| *** markvoelker has quit IRC | 00:12 | |
| openstackgerrit | Nolan Brubaker proposed stackforge/os-ansible-deployment: Add Sphinx machinery and initial docs. https://review.openstack.org/183687 | 00:31 |
|---|---|---|
| openstackgerrit | Nolan Brubaker proposed stackforge/os-ansible-deployment: Add Sphinx machinery and initial docs. https://review.openstack.org/183687 | 00:33 |
| openstackgerrit | Nolan Brubaker proposed stackforge/os-ansible-deployment: Add an extension section to the developer docs. https://review.openstack.org/183857 | 00:35 |
| *** sdake_ has quit IRC | 00:38 | |
| *** sdake has joined #openstack-ansible | 00:38 | |
| *** sigmavirus24 is now known as sigmavirus24_awa | 00:41 | |
| *** jimi|ansible has joined #openstack-ansible | 00:45 | |
| openstackgerrit | Nolan Brubaker proposed stackforge/os-ansible-deployment: Add an extension section to the developer docs. https://review.openstack.org/183857 | 00:46 |
| *** sdake_ has joined #openstack-ansible | 00:52 | |
| *** sdake has quit IRC | 00:57 | |
| *** markvoelker has joined #openstack-ansible | 01:13 | |
| *** sdake_ has quit IRC | 01:17 | |
| *** markvoelker has quit IRC | 01:18 | |
| *** sdake has joined #openstack-ansible | 01:18 | |
| *** dkalleg has quit IRC | 01:21 | |
| *** hoppers99 has left #openstack-ansible | 02:01 | |
| *** javeriak has quit IRC | 02:05 | |
| *** dkalleg has joined #openstack-ansible | 02:22 | |
| *** dkalleg has quit IRC | 02:26 | |
| *** sdake_ has joined #openstack-ansible | 02:40 | |
| *** sdake has quit IRC | 02:43 | |
| *** sdake has joined #openstack-ansible | 02:43 | |
| *** sdake_ has quit IRC | 02:47 | |
| *** galstrom_zzz is now known as galstrom | 02:57 | |
| *** markvoelker has joined #openstack-ansible | 03:02 | |
| *** markvoelker has quit IRC | 03:06 | |
| *** fawadkhaliq has joined #openstack-ansible | 03:07 | |
| *** markvoelker has joined #openstack-ansible | 04:02 | |
| *** markvoelker has quit IRC | 04:07 | |
| *** galstrom is now known as galstrom_zzz | 04:24 | |
| *** sdake_ has joined #openstack-ansible | 04:29 | |
| *** appprod0 has joined #openstack-ansible | 04:30 | |
| *** appprod0 has quit IRC | 04:31 | |
| *** sdake has quit IRC | 04:33 | |
| *** JRobinson__ is now known as JRobinson__afk | 05:14 | |
| *** JRobinson__afk is now known as JRobinson__ | 05:42 | |
| *** sdake has joined #openstack-ansible | 05:46 | |
| *** sdake_ has quit IRC | 05:50 | |
| *** markvoelker has joined #openstack-ansible | 05:52 | |
| *** markvoelker has quit IRC | 05:56 | |
| *** radek__ has joined #openstack-ansible | 06:10 | |
| *** javeriak has joined #openstack-ansible | 06:14 | |
| *** JRobinson__ has quit IRC | 07:34 | |
| *** markvoelker has joined #openstack-ansible | 07:40 | |
| *** igorr1 has joined #openstack-ansible | 07:42 | |
| *** markvoelker has quit IRC | 07:45 | |
| *** sdake_ has joined #openstack-ansible | 07:55 | |
| *** stevemar has quit IRC | 07:55 | |
| *** javeriak has quit IRC | 07:58 | |
| *** sdake has quit IRC | 07:59 | |
| *** sdake_ has quit IRC | 08:17 | |
| odyssey4me | morning all | 08:27 |
| *** igorr1 has quit IRC | 08:30 | |
| svg | morning ^_^ | 08:30 |
| evrardjp | morning | 08:33 |
| *** fawadkhaliq has quit IRC | 08:36 | |
| mattt | morning odyssey4me, svg, evrardjp | 08:39 |
| svg | hi mattt | 08:43 |
| odyssey4me | o/ | 08:44 |
| odyssey4me | svg you're from Belguim right? | 08:57 |
| odyssey4me | *Belgium (sorry) | 08:57 |
| svg | yes | 08:57 |
| odyssey4me | evrardjp where are you from? | 08:59 |
| odyssey4me | (if you don't mind my asking?) | 08:59 |
| evrardjp | from or in? | 09:00 |
| evrardjp | in -> hell, atm | 09:00 |
| evrardjp | cinder causing me pain | 09:00 |
| mattt | haha | 09:00 |
| evrardjp | Europe otherwise | 09:01 |
| evrardjp | why ? | 09:01 |
| evrardjp | why?* | 09:01 |
| odyssey4me | haha, just curious | 09:01 |
| odyssey4me | you're clearly in a similar time zone | 09:01 |
| odyssey4me | or you work some quirky hours ;) | 09:02 |
| evrardjp | I do have some quirky hours sometimes... but that's not the point of the conversation, right? ;) | 09:03 |
| *** igorr1 has joined #openstack-ansible | 09:05 | |
| svg | evrardjp: my rabbits seem to alive and well | 09:09 |
| svg | I used the f5 script - from rpc-extras - to generate lb config, and I notice this generates a lot less than it did in Juno | 09:16 |
| mattt | if i cherry-pick a change, is it acceptable for me to +2 it and merge? | 09:17 |
| svg | e.g. for nova, I only get the metadata en api_compute definitions, whil in juno there's also the spice console and the ec2 service | 09:17 |
| svg | (wondering if this could cause my compute hosts to be marked down) | 09:18 |
| *** dkalleg has joined #openstack-ansible | 09:20 | |
| andymccr | svg: that shouldn't be the case - the spice_console thing is a bug, but I thought we had fixed that (the name on the spice_console group changed so it doesn't find any nodes) | 09:23 |
| andymccr | i believe ec2 was deprecated from nova, so its not setup in kilo | 09:24 |
| svg | I suspected something like that | 09:24 |
| *** dkalleg has quit IRC | 09:24 | |
| andymccr | compute itself should get handled by the scheduler so isn't actually load balanced | 09:25 |
| svg | andymccr: looks like that group name was changed in master but not in kilo: https://github.com/rcbops/rpc-extras/blob/kilo/scripts/f5-config.py#L151 | 09:26 |
| andymccr | svg: ahh good catch - i'll fix it up | 09:27 |
| andymccr | thanks! | 09:27 |
| svg | the funny thing is, I think I used the one from master :-s | 09:28 |
| andymccr | hahaha | 09:28 |
| andymccr | if that didnt work i can take a look and see why | 09:28 |
| *** markvoelker has joined #openstack-ansible | 09:29 | |
| svg | hm, it creates a OPENSTACK_POOL_NOVA_SPICE_CONSOLE, but no matching VS | 09:31 |
| svg | ah, pool is empty | 09:32 |
| *** markvoelker has quit IRC | 09:34 | |
| *** fawadkhaliq has joined #openstack-ansible | 09:49 | |
| odyssey4me | mattt typically a cherry-pick needs 2x+2 from others, much like any other patch | 09:51 |
| mattt | yeah i opted against it :) | 09:55 |
| odyssey4me | urgh, it looks like we're getting build failures in master/kilo - possibly a dependency update somewhere | 09:56 |
| mattt | master also? hrrrmph. | 10:04 |
| mattt | odyssey4me: not understanding this issue | 10:38 |
| odyssey4me | mattt yeah, took a peep and didn't see anything obvious - but on a deadline for log processing so it'll have to wait or someone else will need to look into it | 10:39 |
| mattt | i'm still digging | 10:39 |
| odyssey4me | sweet, thanks | 10:39 |
| *** fawadkhaliq has quit IRC | 10:42 | |
| *** markvoelker has joined #openstack-ansible | 11:18 | |
| *** markvoelker has quit IRC | 11:22 | |
| *** fawadkhaliq has joined #openstack-ansible | 11:43 | |
| *** fawadkhaliq has quit IRC | 11:48 | |
| mattt | ok the repo issue may be due to python-qpid-proton | 11:50 |
| *** markvoelker has joined #openstack-ansible | 11:59 | |
| openstackgerrit | Kevin Carter proposed stackforge/os-ansible-deployment: Updated repo server deps https://review.openstack.org/186740 | 12:21 |
| *** sigmavirus24_awa is now known as sigmavirus24 | 13:01 | |
| mattt | cloudnull: question | 13:02 |
| cloudnull | yes? | 13:02 |
| mattt | cloudnull: why do we build requirements and branches/releases in two steps in our playbooks? | 13:03 |
| cloudnull | you mean https://github.com/stackforge/os-ansible-deployment/blob/master/playbooks/repo-build.yml#L38 and https://github.com/stackforge/os-ansible-deployment/blob/master/playbooks/repo-build.yml#L58 ? | 13:04 |
| mattt | cloudnull: correct, i see there arent' two steps in https://github.com/cloudnull/yaprt/blob/master/EXAMPLE-SCRIPTS/openstack-repo-build.sh | 13:04 |
| mattt | was wondering if it was a timing thing or something? | 13:04 |
| cloudnull | no. | 13:04 |
| cloudnull | the first one resolves all deps and uses a graph to determine in deps can or should be built as they're stated. | 13:05 |
| cloudnull | the second builds all of the git repos that have a setup.py | 13:05 |
| cloudnull | with no deps | 13:05 |
| cloudnull | because they've already been resolved. | 13:05 |
| cloudnull | essentially one is an online build using pip and the other is using bdist_wheel | 13:06 |
| cloudnull | from within the contex of the stored git repo. | 13:06 |
| mattt | hmm, ok | 13:07 |
| mattt | i don't know enough about pip to actually understand your answer | 13:07 |
| * mattt reads | 13:08 | |
| mattt | hehe | 13:08 |
| cloudnull | so the first one creates this http://rpc-repo.rackspace.com/os-releases/master/build_reqs.txt | 13:08 |
| cloudnull | and then builds everything as a wheel. | 13:09 |
| cloudnull | then the second one uses the git repo to build the individual clients and services | 13:09 |
| cloudnull | while we reference upstream global requirements and pull that into our process we also query all of the projects we support to make sure that the global requirements is correct and that the projects are compatible with it. | 13:11 |
| mattt | that's awesome cloudnull, yaprt looks like a pretty cool tool | 13:12 |
| cloudnull | and then build a for real global requirements based on all of the projects checked out sha|branch|tag | 13:12 |
| cloudnull | its getting there. | 13:12 |
| * cloudnull brb | 13:14 | |
| *** fawadkhaliq has joined #openstack-ansible | 13:14 | |
| *** fawadkhaliq has quit IRC | 13:19 | |
| *** KLevenstein has joined #openstack-ansible | 13:21 | |
| *** subscope has quit IRC | 13:35 | |
| *** jimi|ansible has quit IRC | 13:45 | |
| *** galstrom_zzz is now known as galstrom | 13:49 | |
| *** subscope has joined #openstack-ansible | 13:52 | |
| *** fawadkhaliq has joined #openstack-ansible | 13:57 | |
| *** Mudpuppy has joined #openstack-ansible | 13:59 | |
| *** Mudpuppy has quit IRC | 13:59 | |
| *** Mudpuppy has joined #openstack-ansible | 14:00 | |
| *** sigmavirus24 is now known as sigmavirus24_awa | 14:06 | |
| mattt | zuul isn't looking happy | 14:15 |
| *** subscope has quit IRC | 14:15 | |
| odyssey4me | mattt yep, in #openstack-infra they seem to be discussing an issue across both providers | 14:21 |
| *** galstrom is now known as galstrom_zzz | 14:24 | |
| *** galstrom_zzz is now known as galstrom | 14:25 | |
| *** sigmavirus24_awa is now known as sigmavirus24 | 14:25 | |
| *** galstrom is now known as galstrom_zzz | 14:28 | |
| *** galstrom_zzz is now known as galstrom | 14:30 | |
| *** subscope has joined #openstack-ansible | 14:31 | |
| *** stevemar has joined #openstack-ansible | 14:39 | |
| *** jimi|ansible has joined #openstack-ansible | 14:40 | |
| *** jimi|ansible has quit IRC | 14:45 | |
| *** sdake has joined #openstack-ansible | 14:46 | |
| *** galstrom is now known as galstrom_zzz | 15:07 | |
| *** alextricity has joined #openstack-ansible | 15:10 | |
| *** sdake_ has joined #openstack-ansible | 15:26 | |
| *** sdake has quit IRC | 15:30 | |
| *** fawadkhaliq has quit IRC | 15:30 | |
| *** subscope has quit IRC | 15:35 | |
| *** jimi|ansible has joined #openstack-ansible | 15:37 | |
| *** daneyon has joined #openstack-ansible | 15:50 | |
| *** daneyon_ has joined #openstack-ansible | 15:53 | |
| *** jwagner_away is now known as jwagner | 15:54 | |
| *** daneyon has quit IRC | 15:55 | |
| svg | afaics, the run-upgrade.sh script does not touch the providers_network in openstack_user_config.yml, whilst there is a difference between the juno and kilo version | 16:09 |
| svg | in particular, the net_name: "flat" was updated (originally net_name: "vlan") | 16:10 |
| svg | noticed, as the neutron agents on the compute hosts (at least) fail to start because of an overlapping mapping (vlan and flat) | 16:13 |
| svg | neutron.plugins.linuxbridge.agent.linuxbridge_neutron_agent Parsing physical_interface_mappings failed: Key vlan in mapping not unique. | 16:13 |
| cloudnull | svg: those entries havent changed the examples have been updated with a few more syntatic bits but if it worked in juno itll work in kilo. That said, the mapping issue was a problem in juno too however juno neutron didnt complain about it where in kilo it does . | 16:13 |
| cloudnull | the neutron ml2 driver wont allow for overllapping networks mapped to different network types. | 16:14 |
| cloudnull | which is why we created the overrides in the provider networks and created an example of veth pairs off a single device to show how it could be done. | 16:15 |
| cloudnull | Apsu: worked on that | 16:15 |
| Sam-I-Am | you can still do flat and vlan on the same interface, no? | 16:16 |
| * Apsu nods | 16:16 | |
| Sam-I-Am | just not more than one of those | 16:16 |
| Apsu | Sam-I-Am: Yes. Just takes a little workaround | 16:16 |
| Apsu | Essentially the answer is actually "no". | 16:16 |
| Apsu | You can't tell Neutron to do two mappings to the same interface. | 16:16 |
| Sam-I-Am | or... you can't have one network name mapped to multiple interfaces | 16:16 |
| Apsu | But you can make a veth pair leg off of a bridge and map to the other end of the veth | 16:17 |
| svg | afaics I ned to update the net_name for tha flat one, from vlan to flat | 16:17 |
| Apsu | So Neutron thinks it's different | 16:17 |
| cloudnull | ^ that | 16:17 |
| Sam-I-Am | you'd think it would just map 1 flat and 1 vlan to the same interface, since flat is more or less the untagged vlan... but neutron. | 16:17 |
| *** sigmavirus24 is now known as sigmavirus24_awa | 16:18 | |
| Apsu | Sam-I-Am: Yep, it's too dumb for that at present. | 16:18 |
| svg | also, container_type: "veth", seems to be a new parameter? | 16:19 |
| Sam-I-Am | Apsu: patches welcome | 16:21 |
| *** igorr1 has quit IRC | 16:22 | |
| *** sigmavirus24_awa is now known as sigmavirus24 | 16:28 | |
| cloudnull | svg the default is veth but you can now hand other network types to a container, ie macvlan, vlan, host, etc. | 16:31 |
| odyssey4me | my brain is fried - a time out is overdue | 16:41 |
| cloudnull | take it easy odyssey4me | 16:42 |
| cloudnull | and a beer | 16:42 |
| *** sigmavirus24 is now known as sigmavirus24_awa | 16:43 | |
| *** javeriak has joined #openstack-ansible | 16:54 | |
| *** galstrom_zzz is now known as galstrom | 16:56 | |
| *** fawadkhaliq has joined #openstack-ansible | 16:59 | |
| *** KLevenstein has quit IRC | 17:00 | |
| *** daneyon_ has quit IRC | 17:05 | |
| *** sdake has joined #openstack-ansible | 17:09 | |
| *** jimi|ansible has quit IRC | 17:11 | |
| *** sdake_ has quit IRC | 17:12 | |
| *** sdake_ has joined #openstack-ansible | 17:20 | |
| *** sdake has quit IRC | 17:23 | |
| *** sdake has joined #openstack-ansible | 17:26 | |
| *** dkalleg has joined #openstack-ansible | 17:26 | |
| *** sigmavirus24_awa is now known as sigmavirus24 | 17:26 | |
| *** sdake_ has quit IRC | 17:29 | |
| *** KLevenstein has joined #openstack-ansible | 17:35 | |
| openstackgerrit | Merged stackforge/os-ansible-deployment: Handle proxies through environment variables https://review.openstack.org/181603 | 17:49 |
| openstackgerrit | Merged stackforge/os-ansible-deployment: Allow configuration of Nova SQLAlchemy options https://review.openstack.org/186583 | 17:50 |
| openstackgerrit | Merged stackforge/os-ansible-deployment: Set Rabbit recovery method for network partitions https://review.openstack.org/186243 | 17:50 |
| *** galstrom is now known as galstrom_zzz | 17:54 | |
| openstackgerrit | Merged stackforge/os-ansible-deployment: Perform sort when creating hostnames_ip file https://review.openstack.org/186591 | 17:55 |
| *** sdake_ has joined #openstack-ansible | 18:07 | |
| svg | +2 for handling proxies | 18:10 |
| *** sdake has quit IRC | 18:12 | |
| sigmavirus24 | svg: I'm sorry | 18:22 |
| sigmavirus24 | We had to revert that because you +2'd it ;) | 18:22 |
| * svg slaps sigmavirus24 with a large trout | 18:22 | |
| sigmavirus24 | (svg, now that you're going to be a core, I have to give you a good ribbing) | 18:22 |
| sigmavirus24 | WELCOME TO THE TEAM | 18:23 |
| svg | Well, that was not decided yet :) | 18:23 |
| sigmavirus24 | I've seen no objections | 18:23 |
| openstackgerrit | Kevin Carter proposed stackforge/os-ansible-deployment: Handle proxies through environment variables https://review.openstack.org/186868 | 18:23 |
| svg | I don't see any either :) | 18:36 |
| *** KLevenstein__ has joined #openstack-ansible | 18:40 | |
| *** galstrom_zzz is now known as galstrom | 18:41 | |
| *** KLevenstein has quit IRC | 18:42 | |
| *** KLevenstein__ is now known as KLevenstein | 18:42 | |
| sigmavirus24 | http://lists.openstack.org/pipermail/openstack-dev/2015-May/065144.html | 18:44 |
| cloudnull | sigmavirus24: yup, thats ok with us now, IMO. we've been tracking the stable/* branches for some time so sticking to tags is not going to be something effects us. | 18:53 |
| sigmavirus24 | cloudnull: I know :D | 18:53 |
| sigmavirus24 | I'm just happy to see it is being abandoned | 18:53 |
| cloudnull | but debian, ubuntu, and rdo are fucked. | 18:53 |
| sigmavirus24 | cloudnull: Like I said, I'm happy to see it | 18:53 |
| sigmavirus24 | ;) | 18:53 |
| cloudnull | the last message in that thread is an rdo guy complaining how his life is about to get hard. | 18:54 |
| cloudnull | so im with you | 18:54 |
| cloudnull | :) | 18:54 |
| cloudnull | looks like you won the packaging argument war after all sigmavirus24 | 18:54 |
| *** KLevenstein has quit IRC | 19:15 | |
| *** KLevenstein has joined #openstack-ansible | 19:15 | |
| svg | hmz, so now I have | 19:32 |
| svg | /etc/neutron/plugins/ml2/ml2_conf.ini:physical_interface_mappings = flat:br-vlan,vlan:br-vlan | 19:32 |
| svg | but it still complains about vlan not being unique | 19:32 |
| svg | 2015-05-29 21:32:32.855 33134 ERROR neutron.plugins.linuxbridge.agent.linuxbridge_neutron_agent [-] Parsing physical_interface_mappings failed: Value br-vlan in mapping: 'vlan:br-vlan' not unique. Agent terminated! | 19:32 |
| svg | argh tha valu emust be unique | 19:33 |
| svg | I actually don;t use flat networks so let's just get rid of that | 19:33 |
| sigmavirus24 | svg: good idea :D | 19:37 |
| svg | coz what else could I do on a Friday night? | 19:37 |
| sigmavirus24 | svg: uhm, enjoy liquids? | 19:37 |
| * svg does that, too | 19:37 | |
| svg | now you mention it, I need a refill | 19:38 |
| openstackgerrit | Javeria Khan proposed stackforge/os-ansible-deployment-specs: Add PLUMgrid Neutron Plugin to neutron playbooks https://review.openstack.org/184726 | 19:41 |
| cloudnull | svg: https://www.youtube.com/watch?v=m1vRvW3QUys | 19:41 |
| sigmavirus24 | svg: enjoy! | 19:41 |
| svg | thans guys | 19:42 |
| svg | btw, cloudnull : is there a reason the -dev and -operators list not having an [osad] topic? | 19:42 |
| javeriak | hey guys, i was hoping to get some feedback on my BP: https://review.openstack.org/#/c/184726/6/specs/kilo/add-plumgrid-neutron-plugin.rst, was hoping to target this for 10.1.7 | 19:42 |
| javeriak | should i wait for a group meeting? | 19:43 |
| *** sdake_ is now known as sdake | 19:43 | |
| cloudnull | svg: we could add one. but i failed to add it to the subject line . | 19:43 |
| cloudnull | javeriak: i will give it a review. | 19:43 |
| svg | I don;t have the impression there is lots of discussion per mail for now? | 19:43 |
| cloudnull | there is not. | 19:43 |
| cloudnull | almost everything is here on IRC | 19:43 |
| svg | javeriak: kilo != 10.x | 19:43 |
| svg | whwich works as the group is small I guess | 19:44 |
| cloudnull | it is , likely we'll be going to the ml as we grow. | 19:44 |
| javeriak | svg: yes, BP targets both juno and kilo, was going to implement juno first | 19:44 |
| svg | perhaps I'm missing what | 19:45 |
| javeriak | We have some customers still using the Juno train for Rackspace deployment | 19:45 |
| svg | perhaps I'm missing what 'BP' means? | 19:45 |
| javeriak | svg: sorry BP: blue print, thought that was like a community accepted acronym :) | 19:45 |
| cloudnull | svg: Blue Print | 19:45 |
| *** sdake_ has joined #openstack-ansible | 19:46 | |
| svg | probably, but i consider myyself still new :) | 19:46 |
| svg | so that's on me | 19:46 |
| * svg should file a bug to get reruns from unchanged config all green | 19:47 | |
| javeriak | svg: no worries :), maybe we should take them out from geek lore and publish these :) | 19:48 |
| *** sdake has quit IRC | 19:49 | |
| *** sdake has joined #openstack-ansible | 19:49 | |
| sigmavirus24 | cloudnull: were you at the requirements handling session? | 19:50 |
| cloudnull | no | 19:50 |
| cloudnull | anything to write home about ? | 19:50 |
| sigmavirus24 | https://review.openstack.org/#/c/186635/ | 19:51 |
| sigmavirus24 | I think it'll make our lives easier? | 19:51 |
| sigmavirus24 | easier to read: http://docs-draft.openstack.org/35/186635/3/check/gate-openstack-specs-docs/84eb447//doc/build/html/specs/requirements-management.html | 19:52 |
| *** sdake_ has quit IRC | 19:53 | |
| *** KLevenstein has quit IRC | 19:57 | |
| *** openstack has joined #openstack-ansible | 20:06 | |
| sigmavirus24 | I think g-r will continue to operate as it is now | 20:11 |
| sigmavirus24 | and test-requirements will be g-r pinned + test requirements pinned | 20:11 |
| sigmavirus24 | I'm guessing | 20:12 |
| sigmavirus24 | I haven't finished reading the spec | 20:12 |
| *** jmccrory has joined #openstack-ansible | 20:19 | |
| svg | does this ring some bells? https://dl.dropboxusercontent.com/u/13986042/20150529222214.png | 20:23 |
| *** sigmavirus24 is now known as sigmavirus24_awa | 20:23 | |
| *** openstack has joined #openstack-ansible | 20:31 | |
| svg | I suspected that :) | 20:32 |
| svg | the other one is mentioning the agent is out of sync | 20:32 |
| svg | this is from the neutron bridge agent on the compute hosts | 20:33 |
| Apsu | Looks like that's right after an agent restart | 20:33 |
| svg | ok | 20:33 |
| svg | that service was crashing constantl bc of the mapping error I mentioned previously | 20:33 |
| svg | so now that is solved | 20:33 |
| * Apsu nods | 20:33 | |
| svg | but my compute hosts still remain in a down state, so that was not ythe problem cuasin git | 20:34 |
| svg | causing it | 20:34 |
| Apsu | Down as seen from "nova service-list" for nova-compute entries? | 20:34 |
| svg | yes | 20:34 |
| svg | if I disable the service, it gets marked up; when I re-enable the service, it goes down again after a couple of minutes | 20:35 |
| Apsu | Check your nova-compute logs, but, you might need to restart the nova-conductor along with the nova-compute services | 20:35 |
| svg | compute logs say nothing except initializing | 20:36 |
| *** sdake_ has joined #openstack-ansible | 20:36 | |
| openstackgerrit | Merged stackforge/os-ansible-deployment: Add Sphinx machinery and initial docs. https://review.openstack.org/183687 | 20:36 |
| Apsu | Might need to add verbose = True or debug = True to ferret out what's up, but first step would be to restart the conductor | 20:36 |
| Apsu | nova-compute communicates via the conductors, and if you had transient rabbit issues or something got mucked up, conductor failures may cause compute to eventually bail out | 20:37 |
| *** sdake has quit IRC | 20:40 | |
| svg | nothing so far | 20:40 |
| Apsu | If you've enabled verbose/debug logging and restarted conductor, let us know if it bails again and if there's any logs to look at and whether conductor/compute are still running when the service-list shows "Down". | 20:42 |
| svg | hold on, redeploying all of nova with debug true | 20:44 |
| Apsu | Alrighty | 20:44 |
| svg | ok, so the compute log basically just starts up and mentions Libvirt host capabilities | 20:49 |
| svg | | 52 | nova-compute | dc1-rk2-ch1-bl9 | nova | enabled | down | 2015-05-29T20:08:37.000000 | - | | 20:50 |
| svg | and so are all the computes | 20:50 |
| *** sdake has joined #openstack-ansible | 20:50 | |
| svg | Any particular logs I should look at? | 20:50 |
| Apsu | Did you restart the conductor service? | 20:50 |
| Apsu | It has its own logs too | 20:50 |
| svg | ansible restarted all teh nova services when adding the debug | 20:51 |
| svg | conductor seems to be overwhelmed with | 20:51 |
| svg | 2015-05-29 22:50:57.902 3656 DEBUG nova.openstack.common.loopingcall [req-a275ac1a-0521-4cc8-a331-4682c9647006 - - - - -] Dynamic looping call <bound method Service.periodic_tasks of <nova.service.Service object at 0x7fa7a1d67f50>> sleeping for 60.00 seconds _inner /usr/local/lib/python2.7/dist-packages/nova/openstack/common/loopingcall.py:132 | 20:51 |
| Apsu | And you're seeing nothing in nova-compute's log? Is it even running? | 20:52 |
| Apsu | pgrep -fa nova-compute | 20:52 |
| svg | yup | 20:52 |
| svg | nothing except it starting up of course | 20:52 |
| Apsu | That's particularly weird. Is it still running? | 20:54 |
| *** sdake_ has quit IRC | 20:54 | |
| svg | conductor? yes, with a bunch of subprocesses | 20:55 |
| Apsu | No, compute | 20:55 |
| Apsu | pgrep -fa nova-compute | 20:56 |
| Apsu | My guess is it's dying not long after starting, probably due to a config error, and that's why it's not spitting out a ton of log message | 20:56 |
| Apsu | Or segfaulting early due to some other issue | 20:57 |
| svg | it is running | 20:57 |
| Apsu | Hrm....check the nova.conf and make sure it says "debug = True" under [DEFAULT] | 20:57 |
| svg | last log entry sya\\ays | 20:57 |
| svg | 2015-05-29 22:47:52.144 22828 INFO nova.compute.resource_tracker [req-8c3fe8e5-b421-4d66-8116-fed4778f2a9d - - - - -] Auditing locally available compute resources for node dc1-rk2-ch1-bl5.cumuli.be | 20:57 |
| *** Mudpuppy_ has joined #openstack-ansible | 20:58 | |
| svg | full compute log: /var/log/nova/nova-compute.log | 20:58 |
| svg | http://sprunge.us/WURd | 20:58 |
| *** Mudpuppy_ has quit IRC | 20:59 | |
| Apsu | So it IS outputting a Ton of debug content. Ok, that's a good sign to start :P | 20:59 |
| *** openstackgerrit has quit IRC | 20:59 | |
| *** Mudpuppy_ has joined #openstack-ansible | 20:59 | |
| *** openstackgerrit has joined #openstack-ansible | 20:59 | |
| svg | ah, sorry, I meant, on error relevant logs :) | 21:00 |
| svg | no | 21:00 |
| Apsu | svg: Any new logs after that point? | 21:00 |
| svg | that was the full file catted | 21:00 |
| Apsu | Understood, I'm asking about right this moment, since it's been a few minutes. | 21:01 |
| svg | last entry 2015-05-29 22:47:52.144 | 21:01 |
| Apsu | So nothing new in the log. hrm. | 21:01 |
| *** Mudpuppy has quit IRC | 21:01 | |
| Apsu | I've got to take off for an appointment, but that's strange. Seems like it's not trying to talk to anything... have you made any custom changes to your nova-related user variables or config templates? | 21:02 |
| Apsu | Might be worth doing a git diff against your OSAD configs | 21:02 |
| svg | not really, except for my ceph patch | 21:02 |
| *** fawadkhaliq has quit IRC | 21:05 | |
| *** jwagner is now known as jwagner_away | 21:11 | |
| *** sigmavirus24_awa is now known as sigmavirus24 | 21:14 | |
| svg | Apsu: tcpdumping during a nova compute restart shows it connects to rabbitmq | 21:14 |
| mattt | svg: so question | 21:20 |
| mattt | actually n/m | 21:21 |
| svg | I was all ears :) | 21:21 |
| *** sdake is now known as steakache | 21:23 | |
| *** steakache is now known as sdake | 21:24 | |
| *** openstack has joined #openstack-ansible | 21:33 | |
| *** stevemar has quit IRC | 21:53 | |
| *** Mudpuppy_ has quit IRC | 21:59 | |
| *** fawadkhaliq has joined #openstack-ansible | 22:05 | |
| *** fawadkhaliq has quit IRC | 22:11 | |
| *** jmccrory has quit IRC | 22:20 | |
| *** sdake_ has joined #openstack-ansible | 22:39 | |
| *** sdake has quit IRC | 22:43 | |
| *** sdake has joined #openstack-ansible | 22:49 | |
| *** sdake has quit IRC | 22:49 | |
| *** sdake has joined #openstack-ansible | 22:49 | |
| *** sdake_ has quit IRC | 22:52 | |
| *** radek__ has quit IRC | 22:54 | |
| *** fawadkhaliq has joined #openstack-ansible | 23:52 | |
| *** sdake_ has joined #openstack-ansible | 23:56 | |
| *** fawadkhaliq has quit IRC | 23:56 | |
Generated by irclog2html.py 2.14.0 by Marius Gedminas - find it at mg.pov.lt!
