| *** ruijie has joined #senlin | 01:05 | |
| *** yanyanhu has joined #senlin | 01:09 | |
| *** haiwei has joined #senlin | 01:48 | |
| *** guoshan has joined #senlin | 02:02 | |
| *** guoshan has quit IRC | 02:09 | |
| *** guoshan has joined #senlin | 02:09 | |
| *** XueFeng has joined #senlin | 02:13 | |
| XueFeng | QiMing, around? | 02:14 |
|---|---|---|
| *** elynn has joined #senlin | 02:16 | |
| elynn | Morning | 02:18 |
| XueFeng | Coverage test fail problem, it caused by " if cfg.CONF.debug is False:" | 02:22 |
| XueFeng | cf.CONF.debug is special | 02:24 |
| elynn | Where the code is? XueFeng | 02:24 |
| XueFeng | http://git.openstack.org/cgit/openstack/senlin/tree/senlin/engine/event.py#n61 | 02:25 |
| XueFeng | In unit test, python -m testtools.run senlin.tests.unit.engine.test_event.TestEvent.test__dump_guarded success. | 02:27 |
| elynn | Why it is special? | 02:27 |
| XueFeng | But in coverage, it will fail in line http://git.openstack.org/cgit/openstack/senlin/tree/senlin/tests/unit/engine/test_event.py#n116 | 02:29 |
| elynn | That's weird | 02:31 |
| XueFeng | Seems linde http://git.openstack.org/cgit/openstack/senlin/tree/senlin/tests/unit/engine/test_event.py#n107 can't overriwde cfg.CONF.debug in coverage test. | 02:31 |
| XueFeng | s/linde/line/g | 02:31 |
| XueFeng | So weird | 02:32 |
| XueFeng | This will cause coverage test cannot run. | 02:35 |
| elynn | Maybe they are reading two different objects, need more investigation. | 02:37 |
| elynn | But I notice that unit tests failed either today, caused by other problem | 02:37 |
| XueFeng | Have looked up cover.sh.coverage will test two branch | 02:39 |
| XueFeng | HEAD^ and cureent branch | 02:39 |
| XueFeng | If we remove http://git.openstack.org/cgit/openstack/senlin/tree/senlin/tests/unit/engine/test_event.py#n116, coverage test will run. | 02:40 |
| XueFeng | I think cfg.CONF.debug is specail. | 02:40 |
| XueFeng | We need set register_options for it.http://git.openstack.org/cgit/openstack/senlin/tree/senlin/tests/unit/engine/test_event.py#n26 | 02:41 |
| elynn | for event? | 02:46 |
| yanyanhu | XueFeng: the coverage test failed for event test failure | 02:54 |
| yanyanhu | just tried to debug it, but haven't found the root cause | 02:54 |
| yanyanhu | that event test passed in python test | 02:55 |
| yanyanhu | but failed for coverage test | 02:55 |
| XueFeng | If we want it pass.We can remove http://git.openstack.org/cgit/openstack/senlin/tree/senlin/tests/unit/engine/test_event.py#n116 | 02:57 |
| XueFeng | temporary | 02:57 |
| XueFeng | YanYan, yes, it can pass in py27 | 02:58 |
| XueFeng | But can't pass in coverage. I have do some debug for it.Seems root cause is some special about cfg.CONF.debug | 02:59 |
| yanyanhu | Yes, removing that test could be a workaround. Just we'd better figure out the root case. | 03:06 |
| yanyanhu | so you mean the configure option of debug could be set incorrect during test? | 03:07 |
| XueFeng | Yes, for coverage test, cfg.CONF.debug may not correct. | 03:08 |
| XueFeng | I have do some test for this. | 03:08 |
| yanyanhu | ok, any way to fix it? beside disabling that test case :) | 03:09 |
| XueFeng | Not yet | 03:12 |
| yanyanhu | ok... | 03:12 |
| XueFeng | I think it will have way soon | 03:12 |
| yanyanhu | XueFeng: ok, I will further dig it to seek better solution. If not, disabling that test case could be a workaround. | 03:14 |
| XueFeng | Hmm, will research too:) | 03:15 |
| openstackgerrit | XueFeng Liu proposed openstack/senlin: Fix coverage test failed problem https://review.openstack.org/416823 | 03:16 |
| openstackgerrit | XueFeng Liu proposed openstack/senlin: Fix coverage test failed problem https://review.openstack.org/416823 | 03:18 |
| *** haiwei has quit IRC | 03:22 | |
| openstackgerrit | Merged openstack/senlin: Hook service clean up to lock breaker https://review.openstack.org/416471 | 03:26 |
| *** Drago has quit IRC | 03:26 | |
| openstackgerrit | Qiming Teng proposed openstack/senlin: Explicitly spell out cloud backend choices https://review.openstack.org/416553 | 03:31 |
| openstackgerrit | Qiming Teng proposed openstack/senlin: Split API tests for cluster actions https://review.openstack.org/416554 | 03:32 |
| openstackgerrit | Qiming Teng proposed openstack/senlin: API test for cluster-collect https://review.openstack.org/416555 | 03:33 |
| openstackgerrit | Merged openstack/senlin: Remove unnecessary coding format in files https://review.openstack.org/416496 | 03:47 |
| openstackgerrit | miaohb proposed openstack/senlin: Api-ref update for filter "user_id" in receiver list https://review.openstack.org/416486 | 03:57 |
| *** guoshan has quit IRC | 03:57 | |
| *** Jeffrey4l has quit IRC | 04:06 | |
| *** Jeffrey4l has joined #senlin | 04:07 | |
| *** elynn has quit IRC | 04:26 | |
| openstackgerrit | miaohb proposed openstack/senlin: Api-ref update for filter "user_id" in receiver list https://review.openstack.org/416486 | 04:30 |
| openstackgerrit | Qiming Teng proposed openstack/senlin: Documentation on API changes https://review.openstack.org/416828 | 04:54 |
| *** guoshan has joined #senlin | 04:57 | |
| *** guoshan has quit IRC | 05:02 | |
| *** elynn has joined #senlin | 05:04 | |
| openstackgerrit | Qiming Teng proposed openstack/senlin: Explicitly spell out cloud backend choices https://review.openstack.org/416553 | 05:13 |
| *** Jeffrey4l_ has joined #senlin | 05:31 | |
| *** Jeffrey4l has quit IRC | 05:34 | |
| *** guoshan has joined #senlin | 05:38 | |
| *** guoshan has quit IRC | 05:43 | |
| *** guoshan has joined #senlin | 05:43 | |
| *** guoshan has quit IRC | 05:55 | |
| *** guoshan has joined #senlin | 06:06 | |
| *** guoshan has quit IRC | 06:19 | |
| openstackgerrit | Qiming Teng proposed openstack/senlin: Split API tests for cluster actions https://review.openstack.org/416554 | 06:26 |
| *** guoshan has joined #senlin | 06:35 | |
| openstackgerrit | Merged openstack/senlin: Explicitly spell out cloud backend choices https://review.openstack.org/416553 | 06:52 |
| openstackgerrit | Qiming Teng proposed openstack/senlin: API test for cluster-collect https://review.openstack.org/416555 | 06:55 |
| openstackgerrit | Qiming Teng proposed openstack/senlin: Split API tests for node actions https://review.openstack.org/416865 | 07:05 |
| openstackgerrit | lvdongbing proposed openstack/senlin: Revise navigation level of "6.3 Reviewing Patches" https://review.openstack.org/416871 | 07:21 |
| *** guoshan has quit IRC | 07:26 | |
| *** guoshan has joined #senlin | 07:27 | |
| openstackgerrit | Merged openstack/senlin: Split API tests for cluster actions https://review.openstack.org/416554 | 07:34 |
| openstackgerrit | Merged openstack/senlin: Revise navigation level of "6.3 Reviewing Patches" https://review.openstack.org/416871 | 07:34 |
| *** openstackgerrit has quit IRC | 07:50 | |
| *** zhurong has joined #senlin | 08:24 | |
| *** openstackgerrit has joined #senlin | 08:29 | |
| *** ChanServ sets mode: +v openstackgerrit | 08:29 | |
| openstackgerrit | Qiming Teng proposed openstack/senlin: Bump max micro-version for V1 API to 1.4 https://review.openstack.org/416884 | 08:29 |
| openstackgerrit | Qiming Teng proposed openstack/senlin: Fix operation handler invocation in node https://review.openstack.org/416885 | 08:29 |
| openstackgerrit | Qiming Teng proposed openstack/senlin: API test for node-operation call https://review.openstack.org/416886 | 08:29 |
| elynn | Hi Qiming yanyanhu | 08:33 |
| *** openstackgerrit has quit IRC | 08:33 | |
| elynn | https://github.com/openstack/python-senlinclient/commit/e01394344f5652020faee7c98b95d8832164b0cb | 08:33 |
| Qiming | yes? | 08:33 |
| elynn | I notice this change in senlinclient, when update_obj , we need to pass a class as parameter? | 08:34 |
| Qiming | problems caused by the changes at sdk side | 08:34 |
| elynn | Why did they do this change? | 08:34 |
| elynn | It will cause two API calls each time. one for get, one for update... | 08:35 |
| Qiming | lemme point you to the dilemma | 08:36 |
| Qiming | http://git.openstack.org/cgit/openstack/python-openstacksdk/tree/openstack/proxy2.py#n46 | 08:37 |
| Qiming | all methods in the new proxy (proxy2) are invoking _get_resource() to build a resource object | 08:38 |
| Qiming | the logic in that function, check if the value you provided is an ID, an object or None | 08:38 |
| Qiming | if it is an ID | 08:38 |
| Qiming | the behavior is to do a resource_type.new() call | 08:39 |
| Qiming | and that new call is making the 'id' property in the resource dirty | 08:39 |
| Qiming | later when you are invoking the xyz_update() call to perform an update, the 'id' is contained in the request body | 08:39 |
| Qiming | I believe team was trying to do things in a more transparent, cleaner way, but ... the reality, as I see it, is dirtier ... | 08:40 |
| Qiming | it is a little tricky to change the logic, however, in senlinclient shell, we were always doing a GET already, so the change was trivial | 08:41 |
| Qiming | I agree this is not near sub-optimal, however, that is life | 08:41 |
| elynn | senlin will reject the request which contain id in body? | 08:42 |
| Qiming | yes | 08:42 |
| Qiming | we don't allow any additional parameters in the request | 08:42 |
| elynn | okay, just to make sure i'm on the right way to use sdk in this patch https://review.openstack.org/#/c/416408/ . | 08:43 |
| Qiming | http://git.openstack.org/cgit/openstack/senlin/tree/senlin/objects/fields.py#n118 | 08:44 |
| *** openstackgerrit has joined #senlin | 08:45 | |
| *** ChanServ sets mode: +v openstackgerrit | 08:45 | |
| openstackgerrit | Merged openstack/senlin: Fix coverage test failed problem https://review.openstack.org/416823 | 08:45 |
| openstackgerrit | XueFeng Liu proposed openstack/senlin: Add test for _create_nodes https://review.openstack.org/416893 | 08:57 |
| openstackgerrit | XueFeng Liu proposed openstack/senlin: Add test for _create_nodes https://review.openstack.org/416893 | 09:00 |
| yanyanhu | hi, elynn, I saw your patch to bump the version of senlinclient | 09:04 |
| yanyanhu | it is mainly for the patch in heat side? | 09:04 |
| elynn | hi yanyanhu | 09:04 |
| elynn | yes | 09:04 |
| elynn | I'm also working on replace senlinclient with sdk, but haven't succeed yet. | 09:05 |
| yanyanhu | sorry, just noticed your last msg, I'm now using web irc. Not as convenient as xchat... | 09:05 |
| yanyanhu | I see | 09:05 |
| elynn | Do you think it | 09:05 |
| elynn | it's necessary? | 09:05 |
| yanyanhu | sdk should always be the better choice in my own opinion | 09:06 |
| yanyanhu | for all services, not only for us | 09:06 |
| elynn | From client side, I think if they wanna use senlin commands, they should use the latest version. | 09:07 |
| elynn | yes, I agree with you. | 09:07 |
| yanyanhu | just there are too many difference between different service apis... it doesn't work as expect sometimes | 09:07 |
| yanyanhu | elynn: yes, we should update the global requirement for senlinclient | 09:07 |
| elynn | I'm working on a patch to use sdk instead of senlinclient in heat, just need some time to get it work. | 09:10 |
| Qiming | elynn, one quetion you will be asked is why we have to bump the lowest version requirement for senlinclient | 09:10 |
| elynn | I hate questions.. | 09:11 |
| elynn | So they don't bump lowest version now? | 09:11 |
| Qiming | you have to prepare an answer in this case | 09:12 |
| elynn | At least they wanna use *-update, they might need 1.1.0... | 09:12 |
| Qiming | that means there are incompatibilities introduced | 09:12 |
| Qiming | and the old version has to be abandoned | 09:12 |
| yanyanhu | so, maybe just update the upper constraint? | 09:13 |
| ruijie | the latest version of oslo.versionedobject is 1.17 | 09:13 |
| ruijie | that need to be updated to 1.19? | 09:13 |
| Qiming | upper constraint is already 1.1.0 | 09:13 |
| elynn | upper constraint is the latest | 09:13 |
| yanyanhu | but if we do need new features supported only by 1.1.0, we can bump the lowest version | 09:13 |
| elynn | okay, Now I see the point | 09:14 |
| Qiming | ruijie, oslo.* packages are shared by all openstack projects | 09:14 |
| elynn | features are the only reasons to bump the lowest version, isn't it | 09:14 |
| Qiming | you cannot bump its version without a good reason | 09:14 |
| Qiming | it may break some other projects | 09:15 |
| Qiming | it will be at the discretion of the oslo.versionedobjects team | 09:15 |
| Qiming | elynn, incompatibility is | 09:15 |
| ruijie | okay, Qiming | 09:15 |
| Qiming | if there are new features, in an ideal case, those features should be back ported, or version negotiation should be employed | 09:16 |
| Qiming | it is not a compelling reason for version bumping | 09:16 |
| elynn | Do you think what is a good reason? Qiming | 09:16 |
| elynn | incompatibility? | 09:17 |
| Qiming | yes, or else why are you throwing away old versions? | 09:17 |
| Qiming | it will break existing users who are using the old versions, right? | 09:17 |
| elynn | yes | 09:18 |
| yanyanhu | so keeping backward compatibility is our responsibility | 09:18 |
| elynn | *-update is some kind of incompatibility, right? | 09:18 |
| yanyanhu | as the maintainer of senlinclient | 09:19 |
| Qiming | yes | 09:19 |
| Qiming | the update behavior has changed, due to upstream changes in sdk | 09:19 |
| yanyanhu | before incompatibility happens, deprecation warning should always be given | 09:19 |
| Qiming | lower version of senlinclient won't work on the newer version of SDK | 09:19 |
| Qiming | yes, yanyanhu, for most cases, a deprecation should be kept there for at least one year, i.e. two cycles | 09:20 |
| Qiming | that will give users sufficient time to do the transition | 09:20 |
| yanyanhu | Qiming: if so, actually sdk doesn't well control their compatibiltiy... for its updating breaks senlinclient | 09:20 |
| yanyanhu | yes | 09:21 |
| Qiming | yes, I am not 100% with the resource2 changes | 09:21 |
| yanyanhu | understand | 09:21 |
| Qiming | e.g. 0.9.11 was marked a failed version | 09:21 |
| Qiming | because the migration to resource2 is not backward compatible | 09:22 |
| Qiming | the new resource2.Resource cannot be treated as a dict | 09:22 |
| yanyanhu | yes, noticed your discussion about that to_dict issue | 09:22 |
| yanyanhu | it's really a headache | 09:22 |
| Qiming | i.e. it is no longer providing __items__() calls | 09:22 |
| Qiming | I wasn't able to convince brian to add back those methods | 09:23 |
| Qiming | so the workaround is adding a to_dict() method for use cases we are aware of | 09:23 |
| yanyanhu | yes... | 09:23 |
| Qiming | after the transition of network resources to resource2 | 09:23 |
| Qiming | openstackclient code was broken | 09:23 |
| Qiming | and the puppet communit started to complain that OSC is broken | 09:24 |
| yanyanhu | yes... saw that bug report... | 09:24 |
| Qiming | so ... SDK 0.9.11 was marked an unusable version ... | 09:24 |
| yanyanhu | actually we had replied on 0.9.11 to recover our client | 09:25 |
| Qiming | it involves a lot of gobal viewpoints | 09:25 |
| yanyanhu | so it's not easy to maintain sdk :) | 09:25 |
| Qiming | that is one of the reason why we are so strictive to senlin APIs | 09:25 |
| elynn | Maybe I should update my commit message to explain why we need to bump version | 09:25 |
| Qiming | yes, that will help, elynn | 09:26 |
| yanyanhu | yep, elynn, dirk asked for it :) | 09:26 |
| yanyanhu | you should update the commit msg | 09:26 |
| yanyanhu | Qiming: yes, thoese efforts are definitely worthy | 09:27 |
| elynn | updated. | 09:28 |
| Qiming | It won't be an easy task, but once developers are well trained to understand the necessity of managing API changes, they will appreciate what we have done in the early days | 09:28 |
| yanyanhu | yep | 09:29 |
| *** openstackgerrit has quit IRC | 09:33 | |
| *** openstackgerrit has joined #senlin | 09:34 | |
| *** ChanServ sets mode: +v openstackgerrit | 09:34 | |
| openstackgerrit | Qiming Teng proposed openstack/senlin: API test for node-operation call https://review.openstack.org/416886 | 09:34 |
| *** zhurong has quit IRC | 09:40 | |
| *** yanyanhu_ has joined #senlin | 09:43 | |
| *** yanyanhu has quit IRC | 09:43 | |
| openstackgerrit | Qiming Teng proposed openstack/senlin: Explicitly state that 406 will be raised https://review.openstack.org/416908 | 09:48 |
| openstackgerrit | Xinhui Li proposed openstack/senlin: Extend Senlin to support mistral workflow service https://review.openstack.org/414979 | 10:08 |
| *** elynn has quit IRC | 10:08 | |
| openstackgerrit | Xinhui Li proposed openstack/senlin: Extend Senlin to support mistral workflow service https://review.openstack.org/414979 | 10:09 |
| *** openstackstatus has quit IRC | 10:13 | |
| *** openstack has joined #senlin | 10:18 | |
| *** guoshan has quit IRC | 10:41 | |
| openstackgerrit | XueFeng Liu proposed openstack/senlin: Remove conditional "if child:" in _create_nodes https://review.openstack.org/416893 | 11:18 |
| openstackgerrit | XueFeng Liu proposed openstack/senlin: Remove conditional "if child:" in _create_nodes https://review.openstack.org/416893 | 11:21 |
| *** guoshan has joined #senlin | 11:42 | |
| *** guoshan has quit IRC | 11:47 | |
| *** yanyanhu_ has quit IRC | 12:05 | |
| *** ruijie has quit IRC | 12:27 | |
| *** catintheroof has quit IRC | 12:38 | |
| *** catintheroof has joined #senlin | 12:44 | |
| *** catintheroof has quit IRC | 12:57 | |
| *** catintheroof has joined #senlin | 12:58 | |
| *** catintheroof has quit IRC | 13:02 | |
| *** catintheroof has joined #senlin | 13:04 | |
| *** catintheroof has quit IRC | 13:11 | |
| *** catintheroof has joined #senlin | 13:47 | |
| *** Jeffrey4l_ has quit IRC | 14:21 | |
| *** Jeffrey4l_ has joined #senlin | 14:35 | |
| *** Jeffrey4l_ has quit IRC | 14:52 | |
| *** Jeffrey4l_ has joined #senlin | 15:05 | |
| *** Drago has joined #senlin | 15:06 | |
| openstackgerrit | XueFeng Liu proposed openstack/senlin: Unit test for ClusterAction.do_update https://review.openstack.org/417007 | 15:14 |
| *** Drago has quit IRC | 15:20 | |
| *** catintheroof has quit IRC | 15:32 | |
| *** Jeffrey4l_ has quit IRC | 15:36 | |
| *** Drago has joined #senlin | 15:39 | |
| *** guoshan has joined #senlin | 15:46 | |
| *** Jeffrey4l_ has joined #senlin | 15:48 | |
| *** guoshan has quit IRC | 15:50 | |
| openstackgerrit | Merged openstack/senlin: Extend Senlin to support mistral workflow service https://review.openstack.org/414979 | 15:52 |
| *** catintheroof has joined #senlin | 16:32 | |
| *** guoshan has joined #senlin | 17:47 | |
| *** guoshan has quit IRC | 17:52 | |
| openstackgerrit | OpenStack Proposal Bot proposed openstack/senlin-dashboard: Updated from global requirements https://review.openstack.org/417176 | 20:31 |
| *** catintheroof has quit IRC | 20:45 | |
| *** guoshan has joined #senlin | 20:49 | |
| *** guoshan has quit IRC | 20:54 | |
| *** Jeffrey4l_ has quit IRC | 21:35 | |
| *** Jeffrey4l_ has joined #senlin | 21:48 | |
| *** guoshan has joined #senlin | 21:50 | |
| *** guoshan has quit IRC | 21:55 | |
| *** guoshan has joined #senlin | 22:51 | |
| *** guoshan has quit IRC | 22:56 | |
| *** Drago has quit IRC | 23:23 | |
| *** guoshan has joined #senlin | 23:52 | |
| *** guoshan has quit IRC | 23:57 | |
Generated by irclog2html.py 2.14.0 by Marius Gedminas - find it at mg.pov.lt!