*** calbers has quit IRC | 01:19 | |
*** calbers has joined #openstack-chef | 01:21 | |
*** Bhujay has joined #openstack-chef | 01:25 | |
*** spheromak has quit IRC | 01:49 | |
*** Ramereth has quit IRC | 01:49 | |
*** openstack has joined #openstack-chef | 02:51 | |
*** ChanServ sets mode: +o openstack | 02:52 | |
*** openstackstatus has joined #openstack-chef | 02:59 | |
*** ChanServ sets mode: +v openstackstatus | 02:59 | |
*** mattgo has joined #openstack-chef | 07:32 | |
*** Bhujay has quit IRC | 09:10 | |
*** mattgo has quit IRC | 14:12 | |
Seb-Solon | scas, I don't remember quite well but you don't use openvswitch with openstack right? | 14:29 |
---|---|---|
Seb-Solon | I have my annoying bug again, considering other option atm | 14:29 |
scas | Seb-Solon: currently my only use of openvswitch is what is upstream in openstack/ | 15:02 |
scas | i have some local patches for calico, but they're not in a place to be released to the wild yet | 15:05 |
scas | i've been looking at tungsten fabric fka opencontrail as well | 15:06 |
scas | the general idea is to keep consistent with what docs.o.o provides | 15:06 |
scas | the latter is probably more effort than keeping ovs working | 15:08 |
scas | Seb-Solon: if you could refresh my memory on what your annoying bug is, i've not found any lp bugs except the mariadb one and logs are spotty | 15:55 |
*** mattgo has joined #openstack-chef | 16:12 | |
Seb-Solon | my bug is more or less : when the packet is bigger thatn the MTu and Comes from my public IP Openstack shink a part of it. In all other case it is fine. | 17:03 |
Seb-Solon | When I reboot the openstack server, it vanishes | 17:04 |
Seb-Solon | The bug comes back within a day | 17:04 |
Seb-Solon | my public IP redirect port XXX to a VM in openstack on that port basically | 17:05 |
*** Ramereth has joined #openstack-chef | 17:05 | |
Seb-Solon | when using the public IP it fails, when going directly to the VM it works. The kind of bug that drive you a bit crazy :) | 17:06 |
Seb-Solon | My workaround is to use the internal IP for now when I have time to debug | 17:06 |
Ramereth | jklare: are you the effective maintainer of the openstackclient cookbook? | 17:09 |
Seb-Solon | By the was the openvswitch docs mentions "Linux distributions often package older releases of Open vSwitch that can introduce issues during operation with the Networking service. We recommend using at least the latest long-term stable (LTS) release of Open vSwitch for the best experience and support from Open vSwitch" | 17:11 |
Seb-Solon | from : https://docs.openstack.org/liberty/networking-guide/scenario-classic-ovs.html | 17:11 |
Seb-Solon | maybe I should do that | 17:12 |
Seb-Solon | the chef use package : *** 2.8.4-0ubuntu0.17.10.1 http://ubuntu-cloud.archive.canonical.com/ubuntu xenial-proposed/pike/main amd64 Packages | 17:12 |
scas | Ramereth: openstackclient is now in the openstack/ space, which falls to me, more or less | 17:13 |
Ramereth | scas: ah ok, it looks like an update to fog-openstack is breaking it now | 17:13 |
Ramereth | (at least for me) | 17:13 |
Ramereth | 0.2.0 was just released yesterday and since that cookbook pulls it in via metadata.rb, it gets updated to the latest release | 17:14 |
scas | Ramereth: yeah, it free-floats versions | 17:14 |
Ramereth | I'm not sure what the best workaround is | 17:15 |
scas | i'll look at it and see if the cookbook needs an update, or if it's a matter of pinning | 17:15 |
Ramereth | It seems using gem in metadata.rb provides some magic that you can't easily override | 17:16 |
scas | i've been more in writing mode and not on the code as of late | 17:16 |
Ramereth | understandable. I'm at a conference this week. I may just implement a temporary workaround for us but can look into a longer term solution if you don't have time | 17:18 |
scas | it's chef_gem under the hood, so it *should* accept the same properties | 17:18 |
Ramereth | however, I probably should create an issue so others hitting it can see they aren't alone | 17:18 |
Ramereth | ya, but it runs before other cookbooks/recipes get run. Even if you use edit_resource it doesn't seem to honor it | 17:19 |
scas | right. i'm thinking it'd be a blob in metadata.rb if we have to version pin | 17:21 |
scas | gem 'fog-openstack' do | 17:21 |
scas | version '0.1.26' | 17:21 |
scas | end | 17:21 |
scas | or something like that totally untested off the cuff thought | 17:21 |
Ramereth | If I do gem 'fog-openstack', '< 0.2.0' on a newly installed system it works fine. However if you already have 0.2.0 that breaks | 17:21 |
scas | i haven't seen it because i'm using chefdk for development, and at that only the stable track | 17:23 |
scas | i did, however, shoehorn it in and kick an integration build locally | 17:24 |
scas | fog-openstack in chefdk 3.1 is at 0.1.26 | 17:25 |
scas | nobody ever updates changelogs anymore | 17:27 |
Ramereth | nope | 17:27 |
Ramereth | it seems the problem I'm running into actually happens in openstack-identity. I see this and then the error: | 17:27 |
Ramereth | * openstack_domain[default] action create[fog][DEPRECATION] An authentication URL including a version is deprecated | 17:27 |
scas | aha! | 17:28 |
scas | https://github.com/fog/fog-openstack/commit/d92fbe0408820d52824fa3a0c4ca7041439aac91 | 17:28 |
scas | it's trying to account for keystone versions now | 17:28 |
scas | it may be a me-thing, might be the endpoint tacking on /v3 | 17:29 |
Ramereth | here's the full error: https://pastebin.osuosl.org/852446/ | 17:29 |
Ramereth | yup | 17:30 |
scas | :path => "/v3/v3/domains" | 17:30 |
Ramereth | :path => "/v3/v3/domains" | 17:30 |
scas | ding | 17:30 |
scas | cookbook-openstack-identity/attributes/default.rb: default['openstack']['endpoints'][ep_type]['identity']['path'] = '/v3' | 17:31 |
scas | gut feeling is that it's chef openstack that needs to adapt | 17:32 |
Ramereth | ya | 17:33 |
scas | what release are you on? | 17:33 |
Ramereth | newton -.- (been trying to find time to do ocata and then pike) | 17:33 |
scas | okay. i'm not sure if i can actually merge a fix upstream for it, but can at least validate a theory | 17:33 |
Ramereth | I've been pushing any changes I need into our fork branches for newton | 17:34 |
scas | if anywhere, it'll go in master, possibly queens since that still works in ci | 17:34 |
Ramereth | ya, if you fix anything, I'll try and back port it into our branch | 17:34 |
Ramereth | bbl, I'll check in later. Thanks for the help! | 17:41 |
scas | looks like it's a keystone-paste problem, too. that goes away in stein | 18:19 |
*** mgagne has joined #openstack-chef | 20:21 | |
*** mattgo has quit IRC | 23:01 |
Generated by irclog2html.py 2.15.3 by Marius Gedminas - find it at mg.pov.lt!