| *** thorst has joined #openstack-ansible | 00:02 | |
| *** thorst has quit IRC | 00:08 | |
| *** woodard has joined #openstack-ansible | 00:09 | |
| *** thorst has joined #openstack-ansible | 00:36 | |
| *** thorst has quit IRC | 00:36 | |
| *** charcol has quit IRC | 00:39 | |
| *** dxiri has quit IRC | 00:41 | |
| *** dxiri has joined #openstack-ansible | 00:41 | |
| *** dxiri has quit IRC | 00:41 | |
| *** dxiri has joined #openstack-ansible | 00:42 | |
| *** yifei has joined #openstack-ansible | 00:46 | |
| *** weezS has joined #openstack-ansible | 00:48 | |
| *** dixiaoli has joined #openstack-ansible | 01:04 | |
| *** gouthamr has quit IRC | 01:13 | |
| *** weezS has quit IRC | 01:16 | |
| *** galstrom_zzz is now known as galstrom | 01:28 | |
| *** fergus has joined #openstack-ansible | 01:42 | |
| *** fergus has quit IRC | 01:44 | |
| *** thorst has joined #openstack-ansible | 01:52 | |
| *** weezS has joined #openstack-ansible | 01:52 | |
| *** charcol has joined #openstack-ansible | 01:54 | |
| *** galstrom is now known as galstrom_zzz | 01:54 | |
| *** thorst has quit IRC | 01:56 | |
| *** schwicht has quit IRC | 02:00 | |
| *** aludwar has quit IRC | 02:02 | |
| *** lbragstad has joined #openstack-ansible | 02:10 | |
| *** weezS has quit IRC | 02:17 | |
| *** weezS has joined #openstack-ansible | 02:30 | |
| *** woodard has quit IRC | 02:43 | |
| *** woodard has joined #openstack-ansible | 02:44 | |
| *** schwicht has joined #openstack-ansible | 02:46 | |
| *** woodard has quit IRC | 02:48 | |
| *** weezS has quit IRC | 02:52 | |
| *** schwicht has quit IRC | 03:34 | |
| *** udesale has joined #openstack-ansible | 03:36 | |
| *** Dinesh_Bhor has joined #openstack-ansible | 03:52 | |
| *** thorst has joined #openstack-ansible | 03:53 | |
| *** thorst has quit IRC | 03:57 | |
| *** woodard has joined #openstack-ansible | 04:00 | |
| *** lbragstad has quit IRC | 04:04 | |
| *** fergus has joined #openstack-ansible | 04:14 | |
| *** fergus has quit IRC | 04:18 | |
| *** hybridpollo has quit IRC | 04:23 | |
| *** fergus has joined #openstack-ansible | 04:37 | |
| openstackgerrit | iswarya vakati proposed openstack/openstack-ansible master: Optimize the link address https://review.openstack.org/481570 | 05:28 |
|---|---|---|
| *** tobberydberg has joined #openstack-ansible | 05:29 | |
| *** jbadiapa has joined #openstack-ansible | 05:43 | |
| *** fergus has quit IRC | 05:50 | |
| *** pvradu has joined #openstack-ansible | 05:50 | |
| *** thorst has joined #openstack-ansible | 05:54 | |
| *** fergus has joined #openstack-ansible | 05:55 | |
| *** fergus has quit IRC | 05:55 | |
| *** thorst has quit IRC | 05:59 | |
| *** dfflanders has joined #openstack-ansible | 06:09 | |
| *** Oku_OS-away is now known as Oku_OS | 06:10 | |
| *** dixiaoli has quit IRC | 06:18 | |
| *** dixiaoli has joined #openstack-ansible | 06:18 | |
| *** cshen has joined #openstack-ansible | 06:27 | |
| *** dfflanders has quit IRC | 06:29 | |
| cshen | w | 06:39 |
| cloudnull | 01:41:53 up 3 days, 17:31, 2 users, load average: 0.20, 0.55, 0.73 | 06:42 |
| cloudnull | :) | 06:42 |
| *** charcol has quit IRC | 07:04 | |
| *** hw_wutianwei has joined #openstack-ansible | 07:09 | |
| *** shardy has joined #openstack-ansible | 07:14 | |
| *** jvidal has joined #openstack-ansible | 07:16 | |
| *** pbandark has joined #openstack-ansible | 07:38 | |
| *** dixiaoli has quit IRC | 07:45 | |
| *** dixiaoli has joined #openstack-ansible | 07:46 | |
| *** mbuil has joined #openstack-ansible | 07:50 | |
| *** jvidal has quit IRC | 07:51 | |
| *** thorst has joined #openstack-ansible | 07:54 | |
| *** thorst has quit IRC | 08:00 | |
| *** jvidal has joined #openstack-ansible | 08:04 | |
| *** shardy has quit IRC | 08:06 | |
| *** shardy has joined #openstack-ansible | 08:07 | |
| openstackgerrit | Jesse Pretorius (odyssey4me) proposed openstack/openstack-ansible master: Tidy up keystone need_db_sync fact https://review.openstack.org/482043 | 08:10 |
| Tahvok | mhayden: so I don't really need it? Or what? | 08:23 |
| openstackgerrit | Merged openstack/openstack-ansible-ops master: Support older versions for getting service ports https://review.openstack.org/481737 | 08:25 |
| *** tots has joined #openstack-ansible | 08:28 | |
| tots | hello | 08:29 |
| tots | reading the doc here https://docs.openstack.org/project-deploy-guide/openstack-ansible/newton/app-config-prod.html i'd like to ask the puprose of having an image_hosts group. | 08:30 |
| tots | i do not have an image hosts in my config and i am experiencing a problem where i can not create more than 2 instances at once , one will fail and so far i create them one by one. is it because i miss the image_hosts ? | 08:31 |
| andymccr | tots - if you have "os-infra_hosts:" specified you shouldn't need the image_hosts: - that will take care of your glance containers/hosts | 08:35 |
| tots | i have that andymccr | 08:36 |
| andymccr | tots: then i think the image_hosts thing may not be the problem :) | 08:36 |
| tots | what else can i check? | 08:36 |
| openstackgerrit | Jesse Pretorius (odyssey4me) proposed openstack/openstack-ansible master: py35: Set virtualenv options for CentOS https://review.openstack.org/482052 | 08:37 |
| tots | creating 10 instances at once , 5 will fail 5 build | 08:37 |
| andymccr | if you can get 2 instances up but then it fails, i'd check your nova logs and see why they are failing to build | 08:37 |
| tots | ok i will show logs in a while andymccr | 08:38 |
| pvradu | hi all! trying to setup an openstack cluster using openstack-ansible, but failing neutron config - trying to do openvswitch + DVR | 08:41 |
| pvradu | tenant network should be vxlan, provider should be flat | 08:42 |
| pvradu | openvswitch agents on both infra and compute nodes are terminated because they don't have anything than br-int created | 08:42 |
| odyssey4me | Tahvok eth12 is explained in the docs: https://github.com/openstack/openstack-ansible/blob/stable/ocata/etc/network/interfaces.d/openstack_interface.cfg.prod.example#L117-L133 | 08:43 |
| odyssey4me | tots that's more than likely because your image is only available on one of your two control plane hosts | 08:44 |
| odyssey4me | glance needs a shared storage back-end, and it would appear to me that you do not have one | 08:44 |
| pvradu | thanks, I'll have a look | 08:44 |
| odyssey4me | so either you need to copy the glance store over to the other glance container, or you need to setup a common store | 08:45 |
| odyssey4me | pvradu see https://docs.openstack.org/project-deploy-guide/openstack-ansible/ocata/overview-storage-arch.html#image-storage-glance | 08:45 |
| odyssey4me | oops, sorry - that was meant for tots | 08:45 |
| Tahvok | odyssey4me: which I understand is for only selfservice networks? But according to the doc, I only need to set up the veth pair if neutron is not conainerized? So, I don't need to do anything, but only provide the second br-vlan network configuration, as in the same example page? | 08:45 |
| pvradu | @odyssey4me I thought so :) | 08:46 |
| odyssey4me | Tahvok neutron is never containerised on the compute hosts | 08:47 |
| Tahvok | I though only nova isn't | 08:47 |
| odyssey4me | Tahvok nothing is containerised on the compute hosts | 08:47 |
| odyssey4me | neutron-agent and nova-compute need direct access to the host | 08:48 |
| Tahvok | That explains it... But am I correct that I need it for selfservice networks? | 08:48 |
| odyssey4me | well, that depends entirely on your setup | 08:48 |
| pvradu | odyssey4me, is there anything I should add to user_variables.yml? like http://trumant.github.io/openstack-ansible-dvr-with-openvswitch.html ? | 08:48 |
| odyssey4me | but yes, it's a veth pair which is used for self service networks in the sample configs | 08:49 |
| odyssey4me | pvradu unfortunately I have no idea how to make OVS work | 08:49 |
| odyssey4me | pvradu I don't know if https://docs.openstack.org/openstack-ansible-os_neutron/latest/app-openvswitch.html helps at all? | 08:49 |
| tots | odyssey4me: so create an nfs share like here ? https://docs.openstack.org/project-deploy-guide/openstack-ansible/newton/app-config-prod.html | 08:51 |
| odyssey4me | tots that's config to connect to an nfs share | 08:52 |
| odyssey4me | you'll need to setup the nfs share yourself | 08:52 |
| pvradu | odyssey4me, that's only using OVS without DVR, so no HA. I need something that uses OVS + DVR. And preferably explained in detail.. | 08:52 |
| tots | is there a tutorial what share is needed and on what host ? | 08:53 |
| odyssey4me | tots the share needed is a common store, which can be anywhere | 08:53 |
| odyssey4me | if you read the documentation properly you'll understand better | 08:53 |
| openstackgerrit | Jesse Pretorius (odyssey4me) proposed openstack/openstack-ansible master: py35: Set virtualenv options for CentOS https://review.openstack.org/482052 | 08:54 |
| *** ivveh has quit IRC | 08:56 | |
| *** ivveh has joined #openstack-ansible | 08:56 | |
| *** ivveh has quit IRC | 09:07 | |
| *** tots has quit IRC | 09:11 | |
| hwoarang | good morning | 09:20 |
| *** thorst has joined #openstack-ansible | 09:23 | |
| *** thorst has quit IRC | 09:28 | |
| *** electrofelix has joined #openstack-ansible | 09:41 | |
| *** electrofelix has quit IRC | 09:45 | |
| evrardjp | good morning | 09:50 |
| odyssey4me | I have some final rolling-upgrade role backports which need reviews please: https://review.openstack.org/481916 & https://review.openstack.org/481920 | 09:51 |
| *** DanyC has joined #openstack-ansible | 09:52 | |
| *** DanyC has quit IRC | 09:52 | |
| *** electrofelix has joined #openstack-ansible | 09:55 | |
| *** stuartgr has joined #openstack-ansible | 09:58 | |
| *** admin0 has joined #openstack-ansible | 10:03 | |
| admin0 | is there pros/cons list of of lb vs ovs vs calico for a new platfrom ? | 10:04 |
| admin0 | greenfield env | 10:05 |
| odyssey4me | admin0 no, this is an open source project - not a vendor product | 10:06 |
| odyssey4me | evaluations of one versus another is the sort of thing that vendor products do | 10:06 |
| admin0 | oh .. sorry .. i thought calico was free :D | 10:06 |
| odyssey4me | it is, but you're missing the point | 10:07 |
| odyssey4me | an open source project is not going to do a competitive analysis for you | 10:07 |
| admin0 | aha! .. i asked it wrong .. are there people where who selected either ovs and calico over linunbridge and if they can share why | 10:08 |
| admin0 | *here | 10:08 |
| odyssey4me | you do know that you've asked the very same questions before in this very channel ;) | 10:08 |
| admin0 | as projects mature and more people use it, opinions and experiece change also | 10:09 |
| odyssey4me | sure | 10:09 |
| admin0 | btw, i was able to integrate with qnap .. i will contribute a howto in the cinder section for that this week | 10:09 |
| odyssey4me | I always find myself curious why people want to use OVS, and never seem to get answers which satisfy me. | 10:10 |
| odyssey4me | ah great, thanks | 10:10 |
| admin0 | i am sticking to LB all the time, but sometimes feel if i am missing out :D | 10:11 |
| admin0 | which is why this question pops up from time to time | 10:11 |
| odyssey4me | one reason that does pop up sometimes is to be able to use DVR | 10:11 |
| odyssey4me | personally I think DVR is a tragedy because of the increased attack surface, but I guess for some use-cases the increased scale is worth that | 10:12 |
| admin0 | heard not so good stories about dvr at scale .. how it blogs down rabbit .. maybe the multi-cell in pike might address it | 10:12 |
| *** tots has joined #openstack-ansible | 10:15 | |
| odyssey4me | well, much like using telemetry you can always setup a dedicated cluster for it | 10:15 |
| tots | odyssey4me: can you take a look at this when you can , about what we talked earlier ? http://paste.openstack.org/show/614889/ | 10:15 |
| odyssey4me | tots what about it? | 10:16 |
| tots | the storage backend | 10:16 |
| odyssey4me | tots ah, well, so you're doing the cinder config there - but that doesn't help glance | 10:16 |
| tots | i see | 10:16 |
| odyssey4me | read the architecture documentation again to make sure you understand the 3 different storage areas for consideration | 10:17 |
| tots | then i should add an image_hosts group as i see in the doc | 10:18 |
| odyssey4me | and by the way, setting up storage hosts like that to all speak to one of them gives you absolutely no redundancy whatsoever | 10:18 |
| odyssey4me | cinder-volume is not a highly available service, and you're making all instances talk to the same back-end | 10:18 |
| odyssey4me | so if that back-end dies, all your cinder dies | 10:18 |
| odyssey4me | so I would not bother - just have one storage host | 10:19 |
| odyssey4me | tots you need an image_hosts group, but you also need to configure it to use nfs | 10:19 |
| odyssey4me | yep, as per the example given for the 'production' environment | 10:20 |
| tots | odyssey4me: correct per the example as i read it | 10:20 |
| tots | ok i wond bother with the changes i showed you i will just do an image hosts group , per the docs | 10:22 |
| DimGR | goodmorning osa | 10:28 |
| DimGR | odyssey4me magnum rocks :) | 10:28 |
| odyssey4me | DimGR oh really? | 10:28 |
| odyssey4me | if you like it so much, perhaps you can help us keep it maintained? :) | 10:29 |
| odyssey4me | the role to deploy it I mean | 10:29 |
| DimGR | oh yes it makes deploying kube easier and fast | 10:29 |
| DimGR | yes i can why not , im already involved with fedora atomic hosts on magnum | 10:29 |
| odyssey4me | great! | 10:30 |
| odyssey4me | all that really needs doing is tracking changes upstream in the magnum service/community and ensuring that the roles keeps up to date with them | 10:30 |
| odyssey4me | along with monitoring changes in the other OSA roles which improve things, and applying the patterns/improvements to the magnum role | 10:31 |
| tots | one hopefully last question odyssey4me . The docs show glance_nfs_client : IP . i can run an nfs server on my one and only storage host but what directory should i export ? can this just be an empty directory that i will create like /images ? | 10:39 |
| *** lostRhino has joined #openstack-ansible | 10:41 | |
| *** udesale has quit IRC | 10:44 | |
| openstackgerrit | Jesse Pretorius (odyssey4me) proposed openstack/openstack-ansible master: Set python interpreter for connection:local plays https://review.openstack.org/482090 | 10:44 |
| odyssey4me | tots yes | 10:46 |
| tots | thank you | 10:47 |
| odyssey4me | it must not be co-located with the other services (cinder/nova) otherwise you may end up with a uuid folder name collission at some point | 10:47 |
| tots | that host has been running cinder only and nothing else , it has been the only host under storage_hosts | 10:47 |
| tots | and it s not declared in any other groups | 10:47 |
| DimGR | odyssey4me im down for that , i will ping you in few days about it | 10:59 |
| *** dixiaoli has quit IRC | 11:00 | |
| odyssey4me | tots sure, what I meant is that the folder used for each service must not be shared | 11:01 |
| *** ricardoas1 has quit IRC | 11:08 | |
| *** ricardoas has joined #openstack-ansible | 11:09 | |
| *** vnogin has joined #openstack-ansible | 11:10 | |
| *** shukixxxxx has joined #openstack-ansible | 11:15 | |
| shukixxxxx | I try to deploy the default setup but it fail to install the repo container | 11:16 |
| tots | ok so i run an nfs server on a host and i exported /var/nfs . i created the image_hosts group per the instructions . do i run just the glance role now or setup everything ? | 11:17 |
| shukixxxxx | It is based on 16.04 LTS but it seem that it fail to apt-get and install many packages?? | 11:17 |
| openstackgerrit | Jesse Pretorius (odyssey4me) proposed openstack/openstack-ansible-os_horizon master: Improve inventory scoping for limited play execution https://review.openstack.org/482101 | 11:28 |
| odyssey4me | tots just the glance role | 11:29 |
| odyssey4me | tots you will need to migrate the existing images over to your nfs store though | 11:29 |
| odyssey4me | shukixxxxx so it sounds like you have some sort of networking failure if the repo container can't install packages | 11:30 |
| *** tobberyd_ has joined #openstack-ansible | 11:31 | |
| *** tobberydberg has quit IRC | 11:35 | |
| *** hw_wutianwei has quit IRC | 11:36 | |
| admin0 | tots, you need glance ( to provide images to vm ) and cinder ( block storage ) .. so what you can do is in your nfs server, create /home/glance /home/cinder or /srv/glance /srv/cinder and then export 2 folders over NFS .. and then pass those to the cinder and glance in your file | 11:36 |
| *** tobberyd_ has quit IRC | 11:36 | |
| openstackgerrit | Merged openstack/openstack-ansible-os_swift stable/ocata: Filter services dictionary per host https://review.openstack.org/481912 | 11:36 |
| openstackgerrit | Merged openstack/openstack-ansible-os_swift stable/newton: Filter services dictionary per host https://review.openstack.org/481913 | 11:37 |
| openstackgerrit | Jesse Pretorius (odyssey4me) proposed openstack/openstack-ansible-os_horizon master: Optimise the role execution https://review.openstack.org/482101 | 11:41 |
| openstackgerrit | Merged openstack/openstack-ansible-os_nova stable/newton: Implement data migrations for rolling upgrades https://review.openstack.org/481920 | 11:49 |
| openstackgerrit | Merged openstack/openstack-ansible-os_swift stable/ocata: Optimise the execution of the role https://review.openstack.org/481911 | 11:49 |
| openstackgerrit | Merged openstack/openstack-ansible-os_swift stable/newton: Optimise the execution of the role https://review.openstack.org/481914 | 11:49 |
| *** tots has quit IRC | 11:49 | |
| openstackgerrit | Merged openstack/openstack-ansible-os_nova stable/ocata: Implement data migrations for rolling upgrades https://review.openstack.org/481916 | 11:50 |
| *** aludwar has joined #openstack-ansible | 11:52 | |
| *** thorst has joined #openstack-ansible | 11:53 | |
| openstackgerrit | Jesse Pretorius (odyssey4me) proposed openstack/openstack-ansible stable/newton: Implement rolling upgrades for nova https://review.openstack.org/481925 | 11:56 |
| *** udesale has joined #openstack-ansible | 11:58 | |
| openstackgerrit | Jesse Pretorius (odyssey4me) proposed openstack/openstack-ansible stable/ocata: Implement rolling upgrades for nova https://review.openstack.org/481924 | 11:58 |
| *** smatzek has joined #openstack-ansible | 12:03 | |
| *** schwicht has joined #openstack-ansible | 12:05 | |
| shukixxxxx | odyssey4me, I don't think it is networking issue .. It get the pkg but fail for example | 12:09 |
| shukixxxxx | libssl-dev : Depends: libssl1.0.0 (= 1.0.2g-1ubuntu4.5) but 1.0.2g-1ubuntu4.6 is to be installed" | 12:09 |
| shukixxxxx | And it happen with different pkgs all the way | 12:10 |
| openstackgerrit | Jesse Pretorius (odyssey4me) proposed openstack/openstack-ansible-os_heat master: Optimise role execution https://review.openstack.org/482116 | 12:12 |
| odyssey4me | shukixxxxx if it's not a networking issue, then you have a problem with the mirror you're using | 12:12 |
| odyssey4me | it's possible that the lxc cache image has more up to date packages installed than are available in your mirror | 12:13 |
| *** brad[] has quit IRC | 12:16 | |
| *** brad[] has joined #openstack-ansible | 12:18 | |
| odyssey4me | andymccr how do you feel about having the translations scneario being executed in the periodic? | 12:20 |
| andymccr | odyssey4me: yeah i want to get that setup. | 12:20 |
| odyssey4me | ok, putting a job in now | 12:20 |
| andymccr | haha ok cool thanks | 12:20 |
| shukixxxxx | odyssey4me So if it is the case is there a way to make it work? I tried to deploy openstack with CENTOS 7 and it fail to deploy the galera containers | 12:22 |
| odyssey4me | shukixxxxx that would depend on your mirrors | 12:23 |
| odyssey4me | it has nothing to do with OSA if your mirrors are broken | 12:23 |
| odyssey4me | you can set to use a different mirror on your host before deploying - use one that works | 12:24 |
| *** acormier has joined #openstack-ansible | 12:27 | |
| *** acormier has quit IRC | 12:28 | |
| *** acormier has joined #openstack-ansible | 12:28 | |
| *** aludwar has quit IRC | 12:29 | |
| mhayden | buenos dias | 12:29 |
| *** ricardoas has quit IRC | 12:30 | |
| *** cshen has quit IRC | 12:30 | |
| shukixxxxx | Do you have a mirror setup that works ? Does it mean that I have to install the target host with diffrent mirror again from scratch? | 12:30 |
| openstackgerrit | Jesse Pretorius (odyssey4me) proposed openstack/openstack-ansible-os_horizon master: Optimise the role execution https://review.openstack.org/482101 | 12:32 |
| *** cshen has joined #openstack-ansible | 12:33 | |
| odyssey4me | andymccr https://review.openstack.org/482120 there you go | 12:33 |
| odyssey4me | shukixxxxx yes, the default setup works as long as your host's mirror is correct | 12:34 |
| odyssey4me | the containers inherit from the host | 12:34 |
| odyssey4me | so yes, if you have a broken setup it's cleaner to start from scratcg | 12:34 |
| odyssey4me | I would recommend that your test environment is something you can easily trash and rebuild at will | 12:34 |
| odyssey4me | mgariepy mhayden as a heads-up, I've pushed https://review.openstack.org/482052 to get the bootstrap working right for centos in master again | 12:38 |
| odyssey4me | yesterday's patch broke it | 12:38 |
| * mhayden ganders | 12:38 | |
| *** ricardoas has joined #openstack-ansible | 12:40 | |
| *** schwicht has quit IRC | 12:42 | |
| mhayden | if anyone has a little time, there's a stack of patches waiting in ansible-hardening land -> https://review.openstack.org/#/q/project:openstack/ansible-hardening+status:open | 12:46 |
| *** acormier has quit IRC | 12:47 | |
| openstackgerrit | Merged openstack/openstack-ansible-memcached_server master: Updated from OpenStack Ansible Tests https://review.openstack.org/476557 | 12:48 |
| andymccr | ahh odyssey4me boom - i was gonna look at that in a bit. nice one | 12:51 |
| openstackgerrit | Markos Chandras (hwoarang) proposed openstack/openstack-ansible-tests master: test-prepare-host: Fix POST_{UP,DOWN}_SCRIPT handling for SUSE https://review.openstack.org/482131 | 12:55 |
| *** woodard has quit IRC | 12:58 | |
| *** electrofelix has quit IRC | 12:58 | |
| *** woodard has joined #openstack-ansible | 12:58 | |
| *** aludwar has joined #openstack-ansible | 12:58 | |
| *** schwicht has joined #openstack-ansible | 13:00 | |
| *** fxpester has quit IRC | 13:00 | |
| *** woodard_ has joined #openstack-ansible | 13:01 | |
| *** fxpester has joined #openstack-ansible | 13:01 | |
| *** woodard_ has quit IRC | 13:01 | |
| *** woodard_ has joined #openstack-ansible | 13:02 | |
| *** woodard has quit IRC | 13:03 | |
| * mhayden passes odyssey4me some biltong | 13:05 | |
| odyssey4me | :) | 13:06 |
| *** lbragstad has joined #openstack-ansible | 13:08 | |
| *** esberglu has joined #openstack-ansible | 13:09 | |
| *** kristian__ has joined #openstack-ansible | 13:09 | |
| *** electrofelix has joined #openstack-ansible | 13:10 | |
| andymccr | mhayden: don't make promises you can't deliver | 13:10 |
| mhayden | i need to order some | 13:10 |
| andymccr | we now have expectations for Denver | 13:10 |
| mhayden | the site offered several different form factors | 13:10 |
| mhayden | i was confused on cuts | 13:10 |
| andymccr | mhayden: impress us. | 13:11 |
| kristian__ | Hi, I have a question on deploying for now aio and in the future scale openstack-ansible on ovh. I have a 128gb ram 12 thread bare metal there and Im getting /28 RIPE block from them (because I only have to pay for setup fees). And I have only one interface there (public). It's an enterprise server. Is it possible to deploy for now openstack as aio and in the future scale it up with only public network? | 13:13 |
| mhayden | kristian__: if i remember correctly, mattt has done this | 13:13 |
| kristian__ | mattt: could you help me? | 13:14 |
| odyssey4me | mhayden if you get it sliced, be aware that very thinly sliced dries out quickly and that's not good | 13:14 |
| odyssey4me | biltong should be wet, not dry | 13:14 |
| mhayden | kristian__: you should be able to extend it to new servers from the initial AIO, but you'll need to ensure that you can get your management/vxlan/vlan networks working between the two | 13:14 |
| andymccr | +2 on the biltong tip there. | 13:15 |
| mhayden | for example, neutron agent on server 2 will need to reach rabbitmq on server 1 | 13:15 |
| odyssey4me | kristian__ just do not use the development AIO - go through the deploy guide and build your own config | 13:15 |
| kristian__ | mhayden: thats another problem, because I have only I ext nic | 13:15 |
| mhayden | that is challenging | 13:15 |
| mhayden | you really need some VLANs provided by an external network | 13:15 |
| *** acormier has joined #openstack-ansible | 13:15 | |
| evrardjp | not really | 13:15 |
| mhayden | kristian__: i got my start that way and then did a colo instead :/ | 13:15 |
| mhayden | odyssey4me: i'll keep that in mind | 13:16 |
| kristian__ | I will, but is it possible to do prod openstack on 1 nic that is public? and on ovh net? | 13:16 |
| evrardjp | it should | 13:16 |
| mhayden | you could do prod on 1 nic | 13:16 |
| kristian__ | mhayden: what do you mean by colo? | 13:16 |
| mhayden | depending on your performance requirements ;) | 13:16 |
| mhayden | kristian__: colocation -- where you bring network devices + servers to a datacenter and they host them | 13:17 |
| mhayden | you get more control over networking and such | 13:17 |
| kristian__ | there will not be many vms | 13:17 |
| kristian__ | oh ok, but I dont have any :( | 13:17 |
| mhayden | yeah, the initial cost isn't fun | 13:17 |
| mhayden | but if you don't have many VMs, one server might be enough | 13:17 |
| evrardjp | it's technically doable kristian__ | 13:17 |
| *** udesale has quit IRC | 13:17 | |
| mhayden | especially with 128GB ram | 13:17 |
| kristian__ | yeah, and upgrading later | 13:18 |
| evrardjp | like odyssey4me told you, go through the deploy guide | 13:18 |
| evrardjp | it will help you understand the concepts | 13:18 |
| kristian__ | yeah I know, but I meant I wont have a lot of vms, I meant a long time | 13:18 |
| kristian__ | vms will go and die | 13:19 |
| kristian__ | there might be a huge no of vm requests and openstack will panic because it will not have resources | 13:19 |
| kristian__ | also there will be a k8s cluster | 13:19 |
| kristian__ | I will read the deploy guide, but when I will be deploying (today or tomorrow, currently playing with aio on packet.net), could you help me if I will get stuck? | 13:21 |
| kristian__ | it will be a pike install with magnum | 13:22 |
| admin0 | kristian__, i have a use case documented in www.openstackfaq.com where i take a big server and slice it up for multi-node install | 13:28 |
| admin0 | that way, you can start small, and grow big later on | 13:29 |
| admin0 | and a single nic is also possible | 13:29 |
| kristian__ | I will check it out, and I dont want to slice the server right now, but I will check it out | 13:30 |
| kristian__ | also got my ripe block :D | 13:30 |
| *** weezS has joined #openstack-ansible | 13:31 | |
| *** askb has quit IRC | 13:35 | |
| *** weezS has quit IRC | 13:35 | |
| evrardjp | the hardest is to keep it for you, if it's a large block :D | 13:38 |
| *** weezS has joined #openstack-ansible | 13:39 | |
| openstackgerrit | Andy McCrae proposed openstack/openstack-ansible-os_nova master: [WIP] Move to use UWsgi for Nova https://review.openstack.org/451425 | 13:39 |
| *** Martin___ has joined #openstack-ansible | 13:40 | |
| Martin___ | Hi, I am installing openstack-ansible on a heterogenous setup using also arm boards (more specifically odroids). When running the setup-openstack playbook I get a package conflict on the odroid because python-openssl requires a more recent version than is available in the ubuntu repositories: | 13:41 |
| Martin___ | python-openssl : Depends: python-cryptography (>= 1.3.4) but 1.2.3-1ubuntu0.1 is to be installed | 13:41 |
| Martin___ | It seems that the new version of python-openssl is be provided by the repositories of OSA as it is not available in a fresh system image. Am I right with this assumption? And why is only a new version of python-openssl pushed to the node and not of the required package? | 13:41 |
| *** schwicht has quit IRC | 13:44 | |
| odyssey4me | Martin___ OSA simply uses the repositories configured on the host | 13:45 |
| odyssey4me | we only add extra repositories when needed | 13:45 |
| *** acormier has quit IRC | 13:45 | |
| odyssey4me | especially for the OS repo, we don't add anything - we adjust the config, but use whatever was configured there | 13:45 |
| odyssey4me | the development AIO is a little more forceful | 13:46 |
| *** klamath has joined #openstack-ansible | 13:46 | |
| *** klamath has quit IRC | 13:46 | |
| *** klamath has joined #openstack-ansible | 13:46 | |
| Martin___ | odyssey4me ok, then something else is adding those sources... | 13:48 |
| *** tots has joined #openstack-ansible | 13:48 | |
| mardim | Hello guys I have a question: Does anyone deployed OSA with ovs instead of linux-bridges ? | 13:48 |
| Martin___ | but as I said: on a fresh system the problematic versions are not available. At least, apt-cache says so | 13:48 |
| odyssey4me | Martin___ so with your fresh odroid, what sources are configured - and are there packages installed that are not available in the mirror ? | 13:49 |
| odyssey4me | did you update the cache though | 13:49 |
| odyssey4me | mardim some have, not sure if anyone's using it in production | 13:49 |
| *** schwicht has joined #openstack-ansible | 13:50 | |
| Martin___ | odyssey4me yep, the cache is updated. And sources are those stated in the image by hardkernel, i.e. http://ports.ubuntu.com/ubuntu-ports/. I alwys thought those were correct. | 13:51 |
| mattt | kristian__: oh dear, i did this a few years back, it was a bit hacky tbh | 13:51 |
| mardim | My question is that you need additional playbooks to create the ovs bridges | 13:52 |
| mardim | ? | 13:52 |
| mardim | right ? | 13:52 |
| mardim | right now the ansible-neutron doesn't have this functionality | 13:52 |
| mardim | someone should provide the playbooks right ? | 13:52 |
| odyssey4me | Martin___ odd, are you able to identify whether the problematic sources are different to those origial ones? | 13:53 |
| odyssey4me | mardim the host bridge setup is, and always has been, up to the deployer in an OSA environment | 13:53 |
| mardim | odyssey4me: So I need to create my own playbooks for the ovs setup right ? | 13:55 |
| mattt | kristian__: i think i made some rough notes in a wordpress blog i had at the time, but i've since taken the blog down so i'll need to dig through backups and see if i can find it for you | 13:55 |
| odyssey4me | mardim the OVS install will be done by OSA, but the mapping of the OVS bridges to host bridges - not sure | 13:55 |
| odyssey4me | I've never done it, so I can't really comment. | 13:55 |
| odyssey4me | I would think that if we could do that, then we should, as we do it for lxb too. | 13:56 |
| mardim | Yes I know thanks odyssey4me that was helpfull :) | 13:56 |
| mardim | that is a struggle | 13:56 |
| mardim | :) | 13:56 |
| odyssey4me | mardim most OSA consumers do not use OVS, so if that matters to you it'd be great if you could help validate/update the docs and improve the implementation | 13:57 |
| mardim | Right now I am trying to do the mapping by hand | 13:57 |
| mardim | when I will figure out which is the correct mapping | 13:57 |
| mardim | I will be gloud to do the implementation of that :0 | 13:58 |
| mardim | :) | 13:58 |
| mardim | glad* | 13:58 |
| admin0 | i just had an environment up with QNAP and OVS | 13:58 |
| admin0 | OVS initial bridge mappings = manual | 13:59 |
| *** woodard_ has quit IRC | 13:59 | |
| admin0 | but once its done, rest works | 13:59 |
| mardim | admin0: So you added the bridges by hand right ? | 13:59 |
| mardim | ovs-bridges* | 13:59 |
| admin0 | correct | 14:00 |
| mardim | thanks a lot :0 | 14:00 |
| mardim | :) | 14:00 |
| admin0 | though i was planning to stick to LB, ovs was a company requirement for some stuff, so there were a few hours of figuring it out stuff | 14:01 |
| Martin___ | odyssey4me I do not see new sources popping up in /etc/apt/sources.list. I will check again on the packages, maybe I get an idea where they are coming from. | 14:01 |
| *** esberglu has quit IRC | 14:02 | |
| odyssey4me | admin0 if you could register a bug with the mappings required and how they can be implemented, then it can be easily converted into a patch | 14:02 |
| admin0 | mardim, i am not sure which part you are trying to automate in ovs setup | 14:02 |
| *** esberglu has joined #openstack-ansible | 14:02 | |
| mardim | admin0: Actually I am trying to do the mapping according to that tutorial : https://docs.openstack.org/liberty/networking-guide/scenario-classic-ovs.html | 14:03 |
| admin0 | in my case, br-mgmt and br-storage is LB, br-vlan and br-vxlan is oVS | 14:03 |
| mardim | I am trying to automate the creation of br-ex so I can have external connectivity | 14:03 |
| admin0 | aah .. | 14:03 |
| admin0 | well, people have diff network requiremnet .. under that br could be a single nick vlan, a bond, a group of nics , etc .. so that part is left manual based on what operator wants to setup .. OSA just cares if there is a bridge or not | 14:04 |
| mardim | I am trying this tutorial because with this tutorial I succedee to create vms in a teneant network | 14:04 |
| admin0 | anyone knows what software was used to create that diagram ? | 14:05 |
| admin0 | in the wiki ? | 14:05 |
| mardim | What do you mean waht software ? | 14:05 |
| mardim | which agent ? | 14:05 |
| admin0 | nah .. it was a diff question .. if to contribute diagrams, use the same software | 14:05 |
| Martin___ | odyssey4me So, after running setup -hosts, there is a new package of python-openssl (16.1.0-1~cloud0 in addition to 0.15.1-2build1). And there is a new new source: deb http://ubuntu-cloud.archive.canonical.com/ubuntu xenial-updates/newton main | 14:06 |
| admin0 | ok .. so in your case, you can use ansible to create the bridge and map it accordingly, or use MaaS or use the kickstart | 14:06 |
| mardim | oh ok :) | 14:06 |
| Martin___ | I didnt see the latter before because it is located in a subdir of /etc/apt | 14:06 |
| odyssey4me | Martin___ aha, yes - we do add UCA to get more up to date packages | 14:06 |
| mardim | Yes I think I will use ansible aand when I will figure it out the correct mapping by hand | 14:07 |
| odyssey4me | Martin___ yes, typically /etc/apt/sources.list is for the distro and /etc/apt/sources.list.d/ contains files with extra sources | 14:07 |
| mardim | I will contribute a playbook for that to the ansible neutron project | 14:07 |
| mardim | that's the plan | 14:07 |
| odyssey4me | Martin___ ok, so now we know where the source comes from, so back to the original problem... you have no source available which has a package that's already installed... how is that the case now if the extra sources was added | 14:08 |
| mardim | admin0: Do you have any helpful guide for that ? | 14:09 |
| mardim | from your own experience / | 14:09 |
| mardim | ? | 14:09 |
| admin0 | i have some of my setup documented in www.openstackfaq.com | 14:09 |
| kristian__ | evrardjp: it can be extended in the future or no? mattt: I will propably do it tomorrow (Slovak timezone) if you would be here and help me? (pm will be great). I also found some blogs too, that spoof the routers mac address. We will see. | 14:10 |
| admin0 | right now, i am working to make the same guide but on ocata and including what i learnt on the ovs+qnap setup | 14:10 |
| *** gouthamr has joined #openstack-ansible | 14:10 | |
| mardim | oh ok thanks | 14:10 |
| Martin___ | odyssey4me No, the problem is that python-openssl(16.1.0) requires python-crytography in a newer version than is available in the system repos. The extra sources are not added to the node. | 14:10 |
| Martin___ | odyssey4me I could try to add it manually... | 14:11 |
| odyssey4me | mhayden posed a question in https://review.openstack.org/#/c/481203/1/doc/source/getting-started.rst | 14:11 |
| Martin___ | odyssey4me And when I do so, I also only get the new python-openssl package but not the python-cryptography one | 14:14 |
| Martin___ | odyssey4me i think I know why: on the x86 control node I get the newer version of python-cryptography but that one is stated to be for amd64 | 14:16 |
| *** weezS has quit IRC | 14:17 | |
| spotz | Y'all miss me?:) | 14:23 |
| Adri2000 | hi | 14:27 |
| *** galstrom_zzz is now known as galstrom | 14:28 | |
| andymccr | spotz: how'd the pony show go?! | 14:28 |
| *** galstrom is now known as galstrom_zzz | 14:28 | |
| *** galstrom_zzz is now known as galstrom | 14:28 | |
| spotz | 6th place out of 18 in our division. Very proud of the pony. She's exhausted | 14:28 |
| *** cpuga has joined #openstack-ansible | 14:29 | |
| andymccr | good job | 14:29 |
| admin0 | \o spotz | 14:29 |
| Adri2000 | in `openstack-ansible -e foo=bar playbook.yml`, is `foo=bar` supposed to override the value of `foo` that may have been set in user_variables.yml for example? my tests with debug=True (when user_variables.yml has debug=False) imply that it's not overriding anything... | 14:29 |
| odyssey4me | Adri2000 that's likely because the /etc/openstack_deploy extra vars are added to the end of the command, and thus override | 14:30 |
| andymccr | Adri2000: yeah - although user_variables counts as a -e too, so i guess that would be a precedence thing. | 14:30 |
| spotz | andymccr: thanks:) | 14:30 |
| spotz | hey admin0! | 14:30 |
| admin0 | whats a pony show ? | 14:31 |
| openstackgerrit | git-harry proposed openstack/openstack-ansible-ops master: Add support for separate neutron network hosts https://review.openstack.org/482159 | 14:31 |
| admin0 | a show where a pony is shown :D ? | 14:31 |
| *** dixiaoli has joined #openstack-ansible | 14:33 | |
| *** dixiaoli has quit IRC | 14:33 | |
| *** dixiaoli has joined #openstack-ansible | 14:34 | |
| Adri2000 | odyssey4me, andymccr: I'll try with moving ${VAR1} before "${@}" | 14:34 |
| odyssey4me | Adri2000 better to just set the override in one place instead of mucking with precedence | 14:35 |
| mhayden | odyssey4me: reasonable request - i'll rework that | 14:36 |
| *** dixiaoli has quit IRC | 14:38 | |
| *** cshen has quit IRC | 14:39 | |
| openstackgerrit | Merged openstack/openstack-ansible-tests master: test-prepare-host: Fix POST_{UP,DOWN}_SCRIPT handling for SUSE https://review.openstack.org/482131 | 14:39 |
| openstackgerrit | Major Hayden proposed openstack/ansible-hardening master: Fix grep for sudoers w/o password https://review.openstack.org/481226 | 14:40 |
| *** cshen has joined #openstack-ansible | 14:41 | |
| *** woodard has joined #openstack-ansible | 14:41 | |
| *** phalmos has joined #openstack-ansible | 14:42 | |
| *** woodard has quit IRC | 14:45 | |
| *** woodard_ has joined #openstack-ansible | 14:45 | |
| openstackgerrit | Major Hayden proposed openstack/ansible-hardening master: [Docs] Make install/usage docs more clear https://review.openstack.org/481203 | 14:45 |
| *** weezS has joined #openstack-ansible | 14:45 | |
| *** woodard_ has quit IRC | 14:45 | |
| *** woodard has joined #openstack-ansible | 14:46 | |
| *** weezS has quit IRC | 14:46 | |
| cloudnull | mornings | 14:48 |
| *** chyka has joined #openstack-ansible | 14:48 | |
| *** toddnni has quit IRC | 14:48 | |
| admin0 | cloudnull \o | 14:48 |
| *** chyka has quit IRC | 14:48 | |
| cloudnull | o/ | 14:49 |
| *** chyka has joined #openstack-ansible | 14:49 | |
| openstackgerrit | Jesse Pretorius (odyssey4me) proposed openstack/openstack-ansible stable/newton: Implement rolling upgrades for nova https://review.openstack.org/481925 | 14:53 |
| *** udesale has joined #openstack-ansible | 14:57 | |
| *** weezS has joined #openstack-ansible | 15:01 | |
| *** udesale has quit IRC | 15:02 | |
| *** admin0 has quit IRC | 15:04 | |
| *** Oku_OS is now known as Oku_OS-away | 15:16 | |
| *** lkoranda has quit IRC | 15:19 | |
| *** nollide has joined #openstack-ansible | 15:19 | |
| Adri2000 | odyssey4me: well, don't you think my `debug` variable example is a good use case of wanting to override a var just for one run? in this case I wanted to debug Horizon temporarily, and would rather not change user_variables.yml and risk forgetting I made the change there | 15:22 |
| *** lkoranda has joined #openstack-ansible | 15:22 | |
| odyssey4me | Adri2000 but debug is disabled by default, so you don't need it in user_variables.yml if the value is the same as the default value | 15:23 |
| Adri2000 | ah, true | 15:23 |
| shukixxxxx | odyssey4me OK; so my mirror is broken? But how? I just installed 16.04.02 LTS and use the deployment guide; apt update etc.. how did it reach this state that it have newer version than the version required by the ansible-openstack dep[loyment? | 15:27 |
| *** cshen has quit IRC | 15:28 | |
| odyssey4me | shukixxxxx good question, and that's going to be up to you to figure out | 15:28 |
| odyssey4me | if it's the container that does, then it's likely that the lxc base cache has updated packages - the base cache is pulled from an upstream resource and therefore uses different apt sources | 15:29 |
| odyssey4me | so if it is the container, then you either need to update your mirror or you need to implement config to downgrade the packages to the versions you have | 15:29 |
| openstackgerrit | Juan Vidal Allende proposed openstack/openstack-ansible-os_neutron master: [WIP] Add option to deploy OpenDaylight SDN Controller https://review.openstack.org/480128 | 15:30 |
| shukixxxxx | I just followed the instructions in the guide and I use deb http://us.archive.ubuntu.com/ubuntu/ in the /etc/apt/source.list | 15:30 |
| odyssey4me | shukixxxxx by way of example, this is what we use to downgrade the cache packages when we build out container artifacts: https://github.com/rcbops/rpc-openstack/blob/master/scripts/artifacts-building/user_rcbops_artifacts_building.yml#L34-L96 | 15:30 |
| odyssey4me | shukixxxxx ok, then I do not know - you'll have to work it out yourself | 15:31 |
| *** vnogin has quit IRC | 15:31 | |
| shukixxxxx | Is this repository is broken?? | 15:31 |
| *** vnogin has joined #openstack-ansible | 15:31 | |
| odyssey4me | shukixxxxx I doubt it. I can tell you though that we are building AIO's every day without the issue you're describing, so the issue is particular to your situation and only you're going to be able to figure it out. | 15:32 |
| odyssey4me | That URL, I think, is directed at a bunch of many mirrors - so it's possible that you got caught with a bad one. | 15:33 |
| odyssey4me | IDK, you'll have to do more digging. | 15:33 |
| openstackgerrit | Hugh Saunders proposed openstack/openstack-ansible-os_tempest stable/newton: Add block/always to tempest result file tasks https://review.openstack.org/482186 | 15:34 |
| *** vnogin has quit IRC | 15:36 | |
| *** vnogin has joined #openstack-ansible | 15:37 | |
| shukixxxxx | OK; which repository address you are using to deploy AIO; I would try it | 15:37 |
| Martin___ | odyssey4me I tried moving to the yakkety sources, that seems to solve this issue but gives me new errors :). Thanks for your advice | 15:38 |
| *** flaviosr has joined #openstack-ansible | 15:41 | |
| *** jvidal has quit IRC | 15:41 | |
| *** vnogin has quit IRC | 15:42 | |
| *** dxiri has quit IRC | 15:44 | |
| openstackgerrit | Andy McCrae proposed openstack/openstack-ansible-os_nova master: [WIP] Move to use UWsgi for Nova https://review.openstack.org/451425 | 15:46 |
| *** shukixxxxx has quit IRC | 15:51 | |
| *** jvidal has joined #openstack-ansible | 15:57 | |
| openstackgerrit | Nolan Brubaker proposed openstack/openstack-ansible master: Install swift serially for rolling updates https://review.openstack.org/478580 | 15:57 |
| *** kristian__ has quit IRC | 15:57 | |
| *** kristian__ has joined #openstack-ansible | 15:58 | |
| *** dxiri has joined #openstack-ansible | 15:58 | |
| *** cathrichardson has joined #openstack-ansible | 15:59 | |
| dmsimard | odyssey4me: found the issue for your subunit patch, added a comment | 16:01 |
| odyssey4me | ORLY?! :) | 16:01 |
| *** cathrich_ has quit IRC | 16:02 | |
| dmsimard | oh man, there should totally be an ascii logo of the ARA bird like the orly owl | 16:03 |
| *** nollide has quit IRC | 16:10 | |
| odyssey4me | dmsimard great feedback, thanks - popped a few extra questions in there before I do another round | 16:10 |
| *** nollide has joined #openstack-ansible | 16:11 | |
| odyssey4me | I guess I suspected the play path more than the task path - I did try manually assigning both in a test to a static string, but the error still came out... so I dunno wtf is going on | 16:11 |
| dmsimard | odyssey4me: on a second thought, perhaps we'd rather hash the full path to the file rather than just it's basename ? | 16:13 |
| odyssey4me | that's one of the questions I posed ;) | 16:13 |
| dmsimard | odyssey4me: yeah, that's why I'm bringing it up :p | 16:13 |
| dmsimard | odyssey4me: two files can have the same name rather easily (especially for i.e, main.yml) | 16:14 |
| odyssey4me | yeah, my only concern would be a situation where it's a CI job which uses a dynamically generated workspace folder to hold the playbooks/roles | 16:15 |
| odyssey4me | so the ideal is for it to be a relative path of some sort | 16:16 |
| odyssey4me | ie relative to the root of the roles dir and relative to the playbook dir | 16:17 |
| *** pvradu has quit IRC | 16:17 | |
| odyssey4me | but that's not really something we can figure out, I don't think | 16:17 |
| *** nollide has quit IRC | 16:19 | |
| dmsimard | odyssey4me: eh, I mean relative up to which point ? I mean, if roles are in /etc/ansible/roles and the playbook is in /home/jenkins for example.. | 16:20 |
| dmsimard | odyssey4me: basename or not, there are cases where the uniqueness will be off I guess :p | 16:21 |
| odyssey4me | yea, that's why the concatenation of both paths hopefully gives a better value than relying on only one | 16:22 |
| odyssey4me | well, play_path + play_name + task_path + task_name | 16:22 |
| *** nollide has joined #openstack-ansible | 16:22 | |
| dmsimard | so I'll be a troublemaker | 16:23 |
| dmsimard | What if I run another playbook which ends up including those same plays and tasks | 16:23 |
| dmsimard | Would I want the plays and tasks from playbook A to end up trending with the ones from playbook B ? :) | 16:24 |
| dmsimard | I think the answer is "maybe" ? | 16:24 |
| odyssey4me | well, if those tasks are 'install packages' then the resulting times will vary massively | 16:24 |
| openstackgerrit | Jean-Philippe Evrard proposed openstack/openstack-ansible-ops master: Do not delete containers named rpc.* https://review.openstack.org/482208 | 16:25 |
| odyssey4me | maybe the really right way to go about this is to allow the identifier for the task to be customised when generating | 16:25 |
| odyssey4me | that therefore allows the use of a default, but the ability to override | 16:25 |
| dmsimard | odyssey4me: exactly what I was going to say. I think it'd be hard to come up with a perfect solution, let's come up with a sane default and then allow it to be overidden | 16:26 |
| dmsimard | odyssey4me: let's land the default in a first patch, we can take care of the config later | 16:27 |
| dmsimard | odyssey4me: fwiw we should probably check with hughsaunders (would this approach be cool with junit?) and see how this ends up fitting with how openstack-health expects the data | 16:28 |
| *** weezS_ has joined #openstack-ansible | 16:29 | |
| odyssey4me | yeah, I haven't looked into how openstack health extracts the data and what data sets it therefore may expect, if it's even that sophisticated | 16:29 |
| *** weezS has quit IRC | 16:30 | |
| *** weezS_ is now known as weezS | 16:30 | |
| *** cshen has joined #openstack-ansible | 16:34 | |
| * dmsimard nods | 16:35 | |
| dmsimard | I'm admittedly not very knowledgeable around junit/subunit and definitely not a user (yet) | 16:35 |
| dmsimard | I can try and rope in mtreinish and ask for his opinion once we get the patch working | 16:36 |
| odyssey4me | yeah, let me work up a revision quickly | 16:37 |
| *** shardy is now known as shardy_afk | 16:37 | |
| *** dxiri has quit IRC | 16:38 | |
| *** dxiri has joined #openstack-ansible | 16:39 | |
| *** stuartgr has quit IRC | 16:46 | |
| *** dxiri has quit IRC | 16:55 | |
| *** dxiri has joined #openstack-ansible | 16:56 | |
| *** weezS has quit IRC | 16:56 | |
| *** weezS has joined #openstack-ansible | 17:02 | |
| *** acormier has joined #openstack-ansible | 17:05 | |
| *** cmart has quit IRC | 17:13 | |
| *** acormier has quit IRC | 17:14 | |
| *** acormier has joined #openstack-ansible | 17:14 | |
| tots | hello , any idea what this means? http://paste.openstack.org/show/614953/ | 17:16 |
| tots | odyssey4me: i sorted the glance issue with the nfs shares etc | 17:16 |
| *** acormier has quit IRC | 17:17 | |
| tots | now i get the above error when trying to lunch an instance | 17:17 |
| dmsimard | Just noticed an interesting behavior for blocks | 17:17 |
| dmsimard | If you have a single task with a "when" condition that is false, it will trigger the task but mark it as skipped | 17:18 |
| dmsimard | If you have tasks inside a block with a "when" condition that is false, the tasks do not register at all | 17:18 |
| *** mbuil has quit IRC | 17:20 | |
| openstackgerrit | Merged openstack/openstack-ansible master: Tidy up keystone need_db_sync fact https://review.openstack.org/482043 | 17:27 |
| *** weezS has quit IRC | 17:34 | |
| *** lostRhino has quit IRC | 17:50 | |
| *** acormier has joined #openstack-ansible | 17:50 | |
| *** acormier has quit IRC | 17:50 | |
| *** acormier has joined #openstack-ansible | 17:51 | |
| *** kristia__ has joined #openstack-ansible | 17:52 | |
| openstackgerrit | Nolan Brubaker proposed openstack/openstack-ansible master: Cleanly catch keyboard interrupts during tests https://review.openstack.org/481199 | 17:53 |
| *** kristian__ has quit IRC | 17:56 | |
| mhayden | lhinds: https://review.openstack.org/#/c/482231/ | 18:00 |
| *** jamesdenton has quit IRC | 18:01 | |
| *** dmellado has quit IRC | 18:02 | |
| *** dmellado has joined #openstack-ansible | 18:02 | |
| *** schwicht has quit IRC | 18:02 | |
| *** jamesdenton has joined #openstack-ansible | 18:03 | |
| *** kristia__ has quit IRC | 18:03 | |
| *** poopcat has joined #openstack-ansible | 18:04 | |
| openstackgerrit | Jesse Pretorius (odyssey4me) proposed openstack/openstack-ansible stable/ocata: Tidy up keystone need_db_sync fact https://review.openstack.org/482232 | 18:06 |
| openstackgerrit | Jesse Pretorius (odyssey4me) proposed openstack/openstack-ansible stable/newton: Tidy up keystone need_db_sync fact https://review.openstack.org/482233 | 18:07 |
| *** schwicht has joined #openstack-ansible | 18:07 | |
| *** electrofelix has quit IRC | 18:09 | |
| *** albertcard1 has joined #openstack-ansible | 18:09 | |
| *** jvidal has quit IRC | 18:10 | |
| lhinds | thx mhayden | 18:11 |
| *** kylek3h has joined #openstack-ansible | 18:11 | |
| *** poopcat has quit IRC | 18:11 | |
| *** kristian__ has joined #openstack-ansible | 18:15 | |
| *** skape has joined #openstack-ansible | 18:15 | |
| *** kristian__ has quit IRC | 18:15 | |
| *** pbandark1 has joined #openstack-ansible | 18:18 | |
| *** pbandark has quit IRC | 18:18 | |
| *** pbandark1 is now known as pbandark | 18:18 | |
| *** cmart has joined #openstack-ansible | 18:19 | |
| skape | Hi to all ! Is there an example file on how to configure the file openstack_user_config.yml with glance using a ceph cluster as storage ? or where do I set glance variables like stores default_store rbd_store_pool | 18:20 |
| mgariepy | skape, https://github.com/openstack/openstack-ansible-os_glance/blob/stable/ocata/defaults/main.yml#L194-L198 | 18:32 |
| mgariepy | skape, and glance_default_store: rbd into your user_variables.yml | 18:33 |
| mgariepy | skape, you probably use this as a base : http://www.openstackfaq.com/openstack-ansible-ceph/ | 18:34 |
| *** weezS has joined #openstack-ansible | 18:40 | |
| skape | mgariepy , thanks ! I will take a look | 18:40 |
| tots | after creating an image_hosts and creating the nfs shares for glance and copying my images onto my nfs server the glance role finished with no errors but i can not launch a new instance i get this error :http://paste.openstack.org/show/614953/ somehow it does not detect my compute hosts | 18:43 |
| openstackgerrit | Miguel Alex Cantu (alextricity25) proposed openstack/openstack-ansible-ops master: Remove upgrades from Juno https://review.openstack.org/482242 | 18:46 |
| openstackgerrit | Miguel Alex Cantu (alextricity25) proposed openstack/openstack-ansible-ops master: Remove ability to upgrade from Juno https://review.openstack.org/482242 | 18:47 |
| *** shardy_afk is now known as shardy | 18:59 | |
| *** galstrom is now known as galstrom_zzz | 19:01 | |
| openstackgerrit | Merged openstack/openstack-ansible master: py35: Set virtualenv options for CentOS https://review.openstack.org/482052 | 19:05 |
| openstackgerrit | Jesse Pretorius (odyssey4me) proposed openstack/openstack-ansible master: Remove run-playbooks https://review.openstack.org/482253 | 19:14 |
| odyssey4me | tots have you checked the scheduler log to see why it's rejecting the compute hosts? | 19:15 |
| odyssey4me | or the compute host logs to see whether there are attempts, but they're failing? | 19:15 |
| tots | im there give me a sec | 19:18 |
| tots | all previous instances are still running though | 19:18 |
| tots | sec | 19:18 |
| tots | odyssey4me: all i see in the scheduler logs are lines like this2017-07-10 21:10:38.335 24061 INFO nova.scheduler.host_manager [req-f7257961-953a-40b3-b89b-49fe5eb24c2c - - - - -] Successfully synced instances from host hostnamehere and no errors at all | 19:21 |
| *** acormier has quit IRC | 19:24 | |
| tots | odyssey4me: nova conductor is showing these errors http://paste.openstack.org/show/614957/ | 19:24 |
| *** kristian__ has joined #openstack-ansible | 19:34 | |
| *** tobberydberg has joined #openstack-ansible | 19:38 | |
| *** kristian__ has quit IRC | 19:39 | |
| *** tobberydberg has quit IRC | 19:43 | |
| *** hybridpollo has joined #openstack-ansible | 19:45 | |
| *** kristian__ has joined #openstack-ansible | 19:45 | |
| *** cathrichardson has quit IRC | 19:49 | |
| *** cathrichardson has joined #openstack-ansible | 19:49 | |
| *** pbandark has quit IRC | 19:50 | |
| *** albertcard1 is now known as poopcat | 19:51 | |
| *** kristian__ has quit IRC | 19:52 | |
| openstackgerrit | Jesse Pretorius (odyssey4me) proposed openstack/openstack-ansible master: [WIP] Just-in-time container creation https://review.openstack.org/471026 | 19:52 |
| *** kristian__ has joined #openstack-ansible | 20:01 | |
| *** cshen has quit IRC | 20:02 | |
| hughsaunders | dmsimard / odyssey4me added some comments to the subunit review | 20:04 |
| openstackgerrit | Jesse Pretorius (odyssey4me) proposed openstack/openstack-ansible master: [WIP] Just-in-time container creation https://review.openstack.org/471026 | 20:05 |
| *** cshen has joined #openstack-ansible | 20:06 | |
| openstackgerrit | weezer su proposed openstack/openstack-ansible-ops master: Be sure mysql_config is really installed. https://review.openstack.org/482279 | 20:26 |
| *** acormier has joined #openstack-ansible | 20:33 | |
| *** acormier has quit IRC | 20:37 | |
| openstackgerrit | Merged openstack/openstack-ansible stable/newton: Tidy up keystone need_db_sync fact https://review.openstack.org/482233 | 20:37 |
| *** asura has joined #openstack-ansible | 20:39 | |
| odyssey4me | anyone other than mhayden got a gap to review the final two rolling-upgrade backports? https://review.openstack.org/481924 & https://review.openstack.org/481925 | 20:42 |
| mhayden | oh i see how it is... | 20:42 |
| *** thorst has quit IRC | 20:42 | |
| odyssey4me | you're done mhayden, just done | 20:43 |
| asura | Hi, inside of /etc/openstack_deploy/openstack_user_config.yml I'm configuring storage_hosts and I'm curious if I use the following https://gist.github.com/michaelbarkdoll/79d1eb2506c713573d610011e5e11731 definition would I need to first create a LVM volume group named cinder-volumes prior to running the playbook? I'm curious how I'd just use local file storage instead of LVM's for a test environment? Thanks. | 20:43 |
| odyssey4me | asura yes, you would have to configure the cinder-volumes vg | 20:43 |
| asura | Thanks | 20:44 |
| odyssey4me | if you want just file system, then rather use an NFS service and configure cinder to use nfs | 20:44 |
| tots | odyssey4me: nova conductor is showing these errors http://paste.openstack.org/show/614957/ any chance you can point me where to look for solution? | 20:44 |
| odyssey4me | as shown in the prod env example | 20:44 |
| odyssey4me | tots keep digging - that's only showing you that it's not working - not why | 20:44 |
| odyssey4me | one of the other services will show why | 20:44 |
| odyssey4me | you may need debug logging to see | 20:45 |
| *** thorst has joined #openstack-ansible | 20:45 | |
| odyssey4me | tots conductor just relays messages, it doesn't make decisions | 20:45 |
| tots | if i delete all images and recreate them , would this go away ? | 20:45 |
| odyssey4me | decisions are made by the scheduler, placement service and compute hosts | 20:45 |
| odyssey4me | tots that would depend if the issue relates ot images | 20:46 |
| odyssey4me | you have to establish the problem before you attempt to solve it | 20:46 |
| tots | this happened after i did the image_hosts and copied image ids from /var/lib/glance/images to the nfs share | 20:46 |
| odyssey4me | you need to trace it through each service in the chain | 20:47 |
| odyssey4me | time to learn how openstack works ;) | 20:47 |
| tots | nova scheduler logs have no errors onlySuccessfully synced instances | 20:49 |
| *** thorst has quit IRC | 20:49 | |
| *** galstrom_zzz is now known as galstrom | 20:52 | |
| tots | just did 3 ciros instances at once and they booted fine , tried 3 ubuntu insances and these failed with no available hosts | 20:54 |
| openstackgerrit | Markos Chandras (hwoarang) proposed openstack/openstack-ansible-os_ceilometer master: test: test-install-mongodb: Do not mask mongodb_user failures https://review.openstack.org/482310 | 20:56 |
| openstackgerrit | Markos Chandras (hwoarang) proposed openstack/openstack-ansible-os_ceilometer master: tests: Provide mongodb configuration template https://review.openstack.org/482311 | 20:56 |
| openstackgerrit | Markos Chandras (hwoarang) proposed openstack/openstack-ansible-os_ceilometer master: Add support for the openSUSE Leap distributions https://review.openstack.org/482312 | 20:56 |
| openstackgerrit | Jesse Pretorius (odyssey4me) proposed openstack/openstack-ansible master: [WIP] Just-in-time container creation https://review.openstack.org/471026 | 20:56 |
| odyssey4me | tots so it sounds like the glance services are just as they were before | 20:59 |
| odyssey4me | ie not using the nfs store - so check it again and be sure | 20:59 |
| odyssey4me | I'm out for the night, cheers all | 21:00 |
| *** thorst has joined #openstack-ansible | 21:00 | |
| *** smatzek has quit IRC | 21:09 | |
| *** phalmos has quit IRC | 21:11 | |
| *** askb has joined #openstack-ansible | 21:12 | |
| *** tots has quit IRC | 21:14 | |
| *** cshen has quit IRC | 21:14 | |
| *** tonytan4ever has joined #openstack-ansible | 21:19 | |
| *** john51 has quit IRC | 21:20 | |
| *** john51 has joined #openstack-ansible | 21:22 | |
| *** john51 has quit IRC | 21:23 | |
| openstackgerrit | Hugh Saunders proposed openstack/openstack-ansible-os_tempest stable/newton: Add block/always to tempest result file tasks https://review.openstack.org/482186 | 21:24 |
| *** john51 has joined #openstack-ansible | 21:27 | |
| *** cpuga has quit IRC | 21:27 | |
| *** cpuga_ has joined #openstack-ansible | 21:27 | |
| *** oanson has quit IRC | 21:28 | |
| *** cpuga_ has quit IRC | 21:28 | |
| *** oanson has joined #openstack-ansible | 21:29 | |
| *** cpuga has joined #openstack-ansible | 21:29 | |
| *** acormier has joined #openstack-ansible | 21:30 | |
| *** cpuga has quit IRC | 21:33 | |
| *** cmart has quit IRC | 21:34 | |
| *** acormier has quit IRC | 21:38 | |
| *** kristian__ has quit IRC | 21:38 | |
| *** acormier has joined #openstack-ansible | 21:38 | |
| openstackgerrit | Markos Chandras (hwoarang) proposed openstack/openstack-ansible-os_aodh master: templates: aodh-httpd: Ensure proper user control in aodh root https://review.openstack.org/482333 | 21:45 |
| openstackgerrit | Markos Chandras (hwoarang) proposed openstack/openstack-ansible-os_aodh master: Add support for the openSUSE Leap distributions https://review.openstack.org/482334 | 21:45 |
| *** woodard_ has joined #openstack-ansible | 21:48 | |
| *** esberglu has quit IRC | 21:49 | |
| *** woodard has quit IRC | 21:52 | |
| *** vishwanathj has joined #openstack-ansible | 21:52 | |
| *** woodard_ has quit IRC | 21:52 | |
| *** tots has joined #openstack-ansible | 21:53 | |
| *** cpuga has joined #openstack-ansible | 21:59 | |
| *** kylek3h has quit IRC | 22:00 | |
| *** thorst has quit IRC | 22:02 | |
| *** acormier has quit IRC | 22:10 | |
| *** galstrom is now known as galstrom_zzz | 22:10 | |
| *** schwicht has quit IRC | 22:12 | |
| *** LiterateHawk has joined #openstack-ansible | 22:15 | |
| LiterateHawk | Greetings, OSA fans -- trying to install a PoC cluster yields the following error: https://pastebin.com/cDjvS9U1 The deployment and installation hosts are all fresh xenial machines. I'm not good enough with Ansible to be able to grok the error, though. Any suggestions? | 22:17 |
| *** deep-book-gk_ has joined #openstack-ansible | 22:18 | |
| LiterateHawk | I should add, I'm using the latest stable/ocata branch and verified that bootstrap-ansible.sh succeeds | 22:18 |
| *** acormier has joined #openstack-ansible | 22:19 | |
| *** deep-book-gk_ has left #openstack-ansible | 22:20 | |
| *** nollide has quit IRC | 22:22 | |
| *** shardy has quit IRC | 22:22 | |
| *** dxiri has quit IRC | 22:23 | |
| *** dxiri has joined #openstack-ansible | 22:24 | |
| *** phalmos has joined #openstack-ansible | 22:27 | |
| *** schwicht has joined #openstack-ansible | 22:28 | |
| *** phalmos_ has joined #openstack-ansible | 22:31 | |
| *** phalmos has quit IRC | 22:31 | |
| *** DimGR has quit IRC | 22:39 | |
| *** weezS has quit IRC | 22:44 | |
| *** acormier has quit IRC | 23:06 | |
| *** acormier has joined #openstack-ansible | 23:07 | |
| *** schwicht has quit IRC | 23:07 | |
| *** acormier has quit IRC | 23:11 | |
| *** thorst has joined #openstack-ansible | 23:17 | |
| *** aludwar has quit IRC | 23:19 | |
| *** woodard has joined #openstack-ansible | 23:20 | |
| *** thorst has quit IRC | 23:21 | |
| *** charcol has joined #openstack-ansible | 23:23 | |
| tots | anyone around ? | 23:29 |
| *** acormier has joined #openstack-ansible | 23:30 | |
| *** weezS has joined #openstack-ansible | 23:43 | |
| *** aslaen has joined #openstack-ansible | 23:45 | |
| *** tots has quit IRC | 23:45 | |
| *** chyka has quit IRC | 23:54 | |
| *** chyka has joined #openstack-ansible | 23:55 | |
| *** thorst has joined #openstack-ansible | 23:57 | |
| *** weezS has quit IRC | 23:58 | |
| *** oanson has quit IRC | 23:59 | |
| *** chyka has quit IRC | 23:59 | |
Generated by irclog2html.py 2.15.3 by Marius Gedminas - find it at mg.pov.lt!