*** luksky has joined #openstack-ansible | 07:09 | |
*** cshen has joined #openstack-ansible | 07:11 | |
*** jcath has joined #openstack-ansible | 07:23 | |
*** tosky has joined #openstack-ansible | 07:42 | |
openstackgerrit | Jonathan Rosser proposed openstack/openstack-ansible-lxc_hosts stable/ussuri: Add functional tests for ubuntu focal https://review.opendev.org/737770 | 07:42 |
jrosser | noonedeadpunk: if you're around we can start merging ussuri stuff for focal starting here https://review.opendev.org/#/c/737770/ | 08:28 |
*** jawad_ax_ has joined #openstack-ansible | 08:35 | |
*** spatel has joined #openstack-ansible | 08:37 | |
*** jawad_axd has quit IRC | 08:39 | |
*** spatel has quit IRC | 08:42 | |
admin0 | good morning .. does osa work with external ceph ? | 08:55 |
jrosser | admin0: absolutely yes - though i think the documentation is a little weak for this | 08:58 |
jrosser | admin0: at the most basic level the integration is done here https://github.com/openstack/openstack-ansible-ceph_client/blob/master/defaults/main.yml#L48-L50 | 09:00 |
jrosser | you don't do any ceph deployment with OSA | 09:00 |
jrosser | but you tell it a list of IP addresses for the mons in the external ceph, and it will then ssh to them and recover the ceph.conf and ceph keys | 09:00 |
jrosser | if it is not possible to ssh to the external mons (different team maintains ceph for example) then you can provide manual ceph conf and keys on the deploy host | 09:01 |
jrosser | you have to have all the networking setup up properly so that the containers/hosts can contact the external ceph cluster over whichever interface you need it to | 09:02 |
admin0 | right now, i have 2 platforms with osa+ceph integrated .. i am think if i should continue let osa manage ceph, or decouple ceph from osa | 09:03 |
jrosser | personally i separate them | 09:10 |
jrosser | becasue then i can have whatever version of ceph i want with whatever version of openstack | 09:10 |
jrosser | and the upgrade cycles are decoupled | 09:10 |
jrosser | but i did plan right from the start that the ceph hardware / deployment would be independant | 09:11 |
openstackgerrit | Merged openstack/openstack-ansible-lxc_container_create stable/ussuri: Add lxc config key mapping for version 4.x https://review.opendev.org/737779 | 09:27 |
openstackgerrit | Merged openstack/openstack-ansible-lxc_container_create stable/ussuri: Use a single vars file for all ubuntu releases https://review.opendev.org/737778 | 09:27 |
openstackgerrit | Jonathan Rosser proposed openstack/openstack-ansible-tests master: Bump ansible version to 2.9.10 https://review.opendev.org/737935 | 09:33 |
openstackgerrit | Jonathan Rosser proposed openstack/openstack-ansible master: Bump ansible version to 2.9.10 https://review.opendev.org/737936 | 09:34 |
openstackgerrit | Merged openstack/openstack-ansible-lxc_container_create stable/ussuri: Use btrfs-progs instead of btrfs-tools in functional tests https://review.opendev.org/737767 | 09:39 |
*** stingrayza has joined #openstack-ansible | 09:48 | |
*** also_stingrayza has quit IRC | 09:51 | |
openstackgerrit | Merged openstack/openstack-ansible-tests stable/ussuri: Rewrite CI provided apt sources to be http instead of https on focal https://review.opendev.org/737907 | 09:56 |
*** cshen has quit IRC | 10:08 | |
openstackgerrit | Merged openstack/openstack-ansible-plugins master: Optimise linear strategy access to physical host variables https://review.opendev.org/737543 | 10:14 |
*** rpittau is now known as rpittau|bbl | 10:20 | |
openstackgerrit | Jonathan Rosser proposed openstack/openstack-ansible-plugins stable/ussuri: Optimise linear strategy access to physical host variables https://review.opendev.org/737963 | 10:52 |
openstackgerrit | Jonathan Rosser proposed openstack/openstack-ansible-plugins stable/train: Optimise linear strategy access to physical host variables https://review.opendev.org/737964 | 10:52 |
openstackgerrit | Jonathan Rosser proposed openstack/openstack-ansible-plugins stable/stein: Optimise linear strategy access to physical host variables https://review.opendev.org/737965 | 10:52 |
jrosser | noonedeadpunk: looks like we have an issue in galera role https://opendev.org/openstack/openstack-ansible-galera_server/commit/3d405dfd52c0a5059cefd877fd578114bcdd912d | 11:10 |
jrosser | 2020-06-25 10:15:55.804604 | ubuntu-bionic | RUNNING HANDLER [galera_server : Start new cluster] **************************** | 11:10 |
jrosser | 2020-06-25 10:15:56.291103 | ubuntu-bionic | fatal: [aio1_galera_container-019ae3c4]: FAILED! => {"msg": "'hostvars' is undefined"} | 11:10 |
jrosser | i think that it's not allowed to access hostvars in task variables like that | 11:10 |
jrosser | but this only becomes a problem with ansible 2.9.10 | 11:11 |
openstackgerrit | Merged openstack/openstack-ansible-lxc_hosts stable/ussuri: Stop installing python2.7 on debian/focal https://review.opendev.org/737769 | 11:14 |
openstackgerrit | Jonathan Rosser proposed openstack/openstack-ansible-lxc_hosts master: Remove support for Unubut 16.04 https://review.opendev.org/737978 | 11:20 |
openstackgerrit | Jonathan Rosser proposed openstack/openstack-ansible-lxc_hosts master: Do not install python2 or its dev package on Ubuntu Bionic hosts https://review.opendev.org/737979 | 11:20 |
openstackgerrit | Merged openstack/openstack-ansible-lxc_hosts stable/ussuri: Add functional tests for ubuntu focal https://review.opendev.org/737770 | 11:23 |
openstackgerrit | Jonathan Rosser proposed openstack/openstack-ansible-plugins master: Update CI tests for currently supported operating systems https://review.opendev.org/737981 | 11:24 |
openstackgerrit | Jonathan Rosser proposed openstack/openstack-ansible-galera_server master: Update CI tests for currently supported operating systems https://review.opendev.org/737982 | 11:28 |
openstackgerrit | Jonathan Rosser proposed openstack/openstack-ansible-tests master: Remove opensuse-15 jobs https://review.opendev.org/737985 | 11:32 |
noonedeadpunk | jrosser: btw, do we have speedup on master after having https://review.opendev.org/#/c/737543 merged? | 11:39 |
*** ghanima has quit IRC | 11:44 | |
openstackgerrit | Dmitriy Rabotyagov (noonedeadpunk) proposed openstack/openstack-ansible-os_rally master: Do not install development packages on the target https://review.opendev.org/731459 | 11:44 |
openstackgerrit | Dmitriy Rabotyagov (noonedeadpunk) proposed openstack/openstack-ansible master: [WIP] Centos-8 support https://review.opendev.org/689629 | 11:48 |
*** udesale_ has joined #openstack-ansible | 11:52 | |
openstackgerrit | Maksim Malchuk proposed openstack/openstack-ansible stable/train: Fix the name of the renamed network for trove https://review.opendev.org/737858 | 11:54 |
*** udesale has quit IRC | 11:54 | |
*** cshen has joined #openstack-ansible | 11:55 | |
jrosser | noonedeadpunk: i've not really managed to take a look yet | 12:01 |
noonedeadpunk | yeah, that's ok, will see on graphs later | 12:02 |
*** cshen has quit IRC | 12:04 | |
*** cshen has joined #openstack-ansible | 12:06 | |
jrosser | noonedeadpunk: there is a really big heap of stuff backported to ussuri for review - maybe over the first 3-4 pages of the review dashboard | 12:10 |
jrosser | plus i found the centos-8 patches being a bit random between python3 and python36 in one of the package names | 12:10 |
openstackgerrit | Merged openstack/openstack-ansible-plugins stable/ussuri: Optimise linear strategy access to physical host variables https://review.opendev.org/737963 | 12:11 |
openstackgerrit | Jonathan Rosser proposed openstack/openstack-ansible-lxc_hosts master: Remove support for Ubuntu 16.04 https://review.opendev.org/737978 | 12:11 |
*** rh-jelabarre has joined #openstack-ansible | 12:13 | |
ioni | is there a way to force python interpretor in openstack-ansible ? | 12:28 |
ioni | i'm trying to get magnum ussuri on top of stein :) | 12:29 |
jrosser | ioni: this is a starting point https://github.com/openstack/openstack-ansible-os_magnum/blob/master/defaults/main.yml#L20 | 12:33 |
ioni | jrosser, thanks | 12:33 |
jrosser | but i think you might find in stein the repo server wheel build process does not account for python version | 12:33 |
jrosser | so that might need some looking at too | 12:33 |
jrosser | as in, it makes a venv for the wheel building, and i beleive in train it can make both a py2 and py3 one in parallel if necessary | 12:33 |
openstackgerrit | Merged openstack/openstack-ansible-plugins stable/train: Optimise linear strategy access to physical host variables https://review.opendev.org/737964 | 12:34 |
*** rpittau|bbl is now known as rpittau | 12:42 | |
openstackgerrit | Georgina Shippey proposed openstack/openstack-ansible-os_keystone master: Identity providers can be created with specified domain https://review.opendev.org/736741 | 12:54 |
*** gshippey has joined #openstack-ansible | 12:58 | |
gshippey | cheers for getting on that review so fast @noonedeadpunk | 12:59 |
*** also_stingrayza has joined #openstack-ansible | 13:03 | |
*** stingrayza has quit IRC | 13:06 | |
openstackgerrit | Jonathan Rosser proposed openstack/openstack-ansible-tests stable/stein: Add pre-run playbook to clean up CI nodes before OSA tests run https://review.opendev.org/738012 | 13:16 |
ioni | jrosser, openstack-ansible os-magnum-install.yml -e 'venv_rebuild=true' -e 'venv_python_executable=python3' -e 'uwsgi_python_executable=python3' | 13:41 |
ioni | :D | 13:42 |
ioni | works fine now :) | 13:42 |
jrosser | awesome :) | 13:42 |
ioni | jrosser, i just needed: https://opendev.org/openstack/ansible-role-python_venv_build/commit/fb243ea7583a1d64c2ecbc51c841c6a2eedabbf9 | 13:42 |
jrosser | powerful tools..... | 13:42 |
ioni | indeed | 13:43 |
jrosser | yeah, i remember we had to make seperate venv for that otherwise something really odd happened | 13:43 |
ioni | still on stein, with magnum from ussuri | 13:43 |
ioni | with heat and octavia from train | 13:43 |
ioni | i have a frankenstein | 13:43 |
jrosser | blimey | 13:43 |
openstackgerrit | Merged openstack/openstack-ansible-nspawn_hosts stable/ussuri: Update minimum tox version and remove old workaround https://review.opendev.org/737801 | 13:52 |
openstackgerrit | Jonathan Rosser proposed openstack/openstack-ansible master: Remove opensuse jobs and project templates https://review.opendev.org/738031 | 13:57 |
openstackgerrit | Jonathan Rosser proposed openstack/openstack-ansible master: Remove opensuse jobs and project templates https://review.opendev.org/738031 | 14:02 |
mgariepy | if you need me to review something you can ping me, i have a lot to do on otehr stuff today be i can take the time to review a few patch | 14:03 |
openstackgerrit | Georgina Shippey proposed openstack/openstack-ansible-os_keystone stable/ussuri: Add Paramaters to httpd.conf template https://review.opendev.org/738034 | 14:03 |
jrosser | mgariepy: cool thanks - almost everything for focal is now backported to ussuri, maybe not merged yet | 14:04 |
mgariepy | great :D | 14:04 |
mgariepy | just in time for me new deployment. ;) haha | 14:05 |
openstackgerrit | Jonathan Rosser proposed openstack/openstack-ansible master: Drop support for openSUSE https://review.opendev.org/725541 | 14:05 |
*** spatel has joined #openstack-ansible | 14:10 | |
spatel | jrosser: I have replied on your comments here - https://review.opendev.org/#/c/736850/2/tasks/lxc_container_network_new.yml | 14:12 |
openstackgerrit | Dmitriy Rabotyagov (noonedeadpunk) proposed openstack/openstack-ansible-os_barbican master: Add Centos-8 support https://review.opendev.org/736932 | 14:12 |
jrosser | spatel: have you pressed "reply?" | 14:13 |
spatel | Yes | 14:13 |
noonedeadpunk | and post ?:) | 14:13 |
spatel | i can see i can see my reply on that link let me check | 14:13 |
noonedeadpunk | regardig this topic - we should use python36-devel http://paste.openstack.org/show/795210/ | 14:14 |
noonedeadpunk | yeah, you will see it even without publishing it | 14:14 |
noonedeadpunk | but it will be visable only for you | 14:14 |
spatel | I think i forgot to click on "done" | 14:14 |
noonedeadpunk | no, that't not the reason | 14:14 |
noonedeadpunk | let me make a screenshot for you or smth | 14:15 |
spatel | check this out - https://imgur.com/a/gG2XL4X | 14:15 |
noonedeadpunk | so after that you need this https://imgur.com/8tAXwvM | 14:17 |
noonedeadpunk | as you saved this as draft | 14:17 |
spatel | Oh! i gave answer in comments section, instead of "Reply" let me do with reply | 14:18 |
noonedeadpunk | no-no | 14:18 |
jrosser | spatel: its the same :) | 14:18 |
noonedeadpunk | you did right | 14:18 |
noonedeadpunk | but you need to publish them | 14:18 |
jrosser | you can either reply to someone elses comment just as a comment | 14:18 |
jrosser | or you can put stuff associtated with particular lines of code like i did originally | 14:19 |
jrosser | both are fine | 14:19 |
noonedeadpunk | that reply button will publish your draft | 14:19 |
*** markvoelker has joined #openstack-ansible | 14:19 | |
jrosser | you get to go back and fiddle with it all, maybe in many files before publishing it | 14:19 |
spatel | I think i found it, check it out again | 14:20 |
noonedeadpunk | yeah | 14:20 |
jrosser | spatel: anyway - take a look at this https://github.com/openstack/openstack-ansible-lxc_hosts/blob/master/templates/prep-scripts/centos_7_prep.sh.j2#L38 | 14:20 |
jrosser | in the centos-7 cache prep script systemd-networkd is enabled right there | 14:20 |
jrosser | so i figure we need to do same for centos-8? | 14:21 |
noonedeadpunk | jrosser: I guess you pathced it for centos 8? | 14:22 |
jrosser | i suspect it isn't in the centos-8 patch, lets look | 14:22 |
jrosser | missing from here https://review.opendev.org/#/c/735781/4/templates/prep-scripts/centos_8_prep.sh.j2 | 14:23 |
jrosser | also take a look at why i -1 the patch, the mirrors are all wrong | 14:23 |
spatel | oh possible its missing centos8 file | 14:24 |
jrosser | because we set a var in the CI job which overrides all the settings in lxc_hosts and is not accounting for centos-7 vs. centos-8 | 14:24 |
*** markvoelker has quit IRC | 14:24 | |
jrosser | we need to unblock stein with this https://review.opendev.org/#/c/738012/ | 14:26 |
noonedeadpunk | jrosser: maybe we can wait for centos8 before updating tests? | 14:28 |
jrosser | hmm? | 14:29 |
noonedeadpunk | https://review.opendev.org/#/c/737982/1 | 14:29 |
noonedeadpunk | and like drop suse+centos7& | 14:29 |
noonedeadpunk | just not to do the work twice | 14:30 |
openstackgerrit | Dmitriy Rabotyagov (noonedeadpunk) proposed openstack/openstack-ansible-openstack_hosts master: Enable PowerTools repository for distro installs https://review.opendev.org/737284 | 14:30 |
jrosser | oh i see yes, we can do that | 14:30 |
*** jcath has quit IRC | 14:30 | |
jrosser | noonedeadpunk: opensuse is properly broken for train https://zuul.openstack.org/builds?job_name=openstack-ansible-deploy-aio_metal-opensuse-15 | 14:35 |
jrosser | did we agree to make that non-voting? | 14:36 |
noonedeadpunk | yeah I saw that. we proably should set it to non-voting | 14:36 |
noonedeadpunk | yeah | 14:36 |
jrosser | ok i will do that | 14:36 |
openstackgerrit | Jonathan Rosser proposed openstack/openstack-ansible stable/train: Remove pre-gate cleanup tasks for the new plain CI images https://review.opendev.org/735785 | 14:38 |
spatel | jrosser: noonedeadpunk i think we are good with that patch (we just need to add re-enable in prep-scripts/centos_8_prep.sh.j2 file) | 14:41 |
spatel | Just drop my patch (i don't think we need that anymore) | 14:42 |
openstackgerrit | Jonathan Rosser proposed openstack/openstack-ansible-lxc_hosts master: Add centos-8 support https://review.opendev.org/735781 | 14:45 |
jrosser | spatel: ^ i added it, lets see what happens | 14:46 |
spatel | i can verify in my lab and see how it goes.. | 14:46 |
noonedeadpunk | jrosser: btw I've covered that override https://review.opendev.org/#/c/689629/70/tests/roles/bootstrap-host/tasks/prepare_aio_config.yml | 14:48 |
openstackgerrit | Merged openstack/openstack-ansible-lxc_hosts stable/ussuri: Download yum keys to host before installing https://review.opendev.org/737777 | 14:49 |
openstackgerrit | Merged openstack/openstack-ansible-lxc_hosts stable/ussuri: Do not set PYTHONPATH for tox tests https://review.opendev.org/737776 | 14:49 |
openstackgerrit | Merged openstack/openstack-ansible-lxc_hosts stable/ussuri: copy the actual keyring https://review.opendev.org/737280 | 14:49 |
openstackgerrit | Merged openstack/openstack-ansible-os_nova stable/ussuri: Install netcat-openbsd for live migration https://review.opendev.org/737787 | 14:49 |
*** watersj has joined #openstack-ansible | 14:50 | |
jrosser | noonedeadpunk: is it right to have the rdo mirror url inside the centos-7 test here https://review.opendev.org/#/c/689629/70/tests/roles/bootstrap-host/tasks/prepare_aio_config.yml ? | 14:54 |
noonedeadpunk | uh.... dunno, it was there.... But eventually I think it just brongs no effect... | 14:58 |
noonedeadpunk | *brings | 14:59 |
noonedeadpunk | as it just equal to default fromwhat I see | 14:59 |
noonedeadpunk | oh, it's master for all branches... | 15:03 |
jrosser | almost - i have a patch for that here https://review.opendev.org/#/c/734418/ | 15:05 |
openstackgerrit | Maksim Malchuk proposed openstack/openstack-ansible stable/train: Fix the name of the renamed network for trove https://review.opendev.org/737858 | 15:05 |
jrosser | we need a similar one for ussuri i guess | 15:05 |
noonedeadpunk | but we don't have them | 15:05 |
noonedeadpunk | there're no ussuri rdo packages for centos 7 | 15:06 |
openstackgerrit | Jonathan Rosser proposed openstack/openstack-ansible-openstack_hosts stable/ussuri: Use Ussuri branch RDO repo on OSA Ussuri branch https://review.opendev.org/738048 | 15:07 |
jrosser | oh... | 15:07 |
jrosser | well i guess this is a thing for the distro jobs | 15:08 |
noonedeadpunk | yeah | 15:08 |
jrosser | the source jobs shouldnt be affected | 15:08 |
noonedeadpunk | so upgrade for centos distro is not present | 15:08 |
openstackgerrit | Merged openstack/openstack-ansible-os_swift stable/ussuri: Cleanup ansible_python_interpreter https://review.opendev.org/737793 | 15:09 |
openstackgerrit | Merged openstack/openstack-ansible-os_ironic stable/ussuri: inspector: fix service catalog creation https://review.opendev.org/737785 | 15:13 |
openstackgerrit | Merged openstack/openstack-ansible-os_placement stable/ussuri: Cleanup ansible_python_interpreter https://review.opendev.org/737790 | 15:13 |
openstackgerrit | Jonathan Rosser proposed openstack/openstack-ansible stable/ussuri: Add Ubuntu Focal support https://review.opendev.org/737771 | 15:14 |
openstackgerrit | Merged openstack/openstack-ansible-tests stable/stein: Add pre-run playbook to clean up CI nodes before OSA tests run https://review.opendev.org/738012 | 15:14 |
*** jawad_ax_ has quit IRC | 15:16 | |
noonedeadpunk | jrosser: focal distro packages are going to appear only for victoria http://ubuntu-cloud.archive.canonical.com/ubuntu/dists/focal-proposed/ | 15:17 |
jrosser | yes, i think that means that UCA is not a thing for this cycle | 15:17 |
noonedeadpunk | and bionic will stop on ussuri... | 15:17 |
jrosser | it's set to an empty var for ussuri | 15:17 |
*** jawad_axd has joined #openstack-ansible | 15:18 | |
jrosser | so does that mean we have bionic distro jobs that are not what we think they are? | 15:18 |
noonedeadpunk | I wean they include ussuri but bionic won't have victoria distro packages | 15:19 |
* noonedeadpunk wondering how canonical see their customer upgrade workflow | 15:19 | |
*** jawad_axd has quit IRC | 15:22 | |
jrosser | yes so on our ussuri branch we have ussuri UCA for bionic and no repo for focal | 15:22 |
noonedeadpunk | yeah | 15:23 |
jrosser | i'll update master for victoria | 15:23 |
openstackgerrit | Merged openstack/openstack-ansible-os_octavia stable/ussuri: Cleanup ansible_python_interpreter https://review.opendev.org/737788 | 15:24 |
openstackgerrit | Jonathan Rosser proposed openstack/openstack-ansible-openstack_hosts master: Update UCA repo for Victoria https://review.opendev.org/738050 | 15:28 |
*** udesale_ has quit IRC | 15:28 | |
openstackgerrit | Merged openstack/openstack-ansible-os_ironic stable/ussuri: Cleanup ansible_python_interpreter https://review.opendev.org/737786 | 15:30 |
*** udesale has joined #openstack-ansible | 15:34 | |
mgariepy | focal has usuri directly in the main repos. https://wiki.ubuntu.com/FocalFossa/ReleaseNotes#OpenStack_Ussuri | 15:39 |
mgariepy | so it doesn't need uca for focal + usuri | 15:40 |
mgariepy | ussuri* | 15:40 |
mgariepy | from now on i'll call it U. | 15:41 |
*** cshen has quit IRC | 15:43 | |
noonedeadpunk | so, we just should try to add distro job for focal? | 15:47 |
*** cshen has joined #openstack-ansible | 15:47 | |
noonedeadpunk | interesting... | 15:47 |
openstackgerrit | Dmitriy Rabotyagov (noonedeadpunk) proposed openstack/openstack-ansible master: Add ubuntu focal distro jobs https://review.opendev.org/738054 | 15:49 |
noonedeadpunk | let's see... | 15:49 |
*** cshen has quit IRC | 15:52 | |
*** jawad_axd has joined #openstack-ansible | 15:53 | |
openstackgerrit | Merged openstack/openstack-ansible-ceph_client stable/ussuri: Cleanup ansible_python_interpreter https://review.opendev.org/737783 | 15:54 |
*** jawad_axd has quit IRC | 15:57 | |
*** rpittau is now known as rpittau|afk | 16:11 | |
vakuznet | what is changed in 20.1.2 that caused octavia to fail http://paste.openstack.org/show/795218/ | 16:12 |
vakuznet | request goes to publicURL despite neutron.endpoint_type is internalURL (no SSL) | 16:14 |
jrosser | out of interest which branch? | 16:14 |
jrosser | ok stable/train? | 16:16 |
vakuznet | stable train from 20.1.1 to 20.1.2 | 16:16 |
jrosser | vakuznet: as far as i know nothing ha changed in train, though i'm surprised to see a python 2.7 venv there | 16:16 |
openstackgerrit | Kevin Carter (cloudnull) proposed openstack/ansible-config_template master: Update config_template to use the builtin _get_remote_user function https://review.opendev.org/738058 | 16:16 |
vakuznet | it's centos7 -> so python 2.7 | 16:17 |
noonedeadpunk | octavia sha has changed | 16:18 |
openstackgerrit | Kevin Carter (cloudnull) proposed openstack/ansible-config_template master: Update config_template to use the builtin _get_remote_user function https://review.opendev.org/738058 | 16:18 |
noonedeadpunk | role is exactly the same | 16:19 |
noonedeadpunk | https://opendev.org/openstack/openstack-ansible/src/tag/20.1.2/playbooks/defaults/repo_packages/openstack_services.yml#L274 | 16:19 |
noonedeadpunk | from dcd80d3028e989fe5f2147b3ea8f074765eb953d to 09e863eb539e29c8950ddfce12d602c0e77c1b7e | 16:20 |
jrosser | oh hah | 16:21 |
jrosser | doh https://github.com/openstack/octavia/commit/09e863eb539e29c8950ddfce12d602c0e77c1b7e | 16:21 |
jrosser | argh this is going to break my deployment too :( | 16:22 |
noonedeadpunk | so we kida need to set endpoint in neutron section | 16:22 |
noonedeadpunk | I won't expect such changes to be backported tbh | 16:23 |
jrosser | johnsom: you around? | 16:24 |
johnsom | o/ | 16:24 |
noonedeadpunk | vakuznet: so it seems that instead of neutron.endpoint_type we need to set neutron.endpoint | 16:25 |
*** jawad_axd has joined #openstack-ansible | 16:25 | |
jrosser | johnsom: read back about 10 lines up there ^^^ | 16:25 |
johnsom | Ah, I see the bug in that patch yeah, it's not honoring the endpoint type... Sigh | 16:27 |
*** cshen has joined #openstack-ansible | 16:27 | |
jrosser | johnsom: so i think from deployment tool point of view there needs to be very strong understanding of which endpoint is used in which circumstances | 16:27 |
johnsom | This was backported as it is a low level security fix with neutron | 16:27 |
jrosser | we have users where the services incl octavia cannot reach the public endpoint | 16:28 |
jrosser | only internal | 16:28 |
johnsom | Yeah, the new code missed a few parameters when it looks up the endpoint in keystone. | 16:28 |
johnsom | They left out a few parameters it looks like | 16:31 |
jrosser | should this change what we need in the config file? | 16:31 |
johnsom | No, they messed up this patch | 16:32 |
*** cshen has quit IRC | 16:32 | |
johnsom | You can work around the bug by changing your config file, but there is no reason you should have to | 16:32 |
jrosser | if it's possible to give vakuznet a pointer to get unstuck, that would be cool | 16:33 |
johnsom | vakkuznet Yeah, so it looks like this patch didn't include some of the criteria it should have when it looks up the neutron endpoint in keystone. I will create a patch to fix this now. To work around it until we can get a patch out, you can set your desired neutron endpoint in the octavia.conf file on the API instances in the [neutron] endpoint setting. | 16:37 |
johnsom | That will override this broken keystone lookup | 16:37 |
jrosser | vakuznet: actually you can put octavia_git_install_branch: dcd80d3028e989fe5f2147b3ea8f074765eb953d into your /etc/openstack_deploy/user_variables.yml and that will wind you back to what you had before | 16:37 |
vakuznet | sorry for not responding, in a meeting, but i'm reading :) | 16:39 |
jrosser | or to do what johnsom and add a new config option for the endpoint you would put this in user_variables.yml http://paste.openstack.org/show/795220/ | 16:40 |
jrosser | so you've got two options there | 16:40 |
*** gshippey has quit IRC | 16:41 | |
vakuznet | i've downdraded to 20.1.1. I think the issue will surface only with self-signed certs. | 16:44 |
jrosser | vakuznet: i will have a different problem, theres no network path at all for that endpoint | 16:45 |
*** gyee has joined #openstack-ansible | 16:46 | |
*** udesale has quit IRC | 16:53 | |
johnsom | It looks like they left out both interface/endpoint_type and region_name from the keystone endpoint lookup. Sigh | 16:55 |
*** jawad_axd has quit IRC | 16:57 | |
johnsom | I have a fix I am going to propose and will poke cores to review and get it out. | 17:00 |
jrosser | johnsom: awesome, thankyou for taking a look | 17:02 |
johnsom | NP. FYI I am going track this here: https://storyboard.openstack.org/#!/story/2007863 | 17:05 |
vakuznet | cool. thanks | 17:42 |
openstackgerrit | Kevin Carter (cloudnull) proposed openstack/ansible-config_template master: Update config_template to use the builtin _get_remote_user function https://review.opendev.org/738058 | 17:52 |
openstackgerrit | Merged openstack/openstack-ansible-rabbitmq_server master: Add Centos-8 support https://review.opendev.org/735552 | 17:55 |
noonedeadpunk | jrosser: I'm wondering if didn't we break stein or smth with https://review.opendev.org/#/c/732750/1 | 18:03 |
jrosser | oh hmm | 18:03 |
jrosser | that has +2 +W on it from before we had other trouble with branchless there | 18:04 |
jrosser | uncool | 18:04 |
noonedeadpunk | eventually it can break only train | 18:05 |
noonedeadpunk | as we didn't use required-repos on stein | 18:06 |
jrosser | should we just revert? | 18:06 |
noonedeadpunk | or backport py3 bits... | 18:07 |
jrosser | but centos | 18:08 |
* jrosser end of day now | 18:08 | |
noonedeadpunk | We have everything we need to run ansible on py3 on centos, aren't we? | 18:08 |
noonedeadpunk | like libselinux module | 18:09 |
jrosser | if it's just the ansible runtime venv then maybe yes | 18:09 |
noonedeadpunk | I'm not sure, but iirc action plugins are supposed to run on ansible host... | 18:10 |
noonedeadpunk | also we probabbly will have issues soon with py2 deployment because of missing py2 pip in default repos... | 18:11 |
noonedeadpunk | for centos | 18:11 |
noonedeadpunk | actually, https://review.opendev.org/#/c/735785/ seems to be ok, so probably I just started panic too early | 18:12 |
*** spatel has quit IRC | 18:15 | |
*** spatel has joined #openstack-ansible | 18:16 | |
*** mmercer has joined #openstack-ansible | 18:17 | |
*** markvoelker has joined #openstack-ansible | 18:21 | |
vakuznet | do you mind reviewing https://review.opendev.org/#/c/737375/ ? | 18:22 |
*** Open10K8S has quit IRC | 18:24 | |
*** Open10K8S has joined #openstack-ansible | 18:25 | |
noonedeadpunk | jamesdenton: I guess there's a reply to your comment, so once you have time on your hands would be cool if you can check this out ^ | 18:26 |
jamesdenton | ahh yes, sure thing | 18:26 |
*** markvoelker has quit IRC | 18:26 | |
*** cshen has joined #openstack-ansible | 18:28 | |
*** cshen has quit IRC | 18:32 | |
*** jawad_axd has joined #openstack-ansible | 18:41 | |
*** itandops has quit IRC | 18:43 | |
openstackgerrit | Merged openstack/openstack-ansible-galera_server stable/ussuri: Use temporary repository for MariaDB 10.4.14 for Ubuntu focal https://review.opendev.org/737782 | 19:00 |
openstackgerrit | Merged openstack/openstack-ansible-galera_server master: Remove Suse tests. https://review.opendev.org/736116 | 19:00 |
openstackgerrit | Merged openstack/openstack-ansible-galera_server stable/stein: Move to PyMySQL instead of MySQLDB for ansible-2.7 https://review.opendev.org/731291 | 19:00 |
*** jawad_axd has quit IRC | 19:15 | |
*** miloa has quit IRC | 19:18 | |
openstackgerrit | Kevin Carter (cloudnull) proposed openstack/ansible-config_template master: Update config_template to use the builtin _get_remote_user function https://review.opendev.org/738058 | 19:19 |
jamesdenton | vakuznet were you able to replicate the vlan tag change via an ovs-vsctl command? | 19:24 |
jamesdenton | With br-ex:bond1,br-lbaas:bond1.202, the way the task is currently written, it would try to add the bond1.202 interface to br-lbaas and then apply the 202 tag. but bond1.202 doesn't exist | 19:25 |
jamesdenton | and because bond1 is already attached to br-ex, you couldn't use it for br-lbaas | 19:25 |
vakuznet | bond1 and bond1.202 interfaces are linux intarfacea and they are precreated | 19:26 |
jamesdenton | ahh ok | 19:26 |
jamesdenton | so if that's the case, then bond1.202 is already tagged by the OS and you don't need to tag again with ovs | 19:27 |
jamesdenton | br-ex:bond1,br-lbaas:bond1.202 would just work | 19:27 |
vakuznet | http://paste.openstack.org/show/795227/ | 19:29 |
*** markvoelker has joined #openstack-ansible | 19:29 | |
vakuznet | i think it needs to be tagged | 19:29 |
jamesdenton | it is tagged, by the OS | 19:29 |
vakuznet | let me find the error i got | 19:31 |
vakuznet | http://paste.openstack.org/show/795228/ | 19:33 |
*** markvoelker has quit IRC | 19:34 | |
jamesdenton | where's that tag directive coming from? | 19:34 |
jamesdenton | one sec | 19:34 |
*** cshen has joined #openstack-ansible | 19:37 | |
vakuznet | from https://opendev.org/openstack/openstack-ansible-os_neutron/src/branch/master/tasks/providers/ovs_config.yml#L31 | 19:38 |
jamesdenton | yes, but /usr/bin/ovs-vsctl -t 5 set port bond1.566 tag=None and the tag=None is invalid | 19:39 |
jamesdenton | http://paste.openstack.org/show/795229/ | 19:41 |
jamesdenton | i created eno3.3999 and used it for br-ex. the command does not issue the tag argument, and it's null in the args | 19:41 |
jamesdenton | *br-ex2 | 19:41 |
*** cshen has quit IRC | 19:41 | |
jamesdenton | the error you got... was that with modified playbooks? | 19:43 |
vakuznet | no, it's original | 19:43 |
jamesdenton | hmm. what OS and ansible version? | 19:44 |
vakuznet | centos7, ansible comes with osa - it's 2.8.8 | 19:44 |
jamesdenton | any chance you could remove the interfaces from your provider bridges (the bond1 and bond1.x) and rerun os-neutron-install.yml with -vvvv? | 19:53 |
jamesdenton | so we can see what it's doing for both bridges? | 19:54 |
jamesdenton | and use the vanilla role w/o the patch? | 19:54 |
chandankumar | jrosser, noonedeadpunk please have a look at this patch https://review.opendev.org/736507 when free | 19:58 |
*** dave-mccowan has quit IRC | 20:01 | |
*** dave-mccowan has joined #openstack-ansible | 20:05 | |
vakuznet | jamesdenton: i'm running the test | 20:15 |
jamesdenton | thanks! | 20:16 |
*** jawad_axd has joined #openstack-ansible | 20:32 | |
*** markvoelker has joined #openstack-ansible | 20:48 | |
*** markvoelker has quit IRC | 20:53 | |
*** jawad_axd has quit IRC | 21:06 | |
*** markvoelker has joined #openstack-ansible | 21:14 | |
*** markvoelker has quit IRC | 21:18 | |
*** spatel has quit IRC | 21:26 | |
*** cshen has joined #openstack-ansible | 21:37 | |
*** cshen has quit IRC | 21:42 | |
openstackgerrit | Merged openstack/openstack-ansible-os_horizon stable/ussuri: Compile mod-wsgi module for CentOS 7 https://review.opendev.org/737784 | 21:57 |
*** jawad_axd has joined #openstack-ansible | 22:24 | |
*** tosky has quit IRC | 22:40 | |
*** luksky has quit IRC | 22:50 | |
*** jawad_axd has quit IRC | 22:56 | |
*** markvoelker has joined #openstack-ansible | 23:15 | |
*** markvoelker has quit IRC | 23:20 | |
openstackgerrit | Merged openstack/openstack-ansible stable/ussuri: Fix inventory_manage when the component for a host is not defined https://review.opendev.org/737803 | 23:30 |
openstackgerrit | Merged openstack/openstack-ansible stable/ussuri: Correct a typo in the document https://review.opendev.org/737809 | 23:30 |
*** cshen has joined #openstack-ansible | 23:38 | |
*** cshen has quit IRC | 23:42 | |
*** mrda has quit IRC | 23:59 | |
*** gyee has quit IRC | 23:59 |
Generated by irclog2html.py 2.17.2 by Marius Gedminas - find it at https://mg.pov.lt/irclog2html/!