| openstackgerrit | Qiming Teng proposed openstack/senlin: Add a TODO item (unexpected attribute) https://review.openstack.org/303116 | 00:04 |
|---|---|---|
| *** xuhaiwei has quit IRC | 00:23 | |
| *** xuhaiwei has joined #senlin | 00:56 | |
| *** shu-mutou has joined #senlin | 01:46 | |
| *** yanyanhu has joined #senlin | 01:55 | |
| *** yanyanhu has quit IRC | 02:01 | |
| *** yanyanhu has joined #senlin | 02:02 | |
| *** elynn has joined #senlin | 02:02 | |
| *** elynn has quit IRC | 02:07 | |
| *** elynn has joined #senlin | 02:07 | |
| *** Qiming has quit IRC | 02:13 | |
| *** elynn has quit IRC | 02:16 | |
| *** elynn has joined #senlin | 02:19 | |
| *** openstackstatus has quit IRC | 02:19 | |
| *** yanyanhu has quit IRC | 02:21 | |
| *** yanyanhu has joined #senlin | 02:23 | |
| *** gongysh has joined #senlin | 02:25 | |
| *** yuanying has quit IRC | 02:46 | |
| *** zzxwill has joined #senlin | 03:05 | |
| *** Qiming has joined #senlin | 03:07 | |
| *** zzxwill_ has joined #senlin | 03:08 | |
| *** zzxwill has quit IRC | 03:09 | |
| *** zzxwill_ is now known as zzxwill | 03:09 | |
| Qiming | elynn, seems this is caused by trusts : https://review.openstack.org/#/c/302889/ | 03:25 |
| Qiming | please check if we should wrap the cred_create_update operation into a transaction | 03:26 |
| elynn | ok, let me check | 03:26 |
| *** gongysh has quit IRC | 03:41 | |
| openstackgerrit | Ethan Lynn proposed openstack/senlin: Make cred_create_update operation into a transaction https://review.openstack.org/303204 | 03:45 |
| *** zzxwill has quit IRC | 03:46 | |
| *** yuanying has joined #senlin | 03:49 | |
| openstackgerrit | Ethan Lynn proposed openstack/senlin: Initial tempest plugin framework https://review.openstack.org/295647 | 03:52 |
| *** elynn has quit IRC | 04:20 | |
| *** zzxwill has joined #senlin | 04:34 | |
| *** elynn has joined #senlin | 04:40 | |
| *** zzxwill has quit IRC | 04:41 | |
| *** elynn has quit IRC | 04:45 | |
| *** elynn has joined #senlin | 04:45 | |
| *** zzxwill has joined #senlin | 04:45 | |
| openstackgerrit | Merged openstack/senlin: Updated from global requirements https://review.openstack.org/302889 | 04:53 |
| *** shu-mutou is now known as shu-mutou-AFK | 04:57 | |
| openstackgerrit | Merged openstack/senlin: Add senlin-dashboard in doc/source/overview.rst https://review.openstack.org/301683 | 05:13 |
| openstackgerrit | Merged openstack/senlin: Initial tempest plugin framework https://review.openstack.org/295647 | 05:14 |
| *** elynn has quit IRC | 05:31 | |
| *** elynn has joined #senlin | 06:02 | |
| *** elynn has quit IRC | 06:07 | |
| *** elynn has joined #senlin | 06:07 | |
| yanyanhu | hi, elynn, Qiming, I'm thinking whether we should rename the directory "senlin/tests/tempest_tests" to "senlin/tests/tempest/scenarios" and maybe also create "senlin/tests/tempest/api" in future | 06:26 |
| yanyanhu | since we have decided to use tempest to cover both scenario and API test for senlin | 06:26 |
| yanyanhu | and if we decide to use Rally for performance test in future, there could be a directory "senlin/tests/rally/performance" | 06:27 |
| yanyanhu | or "senlin/tests/scenario/tempest" and "senlin/tests/api/tempest" | 06:29 |
| yanyanhu | and "senlin/tests/performance/rally" | 06:29 |
| yanyanhu | since we now have "senlin/tests/functional" which is implemented using pure API request | 06:29 |
| yanyanhu | for functional test | 06:31 |
| elynn | You mean rename tests/tempest_tests/tests/api to tests/tempest/api ? | 06:32 |
| elynn | I think in the future we should remove functional tests if we have tempest tests fully supported. | 06:33 |
| yanyanhu | yes | 06:33 |
| yanyanhu | elynn, it's possible | 06:33 |
| yanyanhu | just feel we should organize our test cases based on its classification | 06:34 |
| elynn | Rename the directory name is ok for me. | 06:34 |
| yanyanhu | its/their | 06:34 |
| yanyanhu | since that will be clearer for developer and user | 06:34 |
| yanyanhu | they will know we are using tempest to support API and scenario and maybe also functional test | 06:35 |
| elynn | The dir structure copied from other project , I haven't think of it too much. | 06:35 |
| yanyanhu | elynn, ok, just feel the current dir structure is not very clear since you can't tell what kind of test it is from the name "tempest_test" :) | 06:36 |
| elynn | But I think it's better put all tempest tests into the tempest/ folder. | 06:37 |
| elynn | Because we only have one tempest entry in setup.cfg. | 06:37 |
| yanyanhu | elynn, yes, if so, the directory should be named like "senlin/tests/tempest/scenario" | 06:37 |
| yanyanhu | and "senlin/tests/tempest/api" and "senlin/tests/tempest/functional" I think | 06:37 |
| elynn | functional and API is not the same thing? | 06:38 |
| Qiming | rename the directory "senlin/tests/tempest_tests" to "senlin/tests/tempest/scenarios" and maybe also create "senlin/tests/tempest/api" in future | 06:38 |
| yanyanhu | elynn, I think there some some differences between them | 06:38 |
| yanyanhu | but there are some overlap with API test in existing functional test | 06:39 |
| elynn | the tests in https://github.com/openstack/senlin/blob/master/senlin/tests/tempest_tests/tests/api/test_cluster_basic.py should belong to API or functional? | 06:40 |
| yanyanhu | may need some reorganization after tempest support is completed down | 06:40 |
| elynn | I'm not so clear about API tests and functional tests. | 06:41 |
| Qiming | API tests should cover also negative tests | 06:41 |
| yanyanhu | elynn, I think functional test is for verification for some important workflow inside senlin service | 06:41 |
| yanyanhu | Qiming, exactly | 06:41 |
| yanyanhu | without talking with other openstack services | 06:42 |
| yanyanhu | elynn, functional test runs much faster than scenario test | 06:43 |
| Qiming | example: http://git.openstack.org/cgit/openstack/tempest/tree/tempest/api/compute/servers/test_delete_server.py#n148 | 06:43 |
| yanyanhu | and the performance/stress test is the lowest one | 06:43 |
| Qiming | ^ this is a negative test | 06:43 |
| elynn | ok, now we still lacks of negative tests for API :) | 06:45 |
| Qiming | oh, sorry, that one is not negative | 06:45 |
| Qiming | this one is: http://git.openstack.org/cgit/openstack/tempest/tree/tempest/api/compute/servers/test_servers_negative.py | 06:45 |
| yanyanhu | elynn, yes. I think API test should be covered using tempest | 06:45 |
| elynn | let me propose a patch to rename the dir structure and let's see if that is what you want. | 06:46 |
| Qiming | great | 06:46 |
| yanyanhu | as well as scenario test which includes interaction between senlin and other openstack services like nova | 06:46 |
| yanyanhu | thanks :) | 06:46 |
| Qiming | we will have a rest client shared by all api tests, functional tests, scenario tests, and integration tests | 06:46 |
| yanyanhu | functional test is very fast and user can easily run it without complete openstack env locally | 06:47 |
| yanyanhu | to verify their change | 06:47 |
| yanyanhu | Qiming, that will be great | 06:47 |
| Qiming | if we have good coverage of all api calls in API tests, we can make functional tests more focused on engine/profile/policies | 06:47 |
| yanyanhu | yes | 06:48 |
| elynn | Qiming, Can we use the rest client from tempest ? https://github.com/openstack/senlin/blob/master/senlin/tests/tempest_tests/services/clustering/clustering_client.py | 06:48 |
| yanyanhu | it doesn't need to cover API part | 06:48 |
| yanyanhu | and also driver | 06:48 |
| Qiming | yes, that is the rest client we will use | 06:48 |
| Qiming | some of yanyan's previous work were actually scenario tests I think | 06:49 |
| yanyanhu | Qiming, yes | 06:50 |
| yanyanhu | we need to split them from existing functional test | 06:50 |
| Qiming | for example a cluster attached with scaling policy and we try scale it in that context | 06:50 |
| yanyanhu | and relocate them :) | 06:50 |
| Qiming | that sounds great | 06:50 |
| elynn | okay, now I know the gaps of what we have now and what you want in future. | 06:51 |
| Qiming | the most complicated test would be integration test | 06:51 |
| yanyanhu | and we can make further discussion in summit to decide the test infrastructure of Senlin | 06:51 |
| elynn | I will focus on API tests for now, then consider other tests. | 06:52 |
| Qiming | where we will be using real heat/nova drivers to exercise some stories | 06:52 |
| yanyanhu | elynn, once tempest support is done, adding more tests will be easy | 06:52 |
| Qiming | sure, that would be really great | 06:52 |
| Qiming | agreed | 06:52 |
| yanyanhu | elynn, find an interesting phenomenon. After I reinstall tempest, I have to reinstall Senlin as well to make tempest discover Senlin's test cases | 06:54 |
| elynn | Hmm, that | 06:55 |
| elynn | that's interesting... | 06:55 |
| yanyanhu | :) | 06:55 |
| yanyanhu | not sure whether this is a strict limitation | 06:55 |
| elynn | You can open a bug first in case I forget it :) | 06:55 |
| yanyanhu | ah, looks like it only happened when tempest version changed | 06:56 |
| yanyanhu | elynn, ok | 06:56 |
| elynn | Most likely it's not a bug in senlin :) | 06:57 |
| yanyanhu | yes, I don't think so | 06:57 |
| *** yuanying has quit IRC | 06:57 | |
| openstackgerrit | Ethan Lynn proposed openstack/senlin: Redefine the tempest tests directory structure https://review.openstack.org/303264 | 07:08 |
| elynn | yanyanhu, Qiming ^^ | 07:09 |
| Qiming | looks good to me | 07:33 |
| Qiming | just curious about the services subdir | 07:34 |
| Qiming | and the services/clustering subdir | 07:34 |
| Qiming | files under them can be gathered into common as well? | 07:34 |
| Qiming | elynn, based on this: http://logs.openstack.org/55/302355/1/experimental/gate-senlin-dsvm-functional/74761d5/logs/screen-sl-eng.txt.gz#_2016-04-08_06_34_56_539 | 07:39 |
| Qiming | I think we should at least catch the DBDuplicateEntry exception somehow | 07:40 |
| Qiming | the context manager way of handling DB concurrency is not a silver bullet | 07:41 |
| Qiming | for example, it won't work because in our api pipeline, trusts are check before the current thread/eventlet even gets a context | 07:41 |
| Qiming | so a simple workaround could be just catch the exception and let it pass | 07:42 |
| elynn | Qiming, hmm, that could be a quick fix. | 07:45 |
| elynn | about services/clustering | 07:45 |
| elynn | I'm not so sure, just because all other projects using this structure. I might need some check. | 07:46 |
| elynn | Will leave for a while. | 07:46 |
| *** elynn has quit IRC | 07:47 | |
| openstackgerrit | Merged openstack/senlin: Define context.roles with base class https://review.openstack.org/302355 | 07:55 |
| openstackgerrit | Yanyan Hu proposed openstack/senlin: Reorganize profile example for nova server https://review.openstack.org/303279 | 08:11 |
| openstackgerrit | Yanyan Hu proposed openstack/senlin: Reorganize profile example dir https://review.openstack.org/303279 | 08:18 |
| *** openstackgerrit has quit IRC | 08:33 | |
| *** openstackgerrit has joined #senlin | 08:34 | |
| openstackgerrit | Qiming Teng proposed openstack/senlin: Add unit test for event module (1) https://review.openstack.org/303287 | 08:37 |
| *** zzxwill has quit IRC | 08:56 | |
| *** zzxwill has joined #senlin | 08:58 | |
| *** zzxwill has quit IRC | 09:18 | |
| *** zzxwill has joined #senlin | 09:19 | |
| yanyanhu | tried to use Rally plugin, but it totally didn't work as described in Rally document... | 09:32 |
| yanyanhu | sigh | 09:32 |
| Qiming | ... | 09:36 |
| *** zzxwill has quit IRC | 09:38 | |
| yanyanhu | really depressed... | 09:40 |
| yanyanhu | trying to read the code of plugin loading in rally | 09:40 |
| yanyanhu | hope can find some hint | 09:40 |
| Qiming | seems not a piece of cake | 09:46 |
| *** yuanying has joined #senlin | 09:50 | |
| xuhaiwei | Qiming, sent you a mail to your company mail box | 09:53 |
| Qiming | got it | 09:53 |
| Qiming | haven't got cycles to read it carefully, will do so when back home | 09:54 |
| yanyanhu | finally, it works... | 09:54 |
| xuhaiwei | ok, thanks | 09:54 |
| Qiming | in the right way, yanyanhu? | 09:54 |
| yanyanhu | perfect, if so, we only need the first patch to Rally | 09:54 |
| yanyanhu | yes | 09:54 |
| yanyanhu | we can hold the plugin by ourself | 09:54 |
| Qiming | great | 09:54 |
| yanyanhu | no need to contribute them to Rally repo | 09:54 |
| Qiming | glad we figured out the 'right' way to save our asses | 09:54 |
| Qiming | :) | 09:54 |
| yanyanhu | much less dependency on Rally repo | 09:54 |
| yanyanhu | oh, it's still based on client now | 09:55 |
| yanyanhu | tempest way still have some problems... | 09:55 |
| yanyanhu | but I think this give us another choice | 09:55 |
| yanyanhu | we can easily adjust it inside Senlin | 09:55 |
| yanyanhu | it actually works as well for tempest scenario | 09:56 |
| Qiming | okay | 09:56 |
| yanyanhu | so we can use plugin to support Rally test of Senlin(either using client or tempest) | 09:56 |
| yanyanhu | a little bit like heat resource plugin | 09:57 |
| Qiming | right, that was what we expected | 09:57 |
| yanyanhu | yes | 09:58 |
| Qiming | \o/ | 09:58 |
| yanyanhu | will try to add more support to the plugin to see what we can do using rally now | 09:58 |
| Qiming | great | 09:59 |
| Qiming | starting to contribute to openstack-ansible now, ;) | 10:03 |
| yanyanhu | cool! | 10:03 |
| Qiming | it helps understand how it works | 10:03 |
| yanyanhu | maybe we can find a day to sit together and share our recent work :) | 10:04 |
| Qiming | yes | 10:04 |
| yanyanhu | maybe in R_lixh's meeting room :) | 10:04 |
| yanyanhu | to vmware :P | 10:05 |
| R_lixh | More Than Welcome :P | 10:05 |
| Qiming | yanyanhu, you do have a coffee addict | 10:05 |
| yanyanhu | heihei | 10:06 |
| R_lixh | :) | 10:06 |
| yanyanhu | ls | 10:07 |
| yanyanhu | mis click... | 10:07 |
| Qiming | bin | 10:07 |
| Qiming | lib | 10:07 |
| Qiming | etc | 10:08 |
| Qiming | usr | 10:08 |
| yanyanhu | 。。。 | 10:08 |
| Qiming | var | 10:08 |
| R_lixh | ... | 10:08 |
| Qiming | exit | 10:14 |
| *** Qiming has quit IRC | 10:18 | |
| *** yanyanhu has quit IRC | 10:18 | |
| *** yanyanhu has joined #senlin | 10:19 | |
| *** yanyanhu has quit IRC | 10:19 | |
| openstackgerrit | Merged openstack/senlin: [doc] Removed the invalid link for Module Index https://review.openstack.org/302180 | 11:02 |
| *** Qiming has joined #senlin | 11:23 | |
| nk2527 | xuhaiwei: I'm back | 12:28 |
| openstackgerrit | OpenStack Proposal Bot proposed openstack/senlin: Updated from global requirements https://review.openstack.org/303365 | 13:29 |
| *** zzxwill has joined #senlin | 15:31 | |
| *** zzxwill has quit IRC | 16:00 | |
| *** zzxwill has joined #senlin | 16:01 | |
| openstackgerrit | Nick Klenke proposed openstack/senlin: Fixed typos in the Mitaka Series Release Notes https://review.openstack.org/303526 | 16:22 |
| *** Qiming has quit IRC | 16:29 | |
| openstackgerrit | Merged openstack/senlin: Fixed typos in the Mitaka Series Release Notes https://review.openstack.org/303526 | 17:04 |
| *** zzxwill has quit IRC | 17:26 | |
| *** zzxwill has joined #senlin | 17:28 | |
| *** zzxwill has quit IRC | 17:38 | |
| *** zzxwill has joined #senlin | 17:43 | |
| *** zzxwill has quit IRC | 18:31 | |
| openstackgerrit | OpenStack Proposal Bot proposed openstack/senlin: Updated from global requirements https://review.openstack.org/303365 | 18:43 |
Generated by irclog2html.py 2.14.0 by Marius Gedminas - find it at mg.pov.lt!