| *** rromans has joined #openstack-ansible | 00:02 | |
| *** cfarquhar has quit IRC | 00:04 | |
| *** cfarquhar has joined #openstack-ansible | 00:15 | |
| *** cfarquhar has quit IRC | 00:15 | |
| *** cfarquhar has joined #openstack-ansible | 00:15 | |
| *** shoutm has quit IRC | 00:22 | |
| *** darrenc is now known as darrenc_afk | 00:22 | |
| *** shoutm has joined #openstack-ansible | 00:23 | |
| *** woodard has joined #openstack-ansible | 00:34 | |
| *** woodard has quit IRC | 00:38 | |
| *** mrstanwell has quit IRC | 00:39 | |
| *** shoutm_ has joined #openstack-ansible | 00:40 | |
| *** shoutm has quit IRC | 00:41 | |
| *** tlian2 has joined #openstack-ansible | 00:54 | |
| *** skamithi12 has joined #openstack-ansible | 00:56 | |
| *** skamithi12 has quit IRC | 00:57 | |
| *** tlian has quit IRC | 00:58 | |
| *** skamithi12 has joined #openstack-ansible | 01:10 | |
| *** darrenc_afk is now known as darrenc | 01:12 | |
| *** Mudpuppy has joined #openstack-ansible | 01:13 | |
| *** Mudpuppy has quit IRC | 01:16 | |
| *** Mudpuppy has joined #openstack-ansible | 01:16 | |
| *** skamithi12 has quit IRC | 01:23 | |
| *** sdake has quit IRC | 01:26 | |
| *** tlian2 has quit IRC | 01:28 | |
| *** daneyon_ has quit IRC | 01:32 | |
| *** tlian has joined #openstack-ansible | 01:34 | |
| *** woodard has joined #openstack-ansible | 01:35 | |
| *** woodard has quit IRC | 01:39 | |
| *** jlvillal has quit IRC | 01:49 | |
| *** shoutm_ has quit IRC | 02:02 | |
| *** shoutm has joined #openstack-ansible | 02:09 | |
| *** daneyon has joined #openstack-ansible | 02:09 | |
| *** daneyon has quit IRC | 02:10 | |
| *** daneyon has joined #openstack-ansible | 02:12 | |
| *** daneyon has quit IRC | 02:13 | |
| *** daneyon has joined #openstack-ansible | 02:13 | |
| *** jlvillal has joined #openstack-ansible | 02:18 | |
| *** daneyon has quit IRC | 02:19 | |
| *** jlvillal has quit IRC | 02:35 | |
| *** arbrandes1 has joined #openstack-ansible | 02:48 | |
| *** arbrandes has quit IRC | 02:49 | |
| *** sigmavirus24_awa is now known as sigmavirus24 | 02:56 | |
| *** sigmavirus24 is now known as sigmavirus24_awa | 02:58 | |
| *** tlian has quit IRC | 03:31 | |
| *** sdake has joined #openstack-ansible | 03:55 | |
| *** daneyon has joined #openstack-ansible | 04:00 | |
| *** jbweber has joined #openstack-ansible | 04:42 | |
| *** skamithi12 has joined #openstack-ansible | 04:53 | |
| *** daneyon has quit IRC | 04:54 | |
| *** daneyon has joined #openstack-ansible | 04:56 | |
| *** daneyon has quit IRC | 04:58 | |
| *** daneyon has joined #openstack-ansible | 05:00 | |
| *** jlvillal has joined #openstack-ansible | 05:06 | |
| *** sdake has quit IRC | 05:13 | |
| *** daneyon has quit IRC | 05:30 | |
| *** shausy has joined #openstack-ansible | 05:37 | |
| *** shausy has quit IRC | 05:48 | |
| *** javeriak has joined #openstack-ansible | 05:57 | |
| *** javeriak has quit IRC | 06:05 | |
| *** javeriak has joined #openstack-ansible | 06:11 | |
| *** markvoelker has quit IRC | 06:16 | |
| *** skamithi12 has quit IRC | 06:27 | |
| *** javeriak_ has joined #openstack-ansible | 06:55 | |
| *** javeriak has quit IRC | 06:57 | |
| *** markvoelker has joined #openstack-ansible | 07:17 | |
| *** markvoelker has quit IRC | 07:22 | |
| *** javeriak has joined #openstack-ansible | 07:25 | |
| *** javeriak_ has quit IRC | 07:25 | |
| *** javeriak_ has joined #openstack-ansible | 07:35 | |
| *** javeriak has quit IRC | 07:37 | |
| *** gparaskevas has joined #openstack-ansible | 07:41 | |
| *** javeriak_ has quit IRC | 08:19 | |
| *** cbaesema_ has left #openstack-ansible | 08:25 | |
| *** shoutm has quit IRC | 08:32 | |
| openstackgerrit | Jean-Philippe Evrard proposed stackforge/os-ansible-deployment: Implementation of keepalived for haproxy https://review.openstack.org/218818 | 08:34 |
|---|---|---|
| *** Mudpuppy has quit IRC | 08:36 | |
| openstackgerrit | Jean-Philippe Evrard proposed stackforge/os-ansible-deployment: [WIP] Implementation of keepalived for haproxyiImplementation of keepalived for haproxy https://review.openstack.org/218818 | 08:37 |
| evrardjp | good morning | 08:37 |
| evrardjp | sorry for this ^(my vim went AWOL) | 08:37 |
| openstackgerrit | Jean-Philippe Evrard proposed stackforge/os-ansible-deployment: [WIP] Implementation of keepalived for haproxy https://review.openstack.org/218818 | 08:38 |
| *** javeriak has joined #openstack-ansible | 08:41 | |
| *** javeriak has quit IRC | 08:47 | |
| hughsaunders | odyssey4me: I may be missing something here, but why do we need a blank line before </virtualhost> ? https://review.openstack.org/#/c/221974/7/playbooks/roles/os_keystone/templates/keystone-httpd.conf.j2,cm | 08:51 |
| *** tiagogomes has joined #openstack-ansible | 08:57 | |
| tiagogomes | hi | 08:57 |
| hughsaunders | HI tiagogomes | 08:58 |
| openstackgerrit | Matt Thompson proposed stackforge/os-ansible-deployment: Add neutron_migrations_facts module https://review.openstack.org/219759 | 09:05 |
| tiagogomes | I am trying to use os-ansible-deployment to install a new stack. I had configured my hosts with some interfaces with IP addresses with a /16 prefix for the different required networks (management, vlanx, storage). However after running the setup-hosts playbook I verified the prefix changed to be /24. This is causing the packets to not be routed to the container properly. Anyone knows what could have caused the change of pref | 09:05 |
| tiagogomes | ix? | 09:05 |
| tiagogomes | On the cidr_networks key of /etc/openstack_deploy/openstack_user_config.yml, I am using a /16 prefix as well | 09:06 |
| andymccr | tiagogomes: you mean the cidr of the interfaces inside the containers is a /24 when it should be a /16? | 09:15 |
| odyssey4me | hughsaunders we don't, but what I have seen sometimes is that the whitespace removal (ie the dash in {%-) ends up removing whitespace including the newline, so the </VirtualHost> tag ends up on the same line as the WSGI... | 09:19 |
| odyssey4me | hughsaunders I was not really in agreement with the first patch, but it was voted through - this is a backport | 09:19 |
| hughsaunders | odyssey4me: I thought it was totally unnecessary so I tested it, but it does fail as described :/ | 09:20 |
| hughsaunders | However the cherry pick id doesn't match the commit it was actually cherry picked from | 09:21 |
| hughsaunders | really unintuitive that the - on the tag effects newlines that are visually outside the tag. | 09:21 |
| odyssey4me | hughsaunders the original patch was done as a patch to master and kilo at the same time - I edited the commit message once the master patch had merged | 09:22 |
| tiagogomes | andyhky, the interfaces outside the containers | 09:22 |
| odyssey4me | did I do a bad edit? | 09:22 |
| openstackgerrit | Jean-Philippe Evrard proposed stackforge/os-ansible-deployment: [WIP] Implementation of keepalived for haproxy https://review.openstack.org/218818 | 09:22 |
| odyssey4me | hughsaunders if you can just fix up the commit message that'd be great - the content of the patch itself is the same | 09:23 |
| *** arbrandes has joined #openstack-ansible | 09:23 | |
| *** arbrandes1 has quit IRC | 09:24 | |
| openstackgerrit | Hugh Saunders proposed stackforge/os-ansible-deployment: Fixes jinja typo in keystone-httpd.conf.j2 https://review.openstack.org/221974 | 09:24 |
| odyssey4me | hughsaunders looks good to me | 09:26 |
| hughsaunders | haha +2 race | 09:27 |
| openstackgerrit | Matt Thompson proposed stackforge/os-ansible-deployment: Add neutron_migrations_facts module https://review.openstack.org/219759 | 09:27 |
| andymccr | tiagogomes: the existing interfaces on the physical hosts aren't managed by os-ansible-deployment at all, so i dont think they could/would've been changed by a setup-hosts run. The only change that happens there would be the lxc-br0 which is added by the run, but that shouldn't determine container connectivity on the other interfaces. | 09:29 |
| tiagogomes | andymccr right, ok thanks. That's what I thought but I wanted to me sure. I was deploying to another host with identical network setup and the prefixes haven't changed. Odd, but restarting the network fixed the issue | 09:31 |
| andymccr | tiagogomes: no problem! glad thats working now | 09:31 |
| openstackgerrit | Jean-Philippe Evrard proposed stackforge/os-ansible-deployment: Implementation of keepalived for haproxy https://review.openstack.org/218818 | 09:32 |
| tiagogomes | I have another question, does the deployment host has to have access to the container management network directly? Or does it ssh first the the container host using the IP in the inventory, and then to the container | 09:32 |
| tiagogomes | It looks like that is trying to SSH to the container directly | 09:33 |
| evrardjp | cloudnull: when you'll be there, could you check the inventory of this is what you expected? : https://review.openstack.org/218818 | 09:33 |
| evrardjp | it works here for me | 09:34 |
| andymccr | tiagogomes: it needs access to the management network directly - if you look in the openstack_inventory.json file, which is the inventory file that is created by taking your user_config.yml and the environment config files, you'll see each container is actually a host. | 09:40 |
| andymccr | there is a file in the scripts directory called "inventory-manage.py" you can run this with a --list and you'll see all the hosts (which includes containers) as well as their ansible_ssh_address | 09:40 |
| andymccr | which is the address ansible will try connect directly to. | 09:41 |
| tiagogomes | bugger, I would be nice to mention that the deployment host has to have access to the management network in the docs | 09:41 |
| tiagogomes | I'll have to change the deployment to one of the target hosts then | 09:41 |
| tiagogomes | thanks andymccr | 09:41 |
| andymccr | tiagogomes: thats a good point - that should be mentioned in the docs/guide. i guess your deployment host would only need access on that network, but none the less it does require access to that network. | 09:42 |
| openstackgerrit | Serge van Ginderachter proposed stackforge/os-ansible-deployment: Allow Horizon setup with external SSL termination https://review.openstack.org/214647 | 09:42 |
| hughsaunders | mattt: your tempest cinder multi backend patch, allows a deployer to test multi backend but without changing anything for the current gate? | 09:54 |
| odyssey4me | hughsaunders yup | 09:58 |
| odyssey4me | it enables alternative gate tests | 09:59 |
| openstackgerrit | git-harry proposed stackforge/os-ansible-deployment: Remove unused Juno-specific groups from inventory https://review.openstack.org/222186 | 10:08 |
| tiagogomes | hi again, anyone knows what is going on here: http://paste.openstack.org/show/456579/ | 10:22 |
| *** javeriak has joined #openstack-ansible | 10:34 | |
| andymccr | tiagogomes: ansible cant reach those 2 memcache containers so it fails out and then has no hosts to run the next task against (because the 2 eligible hosts have already failed) | 10:36 |
| tiagogomes | I got that, but I managed to ssh to the containers without ansible, so I don't don't know why ansible can't | 10:37 |
| andymccr | hmm. | 10:40 |
| andymccr | if you run the ansible command with a -vvv it could give you more of a clue. I have had an issue before where it tries to execute /usr/bin/python and the containers i had were not deployed properly (mostly because i stopped a run midway through). | 10:42 |
| tiagogomes | I don't have a python executable on the container, only python3 | 10:43 |
| andymccr | if you check with -vvv it'll give a better idea, but if there is an issue during the container deployment process that could be a cause | 10:45 |
| tiagogomes | The output with that I posted is already with the verbose option | 10:47 |
| tiagogomes | right, the is a broken symbolic link in the container rootfs python -> /usr/bin/python2.7. The python2.7 file it is not on the container rootfs | 10:51 |
| openstackgerrit | Merged stackforge/os-ansible-deployment: Add profiling for Ansible tasks https://review.openstack.org/222348 | 10:51 |
| tiagogomes | would be worth so run teardown and start all over? Is this a know bug? | 10:52 |
| * tiagogomes notes that there is a task that created the python symbolic link | 10:55 | |
| hughsaunders | tiagogomes: ansible has problems with ssh connections sometimes, and doesn't have a mechanism to retry them until v2 when ssh-retry got merged: https://github.com/ansible/ansible/commit/42467777593e3a4897c86362d3ec9fb09f517862 | 10:59 |
| *** javeriak_ has joined #openstack-ansible | 11:00 | |
| mattt | hughsaunders: yeah, that backend1 stuff we were setting was actually not being used | 11:01 |
| hughsaunders | mattt: are you using those vars in something? | 11:02 |
| odyssey4me | tiagogomes that's odd, the task right before that installs the python2.7 package | 11:02 |
| *** javeriak has quit IRC | 11:02 | |
| mattt | hughsaunders: in a gate? no | 11:03 |
| mattt | hughsaunders: it was a case of removing the pointless entries from tempest.conf, or adding some vars so that functionality can actually be used | 11:03 |
| mattt | hughsaunders: and since we may have deployments w/ lvm and ceph backends, i figured it may get used down the line | 11:03 |
| *** markvoelker has joined #openstack-ansible | 11:04 | |
| hughsaunders | Ah, it looked prep work for something, but it was actually tidy up | 11:04 |
| mattt | hughsaunders: yep, just cleanup, that's why i was on the fence about backporting to kilo | 11:05 |
| *** Mudpuppy has joined #openstack-ansible | 11:08 | |
| *** markvoelker has quit IRC | 11:09 | |
| *** Mudpuppy has quit IRC | 11:12 | |
| *** shoutm has joined #openstack-ansible | 11:13 | |
| openstackgerrit | Matt Thompson proposed stackforge/os-ansible-deployment: Add neutron_migrations_facts module https://review.openstack.org/219759 | 11:13 |
| openstackgerrit | Merged stackforge/os-ansible-deployment: Fixes jinja typo in keystone-httpd.conf.j2 https://review.openstack.org/221974 | 11:14 |
| *** ggillies has quit IRC | 11:23 | |
| *** Mudpuppy has joined #openstack-ansible | 11:26 | |
| hughsaunders | just looking at log slave updates patch for kilo, does that make sense with galera? | 11:28 |
| hughsaunders | If a galera node dies, when its rebuilt, it can be brought back up to speed via galera's SST mechanism and doesn't need local binlogs | 11:29 |
| hughsaunders | So why add load making every node write binlogs for each tx? | 11:30 |
| hughsaunders | ^ mancdaz | 11:31 |
| mancdaz | hughsaunders I keep meaning to follow up with DBAs on why this was recommended | 11:31 |
| mancdaz | this and server_id uniqueness | 11:32 |
| *** Mudpuppy has quit IRC | 11:32 | |
| mancdaz | which is generally not recommended with galera clusters | 11:32 |
| *** Mudpuppy has joined #openstack-ansible | 11:33 | |
| mancdaz | hughsaunders I guess in a case where the entire db is screwed, being able to restore from binlongs up to a point is useful | 11:33 |
| mancdaz | screwed and also replicated in it's screwed state | 11:33 |
| mancdaz | so other nodes are not good donors here | 11:33 |
| hughsaunders | ahh, yes in the raid-is-not-backup sense | 11:34 |
| hughsaunders | however binlogs are not backups either | 11:34 |
| mancdaz | hughsaunders they kind of are a bit | 11:34 |
| hughsaunders | if you happen to catch them at the right time | 11:34 |
| mancdaz | hughsaunders well you can replay up to whenever you like | 11:34 |
| mancdaz | ie before catastrophe | 11:35 |
| hughsaunders | yeah, depending on how long ago it was and how much binlog you store | 11:35 |
| mancdaz | well quite | 11:35 |
| hughsaunders | but I can see there can be a use and that it would be hard to try and reconstruct that from multiple nodes | 11:35 |
| mancdaz | I'm guessing though that a full restore on all nodes because of catastrophe will be done inside 5 days | 11:35 |
| mancdaz | so this is why we need server_id uniques | 11:36 |
| mancdaz | otherwise all updates get logged multiple times | 11:36 |
| mancdaz | else you can't decipher your own updates | 11:36 |
| mancdaz | or something | 11:36 |
| mancdaz | I think possibly I just talked myself into understanding why this change is needed, but also that without the server_id change this could be bad...? | 11:37 |
| mancdaz | hughsaunders walk down that logic path with me | 11:37 |
| hughsaunders | but you wouldn't restore on multiple nodes, you'd restore on one, re initialise it as a new cluster, and let it update the others | 11:39 |
| hughsaunders | then you don't have to sync positions and whatnot, let galera do its thang | 11:39 |
| mancdaz | hughsaunders sure. I don't think that's what I was saying | 11:39 |
| mancdaz | I think the server_id helps in determining which other updates to log/ignore | 11:40 |
| hughsaunders | sounds plausible, don't know though | 11:41 |
| *** woodard has joined #openstack-ansible | 11:42 | |
| *** sdake has joined #openstack-ansible | 11:42 | |
| *** sdake has quit IRC | 11:44 | |
| *** woodard has quit IRC | 11:46 | |
| openstackgerrit | Merged stackforge/os-ansible-deployment: Replaced the copy_update module https://review.openstack.org/216790 | 11:50 |
| hughsaunders | git-harry: didn't know about viewitems, interesting. | 11:58 |
| git-harry | hughsaunders: yeah it's python3 items | 12:01 |
| mgariepy | good morning | 12:02 |
| openstackgerrit | Merged stackforge/os-ansible-deployment: Allow testing of cinder multi backends w/ tempest https://review.openstack.org/222119 | 12:10 |
| *** markvoelker has joined #openstack-ansible | 12:15 | |
| *** kukacz has joined #openstack-ansible | 12:17 | |
| tiagogomes | I think that python2.7 was not installed on the containers because apt-transport-https was not installed in the rootfs, and the repository URLs were using https | 12:28 |
| *** tlian has joined #openstack-ansible | 12:30 | |
| openstackgerrit | Merged stackforge/os-ansible-deployment: Enable log_slave_updates https://review.openstack.org/222198 | 12:39 |
| odyssey4me | mancdaz hughsaunders as I recall the bil_logs are useful when a server has simply been offline for a period of time, not entirely destroyed | 12:49 |
| mancdaz | odyssey4me kind of, except when you have galera you don't need them | 12:50 |
| mancdaz | in a normal master/slave replication env, sure | 12:50 |
| hughsaunders | odyssey4me: very much so in the mysql replication case, in the g.... what mancdaz said | 12:50 |
| odyssey4me | tiagogomes both of those packages are pre-installed into the container image that is downloaded, so how is it that your container image doesn't have those? | 12:50 |
| *** sigmavirus24_awa is now known as sigmavirus24 | 12:50 | |
| odyssey4me | mancdaz hughsaunders I'm just giving you the justification that was given me. :p | 12:51 |
| *** tlian has quit IRC | 12:51 | |
| mancdaz | odyssey4me in this case, I think we're looking at using them to restore n entire cluster from "last nightly full + latest binlogs" | 12:51 |
| odyssey4me | mancdaz that seems odd though, why bother restore when you can just add the node to the cluster? | 12:52 |
| *** javeriak_ has quit IRC | 12:54 | |
| *** tlian has joined #openstack-ansible | 12:56 | |
| *** javeriak has joined #openstack-ansible | 12:59 | |
| mancdaz | enitire dead cluster data...need to start from last good point in time | 13:04 |
| *** cloudtrainme has joined #openstack-ansible | 13:05 | |
| *** cloudtrainme has quit IRC | 13:05 | |
| openstackgerrit | Jesse Pretorius proposed stackforge/os-ansible-deployment: Adds the config_template to OpenStack Services https://review.openstack.org/217030 | 13:09 |
| odyssey4me | mattt in light of our earlier discussion about independent role repositories, can you please review https://review.openstack.org/213779 ? | 13:10 |
| kukacz | hi, I've just recently noticed this project and on first look I'm excited of it's scope | 13:10 |
| kukacz | I wonder what is the maturity of the project, is it something being used in production already? | 13:11 |
| odyssey4me | kukacz yes, absolutely | 13:12 |
| mattt | odyssey4me: yeah sure | 13:12 |
| *** hrvojem has left #openstack-ansible | 13:12 | |
| *** javeriak_ has joined #openstack-ansible | 13:12 | |
| kukacz | oddysey4me, mattt: that's so nice to hear :-) | 13:12 |
| odyssey4me | it's been used for Rackspace Private Cloud in production for over a year now (since Icehouse) and has been growing and maturing over that time. In the Liberty cycle we have had several other deployers adopt it to run Kilo. | 13:12 |
| *** javeria__ has joined #openstack-ansible | 13:13 | |
| *** javeria__ has quit IRC | 13:13 | |
| mgariepy | out of curiosity how big private cloud are at rackspace ? | 13:14 |
| kukacz | when considering deploying it for own production - do I have to bother with some Rackspace specifics? | 13:14 |
| odyssey4me | kukacz our docs are a little dated and need some attention, but if you're able to explore the ansible defaults files you can often find the right changes - we'd love to get more doc update patches! | 13:14 |
| *** logan2 has quit IRC | 13:15 | |
| odyssey4me | kukacz There are no Rackspace specifics - those were pulled out for Kilo and RPC now consumes openstack-ansible as an upstream project. Anything RPC specific is handled in rpc-openstack and not in openstack-ansible | 13:15 |
| *** javeriak has quit IRC | 13:15 | |
| kukacz | odyssey4me: if I consider using OSAD in place of eg. RedHat's OSP - is it something one would seriously recommend? (ignoring the support) | 13:15 |
| *** javeriak has joined #openstack-ansible | 13:15 | |
| openstackgerrit | Merged stackforge/os-ansible-deployment: Ensure cinder-backup-related variables are defined https://review.openstack.org/221705 | 13:16 |
| openstackgerrit | Merged stackforge/os-ansible-deployment: Be more clear about used_ips, mostly in the example file https://review.openstack.org/222121 | 13:16 |
| odyssey4me | kukacz take note that the repository is being renamed to openstack/openstack-ansible tonight - it will no longer be a stackforge project | 13:16 |
| kukacz | odyssey4me:yeah, I've noticed. congratulations to such a move! | 13:16 |
| odyssey4me | kukacz if you're happy with using Ubuntu instead of RHEL then yes | 13:16 |
| *** javeriak_ has quit IRC | 13:16 | |
| odyssey4me | if not, then we have some work being done in the Mitaka cycle to implement broader OS support and we'd love more hands and eyes to be involved | 13:17 |
| odyssey4me | mgariepy big in terms of people, or in terms of deployments? | 13:17 |
| mgariepy | in size of the cluster | 13:18 |
| kukacz | odyssey4me: so similarly to RAX, one can thing of consuming OSAD as an upstream source for own private cloud solution? | 13:18 |
| odyssey4me | mgariepy RPC has builds that vary in size from small (ie 10 compute nodes or less) to large (50+ nodes) | 13:18 |
| *** javeriak has quit IRC | 13:18 | |
| odyssey4me | mgariepy some builds are in RAX DC's and others are in Customer DC's | 13:18 |
| mgariepy | ok | 13:18 |
| odyssey4me | kukacz yes, absolutely - you can learn from how RAX does it by looking at https://github.com/rcbops/rpc-openstack/ or you can do it completely your own way | 13:19 |
| odyssey4me | kukacz we're also doing work in the Mitaka cycle to make it easier to consume openstack-ansible in more flexible ways 0 see the spec in https://review.openstack.org/213779 | 13:20 |
| *** javeriak has joined #openstack-ansible | 13:20 | |
| odyssey4me | mgariepy I don't know if you've seen the news about the large clusters being built for the OpenStack Innovation Centre? That's 2 x 1000 node clusters, all being built with openstack-ansible | 13:21 |
| kukacz | odyssey4me: amazing. I'll check those links | 13:21 |
| mgariepy | nop i havent seen it. | 13:21 |
| *** javeriak has quit IRC | 13:21 | |
| odyssey4me | the big clusters are specifically being built for larger scale testing, cross DC testing, federation, global clustering, etc | 13:21 |
| odyssey4me | mgariepy http://www.rackspace.com/blog/newsarticles/rackspace-collaborates-with-intel-to-accelerate-openstack-enterprise-feature-development-and-adoption/ | 13:22 |
| *** KLevenstein has joined #openstack-ansible | 13:24 | |
| kukacz | odyssey4me: I start considering using it for our company's private cloud offering. we're just beginning with openstack, consuming distro and support from third parties, but in short term window we'll need to move to customized solutions offered to our customers | 13:24 |
| kukacz | odyssey4me: .... doing those customizations in house. OSAD seems to be a great basement for this | 13:25 |
| odyssey4me | kukacz well, what we offer is really flexible - and ansible is easy to understand and quick to learn and work with, so it makes it really easy to get involved in doing improvements in the project which filter down into your environment after going through feedback, iteration and testing | 13:26 |
| kukacz | will probably spend some time this weekend running an OSAD eval deployment on top our pilot OS env | 13:26 |
| odyssey4me | we understand that some deployers prefer to fork for internal needs, but we do try and encourage upstream commits of whatever's been forked so that we can all help you make it better and everyone can benefit from it | 13:27 |
| kukacz | odyssey4me: I'm not familiar with Ansible but this is a good reason to start learning it | 13:27 |
| *** mrstanwell has joined #openstack-ansible | 13:28 | |
| kukacz | odyssey4me: I understand. I mostly thing of those customizations like doing something required by our customers (integrate ELK stack or whatever....) while contributing that back to the project. that should not be a problem | 13:28 |
| odyssey4me | kukacz if you look carefully, rpc-openstack already has an ELK stack :) | 13:29 |
| odyssey4me | you can always borrow from there, considering that the source is open :) | 13:29 |
| kukacz | odyssey4me: OK, will have to grab for some other task :-) | 13:29 |
| kukacz | odyssey4me: I've noticed a ceph support too | 13:30 |
| mrstanwell | I tried making that ELK go in osad kilo, but failed. | 13:30 |
| odyssey4me | you could also join evrardjp in the call to move those roles into the openstack-ansible umbrella so that we can all participate in making it better :) | 13:30 |
| odyssey4me | kukacz yup, ceph support was merged a month or two ago and has been baking for a few weeks | 13:30 |
| odyssey4me | mrstanwell oh? I'm pretty sure that it works as the team working on Ceph in rpc-openstack included some new filters for Ceph... so I'm fairly certain that it works - there may just be some implementation details that need clarification | 13:31 |
| kukacz | odyssey4me: (thinkin of what else we need in the core) - any experience with OpenContrail integrations? | 13:31 |
| kukacz | I mean with OSAD specifically | 13:32 |
| tiagogomes | odyssey4me I don't know | 13:33 |
| odyssey4me | kukacz nope, we've had no-one work with that - some work is being done for plumgrid enablement - not sure if that helps | 13:33 |
| *** jroll is now known as jtroll | 13:33 | |
| *** jtroll is now known as Guest13730 | 13:34 | |
| *** Guest13730 is now known as jroll | 13:34 | |
| kukacz | odyssey4me: nevermind, still I'm excited :-) | 13:34 |
| kukacz | odyssey4me: thanks for the chat! I'll prepare an environment to test it out myself | 13:35 |
| mrstanwell | odyssey4me: could be. in my case, it looked like the cloud-init scripts defined in the heat template weren't being applied when the master node was being created. could never figure out why. chalked it up to some moving part having revved since then or something. but, I'd never done anything with heat before. so, it's also very possibly a n00b error. | 13:36 |
| mrstanwell | got distracted with ipv6. | 13:37 |
| mrstanwell | and, after all, the rsyslog container in osad gathers all the logs up sooooo nicely!! | 13:37 |
| odyssey4me | mrstanwell I'm confused - was it the heat template that didn't work, or was it the ELK stack that didn't work? | 13:37 |
| mrstanwell | odyssey4me: sorry. it was the heat template. | 13:38 |
| mrstanwell | (the rpc-openstack one) | 13:38 |
| odyssey4me | mrstanwell which heat template did you use? mattt is the primary maintainer of it so he may help you make it go | 13:38 |
| mrstanwell | odyssey4me: it was from the RPC-Heat-ELK project on github. at commit 6f4d52b63d, looks like. | 13:41 |
| odyssey4me | mrstanwell ah, that's meant to be used from inside an RPC installation - it's part of a Horizon solutions tab | 13:42 |
| odyssey4me | it doesn't deploy the same thing as is present in the ELK roles in rpc-openstack (which is developed for an openstack back-end) | 13:43 |
| *** logan2 has joined #openstack-ansible | 13:43 | |
| odyssey4me | mattt can you backport https://review.openstack.org/221705 please? | 13:43 |
| openstackgerrit | Matt Thompson proposed stackforge/os-ansible-deployment: Ensure cinder-backup-related variables are defined https://review.openstack.org/222610 | 13:44 |
| mrstanwell | odyssey4me: aaah. ok. thanks for clarifying! | 13:46 |
| odyssey4me | mattt perhaps it's worth adding a subsequent patch to clean-up the old users - re: https://review.openstack.org/202822 | 13:48 |
| mattt | odyssey4me: are you asking me to do the follow-up patch? | 13:51 |
| mattt | if so, i'd rather it just get done in that review :P | 13:51 |
| evrardjp | about opencontrail: At some point we'll get there | 13:55 |
| evrardjp | I mean: opencontrail is something we think of, where I work ;) | 13:56 |
| kukacz | evrardjp: same with us. | 13:56 |
| evrardjp | so if we need to integrate it with openstack, the odds are high that we are going to use ansible to deploy it, so why not integrate it with osad (if relevant) or in a close repo | 13:57 |
| kukacz | evrardjp: the network guys find it the most suitable and flexible enough for their needs | 13:57 |
| evrardjp | it makes sense ;) | 13:57 |
| openstackgerrit | Major Hayden proposed stackforge/os-ansible-deployment-specs: Adding security hardening spec https://review.openstack.org/222619 | 13:57 |
| evrardjp | working with juniper hardware? | 13:57 |
| evrardjp | not mandatory, but still interesting to know | 13:58 |
| kukacz | partially, we are a mix of cisco/juniper | 13:58 |
| kukacz | but thei favor contrail for the real features, without any preference regarding the hardware vendor | 13:59 |
| *** woodard has joined #openstack-ansible | 13:59 | |
| openstackgerrit | Major Hayden proposed stackforge/os-ansible-deployment-specs: Adding security hardening spec https://review.openstack.org/222619 | 13:59 |
| kukacz | we need to connect various external networks and extend them into openstack as tenant networks. contrail has good ways how to ensure this | 14:00 |
| kukacz | ... among others | 14:00 |
| evrardjp | same use case here | 14:00 |
| kukacz | so the network guys appreciate MPBGP, EVPN and these protocol options | 14:01 |
| *** cloudtrainme has joined #openstack-ansible | 14:03 | |
| kukacz | is there any fast way to deploy a small environment for evaluating OSAD deployment? | 14:06 |
| kukacz | I thought I would start few openstack instances, to deploy control node cluster and a pair of compute nodes | 14:07 |
| kukacz | but seems that I'll spend some time configuring several networks on those instances first | 14:07 |
| evrardjp | I'm afraid I don't have time to help you right now :/ I must go, maybe on Monday if everything goes fine | 14:08 |
| evrardjp | I'm off! | 14:08 |
| kukacz | evrardjp: no problem, enjoy the weekend! | 14:08 |
| kukacz | anybody else? I wonder what people do to repeatedly build and destroy OSAD environment for dev/test purposes? | 14:09 |
| mrstanwell | you can build a basic all-in-one pretty easily. I've got a multi-node that was pretty easy to deploy once I worked out how to set up the network interfaces. the teardown script wipes the containers out, and I've generally been able to just redeploy using osad. | 14:11 |
| gparaskevas | tear down script does the work | 14:13 |
| gparaskevas | and then redeploy | 14:13 |
| gparaskevas | i believe allthough i am not sure that it also deletes openstack_deploy folder | 14:13 |
| kukacz | but first I probably need to create the required bridges and interfaces on my instances, right? | 14:16 |
| mrstanwell | kukacz: the bootstrap-aio script will set up your bridges. then you can just run the setup-everything playbook. | 14:17 |
| mrstanwell | wait. you said on instances...? | 14:17 |
| openstackgerrit | Merged stackforge/os-ansible-deployment: Used named veth pairs that match container https://review.openstack.org/222295 | 14:17 |
| kukacz | mrstanwell: yes, I would like to deploy multinode, using existing openstack environment | 14:18 |
| kukacz | mrstanwell:I thought I will run some 5 instances and run OSAD on top of them | 14:18 |
| *** javeriak has joined #openstack-ansible | 14:19 | |
| palendae | kukacz: That's do-able, we run multinode on top of rackspace public cloud for testing. You need to define the bridges and interfaces, though | 14:19 |
| kukacz | mrstanwell:the bootstrap-aio - I assume that works on single node only, does it? | 14:19 |
| mrstanwell | kukacz: Ah. I've never deployed ostack on ostack. But, yeah, for an all-in-one, the bootstrap-aio will set up the bridges. for a multi-node, you need to configure them yourself, then update the user-settings.yml accordingly. | 14:19 |
| palendae | kukacz: Yes, bootstrap-aio.sh makes the assumption that is one node only | 14:19 |
| palendae | kukacz: For reference, this is what we use to deploy across the rackspace public cloud for multinode: https://github.com/rcbops/rpc-heat | 14:20 |
| kukacz | palendae: are there any heat teplates or ansible scripts available to help with the prepare phase? | 14:20 |
| kukacz | palendae: thanks, will look at it | 14:21 |
| palendae | Within openstack-ansible/OSAD, no, there's no host configuration present right now | 14:21 |
| kukacz | and just to confirm I got it right - I need to create 4 interfaces + bridges | 14:22 |
| kukacz | "br-mgmt" "br-storage" "br-vxlan" and "br-vlan" | 14:22 |
| *** spotz_zzz is now known as spotz | 14:23 | |
| *** javeriak has quit IRC | 14:23 | |
| *** javeriak has joined #openstack-ansible | 14:24 | |
| *** woodard has quit IRC | 14:24 | |
| mrstanwell | kukacz: sounds right. you can use the examples in os-ansible-deployment/etc/network/interfaces.d. Frankly, my interface files for multinode don't look much different from aio, just with different addressing. and bonding, if you're into that sort of thing. | 14:27 |
| kukacz | mrstanwell:yeah. now I'm checking that example interfaces.d config and completing my openstack_user_config.yml | 14:30 |
| kukacz | mrstanwell: I noticed the example uses linuxbridge - is that a default or only integrated networking option? | 14:31 |
| palendae | kukacz: Currently the only one | 14:32 |
| kukacz | ok | 14:32 |
| *** javeriak_ has joined #openstack-ansible | 14:34 | |
| *** javeriak has quit IRC | 14:34 | |
| mrstanwell | kukacz: and, technically, you don't need 4 interfaces. I've done it with just one, using vlans. (intel nucs only have one wired ifc...) | 14:35 |
| *** phalmos has joined #openstack-ansible | 14:35 | |
| palendae | mrstanwell: I've avoided getting a NUC in my homelab for that reason, but sure would reduce the amount of NIC packaging I have to deal with | 14:36 |
| mrstanwell | palendae: I've got some startech usb enet adapters that worked out-of-the-box w/ ubuntu. but they do drive the cpu load a bit. and for some reason they get pretty warm... | 14:37 |
| kukacz | mrstanwell: I see. I thing I'll create 4 OS non-routed networks so those instances will have simple NIC/bridge pairs without vlans in config - to make my config task easier | 14:37 |
| palendae | mrstanwell == Prof Denton? | 14:37 |
| mrstanwell | nope | 14:37 |
| mrstanwell | kukacz: yeah, if you're doing it ostack already, no point in complicating your life with vlans if you don't have to. extra interfaces are cheap! ;-) | 14:39 |
| kukacz | mrstanwell: but in the end - I need to always have those 4 bridges in place to satisfy the default/example config, right? | 14:39 |
| mrstanwell | kukacz: I believe so, yes. or depending on whatever you set up in openstack_user_config.yml. | 14:40 |
| kukacz | mrstanwell: I will leave it as much default as I can to make it work for the first tries | 14:41 |
| kukacz | as I'm new both to OSAD and Ansible | 14:41 |
| mrstanwell | palendae: so far, nucs make a great little platform for osad. if you don't mind the vlan config hassle and don't need serious performance, you can get by with the single nic. or go usb. | 14:42 |
| mrstanwell | kukacz: I'm not much farther ahead of you, frankly. ;-) But osad is hands-down the easiest ostack install I've ever done. | 14:43 |
| mrstanwell | kukacz: I went from git clone to running aio in less than an hour. outstanding! | 14:44 |
| kukacz | mrstanwell: good to hear. as I need to spare some time of the weekend for kids too :-) | 14:44 |
| kukacz | mrstanwell: congrats, I won't beat you. already spent 2 hours chatting here and checking the configs | 14:45 |
| mrstanwell | kukacz: well, multinode took me quite a bit longer. had to figure out what aio had been doing for me, networkwise. ;-) | 14:46 |
| kukacz | network is always the pain. we should get rid off it | 14:47 |
| openstackgerrit | Merged stackforge/os-ansible-deployment: Enabling log compression by default https://review.openstack.org/221924 | 14:49 |
| mhayden | should i cherry pick this one into kilo ^^ ? | 14:51 |
| *** KLevenstein has quit IRC | 14:53 | |
| *** KLevenstein has joined #openstack-ansible | 14:56 | |
| odyssey4me | mhayden yeah, I think so - it looks like a good candidate for a backport | 14:56 |
| mhayden | alrighty | 14:56 |
| odyssey4me | kukacz yes, openstack-ansible assumes that the hosts are setup already with the networking and disks already configured in the way you want to consume them | 14:56 |
| openstackgerrit | Major Hayden proposed stackforge/os-ansible-deployment: Enabling log compression by default https://review.openstack.org/222646 | 14:56 |
| * mhayden did a thing | 14:57 | |
| odyssey4me | kukacz it may be helpful to you to try out some basic ansible work on the side to understand how it works - it really helps you navigate the openstack-ansible project when you understand ansible a little better | 14:58 |
| cloudnull | morning. | 14:58 |
| cloudnull | finally at a compute.r | 14:58 |
| kukacz | odyssey4me: thanks. actually, I've just started trying to use Ansible to configure the networks on my hosts | 15:00 |
| tiagogomes | anyone has seen this http://paste.openstack.org/show/456863/ ? | 15:00 |
| odyssey4me | o/ cloudnull :) | 15:00 |
| odyssey4me | tiagogomes yes, that may have come up if you're using yesterda | 15:00 |
| odyssey4me | the last tag of kilo? there was an upstream issue - we'll likely bump the upstream sha today for both juno and kilo to take care of that and rev the tag | 15:01 |
| odyssey4me | cloudnull shall I go ahead with that? | 15:01 |
| cloudnull | typie typie, make it go ! =) | 15:02 |
| odyssey4me | cloudnull on it :) | 15:02 |
| tiagogomes | oh, I was using master | 15:02 |
| odyssey4me | tiagogomes haha, living dangerously :p | 15:02 |
| cloudnull | DO IT LIVE! | 15:03 |
| odyssey4me | tiagogomes I'm not sure if you're aware of this, but master tracks upstream master - so you're deploying Liberty-3 code with that. :) | 15:03 |
| tiagogomes | I had to clone again the repo when moving from separate deployment host to deploy from a target host and I forgot to checkout the branch | 15:03 |
| odyssey4me | it may break at any time | 15:03 |
| *** agireud has joined #openstack-ansible | 15:03 | |
| tiagogomes | right, I'll try with Kilo | 15:04 |
| odyssey4me | tiagogomes hang ten with that - let us get that upstream sha update done | 15:04 |
| *** phalmos has quit IRC | 15:04 | |
| *** agireud has quit IRC | 15:08 | |
| openstackgerrit | Jesse Pretorius proposed stackforge/os-ansible-deployment: Update kilo to resolve PyECLib compile issue https://review.openstack.org/222656 | 15:11 |
| *** cloudtra_ has joined #openstack-ansible | 15:12 | |
| *** cloudtrainme has quit IRC | 15:13 | |
| openstackgerrit | Jesse Pretorius proposed stackforge/os-ansible-deployment: Update juno to resolve PyECLib compile issue https://review.openstack.org/222666 | 15:21 |
| openstackgerrit | Merged stackforge/os-ansible-deployment: Adds the config_template to OpenStack Services https://review.openstack.org/217030 | 15:25 |
| *** jwagner_away is now known as jwagner | 15:26 | |
| *** phalmos has joined #openstack-ansible | 15:26 | |
| *** javeriak_ has quit IRC | 15:40 | |
| *** javeriak has joined #openstack-ansible | 15:41 | |
| *** gparaskevas has quit IRC | 15:43 | |
| *** javeriak_ has joined #openstack-ansible | 15:45 | |
| *** javeriak has quit IRC | 15:47 | |
| openstackgerrit | Matt Thompson proposed stackforge/os-ansible-deployment: Add neutron_migrations_facts module https://review.openstack.org/219759 | 15:56 |
| *** alop has joined #openstack-ansible | 15:56 | |
| *** javeriak has joined #openstack-ansible | 16:05 | |
| *** devlaps has joined #openstack-ansible | 16:06 | |
| *** daneyon has joined #openstack-ansible | 16:06 | |
| *** javeriak_ has quit IRC | 16:06 | |
| openstackgerrit | Kevin Carter proposed stackforge/os-ansible-deployment: Compartmentalizing RabbitMQ https://review.openstack.org/202822 | 16:09 |
| *** KLevenstein has quit IRC | 16:10 | |
| *** KLevenstein_ has joined #openstack-ansible | 16:10 | |
| cloudnull | mattt: ^^ updated that patch per your review. that removes the "openstack" rabbitmq user | 16:11 |
| *** phalmos has quit IRC | 16:13 | |
| *** phalmos has joined #openstack-ansible | 16:15 | |
| mattt | cloudnull: brilliant, thanks kevin | 16:17 |
| *** woodard has joined #openstack-ansible | 16:17 | |
| cloudnull | thank you sir, that was a good catch | 16:17 |
| *** javeriak_ has joined #openstack-ansible | 16:17 | |
| mattt | cloudnull: we need to write an upgrade task to remove it on existing deploys, palendae had a review in flight to clean up the upgrade script ... any progress on that one palendae so we can start adding to it? | 16:18 |
| mattt | but i have to bounce, i'll be back online this evening | 16:19 |
| * mattt is afk | 16:19 | |
| cloudnull | kk take care mattt | 16:19 |
| *** javeriak has quit IRC | 16:19 | |
| palendae | mattt: Nope, still fixing Juno -> kilo upgrades | 16:20 |
| openstackgerrit | Merged stackforge/os-ansible-deployment: Added post and pre hook script for veth cleanup https://review.openstack.org/220342 | 16:21 |
| odyssey4me | palendae do you mind if I quickly fix the nit on that patch so that we can get on with building upgrade tasks into a fresh version? | 16:22 |
| palendae | odyssey4me: Nope - feel free to take it | 16:22 |
| *** javeriak has joined #openstack-ansible | 16:22 | |
| *** cloudtra_ has quit IRC | 16:22 | |
| *** javeriak_ has quit IRC | 16:23 | |
| *** cloudtrainme has joined #openstack-ansible | 16:24 | |
| openstackgerrit | Jesse Pretorius proposed stackforge/os-ansible-deployment: Remove Juno to Kilo logic in upgrade script https://review.openstack.org/215291 | 16:25 |
| openstackgerrit | Kevin Carter proposed stackforge/os-ansible-deployment: Added post and pre hook script for veth cleanup https://review.openstack.org/222695 | 16:36 |
| odyssey4me | cloudnull happy to make https://review.openstack.org/215291 go? | 16:38 |
| cloudnull | ha! already did it | 16:38 |
| *** sdake has joined #openstack-ansible | 16:51 | |
| *** jwagner is now known as jwagner_away | 16:56 | |
| *** sdake_ has joined #openstack-ansible | 16:57 | |
| *** woodard has quit IRC | 16:58 | |
| *** javeriak_ has joined #openstack-ansible | 16:59 | |
| *** sdake has quit IRC | 17:01 | |
| *** javeriak has quit IRC | 17:01 | |
| openstackgerrit | Jesse Pretorius proposed stackforge/os-ansible-deployment-specs: Add spec for Gate Split https://review.openstack.org/221009 | 17:05 |
| *** javeriak has joined #openstack-ansible | 17:05 | |
| *** javeriak_ has quit IRC | 17:06 | |
| *** woodard has joined #openstack-ansible | 17:08 | |
| odyssey4me | hughsaunders what do you think about skipping the submodule thing and simply switching over to using the role requirements and a sha reference? | 17:08 |
| odyssey4me | as I recall cloudnull favored that approach | 17:08 |
| *** sdake_ has quit IRC | 17:10 | |
| odyssey4me | cloudnull mattt hughsaunders can we make these two go please? https://review.openstack.org/222666 | 17:15 |
| odyssey4me | https://review.openstack.org/222656 | 17:15 |
| odyssey4me | d34dh0r53 if you're around ^ | 17:18 |
| *** sdake has joined #openstack-ansible | 17:19 | |
| *** phalmos has quit IRC | 17:30 | |
| *** KLevenstein_ has quit IRC | 17:32 | |
| *** javeriak_ has joined #openstack-ansible | 17:36 | |
| *** javeriak has quit IRC | 17:38 | |
| *** jwagner_away is now known as jwagner | 17:39 | |
| *** woodard has quit IRC | 17:43 | |
| *** javeriak has joined #openstack-ansible | 17:44 | |
| *** javeriak_ has quit IRC | 17:47 | |
| *** abitha has joined #openstack-ansible | 17:47 | |
| d34dh0r53 | odyssey4me: looking at it now | 17:49 |
| *** KLevenstein has joined #openstack-ansible | 17:55 | |
| *** mgariepy has quit IRC | 18:03 | |
| *** mgariepy has joined #openstack-ansible | 18:06 | |
| *** tiagogomes has quit IRC | 18:23 | |
| *** phalmos has joined #openstack-ansible | 18:35 | |
| *** javeriak has quit IRC | 18:37 | |
| cloudnull | odyssey4me: I do favor the approach of skipping submodules and using the requirement files. | 19:07 |
| *** lkoranda_ has joined #openstack-ansible | 19:07 | |
| *** sdake_ has joined #openstack-ansible | 19:09 | |
| palendae | ^ +1 | 19:10 |
| *** lkoranda has quit IRC | 19:11 | |
| *** lkoranda_ is now known as lkoranda | 19:11 | |
| odyssey4me | cloudnull palendae :) if you could put that into the review it'd be awesome :) | 19:13 |
| *** sdake has quit IRC | 19:13 | |
| palendae | odyssey4me: Which spec was that? The gate spliit one? | 19:14 |
| odyssey4me | palendae yep | 19:14 |
| odyssey4me | palendae cloudnull https://review.openstack.org/221009 | 19:14 |
| cloudnull | odyssey4me: isnt that covered here https://review.openstack.org/#/c/213779/ | 19:16 |
| odyssey4me | palendae cloudnull whoops, you're right | 19:17 |
| palendae | I was searching the page for submodule, wasn't seeing it | 19:17 |
| odyssey4me | I clearly need to have my first sip of scotch so that I can think straight. :p | 19:17 |
| cloudnull | Patch set 3: Line 166: "I don't think we need to or should do the submoduling for anything that is Ansible galaxy compatible. When the role is moved to the external repo we should update our role-requirements.yml file with the data required to make it go using our scripts." | 19:18 |
| odyssey4me | cloudnull ah yes, my bad - youdunnit already | 19:18 |
| cloudnull | but easy to miss , my review was like a stream of consciousness | 19:19 |
| odyssey4me | I am considering taking that whole piece out and just leaving the line that says it should be blueprinted and spec'd per role, and leave it at that | 19:19 |
| cloudnull | yea that makes sense | 19:19 |
| odyssey4me | the method for splitting roles out is contentious and distracting from the main purpose of the spec | 19:19 |
| odyssey4me | which is to allow other repositories for roles in our umbrella | 19:20 |
| *** jmckind has joined #openstack-ansible | 19:20 | |
| *** sdake_ is now known as sdake | 19:20 | |
| cloudnull | as things move out it may be necessary to submodule something like the shared libs or whatever. so it makes sense to have the decision made within the role migration spec. | 19:20 |
| odyssey4me | cloudnull sure, that's under playbook/role impact | 19:21 |
| *** jwagner is now known as jwagner_away | 19:22 | |
| *** javeriak has joined #openstack-ansible | 19:26 | |
| openstackgerrit | Jesse Pretorius proposed stackforge/os-ansible-deployment-specs: Add spec for Independent Role Repositories https://review.openstack.org/213779 | 19:27 |
| *** jwagner_away is now known as jwagner | 19:28 | |
| *** lkoranda has quit IRC | 19:34 | |
| *** lkoranda has joined #openstack-ansible | 19:37 | |
| *** shoutm has quit IRC | 19:46 | |
| *** sdake_ has joined #openstack-ansible | 19:46 | |
| *** sdake has quit IRC | 19:50 | |
| *** javeriak has quit IRC | 19:50 | |
| openstackgerrit | Merged stackforge/os-ansible-deployment: Update kilo to resolve PyECLib compile issue https://review.openstack.org/222656 | 20:06 |
| mhayden | when i finish a draft of a spec, do i flip the blueprint to review or pending approval? | 20:15 |
| mhayden | or neither? | 20:15 |
| *** jmckind has quit IRC | 20:18 | |
| cloudnull | yes | 20:21 |
| cloudnull | :) | 20:21 |
| spotz | heheh | 20:21 |
| cloudnull | I'd flip it to pending approval. it cant hurt | 20:22 |
| openstackgerrit | Jesse Pretorius proposed stackforge/os-ansible-deployment: Remove Juno to Kilo logic in upgrade script https://review.openstack.org/215291 | 20:26 |
| mhayden | cloudnull: thanks | 20:26 |
| odyssey4me | mhayden I'd suggest review | 20:29 |
| odyssey4me | it's likely to go through a few iterations before being accepted, so until you're fairly certain that it's almost done it should be in review. | 20:29 |
| odyssey4me | changing those values is supposed to be a PTL activity though :p | 20:30 |
| *** jmckind has joined #openstack-ansible | 20:30 | |
| *** jmckind has quit IRC | 20:31 | |
| *** jmckind has joined #openstack-ansible | 20:32 | |
| *** lkoranda has quit IRC | 20:37 | |
| *** lkoranda has joined #openstack-ansible | 20:42 | |
| *** phalmos has quit IRC | 20:43 | |
| openstackgerrit | Merged stackforge/os-ansible-deployment: Update juno to resolve PyECLib compile issue https://review.openstack.org/222666 | 20:54 |
| *** jmckind has quit IRC | 21:00 | |
| *** pradk has quit IRC | 21:10 | |
| *** alop has quit IRC | 21:32 | |
| *** alop has joined #openstack-ansible | 21:38 | |
| *** jwagner is now known as jwagner_away | 21:47 | |
| odyssey4me | cloudnull sigmavirus24 d34dh0r53 hughsaunders mattt I would recommend holding off submitting any new reviews or doing any +w on patches from now on. | 21:55 |
| odyssey4me | we're close to the rename timeframe, so I think it'd be prudent to hold off further changes | 21:55 |
| cloudnull | ++ | 21:55 |
| prometheanfire | :D | 21:56 |
| sigmavirus24 | thanks odyssey4me | 21:56 |
| odyssey4me | I think it's safe to quickly do a juno/kilo tag though :) | 21:57 |
| Apsu | Oh man, renamery ahoy | 21:57 |
| odyssey4me | let me get on that | 21:57 |
| Apsu | git clone while you can! | 21:57 |
| *** daneyon has quit IRC | 22:01 | |
| d34dh0r53 | odyssey4me: tys | 22:09 |
| *** woodard has joined #openstack-ansible | 22:10 | |
| *** kukacz has quit IRC | 22:13 | |
| *** woodard has quit IRC | 22:18 | |
| odyssey4me | fyi everyone 11.2.2 has been released | 22:19 |
| -cloudnull- PTL nomination for the upcoming cycle have started flowing into the mailing list. If you're interested in the position please submit your candidacy directly to openstack-dev@lists.openstack.org list. | 22:23 | |
| cloudnull | yay! 11.2.2 thanks odyssey4me! | 22:24 |
| odyssey4me | cloudnull I do find the whole ML candidacy a little odd as it seems it's supposed to be done differently this round: http://lists.openstack.org/pipermail/openstack-dev/2015-September/074171.html | 22:25 |
| cloudnull | yea its a bit odd | 22:25 |
| odyssey4me | I think that people may be misreading, or not reading. ;) | 22:26 |
| cloudnull | it seems that they're following the old process | 22:26 |
| cloudnull | but its also likely that they've sent the message to the ML | 22:27 |
| mattt | so who's going for PTL? | 22:27 |
| cloudnull | and the file to the repo | 22:27 |
| cloudnull | mattt: you said it last | 22:27 |
| cloudnull | :) | 22:27 |
| odyssey4me | lol, I'm happy to admit that I've put my hat into the ring | 22:27 |
| odyssey4me | I do still need to write it up though | 22:27 |
| * odyssey4me watches the clock... 2 mins | 22:29 | |
| palendae | They couldn't possibly run late :) | 22:29 |
| odyssey4me | fyi http://lists.openstack.org/pipermail/openstack-dev/2015-September/074235.html | 22:30 |
| -openstackstatus- NOTICE: 30 minute warning, Gerrit will be offline from 23:00 to 23:30 UTC while some projects are renamed http://lists.openstack.org/pipermail/openstack-dev/2015-September/074235.html | 22:30 | |
| dolphm | odyssey4me: i assume people still want to campaign per usual | 22:31 |
| dolphm | odyssey4me: "vote for me!" is never going away lol | 22:31 |
| odyssey4me | haha dolphm is that am american thing? :p | 22:31 |
| odyssey4me | haha, ok so my timing was out - apparently 23:00 UTC in in 30 mins :p | 22:32 |
| odyssey4me | well, 28 mins | 22:32 |
| mrstanwell | odyssey4me: the American thing is to screw campaigning, and just buy what you need. | 22:33 |
| odyssey4me | best I refresh my scotch :) | 22:33 |
| dolphm | odyssey4me: how many dollars must i give you for your vote, dear sir? | 22:33 |
| odyssey4me | dolphm you're contesting stevemar ? | 22:33 |
| dolphm | odyssey4me: here, just take this blank check, i don't have time to talk petty matters with my constituency | 22:34 |
| dolphm | odyssey4me: no, just being american | 22:34 |
| odyssey4me | dolphm I'll PM you my postal address. Money orders welcome. :p | 22:35 |
| *** spotz is now known as spotz_zzz | 22:38 | |
| cloudnull | confirmed people are campaigning :) https://review.openstack.org/#/q/status:open+project:openstack/election+branch:master,n,z -- Vote for Pedro! | 22:40 |
| odyssey4me | heh, how about you +2 your own patch? https://review.openstack.org/222654 | 22:40 |
| cloudnull | thats how you do it ! | 22:41 |
| palendae | cloudnull: In Juno, do you know if I can tell Glance to use a swift proxy URL that's different from the VIP's? | 22:41 |
| cloudnull | i want to say no. but let me go look | 22:42 |
| palendae | Crap | 22:42 |
| odyssey4me | palendae I want to say yes because you can make glance use cloudfiles | 22:44 |
| *** Mudpuppy_ has joined #openstack-ansible | 22:44 | |
| palendae | odyssey4me: Yeah, it's a special case though | 22:45 |
| palendae | Basically I have 1 swift proxy node, want to use its IP and not redo the F5 | 22:46 |
| palendae | I'm not seeing a swift-specific URL in glance | 22:46 |
| palendae | Oh...is it taking it from the service catalog again? | 22:47 |
| palendae | sigmavirus24: ^ | 22:47 |
| palendae | Though cloudfiles wouldn't be in the service catalog | 22:47 |
| mattt | cloudnull: lolzusofunny | 22:47 |
| *** Mudpuppy has quit IRC | 22:47 | |
| odyssey4me | if it's cloudfiles then you probably won't have object storage in the service catalogue | 22:47 |
| palendae | Yeah... | 22:47 |
| palendae | Not seeing a URL in here | 22:48 |
| cloudnull | so the auth address is defined here: https://github.com/stackforge/os-ansible-deployment/blob/juno/rpc_deployment/roles/glance_common/templates/glance-api.conf#L72 | 22:48 |
| palendae | Yeah, got the auth one | 22:48 |
| cloudnull | and in juno swift_store_auth_address is defined through rax uservars | 22:48 |
| palendae | Oh, is that the *swift* auth, not keystone? | 22:49 |
| palendae | Hrm, not necessarily, looks like it's cloud identity | 22:49 |
| cloudnull | nevermind thats keystone and not the specific proxy address. | 22:50 |
| odyssey4me | well, if that auth address is keystone then it should discover through the service catalogue, surely? | 22:50 |
| *** arbrandes has quit IRC | 22:50 | |
| odyssey4me | I'd suggest changing up the service catalogue for the object store | 22:50 |
| cloudnull | if you want to change the proxy address as stored in the service catalog, in juno, youll need to modify the ./rpc_deployment/playbooks/openstack/swift-proxy.yml file with the value you want | 22:52 |
| cloudnull | remove the endpoint from the service catalog | 22:52 |
| cloudnull | and then run the swift-proxy.yml plan | 22:52 |
| cloudnull | *play | 22:52 |
| palendae | Yeah, so, the auth url is the LB at Keystone's port | 22:52 |
| cloudnull | palendae: https://github.com/stackforge/os-ansible-deployment/blob/juno/rpc_deployment/vars/openstack_service_vars/swift_proxy_endpoint.yml | 22:53 |
| sigmavirus24 | palendae: sorry | 22:54 |
| palendae | Hrm, yeah, my swift changes didn't make it work | 22:58 |
| *** tlian has quit IRC | 22:59 | |
| *** cloudtrainme has quit IRC | 23:00 | |
| -openstackstatus- NOTICE: Gerrit is offline from 23:00 to 23:30 UTC while some projects are renamed. http://lists.openstack.org/pipermail/openstack-dev/2015-September/074235.html | 23:00 | |
| *** ChanServ changes topic to "Gerrit is offline from 23:00 to 23:30 UTC while some projects are renamed. http://lists.openstack.org/pipermail/openstack-dev/2015-September/074235.html" | 23:00 | |
| odyssey4me | popcorn time :) | 23:01 |
| KLevenstein | odyssey4me: it begons? | 23:02 |
| KLevenstein | *begins | 23:02 |
| *** meteorfox has quit IRC | 23:02 | |
| odyssey4me | KLevenstein yep | 23:03 |
| odyssey4me | follow progress in openstack-infra and on https://etherpad.openstack.org/p/project-renames-Septemeber-2015 | 23:03 |
| * KLevenstein goes to pour a drink | 23:03 | |
| *** jlvillal has quit IRC | 23:03 | |
| *** meteorfox has joined #openstack-ansible | 23:06 | |
| odyssey4me | I suppose I may as well get on with changing up the wiki and such. | 23:07 |
| *** arbrandes has joined #openstack-ansible | 23:07 | |
| palendae | Does swift-all register with keystone? | 23:09 |
| palendae | Not seeing my IP in there | 23:10 |
| *** sdake has joined #openstack-ansible | 23:10 | |
| *** sdake_ has quit IRC | 23:14 | |
| odyssey4me | KLevenstein it seems from current discussion that the old repo should be redirected automatically by github - so we may be in some luck with the older deployments | 23:18 |
| KLevenstein | that would be helpful | 23:18 |
| odyssey4me | (where github was the referential git source) | 23:19 |
| odyssey4me | we can double-check when it's all done | 23:19 |
| *** markvoelker has quit IRC | 23:19 | |
| odyssey4me | we can also try to validate whether git.o.o redirects | 23:19 |
| prometheanfire | https://github.com/openstack/openstack-ansible | 23:21 |
| odyssey4me | KLevenstein heh, in a browser enter https://github.com/stackforge/os-ansible-deployment right now | 23:22 |
| KLevenstein | sweet! | 23:22 |
| odyssey4me | and https://github.com/stackforge/os-ansible-deployment-specs | 23:23 |
| palendae | Gonna re-push my rpc-openstack PR and see what happens | 23:23 |
| KLevenstein | I’ll see what happens with the docs. | 23:24 |
| odyssey4me | palendae hang ten - stuff's still going on | 23:25 |
| KLevenstein | yeah, docs still not building right yet. | 23:25 |
| odyssey4me | KLevenstein are you able to poke some people to get attention on https://review.openstack.org/222323 some time this w/end? | 23:25 |
| KLevenstein | I can ping Lana | 23:26 |
| KLevenstein | if you want, I’ll add her to the reviewers? | 23:26 |
| odyssey4me | not right now - we can wait for the target folder to actually be created, but ideally in the Aus Monday? | 23:26 |
| odyssey4me | ah, good plan :) | 23:26 |
| KLevenstein | consider it done | 23:26 |
| odyssey4me | happy to wait for a work day, just wanting it to be the earliest work day possible :) | 23:27 |
| palendae | Ah, yeah, it clones from git.openstack.org which isn't working | 23:27 |
| odyssey4me | palendae yep, they're still busy there | 23:28 |
| odyssey4me | I have to say, this is pretty impressive operational co-ordination. | 23:28 |
| odyssey4me | I've seen teams working in the same room which can't do this well. | 23:29 |
| KLevenstein | odyssey4me: added her; she’ll see it on her Monday/our Sunday | 23:29 |
| odyssey4me | KLevenstein perfect, thanks - you're a star! | 23:29 |
| odyssey4me | and https://review.openstack.org/#/admin/groups/490 and https://review.openstack.org/#/admin/groups/491 :) | 23:31 |
| odyssey4me | now for the project-config / system-config / nodepool changes to merge | 23:32 |
| cloudnull | gerrit is reflecting the change https://review.openstack.org/#/q/status:open+project:openstack/openstack-ansible,n,z | 23:33 |
| odyssey4me | anyone keen to submit the review to gerrit dashboard to change the dashboard? | 23:34 |
| odyssey4me | I forgot about that | 23:34 |
| cloudnull | what ? | 23:35 |
| odyssey4me | ie https://github.com/stackforge/gerrit-dash-creator/blob/master/dashboards/os-ansible-deployment.dash | 23:35 |
| *** alop has quit IRC | 23:35 | |
| odyssey4me | did y'all even know that existed? | 23:35 |
| cloudnull | nope | 23:35 |
| KLevenstein | oh cool, our docs are building | 23:35 |
| odyssey4me | it's the secret sauce behind bit.ly/os-ansible-deployment-review | 23:36 |
| palendae | odyssey4me: Did not | 23:36 |
| cloudnull | also it doesnt seem that there is an equivelant repo in /openstack/ | 23:36 |
| odyssey4me | which I guess needs to change now too :) | 23:36 |
| cloudnull | ill review that | 23:36 |
| odyssey4me | palendae it's a convenience dashboard, primarily for core reviewers :) | 23:37 |
| odyssey4me | it hides the stuff that hasn't successfully built, or already has negative votes | 23:37 |
| odyssey4me | I should probably add the updated version to the wiki or something :) | 23:37 |
| *** ChanServ changes topic to "Topic: Launchpad: https://launchpad.net/openstack-ansible Weekly Meetings: https://wiki.openstack.org/wiki/Meetings/openstack-ansible || Repo rename from stackforge/os-ansible-deployment to openstack/openstack-ansible happens Sept 11 2015 23:00 to 23:30. See https://review.openstack.org/#/c/200730/" | 23:38 | |
| odyssey4me | ok, so the bulk of the work is done - now it's just down to the proposed change merging | 23:38 |
| odyssey4me | *changes | 23:38 |
| odyssey4me | once that's done - meaning that our ACL's will be updated, we should have our full capability back | 23:39 |
| palendae | odyssey4me: I know, I saw that | 23:39 |
| palendae | them* | 23:39 |
| palendae | Also the 'new' view in gerrit is soooooooooooooo much nicer | 23:39 |
| cloudnull | odyssey4me: https://review.openstack.org/#/c/222795/ | 23:40 |
| palendae | https://review.openstack.org/#/settings/preferences > change view to 'New Screen' | 23:40 |
| odyssey4me | palendae yep - it takes some getting used to, but it is great | 23:41 |
| odyssey4me | especially for seeing the history of comments :) | 23:41 |
| odyssey4me | KLevenstein ok, so for the sake of information - whether it's useful or not... | 23:41 |
| dolphm | i had not realized the change in acronym resulting from openstack/ until just now. | 23:41 |
| odyssey4me | openstack-ansible 11.1.2 referred to github for its git source | 23:42 |
| palendae | dolphm: OA | 23:42 |
| odyssey4me | 11.2.0 switched to using git.openstack.org | 23:42 |
| cloudnull | yep OSAD is a clasic | 23:42 |
| cloudnull | *classic | 23:42 |
| palendae | deprecated | 23:42 |
| KLevenstein | odyssey4me: ok | 23:42 |
| dolphm | sosad nomore osad | 23:42 |
| * dolphm apologizes | 23:42 | |
| cloudnull | hahaha | 23:42 |
| palendae | Not sad | 23:42 |
| cloudnull | that'll be the next shirt | 23:42 |
| odyssey4me | I expect that rpc-openstack r11 is using git.openstack.org - so there's not redirect there | 23:42 |
| palendae | Now everything matches | 23:42 |
| dolphm | how do i upgrade my legacy shirt? | 23:43 |
| palendae | odyssey4me: Correct | 23:43 |
| KLevenstein | odyssey4me: not sure where the docs will be affected but I’ll have a look | 23:43 |
| rromans | dolphm: turn it upside down | 23:43 |
| dolphm | plz assign bp zero-downtime-shirt-upgrades to claco kthx | 23:43 |
| palendae | >:( | 23:44 |
| odyssey4me | KLevenstein so for rpc-openstack, nothing changes - the osad rename will be handled transparently by a change in the rpc-openstack submodule poiinter | 23:44 |
| cloudnull | OSA is the new acronym. the D has been dropped. I'm sure Poettering is crying right now. | 23:44 |
| *** openstackgerrit has quit IRC | 23:46 | |
| *** openstackgerrit has joined #openstack-ansible | 23:46 | |
Generated by irclog2html.py 2.14.0 by Marius Gedminas - find it at mg.pov.lt!