| *** shashank_hegde has quit IRC | 00:04 | |
| *** cleverdevil has quit IRC | 00:39 | |
| *** stanchan has quit IRC | 00:56 | |
| *** shashank_hegde has joined #akanda | 01:13 | |
| *** shashank_hegde has quit IRC | 04:04 | |
| *** BharatK_ has joined #akanda | 05:21 | |
| *** BharatK_ has quit IRC | 05:22 | |
| *** BharatK_ has joined #akanda | 05:23 | |
| *** BharatK has quit IRC | 05:24 | |
| *** BharatK_ has quit IRC | 05:24 | |
| *** BharatK_ has joined #akanda | 05:25 | |
| *** BharatK_ has quit IRC | 05:26 | |
| *** BharatK_ has joined #akanda | 05:27 | |
| *** BharatK_ has quit IRC | 05:28 | |
| *** BharatK_ has joined #akanda | 05:29 | |
| *** BharatK_ has quit IRC | 05:34 | |
| *** shashank_hegde has joined #akanda | 05:35 | |
| *** shashank_hegde has quit IRC | 05:42 | |
| *** stanchan has joined #akanda | 05:48 | |
| *** stanchan has quit IRC | 05:55 | |
| *** shashank_hegde has joined #akanda | 06:05 | |
| *** shashank_hegde has quit IRC | 06:10 | |
| strayandy | adam_g Sorry, missed this conv. I had to work elsewhere. | 06:34 |
|---|---|---|
| *** strayandy is now known as mandoonandy | 06:37 | |
| *** shashank_hegde has joined #akanda | 06:39 | |
| *** shashank_hegde has quit IRC | 06:57 | |
| *** openstackgerrit has quit IRC | 07:46 | |
| *** openstackgerrit has joined #akanda | 07:46 | |
| *** ronis has joined #akanda | 08:27 | |
| *** ronis has quit IRC | 08:30 | |
| *** openstackgerrit has quit IRC | 08:31 | |
| *** openstackgerrit has joined #akanda | 08:31 | |
| *** ronis has joined #akanda | 11:53 | |
| *** ronis has quit IRC | 12:42 | |
| *** ronis has joined #akanda | 12:47 | |
| *** ronis has quit IRC | 15:58 | |
| *** shashank_hegde has joined #akanda | 17:18 | |
| *** ronis has joined #akanda | 17:34 | |
| *** shashank_hegde has quit IRC | 17:38 | |
| *** shashank_hegde has joined #akanda | 17:43 | |
| stupidnic | Is there a different set of instructions to follow for installing Akanda now that things are moving to Astara? | 17:50 |
| stupidnic | repos, et. al. | 17:50 |
| *** cleverdevil has joined #akanda | 17:54 | |
| *** shashank_hegde has quit IRC | 18:06 | |
| adam_g | stupidnic, hi, we're in the process of getting docs updated for the rename | 18:16 |
| adam_g | stupidnic, and there is this, https://review.openstack.org/#/c/242629/ to get the devstack deployment going with new repo locations | 18:17 |
| stupidnic | Okay. let me ask you this then. I am not doing a devstack deployment but a full openstack deployment... is there any information on how to go about that? | 18:18 |
| stupidnic | I know that Mark mentioned that there were playbooks available but I haven't found any usable information in them (but that could be my lack of familiarity with Ansible) | 18:19 |
| adam_g | stupidnic, so, unfortunately, there isn't. thats a big todo for our doc work this cycle. but i'd be happy to walkthrough it with you here, and maybe we can get some instructions created from that | 18:20 |
| stupidnic | adam_g: that would be great. | 18:20 |
| adam_g | stupidnic, https://etherpad.openstack.org/p/astara-installation-docs | 18:21 |
| stupidnic | adam_g: are there any limitations with how you are running neutron? | 18:25 |
| stupidnic | like OVS versus LB | 18:25 |
| stupidnic | VXLANs, etc | 18:26 |
| adam_g | stupidnic, we support both OVS, LB, NSX, VXLAN. if you are using something we dont support we'd need to add an interface driver to astara in order to plug it into the management network | 18:26 |
| stupidnic | Alright. | 18:26 |
| stupidnic | adam_g: I am working through this now. | 18:43 |
| stupidnic | One question re: management_prefix | 18:43 |
| adam_g | stupidnic, cool. hope you dont mind being a guinea pig here. im doing this mostly by whats in my head and will probably be wrong/incomplete in some spots | 18:44 |
| stupidnic | is that the neutron network that the VXLANs will work on? Or is this another network all together | 18:44 |
| stupidnic | no problem. This is much better than what we have now which is nothing | 18:44 |
| adam_g | stupidnic, the management network is a specific neutron network+subnet you/admin creates that the astara service uses to reach the appliance VMs | 18:45 |
| adam_g | stupidnic, so if you're using VXLAN L2, it'd be using that as well. its just a normal neutron network in that regard, except that astara will create an interface into it from the control node to be able to reach the appliance instances | 18:46 |
| stupidnic | Ah okay. That's the part I was missing. I wasn't sure if I had to configure it on the controller or not | 18:47 |
| stupidnic | So you are saying that it will hook into neutron and create the network itself and then use VXLAN + LB to communicate with that network | 18:47 |
| adam_g | stupidnic, the interface driver is meant to take care of doing that for you on startup of the service | 18:47 |
| stupidnic | Just making sure I understand all the moving parts here. So thanks for the clarification. | 18:48 |
| adam_g | stupidnic, well, astara doesn't create the actual network in neutron, you need to do that.. but the interface driver takes care of bringin up the port required to tap into that | 18:48 |
| stupidnic | hmm okay... that might be something we need to define in the instructions as well | 18:48 |
| adam_g | stupidnic, sure | 18:49 |
| adam_g | stupidnic, just curious, are you going to be using some config management tool to role this out eventually? it'd be great to distill some of this into a puppet/chef/ansible module | 18:49 |
| adam_g | in fact, the openstack-ansible team are looking for a volunteer to do just that | 18:50 |
| stupidnic | Yes. Our entire OS deployment is defined in SaltStack | 18:50 |
| adam_g | cool | 18:52 |
| *** shashank_hegde has joined #akanda | 18:57 | |
| *** shashank_hegde has quit IRC | 18:59 | |
| adam_g | stupidnic, i think that might be enough to get you started. apologies in advanced for anything missing or snags you might hit | 19:03 |
| adam_g | this is useful, tho, and will result in real docs. thanks for playing | 19:04 |
| stupidnic | no problem I am chugging through it now | 19:05 |
| stupidnic | I will let you know if I run into a show stopper | 19:06 |
| *** shashank_hegde has joined #akanda | 19:10 | |
| davidlenwell | adam_g: https://review.openstack.org/#/c/243782 | 19:38 |
| davidlenwell | preserving the git history was just going to be too time consuming.. mark said to just do a simple copy.. since its just docs | 19:39 |
| adam_g | davidlenwell, ya | 19:41 |
| davidlenwell | adam_g: I'm gonna do the same with the specs folder | 19:42 |
| adam_g | davidlenwell, do you have a local stash of the old repo somewhere? it'd be nice to have 'Co-authored by' tags in the commit msg for the people that did commit to it | 19:42 |
| davidlenwell | I do have it locally | 19:42 |
| davidlenwell | do you know the command off hand to get that list? | 19:42 |
| adam_g | git log docs/ | grep ^Author | 19:43 |
| davidlenwell | adam_g: https://review.openstack.org/#/c/243782/3 | 19:50 |
| adam_g | davidlenwell, cool. LGTM. gonna be a few before it can merge, waiting on infra changes to sync | 19:50 |
| davidlenwell | yeah .. I figured | 19:51 |
| davidlenwell | but I can start laying changes on top of that | 19:51 |
| adam_g | ya | 19:52 |
| davidlenwell | adam_g: here is the review for the specs folder.. https://review.openstack.org/#/c/243802/ | 19:56 |
| adam_g | davidlenwell, did you have any issues pushing those patches? | 20:30 |
| adam_g | i cant seem to push anything, getting rejected from gerrit | 20:30 |
| adam_g | remote: (W) No changes between prior commit 492f56c and new commit d3d4a0e | 20:30 |
| stupidnic | adam_g: do those changes to nova.conf need to be made to the compute nodes as well or just the controller? | 20:35 |
| adam_g | stupidnic, the controller | 20:36 |
| stupidnic | cool, just double checking | 20:36 |
| stupidnic | adam_g: I went through the entire process and went to start astara and I think this is a bug | 20:44 |
| stupidnic | So in the [router] section you can specify the flavor | 20:44 |
| stupidnic | and the default is 1, I don't have a 1 since I edited the m1.tiny, and when you do that, it drops the int and replaces it with an uuid | 20:44 |
| stupidnic | however you can't use the uuid in the instance_flavor value | 20:45 |
| stupidnic | it complains that it isn't an int() base 10 | 20:45 |
| adam_g | stupidnic, let me see | 20:46 |
| adam_g | doh yeah | 20:46 |
| adam_g | one sec ill put up a patch | 20:46 |
| stupidnic | k | 20:46 |
| stupidnic | also... would it be possible to have an option in the appliance disk-image-create to not convert to qcow2? | 20:49 |
| stupidnic | I am using Ceph and it does copy on write within Ceph itself, but to do that it needs the base image in glance to be raw | 20:50 |
| cleverdevil | that's what we do at DreamHost, so its definitely possible :) | 20:51 |
| stupidnic | I can live with it in qcow2, but figured I would mention it | 20:51 |
| cleverdevil | cool to see another person using Ceph / OpenStack and Astara! | 20:52 |
| adam_g | stupidnic, here you go https://review.openstack.org/#/c/243820/ | 20:52 |
| stupidnic | Well not yet, but close :) | 20:52 |
| adam_g | stupidnic, yeah, you can specify that to DIB | 20:52 |
| adam_g | stupidnic, pass '-t raw' | 20:52 |
| adam_g | to disk-image-create | 20:53 |
| stupidnic | adam_g: ah okay. thanks | 20:53 |
| stupidnic | adam_g: so that patch worked and I have orchestrator running. I have a router created, but I am not sure where I should be looking for it? Will it show as an instance on a hypervisor? | 21:07 |
| stupidnic | I might have another issue with my hypervisors if that is the case (ceph related I am sure) | 21:08 |
| adam_g | stupidnic, right. on startup, astara should try to create intsances in nova for existing routers | 21:09 |
| adam_g | a good check to see if things are wired up corretly on the astara/neutron side is to run the orhcestrator and watch the logs, then go create a neutron router. astara should get the notification and start doing some stuff | 21:09 |
| stupidnic | Remote error: AgentNotFoundByTypeHost Agent with agent_type=L3 agent and host=controller01 could not be found | 21:17 |
| stupidnic | where should I be looking for that? | 21:17 |
| stupidnic | l3? | 21:17 |
| adam_g | stupidnic, did you restart neutron after changing its config? | 21:19 |
| stupidnic | Yeah | 21:19 |
| adam_g | one sec | 21:19 |
| stupidnic | Do I need to update the interface_driver in l3_agent.ini? | 21:19 |
| stupidnic | I think that might be where I went wrong | 21:19 |
| adam_g | stupidnic, actually we disable the l3 agent entirely | 21:19 |
| adam_g | you can stop it | 21:20 |
| stupidnic | okay | 21:20 |
| stupidnic | any others? | 21:20 |
| stupidnic | metadata? dhcp? | 21:20 |
| adam_g | ya..dhcp, l3, metadata, vpn, lbaas | 21:20 |
| stupidnic | Okay. shut them down, let me kick astara again | 21:21 |
| stupidnic | hmm nope same error, I mean the orchestrator is still running | 21:23 |
| adam_g | hmm thats one i haven't seen. | 21:23 |
| adam_g | markmcclain, any idea ^? | 21:23 |
| adam_g | stupidnic, where is that error being logged? | 21:23 |
| stupidnic | I can provide a stack trace if you want | 21:23 |
| * markmcclain catchs up | 21:23 | |
| adam_g | sure | 21:23 |
| stupidnic | it's thrown in the console of astara-orchestration | 21:23 |
| stupidnic | http://paste.openstack.org/show/8kAwnurcnWICbOxpjG2q/ | 21:24 |
| stupidnic | Just to let you know how I got there... I saw some errors in the output about amqp being deprecated, so I swtiched them to oslo_messaging_rabbit, but that seemed to throw an error, so I backed those changes out | 21:26 |
| markmcclain | hmmm.. I haven't seen that error before | 21:26 |
| stupidnic | maybe something stuff in the queue? | 21:26 |
| stupidnic | s/stuff/stuck | 21:27 |
| *** ronis has quit IRC | 21:27 | |
| stupidnic | 2015-11-10 16:27:57.416 15534 DEBUG oslo_messaging._drivers.amqpdriver [req-2885ac64-aec2-465c-a9ec-19e637a9ae08 - - - - -] CALL msg_id: c97012e832ec40bc8f7143a9a0f5874b exchange 'neutron' topic 'q-l3-plugin' _send /usr/local/lib/python2.7/dist-packages/oslo_messaging/_drivers/amqpdriver.py:448 | 21:28 |
| stupidnic | that seems to be what the queue is passing in | 21:28 |
| adam_g | stupidnic, the errors are coming from the messaging layer because they are actually neutron errors being sent over via RPC | 21:30 |
| stupidnic | Okay. So something in my neutron is emitting this error? | 21:31 |
| davidlenwell | markmcclain: adam_g: https://review.openstack.org/243831 <-- new readme file mostly derived from the old one in openstack-attic/akanda .. I think its close ..but the name section might need some eyes.. | 21:32 |
| adam_g | stupidnic, yeah. is this happening on startup, or after creating a new router? | 21:33 |
| stupidnic | startup | 21:33 |
| stupidnic | which makes me think it is something sitting in the queue that it barfs on | 21:33 |
| adam_g | markmcclain, do routers get created with a binding to the running L3 agent? in other words, would pre-populating a router from pre-astara throw such an error? | 21:34 |
| markmcclain | possibly | 21:34 |
| markmcclain | I haven't tested switching | 21:34 |
| stupidnic | Well this isn't production, so I can tear the router out and add a new one | 21:35 |
| markmcclain | which backend? | 21:35 |
| stupidnic | linuxbridge | 21:35 |
| adam_g | stupidnic, its the call to sync_routers thats throwing it | 21:35 |
| adam_g | stupidnic, mgiht be a good idea, start fresh | 21:35 |
| stupidnic | k | 21:35 |
| stupidnic | Okay. Router deleted... when I did that it emitted the same error | 21:36 |
| stupidnic | (in the logs) | 21:36 |
| markmcclain | also make sure you've got the astara l3 plugin configured in neutron | 21:36 |
| stupidnic | that would be: service_plugins = akanda.neutron.plugins.ml2_neutron_plugin.L3RouterPlugin correct? | 21:38 |
| markmcclain | yes | 21:38 |
| stupidnic | that just goes in [DEFAULT] right? | 21:38 |
| markmcclain | yes | 21:38 |
| markmcclain | because https://github.com/openstack/astara-neutron/blob/master/akanda/neutron/plugins/ml2_neutron_plugin.py#L114 | 21:38 |
| markmcclain | works around that agent issue | 21:39 |
| stupidnic | hmm okay. That's in there. Is it possible that because I am using akanda versus astara? | 21:39 |
| stupidnic | maybe a namespacing issue? | 21:39 |
| *** shashank_hegde has quit IRC | 21:39 | |
| adam_g | stupidnic, pythhon paths havent changed yet | 21:40 |
| stupidnic | alright, figured it was the simpliest reason :() | 21:40 |
| stupidnic | :) | 21:40 |
| stupidnic | let me check my neutron logs | 21:40 |
| stupidnic | Loading Plugin: akanda.neutron.plugins.ml2_neutron_plugin.L3RouterPlugin | 21:41 |
| stupidnic | When I try to add a router, I don't see any activity in the astara log | 21:44 |
| stupidnic | a delete of the router triggers the same error again. | 21:45 |
| davidlenwell | adam_g: did you say this morning that you were in the process of fixing the devstack plugin to work with the name change.. is that in review already? | 21:45 |
| adam_g | davidlenwell, ya https://review.openstack.org/#/c/242629/ | 21:46 |
| davidlenwell | thanks | 21:46 |
| adam_g | theres a neutron change that needs to land first before it'll work tho | 21:46 |
| adam_g | see the Depends-on | 21:46 |
| davidlenwell | k | 21:47 |
| markmcclain | stupidnic: not 100% sure what's going on, but I've got to step away so wont be able to standup fresh env for a good bit | 21:50 |
| stupidnic | markmcclain: no problem. I have been staring at this so long my eyes are going fuzzy | 21:51 |
| stupidnic | I will work on some other stuff in the mean time | 21:51 |
| adam_g | markmcclain, https://review.openstack.org/#/c/242629/ looks ready | 22:44 |
| *** cleverdevil has quit IRC | 23:58 | |
Generated by irclog2html.py 2.14.0 by Marius Gedminas - find it at mg.pov.lt!