| openstackgerrit | Monty Taylor proposed openstack-infra/shade: DNM Testing a devstack revert patch https://review.openstack.org/417925 | 00:25 |
|---|---|---|
| morgan | mordred: how are we looking at the devstack thing thatn sc68cal is working on? if you can point me at the items i can continue with my other work for shade. | 00:30 |
| mordred | morgan: we MAY have an answer | 00:30 |
| morgan | ah | 00:30 |
| mordred | morgan: (we were looking at it in #openstack-qa) | 00:30 |
| morgan | ahh let me join there | 00:31 |
| mordred | morgan: https://review.openstack.org/#/c/417879/ | 00:31 |
| morgan | cool looking | 00:31 |
| mordred | morgan: (and Shrews) shade gate issues were caused by openstacksdk breaking openstackclient and doing $something with quota setting and network setup in devstack | 02:01 |
| mordred | I'm still baffled that we (and rally) seem to have been the only thing broken | 02:01 |
| openstackgerrit | Merged openstack-infra/shade: Remove link to modindex https://review.openstack.org/415485 | 02:22 |
| morgan | mordred: probably because we're the only ones with comprehensive tests that mirror real-world cases (and a lot of edge-cases) | 02:29 |
| morgan | mordred: unit and functional are usable in the other projects but dont really cover extended project interactions like rally and shade do | 02:30 |
| mordred | morgan: good point | 02:41 |
| morgan | Maybe Shade and Rally should be the cross-project gate</only somewhat sarcastic about bringing back massive co-gating experience> | 02:45 |
| openstackgerrit | Merged openstack-infra/shade: Add support for using the default subnetpool https://review.openstack.org/416319 | 03:05 |
| openstackgerrit | Monty Taylor proposed openstack/os-client-config: Swap the order of interface and endpoint_type https://review.openstack.org/418192 | 03:19 |
| dfflanders | fyi mordred, just pinged you on twitter re oaktree support at PTG | 03:41 |
| mordred | dfflanders: yes! it was a travel day ... will get back to you tomorrow when I have a brain | 05:10 |
| *** yfried has joined #openstack-shade | 07:32 | |
| *** pabelanger has quit IRC | 09:35 | |
| *** pabelanger has joined #openstack-shade | 09:35 | |
| *** stevemar has quit IRC | 09:37 | |
| *** sigmavirus has quit IRC | 09:37 | |
| *** stevemar has joined #openstack-shade | 09:37 | |
| *** openstack has joined #openstack-shade | 14:26 | |
| frickler | mordred: seems that shade always does the discovery here http://git.openstack.org/cgit/openstack-infra/shade/tree/shade/openstackcloud.py#n403 and then uses the broken URL it receives :( | 14:30 |
| mordred | frickler: it's not intentional to not be documented - sorry about that | 14:32 |
| mordred | frickler: looking at the other pastes | 14:32 |
| mordred | frickler: oh! hrm. you're right - that function should honor image_endpoint_override. :( | 14:34 |
| mordred | frickler: k. lemme make you a couple of patches - sorry for the trouble, but thanks for pointing it out | 14:34 |
| frickler | mordred: no need to apologize, thanks for your support | 14:36 |
| openstackgerrit | Monty Taylor proposed openstack-infra/shade: Log request ids when debug logging is enabled https://review.openstack.org/417931 | 14:37 |
| openstackgerrit | Merged openstack-infra/shade: Add support for limits https://review.openstack.org/395235 | 14:47 |
| openstackgerrit | Monty Taylor proposed openstack-infra/shade: Honor image_endpoint_override for image discovery https://review.openstack.org/418432 | 14:59 |
| mordred | frickler: ^^ | 14:59 |
| mordred | frickler: that should solve the image_endpoint_override issue | 14:59 |
| mordred | frickler: you should also be able to set api_timeout in clouds.yaml (I think that should make the failure take less time) | 15:01 |
| frickler | mordred: that patch is working fine for me, thanks a lot | 15:14 |
| sc68cal | mordred: so, now that kevinbenton has bug #1654806 I can take another peek at that issue we were looking at yesterday | 15:15 |
| openstack | bug 1654806 in python-novaclient "Functional tests (with neutron) failing with "ERROR (Conflict): Multiple possible networks found, use a Network ID to be more specific. (HTTP 409)" since 1/6" [High,In progress] https://launchpad.net/bugs/1654806 - Assigned to Matt Riedemann (mriedem) | 15:15 |
| mordred | sc68cal: oh! so we're fixed now | 15:15 |
| sc68cal | lol | 15:15 |
| sc68cal | so what did it end up being? | 15:15 |
| mordred | sc68cal: our problem stemmed from the openstacksdk release which broke openstackclient's ability to set quotas | 15:15 |
| mordred | sc68cal: why that didn't break MORE of the gate I don't really know | 15:15 |
| mordred | but pinning openstacksdk fixed all the things | 15:16 |
| sc68cal | Ah, and that same commit that pinned openstacksdk fixed 1654806 | 15:16 |
| sc68cal | so basically all our woes point back to openstacksdk | 15:16 |
| mordred | sc68cal: fwiw, it also manifested as weird things, like subnets owned by admin attached to the demo project but with no network and stuff | 15:16 |
| mordred | yah | 15:16 |
| sc68cal | yeah openstacksdk is .... having issues | 15:17 |
| mordred | sc68cal: clarkb said some stuff around 00:15:15 in #openstack-qa yesterday | 15:17 |
| sc68cal | I love https://review.openstack.org/#/c/364052/4 btw | 15:17 |
| sc68cal | best commit message of all time | 15:17 |
| mordred | sc68cal: wow. that really communicates to me the background, the intent and the implementation | 15:19 |
| sc68cal | oh even better - part of that commit changes the handling of tenant_id to project_id | 15:20 |
| sc68cal | among god knows what else | 15:20 |
| frickler | mordred: wow, cool, so now you also just fixed https://bugs.launchpad.net/python-openstackclient/+bug/1655335 for me :) | 15:22 |
| openstack | Launchpad bug 1655335 in python-openstackclient "image list fails against Icehouse cloud with OSC 3.6.0" [Undecided,New] | 15:22 |
| mordred | frickler: yay!!!! | 15:22 |
| mordred | Shrews: most recent patch from me up there fixes an issue frickler encountered in the wild. I also wrote a test :) | 15:28 |
| openstackgerrit | Merged openstack/os-client-config: Update swift constructor to be Session aware https://review.openstack.org/414086 | 15:48 |
| *** gouthamr has joined #openstack-shade | 16:04 | |
| *** cdent has joined #openstack-shade | 16:06 | |
| yfried | mordred: ping re shade/ansible | 16:15 |
| mordred | yfried: heya | 16:15 |
| yfried | mordred: how are you? | 16:15 |
| yfried | mordred: where should os_* look for clouds.yml? | 16:15 |
| mordred | yfried: I'm good! I'm in an all-day meeting today, so I may respond and then not respond - but also I'm in an all day meeting so I like to ignore the meeting when I can | 16:15 |
| yfried | mordred: ~/.config and /etc/openstack are easy, but what about $PWD? | 16:16 |
| mordred | yfried: it should look in PWD too ... | 16:16 |
| mordred | yfried: http://docs.openstack.org/developer/os-client-config/#config-files | 16:16 |
| yfried | mordred: we have {{ playbook_dir }} and {{ inventory_dir }} and $PWD, so where should it be? | 16:17 |
| yfried | mordred: $PWD doesn't work | 16:17 |
| mordred | yfried: it'll be whatever python thinks os.cwd is when it gets to excuting ... so in ansible's case that's going to be the remote tmp dir, not the dir the playbook was executed in | 16:18 |
| yfried | mordred: assuming localhost? | 16:18 |
| mordred | ansible still makes a tmp dir, copies things to it and then executes from there I believe | 16:19 |
| yfried | mordred: that's bad then... | 16:19 |
| mordred | if we are doing local_connection/localhost, I imagine it might be possible to plumb in playbook_dir and pass it to the constructor to add it to the path | 16:19 |
| yfried | mordred: can we use some env_var for a direct pointer? | 16:19 |
| mordred | probably, yeah. lemme poke at it a little bit | 16:20 |
| mordred | I grok the desire | 16:20 |
| yfried | mordred: not sure which of the 3 dirs is the right one | 16:20 |
| yfried | mordred: I think $PWD | 16:20 |
| yfried | mordred: but a way to choose that explicitly would be best | 16:20 |
| mordred | yfried: out of curiosity ... can you describe your use case a little bit, so that I can understand better where /etc/openstack and ~/.config aren't working for you? (not questioning that they are, just don't want to make assumptions as I look) | 16:27 |
| yfried | mordred: I have a jenkins slave running multiple jobs | 16:28 |
| yfried | mordred: each building openstack clouds | 16:28 |
| mordred | ah. gotcha | 16:28 |
| mordred | thanks - that's super helpful | 16:28 |
| yfried | mordred: also, note that {{ playbook_dir }} and {{ inventory_dir }} and $PWD are all mutually exclusive in my case | 16:34 |
| *** yfried has quit IRC | 16:55 | |
| thingee | mordred +1 to not passing module.params to openstack_cloud | 17:11 |
| openstackgerrit | Monty Taylor proposed openstack-infra/shade: Honor image_endpoint_override for image discovery https://review.openstack.org/418432 | 17:18 |
| mordred | silly pep8 | 17:31 |
| *** thingee has quit IRC | 17:58 | |
| *** thingee has joined #openstack-shade | 18:00 | |
| *** thingee has quit IRC | 18:05 | |
| *** thingee has joined #openstack-shade | 18:05 | |
| *** thingee has quit IRC | 18:07 | |
| *** thingee has joined #openstack-shade | 18:08 | |
| clarkb | sc68cal: you should update https://bugs.launchpad.net/manila/+bug/1652317 if you have more info. I put what I learned yesterday in there but beyond knowing what was broken I had little detail | 18:11 |
| openstack | Launchpad bug 1652317 in Manila "OpenStackSDK refactoring caused various OSC networking commands to fail" [Critical,New] | 18:11 |
| sc68cal | clarkb: ack. I think kevinbenton and mreidem chatted for a bit in #openstack-neutron about how OSC broke python-neutronclient | 18:19 |
| sc68cal | so yeah OSC has really shit the bed for folks | 18:19 |
| mordred | we seem to have all grown a dependency on sdk and sdk has largely been over in the corner and ignored by folks who aren't on the sdk team | 18:33 |
| *** yfried has joined #openstack-shade | 18:44 | |
| openstackgerrit | Merged openstack-infra/shade: Honor image_endpoint_override for image discovery https://review.openstack.org/418432 | 18:48 |
| Shrews | mordred: oh, this might be useful at some point: https://review.openstack.org/418347 | 18:50 |
| openstackgerrit | Monty Taylor proposed openstack-infra/shade: Log request ids when debug logging is enabled https://review.openstack.org/417931 | 18:51 |
| mordred | Shrews: thanks ^^ fixed the nits | 18:51 |
| mordred | Shrews: ooh nice! | 18:52 |
| mordred | Shrews: I was also thinking of writing a little utility to include in shade to print out information about what shade thinks your networking setup is to ease debugging of complicated network configs | 18:53 |
| Shrews | mordred: you had that same thought *several* months back | 18:53 |
| mordred | Shrews: I did? damn | 18:54 |
| openstackgerrit | Merged openstack-infra/shade: Rework limits normalization https://review.openstack.org/417044 | 19:00 |
| *** cdent has quit IRC | 19:23 | |
| *** yfried has quit IRC | 19:41 | |
| *** yfried has joined #openstack-shade | 20:00 | |
| openstackgerrit | Merged openstack-infra/shade: Log request ids when debug logging is enabled https://review.openstack.org/417931 | 20:23 |
| *** dfflanders has joined #openstack-shade | 20:54 | |
| *** gouthamr has quit IRC | 21:34 | |
| *** cdent has joined #openstack-shade | 21:39 | |
| *** yfried has quit IRC | 21:43 | |
| *** cdent has quit IRC | 21:57 | |
| *** gouthamr has joined #openstack-shade | 22:47 | |
| *** larainema has quit IRC | 23:21 | |
| *** morgan has quit IRC | 23:23 | |
| *** larainema has joined #openstack-shade | 23:25 | |
| *** morgan_ has joined #openstack-shade | 23:38 | |
Generated by irclog2html.py 2.14.0 by Marius Gedminas - find it at mg.pov.lt!