| *** lixinhui_ has quit IRC | 00:31 | |
| *** Liu has quit IRC | 01:11 | |
| *** Qiming has joined #senlin | 01:26 | |
| openstackgerrit | Yanyan Hu proposed stackforge/senlin: [WIP]Fix gate job of Senlin functional test https://review.openstack.org/211813 | 01:34 |
|---|---|---|
| *** Yanyanhu has joined #senlin | 01:35 | |
| *** mathspanda has joined #senlin | 01:57 | |
| Qiming | https://review.openstack.org/#/c/212205/ | 02:00 |
| Qiming | this is insane | 02:00 |
| Qiming | my comments may be too harsh ... | 02:01 |
| Yanyanhu | this is the guy who also gave comment on the spec we saw yesterday? | 02:02 |
| Yanyanhu | oh, just the owner of that spec | 02:03 |
| *** ChrisSen has joined #senlin | 02:11 | |
| *** lkarm has joined #senlin | 02:33 | |
| *** lkarm has quit IRC | 02:37 | |
| *** jdandrea has quit IRC | 02:54 | |
| openstackgerrit | xu-haiwei proposed stackforge/senlin: Make ignore_missing default to False https://review.openstack.org/211397 | 03:16 |
| openstackgerrit | Merged stackforge/senlin: Revised node implementation with test cases https://review.openstack.org/212084 | 04:04 |
| openstackgerrit | Merged stackforge/senlin: Updated from global requirements https://review.openstack.org/212285 | 04:11 |
| *** Tennyson has joined #senlin | 04:32 | |
| *** lkarm has joined #senlin | 05:00 | |
| *** lkarm has quit IRC | 05:04 | |
| openstackgerrit | Merged stackforge/senlin: Polish stack profile unit test https://review.openstack.org/211815 | 05:11 |
| openstackgerrit | Merged stackforge/senlin: Make ignore_missing default to False https://review.openstack.org/211397 | 05:14 |
| Qiming | just proposed this | 05:15 |
| Qiming | https://review.openstack.org/212336 | 05:15 |
| Yanyanhu | nice, this is the correct thing I think | 05:15 |
| openstackgerrit | Merged stackforge/senlin: RPC client support to triggers https://review.openstack.org/211989 | 05:20 |
| Qiming | ummm, I had a second thought | 05:34 |
| Qiming | maybe the emulation of find-by-name should be done at sdk layer | 05:34 |
| Yanyanhu | agree with this | 05:35 |
| Yanyanhu | actually what sdk is doing now is not wrong, just because the query logic in different service is distinct | 05:36 |
| Yanyanhu | thus cause the find() method can't work as expected | 05:36 |
| Qiming | the problem is .... find is a classmethod | 05:37 |
| Qiming | can we override it easily in Flavor? | 05:38 |
| Yanyanhu | hmm, I think so | 05:38 |
| Yanyanhu | this is much better than implementing this logic in nova driver | 05:39 |
| Yanyanhu | maybe we can also do similar thing for other nova resource, then we can complete remove those get_xxx_by_name | 05:40 |
| openstackgerrit | Yanyan Hu proposed stackforge/senlin: [WIP]Fix gate job of Senlin functional test https://review.openstack.org/211813 | 05:41 |
| xuhaiwei | I think we should support find-by-name, maybe we should first LIST the resource and then query from the list? | 05:54 |
| Qiming | xuhaiwei, yes, find-by-name is an important feature for flavor resource | 05:54 |
| Qiming | the only question is where should we implement it | 05:55 |
| Qiming | ideally, nova | 05:55 |
| Qiming | sub-optimal, sdk | 05:55 |
| Qiming | finally, senlin | 05:55 |
| xuhaiwei | yes, but I think the easiest way is doing in senlin | 05:55 |
| Qiming | I'd try sdk first | 05:56 |
| Qiming | from sdk user's perspective, we would like it to hide the difference between services | 05:56 |
| Qiming | or else why are we using it | 05:57 |
| xuhaiwei | yes | 05:57 |
| Qiming | it's not easy to override classmethod as far as I can see | 05:57 |
| Qiming | I'm doing some experiments to try it out | 05:57 |
| Qiming | something like this: | 05:58 |
| Qiming | 21 @classmethod | 05:58 |
| Qiming | 22 def class_method(cls): | 05:58 |
| Qiming | 23 super(Child, cls).class_method() | 05:58 |
| Qiming | 24 print "child class_method: cls = %s x = %s" % (cls, cls.x) | 05:58 |
| Yanyanhu | so it's different from overriding a normal method of a class? | 06:00 |
| Qiming | not sure | 06:00 |
| Qiming | overriding classmethod may be different from overriding instance method | 06:01 |
| Qiming | trying | 06:01 |
| Yanyanhu | ok | 06:01 |
| Yanyanhu | trying to build a clean VM to test the senlinclient installation in devstack env | 06:01 |
| Yanyanhu | hope can find some reasons | 06:02 |
| Qiming | ok | 06:02 |
| *** Tennyson has quit IRC | 06:42 | |
| *** Tennyson has joined #senlin | 06:42 | |
| Qiming | changed my minde | 06:43 |
| Qiming | revised the SDK patch: https://review.openstack.org/212336 | 06:43 |
| Qiming | it is not that difficult | 06:44 |
| Qiming | we need to test if it works | 06:44 |
| Yanyanhu | yes, this is better | 06:45 |
| Yanyanhu | will make test using it | 06:45 |
| Qiming | shit | 06:45 |
| Qiming | SDK is throwing DuplicateResource for this | 06:45 |
| Qiming | hold on your test | 06:45 |
| Yanyanhu | ok | 06:45 |
| Qiming | fine now | 06:47 |
| Qiming | patch3 should be okay | 06:47 |
| Qiming | what we learned from their experience ... don't introduce new exception types unless absolutely necessary | 06:48 |
| Yanyanhu | ok | 06:48 |
| Qiming | it you pollute your code like kind of virus | 06:48 |
| openstackgerrit | Qiming Teng proposed stackforge/senlin: Replace flavor_get_by_name with flavor_find https://review.openstack.org/212028 | 06:51 |
| Yanyanhu | hi, Qiming, it works | 06:54 |
| Qiming | cool | 06:54 |
| Qiming | hope sdk guys won't force me to write some ... tests | 06:55 |
| Yanyanhu | if so, maybe we can just leave the flavor_find method | 06:55 |
| Yanyanhu | ... | 06:55 |
| Qiming | yes | 06:55 |
| Qiming | it is a problem that has better solved by sdk | 06:55 |
| Yanyanhu | yes | 06:55 |
| *** lkarm has joined #senlin | 07:16 | |
| *** lkarm has quit IRC | 07:20 | |
| *** Tennyson has quit IRC | 07:24 | |
| xuhaiwei | Qiming, the DuplicateResource exception is not HttpException type, so we should take care of it in senlin I think | 07:25 |
| *** Tennyson has joined #senlin | 07:25 | |
| Qiming | yes, xuhaiwei | 07:27 |
| Qiming | or we should help SDK kill it | 07:27 |
| xuhaiwei | before we can kill it, we should catch it and make it an internalError, besides this exception, ResourceTimeout and ResourceFailure seem to happen in some cases too | 07:32 |
| *** Tennyson has quit IRC | 07:35 | |
| *** Tennyson has joined #senlin | 07:35 | |
| Qiming | okay, those ones needs to be caught as well | 07:38 |
| *** openstackgerrit has quit IRC | 07:41 | |
| *** openstackgerrit has joined #senlin | 07:42 | |
| *** openstackgerrit has quit IRC | 07:45 | |
| *** Qiming_ has joined #senlin | 07:49 | |
| *** Qiming has quit IRC | 07:49 | |
| *** openstackgerrit has joined #senlin | 08:00 | |
| openstackgerrit | Qiming Teng proposed stackforge/senlin: Unit test for custom action module https://review.openstack.org/212398 | 08:00 |
| openstackgerrit | Qiming Teng proposed stackforge/senlin: Remove 'wrap_exception' function https://review.openstack.org/212394 | 08:02 |
| *** Tennyson has quit IRC | 08:02 | |
| *** Tennyson has joined #senlin | 08:03 | |
| openstackgerrit | Yanyan Hu proposed stackforge/senlin: [WIP]Fix gate job of Senlin functional test https://review.openstack.org/211813 | 08:14 |
| Yanyanhu | hi, Qiming_, I think I found the reason why senlinclient installation failed in gate side. | 08:25 |
| Yanyanhu | in our devstack plugin, we use git_clone function of devstack to clone senlinclient code from repo to dest dir, like /opt/stack/new/python-senlinclient | 08:26 |
| Yanyanhu | http://git.openstack.org/cgit/stackforge/senlin/tree/devstack/lib/senlin#n137 | 08:26 |
| Yanyanhu | git_clone func will check whether the source url is a gerrit style url | 08:26 |
| Yanyanhu | http://git.openstack.org/cgit/openstack-dev/devstack/tree/functions-common | 08:26 |
| Yanyanhu | if so, it will do the job correctly | 08:27 |
| Yanyanhu | but if NOT, it will check whether an old repo has been there in the dest dir | 08:27 |
| Yanyanhu | if it found the dest dir is not existing | 08:27 |
| Yanyanhu | it will further check ERROR_ON_CLONE env variable | 08:28 |
| Qiming_ | okay | 08:28 |
| Qiming_ | so, solution? | 08:28 |
| Yanyanhu | if this ERROR_ON_CLONE is false, it will do a complete clone using clone_time | 08:28 |
| Yanyanhu | but if it is True, an error will be given and installation progress is cancelled... | 08:29 |
| Yanyanhu | in our local env, this variable is set to False by default | 08:29 |
| Yanyanhu | so the installation progress can complete correctly | 08:29 |
| Yanyanhu | but in gate, this value is always set to True by default | 08:29 |
| Yanyanhu | for this reason | 08:30 |
| Yanyanhu | http://git.openstack.org/cgit/openstack-infra/devstack-gate/tree/devstack-vm-gate.sh#n249 | 08:30 |
| Yanyanhu | so I think the solution could be manually change the value of this variable during the installation of senlinclient | 08:30 |
| *** Tennyson has quit IRC | 08:30 | |
| Yanyanhu | and change it back after client source code clone is finished | 08:31 |
| Qiming_ | cann't we just avoid using git_clone function? | 08:31 |
| Yanyanhu | I think this won't break the rule that gate wants to have | 08:31 |
| Yanyanhu | hmm, I think we can | 08:31 |
| Yanyanhu | but it is recommned by devstack | 08:31 |
| Yanyanhu | when any service try to implement its own plugin for devstack | 08:32 |
| Yanyanhu | it provides some control like timeout | 08:32 |
| *** Tennyson has joined #senlin | 08:32 | |
| Yanyanhu | or heat check | 08:32 |
| Yanyanhu | so I think we can directly use 'git clone', but maybe not as safe as 'git_clone' | 08:32 |
| Qiming_ | the gate is designed to install and check the service targeted (e.g. senlin), not the other software | 08:32 |
| Yanyanhu | right | 08:32 |
| Yanyanhu | this is the point | 08:32 |
| Yanyanhu | so I think maybe we can use a work around for code cloning of senlinclient | 08:33 |
| Qiming_ | git_clone is obviously not designed for cloning an "irrelevant" software | 08:33 |
| Qiming_ | sure | 08:33 |
| Yanyanhu | right | 08:33 |
| Yanyanhu | so any suggestion? | 08:33 |
| Qiming_ | 1. we don't need senlinclient for functional test, right? | 08:33 |
| Yanyanhu | right | 08:33 |
| Qiming_ | 2. even we do need it, we are expected to do a pip install, suppose python-senlinclient is released with a version | 08:34 |
| Yanyanhu | right | 08:34 |
| Yanyanhu | actually after we have the first release, this problem will go I think | 08:34 |
| Qiming_ | 3. without a release published, everything we do today is a hack, so we choose the quickest way | 08:34 |
| Yanyanhu | ok | 08:35 |
| Qiming_ | in this case, it makes good sense to avoid install senlinclient | 08:35 |
| Yanyanhu | ok, agree with this | 08:36 |
| Yanyanhu | so maybe we just remove the installing of senlinclient from devstack plugin and leave some comments in README? | 08:36 |
| Yanyanhu | to let people know how to enable it or manually do this job? | 08:37 |
| openstackgerrit | Yanyan Hu proposed stackforge/senlin: [WIP]Fix gate job of Senlin functional test https://review.openstack.org/211813 | 08:52 |
| openstackgerrit | Yanyan Hu proposed stackforge/senlin: [WIP]Fix gate job of Senlin functional test https://review.openstack.org/211813 | 08:56 |
| *** Tennyson has quit IRC | 08:57 | |
| *** mathspanda has quit IRC | 09:20 | |
| openstackgerrit | Yanyan Hu proposed stackforge/senlin: Fix gate job of Senlin functional test https://review.openstack.org/211813 | 09:38 |
| *** ChrisSen has quit IRC | 09:43 | |
| *** Liu has joined #senlin | 09:44 | |
| openstackgerrit | xu-haiwei proposed stackforge/senlin: Handle exceptions which happen in heat driver https://review.openstack.org/212441 | 09:47 |
| openstackgerrit | Qiming Teng proposed stackforge/senlin: Unit test for actions in engine service https://review.openstack.org/212443 | 09:48 |
| openstackgerrit | xu-haiwei proposed stackforge/senlin: Modify parse_exception method https://review.openstack.org/211457 | 09:50 |
| *** Yanyan has joined #senlin | 09:59 | |
| *** Yanyanhu has quit IRC | 09:59 | |
| *** Yanyan has quit IRC | 10:04 | |
| openstackgerrit | Qiming Teng proposed stackforge/senlin: Unit test for actions in engine service https://review.openstack.org/212443 | 10:13 |
| *** branw has joined #senlin | 10:14 | |
| openstackgerrit | Merged stackforge/senlin: Modify parse_exception method https://review.openstack.org/211457 | 10:21 |
| openstackgerrit | Merged stackforge/senlin: Add server_metadata_get and server_metadata_update method https://review.openstack.org/210958 | 10:21 |
| *** Qiming_ has quit IRC | 10:26 | |
| *** Liu has quit IRC | 10:35 | |
| *** Liu has joined #senlin | 10:48 | |
| *** Qiming_ has joined #senlin | 11:25 | |
| openstackgerrit | Merged stackforge/senlin: Use Senlin generic driver to manage openstack lbaas driver https://review.openstack.org/212391 | 12:00 |
| *** lkarm has joined #senlin | 12:13 | |
| openstackgerrit | Merged stackforge/senlin: Remove 'wrap_exception' function https://review.openstack.org/212394 | 12:15 |
| openstackgerrit | Qiming Teng proposed stackforge/senlin: Unit test for events in service engine https://review.openstack.org/212516 | 13:01 |
| *** openstackgerrit has quit IRC | 13:01 | |
| *** openstackgerrit has joined #senlin | 13:02 | |
| *** jdandrea has joined #senlin | 13:40 | |
| *** lkarm has quit IRC | 13:51 | |
| *** lkarm has joined #senlin | 13:52 | |
| *** jroyal has joined #senlin | 16:28 | |
| *** openstackgerrit has quit IRC | 16:31 | |
| *** openstackgerrit has joined #senlin | 16:32 | |
| *** Qiming_ has quit IRC | 16:39 | |
| *** jroyal has quit IRC | 17:10 | |
| *** jroyal has joined #senlin | 17:11 | |
| openstackgerrit | Navneet Gupta proposed stackforge/python-senlinclient: Add unit test cases for sdk module https://review.openstack.org/212704 | 19:09 |
| openstackgerrit | Navneet Gupta proposed stackforge/python-senlinclient: Add unit test cases for sdk module https://review.openstack.org/212704 | 20:12 |
| *** lkarm has quit IRC | 21:10 | |
| *** jroyal has quit IRC | 21:21 | |
| *** jroyal has joined #senlin | 21:22 | |
| *** jroyal has quit IRC | 21:26 | |
| *** jroyal has joined #senlin | 21:27 | |
| *** jroyal has quit IRC | 21:32 | |
| *** branw has quit IRC | 22:50 | |
| openstackgerrit | Merged stackforge/senlin: Policy types API test case fix https://review.openstack.org/212034 | 23:35 |
Generated by irclog2html.py 2.14.0 by Marius Gedminas - find it at mg.pov.lt!