| jakeyip | flwang: functional tests still broken :P | 01:40 |
|---|---|---|
| jakeyip | I see you've been re-trying. I have been trying to figure out why but didn't get anywhere. Do you know what's wrong? | 01:41 |
| flwang | jakeyip: i'm testing with this patch https://review.opendev.org/#/c/659280/3 | 01:45 |
| flwang | from the error, seems the job is running py3 job in a py3 environment | 01:46 |
| flwang | but i don't understand what's root cause yet | 01:46 |
| flwang | jakeyip: it would be appreciated if you can take a look | 01:46 |
| *** ykarel has joined #openstack-containers | 02:01 | |
| jakeyip | yeah I did take a look and the thing I found was it fails with `ImportError: No module named tempest.lib.common.utils` | 02:07 |
| *** ricolin has joined #openstack-containers | 02:14 | |
| jakeyip | can't figure out why it doesn't seemed to be installed | 02:15 |
| *** threestrands has joined #openstack-containers | 02:34 | |
| flwang | jakeyip: actually, other tempest cases do run in py3.6, see http://logs.openstack.org/80/659280/3/check/magnum-functional-api/e4aab50/job-output.txt.gz#_2019-05-19_23_13_48_293321 | 02:45 |
| jakeyip | the line I'm looking at is http://logs.openstack.org/80/659280/3/check/magnum-functional-api/e4aab50/job-output.txt.gz#_2019-05-19_23_16_28_239150 it's a py27 thing | 02:47 |
| jakeyip | hm. there was a patch for sphinx being different in py2 and 3 | 02:48 |
| flwang | yep, that's the weird thing, seems only the funciton api testing running in the py27, seems we need to make sure the template magnum plugin is installed in py3.6 | 02:50 |
| jakeyip | I don't understand the last bit - do you think error is in py27 or py36? | 02:52 |
| flwang | jakeyip: why do we need the sphinx change in https://review.opendev.org/#/c/659027/4/test-requirements.txt ? | 02:53 |
| flwang | i'm wondering if we need same change in the magnum-tempest-plugin | 02:53 |
| jakeyip | hm, AFAIK sphinx is deprecated for py2 | 02:55 |
| jakeyip | let's try! | 02:55 |
| openstackgerrit | Jake Yip proposed openstack/magnum-tempest-plugin master: Update json module to jsonutils https://review.opendev.org/638968 | 02:57 |
| jakeyip | rebased so if that fails we know we need the patch | 02:58 |
| jakeyip | zuul says ok | 03:03 |
| flwang | we need to make https://review.opendev.org/#/c/659280/ depends on the one in magnum-tempest-plugin to make sure | 03:04 |
| jakeyip | why? that rebased change doesn't update any requirements? | 03:05 |
| flwang | do you know 'Depends-On' keyword in commit message? | 03:06 |
| flwang | it can help verify cross-project patch | 03:07 |
| flwang | like this | 03:07 |
| jakeyip | yeah I know but I don't think that matters? | 03:07 |
| flwang | we're not on the same page i think | 03:07 |
| flwang | jakeyip: can you please help me understand why we need https://review.opendev.org/#/c/638968/3 ? | 03:08 |
| jakeyip | the rebase change made sure that magnum-tempest-plugin is not broken with upgrade libraries | 03:08 |
| flwang | why it's helpful for our current failure on magnum? | 03:08 |
| jakeyip | *not broken with new libraries in pip | 03:08 |
| jakeyip | you were wondering if we needed to update magnum-tempest-plugin | 03:09 |
| flwang | as you can see, only pep8 job was run for that patch, i can't trust it | 03:09 |
| flwang | jakeyip: not update magnum-tempest-plugin, but we may need a bump in the test-requirements or the tox.ini | 03:10 |
| jakeyip | ah yes. magnum-tempest-plugin only have pep8 tests. | 03:10 |
| flwang | but not the code | 03:10 |
| jakeyip | hm, so I found that sphinx is in test-requirements.txt of magnum-tempest-plugin. maybe that needs an update | 03:11 |
| jakeyip | I'll send up the changes. Can I use your DNM patch with depends on ? | 03:11 |
| flwang | sure, go for it | 03:11 |
| jakeyip | ok | 03:11 |
| *** spiette has quit IRC | 03:13 | |
| *** spiette has joined #openstack-containers | 03:14 | |
| openstackgerrit | Jake Yip proposed openstack/magnum-tempest-plugin master: Cap sphinx on py2 https://review.opendev.org/660009 | 03:16 |
| openstackgerrit | Jake Yip proposed openstack/magnum master: DO NOT MERGE https://review.opendev.org/659280 | 03:19 |
| *** xiangdong-huang has joined #openstack-containers | 03:57 | |
| *** dave-mccowan has quit IRC | 03:58 | |
| *** udesale has joined #openstack-containers | 04:01 | |
| *** ykarel has quit IRC | 04:04 | |
| *** ykarel has joined #openstack-containers | 04:22 | |
| *** janki has joined #openstack-containers | 04:38 | |
| *** ivve has quit IRC | 05:09 | |
| *** lpetrut has joined #openstack-containers | 05:40 | |
| *** udesale has quit IRC | 05:42 | |
| *** udesale has joined #openstack-containers | 05:42 | |
| *** ivve has joined #openstack-containers | 06:12 | |
| *** lxkong has quit IRC | 06:47 | |
| *** lxkong has joined #openstack-containers | 06:48 | |
| *** guilhermesp has quit IRC | 06:48 | |
| *** mnasiadka has quit IRC | 06:48 | |
| *** adrianreza has quit IRC | 06:48 | |
| *** guilhermesp has joined #openstack-containers | 06:48 | |
| *** mnasiadka has joined #openstack-containers | 06:48 | |
| *** adrianreza has joined #openstack-containers | 06:48 | |
| *** udesale has quit IRC | 07:00 | |
| *** udesale has joined #openstack-containers | 07:00 | |
| *** lpetrut has quit IRC | 07:04 | |
| *** zigo has joined #openstack-containers | 07:06 | |
| *** rcernin has quit IRC | 07:08 | |
| *** yolanda_ has quit IRC | 07:13 | |
| *** yolanda_ has joined #openstack-containers | 07:20 | |
| *** udesale has quit IRC | 07:20 | |
| *** udesale has joined #openstack-containers | 07:21 | |
| *** threestrands has quit IRC | 07:23 | |
| *** pcaruana has joined #openstack-containers | 07:23 | |
| *** udesale has quit IRC | 07:26 | |
| *** udesale has joined #openstack-containers | 07:27 | |
| *** ykarel is now known as ykarel|lunch | 07:35 | |
| *** mkuf has quit IRC | 08:28 | |
| *** ykarel|lunch is now known as ykarel | 08:31 | |
| *** ramishra has joined #openstack-containers | 08:38 | |
| *** mkuf has joined #openstack-containers | 08:38 | |
| *** flwang1 has joined #openstack-containers | 09:05 | |
| openstackgerrit | Feilong Wang proposed openstack/magnum-tempest-plugin master: Cap sphinx on py2 https://review.opendev.org/660009 | 09:12 |
| *** ricolin has quit IRC | 09:19 | |
| zigo | flwang: Hi there! Around? | 09:26 |
| zigo | Looks like I got Magnum "almost" working, ie, api and conductor can start, but when I do coe cluster list, I get a crash here: | 09:27 |
| zigo | File "/usr/lib/python3/dist-packages/magnum/common/policy.py", line 110, in add_policy_attributes | 09:27 |
| zigo | trustee_domain_id = admin_osc.keystone().trustee_domain_id | 09:27 |
| zigo | Keystone returns 404, and therefore, magnum-api returns 500. | 09:28 |
| zigo | What did I do wrong in my Keystone / magnum setup? | 09:28 |
| flwang1 | zigo: can you pls show me your magnum.conf? | 09:28 |
| zigo | flwang: http://paste.openstack.org/show/751658/ | 09:29 |
| *** lpetrut has joined #openstack-containers | 09:30 | |
| zigo | Hum ... pybasedir looks wrong to me... :P | 09:31 |
| zigo | Though it shouldn't be the issue. | 09:31 |
| flwang1 | adding 'cluster_user_trust = True' to trust section | 09:32 |
| flwang1 | do you have the magnum domain in keystone? | 09:33 |
| zigo | flwang: Yep. | 09:33 |
| zigo | flwang1: Adding cluster_user_trust = True doesn't fix the issue. | 09:34 |
| zigo | Should I keep this always to true? The comment on the directive says its dangerous ... | 09:34 |
| flwang1 | can you confirm you have the domain in keystone? | 09:38 |
| zigo | Also, what's the goal of the [keystone_auth] section? Isn't it redundant with [keystone_authtoken]? Is it ok if I just set [keystone_auth]/auth_section => keystone_authtoken? | 09:38 |
| flwang1 | you can drop the config option then | 09:38 |
| zigo | I do. | 09:38 |
| flwang1 | how did you install your magnum? | 09:39 |
| zigo | # openstack role assignment list --user magnum_admin --names | 09:39 |
| zigo | +-------+---------------------+-------+---------+--------+--------+-----------+ | 09:39 |
| zigo | | Role | User | Group | Project | Domain | System | Inherited | | 09:39 |
| zigo | +-------+---------------------+-------+---------+--------+--------+-----------+ | 09:39 |
| zigo | | admin | magnum_admin@magnum | | | magnum | | False | | 09:39 |
| zigo | +-------+---------------------+-------+---------+--------+--------+-----------+ | 09:39 |
| zigo | flwang1: I am using the Debian packages that I maintain in Debian (I'm the Debian OpenStack package maintainer since OpenStack started), and my own installer: https://salsa.debian.org/openstack-team/debian/openstack-cluster-installer | 09:40 |
| zigo | This installer uses puppet-openstack internally. | 09:40 |
| zigo | I also tried installing the Magnum package by hand, and I have the exact same problem. | 09:40 |
| zigo | My goal is having Magnum fully integrated into OCI, once and for all, then you'll be 100% sure that it continues to work in Debian forever ... :P | 09:41 |
| flwang1 | zigo: ah, i see | 09:43 |
| flwang1 | let me think | 09:43 |
| flwang1 | we do need keystone_auth and keystone_authtoken, i can't remember the reason now :( | 09:43 |
| zigo | I believe I have it written correctly already anywas. | 09:44 |
| flwang1 | i would suggest add a break point at /usr/lib/python3/dist-packages/magnum/common/policy.py | 09:45 |
| flwang1 | to understand why 404 at trustee_domain_id = admin_osc.keystone().trustee_domain_id | 09:45 |
| zigo | flwang1: Is it ok to run magnum-api over UWSGI ? | 09:46 |
| zigo | I wrote this: https://review.opendev.org/#/c/659764/ | 09:47 |
| zigo | it works well for me, but for a reason, the gate doesn't like it... | 09:47 |
| zigo | Anyway, I'll try debugging... | 09:47 |
| flwang1 | the gate failure is not related to your patch i think | 09:48 |
| jakeyip | zigo: it's not you, functional tests is broken a.t.m. | 09:48 |
| flwang1 | we're have a known issue with magnum-functional-api | 09:48 |
| zigo | Oh, good to know. | 09:48 |
| jakeyip | flwang1: no luck still even with the magnum-test-plugin patch | 09:48 |
| flwang1 | jakeyip: i saw that and i'm doing another testing | 09:48 |
| flwang1 | if we still can't figure out it, then i'm going to change it as non-vote until we fix it | 09:49 |
| zigo | I very much would recommend running functional tests over UWSGI. The current setup with werkzeug isn't IMO for production. | 09:49 |
| jakeyip | I can't figure out how to run the functional tests on my desktop so I do more in-depth debugging | 09:49 |
| jakeyip | *to do more in-depth debugging | 09:50 |
| jakeyip | the docs didn't work for me. if anyone has gotten functional tests working on their desktop please let me know. | 09:50 |
| *** nfakhir is now known as nfakhir_ | 09:51 | |
| jakeyip | desktop or in their own CI pipelines | 09:51 |
| *** nfakhir_ is now known as nfakhir | 09:51 | |
| *** ttsiouts has joined #openstack-containers | 09:54 | |
| flwang1 | ttsiouts: around? | 09:55 |
| *** rcernin has joined #openstack-containers | 10:04 | |
| *** ivve has quit IRC | 10:09 | |
| flwang1 | jakeyip: i just compared the depedencies version in http://logs.openstack.org/49/656649/1/check/magnum-functional-api/95ff2f8/job-output.txt.gz#_2019-05-01_15_43_49_348082 | 10:16 |
| flwang1 | with the current failed one | 10:16 |
| flwang1 | seems a lot of libs changed their versions recently | 10:16 |
| openstackgerrit | Feilong Wang proposed openstack/magnum master: DO NOT MERGE https://review.opendev.org/659280 | 10:21 |
| *** ttsiouts has quit IRC | 10:23 | |
| jakeyip | flwang1: yeah I did compare the dependencies too. that's why I'm trying to build it locally so I can test versions easier | 10:25 |
| flwang1 | i just pin the tox and stestr version to see if it's related | 10:41 |
| *** ramishra has quit IRC | 10:43 | |
| jakeyip | stestr in a successful run is stestr==2.2.0 http://logs.openstack.org/70/653270/4/check/magnum-functional-api/b0be9a1/job-output.txt.gz#_2019-05-02_18_15_02_661951 | 10:50 |
| *** rcernin has quit IRC | 10:57 | |
| *** ykarel is now known as ykarel|afk | 11:00 | |
| *** yolanda_ has quit IRC | 11:04 | |
| *** yolanda_ has joined #openstack-containers | 11:08 | |
| *** udesale has quit IRC | 11:16 | |
| zigo | flwang1: When debugging a little bit, it looks like Magnum is trying to get a token using https://z-api.infomaniak.ch:443/identity/auth/tokens instead of https://z-api.infomaniak.ch:443/identity/v3/auth/tokens :/ | 11:47 |
| zigo | flwang1: Where should I fix this? | 11:47 |
| flwang1 | zigo: let me check your conf again, wait a sec | 11:48 |
| zigo | flwang1: I did this: curl -k -i -H "Content-Type: application/json" -d '{"auth": {"identity": {"methods": ["password"], "password": {"user": {"password": "XXXX", "name": "magnum_admin", "domain": {"name": "magnum"}}}}, "scope": {"domain": {"name": "magnum"}}}}' https://z-api.infomaniak.ch:443/identity/v3/auth/tokens | 11:48 |
| zigo | This works. | 11:48 |
| zigo | But with /v3, it doesn't. | 11:48 |
| zigo | And Magnum, apparently, is trying without /v3. | 11:48 |
| flwang1 | zigo: follow this http://paste.openstack.org/show/751662/ | 11:49 |
| zigo | Right. | 11:49 |
| flwang1 | zigo: your auth_url should be added the /v3 | 11:50 |
| zigo | flwang1: IMO, this is a defect in Magnum, and this should be fixed. | 11:50 |
| zigo | flwang1: Now, I get a different issue ! :) | 11:51 |
| zigo | Ah no, maybe not. | 11:51 |
| zigo | Hang on. | 11:52 |
| zigo | Removing my breakpoint ... | 11:52 |
| zigo | flwang1: Looks like working now ... :) | 11:52 |
| flwang1 | :) | 11:52 |
| flwang1 | congrats | 11:53 |
| zigo | flwang1: If you guys don't fix it, I would suggest that YOU WRITE IN BOLD LETTERS that Magnum needs a versionned API URL. | 11:54 |
| zigo | Because that's very uncommon. | 11:54 |
| flwang1 | zigo: i'm happy to fix it, the only problem is, we don't have enough resource | 11:57 |
| flwang1 | though there are many adopters of Magnum, but the contributors are not much, | 12:00 |
| flwang1 | i have to go, it's 12:00am here | 12:00 |
| nfakhir | I've ran into the same problem. Same thing for the [keystone_auth] section which is not required in other services | 12:00 |
| flwang1 | nfakhir: zigo: if you guys have time, pls feel free to file a bug and propose a patch, i'm happy to review it | 12:01 |
| *** kplant has joined #openstack-containers | 12:05 | |
| *** ykarel|afk is now known as ykarel | 12:05 | |
| openstackgerrit | Thomas Goirand proposed openstack/magnum master: Add note about Keystone URLs must be versionned. https://review.opendev.org/660101 | 12:10 |
| zigo | flwang1: My time is very limited, so let's just get the job done and add just a quick note ... :P | 12:10 |
| flwang1 | zigo: no problem | 12:11 |
| *** ykarel_ has joined #openstack-containers | 12:16 | |
| kplant | is it possible for magnum to create instances using cinder volumes for the root disk instead of nova volumes? | 12:18 |
| *** ykarel has quit IRC | 12:18 | |
| *** henriqueof has joined #openstack-containers | 12:20 | |
| flwang1 | kplant: i'm working on that https://review.opendev.org/#/c/621734/ | 12:24 |
| kplant | awesome | 12:24 |
| kplant | thank you | 12:25 |
| *** ianychoi has quit IRC | 12:28 | |
| *** ianychoi has joined #openstack-containers | 12:29 | |
| *** janki has quit IRC | 12:29 | |
| *** jmlowe has quit IRC | 12:49 | |
| *** udesale has joined #openstack-containers | 13:01 | |
| *** janki has joined #openstack-containers | 13:08 | |
| *** ykarel__ has joined #openstack-containers | 13:17 | |
| *** ykarel_ has quit IRC | 13:19 | |
| *** jmlowe has joined #openstack-containers | 13:26 | |
| *** pcaruana has quit IRC | 13:33 | |
| *** ykarel__ is now known as ykarel | 13:33 | |
| *** lpetrut has quit IRC | 13:53 | |
| *** janki has quit IRC | 14:10 | |
| *** lpetrut has joined #openstack-containers | 14:15 | |
| openstackgerrit | Jake Yip proposed openstack/magnum master: WIP: Blacklist bandit 1.6.0 and cap Sphinx on Python2 https://review.opendev.org/660129 | 14:18 |
| *** lpetrut has quit IRC | 14:26 | |
| *** pcaruana has joined #openstack-containers | 14:53 | |
| *** sapd1_x has joined #openstack-containers | 15:00 | |
| *** ykarel is now known as ykarel|away | 15:05 | |
| *** jmlowe has quit IRC | 15:36 | |
| *** jmlowe has joined #openstack-containers | 15:37 | |
| *** ykarel|away has quit IRC | 15:40 | |
| *** ricolin has joined #openstack-containers | 15:55 | |
| *** ykarel|away has joined #openstack-containers | 16:03 | |
| *** sapd1_x has quit IRC | 16:06 | |
| *** nfakhir is now known as nfakhir_ | 16:12 | |
| *** ricolin has quit IRC | 16:22 | |
| openstackgerrit | zhulingjie proposed openstack/python-magnumclient master: Replace git.openstack.org URLs with opendev.org URLs https://review.opendev.org/654786 | 16:23 |
| *** ivve has joined #openstack-containers | 16:24 | |
| *** itlinux_ has quit IRC | 16:26 | |
| *** dave-mccowan has joined #openstack-containers | 16:29 | |
| *** nfakhir_ is now known as nfakhir | 16:35 | |
| *** nfakhir is now known as nfakhir_ | 16:36 | |
| *** udesale has quit IRC | 16:40 | |
| *** ykarel|away has quit IRC | 16:59 | |
| openstackgerrit | Jake Yip proposed openstack/magnum-tempest-plugin master: wip: do not use sitepackages for functional tests https://review.opendev.org/660177 | 17:00 |
| openstackgerrit | Jake Yip proposed openstack/magnum-tempest-plugin master: Cap sphinx on py2 https://review.opendev.org/660009 | 17:10 |
| *** ykarel|away has joined #openstack-containers | 17:14 | |
| *** ttsiouts has joined #openstack-containers | 17:15 | |
| *** ttsiouts has quit IRC | 17:20 | |
| *** lpetrut has joined #openstack-containers | 17:23 | |
| *** lpetrut has quit IRC | 17:27 | |
| *** ykarel|away has quit IRC | 18:04 | |
| *** kplant has quit IRC | 19:17 | |
| *** jmlowe has quit IRC | 19:30 | |
| *** flwang1 has quit IRC | 19:55 | |
| *** nfakhir_ is now known as nfakhir | 20:33 | |
| *** nfakhir is now known as nfakhir_ | 20:39 | |
| *** ttsiouts has joined #openstack-containers | 20:55 | |
| *** itlinux has joined #openstack-containers | 21:14 | |
| *** jmlowe has joined #openstack-containers | 21:33 | |
| *** ttsiouts has quit IRC | 21:34 | |
| *** henriqueof has quit IRC | 21:35 | |
| *** pcaruana has quit IRC | 21:40 | |
| *** itlinux has quit IRC | 21:49 | |
| *** ttsiouts has joined #openstack-containers | 21:49 | |
| *** ttsiouts has quit IRC | 21:54 | |
| *** ttsiouts has joined #openstack-containers | 22:23 | |
| openstackgerrit | Feilong Wang proposed openstack/magnum master: Blacklist bandit 1.6.0 and cap Sphinx on Python2 https://review.opendev.org/659027 | 23:05 |
| *** itlinux has joined #openstack-containers | 23:18 | |
| *** rcernin has joined #openstack-containers | 23:23 | |
| jakeyip | morning flwang you around? | 23:30 |
| *** itlinux has quit IRC | 23:35 | |
| *** ttsiouts has quit IRC | 23:35 | |
| openstackgerrit | Jake Yip proposed openstack/magnum-tempest-plugin master: Do not use sitepackages for functional tests https://review.opendev.org/660177 | 23:36 |
| jakeyip | don't set functional-api to non-voting yet, I think I've fixed it | 23:37 |
| jakeyip | see https://review.opendev.org/#/c/660129/ | 23:38 |
| jakeyip | going to fix up commit messages and send them up | 23:38 |
| openstackgerrit | Jake Yip proposed openstack/magnum-tempest-plugin master: Cap sphinx on py2 https://review.opendev.org/660009 | 23:38 |
| openstackgerrit | Jake Yip proposed openstack/magnum master: Blacklist bandit 1.6.0 and cap Sphinx on Python2 https://review.opendev.org/660129 | 23:41 |
| openstackgerrit | Jake Yip proposed openstack/magnum master: DO NOT MERGE https://review.opendev.org/659280 | 23:43 |
| flwang | jakeyip: cool | 23:43 |
| flwang | let's see | 23:43 |
| jakeyip | the change that fixed it is https://review.opendev.org/660177 | 23:43 |
| flwang | hah | 23:44 |
| flwang | but i'm a bit confused, using sitepackages=true means a virtual env will be created | 23:46 |
| flwang | jakeyip: https://tox.readthedocs.io/en/latest/config.html#conf-sitepackages | 23:46 |
| flwang | i can't see what's the behaviour of the default value(False?) | 23:47 |
| *** itlinux has joined #openstack-containers | 23:48 | |
| *** ttsiouts has joined #openstack-containers | 23:50 | |
| jakeyip | I think setting to true means the venv will try to use site packages too | 23:54 |
| jakeyip | Your DNM patch failure is something different - on py3 it was timing out | 23:55 |
| jakeyip | I've rebased it to check if my patch that fixes py2 helps in py3 | 23:57 |
Generated by irclog2html.py 2.15.3 by Marius Gedminas - find it at mg.pov.lt!