| *** rromans has quit IRC | 00:14 | |
| *** charz has quit IRC | 00:14 | |
| *** rromans has joined #openstack-ansible | 00:15 | |
| *** tobasco_ has quit IRC | 00:15 | |
| *** tobasco has joined #openstack-ansible | 00:16 | |
| *** charz has joined #openstack-ansible | 00:16 | |
| *** galstrom_zzz has quit IRC | 00:18 | |
| *** xar- has quit IRC | 00:19 | |
| *** galstrom_zzz has joined #openstack-ansible | 00:19 | |
| *** xar- has joined #openstack-ansible | 00:26 | |
| *** richoid has quit IRC | 00:28 | |
| *** shoutm_ has joined #openstack-ansible | 00:56 | |
| *** shoutm has quit IRC | 00:56 | |
| *** jwagner is now known as jwagner_away | 01:21 | |
| *** sdake has quit IRC | 01:36 | |
| *** mcarden has quit IRC | 01:40 | |
| *** mcarden has joined #openstack-ansible | 01:40 | |
| *** ybabenko has joined #openstack-ansible | 01:57 | |
| *** ybabenko has quit IRC | 02:01 | |
| *** sdake has joined #openstack-ansible | 02:28 | |
| *** sdake_ has joined #openstack-ansible | 02:31 | |
| *** k_stev has joined #openstack-ansible | 02:34 | |
| *** k_stev has quit IRC | 02:34 | |
| *** sdake has quit IRC | 02:34 | |
| *** shoutm_ has quit IRC | 03:20 | |
| *** eglute has quit IRC | 03:27 | |
| *** palendae has quit IRC | 03:27 | |
| *** bgmccollum has quit IRC | 03:27 | |
| *** dolphm has quit IRC | 03:29 | |
| *** sigmavirus24 has quit IRC | 03:30 | |
| *** d34dh0r53 has quit IRC | 03:30 | |
| *** jroll has quit IRC | 03:30 | |
| *** d34dh0r53 has joined #openstack-ansible | 03:31 | |
| *** dolphm has joined #openstack-ansible | 03:32 | |
| *** eglute has joined #openstack-ansible | 03:32 | |
| *** bgmccollum has joined #openstack-ansible | 03:32 | |
| *** sigmavirus24 has joined #openstack-ansible | 03:33 | |
| *** jroll has joined #openstack-ansible | 03:33 | |
| *** palendae has joined #openstack-ansible | 03:33 | |
| *** sigmavirus24 is now known as sigmavirus24_awa | 03:34 | |
| *** shoutm has joined #openstack-ansible | 03:41 | |
| *** sdake_ is now known as sdake | 03:42 | |
| *** stevemar has joined #openstack-ansible | 04:23 | |
| *** tlian has quit IRC | 04:24 | |
| *** shoutm has quit IRC | 04:38 | |
| *** stevemar has quit IRC | 04:43 | |
| *** shoutm has joined #openstack-ansible | 04:47 | |
| *** metral is now known as metral_zzz | 05:11 | |
| *** metral_zzz is now known as metral | 05:11 | |
| *** sdake_ has joined #openstack-ansible | 05:13 | |
| *** shausy has joined #openstack-ansible | 05:16 | |
| *** sdake has quit IRC | 05:17 | |
| *** shausy has quit IRC | 05:27 | |
| *** shoutm has quit IRC | 05:36 | |
| *** shoutm has joined #openstack-ansible | 05:37 | |
| openstackgerrit | Jesse Pretorius proposed stackforge/os-ansible-deployment: Update how neutron migrations are handled https://review.openstack.org/215584 | 05:45 |
|---|---|---|
| openstackgerrit | Jesse Pretorius proposed stackforge/os-ansible-deployment: Ensure rsync restarts fully during swift setup https://review.openstack.org/217341 | 05:45 |
| openstackgerrit | Jesse Pretorius proposed stackforge/os-ansible-deployment: Allow cinder-backup to use ceph https://review.openstack.org/209537 | 05:46 |
| openstackgerrit | Jesse Pretorius proposed stackforge/os-ansible-deployment: Adds the ability to provide user certificates to HAProxy https://review.openstack.org/215525 | 05:46 |
| openstackgerrit | Jesse Pretorius proposed stackforge/os-ansible-deployment: Add sorting_method to swift proxy config as needed https://review.openstack.org/208817 | 05:46 |
| openstackgerrit | Jesse Pretorius proposed stackforge/os-ansible-deployment: Add profiling for Ansible tasks https://review.openstack.org/216849 | 05:47 |
| *** sdake has joined #openstack-ansible | 05:48 | |
| *** sdake_ has quit IRC | 05:52 | |
| *** sdake_ has joined #openstack-ansible | 05:54 | |
| *** sdake has quit IRC | 05:57 | |
| *** shoutm has quit IRC | 05:58 | |
| openstackgerrit | Jesse Pretorius proposed stackforge/os-ansible-deployment: [WIP] Set the Ubuntu mirror used based on the environment https://review.openstack.org/218611 | 05:58 |
| *** shoutm has joined #openstack-ansible | 05:58 | |
| openstackgerrit | Jesse Pretorius proposed stackforge/os-ansible-deployment: Allow Horizon setup with external SSL termination https://review.openstack.org/214647 | 06:18 |
| *** shoutm_ has joined #openstack-ansible | 06:25 | |
| *** shoutm has quit IRC | 06:26 | |
| *** javeriak has joined #openstack-ansible | 06:36 | |
| *** javeriak has quit IRC | 06:41 | |
| *** javeriak has joined #openstack-ansible | 06:42 | |
| *** ybabenko has joined #openstack-ansible | 06:51 | |
| *** ybabenko has quit IRC | 06:56 | |
| *** ybabenko has joined #openstack-ansible | 06:56 | |
| openstackgerrit | Merged stackforge/os-ansible-deployment: Removed rpc-repo upstream pip deps https://review.openstack.org/216181 | 07:06 |
| openstackgerrit | Merged stackforge/os-ansible-deployment: Update tempest configuration https://review.openstack.org/210107 | 07:14 |
| evrardjp | mgariepy: Please tell me if you have issues with keepalived | 07:20 |
| *** gparaskevas has joined #openstack-ansible | 07:23 | |
| openstackgerrit | Jesse Pretorius proposed stackforge/os-ansible-deployment: Allow cinder-backup to use ceph https://review.openstack.org/209537 | 07:37 |
| *** javeriak has quit IRC | 07:57 | |
| *** Ti-mo has joined #openstack-ansible | 08:22 | |
| *** javeriak has joined #openstack-ansible | 08:25 | |
| openstackgerrit | Merged stackforge/os-ansible-deployment: Enable tempest testing of ceilometer https://review.openstack.org/217600 | 08:53 |
| *** d0ugal has left #openstack-ansible | 09:01 | |
| openstackgerrit | Jesse Pretorius proposed stackforge/os-ansible-deployment: [WIP] Set client versions deployed to use global requirements https://review.openstack.org/220049 | 09:28 |
| openstackgerrit | Jesse Pretorius proposed stackforge/os-ansible-deployment: [WIP] Set client versions deployed to use global requirements https://review.openstack.org/220050 | 09:34 |
| *** javeriak_ has joined #openstack-ansible | 09:38 | |
| *** javeriak has quit IRC | 09:39 | |
| odyssey4me | evrardjp FYI, considering you've been looking into this lately: https://bugs.launchpad.net/openstack-ansible/trunk/+bug/1424797 | 09:56 |
| openstack | Launchpad bug 1424797 in openstack-ansible trunk "Implement SSL for spice consoles" [Wishlist,Confirmed] | 09:56 |
| odyssey4me | you may wish to assign it to yourself | 09:56 |
| evrardjp | this isn't what I've done myself | 09:57 |
| *** mhayden has quit IRC | 09:58 | |
| odyssey4me | oh ok - I thought that you were :p | 09:58 |
| evrardjp | I've used a loadbalancer with SSL termination (here, haproxy) | 09:58 |
| evrardjp | and changed upstream code | 09:58 |
| evrardjp | of novnc | 09:58 |
| evrardjp | so thereisn't something to do | 09:58 |
| evrardjp | in OSAD | 09:58 |
| evrardjp | so I'm not really sure if this is required or working better than my workaround | 09:59 |
| evrardjp | in all cases, I don't think we should fix upstream code in osad | 10:00 |
| evrardjp | or implement workarounds | 10:00 |
| evrardjp | but I could share | 10:00 |
| odyssey4me | perhaps share your workaround in the bug for any that wish to use it | 10:00 |
| odyssey4me | but yes, if upstream doesn't support it then we can't do it until upstream is fixed | 10:01 |
| evrardjp | odyssey4me: you've found a solution for hpcloud? | 10:03 |
| evrardjp | let me rephrase that | 10:05 |
| odyssey4me | evrardjp it's an ongoing process - I'm just testing an optimisation right now which I hope helps | 10:05 |
| evrardjp | are you sure this is complete? https://review.openstack.org/#/c/218611/ | 10:05 |
| odyssey4me | hpcloud-b4 is still broken for us permanently | 10:05 |
| odyssey4me | evrardjp not sure just yet - the jury's still out on whether using the hpcloud az based ubuntu repositories is better than using ubuntu archive | 10:06 |
| evrardjp | I have variables in my user_variables: lxc_container_template_(main|security)_apt_repo | 10:06 |
| odyssey4me | ubuntu archive worked quite well for the previous set of tests | 10:06 |
| evrardjp | it could be useful to you | 10:06 |
| odyssey4me | yeah, the bootstrap-aio script already adds those variables into it | 10:06 |
| evrardjp | ok | 10:07 |
| evrardjp | I missed that part sorry | 10:07 |
| evrardjp | just trying to help ;) | 10:07 |
| odyssey4me | https://github.com/stackforge/os-ansible-deployment/blob/master/scripts/bootstrap-aio.sh#L357-L359 | 10:07 |
| odyssey4me | :) no worries - it's always good to have extra eyes | 10:08 |
| evrardjp | that's my opinion | 10:08 |
| *** ybabenko has quit IRC | 10:23 | |
| openstackgerrit | Merged stackforge/os-ansible-deployment: Allow cinder-backup to use ceph https://review.openstack.org/209537 | 10:24 |
| openstackgerrit | Matt Thompson proposed stackforge/os-ansible-deployment: Allow cinder-backup to use ceph https://review.openstack.org/220074 | 10:25 |
| *** mhayden has joined #openstack-ansible | 10:27 | |
| *** javeriak_ has quit IRC | 10:34 | |
| odyssey4me | evrardjp perhaps now's a good time to backport https://review.openstack.org/215579 ? | 10:38 |
| openstackgerrit | Jesse Pretorius proposed stackforge/os-ansible-deployment: Switch Nova/Tempest to use/test Cinder API v2 https://review.openstack.org/220081 | 10:38 |
| *** javeriak has joined #openstack-ansible | 10:39 | |
| openstackgerrit | Jesse Pretorius proposed stackforge/os-ansible-deployment: Implement tox.ini config for bashate and pep8 tests https://review.openstack.org/220082 | 10:39 |
| openstackgerrit | Jesse Pretorius proposed stackforge/os-ansible-deployment: Disable python buffering for gate checks https://review.openstack.org/220083 | 10:39 |
| *** javeriak has quit IRC | 10:51 | |
| openstackgerrit | Jesse Pretorius proposed stackforge/os-ansible-deployment: [WIP] Set client versions deployed to use global requirements https://review.openstack.org/220050 | 10:52 |
| openstackgerrit | Jesse Pretorius proposed stackforge/os-ansible-deployment: [WIP] Set client versions deployed to use global requirements https://review.openstack.org/220049 | 10:53 |
| openstackgerrit | Jesse Pretorius proposed stackforge/os-ansible-deployment: Set the Ubuntu mirror used based on the environment https://review.openstack.org/218611 | 10:54 |
| *** gparaskevas has quit IRC | 11:08 | |
| evrardjp | odyssey4me: cool, merged (: | 11:13 |
| openstackgerrit | Jean-Philippe Evrard proposed stackforge/os-ansible-deployment: Fixing haproxy-playbook fails when installing on multiple hosts https://review.openstack.org/220089 | 11:14 |
| openstackgerrit | Merged stackforge/os-ansible-deployment: Updated juno to include fix for CVE-2015-3241 - 26 Aug 2015 https://review.openstack.org/217098 | 11:15 |
| *** shoutm has joined #openstack-ansible | 11:17 | |
| *** shoutm_ has quit IRC | 11:18 | |
| *** gparaskevas has joined #openstack-ansible | 11:21 | |
| openstackgerrit | Merged stackforge/os-ansible-deployment: Remove hardcoded config drive enforcement https://review.openstack.org/218480 | 11:23 |
| *** ybabenko has joined #openstack-ansible | 11:32 | |
| *** ybabenko has quit IRC | 11:32 | |
| mgariepy | good morning ! | 12:02 |
| mgariepy | evrardjp, I don't have issue with keepalived other than i think it need different priority when deploying on 3 hosts. | 12:03 |
| evrardjp | what do you mean mgariepy? | 12:05 |
| evrardjp | 2 backups with different priorities? | 12:05 |
| mgariepy | yep | 12:05 |
| evrardjp | it's doable | 12:05 |
| mgariepy | i override with hosts_var | 12:05 |
| evrardjp | you can define it in conf.d/keepalived.yml | 12:06 |
| mgariepy | http://paste.ubuntu.com/12262483/ | 12:06 |
| evrardjp | yup, it works the same way | 12:06 |
| evrardjp | and it doesn't the change into account? | 12:06 |
| mgariepy | yeah it works. | 12:07 |
| evrardjp | oh ok | 12:07 |
| evrardjp | then it's fine :) | 12:07 |
| evrardjp | I'll remove the WIP tag on it then | 12:07 |
| openstackgerrit | Jean-Philippe Evrard proposed stackforge/os-ansible-deployment: Implementation of keepalived for haproxy https://review.openstack.org/218818 | 12:08 |
| mhayden | morning | 12:20 |
| *** woodard has joined #openstack-ansible | 12:27 | |
| odyssey4me | o/ mhayden i see you had some fun in the sewers of linux/lxc networking yesterday :p | 12:28 |
| *** cfarquhar has joined #openstack-ansible | 12:28 | |
| *** cfarquhar has quit IRC | 12:28 | |
| *** cfarquhar has joined #openstack-ansible | 12:28 | |
| odyssey4me | hughsaunders how do you feel about https://review.openstack.org/220074 ? | 12:29 |
| hughsaunders | odyssey4me: warm and fuzzy | 12:29 |
| odyssey4me | andymccr I think it may be time to review https://review.openstack.org/217341 | 12:29 |
| mhayden | odyssey4me: yes, had to take a long shower last night | 12:29 |
| andymccr | odyssey4me: you already have 2+2's ? | 12:30 |
| mhayden | odyssey4me: i have a lxc patch that seems to be working so far, but needs more testing | 12:30 |
| mhayden | might propose it as a PR to the upstream shortly to get them to gander | 12:30 |
| odyssey4me | andymccr doh, wrong one: https://review.openstack.org/208817 | 12:30 |
| odyssey4me | mhayden nice :) | 12:31 |
| *** woodard has quit IRC | 12:39 | |
| *** woodard has joined #openstack-ansible | 12:43 | |
| *** pradk has quit IRC | 12:45 | |
| *** shoutm_ has joined #openstack-ansible | 12:47 | |
| *** shoutm has quit IRC | 12:48 | |
| evrardjp | nice indeed :) | 13:05 |
| *** tlian has joined #openstack-ansible | 13:06 | |
| evrardjp | mgariepy: just a quick question about your usage of keepalived: you're using on bare metal, right? because I've seen lots of issues in the past with running this kind of sofware virtualized/containerized | 13:07 |
| mgariepy | evrardjp, yep indeed on metal. | 13:13 |
| evrardjp | ok perfect | 13:13 |
| cloudnull | morning | 13:20 |
| evrardjp | o/ cloudnull | 13:28 |
| mhayden | odyssey4me: the PR is in --> https://github.com/lxc/lxc/pull/649 | 13:30 |
| odyssey4me | mhayden sweet, subscribed :) | 13:30 |
| cloudnull | nice mhayden! | 13:31 |
| mhayden | forgot semicolons -- had to rebase a quick fix before i really embarrassed myself | 13:33 |
| mhayden | i still wish i could figure out where the actual veth teardown was going wrong | 13:35 |
| mhayden | that would take some kernel digging | 13:35 |
| mhayden | would need additional coffee for that | 13:35 |
| cloudnull | typie typie ;) | 13:35 |
| mhayden | there has to be some kind of kernel queue for netlink where devices are removed | 13:36 |
| mattt | mhayden: didn't you review a linux kernel hacking book rcently? :P | 13:36 |
| mhayden | mattt: already checked that book -- no netlink mentions :) | 13:36 |
| odyssey4me | cloudnull and others - please don't +w any patches until https://review.openstack.org/220074 is merged :) | 13:38 |
| odyssey4me | if any patch ahead of it fails, it restarts the build process - so things take awfully long | 13:38 |
| cloudnull | only for kilo it seems | 13:39 |
| odyssey4me | cloudnull yup, but unfortunately the zuul chain doesn't seem to differentiate | 13:40 |
| odyssey4me | it queues all patches equally, then resets the queue if one fails in the gate pipeline | 13:40 |
| cloudnull | so your saying a failure in master is causing kilo to regate ? | 13:40 |
| odyssey4me | yup | 13:40 |
| cloudnull | is infra fixing that ? | 13:41 |
| cloudnull | because that seems wrong | 13:41 |
| odyssey4me | I haven't spoken to anyone about it yet - but yes, it seems to me like a bug | 13:41 |
| odyssey4me | I only noticed it yesterday when our gate pipeline was very, very long - and got reset many, many times. | 13:42 |
| odyssey4me | we had some patches in there for over 4 hours when I stopped watching. | 13:42 |
| *** Bjoern_ has joined #openstack-ansible | 13:42 | |
| svg | Seems Packt Publishing found me again for a round. No thanks :) | 13:52 |
| *** k_stev has joined #openstack-ansible | 13:53 | |
| *** pradk has joined #openstack-ansible | 13:54 | |
| cloudnull | svg the author ! | 13:54 |
| *** jroll has quit IRC | 13:54 | |
| *** jroll has joined #openstack-ansible | 13:54 | |
| *** woodard has quit IRC | 13:55 | |
| svg | cloudnull: alas no, only reviewer credits! in the book! on the website! | 13:55 |
| *** KLevenstein has joined #openstack-ansible | 13:57 | |
| *** woodard has joined #openstack-ansible | 13:57 | |
| *** galstrom_zzz is now known as galstrom | 13:58 | |
| Bjoern_ | what's going on with pakt ? They bother me too | 13:59 |
| *** sigmavirus24_awa is now known as sigmavirus24 | 13:59 | |
| *** Bjoern_ is now known as BjoernT | 13:59 | |
| openstackgerrit | Jesse Pretorius proposed stackforge/os-ansible-deployment: Updated juno for new dev work - 3 Sep 2015 https://review.openstack.org/220151 | 14:02 |
| *** Mudpuppy has joined #openstack-ansible | 14:03 | |
| *** gparaskevas has quit IRC | 14:03 | |
| *** spotz_zzz is now known as spotz | 14:05 | |
| palendae | BjoernT: I think they're nagging svg about an ansible book review | 14:11 |
| *** woodard has quit IRC | 14:12 | |
| svg | yeah, I already told them last year to no bother me with that. They're not very professional. | 14:12 |
| palendae | Nope | 14:12 |
| palendae | That's kind of their MO | 14:12 |
| *** shoutm_ has quit IRC | 14:18 | |
| *** woodard has joined #openstack-ansible | 14:21 | |
| *** galstrom is now known as galstrom_zzz | 14:22 | |
| *** cloudtrainme has joined #openstack-ansible | 14:23 | |
| *** jmckind has joined #openstack-ansible | 14:27 | |
| arbrandes1 | Hey guys. So, I'm trying to use Ceph with Cinder. On openstack_user_config.yml, how can I define "cinder_backends" only once for all "storage_hosts"? | 14:31 |
| openstackgerrit | Merged stackforge/os-ansible-deployment: Add sorting_method to swift proxy config as needed https://review.openstack.org/208817 | 14:32 |
| openstackgerrit | Merged stackforge/os-ansible-deployment: Allow cinder-backup to use ceph https://review.openstack.org/220074 | 14:32 |
| arbrandes1 | Yay, I was waiting for that one! ^^^ | 14:33 |
| *** devlaps has quit IRC | 14:33 | |
| *** arbrandes1 is now known as arbrandes | 14:33 | |
| odyssey4me | right, time to tag 11.2.0 :) | 14:34 |
| arbrandes | Woohooo! \o/ \o/ \o/ | 14:34 |
| odyssey4me | arbrandes I might be wrong, but if you're using ceph only, then you should only have one cinder-volume and one cinder-scheduler container (change the cinder-volume to is_metal: false) | 14:39 |
| arbrandes | odyssey4me: no redundancy for cinder-volume? | 14:40 |
| *** k_stev1 has joined #openstack-ansible | 14:42 | |
| *** k_stev has quit IRC | 14:42 | |
| arbrandes | * [new tag] 11.2.0 -> 11.2.0 | 14:43 |
| arbrandes | \o/ \o/ \o/ \o/ \o/ \o/ \o/ | 14:43 |
| odyssey4me | arbrandes well, technically cinder-volume is not redundant in cinder anyway.... but with ceph and in most cases cinder-volume only facilitates the initial connection to the storage | 14:44 |
| odyssey4me | from then on its direct | 14:44 |
| odyssey4me | again, I could be wrong - but that's my understanding | 14:44 |
| openstackgerrit | Kevin Carter proposed stackforge/os-ansible-deployment: [WIP] Compartmentalizing RabbitMQ https://review.openstack.org/202822 | 14:44 |
| openstackgerrit | Kevin Carter proposed stackforge/os-ansible-deployment: Updated MariaDB to the new release version (10.0) https://review.openstack.org/178259 | 14:44 |
| arbrandes | odyssey4me: that's true - I'm just worried about having some redundancy in the sense that if the storage_host running cinder-volume goes down, I'd like to have another option running. | 14:45 |
| odyssey4me | arbrandes if you're only using ceph then I think your storage hosts should all be the same hosts as your controllers | 14:46 |
| odyssey4me | andymccr hughsaunders git-harry is that correct? | 14:46 |
| arbrandes | odyssey4me: yup, that's what I have configured. I may be misuderstanding exactly HOW to configure this, though. | 14:46 |
| andymccr | arbrandes: if you are using cinder-volumes with only a ceph backend (and not an lvm backend) then you can put it in containers and then run it on the infra hosts | 14:47 |
| arbrandes | andymccr: yeah, that's my intention. | 14:48 |
| arbrandes | andymccr: my doubt is regarding how to best configure this in openstack_user_config.yml. | 14:48 |
| andymccr | the only change to be aware of is in your /etc/openstack_deploy/env.d/cinder.yml you will need to change the cinder-volumes container's "is_metal" setting to false | 14:49 |
| mattt | arbrandes: can you specify that in user_variables.yml ? | 14:49 |
| mattt | arbrandes: but if you do ever want to add another backend then you'll be SOL | 14:49 |
| arbrandes | andymccr: I have three hosts (alice, bob, and charlie) that will handle infra, including cinder-volume. I set "storage-infra_hosts" to those three. Next, from the examples, it seems I have to set "storage_hosts" as well, where I'm configuring the cinder_backends property. | 14:49 |
| arbrandes | mattt: no other backend intended. | 14:50 |
| arbrandes | andymccr: ok, thanks, I'll take care with is_metal = false. :) | 14:50 |
| mattt | arbrandes: my personal opinion is to just duplicate it for each storage host | 14:51 |
| mattt | arbrandes: it gives you more flexibility down the line | 14:51 |
| arbrandes | mattt: Alright, I can live with that. I was just wondering if there was a cleaner way. user_variables might be the place, if the playbooks support it. If they don't, it's not a biggie. | 14:52 |
| mattt | arbrandes: sec | 14:52 |
| mattt | arbrandes: just chatting to someone about this right now | 14:52 |
| *** javeriak has joined #openstack-ansible | 14:52 | |
| openstackgerrit | Merged stackforge/os-ansible-deployment: Ensure rsync restarts fully during swift setup https://review.openstack.org/217341 | 14:52 |
| arbrandes | mattt: awesome, thanks | 14:53 |
| mattt | arbrandes: chatting to git-harry about this, ideally you should only run one cinder-volume for each backend | 14:53 |
| mattt | arbrandes: if that cinder-volume service goes down, you'll just have issues w/ scheduling i believe | 14:54 |
| arbrandes | mattt: ok. Any trouble if I have two, though? I know OpenStack is fine about this - just wondering if the playbooks are too. | 14:54 |
| mattt | arbrandes: he actually said that the limitation is from an openstack perspective, not os-ansible-deployment | 14:55 |
| mattt | arbrandes: os-ansible-deployment will let you do it as far as i can tell | 14:55 |
| odyssey4me | arbrandes user_variables have the highest precedence | 14:56 |
| arbrandes | mattt: well, I know of a couple of clusters that, in switching over to Ceph-only, stuck with running a cinder-volume instance on *every* compute node. OpenStack might not be using all those, but I know it's not an issue in and of itself. | 14:56 |
| palendae | A request - please do not take people's in-flight reviews without at least checking with them first. | 14:56 |
| mattt | palendae: agreed | 14:56 |
| arbrandes | odyssey4me: cool, thank you | 14:57 |
| *** jbweber has joined #openstack-ansible | 14:57 | |
| d34dh0r53 | odyssey4me: cloudnull: any reason https://review.openstack.org/#/c/219105/5 wasn't +W? | 14:57 |
| odyssey4me | d34dh0r53 to get more reviewers ;) | 14:58 |
| mattt | arbrandes: yeah i'm not entirely sure about that part, but osad should allow you to do it without an issue | 14:58 |
| palendae | Based on what meetings I have scheduled today I'll try to hide and get another review on that | 14:58 |
| mattt | arbrandes: the only thing to note is if you set a blanket override in user_variables.yml, you'll never to do a per-node config in openstack_user_config.yml as the user_variables.yml file will always take precedence | 14:59 |
| mattt | arbrandes: i got burned w/ this when having a blanket drive configuration for ceph in user_variables.yml, and then wanted a single node to have a custom config | 15:00 |
| odyssey4me | mattt hughsaunders andymccr d34dh0r53 cloudnull palendae juno sha bump ready for review: https://review.openstack.org/220151 | 15:00 |
| mhayden | thanks for the w+1 on the named-veths spec, everybody -- that's my first ever openstack spec! :) | 15:01 |
| openstackgerrit | Merged stackforge/os-ansible-deployment-specs: Adding spec for named veth interfaces https://review.openstack.org/219105 | 15:01 |
| palendae | Argh! I had a comment on that | 15:01 |
| palendae | Should mention that doing this will cause restarts to fail should there be more dangling veths | 15:01 |
| palendae | cloudnull: ^ | 15:01 |
| mhayden | palendae: want me to go and revise the spec right quick? | 15:01 |
| mhayden | i can submit a follow-on review | 15:01 |
| odyssey4me | palendae you can still comment on it | 15:01 |
| palendae | cloudnull: Named veths will *not* re-attach when the container restarts. It'll prevent it from restarting at all | 15:02 |
| *** jwagner_away is now known as jwagner | 15:02 | |
| palendae | odyssey4me: If gerritt will let me sign in, yeah :) | 15:02 |
| arbrandes | mattt: yeah, that's sound advice. However, I doubt I'll ever have to change the configuration of an individual instance of cinder-volume. In any case, how would I set container_vars for all storage_hosts in user_variables.yml? | 15:02 |
| cloudnull | palendae: maybe i missunderstood the other day, but mhayden you wern't able to reproduce the issue with the named veths right? | 15:03 |
| mattt | arbrandes: i'd imagine you can just set cinder_backends: in user_variables.yml, but untested by me :P | 15:03 |
| palendae | cloudnull: No, he wasn't - the issue is different. It hard-fails, the container just won't start | 15:03 |
| mhayden | cloudnull: if the box was under heavy contention, i could reproduce :| | 15:03 |
| arbrandes | mattt: ok, I'll try that and come back with results. :) | 15:03 |
| mhayden | waiting on osad to build out on a new 8GB server so i can test the LXC fix | 15:03 |
| mattt | arbrandes: but obviously those vars will not be limited to the storage nodes, however i don't think it'd be a problem other nodes having those vars | 15:04 |
| arbrandes | mattt: yeah, I figured. | 15:04 |
| cloudnull | IMO hard fail is better than system overload due dangling veth. | 15:04 |
| palendae | cloudnull: Right, and I agree. I just want to be very clear that this is what will happen | 15:04 |
| mattt | arbrandes: but honestly i'd just bung it all in openstack_user_config.yml, more flexible that way | 15:04 |
| palendae | Not that it'll fix it and it's all better | 15:04 |
| mhayden | cloudnull: there's also an opportunity to force a cleanup on container start within the LXC code | 15:05 |
| mhayden | but that's lower priority for me at the moment | 15:05 |
| mhayden | there's already some network cleanup code in lxc-start that gets used when the container start explodes for other reasons | 15:05 |
| palendae | Yeah, and who knows if that'll get into the ubuntu LTS packages | 15:05 |
| arbrandes | mattt: that's the first thing I'll try, of course. Just to see it working. Next, I'll try consolidating it. | 15:05 |
| mattt | arbrandes: cool, let us know how you get on | 15:07 |
| mhayden | palendae: if the fix gets merged into lxc, i'll go open tickets for a backport into LTS | 15:07 |
| *** markvoelker has joined #openstack-ansible | 15:08 | |
| mhayden | but most of my ubuntu tickets get a WONTFIX :P | 15:08 |
| mattt | arbrandes: how are you deploying ceph itself, out of curiousity ? | 15:09 |
| odyssey4me | palendae 11.2.0 is released, and the upstream wheel repo is updated | 15:09 |
| palendae | odyssey4me: Thanks | 15:09 |
| arbrandes | mattt: ceph-ansible | 15:09 |
| mattt | arbrandes: ok, same w/ us | 15:10 |
| arbrandes | mattt: it's not the prettiest thing in the universe, but it seems to work. | 15:10 |
| *** sigmavirus24 is now known as sigmavirus24_awa | 15:12 | |
| *** sigmavirus24_awa is now known as sigmavirus24 | 15:13 | |
| *** phalmos has joined #openstack-ansible | 15:13 | |
| odyssey4me | palendae cloudnull erm, oops: https://github.com/stackforge/os-ansible-deployment/blob/11.2.0/playbooks/inventory/group_vars/all.yml#L17 | 15:17 |
| cloudnull | 11.2.1 here we come :) | 15:17 |
| mgariepy | mhayden, maybe you can go poke stgraber on #lxcontainers about your fix and in the same time ask him to fix it in ubuntu LTS | 15:17 |
| cloudnull | odyssey4me: it happens | 15:17 |
| odyssey4me | haha, yeah | 15:17 |
| palendae | So does stgraber not use #lxc? | 15:18 |
| cloudnull | update-revision.sh in the various branches helps with that | 15:18 |
| palendae | Oh, I just read the topic on #lxc, nm | 15:18 |
| mgariepy | hehe | 15:19 |
| mhayden | mgariepy: not 100% sure that my fix is totally solid -- will hopefully test in < 15-30 minutes | 15:21 |
| mhayden | i'd rather ping stgraber when i know it does something good | 15:21 |
| mhayden | it doesn't break anything, and the code compiles -- i know that much ;) | 15:21 |
| openstackgerrit | Jesse Pretorius proposed stackforge/os-ansible-deployment: Update tag version to 11.2.1 https://review.openstack.org/220195 | 15:21 |
| odyssey4me | cloudnull ah, I didn't know about that script! | 15:22 |
| mgariepy | he might have an idea of what is wrong (or know who is) | 15:22 |
| odyssey4me | cloudnull ^ for fast follow release | 15:22 |
| odyssey4me | hughsaunders ^ | 15:22 |
| palendae | We should probably doc our tagging process ;) | 15:23 |
| cloudnull | odyssey4me: yea update-revision.sh and sources-branch-updater.sh are in all branches to have a common interface to update all the various parts as we've moved bits around. | 15:23 |
| cloudnull | odyssey4me: +2 | 15:23 |
| odyssey4me | cloudnull yeah, I knew about the second, but not the first | 15:24 |
| *** javeriak has quit IRC | 15:25 | |
| *** mcarden has quit IRC | 15:29 | |
| *** mcarden has joined #openstack-ansible | 15:30 | |
| evrardjp | mattt: I'm just back on the conversation, tried to read the reasons why not to have multiple storage-hosts having the same backend using ceph... didn't get it | 15:31 |
| mattt | evrardjp: yeah, i'm just passing over 2nd hand information | 15:32 |
| evrardjp | isn't the scheduler checking first if one cinder-volume is available among all the cinder-volumes? if true, I don't see something wrong having 2 cinder-volumes with ceph | 15:33 |
| evrardjp | it prevents the failure of scheduling if one cinder-volume fails | 15:33 |
| evrardjp | but I may be mistaken | 15:34 |
| *** alop has joined #openstack-ansible | 15:34 | |
| odyssey4me | evrardjp yes, that's my understanding | 15:35 |
| odyssey4me | however, once a volume is attached to an instance the cinder-volume service is not involved unless it needs to re-attach (after an instance reboot or something) | 15:35 |
| odyssey4me | unfortunately that reattachment process has to use the same cinder-volume host | 15:36 |
| odyssey4me | so if it goes down, your instance can't reattach | 15:36 |
| odyssey4me | (unless you force detach and re-attach which will then use a new cinder-volume host) | 15:36 |
| evrardjp | mmm didn't try that | 15:38 |
| evrardjp | that's why I didn't got the issue | 15:38 |
| evrardjp | get* | 15:39 |
| odyssey4me | give it a go :) | 15:39 |
| mattt | evrardjp: aren't volumes tied to a specific cinder-volume tho? | 15:39 |
| mattt | evrardjp: so if you have multiple hosts running cinder-volume, they'd all register differently | 15:39 |
| evrardjp | you are both right | 15:39 |
| evrardjp | >< | 15:39 |
| mattt | git-harry knows the most about this, unfortunately he doesn't like to type | 15:40 |
| odyssey4me | the only way to recover from a broken cinder-volume container/host (assuming a back-end like ceph) is to rename the cinder-volume host entry for the volumes in the DB | 15:40 |
| evrardjp | but that's not the same | 15:41 |
| evrardjp | that's recover from a crash of your cinder-volume | 15:41 |
| evrardjp | I just meant to improve availability of my system, but for example rebooting one cinder-volume at a time | 15:42 |
| *** markvoelker_ has joined #openstack-ansible | 15:42 | |
| evrardjp | in case that every happens | 15:42 |
| evrardjp | ever* | 15:42 |
| mattt | evrardjp: my cinder knowledge is limited but you'll still have problems there | 15:42 |
| evrardjp | so your cinder-volume will be down, you won't be able to schedule anything new on this | 15:42 |
| evrardjp | you'll schedule on the other ones | 15:43 |
| odyssey4me | evrardjp you get no gain from that except for api requests at that particular moment in time | 15:43 |
| mattt | evrardjp: yep, and anyone w/ a volume 'attached' to that node is going to get an error if they attempt to do anything | 15:43 |
| odyssey4me | mattt evrardjp only api requests | 15:43 |
| mattt | yep through API | 15:43 |
| evrardjp | it's a smaller issue than having EVERYONE that can't do anything with a volume | 15:43 |
| evrardjp | it's just a question of sla | 15:43 |
| mattt | evrardjp: that is true | 15:43 |
| odyssey4me | yeah, fair enough | 15:44 |
| odyssey4me | it's not high availability - it's just greater api availability | 15:44 |
| evrardjp | and yes, it's only api requests, but that's where we put sla | 15:44 |
| evrardjp | :p | 15:44 |
| mattt | unless of course you find that one cinder-volume always seems to grab messages first | 15:44 |
| mattt | at which point you're back to square 1 :) | 15:44 |
| odyssey4me | yeah, the sceduler could probably do with a round-robin filter or something | 15:45 |
| openstackgerrit | Jesse Pretorius proposed stackforge/os-ansible-deployment: Update kilo for new dev work - 3 Sep 2015 https://review.openstack.org/220206 | 15:45 |
| evrardjp | I'll need to look at cinder more cautiously | 15:45 |
| *** markvoelker has quit IRC | 15:45 | |
| palendae | evrardjp: That's a very reasonable approach | 15:45 |
| mattt | evrardjp: i believe git-harry said you could set all cinder-volume hosts to have the same hostname | 15:45 |
| mattt | evrardjp: he's our cinder SME, but he's shy | 15:45 |
| odyssey4me | mattt if you do that, then you'll only ever be wanting to run one cinder-volume service at a time | 15:46 |
| evrardjp | I'm looking for a way to avoid that ALL my tenants can't do something with cinder at one moment | 15:46 |
| odyssey4me | so cinder-volume would need to fail over between hosts | 15:46 |
| evrardjp | what we are talking about is doing real HA on cinder-volume ;) | 15:47 |
| mattt | git-harry: think i found a picture of you as a kid: http://image.shutterstock.com/display_pic_with_logo/732076/732076,1305444102,2/stock-photo-shy-boy-77485684.jpg | 15:47 |
| *** phalmos has quit IRC | 15:47 | |
| palendae | evrardjp: That's a good idea | 15:48 |
| hughsaunders | I'm not so sure, why not use something that is already HA? | 15:49 |
| hughsaunders | ah, I misunderstood. | 15:50 |
| *** phalmos has joined #openstack-ansible | 15:56 | |
| palendae | odyssey4me: Are the 11.2.1 mirror builds done? | 16:00 |
| odyssey4me | palendae nope - busy setting that up now | 16:00 |
| palendae | odyssey4me: Ok, let me know when that happens if you could please | 16:01 |
| *** phalmos has quit IRC | 16:01 | |
| *** phalmos has joined #openstack-ansible | 16:01 | |
| openstackgerrit | Kevin Carter proposed stackforge/os-ansible-deployment: Create config_template plugin https://review.openstack.org/220212 | 16:02 |
| odyssey4me | cloudnull meeting? | 16:02 |
| odyssey4me | meeting in #openstack-meeting-4 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 | 16:04 |
| cloudnull | ^ beat me to the paste :) | 16:04 |
| Sam-I-Am | we're all tied up in a meeting | 16:04 |
| mhayden | bam | 16:04 |
| Sam-I-Am | pretty much the same meeting as the last 3 weeks of osad meetings | 16:05 |
| mhayden | currently have a testing loop rolling for the lxc fix... | 16:05 |
| cloudnull | Sam-I-Am: if you can comment here you can join the meeting in #openstack-meeting-4 | 16:05 |
| Sam-I-Am | cloudnull: i'm there | 16:06 |
| Sam-I-Am | as much as i can be | 16:06 |
| Sam-I-Am | also looking for a job as a truck driver | 16:06 |
| cloudnull | hahaha | 16:06 |
| *** jaypipes has joined #openstack-ansible | 16:07 | |
| *** phalmos has quit IRC | 16:38 | |
| evrardjp | are we fully ready for keystone v3? | 16:42 |
| odyssey4me | evrardjp we're waiting for upstream | 16:43 |
| evrardjp | I'll stick with v2 for now :p | 16:43 |
| odyssey4me | we've had a patch in flight for full conversion to it ages ago | 16:43 |
| odyssey4me | we should be able to switch to it as the default for liberty | 16:44 |
| evrardjp | I'm still reluctant to define the variable that pushes the usage to v3 | 16:44 |
| evrardjp | ok | 16:44 |
| *** k_stev1 has quit IRC | 16:47 | |
| *** k_stev has joined #openstack-ansible | 16:50 | |
| openstackgerrit | Kevin Carter proposed stackforge/os-ansible-deployment: [WIP] Replaced the copy_update module https://review.openstack.org/216790 | 16:52 |
| openstackgerrit | Kevin Carter proposed stackforge/os-ansible-deployment: Create config_template plugin https://review.openstack.org/220212 | 16:52 |
| prometheanfire | Sam-I-Am: docs link to generated osad docs? | 16:53 |
| openstackgerrit | Kevin Carter proposed stackforge/os-ansible-deployment: Create config_template plugin https://review.openstack.org/220212 | 16:53 |
| prometheanfire | Sam-I-Am: want to give a link to the ops-meetup page | 16:53 |
| Sam-I-Am | prometheanfire: they're on readthedocs | 16:53 |
| prometheanfire | ah, right | 16:53 |
| Sam-I-Am | and supposedly broken right now | 16:53 |
| Sam-I-Am | somehow | 16:54 |
| prometheanfire | https://osad.readthedocs.org/en/latest/ | 16:54 |
| prometheanfire | ya, 404 or something | 16:54 |
| Sam-I-Am | some patch broke it | 16:54 |
| Sam-I-Am | and i dont have time to figure out why | 16:54 |
| Sam-I-Am | because upgrades | 16:54 |
| prometheanfire | ya | 16:54 |
| prometheanfire | ok | 16:55 |
| prometheanfire | just github/lp links then | 16:55 |
| odyssey4me | Sam-I-Am prometheanfire http://os-ansible-deployment.readthedocs.org/ | 17:01 |
| odyssey4me | just the lp link should be fine - as it links to everything | 17:02 |
| odyssey4me | ie https://launchpad.net/openstack-ansible | 17:02 |
| odyssey4me | maybe the wiki is better though: https://wiki.openstack.org/wiki/OpenStackAnsible | 17:03 |
| odyssey4me | palendae the python wheels are build for 11.2.1 and 11.2.2 in the upstream repo - we're just waiting for the merges of https://review.openstack.org/220195 and then https://review.openstack.org/220206 | 17:04 |
| odyssey4me | stevelle it's probably a good idea to go ahead and backport https://review.openstack.org/217341 and https://review.openstack.org/208817 | 17:07 |
| *** jwagner is now known as jwagner_away | 17:07 | |
| stevelle | odyssey4me: thanks for the rebases on them while I was lost, btw | 17:08 |
| odyssey4me | stevelle yeah, it was a right pain to try and get things back on track for master - happy that we're getting patches through the door again though | 17:08 |
| stevelle | still finding my way out of the woods, as I seem to have missed a lot. Hope to get clean backports in this afternoon | 17:09 |
| *** jwagner_away is now known as jwagner | 17:10 | |
| openstackgerrit | Kevin Carter proposed stackforge/os-ansible-deployment: Adds the crud_template to ceilometer https://review.openstack.org/217030 | 17:13 |
| *** markvoelker_ has quit IRC | 17:16 | |
| openstackgerrit | Kevin Carter proposed stackforge/os-ansible-deployment: Adds the config_template to ceilometer https://review.openstack.org/217030 | 17:19 |
| palendae | odyssey4me: Ok, but shoudln't need the 2nd one for 11.2.1 to be tagged, right? | 17:28 |
| palendae | Obviously necessary for work to carry on for 11.2.2 | 17:30 |
| *** kerwin_bai has joined #openstack-ansible | 17:33 | |
| *** phalmos has joined #openstack-ansible | 17:35 | |
| *** kerwin_bai has quit IRC | 17:35 | |
| *** javeriak has joined #openstack-ansible | 17:58 | |
| openstackgerrit | Steve Lewis proposed stackforge/os-ansible-deployment: Add sorting_method to swift proxy config as needed https://review.openstack.org/220260 | 18:03 |
| openstackgerrit | Steve Lewis proposed stackforge/os-ansible-deployment: Ensure rsync restarts fully during swift setup https://review.openstack.org/220261 | 18:04 |
| *** phalmos has quit IRC | 18:05 | |
| openstackgerrit | Merged stackforge/os-ansible-deployment: Update how neutron migrations are handled https://review.openstack.org/215584 | 18:09 |
| openstackgerrit | Merged stackforge/os-ansible-deployment: Implement tox.ini config for bashate and pep8 tests https://review.openstack.org/220082 | 18:10 |
| *** woodard has quit IRC | 18:19 | |
| *** k_stev has quit IRC | 18:30 | |
| *** harlowja has quit IRC | 18:31 | |
| *** tlian has quit IRC | 18:31 | |
| *** harlowja has joined #openstack-ansible | 18:32 | |
| *** phalmos has joined #openstack-ansible | 18:34 | |
| openstackgerrit | Kevin Carter proposed stackforge/os-ansible-deployment: Shift irqbalance package from lxc_hosts to openstack_hosts https://review.openstack.org/218354 | 18:35 |
| odyssey4me | palendae yep, that's for 11.2.2 work | 18:36 |
| openstackgerrit | Kevin Carter proposed stackforge/os-ansible-deployment: Additional retries for ssh wait check https://review.openstack.org/219638 | 18:36 |
| openstackgerrit | Kevin Carter proposed stackforge/os-ansible-deployment: Adds the ability to provide user certificates to HAProxy https://review.openstack.org/215525 | 18:37 |
| odyssey4me | stevelle you can rebase https://review.openstack.org/220261 and https://review.openstack.org/220260 on top of https://review.openstack.org/220206 to respect the order :) | 18:38 |
| *** tlian has joined #openstack-ansible | 18:38 | |
| odyssey4me | cloudnull it looks like many of the projects are tagging their liberty-3 tags today, so I'll prep a liberty-3 related sha update tomorrow when they're all settled | 18:41 |
| cloudnull | yea i'd imagine we'll be in queue for another round of all the things are broken real soon. | 18:43 |
| odyssey4me | cloudnull fyi - I've created the 11.3.0 and 10.2.0 milestones to add any new bugs to. They're not dated, so they're a good catch-all for any new bugs which we can pull back to the appropriate milestone when someone actually picks up the patch | 18:43 |
| odyssey4me | that way we're not making any false promises by adding it to a milestone, then shifting it every time we release the milestone | 18:44 |
| cloudnull | okiedokie | 18:44 |
| odyssey4me | all master bugs need to now go to the 12.0.0 milestone as there won't be any more until then - if we haven't tackled them by then then we'll have to shift them to 12.1.0 | 18:45 |
| odyssey4me | it is nice having a milestone to tag for trunk though :) | 18:45 |
| cloudnull | it is, allows us to keep better track of things for sure. | 18:48 |
| odyssey4me | cloudnull how do you like https://review.openstack.org/218611 ? | 18:50 |
| odyssey4me | cloudnull time to perhaps update https://review.openstack.org/168976 ? | 18:52 |
| cloudnull | odyssey4me: https://review.openstack.org/218611 lgtm +2 | 18:53 |
| cloudnull | the spec will be updated soon. | 18:53 |
| odyssey4me | cloudnull awesome - it does seem to help improve hpcloud's responses - still not hpcloud-b4, but I have a seperate bug for that | 18:53 |
| odyssey4me | https://bugs.launchpad.net/openstack-ansible/+bug/1442630 | 18:53 |
| openstack | Launchpad bug 1442630 in openstack-ansible "gate: apt install for containers fail on hpcloud-b4" [High,Confirmed] - Assigned to Jesse Pretorius (jesse-pretorius) | 18:53 |
| odyssey4me | palendae if you have a gap to look at https://review.openstack.org/218611 it'd be appreciated - it seems to improve gating success for hpcloud | 18:56 |
| openstackgerrit | Kevin Carter proposed stackforge/os-ansible-deployment-specs: Tunable OpenStack Configuration Specification https://review.openstack.org/168976 | 18:59 |
| cloudnull | odyssey4me: ^ spec updated | 18:59 |
| *** cloudtrainme has quit IRC | 19:01 | |
| *** cloudtrainme has joined #openstack-ansible | 19:02 | |
| *** galstrom_zzz is now known as galstrom | 19:02 | |
| odyssey4me | cloudnull +2 from me :) | 19:02 |
| odyssey4me | I honestly think that's crucial for the Liberty cycle so that we can focus on features and real bugs, instead of silly conf entries. | 19:03 |
| *** jwagner is now known as jwagner_away | 19:07 | |
| *** jwagner_away is now known as jwagner | 19:09 | |
| openstackgerrit | Jesse Pretorius proposed stackforge/os-ansible-deployment: Add variable for cirros url https://review.openstack.org/217310 | 19:10 |
| openstackgerrit | Jesse Pretorius proposed stackforge/os-ansible-deployment: Update kilo for new dev work - 3 Sep 2015 https://review.openstack.org/220206 | 19:15 |
| cloudnull | odyssey4me: i 100% agree | 19:16 |
| *** woodard has joined #openstack-ansible | 19:19 | |
| *** galstrom is now known as galstrom_zzz | 19:22 | |
| *** cloudtrainme has quit IRC | 19:28 | |
| *** thrawn01 has joined #openstack-ansible | 19:28 | |
| *** galstrom_zzz is now known as galstrom | 19:29 | |
| thrawn01 | Hey guys, what backends does RS support for private cloud cinder backends? | 19:30 |
| palendae | thrawn01: To my knowledge, netapp and LVM. NetApp is the one you'll want if you want redundancy | 19:30 |
| cloudnull | thrawn01: we have lvm, nfs, and netapp | 19:31 |
| odyssey4me | thrawn01 and rbd (ceph) | 19:32 |
| thrawn01 | so, if I wanted to play with the cinder netapp driver, is there a place perhaps in the lab that might have a netapp to play with? | 19:32 |
| *** cloudtrainme has joined #openstack-ansible | 19:33 | |
| cloudnull | if you have a netapp or access to the netapp virtual appliance you could | 19:33 |
| palendae | thrawn01: Are you in RAX? | 19:39 |
| thrawn01 | palendae: yeah, I'm hanging with dolphm at the bugbash | 19:40 |
| *** cloudtrainme has quit IRC | 19:43 | |
| alextricity | Anybody know if there is a way to setup keystone SSL in OSAD? | 19:44 |
| odyssey4me | alextricity yes there is :p | 19:45 |
| *** cloudtrainme has joined #openstack-ansible | 19:46 | |
| odyssey4me | alextricity two methods - one is on keystone itself: https://github.com/stackforge/os-ansible-deployment/commit/caa9733788468886c6ac50cd2fde00a4f8a58321 | 19:46 |
| odyssey4me | alextricity the other is using haproxy if you don't have another LB: https://github.com/stackforge/os-ansible-deployment/commit/ff4c7ff746f9eadbdc146016b53ff8f8aabfa17e | 19:47 |
| palendae | alextricity: Yeah, a good question is: In Kilo or Juno? | 19:47 |
| palendae | Since Juno and Icehouse will live forever | 19:48 |
| alextricity | This is Kilo | 19:48 |
| palendae | k | 19:48 |
| palendae | Nice answer | 19:48 |
| alextricity | Thanks odyssey4me. I'm going to try out configuring it through keystone itself. | 19:49 |
| odyssey4me | alextricity sure :) | 19:50 |
| openstackgerrit | Merged stackforge/os-ansible-deployment-specs: Tunable OpenStack Configuration Specification https://review.openstack.org/168976 | 19:50 |
| arbrandes | mattt: Am I going to get in trouble if I co-locate Ceph OSD servers with compute nodes? | 19:51 |
| * arbrandes grins mischievously | 19:52 | |
| palendae | arbrandes: The safe answer there is: yes, don't do that. | 19:52 |
| mgariepy | arbrandes, in dev or in production ? | 19:52 |
| arbrandes | palendae: that's what I expected. ;) | 19:52 |
| arbrandes | mgariepy: dev | 19:52 |
| arbrandes | Oh, I'm not crazy enough to do that in production. | 19:52 |
| mgariepy | lol ok | 19:53 |
| arbrandes | But I'm too lazy to fire up 3 more VMs. :P | 19:53 |
| *** phalmos_ has joined #openstack-ansible | 19:53 | |
| odyssey4me | arbrandes dev is fine to some degree | 19:54 |
| odyssey4me | it depends on the load you put on them | 19:54 |
| odyssey4me | the osd's and instances contend for the cpu heavily, so the host does a lot of context switching | 19:55 |
| odyssey4me | the performance for both dives horribly | 19:55 |
| *** phalmos_ has quit IRC | 19:55 | |
| arbrandes | odyssey4me: agreed, entirely. I'm just wondering if any of the playbooks will mangle my cluster. | 19:55 |
| *** phalmos has quit IRC | 19:56 | |
| arbrandes | My development cluster, to be more specific. :P | 19:56 |
| *** dabernie_ has joined #openstack-ansible | 19:57 | |
| *** phalmos has joined #openstack-ansible | 19:57 | |
| *** dabernie_ has left #openstack-ansible | 19:57 | |
| odyssey4me | arbrandes I don't think it's tested. I'll hold your beer and watch. :) | 19:58 |
| cloudnull | hahaha | 19:58 |
| arbrandes | lol | 19:58 |
| openstackgerrit | Kevin Carter proposed stackforge/os-ansible-deployment: Replaced the copy_update module https://review.openstack.org/216790 | 19:58 |
| arbrandes | Ok, I think I'll just fire up 3 more VMs. :D | 19:58 |
| palendae | arbrandes: Cloud on top of VMs, I'm sure it'll be fine :) | 19:59 |
| cloudnull | afk a min | 19:59 |
| arbrandes | palendae: just for testing. ;) | 19:59 |
| odyssey4me | hahaha, we do that all the time - cloud in cloud :) | 20:00 |
| openstackgerrit | Merged stackforge/os-ansible-deployment: Fixing haproxy-playbook fails when installing on multiple hosts https://review.openstack.org/220089 | 20:01 |
| odyssey4me | hahaha ^ unexpected bonus patch for 11.2.1 | 20:03 |
| *** phalmos has quit IRC | 20:11 | |
| *** cloudtrainme has quit IRC | 20:12 | |
| arbrandes | Awesome | 20:12 |
| arbrandes | I intend to use that as well. :D | 20:12 |
| odyssey4me | arbrandes where're you from if you won't mind me asking? | 20:17 |
| odyssey4me | arbrandes you may also be interested in https://review.openstack.org/215525 then - we'll get that into 11.2.2 I hope :) | 20:20 |
| arbrandes | odyssey4me: I'm from Brazil. I'm guessing you guys are all rackers? | 20:20 |
| odyssey4me | arbrandes several of us are, although we have a few deployers from Belgium too :) | 20:20 |
| arbrandes | odyssey4me: ah, that changeset will indeed be useful. Thanks for pointing it out! | 20:21 |
| odyssey4me | and a few others are joining the party slowly | 20:21 |
| *** phalmos has joined #openstack-ansible | 20:22 | |
| *** cloudtrainme has joined #openstack-ansible | 20:24 | |
| *** sdake has joined #openstack-ansible | 20:25 | |
| *** sdake_ has quit IRC | 20:29 | |
| *** woodard has quit IRC | 20:34 | |
| *** galstrom is now known as galstrom_zzz | 20:35 | |
| *** woodard has joined #openstack-ansible | 20:36 | |
| *** jmckind has quit IRC | 20:38 | |
| *** jmckind has joined #openstack-ansible | 20:42 | |
| *** Mudpuppy has quit IRC | 21:08 | |
| *** cloudtrainme has quit IRC | 21:09 | |
| *** KLevenstein has quit IRC | 21:09 | |
| *** cloudtrainme has joined #openstack-ansible | 21:18 | |
| *** woodard has quit IRC | 21:20 | |
| *** javeriak has quit IRC | 21:21 | |
| *** KLevenstein has joined #openstack-ansible | 21:32 | |
| *** jmckind has quit IRC | 21:34 | |
| *** KLevenstein has quit IRC | 21:37 | |
| *** KLevenstein has joined #openstack-ansible | 21:38 | |
| *** phalmos has quit IRC | 21:46 | |
| *** jwagner is now known as jwagner_away | 21:52 | |
| *** Mudpuppy has joined #openstack-ansible | 21:55 | |
| *** tlian2 has joined #openstack-ansible | 21:56 | |
| openstackgerrit | Kevin Carter proposed stackforge/os-ansible-deployment: Replaced the copy_update module https://review.openstack.org/216790 | 21:58 |
| *** tlian has quit IRC | 21:58 | |
| openstackgerrit | Kevin Carter proposed stackforge/os-ansible-deployment: Disable scatter-gather offload on host bridges https://review.openstack.org/219292 | 22:02 |
| *** woodard has joined #openstack-ansible | 22:05 | |
| *** woodard has quit IRC | 22:07 | |
| *** woodard has joined #openstack-ansible | 22:07 | |
| *** KLevenstein has quit IRC | 22:11 | |
| *** Mudpuppy has quit IRC | 22:28 | |
| *** Mudpuppy has joined #openstack-ansible | 22:28 | |
| odyssey4me | \o/ https://review.openstack.org/186684 :) | 22:29 |
| openstackgerrit | Jesse Pretorius proposed stackforge/os-ansible-deployment: Update tag version to 11.2.1 https://review.openstack.org/220195 | 22:32 |
| *** spotz is now known as spotz_zzz | 22:38 | |
| BjoernT | quick question, which part in our new galaxy roles does the git part, for example neutron_git_repo as part of the defaults/repo_packages/openstack_services.yml ? | 22:56 |
| BjoernT | ok found it oles/py_from_git, lol | 22:57 |
| *** mrstanwell has joined #openstack-ansible | 22:57 | |
| openstackgerrit | Kevin Carter proposed stackforge/os-ansible-deployment: Added post and pre hook script for veth cleanup https://review.openstack.org/220342 | 23:00 |
| *** woodard has quit IRC | 23:02 | |
| *** woodard has joined #openstack-ansible | 23:04 | |
| *** sdake has quit IRC | 23:06 | |
| mrstanwell | hi, all. I've got a multi-node osad kilo install. I was trying to set up an ipv6 tenant network, and noticed that radvd isn't being installed on neutron_agents_container. is that a bug, or are v6 tenants not supported by osad? thanks! | 23:11 |
| *** markvoelker has joined #openstack-ansible | 23:13 | |
| *** cloudtrainme has quit IRC | 23:15 | |
| *** pradk has quit IRC | 23:16 | |
| *** markvoelker has quit IRC | 23:17 | |
| *** markvoelker has joined #openstack-ansible | 23:25 | |
| palendae | mrstanwell: I don't believe so. https://bugs.launchpad.net/openstack-ansible/+bug/1426809 | 23:43 |
| openstack | Launchpad bug 1426809 in openstack-ansible trunk "Neutron agent container fails when IPv6 is disabled on the host but enabled on containers" [Medium,Confirmed] | 23:43 |
| *** shoutm has joined #openstack-ansible | 23:50 | |
| mrstanwell | palendae: well, when I install radvd in neutron_agents_container, SLAAC appears to be working. tenant configured a v6 addr, and can ping the virtual router. I can't ping anything else, but I might have some incorrect setup elsewhere. so it seems like radvd should be getting installed, unless there are bigger obstacles to ipv6 tenants. That bug looks like a different issue. | 23:56 |
| cloudnull | mrstanwell: Thats awesome ! | 23:59 |
| cloudnull | would you mind regestering a lanuchpad bug for that so we can get that in for liberty / kilo | 23:59 |
Generated by irclog2html.py 2.14.0 by Marius Gedminas - find it at mg.pov.lt!
