| *** jmccrory has joined #openstack-ansible | 00:17 | |
| *** jmccrory has quit IRC | 00:17 | |
| *** jmccrory has joined #openstack-ansible | 00:20 | |
| *** jmccrory has quit IRC | 00:20 | |
| *** ccrouch is now known as ccrouch-biab | 00:21 | |
| *** ccrouch-biab is now known as ccrouch | 00:36 | |
| bgmccollum | andymccr: dropping replication_network fixes my problem (since i dont have a dedicated replication_network). but if replication_network is defined and is the same value as storage_network, the *-server-replication.conf files don't get dropped. This seems to have been fixed in master. Thanks again... | 00:38 |
|---|---|---|
| *** vmtrooper has joined #openstack-ansible | 00:41 | |
| *** vmtrooper has quit IRC | 00:46 | |
| *** jmccrory has joined #openstack-ansible | 00:47 | |
| *** ccrouch is now known as ccrouch-bbl | 00:50 | |
| *** jmccrory has quit IRC | 00:52 | |
| *** jmccrory has joined #openstack-ansible | 00:53 | |
| *** sigmavirus24 is now known as sigmavirus24_awa | 01:19 | |
| *** markvoelker has quit IRC | 01:27 | |
| *** markvoelker has joined #openstack-ansible | 01:27 | |
| *** markvoelker has quit IRC | 01:32 | |
| *** sdake_ has quit IRC | 01:42 | |
| *** ccrouch-bbl is now known as ccrouch | 01:52 | |
| *** vmtrooper has joined #openstack-ansible | 02:30 | |
| *** vmtrooper has quit IRC | 02:35 | |
| *** jmccrory has quit IRC | 02:38 | |
| *** sdake_ has joined #openstack-ansible | 02:48 | |
| *** jmccrory has joined #openstack-ansible | 02:53 | |
| *** jmccrory has quit IRC | 02:53 | |
| *** jmccrory has joined #openstack-ansible | 02:53 | |
| *** sdake_ has quit IRC | 03:38 | |
| *** jmccrory has quit IRC | 03:47 | |
| *** docaedo has joined #openstack-ansible | 04:08 | |
| *** sdake_ has joined #openstack-ansible | 04:08 | |
| *** vmtrooper has joined #openstack-ansible | 04:19 | |
| *** vmtrooper has quit IRC | 04:24 | |
| *** devlaps has quit IRC | 04:25 | |
| *** sdake__ has joined #openstack-ansible | 04:55 | |
| *** sdake_ has quit IRC | 04:58 | |
| *** sdake__ has quit IRC | 05:08 | |
| *** alextricity has quit IRC | 05:09 | |
| *** galstrom_zzz has quit IRC | 05:09 | |
| *** alextric_ has joined #openstack-ansible | 05:09 | |
| *** raginbajin has quit IRC | 05:09 | |
| *** galstrom_zzz has joined #openstack-ansible | 05:10 | |
| *** raginbajin has joined #openstack-ansible | 05:13 | |
| *** britthouser has joined #openstack-ansible | 05:16 | |
| *** britthouser has quit IRC | 05:19 | |
| *** britthouser has joined #openstack-ansible | 05:24 | |
| *** markvoelker has joined #openstack-ansible | 05:25 | |
| *** markvoelker has quit IRC | 05:30 | |
| *** britthouser has quit IRC | 05:34 | |
| *** britthouser has joined #openstack-ansible | 05:40 | |
| *** britthouser has quit IRC | 05:46 | |
| *** britthouser has joined #openstack-ansible | 05:50 | |
| *** sdake_ has joined #openstack-ansible | 06:02 | |
| *** sdake__ has joined #openstack-ansible | 06:04 | |
| *** sdake_ has quit IRC | 06:08 | |
| *** vmtrooper has joined #openstack-ansible | 06:08 | |
| *** sdake_ has joined #openstack-ansible | 06:10 | |
| *** vmtrooper has quit IRC | 06:13 | |
| *** sdake__ has quit IRC | 06:13 | |
| *** sdake__ has joined #openstack-ansible | 06:14 | |
| *** sdake_ has quit IRC | 06:18 | |
| *** markvoelker has joined #openstack-ansible | 06:26 | |
| *** markvoelker has quit IRC | 06:32 | |
| *** sdake_ has joined #openstack-ansible | 06:37 | |
| *** sdake__ has quit IRC | 06:41 | |
| *** sdake__ has joined #openstack-ansible | 06:58 | |
| *** sdake_ has quit IRC | 07:02 | |
| *** sdake__ has quit IRC | 07:08 | |
| *** stevemar has quit IRC | 07:22 | |
| *** markvoelker has joined #openstack-ansible | 07:28 | |
| *** markvoelker has quit IRC | 07:33 | |
| *** sdake_ has joined #openstack-ansible | 07:42 | |
| *** sdake_ has quit IRC | 07:49 | |
| *** sdake has joined #openstack-ansible | 07:56 | |
| *** vmtrooper has joined #openstack-ansible | 07:57 | |
| *** vmtrooper has quit IRC | 08:02 | |
| *** markvoelker has joined #openstack-ansible | 08:29 | |
| *** markvoelker has quit IRC | 08:34 | |
| openstackgerrit | Hugh Saunders proposed stackforge/os-ansible-deployment: Revert "Trim ansible's output to console.log" https://review.openstack.org/162984 | 08:38 |
| *** sdake__ has joined #openstack-ansible | 08:54 | |
| *** sdake has quit IRC | 08:57 | |
| openstackgerrit | git-harry proposed stackforge/os-ansible-deployment: Test commit - do not review https://review.openstack.org/161670 | 09:21 |
| *** nosleep77 has quit IRC | 09:30 | |
| *** markvoelker has joined #openstack-ansible | 09:30 | |
| *** markvoelker has quit IRC | 09:35 | |
| *** vmtrooper has joined #openstack-ansible | 09:45 | |
| *** vmtrooper has quit IRC | 09:51 | |
| openstackgerrit | Merged stackforge/os-ansible-deployment: Set backend_alg for swift_proxy to source https://review.openstack.org/162967 | 09:59 |
| openstackgerrit | Merged stackforge/os-ansible-deployment: Fix the br-vlan mode in the example interfaces https://review.openstack.org/163048 | 10:00 |
| openstackgerrit | Jesse Pretorius proposed stackforge/os-ansible-deployment: Revert "Trim ansible's output to console.log" https://review.openstack.org/162984 | 10:15 |
| openstackgerrit | Andy McCrae proposed stackforge/os-ansible-deployment: Set container_networks to a dict rather than None https://review.openstack.org/163083 | 10:31 |
| *** markvoelker has joined #openstack-ansible | 10:31 | |
| *** markvoelker has quit IRC | 10:36 | |
| openstackgerrit | Merged stackforge/os-ansible-deployment: Create external address setting for maas checks https://review.openstack.org/162734 | 10:48 |
| openstackgerrit | Jesse Pretorius proposed stackforge/os-ansible-deployment: Update contributing guidelines with backport guidance https://review.openstack.org/163764 | 11:08 |
| openstackgerrit | Merged stackforge/os-ansible-deployment: Fix the br-vlan mode in the example interfaces https://review.openstack.org/163044 | 11:10 |
| openstackgerrit | Merged stackforge/os-ansible-deployment: Remove unused external_vip_address reference https://review.openstack.org/163238 | 11:12 |
| openstackgerrit | Jesse Pretorius proposed stackforge/os-ansible-deployment: Create ansible log file and directory https://review.openstack.org/163770 | 11:22 |
| openstackgerrit | Jesse Pretorius proposed stackforge/os-ansible-deployment: Create ansible log file and directory https://review.openstack.org/163772 | 11:22 |
| openstackgerrit | Jesse Pretorius proposed stackforge/os-ansible-deployment: Adjust method for trimming ansible output debug information https://review.openstack.org/163775 | 11:29 |
| *** markvoelker has joined #openstack-ansible | 11:33 | |
| *** vmtrooper has joined #openstack-ansible | 11:34 | |
| *** markvoelker has quit IRC | 11:37 | |
| *** vmtrooper has quit IRC | 11:39 | |
| openstackgerrit | Jesse Pretorius proposed stackforge/os-ansible-deployment: Ensure get-pip.py is retrieved properly https://review.openstack.org/162183 | 11:41 |
| openstackgerrit | Jesse Pretorius proposed stackforge/os-ansible-deployment: Revert "Trim ansible's output to console.log" https://review.openstack.org/162984 | 11:43 |
| openstackgerrit | Jesse Pretorius proposed stackforge/os-ansible-deployment: Enable network basic ops tests https://review.openstack.org/163544 | 11:44 |
| *** galstrom_zzz is now known as galstrom | 11:54 | |
| openstackgerrit | git-harry proposed stackforge/os-ansible-deployment: WIP: Testing hash_behaviour https://review.openstack.org/162989 | 12:02 |
| *** markvoelker has joined #openstack-ansible | 12:15 | |
| *** sdake has joined #openstack-ansible | 12:23 | |
| *** sdake__ has quit IRC | 12:23 | |
| openstackgerrit | Andy McCrae proposed stackforge/os-ansible-deployment: Set container_networks to a dict rather than None https://review.openstack.org/163083 | 12:35 |
| odyssey4me | hughsaunders odd, the ansible logging actually works - check the console output here: https://review.openstack.org/163775 | 12:36 |
| odyssey4me | note that in the review I changed the method of doing the processing - perhaps that's made the difference | 12:36 |
| *** galstrom is now known as galstrom_zzz | 12:37 | |
| hughsaunders | odyssey4me: I thought it was odd that run-playbooks was sourced rather than executed previously | 12:38 |
| hughsaunders | maybe we should hold off on the revert and see if your patch fixes logging for other runs also.. | 12:38 |
| odyssey4me | hughsaunders the library is the only one that actually needs sourcing - all the other scripts can be run via bash | 12:44 |
| *** sdake has quit IRC | 12:48 | |
| *** sdake has joined #openstack-ansible | 12:49 | |
| openstackgerrit | git-harry proposed stackforge/os-ansible-deployment: Test commit - do not review https://review.openstack.org/161670 | 12:50 |
| mattt | odyssey4me hughsaunders : we're going to have to bump tempest sha :P https://bugs.launchpad.net/openstack-ansible/+bug/1431052 | 13:06 |
| openstack | Launchpad bug 1431052 in openstack-ansible "Swift failures with upgrade to tempest 7c73dd5" [Undecided,New] | 13:06 |
| hughsaunders | mattt: ok | 13:06 |
| odyssey4me | mattt fair enough - we'll have to bump both juno and master then, but test with master first obviously | 13:17 |
| odyssey4me | is there a minimum change sha that we can bump to - or should we bump to whatever the current sha is? | 13:17 |
| odyssey4me | ie shall we be conservative, or hold on to our pants? | 13:18 |
| andymccr | surely holding onto your pants would be a conservative move regardless :D | 13:18 |
| *** stevemar has joined #openstack-ansible | 13:18 | |
| odyssey4me | haha | 13:19 |
| mattt | odyssey4me: i worked out where it broke and was fixed, so to be conservative we can either go to the sha before it broke or the one after it was fixed | 13:19 |
| mattt | probably better to bump forward | 13:19 |
| odyssey4me | after fixed, I think | 13:20 |
| odyssey4me | before broke would take us to older than the current sha, right? | 13:20 |
| mattt | odyssey4me: yep, i'll change to go to after | 13:22 |
| andymccr | hughsaunders: did we change the subset of tempest tests for commit_multinode? | 13:23 |
| *** vmtrooper has joined #openstack-ansible | 13:23 | |
| mattt | andymccr: b8c2d7c | 13:24 |
| openstackgerrit | Jesse Pretorius proposed stackforge/os-ansible-deployment: Change testr to run in serial https://review.openstack.org/163807 | 13:26 |
| openstackgerrit | Jesse Pretorius proposed stackforge/os-ansible-deployment: Change testr to run in serial https://review.openstack.org/163808 | 13:27 |
| hughsaunders | andymccr: no | 13:27 |
| andymccr | hughsaunders: yeh it was the forbidden stuff that mattt ran into just for keystone and not swift | 13:27 |
| andymccr | but the sha worked perfectly | 13:28 |
| andymccr | so looks good | 13:28 |
| andymccr | commit_multinode worked end ot end | 13:28 |
| *** vmtrooper has quit IRC | 13:29 | |
| mattt | andymccr: ok going to put a change through for that now | 13:29 |
| openstackgerrit | Jesse Pretorius proposed stackforge/os-ansible-deployment: Adjust method for trimming ansible output debug information https://review.openstack.org/163809 | 13:32 |
| openstackgerrit | Jesse Pretorius proposed stackforge/os-ansible-deployment: Adjust method for trimming ansible output debug information https://review.openstack.org/163810 | 13:32 |
| openstackgerrit | Jesse Pretorius proposed stackforge/os-ansible-deployment: Adjust method for trimming ansible output debug information https://review.openstack.org/163809 | 13:35 |
| openstackgerrit | Jesse Pretorius proposed stackforge/os-ansible-deployment: Adjust method for trimming ansible output debug information https://review.openstack.org/163810 | 13:36 |
| openstackgerrit | Matt Thompson proposed stackforge/os-ansible-deployment: Bump tempest sha https://review.openstack.org/163815 | 13:38 |
| *** sigmavirus24_awa is now known as sigmavirus24 | 13:51 | |
| openstackgerrit | Merged stackforge/os-ansible-deployment: Adjust method for trimming ansible output debug information https://review.openstack.org/163775 | 14:11 |
| openstackgerrit | Jesse Pretorius proposed stackforge/os-ansible-deployment: Adjust method for trimming ansible output debug information https://review.openstack.org/163809 | 14:15 |
| openstackgerrit | Jesse Pretorius proposed stackforge/os-ansible-deployment: Change testr to run in serial https://review.openstack.org/163807 | 14:15 |
| openstackgerrit | Jesse Pretorius proposed stackforge/os-ansible-deployment: Simplify and improve bootstrap/gate/run scripts https://review.openstack.org/163837 | 14:15 |
| *** KLevenstein has joined #openstack-ansible | 14:20 | |
| *** Mudpuppy has joined #openstack-ansible | 14:22 | |
| *** Mudpuppy has quit IRC | 14:22 | |
| *** Mudpuppy has joined #openstack-ansible | 14:24 | |
| openstackgerrit | Jesse Pretorius proposed stackforge/os-ansible-deployment: Simplify and improve bootstrap/gate/run scripts https://review.openstack.org/163846 | 14:32 |
| openstackgerrit | Jesse Pretorius proposed stackforge/os-ansible-deployment: Create ansible log file and directory https://review.openstack.org/163770 | 14:42 |
| openstackgerrit | Jesse Pretorius proposed stackforge/os-ansible-deployment: Create ansible log file and directory https://review.openstack.org/163772 | 14:45 |
| openstackgerrit | Jesse Pretorius proposed stackforge/os-ansible-deployment: Adjust method for trimming ansible output debug information https://review.openstack.org/163810 | 14:50 |
| openstackgerrit | Matt Thompson proposed stackforge/os-ansible-deployment: Bump tempest sha https://review.openstack.org/163815 | 14:55 |
| *** BjoernT has joined #openstack-ansible | 15:00 | |
| *** alextric_ is now known as alextricity | 15:10 | |
| *** vmtrooper has joined #openstack-ansible | 15:12 | |
| *** erikmwilson has quit IRC | 15:13 | |
| *** erikmwilson has joined #openstack-ansible | 15:13 | |
| *** erikmwilson_ has joined #openstack-ansible | 15:13 | |
| *** vmtrooper has quit IRC | 15:17 | |
| *** jwagner_away is now known as jwagner | 15:20 | |
| openstackgerrit | Merged stackforge/os-ansible-deployment: Enable network basic ops tests https://review.openstack.org/163544 | 15:29 |
| *** daneyon_ has joined #openstack-ansible | 15:39 | |
| *** daneyon has quit IRC | 15:42 | |
| sdake | morning | 15:45 |
| sigmavirus24 | morning sdake | 15:46 |
| -sigmavirus24- Meeting in 14 minutes in #openstack-meeting-4 | 15:47 | |
| *** nermik has joined #openstack-ansible | 15:57 | |
| cloudnull | ping odyssey4me | 16:05 |
| odyssey4me | cloudnull yebo? | 16:05 |
| cloudnull | meetings times | 16:05 |
| cloudnull | git-harry meeting time | 16:10 |
| openstackgerrit | Merged stackforge/os-ansible-deployment: Set container_networks to a dict rather than None https://review.openstack.org/163083 | 16:12 |
| *** jmccrory has joined #openstack-ansible | 16:30 | |
| *** daneyon has joined #openstack-ansible | 16:42 | |
| *** daneyon_ has quit IRC | 16:44 | |
| *** vmtrooper has joined #openstack-ansible | 17:01 | |
| Apsu | BjoernT: So, a few things. The project does have a load balancing component provided that we use a lot for our testing and dev, https://github.com/stackforge/os-ansible-deployment/blob/master/playbooks/haproxy-install.yml | 17:01 |
| cloudnull | also if we use haproxy 1.5 we can do ssl within it | 17:02 |
| Apsu | BjoernT: However, that isn't something we support or document for the product that we, rackspace, sell and maintain. | 17:02 |
| BjoernT | apsu: then we need a keepalive component too | 17:02 |
| cloudnull | BjoernT we as in OSAD but not for rax | 17:02 |
| Apsu | BjoernT: I think that's a worthwhile discussion to have. | 17:02 |
| cloudnull | sdake you around ? | 17:03 |
| Apsu | haproxy-install is minimal effort. | 17:03 |
| cloudnull | i know that you guys we're looking into updating the haproxy bits | 17:03 |
| Apsu | There are many reasons why we don't want to ship or support it for the product we sell. | 17:03 |
| BjoernT | yes, some customers are using this already in their DC | 17:03 |
| Apsu | But for our dev needs it was plenty sufficient. | 17:03 |
| cloudnull | is that something that we can look into for the process of testing ssl offloading ? | 17:03 |
| sdake | cloudnull wound me | 17:03 |
| Apsu | BjoernT: Yes, and we've told them they should not be, and trying hard to not allow/support them if they choose to. | 17:03 |
| BjoernT | Yes we know. We fight this every day. From the SSL perspective it would help us to remove all the hard coded stuff and replace it with variables so we can change it easily | 17:04 |
| sdake | cloudnull sorry you possibly didn't get the joke - I often say shoot and wound me when people ping me - what can I do for you | 17:05 |
| BjoernT | The offloading stuff is a different story, similar to FULL ssl support down to all services | 17:05 |
| cloudnull | haha. no worries sdake | 17:05 |
| cloudnull | was just curious if you've had a opportunity to look at our s/w lb roles/plays ? | 17:05 |
| cloudnull | and if you | 17:05 |
| cloudnull | have done any of the haproxy 1.5 ssl things | 17:05 |
| sdake | I have spent 10-15 hours analyzing the source base | 17:05 |
| sdake | keep in mind I have zero ansible experience | 17:06 |
| cloudnull | in master? or in the other branches? | 17:06 |
| sdake | but it looks pretty solid to me | 17:06 |
| *** vmtrooper has quit IRC | 17:06 | |
| sdake | I know zero about ssl haproxy | 17:06 |
| cloudnull | kk. was just curious . | 17:06 |
| Apsu | BjoernT: Separating out these concerns goes like this: A) SSL offloading (haproxy, F5, Apache, Nginx, etc), B) Load Balancing, C) High Availability, D) Endpoint flexibility, E) Project vs Product needs | 17:07 |
| daneyon | I stood up a multi-node cluster using v10. Roughly how many other folks outside of RAX have used osad to deploy a multi node cluster? | 17:07 |
| sdake | also until after VC ODS my plate is so full I am going to gain weight if I finish :) | 17:07 |
| palendae | daneyon: That's a great question :p | 17:07 |
| sdake | daneyon that might be worht a post to the ml to share your experiences | 17:07 |
| palendae | I know approd0 had, but they're not on right now | 17:07 |
| Apsu | sdake: The short version is easy. < 1.5, ssl was a separate binary essentially, that had some performance and behavior issues in some cases. >= 1.5, it's native and seems to be awesome. | 17:07 |
| Apsu | BjoernT: As far as I'm concerned, D all by itself is reasonable and an easy win. I'm happy to help make it go. | 17:08 |
| odyssey4me | daneyon no idea, but I did see someone state that they're doing R&D to do so | 17:08 |
| Apsu | BjoernT: A-C and E are separate concerns and need separate consideration and discussion :) | 17:08 |
| odyssey4me | to be fair, the removal of all the rax bits was only implemented very recently | 17:08 |
| palendae | odyssey4me: And not even in an actual release yet | 17:09 |
| palendae | Just master | 17:09 |
| BjoernT | apsu: Yes that's OK, that's why I asked to slice it up and just start with the removal of the http:// hard coded variables | 17:09 |
| BjoernT | that helps me the rest is not so much an issue for us since we use SSL on F5 | 17:09 |
| daneyon | sdake: I am planning on doing a write-up and will post for feedback. | 17:09 |
| BjoernT | But also telling you that full ssl requirements are coming with big customers who just have HIPAA or PCI requirements | 17:10 |
| Apsu | BjoernT: Sure. For us, that means prioritizing our own dev work and planning around our product needs. Which will of course benefit the project by way of us implementing them. | 17:11 |
| Apsu | But those types of things may or may not be project priorities, and our public meetings will reflect that. :) | 17:11 |
| BjoernT | Personally I would like to start tackling SSL stuff before messing with Kilo. But that's my opinion. Without full SSL implementation we can't sell the product to specific customers | 17:11 |
| Apsu | Likewise our public bug triages. | 17:11 |
| odyssey4me | BjoernT that is an issue for the organisation doing the selling. If that organisation wants to prioritise it, then it needs to allocate the resourcing to make it happen. | 17:13 |
| Apsu | ^ | 17:13 |
| BjoernT | Right I know | 17:13 |
| odyssey4me | I would recommend that you highlight this priority through the appropriate internal paths - otherwise you'll have to wait for someone in the project to take an interest and get it done. | 17:14 |
| sdake | I dont know enough about ssl to provide any educated opnions | 17:14 |
| sdake | daneyon any chance you can find out ? | 17:14 |
| Apsu | sdake: It's terrible, insufficient, but necessary. And openstack projects are inconsistent with it. | 17:15 |
| Apsu | That's the elevator pitch. Delightful, really | 17:15 |
| odyssey4me | A simple workaround for anyone at this stage is simply to adjust the endpoints in keystone after the deployment to reflect what's needed. So I don't really see the real problem here. | 17:15 |
| sdake | apsu our priority for Lamar is likely to be around native packaging support for debuntian and RPM, as well as out-of-band CI/CD/container building | 17:16 |
| BjoernT | odyssey4me: Yeah I got this feeling too, we will talk internally about priorities with the product people. No need to spin here | 17:16 |
| odyssey4me | AFAIK the endpoints are not overridden if they're set. If they are, then we should deal with that as a bug or at least make it a configurable behaviour. | 17:16 |
| Apsu | odyssey4me: The issue is the protocol. | 17:16 |
| sdake | lamar | 17:16 |
| Apsu | That's purely what Bjoern wants. Make the protocol a variable. | 17:16 |
| sdake | where did that come from | 17:16 |
| sdake | old product name 10 years ago :) | 17:16 |
| Apsu | sdake: No idea. I was curious if we were getting a new name for L somehow | 17:17 |
| Apsu | lol | 17:17 |
| sdake | I meant liberty ;-) | 17:17 |
| Apsu | Freedom. 'murica. | 17:17 |
| daneyon | sdake: i'm not completely following. Find out if SSL is a priority? | 17:17 |
| Apsu | odyssey4me: Also, there's no API-ability to update a keystone endpoint URL. You have to add a new one, delete the old one, or mangle the database. | 17:18 |
| Apsu | odyssey4me: It's not that it's impossible, just crufty | 17:19 |
| sdake | there was a technical question in there | 17:19 |
| sdake | unfortunately I dont know anything about it to even know what quesiton to ask :) | 17:19 |
| odyssey4me | Apsu agreed, but this is hardly a difficult thing to do - unless something's changed recently... | 17:19 |
| sdake | perhaps someone oculd post a ml q to bring in the broader openstack community expertise | 17:19 |
| Apsu | odyssey4me: No, add/remove isn't difficult to do. But we fix all of this with simply making the protocol variable. | 17:20 |
| Apsu | It's really not even worthy of discussion. Make variable, set default to http, commit, go. | 17:20 |
| sdake | apsu wfm :) | 17:21 |
| Apsu | Then deployers can put in https if they've solved all the other SSL parts elsewhere. | 17:21 |
| odyssey4me | Apsu sure - code wins... would love to see the patch :) | 17:21 |
| sdake | at this point i'm just a spectator - but plan to get more engaged in code writing during Liberty | 17:22 |
| daneyon | Apsu: Although adding https as a var, I thought OpenStack services performed and scaled poorly when using https instead of http and the recommended approach was SSL offload? | 17:25 |
| *** britthouser has quit IRC | 17:26 | |
| Apsu | daneyon: Setting the endpoint to https doesn't mean the service has to be the SSL endpoint. | 17:26 |
| daneyon | Apsu: ^ adding https as a var should be simple.... | 17:26 |
| *** britthouser has joined #openstack-ansible | 17:26 | |
| *** devlaps has joined #openstack-ansible | 17:26 | |
| Apsu | If you use an SSL offloader with service VIPs through some variety of proxy or load-balancer, then the services can listen on http and perform as normal, while the clients will use SSL because the endpoint protocol says https | 17:26 |
| Apsu | That's how we (RPC) deploy, with F5s as LB/SSL offload | 17:27 |
| daneyon | Apsu: I see, you re: to setting the ks endpoints as https. | 17:27 |
| Apsu | Yep | 17:27 |
| BjoernT | daneyon: Do you have more information around that. SSL should not be an issue nowadays | 17:27 |
| Apsu | Super easy win, solves BjoernT's need, provides a starting point for further SSL work | 17:27 |
| BjoernT | apsu: yess | 17:27 |
| daneyon | BjoernT: I may be 1 or 2 releases behind in my knowledge... it's been a while since I've look into it. | 17:28 |
| openstackgerrit | Jesse Pretorius proposed stackforge/os-ansible-deployment: Revise gate script library to report correctly https://review.openstack.org/163914 | 17:28 |
| Apsu | Lunching! | 17:29 |
| daneyon | Apsu: Right about the the service not needing to be the SSL endpoint, it could be offloaded which is what I thought the openstack best practice. I need to do some homework to see if openstack services provide better native ssl support. | 17:31 |
| odyssey4me | daneyon don't waste your time | 17:33 |
| odyssey4me | there's a general view that the native SSL/TLS support is generally best for development testing only | 17:33 |
| odyssey4me | and it's very well founded | 17:33 |
| daneyon | odyssey4me: that's what i recall. So, SSL offload is still the preferred method. If that's what the discussion is about, then +1 from me. | 17:35 |
| odyssey4me | well, the discussion is more around getting a patch in to allow endpoints to be defined with the HTTPS protocol from the get-go | 17:35 |
| *** nermik has quit IRC | 17:36 | |
| odyssey4me | I think we all agree that SSL offloading is required, but right now within the ansible code we don't make room to define the endpoints using SSL. It has to be done after the initial deployment manually. | 17:36 |
| odyssey4me | While this is not hard to do, it's not exactly optimal. | 17:36 |
| *** nosleep77 has joined #openstack-ansible | 17:43 | |
| *** KLevenstein has quit IRC | 17:56 | |
| cloudnull | just to inject some shinanigans here in master all endpoints have a proto extension that is a default set within all service roles and can be overridden via user_variables.yml | 17:56 |
| *** KLevenstein has joined #openstack-ansible | 17:56 | |
| cloudnull | so if you had ssl offloading on a predefined lb it could be done at deployment with no additional steps. | 17:57 |
| odyssey4me | haha cloudnull in that case, perhaps the discussion goes around to a backport :p | 17:57 |
| odyssey4me | could you perhaps make a note in the bug about that being available in master and point to the variable to override? it may give incentive to someone to handle a backport | 17:58 |
| cloudnull | if we backport in juno the variable structure would be radically different. | 17:58 |
| odyssey4me | yeah, but it does change the conversation a bit :) | 17:59 |
| odyssey4me | I gotta run - trainerating. | 17:59 |
| *** Mudpuppy_ has joined #openstack-ansible | 18:02 | |
| cloudnull | see you later odyssey4me | 18:02 |
| cloudnull | odyssey4me before you bounce do you have a lp bug id # ? | 18:03 |
| *** galstrom_zzz is now known as galstrom | 18:04 | |
| bgmccollum | odyssey4me, Apsu: mental dump - https://etherpad.openstack.org/p/Ansible_Tunable_OpenStack_Configuration_Options | 18:05 |
| *** Mudpuppy has quit IRC | 18:06 | |
| *** Mudpuppy_ has quit IRC | 18:06 | |
| *** Mudpuppy has joined #openstack-ansible | 18:07 | |
| *** nermik has joined #openstack-ansible | 18:26 | |
| *** vmtrooper has joined #openstack-ansible | 18:50 | |
| *** vmtrooper has quit IRC | 18:55 | |
| *** jmccrory has quit IRC | 19:20 | |
| *** jmccrory has joined #openstack-ansible | 19:24 | |
| *** sigmavirus24 is now known as sigmavirus24_awa | 19:27 | |
| *** sigmavirus24_awa is now known as sigmavirus24 | 19:28 | |
| palendae | cloudnull: If we transition to writing specs, I've kinda changed my mind - I'd rather do it in tree in specs/ than have to wait for a new repo | 19:29 |
| palendae | So I'm +1 for specs and getting them in there, but not sure if there's consensus | 19:29 |
| cloudnull | i dont want specs in the main repo. that would require a full gate job in order to add a spec. | 19:31 |
| *** nermik has quit IRC | 19:35 | |
| b3rnard0 | The first hackathon shipment order is away! | 19:35 |
| b3rnard0 | whoops wrong channel lol | 19:36 |
| palendae | cloudnull: That's fair | 19:39 |
| palendae | +1 for specs in any case | 19:40 |
| *** galstrom is now known as galstrom_zzz | 19:40 | |
| cloudnull | no doubt. | 19:40 |
| *** openstackgerrit has quit IRC | 19:40 | |
| *** openstackgerrit has joined #openstack-ansible | 19:40 | |
| BjoernT | cloudnull: Just pinged you with one problem | 19:54 |
| cloudnull | i have no ping | 19:55 |
| BjoernT | you do, you just don't know it | 19:55 |
| cloudnull | hahaha | 19:55 |
| BjoernT | hold on, might have found it | 19:56 |
| cloudnull | so still no ping ? | 20:00 |
| cloudnull | :D | 20:00 |
| BjoernT | lol, found it. Don't need your ping anymore | 20:01 |
| BjoernT | lol | 20:01 |
| cloudnull | boom! | 20:01 |
| cloudnull | your welcome | 20:01 |
| * cloudnull a helper | 20:01 | |
| cloudnull | hahahaha | 20:01 |
| lbragstad | palendae: cloudnull fwiw ttx had a interesting post relating to the ability to add repos faster/easier http://ttx.re/stepping-out-of-the-way.html | 20:02 |
| cloudnull | ah nice. will read that too. thanks lbragstad | 20:03 |
| palendae | lbragstad: Thanks! | 20:05 |
| palendae | that's definitely relevant as we figure out how to grow this project's community base | 20:05 |
| *** Mudpuppy has quit IRC | 20:05 | |
| *** Mudpuppy has joined #openstack-ansible | 20:06 | |
| BjoernT | does anyone know this swift-recon error by chance: status error in()\n File "./swift-recon.py", line 352, in main\n stats = swift_replication(args.ring)\n File "./swift-recon.py", line 156, in swift_replication\n replication_statistics[rep_dict.pop('replication_type')] = rep_dict\nKeyError: 'replication_type'\n | 20:14 |
| odyssey4me | BjoernT that sounds like it may relate to an issue that's been patched where if you define a replication network as being the same as (not sure, andymccr knows best) then stuff breaks. | 20:15 |
| odyssey4me | There's a patch in for it, which I think has hit master - not sure if it's been backported. | 20:15 |
| BjoernT | ok | 20:15 |
| BjoernT | https://github.com/rcbops/rpc-maas/issues/172 | 20:16 |
| docaedo | Hi - I was playing around with the AIO deployment and wanted to add the "Solutions" tab to the environment - is that part of the aoi build script or is it a different playbook to run afterwards? | 20:17 |
| docaedo | (aio script I mean) | 20:18 |
| odyssey4me | docaedo the 'solutions' tab is part of rcbops/rpc-extras | 20:18 |
| BjoernT | is the solution stuff already good to go? | 20:19 |
| odyssey4me | work there is still in progress for master, but I'm not sure of the status for juno/10.x | 20:19 |
| palendae | I think miguelgrinberg would be the one to ask on that | 20:20 |
| odyssey4me | http://docs.rackspace.com/rpc/api/v10/bk-rpc-installation/content/ch-solutions.html | 20:20 |
| odyssey4me | note that it's specifically an RPC thing, not an os-ansible-deployment thing | 20:20 |
| docaedo | odyssey4me: thanks - understood it's part of the rackspace "product". os-ansible-deployment is nice, enjoying poking around in it | 20:21 |
| cloudnull | docaedo: have a look at https://github.com/rcbops/rpc-extras | 20:21 |
| cloudnull | thats where those bits live | 20:21 |
| odyssey4me | docaedo :) I'm sure that rpc-extras would be happy to accept issue reports for anything that's not working against master... if that's your current experimental base :) | 20:22 |
| cloudnull | ^ + 1 | 20:22 |
| docaedo | got it, thanks - figured I should ask first as I know there's been a bunch of great work you guys have done to make the playbooks generic | 20:22 |
| Apsu | Blame.. er, I mean thank cloudnull for that. | 20:22 |
| docaedo | and yes, I'm working off master right now, so thought I would check first in case there were known issues | 20:22 |
| cloudnull | Apsu your welcome ? | 20:23 |
| Apsu | :D | 20:23 |
| odyssey4me | :) docaedo very happy to have input, and even more happy to see bug reports and patches coming in :D | 20:23 |
| cloudnull | docaedo all issues are just unknown features :) | 20:23 |
| Apsu | Bugs are surprise features. | 20:24 |
| docaedo | haha | 20:26 |
| *** Mudpuppy has quit IRC | 20:29 | |
| odyssey4me | hmm, some warnings which may bite us in the near future: http://logs.openstack.org/14/163914/1/check/os-ansible-deployment-dsvm-check-commit/84c8d9a/console.html#_2015-03-12_18_21_21_976 | 20:34 |
| BjoernT | Hey guys, the swift playbooks should configure object-server-replicator.conf right ? | 20:35 |
| *** Mudpuppy has joined #openstack-ansible | 20:35 | |
| odyssey4me | BjoernT as I understand it, that should only be configured in the situation where the replicator network is different to the (proxy?) network | 20:36 |
| odyssey4me | let me find that bug/commit to clarify | 20:36 |
| BjoernT | hmm, shouldn't be the object replicator process running in any case ? | 20:36 |
| odyssey4me | BjoernT are you working with master, juno or 10.x? | 20:37 |
| *** Mudpuppy_ has joined #openstack-ansible | 20:37 | |
| BjoernT | juno, 10.1.2 | 20:37 |
| BjoernT | the swift replication maas checks are failing due to https://github.com/rcbops/rpc-maas/pull/173 and the last comment made me nervous that replication wasn't working | 20:38 |
| BjoernT | BTW, how can I cherry pick these when I have to specify a parent-number | 20:39 |
| *** vmtrooper has joined #openstack-ansible | 20:39 | |
| odyssey4me | BjoernT cherry pick for github or gerrit? | 20:40 |
| palendae | Guessing github if he's working rpc-maas | 20:40 |
| BjoernT | guthub https://github.com/rcbops/rpc-maas/pull/173 | 20:40 |
| *** Mudpuppy has quit IRC | 20:40 | |
| sigmavirus24 | odyssey4me: do you knwo where that's coming from? | 20:41 |
| odyssey4me | you should be able to cherry-pick 9c9483710b1bdc0f76fb049762de85ae3ccf36fe (the commit from that PR) | 20:41 |
| sigmavirus24 | oh i bet that's setuptools | 20:41 |
| odyssey4me | sigmavirus24 not sure - haven't worked with swift at all | 20:41 |
| odyssey4me | I know about these relating to the swift replicator network, but they should be in 10.1.2: | 20:42 |
| sigmavirus24 | odyssey4me: sorry, I meant the deprecation warnings: http://logs.openstack.org/14/163914/1/check/os-ansible-deployment-dsvm-check-commit/84c8d9a/console.html#_2015-03-12_18_21_21_976 | 20:42 |
| odyssey4me | https://bugs.launchpad.net/openstack-ansible/+bug/1414705 | 20:42 |
| openstack | Launchpad bug 1414705 in openstack-ansible trunk "swift: replicator init script deployed unnecessarily" [Medium,Fix committed] - Assigned to Andy McCrae (andrew-mccrae) | 20:42 |
| BjoernT | I did a patch now but the checks fail with status error Cannot parse '[replication_time] - No hosts returned valid data.' for statistics. | 20:42 |
| BjoernT | so replication is not running | 20:42 |
| odyssey4me | https://bugs.launchpad.net/openstack-ansible/+bug/1411465 | 20:42 |
| openstack | Launchpad bug 1411465 in openstack-ansible trunk "If replication_network is the same as storage_network swift will fail" [Medium,Fix committed] - Assigned to Andy McCrae (andrew-mccrae) | 20:42 |
| odyssey4me | https://bugs.launchpad.net/openstack-ansible/+bug/1427599 | 20:42 |
| openstack | Launchpad bug 1427599 in openstack-ansible "swift replication network should not be required" [High,Fix committed] - Assigned to Andy McCrae (andrew-mccrae) | 20:42 |
| odyssey4me | the last one is master only | 20:42 |
| odyssey4me | BjoernT as I understand it, the replication service itself should not run unless the replication network is different | 20:43 |
| *** erikmwilson has quit IRC | 20:43 | |
| odyssey4me | (isolated) | 20:43 |
| odyssey4me | from storage_network | 20:43 |
| BjoernT | Hmm, I though the service has to run always no matter what network. In the end you either specify a replication IP inside the ring or not | 20:44 |
| odyssey4me | as I recall from a message andymccr sent some time last night, the trick was to not specify both replication_network and storage_network if the two are the same - but I may stand corrected by people who know better :) | 20:44 |
| *** vmtrooper has quit IRC | 20:44 | |
| odyssey4me | BjoernT not sure - just trying to provide options for consideration :) | 20:45 |
| BjoernT | so should I apply the commit 8cf39b11ff0dab230b6ad2698b7512679d05e2e1 or remove replication network | 20:45 |
| BjoernT | ? | 20:45 |
| odyssey4me | BjoernT that commit only applies to master | 20:46 |
| BjoernT | damn | 20:46 |
| odyssey4me | 10.1.2 should already have it, so I guess the solution is to remove replication_network as per https://bugs.launchpad.net/openstack-ansible/+bug/1411465 | 20:46 |
| openstack | Launchpad bug 1411465 in openstack-ansible trunk "If replication_network is the same as storage_network swift will fail" [Medium,Fix committed] - Assigned to Andy McCrae (andrew-mccrae) | 20:46 |
| odyssey4me | although, from the commit message it would appear that it's supposed to ignore the replication_network if it's the same as the storage_network | 20:47 |
| odyssey4me | maybe there is a bug there after all - try removing the replication_network and seeing | 20:48 |
| daneyon | I see the following error from Kibana OpenStack API Response (MS) after deploying a v10 cluster: × Oops! FacetPhaseExecutionException[Facet [7]: (value) field [httptime] not found] Anyone see this before? | 20:49 |
| odyssey4me | sigmavirus24 it likely is setuptools - git-harry has been poking through a problem today around our versions of pip and setuptools for juno/icehouse | 20:49 |
| BjoernT | ok, i try removing the setting | 20:49 |
| odyssey4me | daneyon yes - that would seem that logstash is not working | 20:49 |
| odyssey4me | daneyon logstash processes the logs from all the nodes and creates indexes in elasticsearch - one of the fields it creates for some openstack services is 'httptime' (response time) | 20:50 |
| miguelgrinberg | docaedo, BjoernT: feel free to hit me with questions about the solutions stuff. It should be fully working for both juno and master. | 20:50 |
| BjoernT | ok | 20:50 |
| miguelgrinberg | but master uses the new ansible-galaxy style, we haven't done the docs for that yet | 20:51 |
| odyssey4me | if there is no such value, then either logstash hasn't yet gotten to process an openstack log yet, or elasticsearch has crashed (which happens on an AIO, especially after it's been up for a few days - especially if it's out of disk space) | 20:51 |
| daneyon | odyssey4me: I am running logstash on my deploy node. The logstash container is running, the service is running inside the container. i do not see any logs in the logstash log | 20:51 |
| sigmavirus24 | odyssey4me: actually it's pip | 20:52 |
| * sigmavirus24 just talked to the pip maintainer | 20:52 | |
| odyssey4me | daneyon if you set kibana to look for logs over as long a period as possible, does it show any log entries being processed? | 20:52 |
| daneyon | odyssey4me: let me try | 20:52 |
| cloudnull | when you guys get a chance this is the review to add specs / docs https://review.openstack.org/#/c/163960/ | 20:53 |
| odyssey4me | sigmavirus24 ok - so that log's from juno, and in juno we use the get-pip script from the rax mirror | 20:53 |
| sigmavirus24 | odyssey4me: yeah it's for a good reaso | 20:53 |
| sigmavirus24 | we'll have time for that to go away. It's probably a year+ deprecation warning fwiw | 20:54 |
| sigmavirus24 | (requests and pip have very similar methodologies) | 20:54 |
| odyssey4me | cloudnull awesome, +1 for sure - just to get it attention, of course :) | 20:54 |
| daneyon | odyssey4me: i set the duration to 30 days and still the same oops message | 20:54 |
| odyssey4me | daneyon sure, but is it showing events in the histogram for events? | 20:54 |
| daneyon | odyssey4me: yes | 20:55 |
| odyssey4me | sigmavirus24 cloudnull ok, but a question has been raised recently in our team - why do we use the get-pip script from the rax mirror instead of the same approach as master (from the source), and could we now uncap setuptools for juno/icehouse? | 20:56 |
| daneyon | odyssey4me: i see info for raw events, severity, top 5 event sources and infra events | 20:56 |
| odyssey4me | daneyon ok, are there events consistantly coming through, or are they old? | 20:56 |
| sigmavirus24 | odyssey4me: I frankly don't know | 20:57 |
| odyssey4me | daneyon is your kibana available publically - feel free to prv msg me the URL and creds, if you're happy to do that | 20:57 |
| cloudnull | odyssey4me, we cant uncap setuptools in juno / icehouse until we move to 2014.2.2 and 2014.1.4 respectively . | 20:57 |
| odyssey4me | (it's been a while since I looked at logging :)) | 20:57 |
| daneyon | odyssey4me: consistently coming through | 20:57 |
| cloudnull | there is a requirements logic bomb | 20:57 |
| cloudnull | in the earlier juno icehouse releases from upstream sources. | 20:58 |
| odyssey4me | cloudnull it seems that we're hitting that requirements logic bomb on anything other than the deployment node - well, from git-harry's testing | 20:58 |
| *** Mudpuppy_ is now known as Mudpuppy | 20:58 | |
| cloudnull | if git-harry installed pip before hand and didnt allow our deployment process to do it he got pip6 | 20:58 |
| cloudnull | which will never work . | 20:59 |
| odyssey4me | the aio is passing because it is the deployment node too - so pip and setuptools is at the right level | 20:59 |
| cloudnull | thats the same issue bgmccollum ran into | 20:59 |
| odyssey4me | cloudnull he's working on the external CI jenkins, which I believe he uncovered is deploying pip in a different process - so that is likely the issue | 20:59 |
| daneyon | As part of my eval write-up/demo, I would like to show the logging but the error doesn't look so good. So, do I have to run trunk for the logging error to go away? | 20:59 |
| odyssey4me | daneyon nope, the logging does work - there's likely something simple/silly that went wrong | 21:00 |
| daneyon | odyssey4me: OK | 21:00 |
| cloudnull | odyssey4me yea if we update to 2014.2.2 and 2014.1.4 (when released) those issues should go away. upstream openstack fixed their logic issues in requirements. | 21:01 |
| odyssey4me | daneyon in the absence of access to your system, let me fire up an AIO | 21:01 |
| BjoernT | odyssey4me: did we file a documentation bug for 10.1.2 ? | 21:01 |
| daneyon | odyssey4me: I can fire-up a webex, share my screen and even have you take control | 21:01 |
| odyssey4me | cloudnull quite honestly I think the juno branch should update to the latest tag | 21:01 |
| cloudnull | me too | 21:02 |
| odyssey4me | daneyon happy to help if you're happy to do so :) | 21:02 |
| daneyon | odyssey4me: give me 2 min | 21:02 |
| cloudnull | same with icehouse, while were still make fixes in it. | 21:02 |
| odyssey4me | cloudnull icehouse should perhaps not, but we are due to tag 9.0.7 which has some essential fixes (pinning especially), so perhaps the update to the latest openstack tag would be good too | 21:02 |
| cloudnull | but as they're rax technical debt thats a bit of a harder negotiation. | 21:02 |
| odyssey4me | BjoernT doc bug for what? the replication network? no idea - if you don't see it, then you know the drill :) | 21:03 |
| cloudnull | i'd drop 9.0.7 with the couple fixes. and then drop 9.1.0 with the updated stable tags. | 21:03 |
| BjoernT | yes | 21:03 |
| odyssey4me | haha, how am I typing so quickly.... I'm supposed to be relaxing for the evening :p | 21:04 |
| cloudnull | because icehouse as is , is a bunch of disjointed shas because early icehouse was kinda broken from upstream . | 21:04 |
| cloudnull | but the tags "should"-tm be better :) | 21:04 |
| daneyon | odyssey4me: go to http://cisco.webex.com meeting# 209 879 429 meeting pswd: osad123 | 21:04 |
| odyssey4me | daneyon for future - it's probably best to have done that privately | 21:04 |
| odyssey4me | this is a logged channel, after all | 21:05 |
| daneyon | odyssey4me: my bad | 21:05 |
| odyssey4me | in fact, I'd suggest cancelling that right now and setting up a seperate session/password and sending it to me privately | 21:05 |
| daneyon | odyssey4me: OK | 21:05 |
| odyssey4me | before you end up with a troll :p | 21:05 |
| stevelle | paging sigmavirus24? | 21:05 |
| * odyssey4me looks at sigmavirus24 ;) | 21:06 | |
| odyssey4me | hahaha | 21:06 |
| * sigmavirus24 shakes head | 21:06 | |
| sigmavirus24 | o-sad | 21:06 |
| *** erikmwilson has joined #openstack-ansible | 21:20 | |
| *** erikmwilson has quit IRC | 21:23 | |
| *** erikmwilson_ is now known as erikmwilson | 21:23 | |
| *** KLevenstein has quit IRC | 21:24 | |
| *** KLevenstein has joined #openstack-ansible | 21:24 | |
| BjoernT | cloudnull: what was the syntax to specify multiple networks in the rpc_user_config for one network type. I thought it was comma separated if not a range | 21:26 |
| openstackgerrit | Steve Lewis proposed stackforge/os-ansible-deployment: Fixes config name for Cinder default availability zone https://review.openstack.org/163983 | 21:29 |
| *** jwagner is now known as jwagner_away | 21:30 | |
| cloudnull | BjoernT huh? | 21:31 |
| cloudnull | do you want to specify multiple used_ips ? | 21:31 |
| cloudnull | or more provider networks ? | 21:31 |
| BjoernT | multiple provider vlans | 21:32 |
| BjoernT | for one network type | 21:32 |
| BjoernT | looks like we have to repeat each -network stanza | 21:32 |
| BjoernT | I refer to type: "vlan" | 21:33 |
| BjoernT | range: "752" | 21:33 |
| * cloudnull looking | 21:33 | |
| BjoernT | comma separated did not work | 21:34 |
| cloudnull | yup multiple stanzas | 21:35 |
| cloudnull | its doing a merge down . | 21:35 |
| cloudnull | and will append the additional ranges | 21:35 |
| BjoernT | ok | 21:35 |
| openstackgerrit | Steve Lewis proposed stackforge/os-ansible-deployment: Fix config name for Cinder default availability zone https://review.openstack.org/163983 | 21:38 |
| docaedo | apologies for my ansible ignorance - finally circled back to try adding "solutions" tab to the all in one deployment and ansible is looking for an inventory file, but since this is all in one I don't think it's needed | 21:52 |
| docaedo | (which means maybe I'm just missing an environment var somewhere?) | 21:53 |
| docaedo | # ansible-playbook playbooks/horizon_extensions.yml | 21:53 |
| docaedo | ERROR: Unable to find an inventory file, specify one with -i ? | 21:53 |
| *** BjoernT has left #openstack-ansible | 21:54 | |
| bgmccollum | docaedo, make sure you are in /opt/rpc-extras/rpc_deployment | 21:57 |
| docaedo | I started with the "run-aio-build.sh" script (https://github.com/stackforge/os-ansible-deployment/blob/master/development-stack.rst), which checked out the repo into /opt/stackforge/os-ansible-deploy | 21:58 |
| docaedo | Then the docs for the solutions tab (http://docs.rackspace.com/rpc/api/v10/bk-rpc-installation/content/sec-install-solutions.html) tell me to pull the rpc-extras repo down into /opt/rpc-extras (http://docs.rackspace.com/rpc/api/v10/bk-rpc-installation/content/sec-install-solutions.html) | 21:59 |
| *** sigmavirus24 is now known as sigmavirus24_awa | 22:02 | |
| bgmccollum | docaedo: then from `/opt/rpc-extras/rpc_deployment`, run `ansible-playbook playbooks/horizon_extensions.yml` | 22:02 |
| docaedo | ah I suspect I checked out master instead of juno branch (and rpc_deployment wasn't there) | 22:03 |
| cloudnull | this is the review that adds docs / specs repos . please review it as you get a chance https://review.openstack.org/#/c/163960/ | 22:07 |
| Apsu | docaedo: Make sure you're in the playbooks directory before you run ansible-playbook | 22:10 |
| Apsu | docaedo: For master on both os-a-d and rpc-extras | 22:10 |
| Apsu | The other branches expect you to be in rpc_deployment (hence the docs refs) | 22:10 |
| Apsu | Master isn't a release yet, and those changes aren't in the public product docs :) | 22:10 |
| docaedo | Understood, thanks for your help! Soon as I get my VPN connection back I'll try again | 22:12 |
| Apsu | Sure thing! | 22:13 |
| odyssey4me | docaedo if you're on master, try using 'openstack-ansible' from the playbooks directory :) | 22:16 |
| odyssey4me | 'openstack-ansible' is a convenience script deployed by the 'bootstrap-ansible' script which takes care of the /etc/ bits for you - all you have to do is make sure that you run it from the place where the dynamic inventory/playbooks/roles sit | 22:17 |
| Apsu | (which is $gitCloneDir/playbooks/) | 22:18 |
| *** logan2 has quit IRC | 22:26 | |
| *** vmtrooper has joined #openstack-ansible | 22:28 | |
| *** KLevenstein has quit IRC | 22:30 | |
| bgmccollum | odyssey4me: thoughts on how to handle templating the memcached_servers? | 22:31 |
| bgmccollum | maybe via a lookup | 22:32 |
| *** KLevenstein has joined #openstack-ansible | 22:32 | |
| *** KLevenstein has quit IRC | 22:33 | |
| *** vmtrooper has quit IRC | 22:33 | |
| odyssey4me | bgmccollum I haven't looked that far into the details - it might require setting facts in plays or using other means to do the lookups ahead of time, but I don't see that as too hard to resolve | 22:40 |
| bgmccollum | odyssey4me: like this maybe, memcached_servers: "{{ lookup('template', 'memcached_servers.j2') }}" | 22:41 |
| odyssey4me | I am, however, intrigued at the idea of using an ansible module instead of doing a dict merge via jina | 22:41 |
| odyssey4me | *jinja | 22:41 |
| bgmccollum | with the jinja bits in the template to format it correctly | 22:42 |
| odyssey4me | bgmccollum yup, that's an option - although it may not be necessary... it'll be hard to know until we get an example worked up | 22:46 |
| odyssey4me | you're welcome to take a crack at it if you want to :) | 22:46 |
| *** sdake has quit IRC | 23:00 | |
| *** Mudpuppy has quit IRC | 23:20 | |
| *** jmccrory has quit IRC | 23:43 | |
Generated by irclog2html.py 2.14.0 by Marius Gedminas - find it at mg.pov.lt!