| *** puranamr_ has joined #akanda | 00:03 | |
| *** puranamr has quit IRC | 00:03 | |
| *** puranamr_ has quit IRC | 00:17 | |
| adam_g | davidlenwell, fyi looks like your editor is saving files 0755 again | 00:26 |
|---|---|---|
| openstackgerrit | Adam Gandelman proposed stackforge/akanda-rug: Debugging integranion test (DO NOT MERGE) https://review.openstack.org/221541 | 00:30 |
| *** puranamr has joined #akanda | 01:22 | |
| *** puranamr has quit IRC | 01:38 | |
| *** ryanpetrello has quit IRC | 03:29 | |
| *** adam_g has quit IRC | 03:32 | |
| *** sarob has quit IRC | 03:33 | |
| *** sarob has joined #akanda | 03:34 | |
| *** ryanpetrello has joined #akanda | 03:34 | |
| *** adam_g has joined #akanda | 03:35 | |
| *** adam_g has quit IRC | 03:35 | |
| *** adam_g has joined #akanda | 03:35 | |
| *** stanchan_ has joined #akanda | 04:47 | |
| *** stanchan_ has quit IRC | 07:27 | |
| *** stanchan has joined #akanda | 08:03 | |
| *** stanchan has quit IRC | 15:08 | |
| *** thingee has left #akanda | 15:16 | |
| *** puranamr has joined #akanda | 17:08 | |
| *** stanchan has joined #akanda | 17:15 | |
| *** stanchan has quit IRC | 17:27 | |
| openstackgerrit | Adam Gandelman proposed stackforge/akanda-rug: Debugging integration test (DO NOT MERGE) https://review.openstack.org/221541 | 17:27 |
| *** puranamr has quit IRC | 17:53 | |
| *** puranamr has joined #akanda | 18:05 | |
| *** puranamr has quit IRC | 18:10 | |
| *** puranamr has joined #akanda | 18:12 | |
| *** puranamr has quit IRC | 18:16 | |
| *** puranamr has joined #akanda | 18:23 | |
| *** puranamr has quit IRC | 18:37 | |
| *** drwahl has joined #akanda | 18:53 | |
| sarob | davidlenwell: updated router patch soon? | 18:56 |
| sarob | so adam_g can help you debug the state machine | 18:57 |
| sarob | markmcclain: i need a review of https://review.openstack.org/#/c/219564/ | 18:58 |
| markmcclain | right | 19:01 |
| adam_g | markmcclain, can you think of any issues if the instance ports we create in the make_ports_callback get attached to the instance after its been booted? | 19:06 |
| adam_g | late-binding the management port wouldnt work | 19:06 |
| drwahl | yo akanda doods: we were talking about a tenant having multiple networks. i know adding additional networks works, but wasn't sure about deleting | 19:08 |
| drwahl | i know it was an issue at one point. has that been fixed? | 19:08 |
| adam_g | drwahl, theres currently an issue where the rug will create a router and some ports (ie the management port) which is owned by the service tenant and not user tenant. when the user tries to delete network, it fails because of the active port that it doesnt own | 19:10 |
| adam_g | i have a patch up to fix it in master, im not sure if the dreamhost fork is affected | 19:10 |
| adam_g | i imagine it would be, at least for the management port. | 19:10 |
| drwahl | ya, that sounds about right | 19:10 |
| drwahl | it's been at least 12 months since i last looked at this bug | 19:10 |
| adam_g | https://bugs.launchpad.net/akanda/+bug/1491593 | 19:11 |
| openstack | Launchpad bug 1491593 in akanda "tenant created networks cannot be deleted by the tenant" [High,New] | 19:11 |
| adam_g | i actually hit it last week when i was adding your guys' tests on top of our branch | 19:11 |
| drwahl | doh! | 19:11 |
| adam_g | add tests, find bugs, fix bugs, rinse, repeat | 19:11 |
| adam_g | as of kilo we started to create VRRP ports which also get left behind and block deletion | 19:12 |
| openstackgerrit | Adam Gandelman proposed stackforge/akanda-rug: Delete VRRP and MGT ports on router delete https://review.openstack.org/219951 | 19:13 |
| adam_g | ^ | 19:13 |
| openstackgerrit | David Lenwell proposed stackforge/akanda-rug: Router driver and required changes to instance manager (WIP) https://review.openstack.org/215227 | 20:05 |
| davidlenwell | sarob: adam_g^^ | 20:05 |
| davidlenwell | adam_g: I'm in the middle of reworking the event to contain a resource object like we chatted about.. I'm going to get that in later tonight.. right now the process seems to start and work fine.. | 20:06 |
| adam_g | davidlenwell, cool ill take a look in a bit | 20:07 |
| davidlenwell | sarob adam_g, So I have to go pick up devon from school shortly .. will probably be afk till this evening | 20:07 |
| markmcclain | adam_g: none that I can think of | 20:08 |
| markmcclain | adam_g: adding the ports later is like the hot plug stuff we support and with the delete fix we should be good to go | 20:09 |
| *** puranamr has joined #akanda | 20:09 | |
| openstackgerrit | David Lenwell proposed stackforge/akanda-rug: Router driver and required changes to instance manager (WIP) https://review.openstack.org/215227 | 20:17 |
| openstackgerrit | Merged stackforge/akanda: docs update describing what is akanda https://review.openstack.org/219564 | 20:28 |
| *** puranamr has quit IRC | 20:39 | |
| adam_g | markmcclain, so a appliance booted *only* with a mgt address should get its address and be reachable via that address, right? | 20:40 |
| markmcclain | right | 20:40 |
| markmcclain | currently you can simulate that by creating a router w/o attaching any interfaces to it | 20:40 |
| markmcclain | so it would be the same | 20:41 |
| markmcclain | work flow as when the first interface is added to a router w/o any vifs | 20:41 |
| adam_g | yea | 20:42 |
| adam_g | the thing comes up okay but not reachable. wonder if im borking cloud-init somehow | 20:43 |
| adam_g | markmcclain, also, im curious what the idea was with passing the make_vrrp_ports around as a callback.. was that just to share the instance manager's neutronclient with rug.api.nova, or was there other plans ? | 20:45 |
| markmcclain | yes | 20:46 |
| adam_g | markmcclain, yes to A or yes to B? :) | 20:46 |
| adam_g | hmm so booting with only the management port and no instance ports seems to break something. | 20:55 |
| markmcclain | adam_g: sorry too many calls | 21:28 |
| adam_g | markmcclain, no worries | 21:28 |
| adam_g | im clearly overlooking something. ive hacked rug's port callback to skip creating nstance ports, and only create and bind the mgt port. router comes up with a reachable mgt address. doing the same thing in another process (pez) and its unreachable | 21:30 |
| markmcclain | so part of the idea of the callback was that the handle we had at that moment may/may not be valid also to late bind when the ports are created | 21:31 |
| markmcclain | are you using the same cloud-init config? | 21:31 |
| markmcclain | although the default is to listen on all interfaces | 21:31 |
| markmcclain | I'm not married to the callback if we want to pass the mgt port and any instance ports as individual vars or a namedtuple | 21:33 |
| adam_g | markmcclain, ill eventually at least need to split hte callback up to mgt ports and instance ports | 21:38 |
| markmcclain | right makes sense | 21:38 |
| adam_g | and yeah, everything looks the same in both cases (cloud-init, neutron ports are equiv) | 21:38 |
| adam_g | one somewhat strange thing ive noticed is that when you boot without the full set of interfaces, cloud-init doesnt set the instances hostname | 21:39 |
| adam_g | but that doesnt seem to be related to connectivity | 21:40 |
| openstackgerrit | Adam Gandelman proposed stackforge/akanda-rug: Split out instance provisioning to on demand and pez (WIP) https://review.openstack.org/221999 | 21:44 |
| markmcclain | adam_g: interesting.. there's nothing special that looks for more than 1 vif | 21:59 |
| markmcclain | so not sure why c-i is acting up | 22:00 |
| adam_g | markmcclain, wheres it actin up | 22:18 |
| openstackgerrit | Adam Gandelman proposed stackforge/akanda: Add spec for appliance node pooling https://review.openstack.org/194882 | 22:22 |
| markmcclain | adam_g: sorry cloud-init not the CI system | 22:25 |
| adam_g | oh :) | 22:25 |
| markmcclain | ok.. weekend time.. have a great weekend | 22:35 |
| adam_g | davidlenwell, so service startup doesnt get much further /w patchset 11 http://paste.ubuntu.com/12350182/ a lot of the issues i pointed out in patchset 9 are still valid | 22:36 |
| adam_g | markmcclain, cheers | 22:36 |
| *** openstackgerrit has quit IRC | 23:46 | |
| *** openstackgerrit has joined #akanda | 23:46 | |
Generated by irclog2html.py 2.14.0 by Marius Gedminas - find it at mg.pov.lt!