adam_g | gate-functional-dsvm-akandaSUCCESS in 32m 06s | 00:38 |
---|---|---|
*** elo has quit IRC | 00:45 | |
davidlenwell | sweet! | 00:46 |
sarob | awesome | 01:08 |
*** elo has joined #akanda | 01:42 | |
*** elo has quit IRC | 01:43 | |
*** elo has joined #akanda | 06:28 | |
*** elo has quit IRC | 07:22 | |
*** elo has joined #akanda | 07:32 | |
*** elo has quit IRC | 08:23 | |
ryanpetrello | found an akanda-rug bug from a customer | 13:08 |
ryanpetrello | https://bugs.launchpad.net/akanda/+bug/1496848 | 13:08 |
openstack | Launchpad bug 1496848 in akanda "Router hostnames need validation" [Undecided,New] - Assigned to Ryan Petrello (ryan-petrello) | 13:08 |
ryanpetrello | working on a patch on our end and one for upstream | 13:08 |
*** openstackgerrit has quit IRC | 13:31 | |
*** openstackgerrit has joined #akanda | 13:31 | |
rods | adam_g you around? | 14:29 |
*** elo has joined #akanda | 15:15 | |
*** elo has quit IRC | 15:16 | |
*** elo has joined #akanda | 15:18 | |
*** puranamr has joined #akanda | 15:25 | |
markmcclain | ryanpetrello: what kind of problem have you see with the hostname? | 15:40 |
markmcclain | I'm guessing teh tenant changed theirs? | 15:40 |
ryanpetrello | name your router: marks_router | 15:40 |
ryanpetrello | yep | 15:40 |
markmcclain | ah... ok | 15:40 |
ryanpetrello | I'm going to change the code to just normalize it to ak-<tenant> | 15:40 |
ryanpetrello | if that makes sense to you | 15:40 |
ryanpetrello | and not actually use the router name | 15:41 |
markmcclain | yeah.. the uuid makes sense | 15:41 |
ryanpetrello | we can rely on that value to be hostname-valid | 15:41 |
ryanpetrello | k | 15:41 |
ryanpetrello | working on a review for our repo and will open an openstack.org review | 15:41 |
markmcclain | it's not like the hostname is visible to anyone except admins | 15:41 |
markmcclain | k | 15:41 |
ryanpetrello | right | 15:42 |
ryanpetrello | we really only added it for convenience when debugging a router | 15:42 |
markmcclain | yeah | 15:42 |
ryanpetrello | so did the neutron pecan merge not happen? | 15:44 |
*** elo has quit IRC | 15:48 | |
*** puranamr has quit IRC | 15:51 | |
ryanpetrello | man, the unit tests on master take a *really* long time to run | 16:12 |
openstackgerrit | Ryan Petrello proposed stackforge/akanda-rug: Use a standard hostname to prevent setting invalid hostnames. https://review.openstack.org/224808 | 16:14 |
*** stanchan has joined #akanda | 16:33 | |
ryanpetrello | markmcclain adam_g so the akanda-rug tests for me at HEAD are currently failing :\ | 16:35 |
ryanpetrello | are you all aware of this? | 16:36 |
markmcclain | ryanpetrello: no.. style, unit or devstack? | 16:37 |
ryanpetrello | tox -epy27 | 16:37 |
ryanpetrello | I'm running on OSX, so some are related to that | 16:37 |
ryanpetrello | but some don't seem related | 16:37 |
ryanpetrello | e.g., a bunch of | 16:37 |
ryanpetrello | >> raise AssertionError('Expected call: %s\nNot called' % ("run(['--debug', 'tenant', 'manage', 'TENANT123'])",)) | 16:37 |
ryanpetrello | testing on Linux to be sure, but some of these don't look platform-related | 16:37 |
markmcclain | yeah.. there is a linux specific test that really should be function rather than unit | 16:45 |
ryanpetrello | hrm | 16:45 |
ryanpetrello | yea | 16:45 |
ryanpetrello | I see it | 16:45 |
ryanpetrello | running it a few times in Linux to make sure it's not some weird race | 16:45 |
ryanpetrello | if it doesn't work in OSX, that doesn't bother me | 16:46 |
markmcclain | ok... so checked out a fresh head in a new virtualenv and everything worked | 16:47 |
ryanpetrello | it's not like I have a shortage of Linux environments I can work in | 16:47 |
ryanpetrello | yea, same | 16:47 |
markmcclain | haha | 16:47 |
*** elo has joined #akanda | 17:12 | |
drwahl__ | hey yall | 19:31 |
drwahl__ | quick question | 19:31 |
drwahl__ | is the version of akanda tracking/pinned to the version of openstack? | 19:31 |
drwahl__ | we're trying to get some versioning numbers/names on our akanda appliance builds | 19:32 |
drwahl__ | and wondering if there will simply be an "akanda-juno" appliance, and an "akanda-kilo" appliance | 19:32 |
drwahl__ | or is the version going to be disassociated with the OpenStack version? i.e. akanda-1.0, akanda-2.0 | 19:33 |
*** drwahl__ is now known as drwahl | 19:38 | |
markmcclain | drwahl_: so the appliance (and the rug) will track the neutron version numbers | 19:40 |
markmcclain | so 7.0.0 is liberty | 19:40 |
markmcclain | 2015.1.x is kilo etc | 19:40 |
drwahl | thx | 19:42 |
ryanpetrello | adam_g: re: that ticket | 19:59 |
ryanpetrello | my reasoning for ak-<tenant-id> | 19:59 |
ryanpetrello | the reason is that we generally use this for logging into a routing and quickly knowing "who the hell is this?" | 19:59 |
*** elo has quit IRC | 20:06 | |
ryanpetrello | s/that ticket/https://review.openstack.org/#/c/224808/1 | 20:06 |
*** elo has joined #akanda | 20:06 | |
adam_g | ryanpetrello, oh wait im dumb | 20:06 |
ryanpetrello | http://reactiongif.org/wp-content/uploads/GIF/2015/06/I-dunno-man-GIF.gif | 20:06 |
adam_g | ryanpetrello, i read the patch wrong. thats only setting the hostname, i thought you were naming the nova instances themselves ak-$tenant_id | 20:07 |
ryanpetrello | right | 20:07 |
ryanpetrello | just setting the hostname on the VM | 20:07 |
ryanpetrello | we use this when we log in as a gut-check | 20:07 |
ryanpetrello | just to make sure we're working on the tenant we think we are | 20:07 |
adam_g | ryanpetrello, makes sense. +2 | 20:07 |
ryanpetrello | the issue is that a customer of ours decided to name their router some crazy name | 20:07 |
ryanpetrello | cool, hx | 20:07 |
ryanpetrello | *thx | 20:07 |
ryanpetrello | what should I backport it to? | 20:08 |
ryanpetrello | Kilo? | 20:08 |
*** stanchan has quit IRC | 20:08 | |
*** stanchan has joined #akanda | 20:09 | |
adam_g | ryanpetrello, both kilo and juno i guess | 20:09 |
ryanpetrello | k, wfm | 20:09 |
openstackgerrit | Adam Gandelman proposed stackforge/akanda-rug: Various fixes to get drivers functional (WIP) https://review.openstack.org/223884 | 20:10 |
*** elo has quit IRC | 20:11 | |
adam_g | davidlenwell, markmcclain okay so i think that hsould be good to start building the LBAAS thing on top of. i moved some more stuff into the drivers, stuff that will probably be functionally different with other appliances | 20:11 |
adam_g | specifically the stuff that uses akanda.rug.api.akanda_client now, which ATM is effectively a client library to the akanda router | 20:11 |
markmcclain | adam_g: makes sense | 20:12 |
davidlenwell | sweet.. I've been working out a list of notifications lbaas will need to watch for.. will start building the lbaas driver tonight | 20:12 |
adam_g | cool | 20:12 |
adam_g | i need to prep for this panel tonight. ill see what i can do to help with LB stuff tomorrow | 20:13 |
*** elo has joined #akanda | 20:31 | |
*** elo has quit IRC | 21:06 | |
sarob | i have landed | 21:11 |
sarob | dum duh dum | 21:11 |
openstackgerrit | Sean Roberts proposed stackforge/akanda: installation documentation update https://review.openstack.org/224909 | 21:33 |
openstackgerrit | Sean Roberts proposed stackforge/akanda: installation documentation update [WIP] https://review.openstack.org/224909 | 21:34 |
*** elo has joined #akanda | 21:52 | |
*** elo has quit IRC | 23:11 | |
adam_g | davidlenwell, markmcclain i think we're gonna need a neutron-lbaas akanda agent driver | 23:16 |
adam_g | the event notifications themselves dont give us enough context to work with | 23:16 |
adam_g | ie, | 23:16 |
adam_g | event_type: loadbalancer.create.start | 23:16 |
adam_g | payload: {u'loadbalancer': {u'vip_subnet_id': u'8c58b558-be54-45de-9873-169fe845bb80', u'admin_state_up': True}} | 23:16 |
adam_g | is all we get on a create | 23:16 |
davidlenwell | is that not enough information? | 23:18 |
davidlenwell | whats it missing? | 23:18 |
*** stanchan has quit IRC | 23:18 | |
adam_g | davidlenwell, the loadbalancer id, for one | 23:19 |
davidlenwell | hmm | 23:19 |
*** elo has joined #akanda | 23:19 | |
davidlenwell | wouldn't it not have an id until after its created? | 23:20 |
adam_g | should be available | 23:21 |
adam_g | we could probably do something like this https://git.openstack.org/cgit/openstack/neutron-lbaas/tree/neutron_lbaas/drivers/octavia/driver.py#n158 | 23:21 |
adam_g | except send out our own notifications on the neutron side with available context via RPC instead of REST'ing it to akanda | 23:22 |
Generated by irclog2html.py 2.14.0 by Marius Gedminas - find it at mg.pov.lt!