| openstackgerrit | Kevin Carter proposed openstack/openstack-ansible: adds the config_template to nova https://review.openstack.org/223329 | 00:03 |
|---|---|---|
| cloudnull | im out have a good night all | 00:13 |
| coolj | later cloudnull | 00:28 |
| *** abitha has quit IRC | 00:47 | |
| openstackgerrit | Ian Cordasco proposed openstack/openstack-ansible: Make our usage of with_items/with_nested consistent https://review.openstack.org/223365 | 01:22 |
| sigmavirus24 | cloudnull: odyssey4me ^ | 01:29 |
| sigmavirus24 | haven't run it on a lab/aio but we noticed it today with the swift stuff so I went through and audited all ~150 uses of with_items/with_nested | 01:29 |
| sigmavirus24 | I'm tired though so I'll let the gate bash its head against that for now | 01:29 |
| *** sigmavirus24 is now known as sigmavirus24_awa | 01:31 | |
| *** woodard has joined #openstack-ansible | 02:11 | |
| *** woodard has quit IRC | 02:13 | |
| openstackgerrit | Kevin Carter proposed openstack/openstack-ansible: adds the config_template to tempest https://review.openstack.org/223342 | 02:14 |
| *** markvoelker has quit IRC | 03:15 | |
| *** sigmavirus24_awa has quit IRC | 03:27 | |
| *** palendae has quit IRC | 03:27 | |
| *** metral is now known as metral_zzz | 03:27 | |
| *** eglute has quit IRC | 03:27 | |
| *** lbragstad has quit IRC | 03:27 | |
| *** daneyon_ has joined #openstack-ansible | 03:27 | |
| *** metral_zzz is now known as metral | 03:27 | |
| *** bgmccollum has quit IRC | 03:27 | |
| *** spotz_zzz has quit IRC | 03:27 | |
| *** mgagne has quit IRC | 03:28 | |
| *** dolphm has quit IRC | 03:28 | |
| *** Guest60363 has quit IRC | 03:28 | |
| *** metral is now known as metral_zzz | 03:28 | |
| *** daneyon has quit IRC | 03:28 | |
| *** neillc has quit IRC | 03:28 | |
| *** jroll has quit IRC | 03:28 | |
| *** darrenc_ has joined #openstack-ansible | 03:29 | |
| *** gus has quit IRC | 03:29 | |
| *** mattoliverau has quit IRC | 03:29 | |
| *** persia has quit IRC | 03:29 | |
| *** darrenc has quit IRC | 03:29 | |
| *** mattoliverau has joined #openstack-ansible | 03:29 | |
| *** bgmccollum has joined #openstack-ansible | 03:30 | |
| *** spotz_zzz has joined #openstack-ansible | 03:30 | |
| *** persia has joined #openstack-ansible | 03:30 | |
| *** persia has quit IRC | 03:30 | |
| *** persia has joined #openstack-ansible | 03:30 | |
| *** neillc has joined #openstack-ansible | 03:30 | |
| *** eglute has joined #openstack-ansible | 03:30 | |
| *** dolphm has joined #openstack-ansible | 03:30 | |
| *** lbragstad has joined #openstack-ansible | 03:30 | |
| *** palendae has joined #openstack-ansible | 03:31 | |
| *** blewis has joined #openstack-ansible | 03:31 | |
| *** blewis is now known as Guest82354 | 03:31 | |
| *** gus has joined #openstack-ansible | 03:31 | |
| *** mgagne has joined #openstack-ansible | 03:31 | |
| *** jroll has joined #openstack-ansible | 03:31 | |
| *** sigmavirus24_awa has joined #openstack-ansible | 03:32 | |
| *** Guest82354 has quit IRC | 03:39 | |
| *** eglute has quit IRC | 03:39 | |
| *** lbragstad has quit IRC | 03:39 | |
| *** bgmccollum has quit IRC | 03:40 | |
| *** jroll has quit IRC | 03:41 | |
| *** mgagne has quit IRC | 03:41 | |
| *** gus has quit IRC | 03:41 | |
| *** sigmavirus24_awa has quit IRC | 03:41 | |
| *** neillc has quit IRC | 03:41 | |
| *** persia has quit IRC | 03:41 | |
| *** dolphm has quit IRC | 03:41 | |
| *** palendae has quit IRC | 03:41 | |
| *** spotz_zzz has quit IRC | 03:41 | |
| *** mattoliverau has quit IRC | 03:41 | |
| *** d34dh0r53 has quit IRC | 03:42 | |
| *** mattoliverau has joined #openstack-ansible | 03:42 | |
| *** persia has joined #openstack-ansible | 03:42 | |
| *** neillc has joined #openstack-ansible | 03:42 | |
| *** gus has joined #openstack-ansible | 03:44 | |
| *** bgmccollum has joined #openstack-ansible | 04:03 | |
| *** blewis_ has joined #openstack-ansible | 04:03 | |
| *** eglute has joined #openstack-ansible | 04:03 | |
| *** dolphm has joined #openstack-ansible | 04:03 | |
| *** mgagne has joined #openstack-ansible | 04:03 | |
| *** d34dh0r53 has joined #openstack-ansible | 04:04 | |
| *** lbragstad has joined #openstack-ansible | 04:04 | |
| *** jroll has joined #openstack-ansible | 04:04 | |
| *** spotz_zzz has joined #openstack-ansible | 04:04 | |
| *** palendae has joined #openstack-ansible | 04:04 | |
| *** sigmavirus24_awa has joined #openstack-ansible | 04:05 | |
| *** tlian2 has quit IRC | 04:14 | |
| *** markvoelker has joined #openstack-ansible | 04:15 | |
| *** markvoelker has quit IRC | 04:21 | |
| *** darrenc_ is now known as darrenc | 04:21 | |
| openstackgerrit | Kevin Carter proposed openstack/openstack-ansible: adds the config_template to swift https://review.openstack.org/223340 | 04:44 |
| *** shoutm has quit IRC | 04:45 | |
| *** shoutm has joined #openstack-ansible | 04:55 | |
| *** javeriak has joined #openstack-ansible | 05:01 | |
| *** abitha has joined #openstack-ansible | 05:31 | |
| *** shoutm has quit IRC | 05:45 | |
| *** shoutm has joined #openstack-ansible | 05:48 | |
| *** manas has joined #openstack-ansible | 05:52 | |
| *** javeriak has quit IRC | 05:52 | |
| *** javeriak has joined #openstack-ansible | 06:04 | |
| *** abitha has quit IRC | 06:12 | |
| *** markvoelker has joined #openstack-ansible | 06:18 | |
| *** markvoelker has quit IRC | 06:22 | |
| *** javeriak has quit IRC | 06:26 | |
| *** javeriak has joined #openstack-ansible | 07:12 | |
| *** willemgf has joined #openstack-ansible | 07:32 | |
| *** itsuugo has joined #openstack-ansible | 07:32 | |
| *** itsuugo has quit IRC | 07:33 | |
| *** itsuugo has joined #openstack-ansible | 07:34 | |
| *** shoutm_ has joined #openstack-ansible | 07:35 | |
| *** shoutm has quit IRC | 07:35 | |
| svg | mattt: ping | 07:43 |
| mattt | svg: morning2u | 07:44 |
| svg | Hello | 07:44 |
| svg | We're trying out a deploy based on 11.2.2 and bump into an issue related to the cinder-backup /ceph update | 07:45 |
| mattt | svg: you need to update your environment file | 07:45 |
| svg | Ah. That sound slike it might be it: error while evaluating conditional: inventory_hostname in groups[item.0.component] | 07:46 |
| mattt | svg: https://github.com/openstack/openstack-ansible/commit/44d3f25de6daab8073137c9fc76edcdfc23c716f#diff-137efd1f3dd0a7249de05894bd21dd54R26 | 07:47 |
| mattt | svg: depending on when you deployed your env, you may have cinder-volume on metal or not ... so based on that it's not safe to just copy the new file across | 07:49 |
| svg | I can't recall cinder-volume ever was on metal | 07:50 |
| mattt | svg: if a diff doesn't show any further changes then you'll be ok, but just be careful incase your cinder-volume config is different | 07:50 |
| mattt | svg: it defaults to on metal now, so you'll probably want to merge those changes in manually | 07:50 |
| mattt | svg: i really wanted to avoid making changes to the env for cinder-backup, but without it being in a specific group we couldn't target it properly in that task file | 07:51 |
| svg | the setup was initially based on a version some weeks before 11.1.0 | 07:51 |
| evrardjp | good morning everyone | 07:54 |
| mattt | svg: https://review.openstack.org/#/c/195181/ | 07:54 |
| mattt | merged july 7 | 07:54 |
| svg | thx | 07:54 |
| mattt | svg: no worries, but yeah def. the env file needs updating | 07:54 |
| svg | this is all damn confusing :-s | 07:55 |
| mattt | svg: yeah :( | 07:56 |
| mattt | that's why we deliberated over that review for so long, i wanted to avoid changes to env file :( | 07:56 |
| mattt | but the change was the cleanest long-term solution | 07:56 |
| svg | sure | 07:57 |
| svg | how osad's inventory works and how the envuronment is used where I remain confused | 07:58 |
| svg | I get the basic usage, but how OSA(D) works wrt a changing environment, that I don;t get | 07:59 |
| mattt | i have to be honest, i'm not an authority on osad inventory | 08:01 |
| *** gparaskevas has joined #openstack-ansible | 08:04 | |
| svg | Yeah, this project itself is already very complex. | 08:05 |
| mattt | i think a lot of people feel that way | 08:05 |
| mattt | specifically around the inventory | 08:06 |
| mattt | we've had some discussions about simplifying things, especially around how we do roles and playbooks etc. ... i hope that work extends to the inventory also | 08:06 |
| *** sdake has quit IRC | 08:07 | |
| *** itsuugo has quit IRC | 08:11 | |
| *** itsuugo has joined #openstack-ansible | 08:19 | |
| *** markvoelker has joined #openstack-ansible | 08:19 | |
| *** markvoelker has quit IRC | 08:23 | |
| odyssey4me | svg I know that palendae has been wanting to document how the dynamic inventory all fits together with regards to precedence, etc. | 08:29 |
| odyssey4me | We'd love to figure out how we can simplify the dynamic inventory though. Right now it's a powerful tool, but with power comes complexity. | 08:29 |
| svg | ack; I guess document how it works in detail would be a first step | 08:32 |
| odyssey4me | svg considering your own discussion around how best to handle inventory at scale, perhaps you've come up with some good ideas? | 08:38 |
| odyssey4me | I would imagine that using the dynamic inventory is a better option than most, but simplifying how we do it would be better. | 08:39 |
| svg | odyssey4me: that was an early opinion based on hoa I was used to handle inventory in general; given how inventory in osad is shaped, there is quite some complexity and I assume needs which I still don't grasp well enough to be able to come up with ideas, TBH | 08:41 |
| evrardjp | about the inventory at scale, it's definitely something we have expertise on, because 100% of our hosts are manageable with an ansible dynamic inventory | 08:51 |
| evrardjp | just a different than ansible ;) | 08:51 |
| evrardjp | osad* | 08:51 |
| svg | nice evrardjp - what tool do you use for that? | 08:56 |
| evrardjp | cobbler | 08:56 |
| evrardjp | and our own python scripts | 08:56 |
| svg | urgh, cobbler of all tjings | 08:56 |
| evrardjp | it works perfectly fine | 08:56 |
| svg | prolly | 08:57 |
| svg | I tend to see inventory as a single source of truth, from where everything comes from | 08:57 |
| svg | sth that should drive cobbler | 08:57 |
| svg | not the other way around | 08:58 |
| svg | but that's just me :) | 08:58 |
| svg | I see to many people focusing on inventory as just a simple host list | 08:58 |
| evrardjp | svg: it depends on what your source of truth is | 09:00 |
| evrardjp | ours is cobbler | 09:01 |
| evrardjp | yours could be monitoring | 09:01 |
| evrardjp | ansible is kinda flexible | 09:01 |
| evrardjp | and if you're using cobbler, remember it has a good xmlrpc interface | 09:01 |
| svg | sure | 09:03 |
| *** shausy has joined #openstack-ansible | 09:03 | |
| svg | my point is, as in your example, monitoring hsould not be the single source of truth | 09:03 |
| svg | you need a SSoT, then use it to deploy monitoring - IMHO | 09:04 |
| evrardjp | indeed | 09:04 |
| evrardjp | it just happens that our cobbler serves as a SSoT ;) | 09:05 |
| *** shausy has quit IRC | 09:08 | |
| *** javeriak has quit IRC | 09:08 | |
| *** shausy has joined #openstack-ansible | 09:09 | |
| *** javeriak has joined #openstack-ansible | 09:09 | |
| *** itsuugo has quit IRC | 09:09 | |
| svg | evrardjp: so, when do you go into production with OS? | 09:12 |
| evrardjp | September 21 | 09:12 |
| evrardjp | 2015 :D | 09:13 |
| svg | βΈ® ? | 09:13 |
| evrardjp | yeah it's close | 09:13 |
| *** markvoelker has joined #openstack-ansible | 09:19 | |
| *** itsuugo has joined #openstack-ansible | 09:23 | |
| *** markvoelker has quit IRC | 09:24 | |
| openstackgerrit | Merged openstack/openstack-ansible: Fix the heat stack user create https://review.openstack.org/223249 | 09:26 |
| openstackgerrit | Merged openstack/openstack-ansible: Disable scatter-gather offload on host bridges https://review.openstack.org/223292 | 09:36 |
| *** javeriak has quit IRC | 09:50 | |
| *** pradk has quit IRC | 09:59 | |
| *** javeriak has joined #openstack-ansible | 10:00 | |
| *** javeriak_ has joined #openstack-ansible | 10:09 | |
| *** javeriak has quit IRC | 10:10 | |
| *** javeriak has joined #openstack-ansible | 10:10 | |
| *** javeriak_ has quit IRC | 10:12 | |
| *** pradk has joined #openstack-ansible | 10:14 | |
| tiagogomes | hi, I am seeing this http://paste.openstack.org/show/462758/ when running osad . Should I be worried? | 10:30 |
| mattt | tiagogomes: what version of ansible are you using? | 10:35 |
| mattt | (ansible --version_ | 10:35 |
| mattt | (ansible --version) | 10:35 |
| tiagogomes | mattt, ansible 1.9.3 | 10:39 |
| *** shoutm_ has quit IRC | 10:40 | |
| mattt | tiagogomes: ok, wanted to make sure you weren't using an old problematic ansible version :) | 10:41 |
| *** shoutm has joined #openstack-ansible | 10:41 | |
| mattt | tiagogomes: wonder if you don't have bridge kernel module loaded | 10:45 |
| tiagogomes | mm, I have bridges interfaces up and running | 10:45 |
| mattt | tiagogomes: do you have ip_tables, ip6_tables, and ebtables loaded? | 10:47 |
| mattt | (not saying to load, i'm just curious) | 10:47 |
| tiagogomes | not ebtables | 10:48 |
| mattt | actually taht should be fine, ebtables shouldn't be loaded on a controller node i don't think | 10:49 |
| mattt | (my test node is an AIO so we have some overlap here) | 10:49 |
| *** javeriak has quit IRC | 10:56 | |
| *** itsuugo has quit IRC | 10:56 | |
| gparaskevas | what branch are you using and what kind of installation. AIO? | 11:00 |
| *** itsuugo has joined #openstack-ansible | 11:02 | |
| *** javeriak has joined #openstack-ansible | 11:05 | |
| *** shoutm has quit IRC | 11:05 | |
| tiagogomes | kilo with a multi-node installation | 11:09 |
| *** shoutm has joined #openstack-ansible | 11:14 | |
| openstackgerrit | Merged openstack/openstack-ansible: Install nfs-common with nova-compute https://review.openstack.org/221844 | 11:18 |
| *** markvoelker has joined #openstack-ansible | 11:20 | |
| openstackgerrit | Merged openstack/openstack-ansible: Install logrotate with ryslog https://review.openstack.org/222143 | 11:22 |
| *** markvoelker has quit IRC | 11:25 | |
| *** sdake has joined #openstack-ansible | 11:25 | |
| *** itsuugo has quit IRC | 11:27 | |
| *** sdake_ has joined #openstack-ansible | 11:30 | |
| *** sdake has quit IRC | 11:31 | |
| *** itsuugo has joined #openstack-ansible | 11:36 | |
| *** itsuugo has quit IRC | 11:48 | |
| *** javeriak_ has joined #openstack-ansible | 11:49 | |
| *** javeriak has quit IRC | 11:51 | |
| *** javeriak_ has quit IRC | 12:08 | |
| *** pradk has quit IRC | 12:14 | |
| *** markvoelker has joined #openstack-ansible | 12:19 | |
| *** woodard has joined #openstack-ansible | 12:20 | |
| *** itsuugo has joined #openstack-ansible | 12:22 | |
| *** shoutm has quit IRC | 12:24 | |
| *** sdake_ has quit IRC | 12:30 | |
| *** woodard has quit IRC | 12:32 | |
| tiagogomes | Hi, I got this error http://paste.openstack.org/show/462899/ when using osad. Any ideas? I am using the kilo branch | 12:33 |
| mgariepy | tiagogomes, ssh to bl002-test0_galera_container-4d824b40, and run : /usr/local/bin/pip install MySQL-python | 12:35 |
| mgariepy | you will see what's wrong i guess. | 12:35 |
| odyssey4me | tiagomes - did the setup-infrastructure complete? or did you use run-playbooks to go through the playbooks run? | 12:36 |
| odyssey4me | tiagogomes ^ | 12:36 |
| odyssey4me | also, are you using a set of repo servers which you've setup using a different branch before - or is the environment entirely fresh? | 12:37 |
| tiagogomes | mgariepy I got this http://paste.openstack.org/show/462909/ | 12:39 |
| tiagogomes | odyssey4me I ran setup-hosts, and them setup-infrastructure | 12:39 |
| tiagogomes | The environment is fresh | 12:39 |
| *** pradk has joined #openstack-ansible | 12:40 | |
| odyssey4me | tiagogomes if you look at your repo container in /var/www/repo can you find the python wheel for that? | 12:41 |
| odyssey4me | verify it on both your repo containers | 12:41 |
| *** itsuugo has quit IRC | 12:42 | |
| *** itsuugo has joined #openstack-ansible | 12:43 | |
| tiagogomes | I found this file /var/www/repo/os-releases/11.2.2/MySQL_python-1.2.5-cp27-none-linux_x86_64.whl | 12:44 |
| odyssey4me | is it in both repo containers? | 12:44 |
| tiagogomes | odyssey4me yes | 12:45 |
| odyssey4me | hmm, 'Connection refused' tells me that your load balancer is either not running or not properly setup | 12:45 |
| gparaskevas | if i am not mistakes you should have HAProxy installed if no hardware proxy is present. | 12:45 |
| odyssey4me | are you using an external LB or HAproxy ? | 12:45 |
| mgariepy | did you configured haproxy and forgot to run the playbook ? | 12:46 |
| gparaskevas | setup-hosts.yml then haproxy-install.yml and then the rest | 12:46 |
| odyssey4me | if you are using haproxy then you need to run the haproxy playbook before running setup-infrastructure | 12:46 |
| odyssey4me | :) what gparaskevas said | 12:46 |
| gparaskevas | everything relies on VIP , repo, enedpoints. | 12:47 |
| tiagogomes | odyssey4me ah, it could be that, I have some haproxy_hosts setup yes | 12:47 |
| gparaskevas | ok then run haproxy-install.yml | 12:48 |
| tiagogomes | gparaskevas ok, thanks, I'll try | 12:49 |
| gparaskevas | and then try again to run the infrastructure playbook | 12:49 |
| gparaskevas | tiagogomes: ok! np | 12:49 |
| *** tlian has joined #openstack-ansible | 12:54 | |
| odyssey4me | tiagogomes confirm that haproxy is running on the intended servers, and that the vip is bound on one of them | 12:55 |
| willemgf | I'm having following error during the openstack-ansible setup-openstack.yml for cinder_scheduler_container. Installing 11.2.2 | 12:56 |
| willemgf | failed: [dc2-rk1-ch1-bl2_cinder_scheduler_container-3fbb44ef] => (item=cinder) => {"attempts": 5, "cmd": "/usr/local/bin/pip install cinder", "failed": true, "item": "cinder"} | 12:56 |
| willemgf | In the container: root@dc2-rk1-ch1-bl2_cinder_scheduler_container-3fbb44ef:~# /usr/local/bin/pip install cinder | 12:56 |
| willemgf | Ignoring indexes: https://pypi.python.org/simple | 12:56 |
| willemgf | Collecting cinder | 12:56 |
| willemgf | Could not find a version that satisfies the requirement cinder (from versions: ) | 12:56 |
| willemgf | No matching distribution found for cinder | 12:56 |
| willemgf | Where should the version be defined? | 12:57 |
| *** gparaskevas has quit IRC | 12:57 | |
| odyssey4me | willemgf can you verify whether your repo containers has the pip wheel (all repo containers) and whether your load balancer is listening on the right ip and directing traffic to the repo containers | 12:57 |
| *** woodard has joined #openstack-ansible | 13:00 | |
| *** gparaskevas has joined #openstack-ansible | 13:04 | |
| tiagogomes | ok, I don't have that error anymore after running the ha_proxy playbook. But I have another one now: http://paste.openstack.org/show/462947/ | 13:04 |
| willemgf | odyssey4me: the package to be installed is present: http://10.16.8.10:8181/os-releases/11.2.2/cinder-2015.1.2.dev20-py2-none-any.whl | 13:06 |
| odyssey4me | tiagogomes seeing as your cluster is fresh and has nothing in it, you can follow the instructions in the message about the galera cluster not being healthy | 13:07 |
| willemgf | Also the pip-package: http://10.16.8.10:8181/os-releases/11.2.2/pip-7.1.2-py2.py3-none-any.whl | 13:07 |
| willemgf | as well for wheel: http://10.16.8.10:8181/os-releases/11.2.2/wheel-0.24.0-py2.py3-none-any.whl | 13:07 |
| odyssey4me | willemgf ok, from inside the cinder container do a verbose pip install of the package to see what it says | 13:07 |
| willemgf | all other pip installs went well so far, so it can not be a problem with loadbalancer. | 13:08 |
| odyssey4me | willemgf it could be - as the other packages may already be on the boxes and therefore they were just skipped | 13:08 |
| willemgf | root@dc2-rk1-ch1-bl2_cinder_scheduler_container-3fbb44ef:~# /usr/local/bin/pip install cinder -v | 13:09 |
| willemgf | Ignoring indexes: https://pypi.python.org/simple | 13:09 |
| willemgf | Collecting cinder | 13:09 |
| willemgf | 1 location(s) to search for versions of cinder: | 13:09 |
| willemgf | * http://10.16.8.10:8181/os-releases/11.2.2/ | 13:09 |
| willemgf | Skipping link http://10.16.8.10:8181/os-releases/11.2.2/ (from -f); unsupported archive format: .2 | 13:09 |
| willemgf | Getting page http://10.16.8.10:8181/os-releases/11.2.2/ | 13:09 |
| willemgf | Starting new HTTP connection (1): 10.16.8.10 | 13:09 |
| willemgf | "GET /os-releases/11.2.2/ HTTP/1.1" 200 None | 13:09 |
| willemgf | Analyzing links from page http://10.16.8.10:8181/os-releases/11.2.2/ | 13:09 |
| willemgf | Skipping link http://10.16.8.10:8181/os-releases/ (from http://10.16.8.10:8181/os-releases/11.2.2/); not a file | 13:09 |
| odyssey4me | willemgf it's usually better to use paste.openstack.org or pastebin to post log results | 13:10 |
| willemgf | yeah, sorry | 13:10 |
| tiagogomes | odyssey4me, that didn't solve the problem. Still having issues with the galera cluster : http://paste.openstack.org/show/462970/ | 13:12 |
| odyssey4me | tiagogomes so your galera cluster is in some sort of unhealthy state - you'll need to figure out the issue and get it into a healthy state before you can continue | 13:13 |
| willemgf | odyssey4me: here the complete output: http://paste.openstack.org/show/462974/ | 13:13 |
| *** itsuugo has quit IRC | 13:13 | |
| odyssey4me | willemgf and you're sure that every repo server has that cinder package? | 13:16 |
| odyssey4me | ie the LB may be doing round robin and some repo servers have not properly synchronised the repo | 13:16 |
| willemgf | Will check the 2 other containers as well | 13:16 |
| tiagogomes | odyssey4me looking at the log, is it expected to have `password: "" ` when connecting to the server | 13:17 |
| odyssey4me | tiagogomes did you generate your passwords? | 13:18 |
| odyssey4me | ie are they in /etc/openstack_deploy/user_secrets.yml ? | 13:18 |
| tiagogomes | no, I started fresh and I forgot to do it again | 13:18 |
| odyssey4me | tiagogomes haha, that'll cause issues with the setup process for sure :p | 13:19 |
| tiagogomes | no I need to re-run everything? | 13:19 |
| tiagogomes | re-do | 13:19 |
| odyssey4me | generate the passwords | 13:19 |
| odyssey4me | it's probably best to re-run each of the playbooks to be sure | 13:19 |
| evrardjp | hello | 13:19 |
| evrardjp | I don't know what happened to my kilo install but it seems that some of my repos are out of sync | 13:20 |
| evrardjp | re-running setup-infra doesn't change anything | 13:20 |
| evrardjp | did some of you got that issue? | 13:21 |
| evrardjp | get* | 13:21 |
| *** itsuugo has joined #openstack-ansible | 13:23 | |
| *** shausy has quit IRC | 13:25 | |
| odyssey4me | evrardjp that is possible if the lsync (not sure if that's the right name) stops running for some reason on the secondary repo servers | 13:26 |
| odyssey4me | there's a service that runs on all repo servers but the primary which handles the sync of changes | 13:27 |
| odyssey4me | I've seen it stop sometimes | 13:27 |
| evrardjp | if I run the setup-infra again, shouldn't it make sure the service is running? | 13:27 |
| odyssey4me | evrardjp unfortunately it does not seem to | 13:27 |
| evrardjp | I'll check the process lsync, will tell you | 13:27 |
| odyssey4me | it only verifies that the config is right and only restarts it if config changes - which it never does | 13:27 |
| evrardjp | yeah it's not running | 13:28 |
| evrardjp | I'll check in the playbooks, to see if we need a state=started | 13:28 |
| odyssey4me | we could probably do something better to ensure that the process restarts or have some sort of check to verify that it runs and keeps running | 13:29 |
| evrardjp | that's what I'll do with my monitoring system | 13:30 |
| odyssey4me | evrardjp I think that's the best approach | 13:30 |
| evrardjp | it doesn't seem to want to run, even manually | 13:31 |
| evrardjp | nothing in logs | 13:31 |
| odyssey4me | bear in mind that on one server (the primary) it should not run | 13:31 |
| odyssey4me | it should run on all the others though | 13:32 |
| evrardjp | ok my bad | 13:32 |
| odyssey4me | it's been a while since I tried it, but I do seem to recall that you can run it in the foreground or with a debug/verbose flag | 13:32 |
| evrardjp | you sure it's that way? | 13:32 |
| odyssey4me | yup | 13:32 |
| tiagogomes | ok, while I am waiting to redeploy. I don't possess an hardware balancer. If I setup haproxy hosts does that mean if one infrastructure node goes down, the other one is there to server? Do the haproxy hosts negotiates who has the virtual IP? Or I need something like corosync? | 13:33 |
| evrardjp | It's not the opposite? like only the master that sends the rsync to the others? | 13:33 |
| evrardjp | that send the command to rsync to the others* | 13:33 |
| *** jmckind has joined #openstack-ansible | 13:33 | |
| odyssey4me | hmm, evrardjp you may be right: https://github.com/openstack/openstack-ansible/blob/master/playbooks/roles/repo_server/tasks/main.yml#L28 | 13:34 |
| odyssey4me | odd, I always thought it was the other way around | 13:34 |
| evrardjp | only one of the repo has the lua file | 13:35 |
| evrardjp | one of the repo containers* | 13:35 |
| evrardjp | but I don't see the issue of running rsync accross different servers | 13:35 |
| *** woodard has quit IRC | 13:39 | |
| willemgf | OK, only one of the 3 repo-containers has the 11.2.2 repo. Looks like we have the same repo-sync problem as evrardjp has. | 13:40 |
| *** manas has quit IRC | 13:40 | |
| evrardjp | so now you have the answer :) | 13:40 |
| odyssey4me | hughsaunders can you please backport https://review.openstack.org/221844 and https://review.openstack.org/222143 | 13:40 |
| evrardjp | if you are using haproxy or hardware LB, you can disable the other backends temporarily | 13:41 |
| gparaskevas | tiagogomes: do you mean if one haproxy goes down? | 13:41 |
| evrardjp | willemgf: ^ | 13:41 |
| *** woodard has joined #openstack-ansible | 13:42 | |
| evrardjp | the other way would be to start lsyncd on the appropriate server and let the thing run | 13:42 |
| odyssey4me | tiagogomes that functionality is being worked on by evrardjp and is up for review at https://review.openstack.org/218818 | 13:42 |
| tiagogomes | gparaskevas yes, I assume that typically haproxy will run in every infrastructure node | 13:44 |
| gparaskevas | it does run where you say it will run | 13:45 |
| tiagogomes | odyssey4me ah very cool. I need something like that. I may cherry-pick the change | 13:45 |
| odyssey4me | tiagogomes both gparaskevas and evrardjp have been testing and improving hat patch :) | 13:46 |
| odyssey4me | so thank them :) | 13:46 |
| tiagogomes | thanks gparaskevas and evrardjp | 13:46 |
| tiagogomes | :) | 13:46 |
| gparaskevas | haha the osa police has spoken | 13:46 |
| gparaskevas | you are all wellcome | 13:46 |
| evrardjp | what? what did I do? | 13:46 |
| gparaskevas | lol | 13:47 |
| odyssey4me | lol, only good things evrardjp :) | 13:47 |
| evrardjp | ;) | 13:47 |
| *** itsuugo has quit IRC | 13:47 | |
| evrardjp | mgariepy is also a tester | 13:47 |
| evrardjp | he really helped improving the product ;) | 13:47 |
| odyssey4me | ah yes, I apologise for leaving mgariepy out | 13:48 |
| evrardjp | I'll ask cloudnull to check if it (at least the inventory part) sounds good to him | 13:48 |
| evrardjp | it's not because he's living far away and have speaks french, you forgot him, right? ;) | 13:49 |
| mgariepy | i don't do that for the fame ;p | 13:49 |
| *** itsuugo has joined #openstack-ansible | 13:49 | |
| mgariepy | is there an easy way to update patch set in my tree ? | 13:49 |
| mgariepy | i cherry-picket patch set 5 and would like to update to patch set 10.. | 13:50 |
| odyssey4me | mgariepy as in a patch which you cherry-picked into a fork? | 13:50 |
| odyssey4me | personally I rebase to remove the old patch, then cherry-pick the new one - but you could revert the old and cherry-pick the new | 13:50 |
| mgariepy | i use the cherry-pick command in gerrit | 13:50 |
| evrardjp | is it the last commit? | 13:51 |
| mgariepy | no | 13:51 |
| mgariepy | not the last | 13:51 |
| evrardjp | if it was it's simple, reset HEAD^and cherry pick again | 13:51 |
| willemgf | evrardjp: thnx, running lsyncd solved my problem | 13:51 |
| evrardjp | willemgf: no worries | 13:51 |
| cloudnull | morning | 13:51 |
| cloudnull | evrardjp: looking now | 13:52 |
| evrardjp | good morning cloudnull | 13:52 |
| cloudnull | hows it going today ? | 13:52 |
| evrardjp | cloudnull: we are not in a rush, don't worry | 13:52 |
| evrardjp | I just don't want it to be lost in oblivion | 13:52 |
| odyssey4me | o/ cloudnull | 13:53 |
| cloudnull | oblivion is never good :) | 13:53 |
| evrardjp | cloudnull: it's going fine in Belgium. How is life in texas? beautiful weather? It's rainy here :p | 13:53 |
| cloudnull | it is a bit rainy here. and the weather has been a lovly 23-28 the last couple of days. | 13:55 |
| cloudnull | so no complaints from me :) | 13:55 |
| openstackgerrit | Hugh Saunders proposed openstack/openstack-ansible: Install logrotate with ryslog https://review.openstack.org/223603 | 13:58 |
| openstackgerrit | Hugh Saunders proposed openstack/openstack-ansible: Install nfs-common with nova-compute https://review.openstack.org/223604 | 13:58 |
| mgariepy | evrardjp, so you are belgian ? | 13:59 |
| *** KLevenstein has joined #openstack-ansible | 13:59 | |
| evrardjp | mgariepy: why? what did you expect? | 14:00 |
| mgariepy | i'm half belgian, my mother is belgian. | 14:00 |
| evrardjp | cool! | 14:00 |
| mgariepy | i'm not expecting anything ;p haha | 14:00 |
| mgariepy | she came to canada when she was 5 or 6. | 14:01 |
| evrardjp | FYI, svg and willemgf are Belgians too if I'm not mistaken | 14:01 |
| *** cloudtrainme has joined #openstack-ansible | 14:01 | |
| evrardjp | but we are getting personal, it's maybe not the purpose of this chan ;) | 14:01 |
| svg | ASL? | 14:02 |
| mgariepy | lol | 14:02 |
| evrardjp | >< | 14:02 |
| odyssey4me | oh dear | 14:02 |
| cloudnull | evrardjp: looks good to me. seems to be real clean. +1 for now, I hope to have some time today to give it a test. | 14:02 |
| *** willemgf has quit IRC | 14:03 | |
| evrardjp | other topic, I'm getting this: error while evaluating conditional: inventory_hostname in groups[item.0.component] | 14:03 |
| evrardjp | on glance install right now (I'm using ceph backend) | 14:03 |
| evrardjp | it comes from ceph_install | 14:04 |
| evrardjp | found it | 14:05 |
| mattt | evrardjp: you also getting burned by same issue as svg ? | 14:05 |
| mattt | (needing to update cinder env.d file ?) | 14:05 |
| *** cloudtra_ has joined #openstack-ansible | 14:05 | |
| evrardjp | found the commit to blame | 14:06 |
| evrardjp | it's indeed yours | 14:06 |
| evrardjp | :p | 14:06 |
| *** phalmos has joined #openstack-ansible | 14:06 | |
| evrardjp | I'll update my env.d | 14:06 |
| mattt | hehehe | 14:06 |
| evrardjp | and see | 14:06 |
| *** cloudtra_ has quit IRC | 14:06 | |
| odyssey4me | evrardjp you gotta follow you own docs ;) | 14:06 |
| evrardjp | what do you mean? | 14:06 |
| *** cloudtra_ has joined #openstack-ansible | 14:07 | |
| odyssey4me | evrardjp https://github.com/openstack/openstack-ansible/commit/4bf2a1f37c17433ff37246acd0fd292aabd02d70 ;) | 14:07 |
| mattt | odyssey4me: ah, different issue here | 14:08 |
| evrardjp | odyssey4me: I'm using both backends ;) | 14:08 |
| odyssey4me | ah ok :p | 14:08 |
| *** itsuugo has quit IRC | 14:09 | |
| *** cloudtrainme has quit IRC | 14:09 | |
| *** cloudtra_ has quit IRC | 14:12 | |
| evrardjp | mattt: "inventory_hostname in groups[item.0.component]" | 14:12 |
| mattt | evrardjp: ? | 14:13 |
| *** spotz_zzz is now known as spotz | 14:13 | |
| evrardjp | mmm wait | 14:13 |
| *** sigmavirus24_awa is now known as sigmavirus24 | 14:13 | |
| *** Mudpuppy has joined #openstack-ansible | 14:14 | |
| evrardjp | I'll re-test and tell you what really bit me | 14:14 |
| mattt | evrardjp: that error looks right | 14:15 |
| mattt | as in, i'd expect it if you don't have the current cinder env | 14:15 |
| *** Mudpuppy has joined #openstack-ansible | 14:15 | |
| *** cloudtrainme has joined #openstack-ansible | 14:15 | |
| *** tiagogomes has quit IRC | 14:19 | |
| evrardjp | even if running glance? | 14:19 |
| evrardjp | but you're right, that's the env | 14:20 |
| *** tiagogomes has joined #openstack-ansible | 14:20 | |
| evrardjp | so, forget my previous message :p | 14:20 |
| *** itsuugo has joined #openstack-ansible | 14:21 | |
| *** alextricity has quit IRC | 14:28 | |
| *** tiagogomes has quit IRC | 14:31 | |
| *** tiagogomes has joined #openstack-ansible | 14:31 | |
| sigmavirus24 | git-harry: if by "not actually doing what the swift play should be doing" is working, then I'll just happily abandon that patch | 14:39 |
| mattt | evrardjp: i am a bit surprised it triggers on glance tbh | 14:39 |
| git-harry | sigmavirus24: If I've misunderstood I'm happy to be corrected but from what I could tell it was working as expected | 14:41 |
| sigmavirus24 | git-harry: it's not working as intended, it's skipping those tasks because the account '{' does not exist | 14:41 |
| sigmavirus24 | Although perhaps we don't want to run those tasks because their broken so it is working as intended? | 14:41 |
| *** itsuugo has quit IRC | 14:41 | |
| git-harry | sigmavirus24: to confirm, are you referring to the task "Ensure contents file matches ring after ring sync for account/container" | 14:42 |
| *** tiagogomes has quit IRC | 14:44 | |
| sigmavirus24 | Sorry misspoke, that file was the first we noticed it in though, yes | 14:47 |
| sigmavirus24 | (Also your point about the docs is valid, except that it's very clear that we're iterating over a string, not a list in these tasks) | 14:47 |
| git-harry | sigmavirus24: okay, so the reason that task is skipped is because swift_managed_regions is not defined, it's nothing to do with the random values the command would receive | 14:49 |
| git-harry | sigmavirus24: so the logs look a bit weird but it's not exactly broken, if you define the variable and set it to an empty list everything will work as expected. | 14:50 |
| sigmavirus24 | git-harry: but that's showing the parameter to the task (as its running the task over the items it thinks it has) | 14:50 |
| sigmavirus24 | If "swift_managed_regions" is not defined why is it still iterating over the string like that? | 14:50 |
| git-harry | sigmavirus24: because that's how ansible works | 14:51 |
| *** cloudtrainme has quit IRC | 14:51 | |
| sigmavirus24 | I'm not quite convinced sir | 14:51 |
| *** tiagogomes has joined #openstack-ansible | 14:52 | |
| sigmavirus24 | especially given the majority of our uses are sans "{{ }}" | 14:52 |
| git-harry | sigmavirus24: the better way to do this would be to git swift_managed_regions a default value of [] | 14:52 |
| git-harry | sigmavirus24: and as I pointed out in the documentation that is not how ansible suggests. | 14:52 |
| sigmavirus24 | git-harry: except documentation is always a lie | 14:53 |
| sigmavirus24 | I thought you knew that | 14:53 |
| git-harry | sigmavirus24: also changing it to what you suggest doesn't actually make any difference to what ansible is doing in the sense that the task is still being skipped for the same reason only now is shows a junk value of the variable name instead of splitting it into individual characters | 14:54 |
| *** shoutm has joined #openstack-ansible | 14:54 | |
| *** phalmos has quit IRC | 14:54 | |
| git-harry | sigmavirus24: so I stand by my point that this is only really adjusting the skipped log entries and even then it still shows something that isn't quite as nice as it should be | 14:55 |
| *** phalmos has joined #openstack-ansible | 14:59 | |
| *** itsuugo has joined #openstack-ansible | 14:59 | |
| tiagogomes | I still didn't manage to have the galera container correctly configured. These are the first errors related to MySQL that I am seeing: http://paste.openstack.org/show/463126/ | 15:00 |
| *** cloudtrainme has joined #openstack-ansible | 15:03 | |
| odyssey4me | tiagogomes notice the ..ignoring - those are meant to fail on a fresh cluster | 15:07 |
| *** jwagner_away is now known as jwagner | 15:07 | |
| tiagogomes | ah, that's good to know. So far I am only seeing ..ignoring. The playbook is still running | 15:08 |
| mattt | evrardjp: get it sorted ? | 15:08 |
| *** galstrom_zzz is now known as galstrom | 15:09 | |
| mattt | tiagogomes: did you figure the sysctl -p issue you mentioned earlier ? | 15:09 |
| evrardjp | mattt: I didn't check further | 15:09 |
| evrardjp | it was solved with the env.d overwriting | 15:09 |
| evrardjp | I'm on another thing | 15:09 |
| evrardjp | ERROR: openstack The request you have made requires authentication | 15:09 |
| evrardjp | during heat install | 15:09 |
| odyssey4me | evrardjp https://review.openstack.org/223249 may be useful to you? | 15:11 |
| mattt | odyssey4me: i was just looking at that | 15:11 |
| mattt | odyssey4me: but i'd have guessed that our gate would fail if that was absolutely required | 15:11 |
| odyssey4me | mattt it typically only fails if the playbook is re-run after failing the first time | 15:12 |
| odyssey4me | the above patch sorts it out properly | 15:12 |
| evrardjp | my issue happen at Create heat domain admin user | 15:12 |
| odyssey4me | (ie so that it works on every run, or doesn't work at all) | 15:12 |
| evrardjp | I'll check | 15:12 |
| evrardjp | if this introduced the bug | 15:12 |
| odyssey4me | evrardjp confirm that the heat container in question has an appropriate openrc | 15:12 |
| *** gparaskevas has quit IRC | 15:13 | |
| evrardjp | I think it's maybe my haproxy ssl that could cause the issue | 15:13 |
| prometheanfire | odyssey4me: about your candidacy, what are you looking to do about the stability of the existing project? | 15:13 |
| odyssey4me | are you using self-signed certs? | 15:13 |
| prometheanfire | moving forward is nice and all, but I'm concerned about moving too fast and turning into the next neutron | 15:14 |
| *** cloudtra_ has joined #openstack-ansible | 15:14 | |
| prometheanfire | gaining tech debt etc | 15:14 |
| odyssey4me | prometheanfire each area has an implication of managing that - I didn't want to write more of an essay than I already did | 15:14 |
| mattt | odyssey4me: you sent this to -dev ? | 15:15 |
| odyssey4me | mattt no, I'm not campaigning as it appears that there is no-one else raising their hand | 15:15 |
| evrardjp | odyssey4me: nope I'm not ;) | 15:15 |
| odyssey4me | http://git.openstack.org/cgit/openstack/election/plain//candidates/mitaka/OpenStackAnsible/jesse_pretorius.txt | 15:16 |
| mattt | ah cool | 15:16 |
| odyssey4me | I believe that the actual voting happens next week | 15:16 |
| tiagogomes | mattt no I am ignoring it for the time being | 15:17 |
| *** cloudtrainme has quit IRC | 15:17 | |
| *** shoutm has quit IRC | 15:17 | |
| evrardjp | I'm completely clueless with this heat stuff | 15:17 |
| odyssey4me | prometheanfire also, I question calling OSA 'unstable' at this point in time - the issues experienced in testing upgrades from Juno to Kilo were expected as it was the transition between the RPC product and a community product which included a mass of architectural changes... So if there's a perceived issue of instability, then I'd like that to be a constructive discussion at the summit. Can you please add it as a topic in https://etherpad. | 15:22 |
| odyssey4me | openstack.org/p/openstack-ansible-mitaka-summit - perhaps in the Workroom Slots? | 15:22 |
| odyssey4me | re-link: https://etherpad.openstack.org/p/openstack-ansible-mitaka-summit | 15:22 |
| evrardjp | I'll check heat stuff tomorrow | 15:23 |
| mattt | evrardjp: which tag you using again? | 15:23 |
| evrardjp | last kilo | 15:23 |
| mattt | evrardjp: ok let me try a build off that and see if i can duplicate | 15:23 |
| evrardjp | if you don't mind I'll check tomorrow | 15:23 |
| palendae | odyssey4me: Let me answer that with a question - how many OSA juno -> kilo upgrades have you done? | 15:23 |
| *** alextricity has joined #openstack-ansible | 15:23 | |
| mattt | evrardjp: yeah i'm not going to be here much longer | 15:23 |
| evrardjp | we can check tomorrow, maybe not lose time for me right now ;) | 15:23 |
| palendae | Just OSA | 15:23 |
| prometheanfire | odyssey4me: I don't call it unstable | 15:24 |
| odyssey4me | palendae the question is not relevant - juno is rax tech debt... the upgrade to kilo was expected to be an issue | 15:24 |
| prometheanfire | odyssey4me: just worried about adding non-working code, review quality, etc | 15:24 |
| palendae | odyssey4me: I think it is; just because juno is rackspace tech debt doesn't mean others aren't used it | 15:24 |
| *** itsuugo has quit IRC | 15:24 | |
| palendae | using* | 15:25 |
| prometheanfire | new features are what I'm most worried about | 15:25 |
| odyssey4me | palendae it is not something that will be repeated, so it's not a topic relevant to the next cycle | 15:25 |
| prometheanfire | but is someing that's not finished :P | 15:25 |
| odyssey4me | the only topic of relevance would be the design of an upgrade framework, I would think | 15:25 |
| prometheanfire | remaining tech debt should be a topic | 15:25 |
| palendae | So juno -> kilo upgrades being unstable remains | 15:25 |
| palendae | And will | 15:25 |
| palendae | That's what I'm hearing | 15:25 |
| *** Mudpuppy has quit IRC | 15:27 | |
| palendae | Until there are consistent upgrade gate jobs, hard for it to be considered reliable, IMO | 15:27 |
| palendae | For any version to any one | 15:27 |
| prometheanfire | +1 | 15:27 |
| palendae | I know it's in the works, and the gate job split will help in general, but that's where my perception of instability in OSA comes from | 15:28 |
| palendae | The gate job doesn't reflect reality, as evidenced in the many work arounds implemented to get it to run on AIOs | 15:28 |
| odyssey4me | I think that the definition of 'unstable' needs to be agreed on then. | 15:28 |
| *** Mudpuppy has joined #openstack-ansible | 15:28 | |
| palendae | Bah, my UPS battery's almost dead, moving to different location | 15:29 |
| *** cloudtrainme has joined #openstack-ansible | 15:29 | |
| *** cloudtr__ has joined #openstack-ansible | 15:30 | |
| *** cloudtra_ has quit IRC | 15:31 | |
| odyssey4me | prometheanfire to go back to the question you raised, do you have a proposed session in mind to cover that? do you think we can effectively have a discussion at the summit about it? | 15:32 |
| *** cloudtrainme has quit IRC | 15:33 | |
| prometheanfire | that's a good idea, but I don't think saying 'make sure new features have tests' is enough for a session | 15:34 |
| prometheanfire | a session on upgrades would be nice though, that's a community concern in general and would pull them in | 15:34 |
| odyssey4me | prometheanfire :) yeah, I'm seeing it more like some documentation to create | 15:34 |
| prometheanfire | ? | 15:35 |
| odyssey4me | so I think we should have a discussion about documentation for developers - what's missing, etc | 15:35 |
| prometheanfire | maybe | 15:35 |
| odyssey4me | that would be to cover stuff like policies, etc | 15:35 |
| prometheanfire | doc best practices and general requirements in spec / feature submission? | 15:35 |
| odyssey4me | yeah, let me put something together along that line | 15:35 |
| prometheanfire | cool | 15:35 |
| *** mgariepy has quit IRC | 15:37 | |
| *** itsuugo has joined #openstack-ansible | 15:38 | |
| *** mgariepy has joined #openstack-ansible | 15:39 | |
| palendae | Would that I had time to finish the developer docs I had in mind :( | 15:49 |
| palendae | I wanted to do a quickstart "here's how to have an AIO up and running" | 15:51 |
| palendae | Bring-your-own-12GB minimum machine (because 8's too little now, realistically), checkout, run this, done | 15:51 |
| odyssey4me | palendae you mean like https://github.com/openstack/openstack-ansible/blob/master/development-stack.rst ? | 15:51 |
| palendae | Yeah, I guess that really is what I had in mind | 15:52 |
| palendae | Not sure how many people have read it, though. I'd probably move the instructions for starting up a stack into a 'Quickstart' or something | 15:52 |
| palendae | Or have a 'Quickstart' link back to that | 15:52 |
| palendae | Of course there's the problem that OpenStack in general fights where people unfamiliar with Linux are coming to it and expecting it to be easy | 15:53 |
| -cloudnull- -- bug triage in 5 min | 15:54 | |
| prometheanfire | neat | 15:55 |
| *** Bjoern_ has joined #openstack-ansible | 15:55 | |
| *** alop has joined #openstack-ansible | 15:57 | |
| *** fawadkhaliq has joined #openstack-ansible | 15:58 | |
| tiagogomes | Is there an easy way to disable heat? | 15:59 |
| odyssey4me | prometheanfire I'm trying to put together a thingy about code quality, but it seems to overlap too much with an existing topic I suggested - the Keystone Role Analysis. Can you see what I mean? | 15:59 |
| odyssey4me | tiagogomes not right now - it's expected to be a core deployment service | 16:00 |
| cloudnull | bug triage time cloudnull, mattt, andymccr, d34dh0r53, hughsaunders, b3rnard0, palendae, Sam-I-Am, odyssey4me, serverascode, rromans, mancdaz, dolphm, _shaps_, BjoernT, claco, echiu, dstanek, jwagner, ayoung, prometheanfire, evrardjp, arbrandes, mhayden | 16:00 |
| prometheanfire | ohai | 16:00 |
| * mhayden woots | 16:00 | |
| d34dh0r53 | o/ | 16:00 |
| Sam-I-Am | moos | 16:00 |
| * dstanek yawns | 16:00 | |
| prometheanfire | odyssey4me: link? | 16:00 |
| mhayden | hold on, need to put away the selfie stick | 16:00 |
| cloudnull | first up https://bugs.launchpad.net/openstack-ansible/+bug/1435808 | 16:01 |
| openstack | Launchpad bug 1435808 in openstack-ansible trunk "glance module doesn't support multiple api versions" [Wishlist,Confirmed] - Assigned to Darren Birkett (darren-birkett) | 16:01 |
| prometheanfire | odyssey4me: I think a generic talk would be good, also one on upgrades | 16:01 |
| prometheanfire | sorry, will stop distracting now | 16:01 |
| odyssey4me | prometheanfire https://etherpad.openstack.org/p/openstack-ansible-mitaka-summit | 16:01 |
| cloudnull | mancdaz: you around? | 16:01 |
| cloudnull | sems like this was confirmed already, idk if this is an issue or a wishlist item? | 16:02 |
| mancdaz | cloudnull o/ | 16:02 |
| cloudnull | ohai | 16:02 |
| mancdaz | sup | 16:02 |
| cloudnull | re: https://bugs.launchpad.net/openstack-ansible/+bug/1435808 <- what say you ? | 16:02 |
| openstack | Launchpad bug 1435808 in openstack-ansible trunk "glance module doesn't support multiple api versions" [Wishlist,Confirmed] - Assigned to Darren Birkett (darren-birkett) | 16:02 |
| mancdaz | cloudnull I think I filed a patch for this | 16:02 |
| mancdaz | lemme find | 16:03 |
| rromans | . | 16:03 |
| cloudnull | yup and its merged. .. https://review.openstack.org/#/c/167959/ | 16:03 |
| mancdaz | https://github.com/openstack/openstack-ansible/commit/53cc76b15e55ab36ff3c9b434ab3f272bdea6099 | 16:03 |
| mancdaz | yep | 16:04 |
| cloudnull | can you cherry pick athat commit back to kilo? | 16:04 |
| mancdaz | just bad gerrit bug updating | 16:04 |
| mancdaz | right did it not ever get into kilo? | 16:04 |
| mancdaz | ok I will do | 16:04 |
| cloudnull | not according to gerrit | 16:04 |
| cloudnull | maybe it went in without the cherry-pick ? | 16:04 |
| * cloudnull looking | 16:05 | |
| odyssey4me | nothing with the same change id | 16:05 |
| cloudnull | yup | 16:05 |
| mancdaz | well it's in kilo | 16:06 |
| mancdaz | it possibly went into master before we branched? | 16:06 |
| cloudnull | maybe , its an old issue. | 16:07 |
| cloudnull | thats been sitting there for a while. | 16:07 |
| mancdaz | https://github.com/openstack/openstack-ansible/blame/kilo/playbooks/library/glance | 16:07 |
| cloudnull | well then we're done here :) | 16:07 |
| mancdaz | ok I will update the issue with the things | 16:07 |
| mancdaz | not sure what release to target that at | 16:07 |
| mancdaz | since the release will probably be closed for new issues on launchpad | 16:08 |
| cloudnull | yup, i set it fix released. it should be good for now. | 16:08 |
| cloudnull | next https://bugs.launchpad.net/openstack-ansible/+bug/1487155 | 16:08 |
| openstack | Launchpad bug 1487155 in openstack-ansible "Juno to Kilo upgrade: nova spice console services should be removed" [Undecided,New] | 16:08 |
| cloudnull | Bjoern_: cc | 16:08 |
| *** Bjoern_ is now known as BjoernT | 16:08 | |
| cloudnull | this is something that "should" be resovled within the upgrade script | 16:08 |
| cloudnull | we're you seeing something different ? | 16:09 |
| BjoernT | I have not real preference where this gets done. Either playbooks or upgrade script | 16:09 |
| BjoernT | I think the better place might be playbooks since we going away from the upgrade anyway | 16:10 |
| cloudnull | this is where its currently handled https://github.com/openstack/openstack-ansible/blob/kilo/scripts/run-upgrade.sh#L384-L401 | 16:10 |
| BjoernT | it's not about the containers its about the nova service entry | 16:10 |
| odyssey4me | cloudnull this is to remove the keystone entries, not the containers | 16:10 |
| odyssey4me | sorry, nova entries | 16:10 |
| cloudnull | ah. | 16:11 |
| cloudnull | my bad | 16:11 |
| cloudnull | was not reading so good :) | 16:11 |
| BjoernT | yeah it's early in the morning, lol | 16:11 |
| cloudnull | in the current structure, this would have to be added to the upgrade script or simply documented. | 16:12 |
| *** itsuugo has quit IRC | 16:12 | |
| cloudnull | Sam-I-Am: palendae: do you guys have a min to go have a look at this and make sure we're covered from an upgrade perspective ? | 16:12 |
| palendae | Looking | 16:13 |
| cloudnull | while thats happening , https://bugs.launchpad.net/openstack-ansible/+bug/1489451 | 16:13 |
| openstack | Launchpad bug 1489451 in openstack-ansible "Not able to connect memcached_container while running setup-infrastructure.yml " [Undecided,New] | 16:13 |
| palendae | BjoernT: Does this apply to the ec2 entries too? | 16:13 |
| palendae | From my perspective, this'll be handled in docs for the time being, but a patch can be worked | 16:14 |
| Sam-I-Am | do we have a docs issue somewhere? | 16:14 |
| *** phalmos has quit IRC | 16:14 | |
| palendae | Sam-I-Am: Probably not yet, I can make one | 16:14 |
| cloudnull | kk will update that issue as a doc bug for now. | 16:15 |
| Sam-I-Am | if its an upgrade-specific thing, the only place for that now is in rpc docs | 16:15 |
| palendae | Or that | 16:15 |
| cloudnull | on https://bugs.launchpad.net/openstack-ansible/+bug/1489451 - thers been a few comments on this issue. i've not personally seen this in the wild however it seems like some have. I think with the chroot / update image solution that odyssey4me majorhaden and andymccr have been pondering we might be able to better solve that issue | 16:16 |
| openstack | Launchpad bug 1489451 in openstack-ansible "Not able to connect memcached_container while running setup-infrastructure.yml " [Undecided,New] | 16:16 |
| odyssey4me | yeah, it does seem like somehow people are getting the wrong image or the lxc-create playbooks aren't completing properly | 16:17 |
| odyssey4me | https://github.com/openstack/openstack-ansible/blob/master/playbooks/roles/lxc_container_create/tasks/container_create.yml#L321-L330 is meant to ensure that python2.7 is installed and linked | 16:18 |
| cloudnull | in truth im not sure how that is, but it looks like something we need to investigate and fix. | 16:18 |
| odyssey4me | but it seems like it sometimes fails to execute properly | 16:18 |
| cloudnull | confirmed the issue and marked it high | 16:19 |
| odyssey4me | so one way is obviously to switch to creating a base lxc cache/image which contains everything required... | 16:19 |
| cloudnull | ++ | 16:19 |
| palendae | I'm +1 | 16:19 |
| odyssey4me | but this is a change in the way we do things, so I'm not sure it's an appropriate way to resolve this particular issue | 16:19 |
| palendae | When using images, build up a base with the thing in it | 16:19 |
| palendae | things* | 16:20 |
| odyssey4me | but I can propose a review and we can take it from there | 16:20 |
| tiagogomes | mm, I just finish the OSAD deployment but I can't access horizon. I ssh'ed to the container but and there is nothing listening on port 80/8080 | 16:22 |
| cloudnull | next - https://bugs.launchpad.net/openstack-ansible/+bug/1494097 | 16:23 |
| openstack | Launchpad bug 1494097 in openstack-ansible " Be more clear about used_ips, mostly in the example file" [Undecided,New] | 16:23 |
| cloudnull | doc issue | 16:24 |
| cloudnull | next https://bugs.launchpad.net/openstack-ansible/+bug/1494109 | 16:25 |
| openstack | Launchpad bug 1494109 in openstack-ansible " Adds the ability to provide user certificates to HAProxy" [Undecided,New] | 16:25 |
| cloudnull | doc issue | 16:25 |
| *** itsuugo has joined #openstack-ansible | 16:25 | |
| cloudnull | https://bugs.launchpad.net/openstack-ansible/+bug/1494227 | 16:26 |
| openstack | Launchpad bug 1494227 in openstack-ansible "neutron metadata service running but not responding properly" [Undecided,New] | 16:26 |
| cloudnull | mancdaz: cc ^ - was that a Juno build ? | 16:26 |
| Sam-I-Am | looks like rabbit crud | 16:27 |
| mancdaz | cloudnull this was juno - there is a corresponding rpc-openstack bug | 16:27 |
| mancdaz | I filed it on behalf of support | 16:27 |
| Sam-I-Am | or db | 16:27 |
| *** sdake has joined #openstack-ansible | 16:28 | |
| cloudnull | ok juno issue confirmed / medium - trunk invalid - | 16:28 |
| Sam-I-Am | dont think this is an osad-specific thing | 16:29 |
| cloudnull | we'll have to try and recreate it within a Juno lab / AIO and see if we can figure out what caused that problem, however I suspect its an oslo messaging issue. | 16:30 |
| *** daneyon_ has left #openstack-ansible | 16:31 | |
| cloudnull | next https://bugs.launchpad.net/openstack-ansible/+bug/1494295 | 16:31 |
| openstack | Launchpad bug 1494295 in openstack-ansible "Rabbitmq must forget node during infra node recovery" [Undecided,New] | 16:31 |
| *** sdake_ has joined #openstack-ansible | 16:31 | |
| cloudnull | it seems to me this is a wishlist item . thoughts? | 16:31 |
| odyssey4me | yeah, it looks like something that should perhaps be considered in the upgrade framework too | 16:32 |
| odyssey4me | similar work to that done in the galera cluster - ie don't touch it if it's broken | 16:32 |
| cloudnull | yea. | 16:33 |
| git-harry | Repair seems like a separate use case for the project to take on, are there any existing examples in the project? | 16:33 |
| odyssey4me | nope, although I'm happy to see a repair handled in a playbook if there are known problems with solutions that can be automated | 16:34 |
| *** cloudtr__ has quit IRC | 16:34 | |
| odyssey4me | for now though I'd advocate the same as the galera playbook - check status and bail if the cluster's broken | 16:34 |
| cloudnull | ++ i agree with that | 16:35 |
| *** sdake has quit IRC | 16:35 | |
| *** cloudtrainme has joined #openstack-ansible | 16:35 | |
| cloudnull | next https://bugs.launchpad.net/openstack-ansible/+bug/1495139 | 16:36 |
| openstack | Launchpad bug 1495139 in openstack-ansible "Getting Error while performing galera-install.yml in OSAD "error: 'Can't connect to local MySQL server through socket '/var/run/mysqld/mysqld.sock' (111)'" [Undecided,New] | 16:36 |
| *** cloudtra_ has joined #openstack-ansible | 16:36 | |
| cloudnull | is chandu aroun d? | 16:37 |
| cloudnull | i asked in the issue if they've run the command noted in the failure message. | 16:38 |
| cloudnull | issue marked as incomplete | 16:38 |
| cloudnull | for now | 16:38 |
| cloudnull | next https://bugs.launchpad.net/openstack-ansible/+bug/1496001 | 16:39 |
| openstack | Launchpad bug 1496001 in openstack-ansible "Add SSL listener to RabbitMQ" [Undecided,New] - Assigned to Major Hayden (rackerhacker) | 16:39 |
| cloudnull | its a wishlist item but does anyone have any thoughts about the issue ? | 16:39 |
| *** cloudtrainme has quit IRC | 16:40 | |
| cloudnull | cc sigmavirus24 ? | 16:40 |
| sigmavirus24 | cloudnull: I'm not against it | 16:40 |
| sigmavirus24 | I just want to be sure our threat models are understood correctly | 16:40 |
| sigmavirus24 | mhayden: is in here too | 16:40 |
| *** cloudtra_ has quit IRC | 16:41 | |
| cloudnull | mhayden: what say you ? :) | 16:42 |
| openstackgerrit | Kevin Carter proposed openstack/openstack-ansible: Add auth version for legacy OpenStack clients https://review.openstack.org/223692 | 16:43 |
| cloudnull | ok well confirmed, wishlist. we'll have to discuss on the ML or in the issue. | 16:44 |
| cloudnull | next https://bugs.launchpad.net/openstack-ansible/+bug/1495687 | 16:45 |
| openstack | Launchpad bug 1495687 in openstack-ansible juno "Don't create cinder_volumes containers on hosts without cinder-volumes VG" [Undecided,New] | 16:45 |
| cloudnull | this is a Juno only specific item, IMO this is a wishlist issue . | 16:45 |
| cloudnull | cc rackertom mancdaz | 16:46 |
| odyssey4me | this seems a little dangerous | 16:46 |
| odyssey4me | if servers are designated as storage servers then they should be properly prepared | 16:47 |
| *** gparaskevas has joined #openstack-ansible | 16:47 | |
| cloudnull | the cinder hosts can already be defined by the various groups and be instructed to use different cinder backend s | 16:47 |
| odyssey4me | the containers can be created for non lvm too | 16:48 |
| cloudnull | IMO if the deployer does not want the cinder volume container on a given host, the host should not be defined within the storage host group. | 16:48 |
| odyssey4me | also, I don't see how this is juno only | 16:48 |
| cloudnull | the issue was opened with juno as the only series targetted. | 16:49 |
| odyssey4me | yeah, this seems invalid to me - it makes no sense to add cinder containers on a host that's not meant to manage storage - do if the environment is correctly defined then this is a non-issue | 16:49 |
| *** cloudtrainme has joined #openstack-ansible | 16:49 | |
| cloudnull | ok , with nobody lobbying further on it, marked invalid. | 16:50 |
| cloudnull | thats it | 16:50 |
| cloudnull | does anyone else have anything they want to cover ? | 16:50 |
| rackertom | cloudnull: That's ok with me. I created the bug to track the fact that I saw the issue. | 16:50 |
| cloudnull | ok. so we're done here. have a good one everyone. | 16:51 |
| cloudnull | and thank you | 16:51 |
| *** itsuugo has quit IRC | 16:51 | |
| cloudnull | tiagogomes: you still having issues with horizon ? | 16:52 |
| cloudnull | is apache running within the container ? | 16:52 |
| *** gparaskevas has quit IRC | 17:04 | |
| mhayden | cloudnull: oops, i disappeared when 1496001 came up ;) | 17:04 |
| *** itsuugo has joined #openstack-ansible | 17:06 | |
| *** woodard has quit IRC | 17:17 | |
| *** arbrandes has quit IRC | 17:18 | |
| openstackgerrit | Major Hayden proposed openstack/openstack-ansible: Add SSL/TLS listener to RabbitMQ https://review.openstack.org/223717 | 17:26 |
| *** abitha has joined #openstack-ansible | 17:28 | |
| openstackgerrit | Major Hayden proposed openstack/openstack-ansible: Add SSL/TLS listener to RabbitMQ https://review.openstack.org/223717 | 17:29 |
| mhayden | Sam-I-Am: ^^ first time for me writing docs -- feel free to throw darts ;) | 17:30 |
| *** abitha has left #openstack-ansible | 17:31 | |
| *** phalmos has joined #openstack-ansible | 17:32 | |
| *** sdake_ is now known as sdake | 17:33 | |
| Sam-I-Am | mhayden: so this is how patches are supposed to work | 17:38 |
| Sam-I-Am | write the code, write the docs | 17:39 |
| Sam-I-Am | dont throw shit over the fence | 17:39 |
| mhayden | Sam-I-Am: wait, did i throw poo? | 17:40 |
| *** sdake_ has joined #openstack-ansible | 17:41 | |
| Sam-I-Am | mhayden: no, you didnt. most people just do DocImpact in commit messages and hope docs will a) understand the code b) find time to write docs | 17:41 |
| mhayden | ah okay | 17:42 |
| Sam-I-Am | inclusion of at least some docs makes me happy. i can look at it, +1 for readablity without digging through code | 17:42 |
| Sam-I-Am | maybe you want to hit the long list of DocImpact bugs? :) | 17:42 |
| mhayden | haha, well i was going to toss in some docs on how to quickly do an AIO build on rackspace cloud | 17:43 |
| mhayden | people keep asking me | 17:43 |
| *** phalmos has quit IRC | 17:43 | |
| Sam-I-Am | mhayden: dont we already have that? | 17:44 |
| Sam-I-Am | launch >= p1.8, then follow AIO instructions in-tree | 17:44 |
| * mhayden looks again | 17:44 | |
| *** sdake has quit IRC | 17:44 | |
| Sam-I-Am | they're not in the install guide specifically | 17:44 |
| Sam-I-Am | they're in the development-stack.rst or something | 17:44 |
| mhayden | Sam-I-Am: i was thinking more a one-command launch | 17:45 |
| mhayden | using user data | 17:45 |
| Sam-I-Am | ahhh | 17:46 |
| Sam-I-Am | yeah we dont have that stuff | 17:46 |
| * mhayden opens a bug | 17:46 | |
| *** itsuugo has quit IRC | 17:47 | |
| mhayden | Sam-I-Am: would that make sense as an appendix or somewhere in the install guide itself? | 17:49 |
| *** phalmos has joined #openstack-ansible | 17:53 | |
| BjoernT | palendae: this issue (1489451) is similar what we found in out kilo green field install where apt-transport-https was missing from the container and therefore basic packages could not be installed | 17:57 |
| openstackgerrit | Major Hayden proposed openstack/openstack-ansible: Small docs fix for adding compute https://review.openstack.org/223727 | 17:57 |
| Sam-I-Am | mhayden: hold on, grandma died | 17:57 |
| mhayden | Sam-I-Am: uh | 17:57 |
| *** woodard has joined #openstack-ansible | 18:06 | |
| Sam-I-Am | mhayden: probably works best in the development-stack document | 18:06 |
| mhayden | Sam-I-Am: that sounds good | 18:07 |
| *** woodard has quit IRC | 18:11 | |
| *** cloudtra_ has joined #openstack-ansible | 18:13 | |
| *** cloudtra_ has quit IRC | 18:13 | |
| *** cloudtrainme has quit IRC | 18:15 | |
| *** phalmos has quit IRC | 18:20 | |
| *** fawadkhaliq has quit IRC | 18:25 | |
| *** phalmos has joined #openstack-ansible | 18:31 | |
| openstackgerrit | Major Hayden proposed openstack/openstack-ansible: Add AIO build docs https://review.openstack.org/223742 | 18:33 |
| *** cloudtrainme has joined #openstack-ansible | 18:40 | |
| *** jmckind has quit IRC | 18:41 | |
| *** sdake_ is now known as sdake | 18:45 | |
| *** jmckind has joined #openstack-ansible | 18:47 | |
| *** phalmos has quit IRC | 18:48 | |
| *** woodard has joined #openstack-ansible | 18:53 | |
| *** woodard_ has joined #openstack-ansible | 18:55 | |
| *** woodard has quit IRC | 18:55 | |
| *** phalmos has joined #openstack-ansible | 19:04 | |
| openstackgerrit | Robb Romans proposed openstack/openstack-ansible: Add AIO build docs https://review.openstack.org/223742 | 19:07 |
| openstackgerrit | Matthew Thode proposed openstack/openstack-ansible-specs: Add spec for IPv6 support for projects https://review.openstack.org/221516 | 19:17 |
| prometheanfire | odyssey4me: cloudnull: let me know if I did that spec correctly, was told to put into the mikata folder when none existed | 19:18 |
| openstackgerrit | Major Hayden proposed openstack/openstack-ansible: Fixing broken networking link https://review.openstack.org/223783 | 19:44 |
| *** devlaps has joined #openstack-ansible | 19:51 | |
| openstackgerrit | venkatamahesh proposed openstack/openstack-ansible-specs: Added the home-page with openstack.org in setup.cfg https://review.openstack.org/223785 | 19:58 |
| *** jmckind has quit IRC | 20:04 | |
| openstackgerrit | Jesse Pretorius proposed openstack/openstack-ansible-specs: Add spec for IPv6 support for projects https://review.openstack.org/221516 | 20:09 |
| prometheanfire | thought I rebased... | 20:13 |
| prometheanfire | I know I fetched and rebased -i | 20:14 |
| * prometheanfire shrugs | 20:14 | |
| prometheanfire | odyssey4me: would have been nice not to stomp on my commit though :P | 20:14 |
| *** jwagner is now known as jwagner_lunch | 20:17 | |
| *** daneyon has joined #openstack-ansible | 20:17 | |
| *** jmckind has joined #openstack-ansible | 20:22 | |
| openstackgerrit | Major Hayden proposed openstack/openstack-ansible: Docs for named veths + troubleshooting https://review.openstack.org/223792 | 20:27 |
| *** keopp has joined #openstack-ansible | 20:27 | |
| openstackgerrit | Matthew Thode proposed openstack/openstack-ansible-specs: Add spec for IPv6 support for projects https://review.openstack.org/221516 | 20:32 |
| *** daneyon_ has joined #openstack-ansible | 20:36 | |
| *** jwagner_lunch is now known as jwagner | 20:36 | |
| *** daneyon has quit IRC | 20:39 | |
| openstackgerrit | Major Hayden proposed openstack/openstack-ansible: Adding docs for configuring horizon https://review.openstack.org/223804 | 20:50 |
| *** pradk has quit IRC | 20:59 | |
| *** jmckind has quit IRC | 21:04 | |
| logan2 | after running setup-hosts, I end up with a bunch of containers that have 69.20.0.164 and 69.20.0.196 set as resolvers, and it is not setting a default gateway in the eth1 br-mgmt interface. | 21:06 |
| logan2 | i was able to get a default route in the containers by using static_routes int he provider_networks, but the resolver issue I am having a lot of trouble with-- not seeing where that is coming from in the playbooks anywhere | 21:07 |
| *** woodard_ has quit IRC | 21:08 | |
| *** galstrom is now known as galstrom_zzz | 21:13 | |
| *** cloudtra_ has joined #openstack-ansible | 21:18 | |
| *** cloudtrainme has quit IRC | 21:19 | |
| *** Mudpuppy has quit IRC | 21:27 | |
| logan2 | nevermind regarding the resolvers issue.. I see where resolvconf is getting its configuration from now | 21:30 |
| *** cloudtra_ has quit IRC | 21:33 | |
| *** cloudtrainme has joined #openstack-ansible | 21:38 | |
| *** jmckind has joined #openstack-ansible | 21:39 | |
| *** phalmos has quit IRC | 21:47 | |
| *** jmckind has quit IRC | 21:57 | |
| *** keopp has quit IRC | 21:59 | |
| *** spotz is now known as spotz_zzz | 22:32 | |
| *** KLevenstein has quit IRC | 22:45 | |
| *** darrenc is now known as darrenc_afk | 22:52 | |
| *** sdake has quit IRC | 22:56 | |
| *** darrenc_afk is now known as darrenc | 23:06 | |
| *** Mudpuppy has joined #openstack-ansible | 23:06 | |
| *** markvoelker has quit IRC | 23:10 | |
| *** Mudpuppy has quit IRC | 23:13 | |
| *** Mudpuppy has joined #openstack-ansible | 23:14 | |
| *** openstackgerrit has quit IRC | 23:16 | |
| *** openstackgerrit has joined #openstack-ansible | 23:16 | |
| cloudnull | logan2: sorry for the poor responses most. | 23:23 |
| cloudnull | im glad you got figured. | 23:23 |
| cloudnull | feel free to ping out if something else comes up. | 23:23 |
| logan2 | well i am still confused. i see where it is coming from--the rackspace repo--but i thought that wasn't used if I specified repo hosts in the openstack_user_config | 23:32 |
| logan2 | the whole repo structure and buildout isn't really discussed in the docs. is that outside the scope of osad? I guess maybe I can just modify the rootfs image from rackspace repo, repackage, and feed the url to the lxc host setup playbook. is that the correct way to go about doing that? | 23:34 |
| *** cloudtrainme has quit IRC | 23:36 | |
| palendae | logan2: What are you trying to do? The repo structure will be built by the repo playbooks | 23:42 |
| palendae | But you're right, it's not documented, sadly | 23:42 |
| *** galstrom_zzz is now known as galstrom | 23:43 | |
| logan2 | i did notice the repo playbooks but yea I am not sure which ones to run and which order to run them in. at this point I am still just trying to get the containers bootstrapped with the correct network & resolver config etc. | 23:49 |
| *** alop has quit IRC | 23:53 | |
| *** galstrom is now known as galstrom_zzz | 23:56 | |
| *** markvoelker has joined #openstack-ansible | 23:56 | |
| *** BjoernT has quit IRC | 23:57 | |
Generated by irclog2html.py 2.14.0 by Marius Gedminas - find it at mg.pov.lt!