| openstackgerrit | Sean Dague proposed a change to openstack/tempest: Revert "Add jsonschema to requirements.txt" https://review.openstack.org/39005 | 01:40 |
|---|---|---|
| *** woodspa has joined #openstack-qa | 02:19 | |
| openstackgerrit | A change was merged to openstack/tempest: Revert "Add jsonschema to requirements.txt" https://review.openstack.org/39005 | 02:25 |
| *** woodspa has quit IRC | 02:27 | |
| openstackgerrit | Alex Vasilyev proposed a change to openstack-dev/devstack: Add Midonet plugin support to devstack. https://review.openstack.org/36607 | 02:55 |
| openstackgerrit | Zhang Hua proposed a change to openstack-dev/devstack: enable apache2 server as front end for swift https://review.openstack.org/33946 | 03:12 |
| openstackgerrit | Zhang Hua proposed a change to openstack-dev/devstack: enable apache2 server as front end for swift https://review.openstack.org/33946 | 03:16 |
| afazekas | sdague, dkranz: new bug https://bugs.launchpad.net/tempest/+bug/1205980 | 03:39 |
| *** krtaylor has joined #openstack-qa | 03:44 | |
| *** Anju has joined #openstack-qa | 04:28 | |
| *** SergeyLukjanov has joined #openstack-qa | 04:32 | |
| *** prekarat has joined #openstack-qa | 04:35 | |
| openstackgerrit | A change was merged to openstack/tempest: skip stamp test until race in nova can be debugged https://review.openstack.org/38923 | 04:40 |
| openstackgerrit | afazekas proposed a change to openstack/tempest: Use assertIn and asserNotIn in the api tests https://review.openstack.org/39019 | 04:43 |
| openstackgerrit | afazekas proposed a change to openstack/tempest: Use assertIn and asserNotIn in the api tests https://review.openstack.org/39019 | 04:46 |
| openstackgerrit | afazekas proposed a change to openstack/tempest: Use assertIn and assertNotIn instead of assertTrue/assertFalse https://review.openstack.org/39019 | 04:50 |
| openstackgerrit | anju Tiwari proposed a change to openstack/tempest: Negative attribute added in volume negative test https://review.openstack.org/39021 | 05:13 |
| openstackgerrit | Chris Behrens proposed a change to openstack-dev/devstack: Fix various things when n-cell is enabled https://review.openstack.org/38897 | 05:17 |
| openstackgerrit | raiesmh08 proposed a change to openstack/tempest: Add network api test cases https://review.openstack.org/39023 | 05:20 |
| *** prekarat1 has joined #openstack-qa | 05:28 | |
| *** prekarat has quit IRC | 05:31 | |
| Anju | afazekas:ping | 05:46 |
| *** prekarat has joined #openstack-qa | 06:02 | |
| *** prekarat1 has quit IRC | 06:04 | |
| afazekas | Anju: I will be back in 1h | 06:07 |
| Anju | ok:afazekas | 06:07 |
| *** afazekas is now known as afazekas__ | 06:07 | |
| *** SergeyLukjanov has quit IRC | 06:09 | |
| openstackgerrit | Marc Koderer proposed a change to openstack/tempest: Add stress test to attach volumes to vm's https://review.openstack.org/36652 | 06:15 |
| *** sileht has quit IRC | 06:32 | |
| *** sileht has joined #openstack-qa | 06:33 | |
| mkoderer | afazekas__: give me a ping when available... great idea with the unit test stress tests ;) | 06:49 |
| *** afazekas has joined #openstack-qa | 07:04 | |
| Anju | afazekas:ping | 07:06 |
| Anju | on time | 07:07 |
| Anju | :) | 07:07 |
| afazekas | Anju: pong | 07:08 |
| Anju | how can i know what is the response of xml | 07:08 |
| afazekas | you can try using the curl command with different ACCEPT header | 07:09 |
| Anju | accept header? | 07:10 |
| afazekas | Anju: I need to eat something for breakfast | 07:15 |
| Anju | afazekas:sure | 07:16 |
| afazekas | the most cli commands has --debug option, and they are expecting json response, you can change the headers to xml to see an xml response | 07:16 |
| afazekas | the cli commands just using json, so you need to use another tool for xml requests | 07:19 |
| *** psedlak has joined #openstack-qa | 07:20 | |
| Anju | afazekas: tool? | 07:21 |
| afazekas | psedlak: ping | 07:35 |
| *** SergeyLukjanov has joined #openstack-qa | 07:36 | |
| Anju | afazekas:understood :) | 07:39 |
| Anju | have taken the respnse of these two | 07:39 |
| *** sileht has quit IRC | 07:39 | |
| *** sileht has joined #openstack-qa | 07:43 | |
| Anju | afazekas: https://review.openstack.org/#/c/39021/ | 07:51 |
| afazekas | psedlak: do you know an easy way for sending custom requests ? | 07:55 |
| *** jgallard has joined #openstack-qa | 07:55 | |
| *** Dmitry-I_ has joined #openstack-qa | 08:41 | |
| *** mkollaro has joined #openstack-qa | 08:44 | |
| openstackgerrit | Mate Lakat proposed a change to openstack-dev/devstack: xenapi: Uninstall instances https://review.openstack.org/39042 | 08:47 |
| *** SergeyLukjanov has quit IRC | 08:48 | |
| *** jgallard has quit IRC | 08:52 | |
| *** jgallard has joined #openstack-qa | 08:52 | |
| *** jgallard has quit IRC | 09:16 | |
| *** jgallard has joined #openstack-qa | 09:16 | |
| psedlak | Anju: hi, here is my script/tool which i use, it's not cleaned nor whatever-friendly yet (and that's why it's not published anywhere else) ... probably it can help you little bit http://fpaste.org/28545/09033213/ ... curl -o craftol.py "http://fpaste.org/28545/9033213/raw/" && chmod +x craftol.py && ./craftol.py --help | 09:35 |
| psedlak | Anju: usage is then like ==> rhosqe-craftol localhost:5000 POST/v3/domains/<domain-id> '{"domain":{"name":"new name"}}' | 09:36 |
| psedlak | Anju: it needs/uses same auth info in env as general keystone/nova/... commands ... it gets token and then issues your API request + prints results, headers, etc ... with "--debug 1" it will print all ongoing communication ... the output is formatted and collored so it simplifies debugging/examination of API requests little bit | 09:38 |
| Anju | psedlak:afaxekas xan help you | 09:38 |
| Anju | psedlak:afazekas xan help you | 09:38 |
| Anju | psedlak:afazekas can help you | 09:38 |
| psedlak | Anju: well afazekas asked me to provide you with my script as it can make API debugging easier for you :) | 09:39 |
| openstackgerrit | A change was merged to openstack-dev/devstack: xenapi: Setup and Rotate text console logs https://review.openstack.org/38444 | 09:39 |
| *** prekarat1 has joined #openstack-qa | 09:40 | |
| *** prekarat has quit IRC | 09:43 | |
| Anju | pseddlak:ohh thanks | 09:43 |
| Anju | psedlak:what i did | 09:44 |
| Anju | i dbug the cli | 09:44 |
| Anju | and take the reonse from them | 09:44 |
| Anju | i change the header from json to xml | 09:44 |
| Anju | psedlak: the other one is link is opening | 09:45 |
| Anju | the other one link is not opening | 09:52 |
| mkoderer | afazekas: | 10:11 |
| mkoderer | ping | 10:11 |
| afazekas | mkoderer: pong | 10:16 |
| *** SergeyLukjanov has joined #openstack-qa | 10:16 | |
| mkoderer | afazekas: thanks for your great idea :) | 10:17 |
| mkoderer | afazekas: did you see my comments.. maybe we could change it to become just integrated in the driver | 10:18 |
| mkoderer | so we don't need a special action for it | 10:18 |
| afazekas | This action can have more option, it is the first useful working version | 10:19 |
| afazekas | and I agree, the run call could be little nicer in the future | 10:20 |
| mkoderer | yes for all special option we could use the args thing | 10:21 |
| afazekas | I would like to add a skip attribute ignore trick in the next version | 10:21 |
| afazekas | The stop on the first failure is the logic I still miss from the stress test core | 10:27 |
| mkoderer | afazekas: yes it's on my tods | 10:46 |
| mkoderer | I already add it to the bp items | 10:47 |
| *** jgallard has quit IRC | 10:47 | |
| *** prekarat has joined #openstack-qa | 10:55 | |
| *** prekarat1 has quit IRC | 10:57 | |
| *** Dmitry-I_ has quit IRC | 10:57 | |
| *** psedlak_ has joined #openstack-qa | 10:58 | |
| openstackgerrit | A change was merged to openstack/tempest: Negative attribute added in volume negative test https://review.openstack.org/39021 | 10:58 |
| *** prekarat has quit IRC | 11:01 | |
| openstackgerrit | A change was merged to openstack/tempest: Add posargs to testr-full tox job https://review.openstack.org/38827 | 11:01 |
| *** psedlak has quit IRC | 11:01 | |
| *** afazekas has quit IRC | 11:04 | |
| *** afazekas has joined #openstack-qa | 11:24 | |
| *** hrebicek_ has joined #openstack-qa | 11:28 | |
| *** hrebicek has quit IRC | 11:32 | |
| *** meteorfox has quit IRC | 11:39 | |
| *** dkranz has quit IRC | 11:54 | |
| *** Dmitry-I has joined #openstack-qa | 11:58 | |
| *** psedlak__ has joined #openstack-qa | 12:00 | |
| *** psedlak_ has quit IRC | 12:04 | |
| *** woodspa has joined #openstack-qa | 12:22 | |
| *** meteorfox has joined #openstack-qa | 12:34 | |
| *** adalbas has joined #openstack-qa | 12:47 | |
| *** acfleury has joined #openstack-qa | 12:48 | |
| *** dkranz has joined #openstack-qa | 12:50 | |
| openstackgerrit | afazekas proposed a change to openstack/tempest: Use assertIn and assertNotIn instead of assertTrue/assertFalse https://review.openstack.org/39019 | 12:52 |
| *** jasondotstar has joined #openstack-qa | 12:53 | |
| openstackgerrit | afazekas proposed a change to openstack/tempest: Handle a possible volume attachment visibility wait/race issue https://review.openstack.org/38841 | 12:56 |
| openstackgerrit | afazekas proposed a change to openstack/tempest: Handle a possible volume attachment visibility wait/race issue https://review.openstack.org/38841 | 12:57 |
| openstackgerrit | Julie Pichon proposed a change to openstack-dev/devstack: Configure horizon and horizon URLs in Tempest https://review.openstack.org/39067 | 12:59 |
| *** rahmu has quit IRC | 13:17 | |
| *** rahmu has joined #openstack-qa | 13:18 | |
| *** afazekas has quit IRC | 13:21 | |
| *** acfleury has quit IRC | 13:21 | |
| openstackgerrit | afazekas proposed a change to openstack/tempest: Use tempest log with tempest/openstack/common/lockutils.py https://review.openstack.org/39071 | 13:27 |
| *** krtaylor has quit IRC | 13:28 | |
| *** afazekas has joined #openstack-qa | 13:36 | |
| *** afazekas has quit IRC | 13:39 | |
| *** afazekas has joined #openstack-qa | 13:39 | |
| sdague | dkranz, mtreinish, jaypipes: https://review.openstack.org/#/c/38859/ would be nice to land, it's an occational gate reset | 13:43 |
| jaypipes | sdague: reviewing now. | 13:43 |
| dkranz | sdague: Done. | 13:46 |
| *** prekarat has joined #openstack-qa | 13:54 | |
| *** meteorfox has quit IRC | 13:57 | |
| mkoderer | dkranz: ping | 14:05 |
| openstackgerrit | Matthew Treinish proposed a change to openstack/tempest: Add locking to test_aggregates https://review.openstack.org/38857 | 14:06 |
| openstackgerrit | Matthew Treinish proposed a change to openstack/tempest: Switch to using oslo logging https://review.openstack.org/38879 | 14:06 |
| mtreinish | afazekas: ping | 14:07 |
| afazekas | mtreinish: pong | 14:08 |
| mtreinish | afazekas: about https://review.openstack.org/#/c/39071/ we can't do that. That code is synced in from oslo and it's basically treated as out of tree | 14:10 |
| mtreinish | making local changes to it kind of breaks the whole model | 14:11 |
| openstackgerrit | nayna-patel proposed a change to openstack/tempest: Added 3 Routers related testcases for Neutron API https://review.openstack.org/36629 | 14:12 |
| *** meteorfox has joined #openstack-qa | 14:12 | |
| afazekas | mtreinish: which model ? | 14:13 |
| mtreinish | the oslo model | 14:13 |
| afazekas | Can you move the lockutil to the tempest/common ? | 14:14 |
| mtreinish | afazekas: that's not what we want to do either. The code is kept common, if there are changes you want to make it should be done in oslo then synced in | 14:14 |
| afazekas | Do you really need to change the logger everywhere ? | 14:15 |
| afazekas | AFAIK the oslo is for copy and modify when you need , otherwise it would be shared library | 14:16 |
| mtreinish | afazekas: I don't think that having 2 loggers is the right way to do it. IMO the point at which we need to use oslo.log it makes sense to move everything to it. | 14:17 |
| *** krtaylor has joined #openstack-qa | 14:17 | |
| afazekas | mtreinish: but we loose feature with this movement, without a gain | 14:17 |
| mtreinish | afazekas: which features do we loose? | 14:18 |
| afazekas | configurable logging | 14:18 |
| mtreinish | the oslo logging is configurable. Everything is done with config options with it. | 14:19 |
| afazekas | https://github.com/openstack/tempest/blob/master/tempest/common/log.py#L49 | 14:19 |
| afazekas | How can you pass the config to the oslo module without using a command line arguments ? | 14:20 |
| mtreinish | https://github.com/openstack/oslo-incubator/blob/master/openstack/common/log.py#L66 | 14:20 |
| mtreinish | cli_opt can also be specified in the config file | 14:20 |
| mtreinish | afazekas: also having the logging config be passed in as a environ is very weird IMO. It should be in a standard location or configurable in some other way | 14:21 |
| afazekas | we do it with the tempest config as well :) | 14:21 |
| mtreinish | afazekas: yeah I know, but at least that has a default location :) | 14:22 |
| mtreinish | afazekas: anyway using oslo.log just moves all of the logging options into the tempest.conf | 14:22 |
| afazekas | So you say everything will be able to work (even with unittest), what was able to work with this? https://github.com/openstack/tempest/blob/master/etc/logging.conf.sample | 14:25 |
| mtreinish | afazekas: I think it should, although I haven't tried to use it. | 14:26 |
| mtreinish | and if it doesn't then there is probably a bug to fix in oslo | 14:27 |
| *** burt has joined #openstack-qa | 14:28 | |
| afazekas | Or the oslo logging just not designed for complex unit test use case | 14:30 |
| *** adalbas has quit IRC | 14:31 | |
| afazekas | mtreinish: I have doubts it is really able to work | 14:33 |
| dkranz | mkoderer: Here now. | 14:38 |
| afazekas | mtreinish: Is it expects you configure the log_config in the DEFAULT section ? | 14:38 |
| *** mlavalle has joined #openstack-qa | 14:38 | |
| mtreinish | yeah | 14:39 |
| openstackgerrit | A change was merged to openstack-dev/devstack: Set lock_path for Tempest to TMPDIR https://review.openstack.org/38861 | 14:40 |
| *** meteorfox has quit IRC | 14:40 | |
| openstackgerrit | A change was merged to openstack/tempest: fix race condition in service list compares https://review.openstack.org/38859 | 14:41 |
| mkoderer | dkranz: do you think it's in general useless to mix up stress test and usual tempest test? | 14:42 |
| mkoderer | dkranz: since there are designed not to be used in the stress test framework? | 14:42 |
| dkranz | mkoderer: I do think that in the sense that I don't see what is really gained. | 14:43 |
| *** adalbas has joined #openstack-qa | 14:43 | |
| dkranz | mkoderer: We can already run targeted stress tests, or tempest in parallel with any tests we like. | 14:43 |
| dkranz | mkoderer: That seems enough to me. | 14:44 |
| dkranz | mkoderer: Introducing a new run-time dependency between stress and regular test cases just seems fragile and not worth the potential mainetnance | 14:44 |
| mkoderer | dkranz: but if you want to build a new stress test scenario you mostly duplicate code that is already in tempest | 14:45 |
| afazekas | mtreinish: if it working just add it to the tempest.conf.sample, I'll try it | 14:45 |
| dkranz | mkoderer: Yes, that is true. | 14:46 |
| dkranz | mkoderer: But I saw the stress cases as being very small resource bashing, not scenario-based. | 14:46 |
| dkranz | mkoderer: The tempest tests can still be run in parallel as a kind of stress test as well. | 14:47 |
| mkoderer | dkranz: ok I never used this parallel mode.. I have to have a look | 14:47 |
| dkranz | mkoderer: It's not a parellel mode. You just run tempest in n different processes from the command line. | 14:48 |
| mkoderer | dkranz: ok... that's easy | 14:48 |
| dkranz | mkoderer: You could ask why we need the existing stress cases at all. | 14:49 |
| mkoderer | dkranz: yes correct.. this is what I am thinking about | 14:49 |
| mkoderer | dkranz: I mean it gives you a better usability | 14:50 |
| mkoderer | since you get all the logs and a statistic | 14:50 |
| mtreinish | afazekas: also you don't have to use the python logging config format. oslo.log breaks out a bunch of useful config options to set the log formatting and output | 14:50 |
| dkranz | mkoderer: I think the reason is so you can apply stress on certain operations. | 14:50 |
| mkoderer | dkranz: so it's more a less a framework to find a bug in a certain area | 14:51 |
| dkranz | mkoderer: The unit test cases clean up after themselves which is a limiting case. | 14:51 |
| dkranz | mkoderer: Yes, and the "areas" that are the most sensitive are apis that allocate/deallocate resources. | 14:52 |
| dkranz | mkoderer: Like instances, volumes, ips, networks, etc. | 14:52 |
| dkranz | dkranz: That said, I am not strongly objjecting to this patch. | 14:54 |
| afazekas | mtreinish: as I remember the logger is tempest is originated from oslo, it just removed the cli options. (does not seams to works so far..) | 14:54 |
| mkoderer | dkranz: so what I would need is a framework that simulates realistic load | 14:55 |
| mkoderer | to find bug's before there occur in the production system | 14:55 |
| mkoderer | I have doubt's that both possible way will give me such a system | 14:56 |
| dkranz | mkoderer: Yes, but that may be hard to do. If stress on an unrealistic load passes that is good enough. | 14:56 |
| dkranz | mkoderer: Got to run to a meeting now. | 14:57 |
| mkoderer | dkranz: ok - thanks ;) | 14:57 |
| *** meteorfox has joined #openstack-qa | 14:58 | |
| *** fbo_away is now known as fbo | 14:58 | |
| *** fbo is now known as fbo_away | 14:59 | |
| afazekas | mtreinish: http://www.fpaste.org/28658/51100341/ It does not seams to work | 15:01 |
| *** meteorfox has quit IRC | 15:02 | |
| afazekas | Am I have a typo in the config ? | 15:02 |
| mtreinish | afazekas: I think its 'log-config' | 15:03 |
| mtreinish | https://github.com/openstack/oslo-incubator/blob/master/openstack/common/log.py#L66 | 15:03 |
| afazekas | mtreinish: not working either, in the nova sample it is used with _ | 15:06 |
| *** adalbas has quit IRC | 15:08 | |
| *** dkranz has quit IRC | 15:09 | |
| openstackgerrit | Alex Gaynor proposed a change to openstack-dev/devstack: Added libffi-dev to the needed packages https://review.openstack.org/39096 | 15:12 |
| mtreinish | afazekas: ok if it doesn't work then there is probably something to fix in oslo | 15:13 |
| mtreinish | but what options do we set in the logging.conf.sample that can't be set with the oslo.log config options? | 15:13 |
| *** meteorfox has joined #openstack-qa | 15:14 | |
| burt | tempest/devstack question: I'd like to write a tempest test for nova that depends on sync_power_state being run. It's periodic and the default interval is 600s. Any recommendations on handling this? I could submit a patch to change it to 30s, but I assume there's an overall performance hit. | 15:17 |
| afazekas | mtreinish: looks like it cannot use the tempest config file, and it even does not reads the specified config file | 15:17 |
| afazekas | mtreinish: in the config file, you can specify multiple handlers | 15:21 |
| *** adalbas has joined #openstack-qa | 15:24 | |
| *** dkranz has joined #openstack-qa | 15:24 | |
| openstackgerrit | Matthew Treinish proposed a change to openstack/tempest: Add locking for adding a ssh security group rule https://review.openstack.org/39099 | 15:26 |
| openstackgerrit | Bob Ball proposed a change to openstack-dev/devstack: Keep dependencies when replacing packages https://review.openstack.org/39100 | 15:26 |
| afazekas | mtreinish: we should not use the oslo log in tempest until it is not configurable | 15:29 |
| *** adalbas has quit IRC | 15:33 | |
| *** SergeyLukjanov has quit IRC | 15:35 | |
| *** SergeyLukjanov has joined #openstack-qa | 15:39 | |
| afazekas | dkranz: If you are lucky you can reuse one of the test cases for targeted run, the setUpClass and tearDown class an be called once, and you just repeat the test_method | 15:41 |
| mtreinish | afazekas: well I disagree, I think it's configurable even if the python logging config file isn't working. (which I'll look into) I still don't see which options we set in the sample that oslo logging doesn't expose through config options. | 15:41 |
| mtreinish | the multiple handlers can be set with other config options. (at least the ones that we are using, syslog, file, and stdout, it might be standard err in oslo but they're both usable and captured) | 15:42 |
| afazekas | mtreinish: It does not working because tempest loading it's config file in a different way | 15:42 |
| afazekas | mtreinish: may be it is possible to hack it do in a oslo compatible way..., last time when I checked (when the actual tempest logger added) , it does not seamed trivial | 15:43 |
| afazekas | dkranz: we still should have targetted test cases for floating ip association + security groups (with connect or ping scan), and volume attachment recolonization (guest side validation, over ssh (or telent ..)) stress test at least | 15:45 |
| afazekas | dkranz: the stress framework is usable to try to reproduce flaky issues with know failing tests IMHO, at a somewhat cleaner than the other ways | 15:47 |
| mtreinish | afazekas: fixed it, just had to add a call to logging.setup() to initialize the logger... | 15:49 |
| *** adalbas has joined #openstack-qa | 15:50 | |
| mtreinish | the projects handle that it the command scripts but since tempest doesn't have that we just need to once somewhere globally | 15:51 |
| *** bclifford has joined #openstack-qa | 15:52 | |
| afazekas | mtreinish: try to add it at the and of the singleton https://github.com/openstack/tempest/blob/master/tempest/config.py#L662, or call at module load time in the tempest/test.py | 15:54 |
| *** adalbas has quit IRC | 16:00 | |
| afazekas | or may be able to add it to some early __init__.py | 16:00 |
| *** Anju_ has joined #openstack-qa | 16:00 | |
| afazekas | or just had the old tempest logger to call it at the same time .. | 16:08 |
| afazekas | I mean on module load it initializes the oslo logger .. | 16:08 |
| afazekas | sdague: https://review.openstack.org/#/c/34866/ what is required for these test case to be merged ? | 16:14 |
| afazekas | sdague: they even does not spawn a VM | 16:15 |
| *** marun has joined #openstack-qa | 16:16 | |
| *** adalbas has joined #openstack-qa | 16:16 | |
| afazekas | https://review.openstack.org/#/c/34866/11/tempest/api/orchestration/stacks/test_non_empty_stack.py just this one creates only one. | 16:16 |
| *** adalbas has quit IRC | 16:27 | |
| *** dkranz has quit IRC | 16:40 | |
| *** adalbas has joined #openstack-qa | 16:43 | |
| *** afazekas has quit IRC | 16:44 | |
| *** prekarat has left #openstack-qa | 16:45 | |
| *** psedlak__ has quit IRC | 16:46 | |
| *** SergeyLukjanov has quit IRC | 16:49 | |
| *** dkranz has joined #openstack-qa | 16:53 | |
| *** adalbas has quit IRC | 16:54 | |
| *** hemnafk is now known as hemna | 16:57 | |
| *** adalbas has joined #openstack-qa | 17:10 | |
| *** SergeyLukjanov has joined #openstack-qa | 17:12 | |
| *** mlavalle has quit IRC | 17:28 | |
| *** burt has quit IRC | 17:34 | |
| dkranz | sdague, mtreinish : Why do we copy files from oslo rather than importing it as a library? Perhaps I don't understrand how it works. | 17:40 |
| sdague | dkranz: because that's how oslo works | 17:43 |
| sdague | oh, wait.. this is cfg isn't it | 17:44 |
| sdague | no, it's logging, never mind | 17:44 |
| sdague | only oslo.cfg has been spun out as a pypi release | 17:44 |
| sdague | the rest is managed copy / paste through the incubator | 17:44 |
| mtreinish | sdague: and sphinx, messaging, etc. | 17:45 |
| mtreinish | dkranz: https://wiki.openstack.org/wiki/Oslo | 17:45 |
| sdague | well sphinx was never in the incubator, that was straight to library | 17:45 |
| mtreinish | the things in oslo-incubator aren't considered mature/stable enough to be spun out as libraries yet, so they keep them managed copy and paste | 17:46 |
| mtreinish | it is a little weird | 17:46 |
| dkranz | mtreinish: OK :) | 17:47 |
| *** mlavalle has joined #openstack-qa | 18:00 | |
| *** Anju_ has quit IRC | 18:06 | |
| dkranz | sdague: I think the comments from Attila and Harika are correct. Can you withdraw your -1 from https://review.openstack.org/#/c/37194/3 ? | 18:22 |
| dkranz | sdague: We don't check specific return codes in class setup but failures will still raise exceptions. | 18:22 |
| dkranz | sdague: Same with https://review.openstack.org/#/c/37179/ I think. | 18:23 |
| sdague | dkranz: on the second one if they fail we leak resources | 18:26 |
| *** mkollaro has quit IRC | 18:26 | |
| sdague | on the first one, I at least want the extra line wrapping removed that just seems silly to add | 18:26 |
| *** mlavalle has left #openstack-qa | 18:27 | |
| dkranz | sdague: We don't do this kind of checking on other tearDownClass methods. | 18:32 |
| dkranz | sdague: If a delete fails aren't we hosed anyway? | 18:33 |
| sdague | dkranz: but if a delete fails, all the rest of the deletes aren't tried | 18:33 |
| dkranz | sdague: But why is that such a big deal? | 18:34 |
| dkranz | sdague: If a delete fails who knows what the state of the system is. We should just stop in that case. | 18:34 |
| sdague | because we have a high priority blueprint to stop leaking :) | 18:34 |
| sdague | we can try to actually clean up everything | 18:35 |
| dkranz | sdague: Yes, in a successful run. | 18:35 |
| sdague | I want us to actually try to delete every resource we create | 18:35 |
| sdague | if a delete fails, fine | 18:35 |
| sdague | but we shouldn't let it cascade and prevent other things from attempting to delete | 18:35 |
| dkranz | sdague: We have many dozens of clases that need to be fixed by tha standarf | 18:35 |
| sdague | yes, we do | 18:35 |
| dkranz | sdague: "that standard" | 18:36 |
| dkranz | sdague: OK, I was not aware we were trying to change this | 18:36 |
| dkranz | sdague: i would still say that if a delete fails we are hosed | 18:37 |
| sdague | afazekas__ had made a pretty big deal about attacking our leak problem, so I'd like to do our best at making our code more bullet proof there. | 18:37 |
| dkranz | sdague: NOt sure what th epoint is for continuing | 18:37 |
| sdague | dkranz: because one delete failure could be something transient in a server | 18:37 |
| sdague | like a deadlock or something, and it gets better later | 18:37 |
| openstackgerrit | Matthew Treinish proposed a change to openstack/tempest: Add locking to test_aggregates https://review.openstack.org/38857 | 18:38 |
| openstackgerrit | Matthew Treinish proposed a change to openstack/tempest: Switch to using Oslo logging https://review.openstack.org/38879 | 18:38 |
| dkranz | sdague: I think this is worth a mailing list message anyway. I bet lots of people are unaware of this strategy/issue. | 18:39 |
| sdague | ok, I didn't think that was news. I've been pretty consistent about that in reviews for the last couple months. | 18:39 |
| dkranz | sdague: If that is the way you think we should go, we should go to more generic resource management where you don't have to worry about the code cleaning up every single thing it does. | 18:41 |
| sdague | sure, we should get there | 18:41 |
| dkranz | sdague: But that is a fair bit of work too. | 18:41 |
| sdague | but right now our most pressing issue is testr | 18:42 |
| dkranz | sdague: But just once rather than on every piece of new code. | 18:42 |
| sdague | and raising the bar on landing code to ensure people are cleaning up manually, moves us in the right direction | 18:42 |
| dkranz | sdague: How is testr related to this. | 18:42 |
| sdague | dkranz: time | 18:42 |
| sdague | there is only so much people time from now until havana | 18:42 |
| sdague | so fragmenting off for other big efforts isn't useful right now | 18:43 |
| dkranz | sdague: RIght. That is why I was suggesting not worrying about this now at all. | 18:43 |
| *** meteorfox has quit IRC | 18:43 | |
| sdague | dkranz: no, because that just builds more debt later | 18:43 |
| sdague | we can do it better now | 18:43 |
| sdague | and we can refactor something better than that later | 18:43 |
| dkranz | sdague: OK, we should make it clear. | 18:44 |
| sdague | but deciding that we shouldn't apply a standard to new code because there is existing cruft leads to madness :) | 18:44 |
| sdague | well, that's why I -1ed it | 18:44 |
| *** burt has joined #openstack-qa | 18:51 | |
| *** mlavalle has joined #openstack-qa | 18:51 | |
| *** meteorfox has joined #openstack-qa | 18:58 | |
| openstackgerrit | Matthew Treinish proposed a change to openstack/tempest: Switch to using Oslo logging https://review.openstack.org/38879 | 19:00 |
| openstackgerrit | Matthew Treinish proposed a change to openstack/tempest: Add locking to test_aggregates https://review.openstack.org/38857 | 19:00 |
| openstackgerrit | minsel proposed a change to openstack/tempest: Add testing of Neutron per tenant quotas API https://review.openstack.org/35724 | 19:20 |
| mlavalle | dkranz: ping | 19:22 |
| *** meteorfox has quit IRC | 19:27 | |
| *** anteaya has joined #openstack-qa | 19:30 | |
| dkranz | mlavalle: Here | 19:30 |
| mlavalle | dkranz: I just pushed another patch set for Neutron quotas testing, addressing your comments of 1 hour ago | 19:31 |
| mlavalle | dkranz: when you have a chance, please take a look and see if you are happy with it | 19:32 |
| dkranz | mlavalle: Yes, I gave it a +2 already | 19:32 |
| mlavalle | ;-) | 19:32 |
| *** jaypipes has quit IRC | 19:33 | |
| *** klrmn has quit IRC | 19:33 | |
| *** jaypipes has joined #openstack-qa | 19:33 | |
| *** ChanServ sets mode: +o jaypipes | 19:33 | |
| *** klrmn has joined #openstack-qa | 19:33 | |
| mlavalle | dkranz: Thanks | 19:34 |
| *** meteorfox has joined #openstack-qa | 19:39 | |
| openstackgerrit | Sean Dague proposed a change to openstack-dev/devstack: Update projects to global requirements before setup https://review.openstack.org/35705 | 19:48 |
| openstackgerrit | Monty Taylor proposed a change to openstack-dev/devstack: Update projects to global requirements before setup https://review.openstack.org/35705 | 19:52 |
| dkranz | mtreinish: I was looking at your image isolated credentials change. Isn't it a bug that the base class BaseTestCase is not initializing isolated_creds? | 19:55 |
| mtreinish | dkranz: I don't think so, because tenant isolation is a config flag so it's optional | 19:56 |
| dkranz | mtreinish: Yes, but the base class uses this variable and things seem like they would break if a subclass did not. | 19:57 |
| dkranz | mtreinish: Did not set it. | 19:57 |
| dkranz | mtreinish: I am not saying this bug would show itself now but we are setting a trap for subclasses, no? | 19:57 |
| mtreinish | oh, you meant the isolated creds list. Yeah that's required in the subclass. I guess that it should be put in the base class as cls.isolated_creds = [] just so we don't trip over it | 19:59 |
| dkranz | mtreinish: That's what I meant. Same for _interface | 19:59 |
| mtreinish | dkranz: yeah although interface is a little trickier because of xml test classes. We don't want to accidently forget it and start running all the tests with json | 20:01 |
| mtreinish | sdague: http://logs.openstack.org/57/38857/7/check/gate-tempest-devstack-vm-full/37073/console.html , @ 2013-07-29 19:43:00.315 the devstack patch merged right? | 20:03 |
| dkranz | mtreinish: Understood. Are you going to submit a revised patch? | 20:04 |
| sdague | mtreinish: which devstack patch | 20:05 |
| mtreinish | dkranz: I think that should be a separate patch that goes in front of the glance test change | 20:05 |
| mtreinish | sdague: the tmpdir lock_path one | 20:05 |
| dkranz | mtreinish: | 20:05 |
| dkranz | mtreinish: Sure | 20:05 |
| mtreinish | yeah it did it broke it @2013-07-29 19:08:30.743 the config file is lock_path= | 20:05 |
| mtreinish | looks like $TMPDIR is blank | 20:06 |
| mtreinish | I'm going to respin it to push it to the data dir like what nova does | 20:06 |
| sdague | mtreinish: yes, that merged | 20:07 |
| *** meteorfox has quit IRC | 20:08 | |
| *** clarkb has joined #openstack-qa | 20:24 | |
| openstackgerrit | Matthew Treinish proposed a change to openstack-dev/devstack: Set tempest lock_path to default in $DATA_DIR https://review.openstack.org/39148 | 20:28 |
| *** torandu has joined #openstack-qa | 20:30 | |
| openstackgerrit | Kyle Mestery proposed a change to openstack-dev/devstack: Update the ML2 configuration in devstack https://review.openstack.org/37963 | 20:35 |
| *** SergeyLukjanov has quit IRC | 20:54 | |
| *** LuzCazares has joined #openstack-qa | 21:06 | |
| *** meteorfox has joined #openstack-qa | 21:13 | |
| *** woodspa has quit IRC | 21:18 | |
| *** jasondotstar has quit IRC | 21:21 | |
| *** LuzCazares has left #openstack-qa | 21:26 | |
| *** lcazares has joined #openstack-qa | 21:27 | |
| *** dkranz has quit IRC | 21:44 | |
| openstackgerrit | A change was merged to openstack-dev/devstack: Added libffi-dev to the needed packages https://review.openstack.org/39096 | 21:54 |
| openstackgerrit | A change was merged to openstack-dev/devstack: Configure horizon and horizon URLs in Tempest https://review.openstack.org/39067 | 21:54 |
| sdague | dtroyer: https://review.openstack.org/#/c/39148/ how's that look to you? | 21:58 |
| sdague | we've got tempest patches sitting behind it | 21:58 |
| dtroyer | looks good, approved | 21:59 |
| *** burt has quit IRC | 22:08 | |
| *** woodspa has joined #openstack-qa | 22:15 | |
| *** mkollaro has joined #openstack-qa | 22:27 | |
| *** krtaylor has quit IRC | 22:39 | |
| *** mkollaro has quit IRC | 22:46 | |
| openstackgerrit | A change was merged to openstack-dev/devstack: Set tempest lock_path to default in $DATA_DIR https://review.openstack.org/39148 | 22:49 |
| *** marun has quit IRC | 22:55 | |
| *** woodspa_ has joined #openstack-qa | 22:58 | |
| *** woodspa has quit IRC | 23:02 | |
| *** marun has joined #openstack-qa | 23:03 | |
| *** mlavalle has left #openstack-qa | 23:08 | |
| openstackgerrit | John Griffith proposed a change to openstack/tempest: Add wait_for_volume_status in upload_to_image test. https://review.openstack.org/39178 | 23:24 |
| *** mberwanger has joined #openstack-qa | 23:25 | |
| *** bclifford has left #openstack-qa | 23:34 | |
| openstackgerrit | Edgar Magana proposed a change to openstack-dev/devstack: Set horizon configuration for security groups https://review.openstack.org/39181 | 23:41 |
| *** lcazares has quit IRC | 23:48 | |
| openstackgerrit | Edgar Magana proposed a change to openstack-dev/devstack: Update to PLUMgrid plugin configuration https://review.openstack.org/39182 | 23:48 |
Generated by irclog2html.py 2.14.0 by Marius Gedminas - find it at mg.pov.lt!