| *** bgmccollum has joined #openstack-ansible | 00:05 | |
| *** macza has quit IRC | 00:13 | |
| *** TxGirlGeek has quit IRC | 00:29 | |
| *** TxGirlGeek has joined #openstack-ansible | 00:40 | |
| *** ThiagoCMC has joined #openstack-ansible | 00:55 | |
| ThiagoCMC | Guys, I'm OSA with Ceph and it's from here: TASK: ceph-mon : test if initial mon keyring is in mon kv store | 00:58 |
|---|---|---|
| ThiagoCMC | The command `ceph --cluster ceph config-key get initial_mon_keyring` enters in a infinity loop... I checked with `strace` | 00:58 |
| *** nurdie has joined #openstack-ansible | 01:04 | |
| *** nurdie has quit IRC | 01:08 | |
| *** ThiagoCMC has left #openstack-ansible | 01:08 | |
| *** ThiagoCMC has joined #openstack-ansible | 01:10 | |
| openstackgerrit | Kevin Carter (cloudnull) proposed openstack/openstack-ansible-ops master: Update delegated setup hosts to support IP delegation https://review.openstack.org/632201 | 01:28 |
| cloudnull | ThiagoCMC oh that sounds all bad | 01:28 |
| cloudnull | did you get it figured out ? | 01:28 |
| ThiagoCMC | No... | 01:29 |
| ThiagoCMC | I'm seeing that the ceph.conf inside of the container, have its "mon host = IPs" pointing to the host itself, NOT to the ceph-mon container's IP! | 01:30 |
| ThiagoCMC | Might be it, right? | 01:30 |
| cloudnull | derks https://review.openstack.org/#/c/632142/ - maybe related - regarding the missing ceph role ? | 01:30 |
| cloudnull | ThiagoCMC sadly I'm garbage w/ ceph . | 01:30 |
| cloudnull | i really have no idea what the issue could be :'( | 01:31 |
| cloudnull | however if it works on one node and not another, i'd say it makes sense some config is busted. | 01:31 |
| ThiagoCMC | Hmm... ok | 01:32 |
| ThiagoCMC | I'll double check everything | 01:32 |
| ThiagoCMC | Anyway, looks like that OSA is missing a more complete doc about how to install Ceph via "openstack-ansible". | 01:33 |
| cloudnull | I know there's a bunch of docs for ceph, though, because I don't use ceph, I'm not really sure what the state of those are . | 01:34 |
| cloudnull | A lot of folks seem to deploy ceph as a stand a lone thing using ceph ansible , then integrate openstack with that. maybe that would be more successful? | 01:34 |
| cloudnull | that said, I dont really know if that would be better, as I've not done it myself. | 01:35 |
| ThiagoCMC | I would prefer to use ceph-ansible with openstack-ansible, because I'm curious about what it does, then, learn from it. | 01:35 |
| ThiagoCMC | A friend of mine that knows more about Ceph, told me that the ceph.conf must have the container's IPs... Not the IP of the host running the container! | 01:36 |
| ThiagoCMC | Not sure about how that IP ended up there... Maybe it will work if I change ceph-mon to is_metal = true | 01:36 |
| ThiagoCMC | something like that | 01:36 |
| *** nurdie has joined #openstack-ansible | 01:37 | |
| cloudnull | I'm not sure, logan- jrosser odyssey4me cjloader theintern would be the folks I'd think would know? | 01:38 |
| *** TxGirlGeek has quit IRC | 01:39 | |
| ThiagoCMC | Hey, I used "monitor_interface: br-storage", so, it WILL take the ip of that interface! LOLOL | 01:39 |
| cloudnull | that does make some sense :) | 01:40 |
| ThiagoCMC | Might be "eth1", -> br-mgmt at host (172.29.236.0/22) | 01:40 |
| ThiagoCMC | :-P | 01:40 |
| ThiagoCMC | It's only locigal! lol | 01:40 |
| ThiagoCMC | logical! ha! ^_^ | 01:40 |
| ThiagoCMC | Nop... :-/ | 01:44 |
| ThiagoCMC | Maybe "mon_group_name: ceph-mon_hosts" is wrong instead... | 01:45 |
| ThiagoCMC | cloudnull, BTW thanks for calling more ppl to help! | 01:46 |
| *** nurdie has quit IRC | 01:51 | |
| *** nurdie has joined #openstack-ansible | 01:52 | |
| ThiagoCMC | No idea about how to tell ceph-ansible (via user_variables.yml of OSA), to use the container's IPs... =\ | 01:53 |
| openstackgerrit | Kevin Carter (cloudnull) proposed openstack/openstack-ansible-ops master: Update delegated setup hosts to support IP delegation https://review.openstack.org/632201 | 01:54 |
| *** nurdie has quit IRC | 01:56 | |
| cloudnull | ThiagoCMC im just not sure how that's all wired up | 01:58 |
| cloudnull | probably someone in the morning, I suspect most folks who would know are sleeping | 01:59 |
| ThiagoCMC | I guess that I just fixed it! | 02:01 |
| ThiagoCMC | Now the error is at: "please provide lvm_volumes to your osd scenario" :-D | 02:01 |
| ThiagoCMC | lol | 02:01 |
| ThiagoCMC | Reading: http://docs.ceph.com/ceph-ansible/master/osds/scenarios.html#osd-scenario-lvm :) | 02:16 |
| ThiagoCMC | ceph-install.yml, worked! Wheee! | 02:21 |
| ThiagoCMC | cloudnull, just curious, how is the nspawn dev evolving?! I can't wait to use it! :-P | 02:37 |
| ThiagoCMC | And also the "install_method: distro" on Ubuntu LTS | 02:38 |
| openstackgerrit | Marc Gariépy (mgariepy) proposed openstack/ansible-hardening stable/rocky: Switch to rtcsync for chrony https://review.openstack.org/632907 | 03:29 |
| openstackgerrit | Marc Gariépy (mgariepy) proposed openstack/ansible-hardening stable/queens: Switch to rtcsync for chrony https://review.openstack.org/632908 | 03:29 |
| *** udesale has joined #openstack-ansible | 03:49 | |
| *** mino_ has joined #openstack-ansible | 04:07 | |
| *** spsurya has joined #openstack-ansible | 04:33 | |
| *** gyee has quit IRC | 05:01 | |
| *** udesale has quit IRC | 05:27 | |
| *** udesale has joined #openstack-ansible | 05:27 | |
| openstackgerrit | Kevin Carter (cloudnull) proposed openstack/openstack-ansible-ops master: Update delegated setup hosts to support IP delegation https://review.openstack.org/632201 | 05:28 |
| *** udesale has quit IRC | 05:29 | |
| *** udesale has joined #openstack-ansible | 05:29 | |
| *** udesale has quit IRC | 05:30 | |
| *** mkuf has quit IRC | 05:31 | |
| *** shyamb has joined #openstack-ansible | 05:36 | |
| chandankumar | odyssey4me: Hello | 06:00 |
| chandankumar | odyssey4me: http://paste.openstack.org/show/743206/ import requests still exists in virtualbox centos vm | 06:00 |
| *** shyamb has quit IRC | 06:07 | |
| *** hwoarang has quit IRC | 06:08 | |
| *** hwoarang has joined #openstack-ansible | 06:09 | |
| *** udesale has joined #openstack-ansible | 06:09 | |
| *** shyamb has joined #openstack-ansible | 06:11 | |
| *** jbadiapa has joined #openstack-ansible | 06:13 | |
| *** markvoelker has joined #openstack-ansible | 06:16 | |
| *** nurdie has joined #openstack-ansible | 06:26 | |
| *** markvoelker has quit IRC | 06:45 | |
| jrosser | chandankumar: I don’t know what’s wrong there but you could try issuing a similar virtualenv command at the cli and see what happens | 06:58 |
| jrosser | Find out if it’s something wrong generally on the host or it only goes wrong via the ansible | 06:58 |
| *** shyamb has quit IRC | 07:12 | |
| *** shyamb has joined #openstack-ansible | 07:13 | |
| *** eumel8 has joined #openstack-ansible | 07:16 | |
| *** kopecmartin|off is now known as kopecmartin | 07:19 | |
| *** mkuf has joined #openstack-ansible | 07:19 | |
| *** udesale has quit IRC | 07:22 | |
| *** markvoelker has joined #openstack-ansible | 07:42 | |
| *** DanyC has joined #openstack-ansible | 07:44 | |
| *** DanyC has quit IRC | 07:45 | |
| *** udesale has joined #openstack-ansible | 07:46 | |
| *** DanyC has joined #openstack-ansible | 07:48 | |
| *** shyamb has quit IRC | 07:48 | |
| *** shyamb has joined #openstack-ansible | 07:49 | |
| *** DanyC has quit IRC | 07:49 | |
| *** nurdie has quit IRC | 07:55 | |
| *** nurdie has joined #openstack-ansible | 07:55 | |
| *** ostackz has quit IRC | 07:58 | |
| *** shyamb has quit IRC | 07:59 | |
| *** nurdie has quit IRC | 08:00 | |
| chandankumar | jrosser: odyssey4me it is easily reproducable | 08:10 |
| chandankumar | jrosser: odyssey4me virtualenv --no-download --python=python2 /openstack/venvs/neutron-testing | 08:10 |
| *** markvoelker has quit IRC | 08:15 | |
| *** rgogunskiy has joined #openstack-ansible | 08:17 | |
| *** fnpanic has joined #openstack-ansible | 08:33 | |
| openstackgerrit | Filippo Inzaghi proposed openstack/openstack-ansible-os_keystone master: fix tox python3 overrides https://review.openstack.org/606820 | 08:34 |
| *** udesale has quit IRC | 08:37 | |
| *** udesale has joined #openstack-ansible | 08:42 | |
| *** udesale has quit IRC | 08:42 | |
| *** tosky has joined #openstack-ansible | 08:48 | |
| *** eumel8 has quit IRC | 08:49 | |
| *** shyamb has joined #openstack-ansible | 08:52 | |
| *** ianychoi has quit IRC | 08:55 | |
| *** nurdie has joined #openstack-ansible | 09:01 | |
| *** hamzaachi has joined #openstack-ansible | 09:04 | |
| *** nurdie has quit IRC | 09:05 | |
| chandankumar | jrosser: Here is the hack I applied I enabled delorean-deps.repo from rdo install virtualenv and pip from there it is working | 09:18 |
| chandankumar | working pip version is 8.0 | 09:18 |
| *** CeeMac has joined #openstack-ansible | 09:18 | |
| *** udesale has joined #openstack-ansible | 09:20 | |
| *** nurdie has joined #openstack-ansible | 09:21 | |
| jrosser | chandankumar: remind me which patch you hit this trouble with? | 09:25 |
| *** nurdie has quit IRC | 09:26 | |
| *** shardy has joined #openstack-ansible | 09:27 | |
| *** errr has quit IRC | 09:30 | |
| *** errr has joined #openstack-ansible | 09:35 | |
| chandankumar | jrosser: there is no patch involved | 09:41 |
| chandankumar | jrosser: I am trying to reproduce os_Tempest scenario tests failure | 09:41 |
| *** nurdie has joined #openstack-ansible | 09:42 | |
| chandankumar | jrosser: I think the problem is here https://github.com/openstack/openstack-ansible-tests/blob/master/run_tests_common.sh#L76 | 09:42 |
| chandankumar | jrosser: it installs pip <18 | 09:42 |
| chandankumar | but we need pip == 8.0 | 09:42 |
| jrosser | i am confused - iirc we took out the get-pip stuff becasue it was so troublesome and tried to rely entirely on the tools on the host certinaly in the openstack-ansible repo | 09:46 |
| *** nurdie has quit IRC | 09:46 | |
| jrosser | this really is odyssey4me area of expertise though | 09:46 |
| jrosser | chandankumar: it would be worth taking a read through this https://github.com/openstack/openstack-ansible/blob/master/scripts/bootstrap-ansible.sh, lots of careful setup of all these tools happens in there for the integrated builds | 09:49 |
| chandankumar | jrosser: sure | 09:50 |
| *** electrofelix has joined #openstack-ansible | 09:51 | |
| odyssey4me | oh dear, I have the feeling my replies got list - or did they? I wrote three lines in reply to chandankumar | 09:56 |
| jrosser | lost in the intertubes i think | 09:57 |
| chandankumar | odyssey4me: the vagrant steps I tried does not worked | 09:58 |
| chandankumar | it stuck with the same error | 09:58 |
| odyssey4me | chandankumar that's odd - I had a perfectly good test with my virtualbox test | 10:01 |
| odyssey4me | but yes, we should probably no longer install pip with get-pip.py any more, and rather prep the host like we do for the integrated build | 10:01 |
| odyssey4me | it's a bit messy in that tests repo though, and I was hoping to just ditch its use entirely :/ | 10:01 |
| *** nurdie has joined #openstack-ansible | 10:03 | |
| chandankumar | odyssey4me: RDO provides https://trunk.rdoproject.org/centos7-master/delorean-deps.repo to provides other dependencies which are not built by delorean | 10:04 |
| chandankumar | it is used in puppet-openstack-integration tripleo and packstack in the gates | 10:04 |
| chandankumar | http://codesearch.openstack.org/?q=delorean-deps.repo&i=nope&files=&repos= | 10:04 |
| *** hamzaachi has quit IRC | 10:05 | |
| *** nurdie has quit IRC | 10:06 | |
| odyssey4me | chandankumar ok, taking a step back - is this an issue for tripleo gates, our gates, or just for human testing? | 10:07 |
| chandankumar | odyssey4me: human testing | 10:07 |
| chandankumar | odyssey4me: the venv issue i fixed by installing python-virtualenv and pip from rpm | 10:09 |
| chandankumar | from delorean-deps repo | 10:09 |
| odyssey4me | chandankumar yeah, let me take a peek for a bit and see - I don't think it'll be an issue to convert it, I just need to spend a little time re-familiarising myself with it | 10:10 |
| *** udesale has quit IRC | 10:12 | |
| *** markvoelker has joined #openstack-ansible | 10:12 | |
| *** udesale has joined #openstack-ansible | 10:13 | |
| *** asettle has joined #openstack-ansible | 10:14 | |
| *** shyamb has quit IRC | 10:15 | |
| odyssey4me | chandankumar jrosser ah, I remember now why we couldn't switch this like we did for the integrated repo | 10:15 |
| odyssey4me | the issue is that all role tests use tox, and for human testing to get tox there we need pip on the host, and there is no native pip package for centos... for that we have to add the RDO repositories, and that's where things get complicated because we're duplicating somethign already done in the playbooks | 10:17 |
| odyssey4me | my plan all along was to get rid of our use of tox for all functional testing by switching all tests to use the integrated repo | 10:17 |
| odyssey4me | chandankumar I'm curious though - with the vagrant test you did, did you use the vagrant file in the repo - or did you use something else? | 10:18 |
| chandankumar | odyssey4me: I used the same vagrant file | 10:18 |
| chandankumar | odyssey4me: I downloaded the latest vagrant images | 10:18 |
| chandankumar | of centos | 10:18 |
| odyssey4me | very strange that it worked for me | 10:18 |
| odyssey4me | yeah, I got my virtualbox and vagrant up to date, then did http://eavesdrop.openstack.org/irclogs/%23openstack-ansible/%23openstack-ansible.2019-01-23.log.html#t2019-01-23T14:22:43 | 10:19 |
| odyssey4me | chandankumar ok, so there is a bonus here - it will only try to install pip if it's not already on the host | 10:20 |
| chandankumar | odyssey4me: yes, | 10:20 |
| chandankumar | odyssey4me: are we using openstack-ansible-pip_install role somewhere? | 10:22 |
| odyssey4me | chandankumar yes | 10:22 |
| odyssey4me | chandankumar for rocky, onwards, though we mainly use it for laying down pip.conf, not to actually install pip any more | 10:23 |
| odyssey4me | why do you ask? | 10:23 |
| *** nurdie has joined #openstack-ansible | 10:23 | |
| chandankumar | odyssey4me: I was just checking the version of pip getting installed | 10:23 |
| odyssey4me | it should be whatever version is pinned here: https://github.com/openstack/openstack-ansible/blob/master/global-requirement-pins.txt | 10:24 |
| odyssey4me | although for role tests it's plausible that it's whatever is latest instead | 10:24 |
| chandankumar | odyssey4me: https://github.com/openstack/openstack-ansible-tests/blob/master/run_tests_common.sh#L76 is installing pip 18.1 so asked | 10:24 |
| chandankumar | sorry pip <18 | 10:25 |
| odyssey4me | yeah, this is all a mess :/ | 10:25 |
| odyssey4me | chandankumar please feel free to change https://github.com/openstack/openstack-ansible-tests/blob/master/run_tests_common.sh#L73-L77 with using packages instead, I would really prefer that anyway | 10:26 |
| chandankumar | odyssey4me: sure | 10:26 |
| chandankumar | odyssey4me: I will push a patch there | 10:26 |
| odyssey4me | it should then do https://github.com/openstack/openstack-ansible/blob/master/scripts/bootstrap-ansible.sh#L132-L143 so that we're using the same pin as the integrated repo, and all distributions are using the same version of pip, setuptools, wheel | 10:27 |
| odyssey4me | otherwise things go weird | 10:27 |
| *** nurdie has quit IRC | 10:27 | |
| odyssey4me | chandankumar another method could actually be to install virtualenv, then install tox & bindep into a virtualenv and go from there - but I'm not sure how much change that wuld need to work | 10:29 |
| *** mkuf has quit IRC | 10:39 | |
| *** mkuf_ has joined #openstack-ansible | 10:39 | |
| *** nurdie has joined #openstack-ansible | 10:44 | |
| *** markvoelker has quit IRC | 10:46 | |
| *** nurdie has quit IRC | 10:48 | |
| *** mino_ has quit IRC | 11:07 | |
| *** noonedeadpunk has joined #openstack-ansible | 11:09 | |
| *** udesale has quit IRC | 11:25 | |
| *** rgogunskiy has quit IRC | 11:35 | |
| *** markvoelker has joined #openstack-ansible | 11:43 | |
| *** nurdie has joined #openstack-ansible | 11:45 | |
| *** nurdie has quit IRC | 11:49 | |
| *** shyamb has joined #openstack-ansible | 11:57 | |
| *** nurdie has joined #openstack-ansible | 12:06 | |
| *** nurdie has quit IRC | 12:10 | |
| *** markvoelker has quit IRC | 12:16 | |
| *** nurdie has joined #openstack-ansible | 12:26 | |
| *** mkuf has joined #openstack-ansible | 12:28 | |
| *** nurdie has quit IRC | 12:31 | |
| *** mkuf_ has quit IRC | 12:31 | |
| *** mkuf_ has joined #openstack-ansible | 12:32 | |
| *** mkuf has quit IRC | 12:33 | |
| *** mkuf has joined #openstack-ansible | 12:34 | |
| *** mkuf_ has quit IRC | 12:37 | |
| *** bgmccollum has quit IRC | 12:39 | |
| *** pcaruana has quit IRC | 12:40 | |
| *** ansmith has quit IRC | 12:41 | |
| *** shyamb has quit IRC | 12:44 | |
| *** nurdie has joined #openstack-ansible | 12:47 | |
| *** shyamb has joined #openstack-ansible | 12:51 | |
| *** nurdie has quit IRC | 12:51 | |
| *** udesale has joined #openstack-ansible | 13:01 | |
| *** nurdie has joined #openstack-ansible | 13:07 | |
| chandankumar | odyssey4me: I have a setup with os_Tempest instaleld buit failed at libselinux issue, | 13:08 |
| chandankumar | odyssey4me: I had made some changes in the role | 13:08 |
| chandankumar | odyssey4me: But I need to rekick the stuff for tempest | 13:09 |
| chandankumar | odyssey4me: I am not sure where the tempest playbook exists? Do I write a new playbook and run it? | 13:09 |
| chandankumar | against tempest1? | 13:09 |
| odyssey4me | chandankumar the tempest playbook is in tests/common/test-install-tempest.yml | 13:10 |
| *** nurdie has quit IRC | 13:12 | |
| *** markvoelker has joined #openstack-ansible | 13:13 | |
| *** mkuf has quit IRC | 13:16 | |
| *** hwoarang has quit IRC | 13:18 | |
| *** hwoarang has joined #openstack-ansible | 13:19 | |
| *** pcaruana has joined #openstack-ansible | 13:20 | |
| *** jrosser has quit IRC | 13:20 | |
| *** shyamb has quit IRC | 13:23 | |
| *** markvoelker has quit IRC | 13:23 | |
| *** jrosser has joined #openstack-ansible | 13:23 | |
| *** shyamb has joined #openstack-ansible | 13:23 | |
| *** markvoelker has joined #openstack-ansible | 13:23 | |
| *** mkuf has joined #openstack-ansible | 13:24 | |
| *** ygk_12345 has joined #openstack-ansible | 13:26 | |
| ygk_12345 | odyssey4me: Hi. Is there any difference in horizon versions between 18.1.1 and 18.1.2 ? | 13:26 |
| ygk_12345 | odyssey4me: especially the folder structure ? | 13:26 |
| odyssey4me | ygk_12345 probably - we bump sha's for each release | 13:27 |
| odyssey4me | ygk_12345 unlikely major changes - usually just bugfixes | 13:27 |
| ygk_12345 | odyssey4me: actually I have made a custom hack with three .py files and applied it to horizon in both version. it is only working for 18.1.2 and not 18.1.1 | 13:28 |
| *** nurdie has joined #openstack-ansible | 13:28 | |
| odyssey4me | ygk_12345 this is the horizon SHA for the next release: https://github.com/openstack/openstack-ansible/blob/stable/rocky/playbooks/defaults/repo_packages/openstack_services.yml#L75 | 13:28 |
| odyssey4me | check the SHA differences in your test env, and work out what changed in horizon between the two | 13:29 |
| jrosser | ygk_12345: you can override the variables in openstack_services.yml and point to a modified repo of your own if that helps you develop/test some changes | 13:31 |
| *** nurdie has quit IRC | 13:32 | |
| CeeMac | anyone come across an issue with dashboard not loading with "ERR_SSL_PROTOCOL_ERROR" error after stable/rocky install on bionic? | 13:33 |
| jamesdenton | mornin | 13:34 |
| CeeMac | hi | 13:35 |
| *** shyamb has quit IRC | 13:36 | |
| jamesdenton | hi CeeMac | 13:37 |
| *** ygk_12345 has quit IRC | 13:39 | |
| *** ansmith has joined #openstack-ansible | 13:43 | |
| *** shyamb has joined #openstack-ansible | 13:43 | |
| *** shyamb has quit IRC | 13:46 | |
| *** shyamb has joined #openstack-ansible | 13:47 | |
| *** nurdie has joined #openstack-ansible | 13:49 | |
| *** nurdie has quit IRC | 13:53 | |
| *** nurdie has joined #openstack-ansible | 13:56 | |
| CeeMac | jamesdenton, i don't suppose you've come across any issues with horizon dashboard not loading because of "ERR_SSL_PROTOCOL_ERROR" | 14:00 |
| openstackgerrit | Michael Vollman proposed openstack/openstack-ansible-os_manila master: Basic working os_manila role https://review.openstack.org/611930 | 14:01 |
| jamesdenton | i have not, sorry | 14:01 |
| jamesdenton | what version? | 14:02 |
| jamesdenton | oh n/m | 14:02 |
| CeeMac | stable/rocky on bionic | 14:02 |
| CeeMac | if i use curl on the command line I dont get any errors, but then I don't get any response either, I'm not curl expert though | 14:02 |
| jamesdenton | do you see any errors in the logs in horizon container? | 14:03 |
| CeeMac | cant' load the page from IE or Chrome on remote (routed) network. Just about to attach a OC to the same lan | 14:03 |
| openstackgerrit | Andy Smith proposed openstack/openstack-ansible master: Add qdrouterd role for rpc messaging backend deployment https://review.openstack.org/624184 | 14:03 |
| CeeMac | good question | 14:03 |
| CeeMac | let me look again | 14:03 |
| *** strattao has joined #openstack-ansible | 14:12 | |
| CeeMac | jamesdenton, in the other_vhosts_access.log i see this http://paste.openstack.org/show/743224/ | 14:13 |
| CeeMac | nothing in access log of note | 14:14 |
| *** mgariepy has joined #openstack-ansible | 14:15 | |
| CeeMac | nothing in journal of note | 14:15 |
| CeeMac | if i try and connect directly to the container IP on port 80, it redirects to https, with /auth/login/?next=/ then throws the same error | 14:19 |
| jamesdenton | ok - are the backends up for haproxy? | 14:20 |
| jamesdenton | hatop should be able to reveal that | 14:20 |
| CeeMac | they were, let me double check as I've restared the container | 14:21 |
| CeeMac | yep | 14:21 |
| CeeMac | front-ends open, back-end is up | 14:22 |
| CeeMac | incidentally, it has the same issue going through the LB address and direct to container IP | 14:22 |
| CeeMac | last time I saw something like this it was a ssl/tls issue | 14:22 |
| jamesdenton | so when you curl https://<container>/ you get an error too? | 14:23 |
| CeeMac | i get nothing i think | 14:23 |
| CeeMac | i'll try again | 14:23 |
| CeeMac | i did this time http://paste.openstack.org/show/743226/ | 14:24 |
| CeeMac | think i tested against LB address originally | 14:24 |
| jamesdenton | Hmm. did you deviate from the deploy guide at all? meaning, did you have overrides related to SSL/certs or anything? | 14:25 |
| jrosser | it would be a surprise if you could https://<container>/ for a couple of reasons, isnt the ssl termination done at the LB | 14:26 |
| jrosser | and if you can reach the containers from the "outside" then I think there are bigger problems :) | 14:26 |
| CeeMac | jamesdenton, i did not | 14:27 |
| CeeMac | jrosser, its the 'internal' network i'm connecting to from a routed management network on the 'inside' | 14:27 |
| jrosser | CeeMac: I have a rocky/bionic AIO running right now, so i'll be able to see how that turns out later when it's cooked | 14:28 |
| CeeMac | if i http://<container>/ it performs a redirect to https://<container>/auth/login/?next=/ | 14:28 |
| jamesdenton | CeeMac you might also try hitting the horizon container on port 80 with 'X-Forwarded-Proto: https' and see if it doesn't redirect you | 14:28 |
| *** nurdie has quit IRC | 14:28 | |
| *** nurdie has joined #openstack-ansible | 14:29 | |
| jamesdenton | haproxy frontend should offload ssl and send that header to the backend to avoid the redirect. default connection to the backend is 80 | 14:29 |
| CeeMac | then I am confused. Saying that, the logs do show the connection on port 80 | 14:32 |
| CeeMac | jamesdenton, how do i do the thing with the header? | 14:33 |
| *** nurdie has quit IRC | 14:34 | |
| jamesdenton | Example to my horizon container directly: curl -v http://172.29.236.167/auth/login/ -H 'X-Forwarded-Proto: https' | 14:34 |
| nowster | jamesdenton: Do you mind if I pick your brains about networking? | 14:34 |
| CeeMac | ah, ok, one sec | 14:34 |
| jamesdenton | CeeMac This is to the VIP: curl -k -v https://172.29.236.100/auth/login/ | 14:34 |
| jamesdenton | In both cases, I got a 200 OK and a page | 14:35 |
| CeeMac | jamesdenton, that returns xml | 14:35 |
| jamesdenton | that's html | 14:35 |
| jamesdenton | or should be, anyway | 14:35 |
| jamesdenton | nowster what's up? | 14:36 |
| *** shyamb has quit IRC | 14:36 | |
| CeeMac | sorry, it is html | 14:36 |
| CeeMac | so, interesting. If I use curl -k -v https://<internal-vip> then it returns valid code HTTP/1.1 200 OK | 14:37 |
| jamesdenton | CeeMac If the curl is working to the VIP and the container directly using those examples (with your IPs), then I think the OSA side is OK. Maybe the browser doesn't like the cert? | 14:37 |
| CeeMac | thats from my deployment node | 14:38 |
| jamesdenton | k | 14:38 |
| CeeMac | i'll try hooking up another VM to the network with a different browser | 14:38 |
| nowster | jamesdenton: All the documentation assumes you're able to allocate IPs from the provider network interface. I'm trying to set up a routed network which exists on a VXLAN, but with the neutron host routing public traffic out to the upstream router. (For extra difficulty, I'm trying to do this with IPv6.) | 14:38 |
| jamesdenton | Can you elaborate on this? >> All the documentation assumes you're able to allocate IPs from the provider network interface | 14:39 |
| nowster | eg. https://docs.openstack.org/newton/install-guide-ubuntu/launch-instance-networks-provider.html | 14:39 |
| nowster | In my case the provider gives you a single IP, but allows net blocks to be routed to it. | 14:40 |
| jamesdenton | Oh i see. | 14:41 |
| jamesdenton | So in that circumstance, normally i would expect some kind of hardware router or something that can take that IP on its WAN interface and have multiple LAN networks, any of which could be those routed nets | 14:42 |
| nowster | we don't have that | 14:42 |
| jamesdenton | The normal Neutron router scenario does not really support that model | 14:42 |
| nowster | Would using OVS work better? | 14:43 |
| *** fnpanic has quit IRC | 14:43 | |
| jamesdenton | No, it's not really an OVS vs LB issue or anything like that. You need some sort of demarc where you can configure that address and provide routing for those add'l networks | 14:44 |
| jrosser | this sounds like the IP on the host is the 'next hop' for this netblock then | 14:45 |
| jamesdenton | it's possible, i suppose, to have a neutron "provider" network with a subnet consisting of a single IP, and then you can configure the add'l networks to hang off that router. But it's sorta limiting | 14:45 |
| jamesdenton | very possible. while it can be made to work, i'm not entirely sure how to make it work within the bounds of the "reference arch", if you know what i mean | 14:46 |
| nowster | How about setting up a virtual bridge network with that address range, then having the kernel know how to route for it? | 14:47 |
| jamesdenton | is this just for a PoC? prod? | 14:47 |
| CeeMac | jamesdenton, must be an internal routing issue, i've just tested access to the external LB VIP and that seems to be working fine | 14:47 |
| CeeMac | thanks anyway :) | 14:47 |
| jamesdenton | glad you eliminated OSA, anyway. | 14:47 |
| jamesdenton | good luck | 14:47 |
| CeeMac | thanks | 14:48 |
| nowster | It would eventually be production config. It appears that my colleague's config uses kolla and OVS. | 14:48 |
| *** nurdie has joined #openstack-ansible | 14:50 | |
| *** mgariepy has quit IRC | 14:50 | |
| jamesdenton | Well, whatever that IP is configured on would be a SPOF unless you also implemented something like keepalived between the hosts, i would think | 14:51 |
| nowster | That's a later concern. ;-) | 14:52 |
| jamesdenton | And the vxlan between the hosts is part of this, too? A non-neutron vxlan mesh? | 14:52 |
| nowster | forget the vxlan | 14:52 |
| jamesdenton | well, how are the hosts connecting to one another for the various networks (br-mgmt, br-vlan, storage, etc)? | 14:53 |
| nowster | those exist | 14:53 |
| nowster | except we can't do br-vlan, because the provider filters tagged traffic | 14:53 |
| jamesdenton | Did they give you vlans for those networks? | 14:56 |
| jamesdenton | or actual interfaces per network (untagged)? | 14:57 |
| nowster | vlans are allowed, but provider assigned (max 12 per project) | 14:57 |
| jamesdenton | k | 14:57 |
| nowster | I'm thinking to bring up a "provider" vlan, and have that be "flat" type. | 14:58 |
| *** nurdie has quit IRC | 14:59 | |
| jrosser | that sounds like a plan - you need to leave the door open to adding another controller | 15:01 |
| nowster | yep | 15:02 |
| nowster | Off to try to set that up... | 15:02 |
| jamesdenton | You ought to be able to configure infra01 as the 'gateway' for one or more of those routed networks by configuring, say, x.x.x.1, on br-vlan. As long as the other hosts have connectivity to infra01 via whatever provider interface you've chosen, it should work (as a flat net) | 15:02 |
| nowster | yep | 15:04 |
| nowster | I think I'll call it br-provider to avoid confusion. | 15:04 |
| jamesdenton | good idea | 15:05 |
| *** hwoarang has quit IRC | 15:07 | |
| *** hwoarang has joined #openstack-ansible | 15:08 | |
| *** strattao has quit IRC | 15:10 | |
| *** ztr has joined #openstack-ansible | 15:10 | |
| nowster | jamesdenton: I notice that sysctl net.ipv4.conf.*.forwarding is all 1, but the ipv6 equivalents are all 0. Should I be enabling them manually, or will neutron do that for me? | 15:16 |
| jamesdenton | It might do it for you, but there's no harm in doing it, too | 15:16 |
| nowster | ta | 15:18 |
| *** kukacz_ has quit IRC | 15:20 | |
| CeeMac | hmm | 15:21 |
| CeeMac | looks like i might have spoken too soon | 15:21 |
| openstackgerrit | James Denton proposed openstack/openstack-ansible-os_neutron master: [WIP] Deploy Vector Packet Processing (VPP) Platform for Neutron https://review.openstack.org/631644 | 15:22 |
| CeeMac | ok, so if i try and connect to the internal VIP from a machine on the same network, I get ERR_CONNECTION_REFUSED | 15:31 |
| *** kukacz has joined #openstack-ansible | 15:31 | |
| *** nurdie has joined #openstack-ansible | 15:31 | |
| CeeMac | If I connect in through the external VIP I can log in, but get little popup errors for things not being able to be retrieved | 15:31 |
| *** pcaruana has quit IRC | 15:35 | |
| *** kukacz has quit IRC | 15:42 | |
| *** ianychoi has joined #openstack-ansible | 15:44 | |
| *** kukacz has joined #openstack-ansible | 15:46 | |
| openstackgerrit | Frank Kloeker proposed openstack/openstack-ansible master: Change docs build to openstackdocstheme logic https://review.openstack.org/623429 | 15:46 |
| *** udesale has quit IRC | 15:53 | |
| CeeMac | ok, so i've now got an ubunto desktop VM connecting on the br-mgmt network, it can log in to the internal VIP but also gets Error: unable to retrieve the <stuff> messages | 15:53 |
| *** udesale has joined #openstack-ansible | 15:54 | |
| CeeMac | any horizon experts in the house? | 15:54 |
| openstackgerrit | James Denton proposed openstack/openstack-ansible-os_neutron master: [WIP] Deploy Vector Packet Processing (VPP) Platform for Neutron https://review.openstack.org/631644 | 15:54 |
| *** pcaruana has joined #openstack-ansible | 15:55 | |
| jamesdenton | CeeMac Do those errors only occur on certain tabs? | 15:55 |
| *** kukacz has quit IRC | 15:58 | |
| *** TxGirlGeek has joined #openstack-ansible | 15:58 | |
| CeeMac | jamesdenton, I'm logged in as the 'admin' user, and it defaults to the 'service' project. If I switch to the 'admin' project and browse through the project | Compute tabs, every one throws an error | 15:59 |
| *** kukacz has joined #openstack-ansible | 16:00 | |
| CeeMac | same for volumes | 16:00 |
| CeeMac | network topology threw a "Something went wrong!" page | 16:00 |
| jamesdenton | You know, it may be worth re-running setup-openstack playbook just incase something went wonky along the way | 16:01 |
| CeeMac | maybe. i'll refresh my snapshots and do that. Can't harm | 16:02 |
| *** gyee has joined #openstack-ansible | 16:06 | |
| CeeMac | something definitely looks wonky. openstack cli isn't working just now either | 16:11 |
| CeeMac | odd | 16:12 |
| CeeMac | getting an error about role os_blazar not being found | 16:12 |
| CeeMac | i'm just going to comment that out for the moment | 16:14 |
| *** gyee has quit IRC | 16:14 | |
| ztr | chandankumar: I'm told that you are working on an issue affecting openstack-ansible-os_tempest on centOS | 16:15 |
| CeeMac | ah, ok, cloudnull had me check out a patch for the remote-journal issue i was having | 16:17 |
| CeeMac | so i'm in that not in stable/rocky | 16:17 |
| CeeMac | if i checkout stable/rocky will i lose the patch? | 16:17 |
| jamesdenton | more than likely, yes | 16:17 |
| CeeMac | i'm a total git n00b, sorry | 16:17 |
| prometheanfire | someone mind testing https://review.openstack.org/#/c/608393/44/tasks/lxc_pre_install.yml | 16:18 |
| CeeMac | so if i check out stable/rocky | 16:18 |
| prometheanfire | via command line I get the correct thing, but via ansible it is truncating the url | 16:18 |
| CeeMac | is there an alternate way I can pulldown the patched playbook for this run so don't undo the fix? | 16:18 |
| *** eumel8 has joined #openstack-ansible | 16:19 | |
| jamesdenton | CeeMac I'll have to defer, sorry. | 16:20 |
| eumel8 | o/ | 16:21 |
| chandankumar | ztr: yes | 16:21 |
| ztr | Is there a bug for this issue? | 16:21 |
| CeeMac | no problem, i'll do this anyway because i need the dashboard to work, then I'll catch cloudnull to see if i can get a cherry pick of his fix instead for afterwards | 16:22 |
| chandankumar | ztr: if you are free, feel free to take a look at this issue | 16:22 |
| chandankumar | ztr: http://logs.openstack.org/08/631708/4/check/openstack-ansible-functional-centos-7/cc0ab67/logs/openstack/tempest1/stestr_results.html | 16:22 |
| chandankumar | ztr: it might be related to network cidr issue | 16:22 |
| *** kukacz has quit IRC | 16:24 | |
| *** gyee has joined #openstack-ansible | 16:25 | |
| eumel8 | we have now OSA docs with the new openstackdocstheme release 1.29 in the pipeline. sneak preview: http://logs.openstack.org/29/623429/3/check/openstack-tox-docs/0795a7f/html/de/index.html | 16:25 |
| eumel8 | new is that language menu on each page after sphinx build with the script in openstackdocstheme repo | 16:26 |
| *** kukacz has joined #openstack-ansible | 16:26 | |
| ztr | chandankumar: I think that is the same failure that is blocking me, if it is I'll file a bug for tracking purposes | 16:27 |
| chandankumar | ztr: sure | 16:31 |
| *** macza has joined #openstack-ansible | 16:33 | |
| FrankZhang | anyone has idea why centOS gating consistently failed at tempest 'test_server_basic_ops'? 2019-01-23 18:42:09,306 4664 WARNING [tempest.lib.common.ssh] Failed to establish authenticated ssh connection to cirros@10.1.3.24 (timed out). | 16:33 |
| chandankumar | FrankZhang: We are still debugging why it is happening | 16:34 |
| FrankZhang | chandankumar: thanks, can you share the card we're tracking it? | 16:34 |
| chandankumar | ztr: is creating a ticket for that | 16:35 |
| FrankZhang | Cool thank you | 16:36 |
| *** pcaruana has quit IRC | 16:48 | |
| *** asettle has quit IRC | 16:50 | |
| odyssey4me | eumel8 lol @ english vs english(uk) | 16:51 |
| odyssey4me | But that's awesome. :) | 16:51 |
| ztr | chandankumar: https://bugs.launchpad.net/openstack-ansible/+bug/1813187 | 16:52 |
| openstack | Launchpad bug 1813187 in openstack-ansible "CentOS tempest test_server_basic_ops failure" [Undecided,New] | 16:52 |
| eumel8 | odyssey4me: I'm not the super specialist for such kinds of differents ;) | 16:52 |
| odyssey4me | chandankumar was cloudnull trying to help you isolate the issue there? | 16:52 |
| *** antonym has quit IRC | 16:53 | |
| CeeMac | looks like i done broke it | 16:57 |
| CeeMac | http://paste.openstack.org/show/743236/ | 16:57 |
| CeeMac | any help would be appreciated | 16:57 |
| *** eumel8 has quit IRC | 16:57 | |
| CeeMac | all front/back open/up in haproxy | 16:57 |
| CeeMac | openstack cli wont connect now, dashboard won't work | 16:58 |
| *** strattao has joined #openstack-ansible | 16:58 | |
| *** bgmccollum has joined #openstack-ansible | 17:00 | |
| odyssey4me | CeeMac what exactly was done that resulted in requiring os_blazar? | 17:08 |
| *** antonym has joined #openstack-ansible | 17:11 | |
| *** kopecmartin is now known as kopecmartin|off | 17:18 | |
| CeeMac | odyssey4me, i think that came down from the git checkout cloudnull had me do for his journal patch | 17:22 |
| odyssey4me | CeeMac what exactly did you do, because that patch doesn't include any changes that would makr blazar required | 17:22 |
| CeeMac | odyssey4me, git fetch https://git.openstack.org/openstack/openstack-ansible refs/changes/05/632505/2 && git checkout FETCH_HEAD | 17:22 |
| odyssey4me | right, that's why - you did the 'check out' option, not the 'cherry-pick' option | 17:23 |
| odyssey4me | so basically you changed from rocky to master | 17:23 |
| CeeMac | *shrugs* i just did what I was told :/ | 17:23 |
| prometheanfire | machinectl import-image is timing out, it seems to have completed | 17:23 |
| prometheanfire | Jan 24 17:05:37 os-infra-test systemd-importd[29126]: (transfer1) Operation completed successfully. | 17:23 |
| CeeMac | so, i've checked out stable/rocky again | 17:23 |
| prometheanfire | Jan 24 17:05:37 os-infra-test systemd-importd[29126]: (transfer1) Exiting. | 17:23 |
| odyssey4me | CeeMac from the patch in question - https://review.openstack.org/#/c/632505/ | 17:24 |
| odyssey4me | download, o the top right | 17:24 |
| odyssey4me | click on the clip-board next to 'cherry-pick' - not checkout | 17:24 |
| CeeMac | ah, gotcha | 17:24 |
| CeeMac | ok, i'l make a note of that for when things are working again | 17:24 |
| CeeMac | right now something is hinky and dashboard/keystone don't seem to work | 17:25 |
| odyssey4me | CeeMac well, if you're in some weird mix of rocky and stein then it's not surprising | 17:25 |
| FrankZhang | chandankumar: it seems centOS gating error became github accessing failure http://logs.openstack.org/19/617719/2/check/openstack-ansible-functional-centos-7/332b944/job-output.txt.gz#_2019-01-24_17_15_17_915497 after you recheck https://review.openstack.org/#/c/617719/ | 17:25 |
| CeeMac | i only ran the infra-journal-remote.yml playbook from that checkout | 17:26 |
| CeeMac | and openstack cli has been working today since then | 17:26 |
| CeeMac | just appears to have stopped recently | 17:26 |
| CeeMac | while I've been trying to troubleshoot dashboard access | 17:26 |
| odyssey4me | CeeMac that's always nice :/ | 17:27 |
| CeeMac | yeah | 17:27 |
| CeeMac | the only thing I can remember doing is restart apache2 in the horizon container, then restart the horizon container | 17:27 |
| CeeMac | and I'm sure it was working after that | 17:27 |
| CeeMac | then it wasn't | 17:27 |
| CeeMac | did you see the error I pasted? | 17:28 |
| prometheanfire | odyssey4me: cloudnull (I think both of you worked on the machinectl stuff most), import-tar is timing out because it expects uesr input... | 17:34 |
| odyssey4me | odyssey4me Is that for a new patch or what, because it hasn't required input before? Also, I did no work on anything relating to machinectl except to fix some ansible bugs here and there. | 17:35 |
| ThiagoCMC | Hey guys, I just installed OSA Rocky on Ubuntu, everything looks good, except Heat API, there is an error on heat.log: | 17:36 |
| ThiagoCMC | ERROR oslo.messaging._drivers.impl_rabbit [-] Unable to connect to AMQP server on 172.29.238.12:5671 after None tries: Connection.open: (530) NOT_ALLOWED - access to vhost '/heat' refused for user 'heat': NotAllowed: Connection.open: (530) NOT_ALLOWED - access to vhost '/heat' refused for user 'heat' | 17:36 |
| ThiagoCMC | Any idea? | 17:36 |
| prometheanfire | odyssey4me: ah, ok, not for a new patch | 17:37 |
| prometheanfire | :q | 17:37 |
| prometheanfire | found the error though (it's in systemd itself, and fixed already) | 17:40 |
| *** Darcidride has joined #openstack-ansible | 17:48 | |
| *** udesale has quit IRC | 18:01 | |
| openstackgerrit | Kevin Carter (cloudnull) proposed openstack/openstack-ansible-os_tempest master: Adds tempest run command with --test-list option https://review.openstack.org/631351 | 18:03 |
| chandankumar | cloudnull: once we have env ready using run-test.sh functional | 18:12 |
| chandankumar | when i sourced .tox/functional env | 18:13 |
| chandankumar | cloudnull: ansible-playbook does not able to find hosts | 18:13 |
| *** Darcidride has quit IRC | 18:17 | |
| *** Darcidride has joined #openstack-ansible | 18:18 | |
| odyssey4me | chandankumar run_tests sets out some env vars and also provide CLI options which you need to add to make that work | 18:21 |
| odyssey4me | IIRC you need to set TESTING_DIR, and you need to set where the inventory is, and you need to supply the override file | 18:22 |
| *** Darcidride has quit IRC | 18:26 | |
| chandankumar | odyssey4me: got it | 18:26 |
| *** Darcidride has joined #openstack-ansible | 18:26 | |
| *** Darcidride has quit IRC | 18:28 | |
| chandankumar | odyssey4me: it worked thanks, https://github.com/openstack/openstack-ansible-tests/blob/master/test-ansible-syntax.sh#L31 | 18:28 |
| chandankumar | i ecported first 3 vars | 18:29 |
| *** Darcidride has joined #openstack-ansible | 18:29 | |
| chandankumar | run ansible playbook | 18:29 |
| odyssey4me | chandankumar ok, then it should be something like: ./tox/functional/bin/ansible-playbook -i tests/inventory -e@tests/????_overrides.yml tests/test.yml | 18:30 |
| odyssey4me | oh, with ANSIBLE_INVENTORY set, you don't need the -i | 18:30 |
| *** Darcidride has quit IRC | 18:31 | |
| *** Darcidride has joined #openstack-ansible | 18:31 | |
| *** Darcidride has quit IRC | 18:34 | |
| *** Darcidride has joined #openstack-ansible | 18:34 | |
| *** shardy has quit IRC | 18:43 | |
| *** TxGirlGe_ has joined #openstack-ansible | 18:59 | |
| *** TxGirlGeek has quit IRC | 19:00 | |
| *** ztr has quit IRC | 19:01 | |
| *** nurdie has quit IRC | 19:02 | |
| *** nurdie has joined #openstack-ansible | 19:02 | |
| chandankumar | odyssey4me: there is a bug here http://git.openstack.org/cgit/openstack/openstack-ansible-tests/tree/test-vars.yml#n449 | 19:04 |
| odyssey4me | chandankumar how so? | 19:04 |
| chandankumar | odyssey4me: when we use run run_tests.sh it create this file in /root/src | 19:04 |
| chandankumar | not in development directoru | 19:05 |
| odyssey4me | that will only be if TESTING_HOME=/root | 19:05 |
| *** nurdie has quit IRC | 19:06 | |
| chandankumar | odyssey4me: yes correct, I exported Testing home it worked | 19:07 |
| broken_one | hey guys, you happen to have an legalese for contributors and commiters? we have reach the sharks and they want to look at docs | 19:08 |
| chandankumar | odyssey4me: I think we need a additonal script where we export folder so that one can run a playbook | 19:08 |
| chandankumar | odyssey4me: I will put up these things tomorrw in docs | 19:09 |
| odyssey4me | chandankumar I'll wait for those docs, or an etherpad or something. I'm not sure what you mean, but it can wait for tomorrow. | 19:09 |
| chandankumar | odyssey4me: sure | 19:09 |
| odyssey4me | broken_one I think you might mean the Contributor License Agreement? | 19:09 |
| odyssey4me | broken_one https://review.openstack.org/static/cla.html | 19:11 |
| odyssey4me | broken_one note that https://docs.openstack.org/contributors/common/introduction.html has multiple ways of contributing, and they don't all ask for the CLA to be agreed to | 19:11 |
| broken_one | odyssey4me: thank you | 19:14 |
| broken_one | does the CLA also cover corps? | 19:20 |
| broken_one | or are we working as individuals within a corp? | 19:21 |
| odyssey4me | broken_one oh, I dunno - it's perhaps best to ask in #openstack-dev, which has a broader set of people watching | 19:21 |
| *** nurdie has joined #openstack-ansible | 19:22 | |
| broken_one | will do. thank you again | 19:23 |
| *** noonedeadpunk has quit IRC | 19:24 | |
| openstackgerrit | Matthew Thode proposed openstack/openstack-ansible-lxc_hosts master: add gentoo support https://review.openstack.org/608393 | 19:28 |
| ThiagoCMC | Guys, I'm trying to upload a raw Ubuntu image to Glance+Ceph, like this: `openstack image create --disk-format raw --container-format bare --public --file ubuntu-18.04-server-cloudimg-amd64.raw ubuntu1804` - But there is an error on glance-api.log: "since image size is zero we will be doing resize-before-write for each chunk which will be considerably slower than normal", any tips? =P | 19:29 |
| ThiagoCMC | I converted the ubuntu "img" to "raw" using `qemu-img convert`... | 19:30 |
| ThiagoCMC | `openstack image list` is empty | 19:30 |
| *** TxGirlGe_ has quit IRC | 19:32 | |
| prometheanfire | cloudnull: looks like networkd isn't used within containers? | 19:33 |
| prometheanfire | lxc_container_create that is | 19:33 |
| *** Darcidride has quit IRC | 19:40 | |
| *** tosky has quit IRC | 19:44 | |
| *** Darcidride has joined #openstack-ansible | 19:54 | |
| *** strattao has quit IRC | 20:02 | |
| *** strattao has joined #openstack-ansible | 20:06 | |
| *** TxGirlGeek has joined #openstack-ansible | 20:26 | |
| ThiagoCMC | Some ceph vars where wrong on cinder config (rbd_secret_uuid and the user)... I'm re-deploying everything from scratch again, including Ceph (all under MaaS). | 20:34 |
| *** bgmccollum has quit IRC | 21:02 | |
| *** Darcidride has quit IRC | 21:10 | |
| *** hamzaachi has joined #openstack-ansible | 21:12 | |
| *** bgmccollum has joined #openstack-ansible | 21:24 | |
| *** ansmith has quit IRC | 21:28 | |
| *** hamzaachi has quit IRC | 21:31 | |
| *** cmart has joined #openstack-ansible | 22:15 | |
| cmart | Good afternoon OSA | 22:16 |
| *** cmart has left #openstack-ansible | 22:17 | |
| *** cmart has joined #openstack-ansible | 22:17 | |
| cmart | I'm upgrading from Queens to Rocky, and the Rocky upgrade process is installing things in containers that should have been deprecated / removed in Queens, for example the cinder_scheduler_container and nova_api_os_compute container. If I destroy these containers and remove them from inventory, then the upgrade process re-creates them. | 22:17 |
| cmart | Any idea what may be going on? | 22:17 |
| *** ansmith has joined #openstack-ansible | 22:18 | |
| *** nurdie has quit IRC | 22:18 | |
| *** nurdie has joined #openstack-ansible | 22:19 | |
| *** nurdie has quit IRC | 22:23 | |
| jrosser | cmart: looks like cinder_scheduler_container should have been removed for going from P->Q, which is probably why Q->R upgrade tooling doesnt touch it | 22:27 |
| cmart | jrosser yeah :( | 22:27 |
| cmart | I try to remove them from my inventory JSON manually, but apparently the inventory script is adding them back | 22:27 |
| jrosser | are they defined in /etc/openstack_deploy/env.d at all? | 22:28 |
| ThiagoCMC | Quick question: Why there is a btrfs volume mounted under /var/lib/machines (a loop0 dev) if the OSA containers are being stored under /var/lib/lxc ? | 22:29 |
| jrosser | ThiagoCMC: that contains the container base image | 22:29 |
| cmart | jrosser I see this sort of thing in my /etc/openstack_deploy/env.d/cinder.yml: http://paste.openstack.org/show/743433/ | 22:30 |
| cmart | so cinder_scheduler_container still "contains" "cinder_scheduler". Does that control how the inventory JSON gets built? | 22:31 |
| jrosser | ThiagoCMC: you have many choices of lxc storage backend, dir/overlayfs/btrfs/zfs so depeding on which you use slightly different things happen to the base image | 22:31 |
| jrosser | cmart: yeah, you should see that map to a physical host group there, which you'll then have defined in your openstack_user_config | 22:32 |
| jrosser | that is deciding which containers get mapped to which hosts | 22:32 |
| jrosser | so by fiddling with the env.d contents you can add/delete which containers should exist | 22:33 |
| cmart | OK cool. so in that case my `/etc/openstack_deploy/env.d`is still defining stuff that should have been deprecated in Queens. | 22:33 |
| cmart | I don't think I've changed anything in there manually, so maybe i'll just replace the contents of `/etc/openstack_deploy/env.d` with what comes stock with OSA, and try again. | 22:33 |
| cmart | thanks very much for the breadcrumb jrosser :) | 22:34 |
| openstackgerrit | Merged openstack/openstack-ansible-ops master: Update delegated setup hosts to support IP delegation https://review.openstack.org/632201 | 22:34 |
| jrosser | there is stuff in the upgrade tooling which looks for duplicate /etc/openstack_deploy/env.d/.... and /opt/openstack-ansible/inventory/env.d/..... and tries to remove anything from /etc/.... that isnt changed from the dfault | 22:35 |
| jrosser | however as you seem to have a P version of that file it will be different from the R version in /opt/openstack-ansible so will not be touched | 22:35 |
| jrosser | once you have resolved what is going in in /etc/openstack_deploy/env.d use scripts/inventory_manage.py to delete the inventory elements that you do not need any more | 22:36 |
| *** bgmccollum has quit IRC | 22:36 | |
| cmart | OH this is super helpful. "/opt/openstack-ansible/inventory/env.d/cinder.yml" looks like what I want. | 22:37 |
| jrosser | and take a backup first :) | 22:37 |
| ThiagoCMC | jrosser, where can I check if I have too many choices of str back? Also, is it possible to store / run the OSA LXC containers only from btrfs? | 22:37 |
| cmart | jrosser of course. already have 'em | 22:37 |
| jrosser | ThiagoCMC: you can see the combinations of backing store that get tested for the lxc_hosts role here https://review.openstack.org/#/c/632118/ | 22:40 |
| jrosser | ThiagoCMC: but probably stick with the default unless you have a good reason to change it? | 22:41 |
| *** bgmccollum has joined #openstack-ansible | 22:46 | |
| ThiagoCMC | jrosser, interesting... I didn't knew that this (not run the containers from within the btrfs) is the default, I thought that I was doing something wrong. lol | 22:49 |
| *** spsurya has quit IRC | 23:04 | |
| ThiagoCMC | Just curious, will OSA have LXD support for its containers, someday? :-P | 23:10 |
| *** TxGirlGeek has quit IRC | 23:15 | |
Generated by irclog2html.py 2.15.3 by Marius Gedminas - find it at mg.pov.lt!