| *** markvoelker has quit IRC | 06:20 | |
| *** markvoelker has joined #refstack | 07:20 | |
| *** kopecmartin has joined #refstack | 07:24 | |
| *** tosky has joined #refstack | 07:49 | |
| *** markvoelker has quit IRC | 07:55 | |
| openstackgerrit | Martin Kopec proposed openstack/python-tempestconf master: [DNM] Dummy patch, testing timouts https://review.openstack.org/568286 | 07:58 |
|---|---|---|
| openstackgerrit | Martin Kopec proposed openstack/python-tempestconf master: Make tempestconf easier to use as an library https://review.openstack.org/562672 | 07:59 |
| openstackgerrit | Chandan Kumar proposed openstack/python-tempestconf master: Set auth.use_dynamic_credentials to true be default https://review.openstack.org/568507 | 08:04 |
| openstackgerrit | Chandan Kumar proposed openstack/python-tempestconf master: Set auth.use_dynamic_credentials to true be default https://review.openstack.org/568507 | 08:05 |
| openstackgerrit | Chandan Kumar proposed openstack/python-tempestconf master: Set auth.use_dynamic_credentials to true by default https://review.openstack.org/568507 | 08:06 |
| tosky | kopecmartin, arxcruz, chandankumar: I blocked https://review.openstack.org/#/c/567798, see my comment | 08:15 |
| arxcruz | tosky: me and kopecmartin tested locally in our env | 08:16 |
| tosky | arxcruz: and is it a reason to not test it there in any way? | 08:16 |
| tosky | the change introduce a way to use tempestconf without default-config, which means a duplicate set of variables, and it's not clear that it is working or not | 08:17 |
| tosky | if an additional test is too much, there was an open question about then removing the file and using the new inline file completely | 08:17 |
| tosky | but no more discussion about that | 08:18 |
| tosky | was it decided to keep it? Boh | 08:18 |
| tosky | just an answer like "ok, after some discussion, maybe for now we can keep it" could be nice | 08:19 |
| tosky | not nice; required | 08:19 |
| openstackgerrit | Arx Cruz proposed openstack/python-tempestconf master: DNM - Testing tempestconf without default-overrides.conf https://review.openstack.org/568512 | 08:19 |
| arxcruz | tosky: ^ | 08:19 |
| chandankumar | tosky: https://review.openstack.org/#/c/568512/ changes will break tripleo for sure | 08:20 |
| tosky | sure, and that does not answer the question still | 08:20 |
| tosky | chandankumar: and was it too difficult to write it there in a comment? | 08:20 |
| tosky | and remember that more than others, if this code path is not tested, it *will* break (not may) | 08:21 |
| tosky | or can go easily out of sync with the existing file | 08:21 |
| arxcruz | tosky: done :) | 08:22 |
| arxcruz | tosky: it will not break, and let me explain why | 08:22 |
| chandankumar | arxcruz: commented | 08:22 |
| arxcruz | tempest init creates a tempest dir with an etc/default-overrides.conf | 08:23 |
| arxcruz | this etc/default-overrides.conf comes from python-tempestconf | 08:23 |
| arxcruz | once installed, it adds the default-overrides.conf in /etc/tempest/default-overrides.conf | 08:23 |
| arxcruz | tempest just copy from there, so the default-overrides always comes from python-tempestconf | 08:23 |
| arxcruz | not from tempest side | 08:24 |
| chandankumar | arxcruz: tripleo build the packages from each commit | 08:24 |
| chandankumar | using DLRN and then consume that package | 08:24 |
| arxcruz | tempest checks for /etc/tempest/ ~/.tempest and ~/.config/tempest | 08:24 |
| tosky | I'm not concerned about the case which uses default-overrides.conf, which is tested by the current zuul jobs | 08:24 |
| arxcruz | chandankumar: so ? | 08:24 |
| tosky | I was talking about this new code path | 08:24 |
| chandankumar | arxcruz: https://github.com/rdo-packages/tempestconf-distgit/blob/rpm-master/python-tempestconf.spec#L177 file will be not found it is give error | 08:25 |
| chandankumar | file not found | 08:25 |
| tosky | chandankumar: are you saying that removing default-overrides.conf will break tripleo, or that it would require a change in the packaging, and after that everything else should work | 08:26 |
| chandankumar | but in rdoinfo python-tempestconf is pinned at rdoinfo | 08:27 |
| chandankumar | let' see which one is picked | 08:27 |
| chandankumar | tosky: yup, it requires a change in packaging | 08:28 |
| tosky | chandankumar: but apart from that, would it work? | 08:28 |
| chandankumar | tosky: apart from we need to fix constants.py or the reference where file is used | 08:29 |
| chandankumar | tosky: current DNM patch will work | 08:29 |
| chandankumar | apart from that | 08:30 |
| openstackgerrit | Arx Cruz proposed openstack/python-tempestconf master: DNM - Testing tempestconf without default-overrides.conf https://review.openstack.org/568519 | 08:30 |
| arxcruz | chandankumar: done, it will not remove the default-overrides.conf, just rename it to default-overrides.old.conf | 08:30 |
| arxcruz | so it will not be read by python-tempestconf | 08:30 |
| chandankumar | arxcruz: yup | 08:30 |
| tosky | chandankumar: apparently there are no other references to default-overrides.conf, at least by checking codesearch.openstack.org | 08:33 |
| tosky | not sure about other tools like infrared | 08:33 |
| tosky | but that's less relevant | 08:34 |
| arxcruz | tosky: exactly, default-overrides.conf only exists in python-tempestconf | 08:38 |
| *** rmart04 has joined #refstack | 08:39 | |
| tosky | so, to summarize: apart from the (maybe) required fixes to the packaging (which could be avoided in an initial step with thanks to renaming), do you see any other issues with keeping just one codepath by default? | 08:42 |
| tosky | I'd just mention that another possibility would be not to copy the default-overrides.conf file and refer to the copy installed at the system level | 08:42 |
| tosky | that would solve the use case of using tempestconf outside a tempest workspace | 08:43 |
| tosky | was it considered too? | 08:43 |
| arxcruz | tosky: default-overrides.conf is installed at system level when it comes from the rpm | 08:43 |
| arxcruz | tosky: it was considered | 08:43 |
| arxcruz | tosky: so | 08:43 |
| arxcruz | if you just source the rc file | 08:43 |
| arxcruz | and run the tool | 08:44 |
| arxcruz | this patch will work | 08:44 |
| arxcruz | it will not find the default-overrides.conf and it will use the minimum one from code | 08:44 |
| tosky | yes, sure, and that makes that file useless | 08:44 |
| arxcruz | if you run tempest init, it will copy the default-overrides.conf from /etc/tempest/ | 08:44 |
| arxcruz | and if you run the tool from inside the tempet init directory, it will ready the default-overrides.conf | 08:45 |
| chandankumar | tosky: arxcruz in future we can get rid of that file, what about sticking to default values? | 08:45 |
| arxcruz | i'm okay with that | 08:45 |
| tosky | arxcruz: I understand that, what is not clear right now is: when you run tempest init, why is default-overrides.conf copied? Because it is inside /etc/tempest ? | 08:46 |
| arxcruz | tosky: yes | 08:46 |
| arxcruz | tosky: https://github.com/openstack/tempest/blob/master/tempest/cmd/init.py#L39-L43 | 08:46 |
| tosky | ok, and then tempest does not really care about that file | 08:47 |
| tosky | arxcruz: what I was saying is that the alternative could be: not install default-overrides.conf inside /etc/tempest, but with the rest of the data of tempestconf | 08:47 |
| tosky | tempestconf could always refer to files installed with its own resources | 08:47 |
| tosky | no need for a new code path | 08:48 |
| tosky | that's the other solution that I was wondering | 08:48 |
| arxcruz | tosky: imho the best would be hardcode the defaults, and get rid of d-o.conf | 08:48 |
| arxcruz | sorry, too big to keep typing the name of this file | 08:48 |
| tosky | oh, it's fine | 08:49 |
| tosky | my point in all of this discussion, and I hope it is clear, is just to avoid the case when one of the two cases is used more, the other is not normally used, and it breaks at the wrong time | 08:49 |
| arxcruz | np | 08:50 |
| arxcruz | so, once the dnm pass, please, give me your blessing :) | 08:50 |
| *** markvoelker has joined #refstack | 08:52 | |
| tosky | so no plan to remove that file for now? Just to understand the outcome | 08:52 |
| chandankumar | tosky: remove the file in such a way it will not break anything | 08:53 |
| chandankumar | once this patch will land, we can first remove file from rdo package and then from the codebase | 08:54 |
| chandankumar | tosky: there is no reference of default-overrides.conf in infrared, so we are good there | 08:55 |
| tosky | it seems (luckily) an implementation detail of tempestconf, not part of any "public interface" | 08:56 |
| arxcruz | tosky: if i reply the gerrit email, do my comments goes to gerrit web interface? | 09:20 |
| * arxcruz receives a lot of messages from gerrit in email | 09:20 | |
| tosky | arxcruz: good question, I don't know; I don't think I tried | 09:22 |
| tosky | I would say: try | 09:22 |
| tosky | in the worst case, you will get a reply from the mail server | 09:22 |
| tosky | ("what could go wrong?") | 09:22 |
| *** markvoelker has quit IRC | 09:25 | |
| arxcruz | tosky: kopecmartin http://logs.openstack.org/19/568519/1/check/python-tempestconf-tempest-devstack-demo/4ed6bc8/job-output.txt.gz#_2018-05-15_09_06_08_672117 | 09:38 |
| tosky | so something is missing from the internal defaults | 09:41 |
| arxcruz | yup | 09:41 |
| arxcruz | tosky: nothing happens in the email btw | 09:42 |
| *** xinliang has quit IRC | 09:51 | |
| openstackgerrit | Arx Cruz proposed openstack/python-tempestconf master: Load basic default config https://review.openstack.org/567798 | 09:51 |
| kopecmartin | tosky, timeout issue, it was not fixed by turning ceilometer off :/ | 09:55 |
| kopecmartin | https://review.openstack.org/#/c/568286/ | 09:55 |
| tosky | kopecmartin: uhm, I tried; I remember that reducing the concurrency AND disabling ceilometer was working, but maybe just a lucky shot | 09:58 |
| tosky | can you recheck with reducing the concurrency again to see if both things combined works? | 09:58 |
| tosky | shot in the dark, I know, but... | 09:58 |
| kopecmartin | sure | 09:58 |
| openstackgerrit | Martin Kopec proposed openstack/python-tempestconf master: [DNM] Reducing concurrency https://review.openstack.org/568286 | 10:01 |
| *** xinliang has joined #refstack | 10:04 | |
| *** tosky has quit IRC | 10:11 | |
| *** tosky has joined #refstack | 10:12 | |
| openstackgerrit | Chandan Kumar proposed openstack/python-tempestconf master: Set auth.use_dynamic_credentials to true by default https://review.openstack.org/568507 | 10:12 |
| chandankumar | arxcruz: ^^ | 10:12 |
| arxcruz | chandankumar: so, https://github.com/openstack/tempest/blob/master/tempest/config.py#L58 is already true in tempest | 10:14 |
| arxcruz | do we need to set to true here as well ? | 10:14 |
| chandankumar | arxcruz: if we donot set in tempest.conf it will take bydefault | 10:15 |
| chandankumar | ah, I missed that one | 10:15 |
| arxcruz | chandankumar: so, is really needed? | 10:16 |
| chandankumar | arxcruz: since tempest has setted in byfeault we can remove it from TQE | 10:16 |
| arxcruz | chandankumar: and what about your patch ? | 10:17 |
| chandankumar | arxcruz: I will abandon it https://github.com/openstack/tripleo-quickstart-extras/blob/master/roles/validate-tempest/templates/configure-tempest.sh.j2#L142 | 10:17 |
| *** markvoelker has joined #refstack | 10:23 | |
| chandankumar | arxcruz: for volume encryption one method is that | 10:25 |
| chandankumar | https://docs.openstack.org/ocata/config-reference/block-storage/volume-encryption.html | 10:25 |
| chandankumar | if there is a creator role present there, then encryption can be support | 10:25 |
| chandankumar | otherwise set to failse | 10:25 |
| chandankumar | that is the current way | 10:26 |
| arxcruz | tosky: chandankumar is it possible to cleanup a env and reinstall packstack ? | 10:52 |
| arxcruz | chandankumar: not really, the creator role is just if you want non admin users to run encrypted volumes | 10:52 |
| arxcruz | it's not needed for admin for example | 10:52 |
| *** markvoelker has quit IRC | 10:55 | |
| tosky | arxcruz: I have some hand-crafted scripts which should clenaup all packstack traces, let me copy them | 11:33 |
| tosky | arxcruz: http://paste.openstack.org/show/720986 | 11:35 |
| arxcruz | tosky: thanks, i have a machine on rdocloud but it's so freaking hard to get 16gb of ram there that i don't want to give up this machine | 11:40 |
| arxcruz | tosky: https://review.openstack.org/#/c/568519/ passed | 11:41 |
| *** markvoelker has joined #refstack | 11:52 | |
| *** markvoelker_ has joined #refstack | 12:02 | |
| *** markvoelker has quit IRC | 12:04 | |
| tosky | uh | 12:11 |
| *** rmart04 has quit IRC | 12:31 | |
| kopecmartin | tosky, https://review.openstack.org/#/c/568286/ , it was a lucky shot before :D | 13:03 |
| tosky | aufff | 13:09 |
| tosky | then another step is reproduce locally again and again (with 8GiB) | 13:10 |
| kopecmartin | uff :/ | 13:23 |
| tosky | kopecmartin: did you also check with the other people running packstack tests if they hit issues like that? | 13:56 |
| kopecmartin | tosky, not yet, for now rechecking needs to be enough .. I think, yesterday, when i opened logs from random jobs, i saw the same issue | 13:57 |
| openstackgerrit | Chandan Kumar proposed openstack/python-tempestconf master: Allow image creation by default for all users https://review.openstack.org/568252 | 14:27 |
| *** tosky has quit IRC | 14:40 | |
| *** tosky has joined #refstack | 14:44 | |
| openstackgerrit | Merged openstack/python-tempestconf master: Load basic default config https://review.openstack.org/567798 | 15:17 |
| chandankumar | \o/ | 15:33 |
| chandankumar | hogepodge: mguiney Do we have refstack meeting today? | 15:37 |
| hogepodge | yes, we should have a meeting in advance of the summit. Just to catch up a bit | 15:38 |
| hogepodge | chandankumar: tosky: mguiney: ^^ | 15:38 |
| mguiney | ok, can do | 15:41 |
| tosky | oki | 15:44 |
| openstackgerrit | Chandan Kumar proposed openstack/python-tempestconf master: Allow creation/setting of image/flavor-*alt for all users https://review.openstack.org/568252 | 15:45 |
| hogepodge | chandankumar: tosky: https://etherpad.openstack.org/p/refstack-meeting-18-05-15 agenda here | 16:32 |
| hogepodge | I'll let you pick the important reviews you want to cover | 16:41 |
| kopecmartin | arxcruz, are you still here? | 16:54 |
| openstackgerrit | megan guiney proposed openstack/refstack master: Revert eslint version to es5 https://review.openstack.org/568643 | 16:56 |
| arxcruz | kopecmartin: depend | 16:58 |
| kopecmartin | arxcruz, can you rebase this? https://review.openstack.org/#/c/567820/ | 16:59 |
| chandankumar | tosky: regarding this review https://review.openstack.org/#/c/568252/ you are correct, since user might have permission to create image/flavor if not then it will fail, my main intension is to auto detect flavor and two image id show that instances can be spinned | 17:01 |
| tosky | chandankumar: I'm still having some troubles understanding the scope | 17:01 |
| tosky | are you trying to show that a non-admin user as deployed by devstack can still create flavors and upload the images? | 17:02 |
| tosky | chandankumar, arxcruz, kopecmartin: if you are around, #openstack-meeting-alt | 17:02 |
| chandankumar | tosky: the scope of this change is to auto set image_ref and image_altref and flavor/flaver_alt in compute | 17:02 |
| chandankumar | tosky: yes | 17:03 |
| tosky | I guess I need to recheck the code, because I'm missing something | 17:04 |
| *** kopecmartin has quit IRC | 17:17 | |
| *** tosky has quit IRC | 17:33 | |
| *** tosky has joined #refstack | 17:38 | |
| *** ekhugen has quit IRC | 20:18 | |
| *** ekhugen has joined #refstack | 20:25 | |
| openstackgerrit | Martin Kopec proposed openstack/python-tempestconf master: Filter deprecated versions out https://review.openstack.org/568695 | 20:55 |
| *** tosky has quit IRC | 23:26 | |
Generated by irclog2html.py 2.15.3 by Marius Gedminas - find it at mg.pov.lt!