| *** schwicht has joined #openstack-ansible | 00:00 | |
| *** markvoelker has quit IRC | 00:00 | |
| *** markvoelker has joined #openstack-ansible | 00:01 | |
| *** severion has quit IRC | 00:06 | |
| *** v1k0d3n has joined #openstack-ansible | 00:07 | |
| *** schwicht has quit IRC | 00:14 | |
| *** schwicht has joined #openstack-ansible | 00:14 | |
| *** eil397 has quit IRC | 00:15 | |
| *** kenhui has joined #openstack-ansible | 00:16 | |
| *** kenhui has quit IRC | 00:17 | |
| *** schwicht_at_work has joined #openstack-ansible | 00:19 | |
| *** schwicht has quit IRC | 00:20 | |
| cloudnull | back | 00:21 |
|---|---|---|
| *** thorst has joined #openstack-ansible | 00:24 | |
| stevelle | cloudnull: curious what hacking you did to get the middleware to work. did you try something like this: https://gist.github.com/stevelle/3e34465fb31a21c204c38111982d0804 | 00:24 |
| cloudnull | yes. i added that. | 00:26 |
| cloudnull | no effect | 00:26 |
| cloudnull | I then dove into https://github.com/openstack/ceilometermiddleware/blob/master/ceilometermiddleware/swift.py | 00:27 |
| cloudnull | which i changed/modified bits and bobs, still no effect. | 00:27 |
| cloudnull | then i went to twitter. | 00:27 |
| cloudnull | :) | 00:27 |
| stevelle | ok, the conf object being handed through to the oslo messaging transport looks to be the stock oslo config | 00:28 |
| *** schwicht_at_work has quit IRC | 00:28 | |
| stevelle | it suggest that should have worked but was also the obvious choice | 00:29 |
| *** schwicht has joined #openstack-ansible | 00:30 | |
| *** markvoelker has quit IRC | 00:35 | |
| *** markvoelker has joined #openstack-ansible | 00:44 | |
| *** schwicht has quit IRC | 00:45 | |
| *** schwicht has joined #openstack-ansible | 00:45 | |
| *** Jeffrey4l has joined #openstack-ansible | 00:46 | |
| *** v1k0d3n has quit IRC | 00:48 | |
| *** hybridpolio has joined #openstack-ansible | 00:49 | |
| *** schwicht has quit IRC | 00:49 | |
| *** hybridpollo has quit IRC | 00:49 | |
| stevelle | off to dinner | 00:52 |
| stevelle | will continue poking after | 00:52 |
| *** kenhui has joined #openstack-ansible | 00:52 | |
| *** schwicht has joined #openstack-ansible | 00:53 | |
| *** kenhui has quit IRC | 00:54 | |
| *** hybridpolio is now known as hybridpollo | 00:57 | |
| *** kenhui has joined #openstack-ansible | 00:59 | |
| *** kenhui has quit IRC | 00:59 | |
| *** kenhui has joined #openstack-ansible | 01:03 | |
| *** kenhui has quit IRC | 01:04 | |
| *** schwicht has quit IRC | 01:12 | |
| *** schwicht has joined #openstack-ansible | 01:15 | |
| *** waht has joined #openstack-ansible | 01:21 | |
| *** markvoelker has quit IRC | 01:28 | |
| *** schwicht has quit IRC | 01:30 | |
| *** schwicht has joined #openstack-ansible | 01:34 | |
| *** markvoelker has joined #openstack-ansible | 01:35 | |
| *** markvoelker has quit IRC | 01:40 | |
| *** schwicht has quit IRC | 01:42 | |
| *** schwicht has joined #openstack-ansible | 01:42 | |
| *** schwicht has quit IRC | 01:47 | |
| *** schwicht has joined #openstack-ansible | 01:49 | |
| *** markvoelker has joined #openstack-ansible | 02:04 | |
| *** phalmos has quit IRC | 02:10 | |
| *** sdake has quit IRC | 02:21 | |
| *** markvoelker has quit IRC | 02:23 | |
| *** mgariepy has quit IRC | 02:23 | |
| cloudnull | evenings | 02:29 |
| *** mgariepy has joined #openstack-ansible | 02:31 | |
| *** mgariepy has quit IRC | 02:36 | |
| *** kstev has quit IRC | 02:43 | |
| *** thorst has quit IRC | 02:45 | |
| *** thorst has joined #openstack-ansible | 02:46 | |
| *** mgariepy has joined #openstack-ansible | 02:46 | |
| openstackgerrit | Adam Reznechek proposed openstack/openstack-ansible-os_nova: [WIP] Update PowerVM driver documentation https://review.openstack.org/364627 | 02:52 |
| *** thorst has quit IRC | 02:54 | |
| *** v1k0d3n has joined #openstack-ansible | 02:57 | |
| *** asettle has joined #openstack-ansible | 03:05 | |
| *** schwicht has quit IRC | 03:07 | |
| *** bkreitch has quit IRC | 03:09 | |
| *** asettle has quit IRC | 03:10 | |
| *** v1k0d3n has quit IRC | 03:13 | |
| *** javis has quit IRC | 03:18 | |
| openstackgerrit | Michael Carden proposed openstack/openstack-ansible-ceph_client: Fix incorrect project reference https://review.openstack.org/364641 | 03:19 |
| *** wangqwsh has joined #openstack-ansible | 03:32 | |
| *** zerda2 has joined #openstack-ansible | 03:38 | |
| *** intr1nsic has quit IRC | 03:43 | |
| *** haasn has quit IRC | 03:44 | |
| *** gouthamr_ has quit IRC | 03:45 | |
| *** intr1nsic has joined #openstack-ansible | 03:46 | |
| *** haasn has joined #openstack-ansible | 03:49 | |
| *** thorst has joined #openstack-ansible | 03:52 | |
| openstackgerrit | Adam Reznechek proposed openstack/openstack-ansible-galera_server: Fix arch-specific percona-xtrabackup and qpress https://review.openstack.org/364066 | 03:52 |
| *** thorst has quit IRC | 03:59 | |
| openstackgerrit | Kevin Carter (cloudnull) proposed openstack/openstack-ansible: Enable the opportunistic strategy https://review.openstack.org/349485 | 04:07 |
| *** pmannidi has quit IRC | 04:09 | |
| *** pmannidi has joined #openstack-ansible | 04:10 | |
| *** wangqwsh has quit IRC | 04:13 | |
| *** hblah has quit IRC | 04:20 | |
| *** jperry has joined #openstack-ansible | 04:23 | |
| *** csmart has quit IRC | 04:35 | |
| *** evrardjp has quit IRC | 04:36 | |
| *** d34dh0r53 has quit IRC | 04:36 | |
| *** cloudnull has quit IRC | 04:36 | |
| *** v1k0d3n has joined #openstack-ansible | 04:36 | |
| *** stevelle has quit IRC | 04:36 | |
| *** cjloader has quit IRC | 04:37 | |
| *** sigmavirus has quit IRC | 04:37 | |
| *** FrankZhang has quit IRC | 04:37 | |
| *** palendae has quit IRC | 04:38 | |
| *** stevelle has joined #openstack-ansible | 04:41 | |
| *** cloader89 has joined #openstack-ansible | 04:42 | |
| *** FrankZhang has joined #openstack-ansible | 04:43 | |
| *** _sigmavirus24 has joined #openstack-ansible | 04:43 | |
| *** cloudnull has joined #openstack-ansible | 04:45 | |
| *** dkehn_ has quit IRC | 04:47 | |
| *** d34dh0r53 has joined #openstack-ansible | 04:47 | |
| *** evrardjp has joined #openstack-ansible | 04:48 | |
| *** hybridpollo has quit IRC | 04:50 | |
| *** shausy has joined #openstack-ansible | 04:53 | |
| *** eil397 has joined #openstack-ansible | 04:54 | |
| *** fxpester has quit IRC | 04:55 | |
| *** thorst has joined #openstack-ansible | 04:56 | |
| *** v1k0d3n has quit IRC | 05:01 | |
| *** eil397 has quit IRC | 05:04 | |
| *** thorst has quit IRC | 05:04 | |
| *** v1k0d3n has joined #openstack-ansible | 05:06 | |
| -openstackstatus- NOTICE: The logs volume is full causing jobs to fail with POST_FAILURE. This is being worked on, please do not recheck until notified. | 05:07 | |
| *** ChanServ changes topic to "The logs volume is full causing jobs to fail with POST_FAILURE. This is being worked on, please do not recheck until notified." | 05:07 | |
| *** dolphm has quit IRC | 05:07 | |
| *** lbragstad has quit IRC | 05:09 | |
| openstackgerrit | Steve Lewis (stevelle) proposed openstack/openstack-ansible-os_swift: Refactor and improve swift-rings.py https://review.openstack.org/351227 | 05:10 |
| *** dolphm has joined #openstack-ansible | 05:13 | |
| *** lbragstad has joined #openstack-ansible | 05:18 | |
| *** waht has quit IRC | 05:20 | |
| *** dkehn has joined #openstack-ansible | 05:22 | |
| *** javeriak has joined #openstack-ansible | 05:33 | |
| *** ChanServ changes topic to "Launchpad: https://launchpad.net/openstack-ansible || Weekly Meetings: https://wiki.openstack.org/wiki/Meetings/openstack-ansible || Review Dashboard: https://goo.gl/tTmdgs" | 05:33 | |
| -openstackstatus- NOTICE: Space has been freed up on the log server. If you have POST_FAILURE results it is now safe to issue a 'recheck' | 05:33 | |
| *** v1k0d3n has quit IRC | 05:36 | |
| *** javeriak_ has joined #openstack-ansible | 05:37 | |
| *** javeriak has quit IRC | 05:38 | |
| *** palendae has joined #openstack-ansible | 05:42 | |
| odyssey4me | stevelle still around? | 05:52 |
| stevelle | odyssey4me: vaguely | 05:53 |
| odyssey4me | heh, yeah - so it seems that you're up to speed on the rabbit thingy | 05:53 |
| odyssey4me | are you aware of how all the other services are getting the ssl flag? | 05:53 |
| stevelle | we talking about the rabbit_use_ssl flag? | 05:54 |
| stevelle | or is there another? | 05:54 |
| odyssey4me | no, that - it turns out - is the only way to make it work at the moment | 05:55 |
| odyssey4me | passing that option in the transport_url doesn't work | 05:55 |
| odyssey4me | it has to be in the oslo moddleware config section | 05:55 |
| odyssey4me | if there's something equivalent for swift, then that may work | 05:55 |
| stevelle | oslo middleware config section? | 05:56 |
| stevelle | thought it was looking in the oslo_messaging_rabbit group | 05:57 |
| odyssey4me | apologies, I mean the oslo_messaging_rabbit section | 05:57 |
| * odyssey4me gets coffee | 05:57 | |
| stevelle | cloudnull said he tried adding that group to the swift proxy config but no luck | 05:58 |
| odyssey4me | yeah, swift is different - of course... | 06:00 |
| odyssey4me | any luck in figuring out whether there's a place to put the option, or somewhere to patch it? | 06:00 |
| *** csmart_ has joined #openstack-ansible | 06:00 | |
| stevelle | not yet. | 06:01 |
| stevelle | I've been digging through oslo messaging and config trying to find an explanation for why adding an oslo_messaging_rabbit section isn't working | 06:02 |
| stevelle | I have some hunches but they are not developed | 06:02 |
| *** csmart_ is now known as csmart | 06:02 | |
| stevelle | I was hoping to ask some questions today. | 06:02 |
| *** thorst has joined #openstack-ansible | 06:02 | |
| openstackgerrit | Jesse Pretorius (odyssey4me) proposed openstack/openstack-ansible: Revert "Disable SSL use for RabbitMQ" https://review.openstack.org/364673 | 06:03 |
| stevelle | odyssey4me: the old format, pre transport_url still works but is deprecated? | 06:06 |
| openstackgerrit | Jesse Pretorius (odyssey4me) proposed openstack/openstack-ansible: Fixed hosts inclusion when requiring the lxc_hosts role https://review.openstack.org/363759 | 06:06 |
| *** jperry has quit IRC | 06:07 | |
| openstackgerrit | Jesse Pretorius (odyssey4me) proposed openstack/openstack-ansible: Fixed hosts inclusion when requiring the lxc_hosts role https://review.openstack.org/363759 | 06:08 |
| *** Prabhu has joined #openstack-ansible | 06:08 | |
| Prabhu | Hi csmart, odyssey4me | 06:08 |
| *** thorst has quit IRC | 06:09 | |
| odyssey4me | stevelle yes, apparently the new format's processing is incomplete which is why we have to leave some of the old options in place | 06:09 |
| odyssey4me | o/ Prabhu | 06:09 |
| Prabhu | openstack-ansible os-cinder-install.yml was successful. what is the next action ? | 06:09 |
| odyssey4me | Prabhu you can check the contents of setup-openstack.yml - it shows you the sequence there | 06:11 |
| odyssey4me | or you can just execute setup-openstack.yml | 06:12 |
| odyssey4me | cd /opt/openstack-ansible/playbooks; openstack-ansible setup-openstack.yml | 06:12 |
| *** jperry has joined #openstack-ansible | 06:16 | |
| *** bsv has joined #openstack-ansible | 06:16 | |
| csmart | Prabhu: hi | 06:19 |
| Prabhu | Hi csmart | 06:19 |
| csmart | Prabhu: FYI, I was running my vms with 16GB ram and 8 vcpus | 06:19 |
| openstackgerrit | Merged openstack/openstack-ansible-os_swift: Refactor and improve swift-rings.py https://review.openstack.org/351227 | 06:19 |
| Prabhu | csmart, was it successfull for you? | 06:21 |
| *** rgogunskiy has joined #openstack-ansible | 06:24 | |
| csmart | Prabhu: yes, apart from the error on my blog in generating the NO_PROXY varable (which I mentioned to you and is now updated on the my blog), it has always worked for me | 06:25 |
| csmart | but then I also don't use vmware | 06:25 |
| *** janki has joined #openstack-ansible | 06:26 | |
| *** Prabhu has quit IRC | 06:28 | |
| openstackgerrit | Jesse Pretorius (odyssey4me) proposed openstack/openstack-ansible: Revert "Disable SSL use for RabbitMQ" https://review.openstack.org/364673 | 06:30 |
| openstackgerrit | Jesse Pretorius (odyssey4me) proposed openstack/openstack-ansible: Add hostname adjustment to liberty>mitaka upgrades https://review.openstack.org/360539 | 06:33 |
| *** thorst has joined #openstack-ansible | 06:36 | |
| *** thorst has quit IRC | 06:44 | |
| *** pcaruana has joined #openstack-ansible | 06:48 | |
| *** berendt has joined #openstack-ansible | 06:50 | |
| *** drifterza has joined #openstack-ansible | 06:55 | |
| *** rgogunskiy has quit IRC | 07:03 | |
| *** rgogunskiy has joined #openstack-ansible | 07:04 | |
| *** javeriak_ has quit IRC | 07:05 | |
| *** javeriak has joined #openstack-ansible | 07:05 | |
| *** brad[] has quit IRC | 07:09 | |
| *** brad[] has joined #openstack-ansible | 07:10 | |
| *** fxpester has joined #openstack-ansible | 07:11 | |
| *** javeriak has quit IRC | 07:19 | |
| *** askb has joined #openstack-ansible | 07:23 | |
| *** rgogunskiy_ has joined #openstack-ansible | 07:25 | |
| *** javeriak has joined #openstack-ansible | 07:26 | |
| *** rgogunskiy has quit IRC | 07:28 | |
| *** rgogunskiy_ is now known as rgogunskiy | 07:28 | |
| *** javeriak has quit IRC | 07:31 | |
| *** jperry has quit IRC | 07:31 | |
| *** Oku_OS-away is now known as Oku_OS | 07:33 | |
| *** neilus has joined #openstack-ansible | 07:41 | |
| *** thorst has joined #openstack-ansible | 07:42 | |
| *** thorst has quit IRC | 07:48 | |
| *** jperry has joined #openstack-ansible | 07:52 | |
| *** bsv has quit IRC | 07:58 | |
| *** asettle has joined #openstack-ansible | 08:07 | |
| openstackgerrit | Merged openstack/openstack-ansible: Add hostname adjustment to liberty>mitaka upgrades https://review.openstack.org/360539 | 08:12 |
| *** neilus has quit IRC | 08:12 | |
| *** jperry has quit IRC | 08:13 | |
| drifterza | Hi guys | 08:15 |
| *** neilus has joined #openstack-ansible | 08:21 | |
| *** javeriak has joined #openstack-ansible | 08:26 | |
| *** karimb has joined #openstack-ansible | 08:28 | |
| odyssey4me | o/ drifterza | 08:31 |
| drifterza | hows things today odyssey4me | 08:31 |
| odyssey4me | not too bad, just so much still to do for the newton release | 08:31 |
| odyssey4me | any chance you can help work on it? https://etherpad.openstack.org/p/osa-newton-rc1 | 08:31 |
| *** berendt has quit IRC | 08:32 | |
| drifterza | Yeah sure | 08:32 |
| drifterza | I can give it a go | 08:32 |
| drifterza | Ill give trove a rest for a bit | 08:33 |
| drifterza | would love other people to help working on that role | 08:33 |
| odyssey4me | yeah, no-one's going to place much attention there until newton is released | 08:33 |
| odyssey4me | new feature development is typically done within the first two milestones | 08:34 |
| *** pjm6 has joined #openstack-ansible | 08:34 | |
| pjm6 | good morning all | 08:34 |
| odyssey4me | I know that there are definitely several parties interested in making trove happen, so don't dispair just yet. | 08:34 |
| *** dkehn has quit IRC | 08:39 | |
| *** berendt has joined #openstack-ansible | 08:44 | |
| *** thorst has joined #openstack-ansible | 08:46 | |
| openstackgerrit | Alexandra Settle proposed openstack/openstack-ansible: [DOC]Added missing yml file and example for test and production environments. https://review.openstack.org/363947 | 08:49 |
| *** berendt has quit IRC | 08:50 | |
| *** thorst has quit IRC | 08:54 | |
| *** javeriak has quit IRC | 08:57 | |
| *** dkehn_ has joined #openstack-ansible | 08:57 | |
| *** pmannidi has quit IRC | 09:29 | |
| *** javeriak has joined #openstack-ansible | 09:35 | |
| openstackgerrit | Andy McCrae proposed openstack/openstack-ansible-tests: Default the network-conf template file name https://review.openstack.org/364794 | 09:38 |
| openstackgerrit | Andy McCrae proposed openstack/openstack-ansible-os_keystone: Use the central test repository for Keystone https://review.openstack.org/364796 | 09:41 |
| *** berendt has joined #openstack-ansible | 09:48 | |
| openstackgerrit | Andy McCrae proposed openstack/openstack-ansible-tests: Default the network-conf template file name https://review.openstack.org/364794 | 09:50 |
| openstackgerrit | Andy McCrae proposed openstack/openstack-ansible-os_keystone: Use the central test repository for Keystone https://review.openstack.org/364796 | 09:51 |
| *** thorst has joined #openstack-ansible | 09:52 | |
| *** javeriak has quit IRC | 09:52 | |
| stevelle | odyssey4me: what about your revert for rabbit ssl keeps swift from being broken? was there another patch I didn't find to turn it off? | 09:53 |
| odyssey4me | stevelle the group vars for swift's telemetry bits are unlike all the others - they hard code the port and hard set use_ssl to false: https://github.com/openstack/openstack-ansible/blob/master/playbooks/inventory/group_vars/all.yml#L382-L389 | 09:54 |
| odyssey4me | with that, it's safe to turn rabbitmq_use_ssl back to True because swift doesn't use that setting | 09:54 |
| stevelle | got it | 09:55 |
| stevelle | thx | 09:55 |
| odyssey4me | sure thing :) | 09:55 |
| *** thorst has quit IRC | 09:58 | |
| *** markvoelker has joined #openstack-ansible | 10:10 | |
| *** drifterza has quit IRC | 10:17 | |
| *** hblah has joined #openstack-ansible | 10:19 | |
| odyssey4me | asettle perhaps we should name the 'developer documentaiton' the 'developer guide' and do the same for the upgrade documentation? | 10:21 |
| odyssey4me | ie call them all guides | 10:21 |
| asettle | odyssey4me: my only push back on that is that the developer documentation is *just* that, it is not a guide. Also, keeping in context with the rest of the OS projects, each have developer documentation. I guess you don't want to make the cross over too difficult. | 10:22 |
| asettle | Agree with the Upgrade Guide however | 10:22 |
| *** markvoelker has quit IRC | 10:22 | |
| odyssey4me | asettle ok, down with that :) the dev docs are a hodge podge | 10:23 |
| odyssey4me | a hodge podge of guidance stuff... but hardly a story :p | 10:23 |
| asettle | Pretty much, yeah haha | 10:23 |
| asettle | It's kind of why it was given the overarching term 'dev docs' (What even does that meannnn) | 10:24 |
| asettle | odyssey4me: I can patch up the change for the Upgrade Guide if you wish? | 10:24 |
| odyssey4me | asettle http://i.giphy.com/wi8Ez1mwRcKGI.gif | 10:26 |
| *** markvoelker has joined #openstack-ansible | 10:30 | |
| *** LanceHaig has joined #openstack-ansible | 10:35 | |
| *** markvoelker has quit IRC | 10:40 | |
| *** hblah has quit IRC | 10:45 | |
| openstackgerrit | Alexandra Settle proposed openstack/openstack-ansible: [DOCS] Rename upgrade documentation to upgrade guide https://review.openstack.org/364831 | 10:47 |
| asettle | odyssey4me: ready for mergy mergy | 10:54 |
| cloudnull | mornings | 10:55 |
| evrardjp | morning cloudnull | 10:55 |
| cloudnull | hows it evrardjp | 10:55 |
| evrardjp | fine, what about you? | 10:56 |
| cloudnull | still waking. | 10:56 |
| *** thorst has joined #openstack-ansible | 10:56 | |
| cloudnull | sadly could not sleep more | 10:56 |
| *** berendt has quit IRC | 10:56 | |
| evrardjp | well I know the feeling :) | 10:56 |
| andymccr | damn you centos | 10:58 |
| evrardjp | :p | 11:00 |
| *** neilus has quit IRC | 11:01 | |
| *** thorst has quit IRC | 11:04 | |
| cloudnull | so w/ a multi-node install and no gnocchi we're still seeing :stderr: Could not find a version that satisfies the requirement gnocchiclient (from versions: ) ... | 11:05 |
| andymccr | cloudnull: this needs a new openstack project | 11:05 |
| cloudnull | i had thought that was fixed, but im guessing that its fixed only for the aio | 11:06 |
| openstackgerrit | Merged openstack/openstack-ansible: Fixed hosts inclusion when requiring the lxc_hosts role https://review.openstack.org/363759 | 11:06 |
| openstackgerrit | Merged openstack/openstack-ansible: [DOC]Added missing yml file and example for test and production environments. https://review.openstack.org/363947 | 11:06 |
| cloudnull | andymccr: we could call it the "adventures of the mighty ceilometer" | 11:07 |
| *** haasn has quit IRC | 11:07 | |
| *** lkoranda has quit IRC | 11:07 | |
| *** afred312 has quit IRC | 11:07 | |
| *** darrenc has quit IRC | 11:07 | |
| odyssey4me | cloudnull I was thinking that the utility package list should be populated based on whether there are host group members for the applicable group | 11:08 |
| odyssey4me | it'll be a bit messy, but that should solve the issue | 11:08 |
| *** afred312 has joined #openstack-ansible | 11:09 | |
| *** darrenc has joined #openstack-ansible | 11:09 | |
| *** neilus has joined #openstack-ansible | 11:10 | |
| odyssey4me | cloudnull something like http://pastebin.com/QF6LJLM1 as a quick fix | 11:11 |
| odyssey4me | evrardjp andymccr thoughts? | 11:11 |
| evrardjp | well I don't know gnocchi groups | 11:12 |
| evrardjp | but the pattern seems ok | 11:12 |
| *** lkoranda has joined #openstack-ansible | 11:12 | |
| odyssey4me | seeing as the pip install task joins all the list items together in a long string, the empty items that are there if the group isn't populated will not be an issue | 11:13 |
| odyssey4me | that pattern should then actually be applied to heatclient, ironicclient, saharaclient, etc | 11:13 |
| evrardjp | agreed as a quick fix | 11:14 |
| evrardjp | but I think we could improve our lookup for that | 11:14 |
| evrardjp | I can't find gnocchi_all group so we should probably use ceilometer_all | 11:14 |
| cloudnull | odyssey4me: we know the clients all have a similar name. | 11:14 |
| cloudnull | we could do something like (curl http://172.29.236.200:8181/os-releases/master/requirements_absolute_requirements.txt | grep client) | xargs pip install -U | 11:14 |
| cloudnull | which would pull a list from the known repo and just install them | 11:15 |
| odyssey4me | yeah, I suppose if we have each role itemise the required client packages, then we could use the py_pkgs lookup to actually just run through the roles and install the client packages from each role which has a populated host group | 11:15 |
| evrardjp | clounull, yes, plus we can use group length to remove things | 11:15 |
| odyssey4me | even now we could provide the utility install play with the py_pkgs lookup, then install '.*client$' from each role that has a populated group | 11:16 |
| evrardjp | yup | 11:16 |
| *** haasn has joined #openstack-ansible | 11:16 | |
| odyssey4me | so we build a list from a base of some sort - like 'python-openstackclient', then loop through the roles for each populated host group and add any '.*client$' packages into the list, then finally pip install the list | 11:17 |
| evrardjp | why the hell did gnocchi choose gnocchiclient instead of python-gnocchiclient | 11:19 |
| stevelle | umm, it's shorter? :D | 11:19 |
| * stevelle doesn't know the answer | 11:19 | |
| odyssey4me | because french? ;) | 11:19 |
| cloudnull | ^ | 11:20 |
| evrardjp | we should create a metapackage ( I don't know if that exists) | 11:20 |
| *** javeriak has joined #openstack-ansible | 11:20 | |
| evrardjp | to point with standard name :p | 11:20 |
| stevelle | can we call it python-gnochiclient-client? | 11:20 |
| odyssey4me | andymccr I think the lxc_hosts/lxc_container_create roles have var and template structures which you could model from for the network config that covers both centos and ubuntu? | 11:20 |
| evrardjp | description: "Unofficial Gnocchi client. Because standard name. Also, not french." | 11:21 |
| odyssey4me | evrardjp you make belgium proud :p | 11:21 |
| andymccr | odyssey4me: it should be ok its just that centos you have to have separated conf files, whereas ubuntu you can just drop a "all_my_config.cfg" | 11:22 |
| evrardjp | I didn't get the sarcasm on this one, so I guess "your welcome?" | 11:22 |
| andymccr | its just keystone etc that is centos atm so i will resolve for central testing and move on | 11:22 |
| odyssey4me | andymccr yeah, keystone and almost all the infrastructure bits (rabbit, etc) | 11:23 |
| odyssey4me | and a patch is up for glance already | 11:23 |
| evrardjp | andymccr: which bits? | 11:23 |
| andymccr | odyssey4me: yeah thats ok but they all only use 1 interface | 11:24 |
| openstackgerrit | Kevin Carter (cloudnull) proposed openstack/openstack-ansible: Fixed assumed utility pip install for specific clients https://review.openstack.org/364849 | 11:24 |
| cloudnull | ^ odyssey4me evrardjp like so | 11:24 |
| andymccr | as a first pass i'll get it to work - then i'll change it up to allow multiple interfaces/settings for rhel | 11:24 |
| odyssey4me | andymccr ubuntu can also have multiple files - we do that for the containers | 11:24 |
| cloudnull | now it looks up the repo based on the known pip_link adn installs just that | 11:24 |
| andymccr | odyssey4me: i know it "can" but for the purposes of laying down testing its easier not to | 11:24 |
| odyssey4me | so maybe we should just do a file per interface? | 11:24 |
| odyssey4me | fair enough | 11:24 |
| evrardjp | cloudnull: you're gonna have to install other clients you don't want | 11:25 |
| odyssey4me | cloudnull that is a monstrously dirty hack : | 11:25 |
| odyssey4me | :p | 11:25 |
| cloudnull | but works well. | 11:25 |
| cloudnull | :) | 11:25 |
| cloudnull | evrardjp: like what ? | 11:25 |
| evrardjp | ndg_httpsclient oauth2client | 11:26 |
| evrardjp | os_client_config | 11:26 |
| evrardjp | it's fine 'though | 11:26 |
| cloudnull | http://cdn.pasteraw.com/ksdhtvu5cbqz9mm8ohng26uyujfmzpw | 11:26 |
| cloudnull | thats the list based on my current master. | 11:26 |
| evrardjp | fine :p | 11:27 |
| cloudnull | LGTM | 11:27 |
| evrardjp | prometheus_client | 11:27 |
| evrardjp | interesting :D | 11:27 |
| cloudnull | yea. idk where that came in | 11:27 |
| cloudnull | but there it is | 11:27 |
| evrardjp | probably the monasca | 11:27 |
| *** karimb has quit IRC | 11:27 | |
| evrardjp | sorry | 11:27 |
| evrardjp | probably LE monasca | 11:27 |
| odyssey4me | cloud if you make the regex look for client at the end only, it'll be just a tiny bit cleaner | 11:28 |
| evrardjp | that's what I meant | 11:28 |
| evrardjp | but I think two or 3 packages more isn't a big deal | 11:28 |
| odyssey4me | the downside of this is that we lose any set pip install options | 11:28 |
| odyssey4me | so perhaps the list should be gathered as a fact, then the current pip install task used? | 11:29 |
| evrardjp | what do you mean? | 11:29 |
| cloudnull | evrardjp: its nothing we're specifically specifying so it has to be a dep | 11:29 |
| evrardjp | cloudnull: yes I'm pretty sure of it too | 11:29 |
| odyssey4me | yeah - can we rather gather that list (probably without the versions) as a fact, then use the current pip install task so that any proxy configs, etc are provided | 11:30 |
| odyssey4me | bear in mind that it is possible that the repo server is on the other side of a proxy | 11:30 |
| evrardjp | yes, I'd rather use the shell for gathering the items, and then using the pip module with the registered items | 11:31 |
| odyssey4me | also, the patch should remove the current list from group_vars | 11:31 |
| evrardjp | this way we can still have the extra_args | 11:31 |
| cloudnull | i never noticed before but gnocchi and aodh are bad people | 11:31 |
| cloudnull | "aodhclient" | 11:32 |
| evrardjp | same time | 11:32 |
| evrardjp | team* | 11:32 |
| evrardjp | I guess | 11:32 |
| odyssey4me | although I think we should gather that stuff, but also supply a default list of things that must also be installed - like 'python-openstackclient' and the memcached and crypto clients... although I don't know why they need to be there... but it'd be useful to be able to gather a list, but also supply a default set which must always be installed | 11:33 |
| *** schwicht has joined #openstack-ansible | 11:34 | |
| evrardjp | openstackclient will be installed as part of some project I expect | 11:34 |
| evrardjp | same for os_client_config | 11:34 |
| evrardjp | so the only remaining would be memcached and others we want there | 11:34 |
| odyssey4me | evrardjp probably, but there are some items you may wish to always make sure are there regardless of the deployment layout | 11:34 |
| evrardjp | we could basically override the generated list with a "standard" list | 11:35 |
| evrardjp | not override | 11:35 |
| evrardjp | union | 11:35 |
| odyssey4me | union | 11:35 |
| odyssey4me | heh, what you said | 11:35 |
| evrardjp | hatop should be in pip :/ | 11:36 |
| evrardjp | I meant pypi | 11:36 |
| openstackgerrit | Kevin Carter (cloudnull) proposed openstack/openstack-ansible: Fixed assumed utility pip install for specific clients https://review.openstack.org/364849 | 11:37 |
| evrardjp | haproxy-top ! | 11:37 |
| cloudnull | ^ odyssey4me evrardjp now it uses the pip task w/ the registered list | 11:38 |
| odyssey4me | cloudnull now just cut down https://github.com/openstack/openstack-ansible/blob/master/playbooks/inventory/group_vars/utility_all.yml#L43 to a 'default' set, and make that var union with the list | 11:39 |
| odyssey4me | that'll work best if the gathered list is gathered without the version spec | 11:39 |
| evrardjp | awk instead of grep to the rescue! | 11:40 |
| cloudnull | awk will exit 0 if there are no matches. | 11:41 |
| openstackgerrit | Kevin Carter (cloudnull) proposed openstack/openstack-ansible: Fixed assumed utility pip install for specific clients https://review.openstack.org/364849 | 11:43 |
| odyssey4me | cloudnull a small suggestion for consideration in the review | 11:44 |
| openstackgerrit | Andy McCrae proposed openstack/openstack-ansible-tests: Fix the central repository to work with CentOS https://review.openstack.org/364858 | 11:45 |
| cloudnull | andymccr: https://github.com/cloudnull/os-ansible-deployment/tree/master-rhel/etc/network/network-scripts | 11:46 |
| evrardjp | cloudnull: not sure what you meant there as an issue | 11:47 |
| andymccr | cloudnull: cool thanks! | 11:47 |
| cloudnull | ^ idk if you've seen that but i did some work on the RHEL based network scripts for bridges veths and nonsense | 11:47 |
| openstackgerrit | Andy McCrae proposed openstack/openstack-ansible-tests: Fix the central repository to work with CentOS https://review.openstack.org/364858 | 11:48 |
| cloudnull | andymccr: https://github.com/cloudnull/os-ansible-deployment/blob/master-rhel/scripts/bootstrap-aio.sh#L190-L201 -- that was the impl to make it go w/ the aio for testing | 11:48 |
| andymccr | cloudnull: yeah i have a pretty good idea, having the templates ready is good. | 11:48 |
| *** gouthamr has joined #openstack-ansible | 11:49 | |
| andymccr | for now i want to get keystone working - since its just one interface thats easy and i have some other small blockers to fix | 11:49 |
| cloudnull | cool , | 11:49 |
| openstackgerrit | Kevin Carter (cloudnull) proposed openstack/openstack-ansible: Fixed assumed utility pip install for specific clients https://review.openstack.org/364849 | 11:51 |
| cloudnull | odyssey4me: ^ done | 11:51 |
| andymccr | lunch! | 11:52 |
| odyssey4me | cloudnull the last tweak is to only make that task run on one target | 11:53 |
| odyssey4me | so use run_once or whatever | 11:53 |
| openstackgerrit | Kevin Carter (cloudnull) proposed openstack/openstack-ansible: Fixed assumed utility pip install for specific clients https://review.openstack.org/364849 | 11:54 |
| openstackgerrit | Kevin Carter (cloudnull) proposed openstack/openstack-ansible: [TESTING] all is_metal scenario https://review.openstack.org/352024 | 11:55 |
| *** thorst has joined #openstack-ansible | 11:57 | |
| odyssey4me | cloudnull you'll have to skip linting: http://logs.openstack.org/49/364849/5/check/gate-openstack-ansible-linters-ubuntu-xenial/7506e1e/console.html#_2016-09-02_11_56_47_183478 | 11:59 |
| odyssey4me | or use the get_url or uri module | 11:59 |
| *** schwicht has quit IRC | 12:01 | |
| *** neilus1 has joined #openstack-ansible | 12:02 | |
| openstackgerrit | Kevin Carter (cloudnull) proposed openstack/openstack-ansible: Fixed assumed utility pip install for specific clients https://review.openstack.org/364849 | 12:03 |
| cloudnull | ^ lint skipped | 12:03 |
| *** schwicht has joined #openstack-ansible | 12:04 | |
| odyssey4me | cloudnull do you think you have the time or inclination to tackle this? https://etherpad.openstack.org/p/osa-newton-nova-upgrades | 12:04 |
| *** _sigmavirus24 is now known as sigmavirus | 12:05 | |
| *** sigmavirus has joined #openstack-ansible | 12:05 | |
| *** neilus has quit IRC | 12:05 | |
| cloudnull | odyssey4me: no. likely not next week at leasst | 12:08 |
| odyssey4me | cloudnull well, so I thought that maybe I should put the spec up for Ocata, targeting milestone 1 | 12:09 |
| odyssey4me | once it's done, the pattern is established | 12:09 |
| odyssey4me | the work can be started now, perhaps - but we only target completion at Ocata m1 | 12:10 |
| odyssey4me | lemme put a spec up and we can discuss in there | 12:10 |
| *** schwicht has quit IRC | 12:12 | |
| cloudnull | that'd work. i know the osic folks are pushing hard on getting rolling upgrades as a thing. luzC was going to provide me some details on their requirements today. so if that comes through maybe i can add to that etherpad | 12:12 |
| logan- | that is really cool ^ | 12:12 |
| *** asettle has quit IRC | 12:13 | |
| odyssey4me | cloudnull yep, so I'm thinking that we target rolling upgrades for nova/neutron/keystone - as they're the only ones that support it... then all other services have a simple upgrade as mentioned in https://etherpad.openstack.org/p/osa-newton-rc1 | 12:13 |
| cloudnull | as of late i thought it was nova, cinder, swift | 12:14 |
| odyssey4me | oh yeah, swift does too | 12:14 |
| odyssey4me | cinder, I can't seem to find any details for it | 12:14 |
| odyssey4me | neutron's rolling upgrade is pretty simple from what I gather | 12:15 |
| *** woodard has joined #openstack-ansible | 12:15 | |
| *** schwicht has joined #openstack-ansible | 12:15 | |
| *** askb has quit IRC | 12:15 | |
| cloudnull | odyssey4me: from Szimon https://gist.github.com/antwash/f617cfbbc3ba4b18ef1feb7621cef03f | 12:16 |
| cloudnull | regarding cinder. | 12:16 |
| cloudnull | and http://docs.openstack.org/developer/cinder/devref/rolling.upgrades.html | 12:16 |
| odyssey4me | yeah, the devref is useless | 12:16 |
| *** woodard has quit IRC | 12:16 | |
| odyssey4me | but thanks for the gist, that looks more helpful | 12:16 |
| *** woodard has joined #openstack-ansible | 12:17 | |
| cloudnull | hopfully they'll send more over this afternoon | 12:17 |
| logan- | on the services that don't necessarily support rolling upgrades, it would still be nice if we could intelligently set haproxy endpoints to DRAIN on a rolling basis. keystone would be a good candidate since it already targets [0], and [1-n] execution | 12:17 |
| cloudnull | logan-: +1 | 12:18 |
| odyssey4me | would love that logan- | 12:18 |
| cloudnull | that would be nice, | 12:18 |
| odyssey4me | ok, let me work up a spec for all this - we can then add details and then try and recruit some contributors | 12:19 |
| *** javeriak has quit IRC | 12:23 | |
| *** berendt has joined #openstack-ansible | 12:24 | |
| *** ChanServ sets mode: +o cloudnull | 12:24 | |
| *** schwicht has quit IRC | 12:28 | |
| mgariepy | good morning everyone | 12:30 |
| cloudnull | morning | 12:30 |
| matias | odyssey4me: again issues with lxc_container_backing_method undefined, should i add more checks to see if it is defined like what we have done previously? https://github.com/openstack/openstack-ansible-lxc_container_create/blob/master/tasks/container_create.yml#L88 | 12:31 |
| odyssey4me | matias are you using the b3 tag, or the head of master? | 12:34 |
| matias | odyssey4me: head | 12:35 |
| odyssey4me | hmm, that should have been addressed in https://review.openstack.org/363985 | 12:35 |
| odyssey4me | oh hang on - that's a different one | 12:35 |
| odyssey4me | matias yes please add another check | 12:36 |
| matias | ok :P thanks | 12:36 |
| matias | are you guys having trouble with attaching a floating ip to an instance? i had sent this to horizon but 5 days have gone and it got no attention, maybe i'm doing something wrong https://review.openstack.org/#/c/361754 | 12:37 |
| odyssey4me | matias you may have to lobby in the IRC channel for some attention on the patch | 12:38 |
| *** javeriak has joined #openstack-ansible | 12:38 | |
| *** javeriak has quit IRC | 12:38 | |
| odyssey4me | matias #openstack-horizon | 12:39 |
| matias | odyssey4me: i'm already doing that, but their channel has very low traffic | 12:39 |
| matias | odyssey4me: thanks :D | 12:40 |
| odyssey4me | matias there we go - pinged the PTL to ask for help | 12:40 |
| *** kenhui has joined #openstack-ansible | 12:42 | |
| *** hblah has joined #openstack-ansible | 12:43 | |
| *** neilus1 has quit IRC | 12:49 | |
| *** gouthamr_ has joined #openstack-ansible | 12:51 | |
| *** gouthamr has quit IRC | 12:55 | |
| *** admin0 has joined #openstack-ansible | 12:58 | |
| *** schwicht has joined #openstack-ansible | 13:03 | |
| *** berendt has quit IRC | 13:04 | |
| *** shausy has quit IRC | 13:06 | |
| *** neilus has joined #openstack-ansible | 13:08 | |
| openstackgerrit | Andy McCrae proposed openstack/openstack-ansible-tests: Fix the central repository to work with CentOS https://review.openstack.org/364858 | 13:09 |
| admin0 | hi all | 13:09 |
| *** Andrew_jedi has joined #openstack-ansible | 13:09 | |
| *** klamath has joined #openstack-ansible | 13:09 | |
| admin0 | anyone with experinece in openstack big deployments around :D ? | 13:09 |
| admin0 | question .. what is the real thing that people use SDN for .. i hear neutron does not scale .. what is this scale thing ? | 13:10 |
| admin0 | what are the limitations in neutron, and when do they start to act ? | 13:10 |
| *** klamath has quit IRC | 13:11 | |
| *** klamath has joined #openstack-ansible | 13:12 | |
| *** asettle has joined #openstack-ansible | 13:13 | |
| *** BjoernT has joined #openstack-ansible | 13:14 | |
| *** BjoernT is now known as Bjoern_zZzZzZzZ | 13:14 | |
| *** asettle has quit IRC | 13:19 | |
| openstackgerrit | Marc Gariépy proposed openstack/openstack-ansible-openstack_hosts: Only configure systat on debian based system. https://review.openstack.org/364903 | 13:19 |
| mgariepy | wow, it's really fast this morning ! :D | 13:19 |
| mgariepy | git review used to take 1 minute or 2, only took a sec :) | 13:20 |
| mgariepy | cloudnull, can you review when you have time ? https://review.openstack.org/#/c/362881/ | 13:22 |
| *** Bjoern_zZzZzZzZ is now known as BjoernT | 13:22 | |
| *** rgogunskiy has quit IRC | 13:22 | |
| Andrew_jedi | Hello Folks, | 13:23 |
| odyssey4me | admin0 using the Neutron L3 Agent doesn't scale well, so most large deployments will only allow Neutron to manage L2 and will use proper equipment for L3 | 13:23 |
| Andrew_jedi | Please correct me if i am wrong but CentOS support is still a WIP? Right? | 13:23 |
| admin0 | odyssey4me: what large means large ? | 13:24 |
| admin0 | and if you know, what parts of the l3 agent does nto scale well ? | 13:24 |
| odyssey4me | Andrew_jedi yes, very much so - and it won't be done in time for the Newton release | 13:24 |
| odyssey4me | admin0 the L3 parts :p | 13:24 |
| Andrew_jedi | admin0: More than 200 tenant networks | 13:24 |
| admin0 | :D | 13:24 |
| Andrew_jedi | admin0: Flapping floating ip, router issues etc etc | 13:25 |
| admin0 | “Flapping floating ip” — how does this happen ? | 13:25 |
| cloudnull | admin0: we use provider networks w/ vlans | 13:25 |
| Andrew_jedi | odyssey4me: Ahh okk, thanks for letting me know :) | 13:25 |
| *** fxpester has quit IRC | 13:26 | |
| cloudnull | neutron built in L3 will have issues when there are thousands of tenant networks + routers. | 13:26 |
| odyssey4me | admin0 large may mean number of vxlan/gre networks, or might mean the amount of data going through the networks, or might mean number of instance NIC's on the networks... it all depends | 13:26 |
| admin0 | thing is this .. trying to design a big openstack cluster ( say 400 compute nodes per AZ , 2 AZ to start with ) , 10000 instances to scale to .. i think neutron will not scale, but i do not have valid arguments to say otherwise .. | 13:26 |
| *** bgmccollum has quit IRC | 13:26 | |
| cloudnull | also performance wise provider + vlans is a lot better. | 13:27 |
| *** errr has quit IRC | 13:27 | |
| admin0 | so trying to find out blogs or strong use cases to demonstrate or say here i the case study and why we need to go with a SDN | 13:27 |
| odyssey4me | Andrew_jedi we have most of the infrastructure stuff done, and the first OpenStack role (keystone) is done... there's a patch up for Glance... but nothing else has been done yet - it's been a slow going labour of love | 13:27 |
| *** kencjohnston has quit IRC | 13:27 | |
| *** sonus has quit IRC | 13:27 | |
| odyssey4me | Andrew_jedi we're also improving our testing implementation to make it easier to get on with the rest of it - so we've paused the progress to get that done before we continue | 13:27 |
| mgariepy | Andrew_jedi, odyssey4me the lxc stuff isn't working great on centos7 either. (container won't start at boot etc.. ) | 13:28 |
| admin0 | if using vxlan, for the intial packet discovery/path discovery, does it do a broadcast to every other node in the compute layer-2 domain ? | 13:28 |
| odyssey4me | admin0 Neutron itself is fine, but OVS or LXB can't be expected to handle the kind of load you would typically expect from a commercial piece of tin from a network vendor that implements specialised hardware to improve performance. | 13:29 |
| *** woodard has quit IRC | 13:29 | |
| admin0 | that i understand and trying to make the same arguments, but i need to know what the loads are . where does those limitation lie you know | 13:30 |
| logan- | i've heard of scaling issues with neutron also from the rabbit side when you start trying to launch lots of L3 or DHCP agents. it was something that was brought up during the NYC ops meetup last week | 13:30 |
| odyssey4me | admin0 the recommended vxlan implementation is to enable IGMP on the switches and to use multicast discovery for vxlan afaik... evrardjp knows a bit more about this stuff than I do | 13:30 |
| evrardjp | when you build your vxlan tunnels endpoints, you use a multicast address, so it's better to have igmp snooping enabled on the switches | 13:31 |
| odyssey4me | logan- yep - large scale implementations recommend using rabbit clusters per service, which is why automagically did that work recently to ensure that it was easy to do so | 13:31 |
| Andrew_jedi | odyssey4me: I understand that Xenial support is priority for Newton. It would have been great if somebody from RedHat would have cared for the CentOS support in OSA. | 13:31 |
| admin0 | another argument is “we want to keep openstack as close to standard as possible .. so what happens when X sdn vendor stops “ so what are the general criteria for selecting a sdn — there seems to be a lot of them - and is anyone outshines the other ? | 13:32 |
| odyssey4me | at the very least put telemetry on its own MQ cluster, but if you're doing a lot of instance/port churn then you should probably seperate nova and neutron too | 13:32 |
| *** kencjohnston has joined #openstack-ansible | 13:32 | |
| evrardjp | admin0: so to answer your question, yes, depending on how the grouping is supported on the switches | 13:32 |
| evrardjp | or not | 13:32 |
| evrardjp | all the way I mean | 13:32 |
| odyssey4me | Andrew_jedi we've had some assistance from RH people - but their focus is much more on using the openstack packages and packstack... neithr of which we do | 13:32 |
| *** schwicht has quit IRC | 13:33 | |
| *** bgmccollum has joined #openstack-ansible | 13:33 | |
| logan- | admin0: if you're trying to achieve large scale and don't need L2 between VMs, calico is something to look at. we're using it in a setup where our computes do not have any L2 between VMs, the compute host is a router and it announces the VMs it is hosting into our ibgp | 13:33 |
| *** sonus has joined #openstack-ansible | 13:33 | |
| evrardjp | admin0 all of it depends on the amount of tenants you'll host | 13:33 |
| admin0 | logan-: the requirement is to be able to do HA ( both Public and internal ) spanning across 2 intial and maybe a 3rd dc | 13:33 |
| odyssey4me | admin0 the selection criteria should then be to use a vendor who has a continued commitment to a neutron driver which delivers what they need | 13:34 |
| evrardjp | and weird fancy features you have | 13:34 |
| admin0 | standard openstack, lets say 8000-10000 tenants | 13:34 |
| admin0 | mitaka or neuton | 13:34 |
| admin0 | by the time its decided | 13:34 |
| evrardjp | that's not standard openstack | 13:34 |
| odyssey4me | there is no such thing as a 'standard' openstack implementation | 13:34 |
| evrardjp | 10k tenants is big | 13:34 |
| evrardjp | well there is no miracle there | 13:35 |
| *** errr has joined #openstack-ansible | 13:35 | |
| admin0 | ok .. standard “follow the docs” mitaka-ansible playbooks :) | 13:35 |
| evrardjp | vlan won't scale more than 4k tenants :p | 13:35 |
| admin0 | which is why vxlan | 13:35 |
| odyssey4me | well, 4k networks... so assuming 1 network per tenant, yes | 13:35 |
| evrardjp | yes :p | 13:35 |
| *** galstrom_zzz is now known as galstrom | 13:36 | |
| evrardjp | admin0: well that's architecturing the cloud for your needs, it's a specific job | 13:36 |
| admin0 | does rackspace provides this as well ? | 13:36 |
| evrardjp | if we don't have all the info, it's hard to guess | 13:36 |
| admin0 | like a commercial service ? | 13:36 |
| admin0 | or can i bribe one of you guys with pizza to help me with :D | 13:36 |
| evrardjp | rackspace has a product called rackspace private cloud, based on openstack-ansible | 13:37 |
| cloudnull | Andrew_jedi: ++ RE: RHEL support. | 13:37 |
| cloudnull | a few of us tried to make it go. | 13:37 |
| evrardjp | I'm pretty sure some would be interested to talk with you about how to design this for you :) | 13:37 |
| admin0 | whats the biggest one out there,and does it required sdn or did any outgrew neutron ? | 13:37 |
| cloudnull | i hope to get more of it done but it doesn't look like it will be complete by the end of the cycle | 13:37 |
| *** bkreitch has joined #openstack-ansible | 13:38 | |
| cloudnull | admin0: in the OSIC we're using linuxbridge and we're a ~500 node cloud. | 13:38 |
| evrardjp | not sure to understand what you mean admin0 | 13:38 |
| admin0 | did any of the private clouds in rackspace hit neutron limitations ? | 13:38 |
| evrardjp | cloudnull: but are you a 8k tenants network ? :p | 13:38 |
| evrardjp | cloud* | 13:39 |
| cloudnull | no. | 13:39 |
| mgariepy | cloudnull, Andrew_jedi for centos is there a list of tasks that needs to be done ? | 13:39 |
| cloudnull | we do have 1500 networks | 13:39 |
| evrardjp | still less than 4k | 13:39 |
| openstackgerrit | Andy McCrae proposed openstack/openstack-ansible-tests: Fix the central repository to work with CentOS https://review.openstack.org/364858 | 13:39 |
| cloudnull | not quite 8k but its getting there. | 13:39 |
| admin0 | containers - how do they fit in ? does each have its own network as well ? | 13:39 |
| admin0 | or its a part of a tenant network ? | 13:39 |
| matias | i'm curious about this automagically's multi-cluster rabbit stuff.. is there some documentation on setting this up? | 13:40 |
| evrardjp | anyway, admin0 I think openstack-ansible can have large implementations | 13:40 |
| evrardjp | really large ones | 13:40 |
| cloudnull | mgariepy: I think the multi-distro etherpad is up-todate? | 13:40 |
| evrardjp | then it's a matter of optimisations | 13:40 |
| cloudnull | keystone just merged the otherday | 13:40 |
| odyssey4me | matias not at this point - maybe automagically will put together a blog post, otherwise it's relatively easy to see in the code... | 13:40 |
| admin0 | evrardjp: given that i select openstack-ansible, can i get commercial help then ? | 13:40 |
| cloudnull | so now we can get the rest of the roles into shape | 13:40 |
| admin0 | you guys are into cloud, or 3rd party public cloud is also fine ? | 13:41 |
| odyssey4me | matias we just allow you to define the host group for each RPC service to use for setting things up | 13:41 |
| matias | odyssey4me: do you have a filename/ line number handy? | 13:41 |
| evrardjp | admin0: rackspace private cloud is one commercial implementation :D | 13:41 |
| evrardjp | admin0: you should talk to a rackspace commercial representative :) | 13:42 |
| admin0 | is he here on this chat ? | 13:42 |
| odyssey4me | admin0 no, this is an open source project chat room | 13:42 |
| admin0 | well before i talk .. is it per support, per project or pay-as-you-grow-per-node type :) | 13:42 |
| odyssey4me | sorry, but please don't try and do commercial discussions here | 13:42 |
| admin0 | okie | 13:42 |
| odyssey4me | also, those of us here from RAX are not involved in sales, or aware of how that all works | 13:42 |
| admin0 | ok | 13:43 |
| Andrew_jedi | cloudnull: I tried to convince a certain company to contribute on the centos work in OSA but i was not successful. | 13:43 |
| odyssey4me | I would recommend you go through the standard sales channel. | 13:43 |
| Andrew_jedi | mgariepy: https://etherpad.openstack.org/p/openstack-ansible-multi-os-support | 13:43 |
| cloudnull | Andrew_jedi: :( | 13:43 |
| cloudnull | thanks for trying though :) | 13:43 |
| admin0 | i will stick to known openstack-ansible limitations in terms of neutron-at-scale :) | 13:43 |
| odyssey4me | Andrew_jedi cloudnull mgariepy as far as I recall, we haven't really made a hit list for CentOS - the focus was on Xenial | 13:43 |
| *** jperry has joined #openstack-ansible | 13:44 | |
| odyssey4me | admin0 the openstack limitations are much the same as any standard network limitations - openstack neutron simply manages what you have | 13:44 |
| odyssey4me | admin0 if all you have is linux servers, then you'r stuck with the limits of those servers | 13:44 |
| odyssey4me | if you want broader scale networking, then look for commercial vendor tins with neutron L2/L3 drivers | 13:45 |
| Andrew_jedi | admin0: You can look at the blog posts from Cern, architecture blogs from Rackspace and open contrail for sdn. AFAIR Cern dont use neutron. | 13:45 |
| odyssey4me | Andrew_jedi yeah, most large scale implementations use VLANs or flat networks | 13:45 |
| Andrew_jedi | cloudnull: Ohh i am going to try again in Barcelona :) | 13:45 |
| admin0 | i will be in barcelona | 13:46 |
| *** deadnull has joined #openstack-ansible | 13:46 | |
| admin0 | as well | 13:46 |
| odyssey4me | RAX public cloud does use virtual networking, but has a very heavily customised SDN controller, driver and agent I believe | 13:46 |
| odyssey4me | public cloud scale is not something you can do for free - you need to invest in something with the resources and skills you have when it comes to a public cloud implementation | 13:47 |
| Andrew_jedi | odyssey4me: I think most openstack consulting companies recommend running Neutron with a Vendor plugin. | 13:47 |
| odyssey4me | admin0 here is an example of what might work for you with the right tin: http://blogs.cisco.com/datacenter/scaling-openstack-l3-using-cisco-asr1k-platform | 13:48 |
| *** schwicht has joined #openstack-ansible | 13:48 | |
| *** hblah has quit IRC | 13:48 | |
| odyssey4me | admin0 I believe that plumgrid has a pretty reasonable solution too | 13:49 |
| odyssey4me | admin0 we also have instrumentation for Nuage | 13:49 |
| matias | for a 20 compute node, 3 infra node cloud with < 200 tenant networks, would the default (vxlan + lxb) implementation be enough? | 13:50 |
| Andrew_jedi | odyssey4me: But then i guess one has to use the company specific physical switches or routers? | 13:50 |
| admin0 | odyssey4me: the question is about “what demonstrated scale” we are talking about that we need to look for solutions in the first place .. Andrew_jedi mentioned flapping_floating ips and routers ( so trying to find out exactly what ) to build up a case | 13:50 |
| odyssey4me | matias yeah, that should easily work - our standard LXB implementation is running OSIC cloud1 which is larger than that | 13:50 |
| evrardjp | matias: I'd be enclined to say yes | 13:51 |
| evrardjp | but then depends on what you need | 13:51 |
| matias | nice, thanks guys | 13:51 |
| evrardjp | jitter, BW, etc | 13:51 |
| odyssey4me | admin0 even if you split the L3 Agents on to dedicated hardware, the fact that it all has to go through iptables and use linux namespaces on 'server' hardware becomes a hindrence at larger scale | 13:52 |
| admin0 | doesn’t DVR solve this path ? | 13:52 |
| odyssey4me | then you also have the issue of when you're doing upgrades you may end up with data plane down time for the L3 bits when your packages update or your agents restart | 13:52 |
| logan- | DVR introduces other issues ie the rabbit chatter thing I brought up earlier because you're running a L3 agent on every node | 13:53 |
| odyssey4me | ^ that | 13:53 |
| Andrew_jedi | admin0: I meant that with standard L3 agent people have reported issues like floating ip not accessible for a certain period of time and then coming back automatically. | 13:53 |
| odyssey4me | also, DVR just spreads the same problem to more hosts | 13:53 |
| evrardjp | logan-: you tested calico, how does calico scales? | 13:53 |
| *** schwicht has quit IRC | 13:53 | |
| admin0 | i learnt so much in this last 30 mins then 30 days combined | 13:53 |
| matias | evrardjp: we don't really know which research groups in the university will use the cloud, but there is the chance someone requires to run MPI workloads with low latency | 13:53 |
| matias | evrardjp: do you think calico would be a better option in this case? | 13:53 |
| odyssey4me | so your impact when upgrading is reduced, because you're servicing less things per host - but now you have to manage more on every host and a more complex setup too | 13:53 |
| evrardjp | I can't really say matias | 13:54 |
| evrardjp | I've never tested it | 13:54 |
| *** gouthamr_ has quit IRC | 13:54 | |
| odyssey4me | Andrew_jedi yeah, that happens when the L3 agent reloads because iptables has to recompile all the rules - I hear it's better in mitaka, but that's still a risk that has to be taken | 13:54 |
| logan- | evrardjp: I haven't run it at massive scale yet, but with 100 computes it was fine. they found some scaling issues w/ another customer or in testing last year because they were placing neutron-dhcp-agent on every compute, each of which received network-wide DHCP updates on rabbit, so they wrote a replacement DHCP agent that uses etcd and replaces neutron-dhcp | 13:54 |
| logan- | we have not had any scaling issues with it so far | 13:55 |
| logan- | I think the calico-dhcp-agent stuff fixes the most obvious limitation their architecture had | 13:55 |
| *** gouthamr has joined #openstack-ansible | 13:55 | |
| odyssey4me | matias I'm surprised you've not mentioned SR-IOV considering your use-case? | 13:56 |
| odyssey4me | we've had queries about that from oneswig | 13:56 |
| logan- | a big thing with public cloud (esp one that specializes to some extent in ddos mitigation) is we felt like we'd get killed if we had centralized network nodes. so decentralized data plane was a project req from the start and DVR didn't seem mature enough | 13:56 |
| evrardjp | it obviously has its limits, for a SP it's interesting | 13:56 |
| evrardjp | logan-: makes sense | 13:57 |
| *** hj-hpe has joined #openstack-ansible | 13:58 | |
| admin0 | good point logan .. the DDoS and centralized network nodes stuff | 13:59 |
| *** kstev has joined #openstack-ansible | 14:00 | |
| *** kstev1 has joined #openstack-ansible | 14:01 | |
| matias | odyssey4me: people are trying to get a "mixed" cloud up (scientific + admin/services in the same cloud), so I think we will just recommend ironic for the guys who need to run heavily optimized scientific workload | 14:01 |
| *** strobelight has joined #openstack-ansible | 14:01 | |
| openstackgerrit | Andy McCrae proposed openstack/openstack-ansible-tests: Fix the central repository to work with CentOS https://review.openstack.org/364858 | 14:02 |
| matias | odyssey4me: i don't really think most researchers here have their code sufficiently optimized for this to make a difference, and those who have, we can offload to ironic | 14:02 |
| *** kstev has quit IRC | 14:04 | |
| Andrew_jedi | odyssey4me cloudnull : https://plus.google.com/+FlorianHaas/posts/3crxh7hfuRz | 14:05 |
| admin0 | Andrew_jedi: can you tell more of “flapping floating ip” and exactly why it happens ? | 14:07 |
| admin0 | aha | 14:07 |
| admin0 | its above ^^ s | 14:08 |
| odyssey4me | Andrew_jedi :) | 14:08 |
| admin0 | sorry | 14:08 |
| *** neilus has quit IRC | 14:08 | |
| Andrew_jedi | admin0: Yep i was going to paste the answer from odyssey4me | 14:08 |
| Andrew_jedi | odyssey4me cloudnull : Also this, https://www.hastexo.com/resources/news-releases/hx201-ansible/ | 14:08 |
| *** jamesdenton has joined #openstack-ansible | 14:09 | |
| odyssey4me | admin0 if you're using a real L3 switch, it typically has online upgrade capabilities in the OS so you don't have those issues | 14:09 |
| cloudnull | Andrew_jedi: +1 thats awesome! | 14:09 |
| admin0 | true .. | 14:09 |
| admin0 | brb | 14:09 |
| *** Mudpuppy has joined #openstack-ansible | 14:11 | |
| cloudnull | https://twitter.com/cloudnull/status/771712173491916801 -- https://twitter.com/cloudnull/status/771711555582906368 | 14:13 |
| cloudnull | Andrew_jedi: ^ | 14:13 |
| cloudnull | thanks for the heads up :) | 14:13 |
| logan- | wow! | 14:13 |
| * cloudnull goes off to make friends and influence people | 14:13 | |
| *** deadnull_ has joined #openstack-ansible | 14:14 | |
| Andrew_jedi | cloudnull: :) | 14:15 |
| *** javeriak has joined #openstack-ansible | 14:16 | |
| *** deadnull has quit IRC | 14:18 | |
| *** kenhui has quit IRC | 14:18 | |
| *** schwicht has joined #openstack-ansible | 14:18 | |
| *** spedione|AWAY is now known as spedione | 14:19 | |
| *** kenhui has joined #openstack-ansible | 14:21 | |
| *** janki has quit IRC | 14:23 | |
| *** pfb29 has joined #openstack-ansible | 14:23 | |
| *** spotz_zzz is now known as spotz | 14:24 | |
| odyssey4me | arbrandes :) great to see you were involved in putting the OSA training together | 14:24 |
| *** asettle has joined #openstack-ansible | 14:25 | |
| *** michaelgugino has joined #openstack-ansible | 14:25 | |
| *** hblah has joined #openstack-ansible | 14:29 | |
| *** kenhui has quit IRC | 14:30 | |
| openstackgerrit | Andy McCrae proposed openstack/openstack-ansible-tests: Fix the central repository to work with CentOS https://review.openstack.org/364858 | 14:33 |
| *** wadeholler has joined #openstack-ansible | 14:33 | |
| *** spedione is now known as spedione|AWAY | 14:36 | |
| *** javeriak_ has joined #openstack-ansible | 14:36 | |
| wadeholler | hi all. I followed this to setup rally/tempest https://www.mirantis.com/blog/rally-openstack-tempest-testing-made-simpler/ | 14:36 |
| *** javeriak has quit IRC | 14:37 | |
| wadeholler | could someone who isn't a newb at tempest tell me how to disable the identity v2 tests ( as I think v2 is disabled in openstack-ansible | 14:37 |
| wadeholler | sorry - I meant isn't a newb like me :) | 14:37 |
| *** luzC has joined #openstack-ansible | 14:38 | |
| odyssey4me | wadeholler we still test v2 | 14:38 |
| odyssey4me | we just don't expose it in the service catalog by default | 14:38 |
| *** spedione|AWAY is now known as spedione | 14:39 | |
| odyssey4me | wadeholler have you looked through http://docs.openstack.org/developer/tempest/configuration.html#tempest-configuration though? | 14:39 |
| wadeholler | oh..... great. I think I just have my uri (non v2) configured wrong in tempest.conf | 14:39 |
| odyssey4me | wadeholler also, I think devstack has a keystone v3 only gate check - maybe hunting out the config for that would help | 14:39 |
| cloudnull | we also have https://github.com/openstack/openstack-ansible-os_rally | 14:40 |
| odyssey4me | very specifically though, http://docs.openstack.org/developer/tempest/configuration.html#keystone-connection-info | 14:40 |
| wadeholler | @odyssey4me - thank you! I missed a good portion of that in my research. thanks again! | 14:41 |
| *** Andrew_jedi has quit IRC | 14:42 | |
| spotz | cloudnull was it aio_interfaces.cfg in openstack-ansible or a different repo we fixed last night? | 14:47 |
| *** Andrew_jedi has joined #openstack-ansible | 14:48 | |
| cloudnull | idk if thats needed as the fix | 14:48 |
| cloudnull | seems like adding the forced checksum 'ing made it go | 14:48 |
| spotz | ok I'll write it vague:) | 14:48 |
| cloudnull | maybe both are needed | 14:48 |
| asettle | Ah feck odyssey4me I accidentally gave myself a workflow +1. I reverted it, hopefully it'll stop it | 14:51 |
| asettle | Okay, looks good | 14:51 |
| evrardjp | well it's a good way to make sure it will merge, but it sounds a little anti-democratic :) | 14:52 |
| openstackgerrit | Jesse Pretorius (odyssey4me) proposed openstack/openstack-ansible-specs: [WIP] Implement rolling upgrades from Mitaka to Newton https://review.openstack.org/365019 | 14:52 |
| asettle | hahahhaha thanks evrardjp | 14:52 |
| odyssey4me | lol, asettle that's magical friday for you | 14:53 |
| odyssey4me | perhaps evrardjp will stop giving you a hard time and add his vote to get it done | 14:53 |
| asettle | Gosh, imagine that odyssey4me :p | 14:53 |
| asettle | What do you say, evrardjp ? | 14:53 |
| evrardjp | link? | 14:54 |
| asettle | https://review.openstack.org/#/c/364831/1 | 14:54 |
| evrardjp | now we're talking | 14:54 |
| asettle | :D | 14:54 |
| asettle | Win, tanks man | 14:54 |
| asettle | thanks* | 14:54 |
| evrardjp | gerrit is not helping 'though | 14:55 |
| andymccr | finally: https://review.openstack.org/#/c/364858/ | 14:55 |
| odyssey4me | it's time to beer o clock | 14:56 |
| admin0 | all good information .. thanks guys | 14:56 |
| admin0 | and have a great weekend | 14:56 |
| *** phalmos has joined #openstack-ansible | 14:57 | |
| openstackgerrit | Andy McCrae proposed openstack/openstack-ansible-os_keystone: Use the central test repository for Keystone https://review.openstack.org/364796 | 14:58 |
| evrardjp | good weekend admin0 | 14:59 |
| evrardjp | gerrit is close to failing | 14:59 |
| openstackgerrit | Merged openstack/openstack-ansible: [DOCS] Rename upgrade documentation to upgrade guide https://review.openstack.org/364831 | 15:01 |
| odyssey4me | andymccr needs a tweak on some of the new added tasks | 15:01 |
| *** Andrew_jedi has quit IRC | 15:01 | |
| odyssey4me | andymccr otherwise it looks great to me | 15:02 |
| andymccr | fixing! | 15:02 |
| asettle | Thanks evrardjp :) | 15:02 |
| andymccr | review.openstack is in strug city atm | 15:02 |
| odyssey4me | yup | 15:02 |
| openstackgerrit | Andy McCrae proposed openstack/openstack-ansible-tests: Fix the central repository to work with CentOS https://review.openstack.org/364858 | 15:02 |
| evrardjp | I'd replace pkg_mgr by release | 15:02 |
| evrardjp | forget what I just said | 15:03 |
| andymccr | id like to keep it consistent - i kind of agree release is better, i had initially done os_ansible_family but it seems that everywhere else we've used pkg_mgr | 15:03 |
| evrardjp | yes, because it's yum | 15:04 |
| odyssey4me | andymccr the pip.conf copy task has (rhel) still | 15:04 |
| odyssey4me | it's a nit, but a quick fix | 15:04 |
| evrardjp | I mean it makes sense to test if you have yum to do a yum thing | 15:04 |
| evrardjp | so good for me | 15:04 |
| odyssey4me | so yummy | 15:04 |
| openstackgerrit | Andy McCrae proposed openstack/openstack-ansible-tests: Fix the central repository to work with CentOS https://review.openstack.org/364858 | 15:04 |
| andymccr | i think it makes more sense to base it on the release personally, maybe at some point i'll go through and fix all the testing instances of that | 15:05 |
| odyssey4me | lgtm, once gerrit responds I'll vote | 15:05 |
| mgariepy | anyone tested the package module instead of apt/yum ? | 15:05 |
| evrardjp | andymccr: I blame you for killing gerrit | 15:05 |
| evrardjp | ages ago | 15:05 |
| evrardjp | mgariepy: ^ | 15:05 |
| evrardjp | it used to work weirdly when you had multiple package managers (yes it exists) | 15:06 |
| mgariepy | ok | 15:06 |
| andymccr | it feels like review.openstack is telling me to go have a beer :P | 15:06 |
| arbrandes | odyssey4me: Yeah, this was one of the least exasperating ways to run OpenStack on OpenStack. All on a single eth0, who'd'a thunk it. :D | 15:08 |
| arbrandes | Beats all other deployment solutions hands down, if you ask me. | 15:08 |
| evrardjp | arbrandes: wait until I have a tool to deploy openstack-ansible on openstack in an easy way for developers :p | 15:09 |
| arbrandes | evrardjp: if you make it open source, lookout. We might use it in our courseware. ;D | 15:09 |
| *** vnogin has quit IRC | 15:09 | |
| evrardjp | yup, gonna be on pip | 15:10 |
| evrardjp | I just need to find time :) | 15:10 |
| *** weezS has joined #openstack-ansible | 15:10 | |
| arbrandes | Awesome. When/if you have an alpha, holler - we could lend a hand. | 15:11 |
| evrardjp | Yup the alpha is gonna be ugly -- automagically is interested by it too | 15:11 |
| evrardjp | but hell it's gonna be better than nothing :p | 15:12 |
| arbrandes | Well, it's a good way to gauge interest - and if it turns out to solve enough people's problems... That's how you build a community. ;) | 15:13 |
| *** javeriak_ has quit IRC | 15:13 | |
| *** pcaruana has quit IRC | 15:15 | |
| evrardjp | :) | 15:15 |
| *** sdake has joined #openstack-ansible | 15:16 | |
| openstackgerrit | Marc Gariépy proposed openstack/openstack-ansible-openstack_hosts: Only configure systat on debian based system. https://review.openstack.org/364903 | 15:16 |
| mgariepy | git review slow again :( | 15:17 |
| evrardjp | yup | 15:17 |
| *** kenhui has joined #openstack-ansible | 15:17 | |
| mgariepy | i don't push enough review, i only had time to put one when it was fast :( | 15:17 |
| evrardjp | mgariepy: you may want to look at gertty | 15:18 |
| michaelgugino | ugh, wtf. I was trying to review a patch, and now it's merged. | 15:19 |
| -openstackstatus- NOTICE: The Gerrit service on review.openstack.org is restarting quickly to relieve resource pressure and restore normal performance | 15:20 | |
| *** Andrew_jedi has joined #openstack-ansible | 15:20 | |
| mgariepy | evrardjp, nice thanks for this. i'll try this out. | 15:23 |
| evrardjp | I don't use that myself on a regular basis but it's good to know | 15:24 |
| *** phalmos has quit IRC | 15:28 | |
| *** hblah has quit IRC | 15:28 | |
| *** cloader89 is now known as cjloader | 15:29 | |
| odyssey4me | michaelgugino it looks like the nova lxd tests are reliably successful, so I think I'll promote them to voting | 15:29 |
| *** phalmos has joined #openstack-ansible | 15:29 | |
| michaelgugino | odyssey4me: sounds good. I'm hoping to get some docs typed up soon. Looks like the nova role has transitioned to the new doc style? | 15:30 |
| odyssey4me | andymccr the addition of the multi-region swift tests is merging now (non voting) | 15:30 |
| andymccr | sweet! should be able to test it. | 15:31 |
| odyssey4me | michaelgugino yep, all service specific docs are now in the repo | 15:31 |
| odyssey4me | the role repositories | 15:31 |
| michaelgugino | great. I've been waiting on that to happen. I know it was a large effort, so thanks to whomever tackled that. | 15:31 |
| *** admin0 has quit IRC | 15:33 | |
| openstackgerrit | Merged openstack/openstack-ansible-tests: Fix the central repository to work with CentOS https://review.openstack.org/364858 | 15:33 |
| *** mrjk has joined #openstack-ansible | 15:33 | |
| *** janki has joined #openstack-ansible | 15:33 | |
| *** klamath has quit IRC | 15:37 | |
| *** weezS has quit IRC | 15:40 | |
| odyssey4me | also switching the ovs tests for os_neutron to voting :) | 15:43 |
| *** v1k0d3n has joined #openstack-ansible | 15:48 | |
| *** woodard has joined #openstack-ansible | 15:49 | |
| *** klamath has joined #openstack-ansible | 15:51 | |
| odyssey4me | adreznec it looks we we're getting a neutro db migration failure - that's why the os_nova patch is failing | 15:51 |
| odyssey4me | it's probably not related | 15:51 |
| odyssey4me | so if someone has the time to figure out what's going on there it'd be appreciated | 15:51 |
| adreznec | odyssey4me: Yeah, was just looking through the logs there now | 15:52 |
| odyssey4me | 'cos it's time to weekend | 15:52 |
| odyssey4me | (for me) | 15:52 |
| adreznec | I might have time in the afternoon here | 15:52 |
| * adreznec wishes it was time to weekend here | 15:52 | |
| *** klamath has quit IRC | 15:53 | |
| *** Andrew_jedi has quit IRC | 15:54 | |
| *** karimb has joined #openstack-ansible | 15:55 | |
| openstackgerrit | Bob Taylor proposed openstack/openstack-ansible: Add a doc example for yaml file overrides https://review.openstack.org/364502 | 15:55 |
| openstackgerrit | Merged openstack/openstack-ansible-ceph_client: Fix incorrect project reference https://review.openstack.org/364641 | 15:57 |
| *** phalmos has quit IRC | 15:58 | |
| openstackgerrit | Andy McCrae proposed openstack/openstack-ansible-tests: Remove duplicate memcached_server entry https://review.openstack.org/365043 | 16:00 |
| *** armaan has joined #openstack-ansible | 16:17 | |
| *** Andrew_jedi has joined #openstack-ansible | 16:20 | |
| *** asettle has quit IRC | 16:21 | |
| *** weezS has joined #openstack-ansible | 16:21 | |
| *** Andrew_jedi has quit IRC | 16:22 | |
| *** alwaysatthenoc has joined #openstack-ansible | 16:34 | |
| alwaysatthenoc | Im trying to figure out where I can add the openstack ironic service to AIO config. I don't see anywhere to pass the argument. | 16:35 |
| *** michaelgugino has quit IRC | 16:36 | |
| *** shanec has joined #openstack-ansible | 16:38 | |
| *** shanec_ has joined #openstack-ansible | 16:46 | |
| *** armaan has quit IRC | 16:48 | |
| *** shanec has quit IRC | 16:49 | |
| *** galstrom is now known as galstrom_zzz | 16:49 | |
| *** galstrom_zzz is now known as galstrom | 16:51 | |
| *** Oku_OS is now known as Oku_OS-away | 16:51 | |
| openstackgerrit | Qin Wang (qwang) proposed openstack/openstack-ansible: [docs] added a necessary arg for ansible command after removal of ansible.cfg https://review.openstack.org/363965 | 17:03 |
| *** javeriak has joined #openstack-ansible | 17:03 | |
| openstackgerrit | Merged openstack/openstack-ansible-os_cinder: [DOCS] Clarify is_metal is required if using iSCSI https://review.openstack.org/363310 | 17:08 |
| *** bryan_att has left #openstack-ansible | 17:11 | |
| openstackgerrit | Qin Wang (qwang) proposed openstack/openstack-ansible: [docs] added a necessary arg for ansible command after removal of ansible.cfg https://review.openstack.org/363965 | 17:14 |
| *** BjoernT is now known as Bjoern_zZzZzZzZ | 17:16 | |
| *** kenhui has quit IRC | 17:17 | |
| *** Mudpuppy has quit IRC | 17:19 | |
| *** klamath has joined #openstack-ansible | 17:23 | |
| openstackgerrit | Merged openstack/openstack-ansible-tests: Remove duplicate memcached_server entry https://review.openstack.org/365043 | 17:23 |
| *** markvoelker has joined #openstack-ansible | 17:29 | |
| *** phalmos has joined #openstack-ansible | 17:32 | |
| *** markvoelker has quit IRC | 17:34 | |
| *** Bjoern_zZzZzZzZ is now known as BjoernT | 17:34 | |
| *** markvoelker has joined #openstack-ansible | 17:34 | |
| -openstackstatus- NOTICE: The infrastructure team is taking Gerrit offline for maintenance, beginning shortly after 18:00 UTC for a potentially 4 hour maintenance window. | 17:37 | |
| *** waht has joined #openstack-ansible | 17:39 | |
| *** deadnull_ has quit IRC | 17:40 | |
| *** phalmos has quit IRC | 17:42 | |
| *** eil397 has joined #openstack-ansible | 17:43 | |
| *** shanec_ has quit IRC | 17:44 | |
| *** prometheanfire has quit IRC | 17:45 | |
| openstackgerrit | Nish Patwa(IRC: nishpatwa) proposed openstack/openstack-ansible: Added Xenial support in install guide https://review.openstack.org/364572 | 17:46 |
| *** phalmos has joined #openstack-ansible | 17:47 | |
| *** kjw3 has joined #openstack-ansible | 17:47 | |
| *** adrian_otto has joined #openstack-ansible | 17:49 | |
| *** michaelgugino has joined #openstack-ansible | 17:51 | |
| *** waht has quit IRC | 17:54 | |
| *** AnarchyAo has joined #openstack-ansible | 17:58 | |
| *** poopcat has joined #openstack-ansible | 18:01 | |
| -openstackstatus- NOTICE: Gerrit is now going offline for maintenance, reserving a maintenance window through 22:00 UTC. | 18:02 | |
| *** ChanServ changes topic to "Gerrit is now going offline for maintenance, reserving a maintenance window through 22:00 UTC." | 18:02 | |
| openstackgerrit | Merged openstack/openstack-ansible: run-upgrade.sh is considered as tested now https://review.openstack.org/364448 | 18:04 |
| *** schwicht has quit IRC | 18:07 | |
| *** kenhui has joined #openstack-ansible | 18:09 | |
| *** karimb has quit IRC | 18:12 | |
| *** shanec_ has joined #openstack-ansible | 18:17 | |
| *** schwicht has joined #openstack-ansible | 18:24 | |
| *** kenhui has quit IRC | 18:26 | |
| *** phalmos has quit IRC | 18:26 | |
| *** janki has quit IRC | 18:29 | |
| *** phalmos has joined #openstack-ansible | 18:30 | |
| stevelle | mhayden: I was trying to ask why the role default was changing rather than only changing the default in the integrated gate with a group_var | 18:32 |
| *** shanec_ has quit IRC | 18:40 | |
| *** prometheanfire has joined #openstack-ansible | 18:40 | |
| *** shanec_ has joined #openstack-ansible | 18:46 | |
| *** phalmos has quit IRC | 18:51 | |
| openstackgerrit | Michael Gugino proposed openstack/openstack-ansible-os_nova: Add config template for qemu.conf https://review.openstack.org/365095 | 18:57 |
| *** sdake has quit IRC | 19:02 | |
| *** admin0 has joined #openstack-ansible | 19:03 | |
| openstackgerrit | Michael Gugino proposed openstack/openstack-ansible-os_nova: Update docs for mixed lxd hypervisors https://review.openstack.org/365098 | 19:07 |
| *** Jeffrey4l has quit IRC | 19:07 | |
| alextricity25 | cloudnull: Hey man. Are you around? | 19:10 |
| cloudnull | i am | 19:10 |
| cloudnull | whats up? | 19:10 |
| alextricity25 | I'm looking at https://github.com/openstack/openstack-ansible/blob/stable/mitaka/scripts/upgrade-utilities/scripts/migrate_openstack_vars.py From what I can tell..it doesn't do anything. Is that a correct assumption for L->M upgrades? | 19:11 |
| *** ChanServ changes topic to "Launchpad: https://launchpad.net/openstack-ansible || Weekly Meetings: https://wiki.openstack.org/wiki/Meetings/openstack-ansible || Review Dashboard: https://goo.gl/tTmdgs" | 19:11 | |
| -openstackstatus- NOTICE: Gerrit is completing an online re-index, you may encounter slowness until it is complete | 19:11 | |
| alextricity25 | cloudnull...well..maybe "doesn't do anything" Isn't the right wordage there. I see that it writes out to the openstack_deploy.MITAKA user_variable files, but it doesn't actually migrate any variables there, does it? | 19:12 |
| alextricity25 | Unless I'm missing something.. | 19:12 |
| cloudnull | alextricity25: its used here https://github.com/openstack/openstack-ansible/blob/stable/mitaka/scripts/upgrade-utilities/playbooks/deploy-config-changes.yml | 19:14 |
| alextricity25 | cloudnull: Yes, but since VAR_MAPPINGS is practically empty, it's not really doing anything there is it? I just want to make sure I'm not missing anything :) | 19:14 |
| cloudnull | That was carried forward from kilo>liberty. if its no longer needed we should remove it | 19:15 |
| alextricity25 | cloudnull: ohhh I just looked at the liberty one. Makes more sense now. | 19:15 |
| alextricity25 | hehe | 19:15 |
| *** klamath has quit IRC | 19:16 | |
| admin0 | what happened to the files in env.d ? how do I disable is_metal = true from cinder in the new tag ? | 19:20 |
| admin0 | 13.3.2 | 19:20 |
| *** Andrew_jedi has joined #openstack-ansible | 19:21 | |
| mattt | admin0: not sure i understand the question | 19:22 |
| logan- | it is an override system now so you can drop a file in /etc/openstack_ansible/env.d like http://cdn.pasteraw.com/6qms64rg1lehmtfhwt8as6390x0n31c | 19:22 |
| mattt | admin0: https://github.com/openstack/openstack-ansible/blob/13.3.2/etc/openstack_deploy/env.d/cinder.yml#L62 | 19:22 |
| stevelle | admin0: we moved the default files in-tree to playbooks/inventory/ and you only need a file under /etc/openstack-deploy/ if you need to override | 19:22 |
| stevelle | just as FYI for 14.x | 19:23 |
| admin0 | got it | 19:23 |
| mattt | nice! didn't know that | 19:23 |
| *** karimb has joined #openstack-ansible | 19:24 | |
| logan- | disregard what i said for 13.x, that is for master/newton only | 19:24 |
| *** Andrew_jedi has quit IRC | 19:24 | |
| logan- | the full files should still be in place on 13.x, no change from the old system until newton :) | 19:24 |
| admin0 | i did a git glone , and then switched to the tag | 19:25 |
| admin0 | one more questions. if its for a internal demo setup, can master branch also be used to setup ? | 19:27 |
| admin0 | does it work ? | 19:27 |
| admin0 | till now, i have been doing via tags | 19:27 |
| stevelle | admin0: the 14.0.0.b3 tag is known to be slightly broken, should be fixed since then but it's a gamble. | 19:28 |
| stevelle | we're stabilizing and upstream doesn't have tons of feature-freeze exceptions open so it shouldn't be too bad but no guarantee | 19:29 |
| *** schwicht has quit IRC | 19:30 | |
| admin0 | hmm.. will stick to 13.3.2 :D | 19:30 |
| *** gouthamr has quit IRC | 19:33 | |
| *** dkehn_ has quit IRC | 19:34 | |
| *** schwicht has joined #openstack-ansible | 19:34 | |
| *** thorst has quit IRC | 19:36 | |
| *** thorst has joined #openstack-ansible | 19:37 | |
| *** markvoelker has quit IRC | 19:37 | |
| *** thorst_ has joined #openstack-ansible | 19:39 | |
| *** eil397 has left #openstack-ansible | 19:41 | |
| *** thorst has quit IRC | 19:41 | |
| *** thorst_ has quit IRC | 19:43 | |
| *** dkehn_ has joined #openstack-ansible | 19:46 | |
| *** karimb has quit IRC | 19:48 | |
| *** schwicht has quit IRC | 19:50 | |
| *** schwicht has joined #openstack-ansible | 19:51 | |
| *** karimb has joined #openstack-ansible | 19:52 | |
| *** klamath has joined #openstack-ansible | 19:55 | |
| *** klamath has quit IRC | 19:57 | |
| openstackgerrit | Paulo Matias proposed openstack/openstack-ansible-lxc_container_create: Check if lxc_container_backing_method is defined https://review.openstack.org/365119 | 20:04 |
| *** openstackgerrit has quit IRC | 20:04 | |
| matias | odyssey4me: ^ sorry for the delay | 20:04 |
| *** openstackgerrit has joined #openstack-ansible | 20:05 | |
| *** gouthamr has joined #openstack-ansible | 20:07 | |
| *** spotz is now known as spotz_zzz | 20:07 | |
| *** schwicht has quit IRC | 20:14 | |
| *** spedione is now known as spedione|AWAY | 20:20 | |
| *** BjoernT has quit IRC | 20:20 | |
| *** jamesdenton has quit IRC | 20:21 | |
| *** gouthamr has quit IRC | 20:34 | |
| michaelgugino | fyi, looks like gerrit didn't send a patch to zuul, so if you're waiting on something, you might want to issue a recheck | 20:40 |
| *** askb has joined #openstack-ansible | 20:41 | |
| *** galstrom is now known as galstrom_zzz | 20:42 | |
| *** karimb has quit IRC | 20:45 | |
| *** weezS has quit IRC | 20:45 | |
| *** thorst has joined #openstack-ansible | 20:46 | |
| *** markvoelker has joined #openstack-ansible | 20:47 | |
| *** schwicht has joined #openstack-ansible | 20:48 | |
| *** hj-hpe has quit IRC | 20:48 | |
| *** karimb has joined #openstack-ansible | 20:50 | |
| cloudnull | :\ | 20:51 |
| cloudnull | michaelgugino: good lookin out | 20:51 |
| *** AnarchyAo has quit IRC | 20:53 | |
| *** scarlisle has joined #openstack-ansible | 20:54 | |
| stevelle | cloudnull: looks like we won't be able to fix up ceilo middleware through conf alone, and even then it's going to be tricky so we have to leave that on the std port for newton | 21:02 |
| cloudnull | :'( | 21:02 |
| michaelgugino | cloudnull: odyssey4me looks like neutron has a critical bug out: https://bugs.launchpad.net/neutron/+bug/1619696 | 21:02 |
| openstack | Launchpad bug 1619696 in neutron ""neutron-db-manage upgrade heads" fails with networksegments_ibfk_2" [Critical,In progress] - Assigned to Jakub Libosvar (libosvar) | 21:02 |
| stevelle | this doesn't seem to warrant a feature freeze exception | 21:02 |
| cloudnull | bummer | 21:02 |
| michaelgugino | we need to bump master as soon as that merges | 21:02 |
| cloudnull | odyssey4me: ^ still around? | 21:03 |
| cloudnull | michaelgugino: i can do the needful | 21:03 |
| *** kjw3 has quit IRC | 21:03 | |
| michaelgugino | cool, it's not merged yet, so no rush. Probably want to look at it tomorrow or Monday | 21:03 |
| cloudnull | ++ | 21:04 |
| michaelgugino | I'm out. Have a good one everybody | 21:09 |
| cloudnull | have a good one michaelgugino | 21:09 |
| *** michaelgugino has quit IRC | 21:14 | |
| *** alwaysatthenoc has quit IRC | 21:17 | |
| *** sdake has joined #openstack-ansible | 21:19 | |
| *** woodard has quit IRC | 21:24 | |
| *** spotz_zzz is now known as spotz | 21:32 | |
| *** intr1nsic has quit IRC | 21:35 | |
| *** intr1nsic has joined #openstack-ansible | 21:37 | |
| *** schwicht has quit IRC | 21:37 | |
| *** admin0 has quit IRC | 21:43 | |
| *** schwicht has joined #openstack-ansible | 21:43 | |
| *** kstev has joined #openstack-ansible | 21:43 | |
| *** kstev1 has quit IRC | 21:44 | |
| *** scarlisle has quit IRC | 21:45 | |
| *** kstev has quit IRC | 21:47 | |
| *** karimb has quit IRC | 21:51 | |
| *** fedruantine has quit IRC | 21:58 | |
| *** jperry has quit IRC | 21:59 | |
| *** fedruantine has joined #openstack-ansible | 22:15 | |
| *** david-lyle has quit IRC | 22:15 | |
| *** david-lyle has joined #openstack-ansible | 22:17 | |
| *** javeriak has quit IRC | 22:22 | |
| *** admin0 has joined #openstack-ansible | 22:24 | |
| *** markvoelker has quit IRC | 22:47 | |
| *** kenhui has joined #openstack-ansible | 22:49 | |
| *** asettle has joined #openstack-ansible | 23:05 | |
| *** asettle has quit IRC | 23:10 | |
| *** admin0 has quit IRC | 23:17 | |
| *** openstackgerrit has quit IRC | 23:18 | |
| *** openstackgerrit has joined #openstack-ansible | 23:19 | |
| *** thotypous has quit IRC | 23:21 | |
| *** shanec_ has quit IRC | 23:23 | |
| *** thorst has quit IRC | 23:30 | |
| *** thorst has joined #openstack-ansible | 23:32 | |
| *** jamesdenton has joined #openstack-ansible | 23:37 | |
| *** asettle has joined #openstack-ansible | 23:46 | |
| *** Apsu has joined #openstack-ansible | 23:53 | |
Generated by irclog2html.py 2.14.0 by Marius Gedminas - find it at mg.pov.lt!
