| *** zhurong has quit IRC | 00:04 | |
| *** zzxwill has joined #senlin | 00:51 | |
| *** zzxwill has quit IRC | 00:52 | |
| *** zzxwill has joined #senlin | 00:56 | |
| *** zzxwill has quit IRC | 00:57 | |
| *** zzxwill has joined #senlin | 00:57 | |
| *** elynn has joined #senlin | 00:58 | |
| *** zzxwill has quit IRC | 01:08 | |
| *** zhurong has joined #senlin | 01:09 | |
| *** zzxwill has joined #senlin | 01:10 | |
| *** zzxwill has quit IRC | 01:23 | |
| *** zzxwill has joined #senlin | 01:23 | |
| *** elynn_ has joined #senlin | 01:34 | |
| *** elynn has quit IRC | 01:38 | |
| *** elynn_ has quit IRC | 01:42 | |
| *** elynn has joined #senlin | 01:42 | |
| *** yanyanhu has joined #senlin | 01:44 | |
| Qiming | hi, elynn, if we are gonna land profile/policy validate APIs in Newton release, we should make sure the SDK support is done by Aug 22nd, I mean the next version bump of SDK | 02:16 |
|---|---|---|
| Qiming | that week is the deadline for non-client libraries | 02:16 |
| openstackgerrit | Yanyan Hu proposed openstack/senlin: Add zaqar v2 driver interfaces for claim resource https://review.openstack.org/355830 | 02:16 |
| elynn | Oh, I need to hurry. | 02:17 |
| Qiming | I believe SDK should release a version by then and maybe no further releases before newton release | 02:17 |
| Qiming | the senlin side patches do not have such an urgency because we can control its release version | 02:18 |
| elynn | okay, I need to figure out how much I rely on sdk this week. | 02:19 |
| Qiming | just put the sdk side patches there and make sure 0.9.3 contains what we want, that would be okay | 02:19 |
| Qiming | yep, thx | 02:19 |
| *** zhurong_ has joined #senlin | 02:19 | |
| *** zhurong has quit IRC | 02:23 | |
| yanyanhu | hi, Qiming, just made a test and found senlinclient works correctly with 0.9.3-dev sdk | 02:33 |
| yanyanhu | I think I may didn't setup sdk installation correctly yesterday | 02:34 |
| Qiming | okay, if you are using the latest version, it has the latest code about api_version and updated requirements about keystoneauth1, so everything should be fine | 02:37 |
| yanyanhu | yes, it is | 02:38 |
| *** zhurong_ has quit IRC | 02:39 | |
| *** zhurong has joined #senlin | 02:39 | |
| *** zzxwill has quit IRC | 02:47 | |
| *** yuanying has quit IRC | 02:50 | |
| *** zzxwill has joined #senlin | 02:50 | |
| openstackgerrit | Qiming Teng proposed openstack/senlin: Fix DB setup for unit tests https://review.openstack.org/356193 | 02:57 |
| openstackgerrit | Merged openstack/senlin: Fix service filtering when claiming clusters https://review.openstack.org/355743 | 03:11 |
| *** eldon has joined #senlin | 03:12 | |
| eldon | http://paste.openstack.org/show/558687/ | 03:15 |
| eldon | do you have any idea about this problem? | 03:15 |
| yanyanhu | hi, eldon, let me check it | 03:17 |
| eldon | We use rally to create cluster with 40 concurrencies. | 03:19 |
| yanyanhu | hi, eldon, are you using the latest code of senlin? | 03:19 |
| eldon | not yet. | 03:20 |
| yanyanhu | I found the code in the log is a little different from the one in master | 03:20 |
| yanyanhu | the engine/scheduler.py module | 03:20 |
| eldon | Our deadline is 20 oct. So we don't have enough time to test newest code. | 03:20 |
| yanyanhu | I see | 03:21 |
| yanyanhu | just want to confirm the scheduler is the old one or new one :) | 03:21 |
| eldon | ok~ | 03:21 |
| yanyanhu | since we have a refactoring of senlin scheduler after austin summit | 03:21 |
| eldon | I think it is something about IMPL sessions. | 03:22 |
| yanyanhu | http://git.openstack.org/cgit/openstack/senlin/tree/senlin/engine/scheduler.py | 03:22 |
| yanyanhu | eldon, looks so | 03:22 |
| yanyanhu | this is the new one which enabled tickless scheduling and also batch control | 03:23 |
| yanyanhu | eldon, looks like the error was caused by db concurrency issue? | 03:23 |
| eldon | yes. | 03:24 |
| eldon | I have newest code in our test zone. I will try again in this afternoon | 03:24 |
| yanyanhu | eldon, great | 03:25 |
| yanyanhu | if the same issue happens, plz just ping me | 03:25 |
| yanyanhu | BTW, are you guys using the latest rally code :) | 03:25 |
| yanyanhu | since senlin's cluster creating deleting test support has been merged into rally | 03:26 |
| eldon | But it only have three cpts, we can only test with no node in cluster. | 03:26 |
| yanyanhu | https://review.openstack.org/301522 | 03:27 |
| eldon | rally is also a little old, but I think it doesn't matter:) | 03:27 |
| yanyanhu | cpts compute nodes? | 03:27 |
| yanyanhu | I see | 03:27 |
| yanyanhu | http://git.openstack.org/cgit/openstack/rally/tree/samples/tasks/scenarios/senlin/create-and-delete-profile-cluster.yaml | 03:28 |
| yanyanhu | this is the example file of senlin cluster test job in rally | 03:28 |
| yanyanhu | also I'm now working on context support for profile. Hopefully it can be done in two weeks | 03:29 |
| yanyanhu | if you guys need any help on rally senlin plugin or test jobs, plz feel free to ping me as well :) eldon | 03:29 |
| eldon | ok:) | 03:29 |
| eldon | great | 03:29 |
| *** zhurong has quit IRC | 03:35 | |
| *** zhurong has joined #senlin | 03:36 | |
| *** elynn_ has joined #senlin | 03:37 | |
| *** zzxwill has quit IRC | 03:37 | |
| *** zzxwill has joined #senlin | 03:39 | |
| *** elynn has quit IRC | 03:41 | |
| *** zzxwill has quit IRC | 03:42 | |
| *** elynn_ has quit IRC | 03:47 | |
| *** yuanying has joined #senlin | 03:49 | |
| Qiming | yanyanhu, I'm pushing a patch to the nova delete operation, hope this can fix the gate problem you saw | 03:50 |
| yanyanhu | Qiming, great, will check it | 03:50 |
| openstackgerrit | Qiming Teng proposed openstack/senlin: Make server delete operation back compatible https://review.openstack.org/356212 | 03:51 |
| Qiming | yanyanhu, ^ this one | 03:51 |
| yanyanhu | nice workaround :) | 03:52 |
| *** yuanying has quit IRC | 03:55 | |
| Qiming | the goal is to preserve xinhui's hard work while make the gate happy | 03:55 |
| openstackgerrit | Merged openstack/senlin: Fix health manager initialization https://review.openstack.org/355751 | 03:56 |
| openstackgerrit | Merged openstack/senlin: Allow health check to be enabled/disabled https://review.openstack.org/355790 | 03:58 |
| openstackgerrit | Qiming Teng proposed openstack/senlin: Remove garbage parameters at db layer https://review.openstack.org/356215 | 04:04 |
| *** zhurong has quit IRC | 04:14 | |
| *** zhurong has joined #senlin | 04:16 | |
| *** yuanying has joined #senlin | 04:16 | |
| *** eldon has quit IRC | 04:32 | |
| *** elynn_ has joined #senlin | 04:51 | |
| *** elynn__ has joined #senlin | 05:08 | |
| *** elynn_ has quit IRC | 05:12 | |
| yanyanhu | hi, Qiming, about this patch, I think we may need to add corresponding fake driver interface for this new added "server_force_delete" method | 05:20 |
| yanyanhu | https://review.openstack.org/#/c/356212/ | 05:20 |
| yanyanhu | will try to add it based on your patch | 05:20 |
| *** zzxwill has joined #senlin | 05:24 | |
| yanyanhu | hmm, looks like test env setting up at gate side failed | 05:26 |
| yanyanhu | will wait for recheck result | 05:26 |
| yanyanhu | fake interfaces may not be necessary | 05:27 |
| *** yuanying_ has joined #senlin | 05:29 | |
| *** yuanying has quit IRC | 05:29 | |
| *** eldon has joined #senlin | 05:53 | |
| Qiming | are we touching server_force_delete call in current test cases? | 06:08 |
| Qiming | I'm not against adding that into the fake driver | 06:08 |
| Qiming | just don't believe current failure has anything to do with force delete | 06:09 |
| *** zzxwill has quit IRC | 06:15 | |
| *** zzxwill has joined #senlin | 06:17 | |
| elynn__ | Hi | 06:27 |
| elynn__ | I got some errors like openstack.exceptions.SDKException: Cannot connect to API service. when run cluster-show | 06:27 |
| elynn__ | DEBUG (session) REQ: curl -g -i -X GET http://9.119.148.46:8778./clusters/c1 -H "openstack-api-version: clustering 1.2" -H "User-Agent: python-senlinclient openstacksdk/0.9.2 keystoneauth1/2.10.0 python-requests/2 | 06:28 |
| elynn__ | .10.0 CPython/2.7.5" -H "X-Auth-Token: {SHA1}a5774b06badb28147fc2e234ccfbcc996871eade" | 06:28 |
| elynn__ | Request like ^ | 06:28 |
| elynn__ | All API requests failed. | 06:29 |
| elynn__ | I'm using sdk of 0.9.2 version, have you met before? | 06:29 |
| openstackgerrit | Xinhui Li proposed openstack/senlin: Support Fencing by Health Policy and Cluster Action https://review.openstack.org/354045 | 06:34 |
| openstackgerrit | Ethan Lynn proposed openstack/senlin: Fix do_validate in heat profile https://review.openstack.org/355780 | 06:36 |
| Qiming | elynn__, you are not using the master branch | 06:36 |
| Qiming | the URI is incorrect: http://9.119.148.46:8778./clusters/c1 | 06:37 |
| Qiming | it has been fixed yesterday | 06:37 |
| Qiming | the bug was introduced when migrating micro-version support to the v1 endpoint | 06:37 |
| Qiming | this one: https://review.openstack.org/#/c/355703/ | 06:38 |
| elynn__ | After pulling the latest codes, it works now, I need to rebase my codes. | 06:38 |
| Qiming | okay | 06:38 |
| elynn__ | And another question, all the requests now should contain a header `openstack-api-version: clustering 1.2` now? | 06:39 |
| openstackgerrit | Merged openstack/senlin: Make server delete operation back compatible https://review.openstack.org/356212 | 06:49 |
| Qiming | yes, that means our client supports the latest microversion of senlin api | 06:59 |
| Qiming | if we don't provide such a header when requesting an api such as profile-validate, policy-validate, cluster-collect, we will get an error message | 06:59 |
| Qiming | the plan is to have 1.2 api support these three operations released in newton (2.0.0) version | 07:00 |
| elynn__ | I see, I though if without this header, server will automatically choose the latest version. Looks I'm misunderstanding it. | 07:03 |
| elynn__ | Hi Qiming , I got those errors after updating my codes http://paste.openstack.org/show/558964/ | 07:47 |
| elynn__ | It seems only happened on profile validation, when sdk try to post request to stacks/preview | 07:48 |
| Qiming | according to api-wg spec, the default behavior should go to the lowest version | 07:48 |
| Qiming | http://git.openstack.org/cgit/openstack/api-wg/tree/guidelines/microversion_specification.rst#n67 | 07:49 |
| elynn__ | I should really read this doc thoroughly... | 07:50 |
| Qiming | elynn | 07:50 |
| Qiming | if you want to dump more info for debugging | 07:50 |
| Qiming | you can add these to senlin/drivers/openstack/sdk.py | 07:50 |
| Qiming | import sys | 07:52 |
| Qiming | from openstack import utils | 07:52 |
| Qiming | utils.enable_logging(debug=True, stream=sys.stdout) | 07:52 |
| Qiming | then you can see more details during the interaction between sdk and the backend services | 07:53 |
| elynn__ | Thanks! Let me try | 07:53 |
| elynn__ | Should I put them on the top of other codes? | 07:54 |
| elynn__ | Have you try to create a heat stack cluster with sdk 0.9.2? | 08:09 |
| elynn__ | I got the errors I mentioned above http://paste.openstack.org/show/558964/ Qiming yanyanhu | 08:09 |
| Qiming | yes, elynn__ | 08:16 |
| Qiming | put them on top of the module | 08:16 |
| Qiming | the key is about the not found exception | 08:18 |
| openstackgerrit | Qiming Teng proposed openstack/senlin: Fix node join behavior https://review.openstack.org/356301 | 08:23 |
| openstackgerrit | Qiming Teng proposed openstack/senlin: Improve server metadata format https://review.openstack.org/356302 | 08:23 |
| openstackgerrit | Qiming Teng proposed openstack/senlin: Close health manager loop https://review.openstack.org/356303 | 08:23 |
| Qiming | elynn__, I'm using the master branch of sdk, not 0.9.2 | 08:28 |
| elynn__ | Does it works for you? | 08:29 |
| elynn__ | Create node from a heat stack profile | 08:29 |
| openstackgerrit | Qiming Teng proposed openstack/senlin: Fix DB setup for unit tests https://review.openstack.org/356193 | 08:30 |
| openstackgerrit | Qiming Teng proposed openstack/senlin: Remove garbage parameters at db layer https://review.openstack.org/356215 | 08:31 |
| Qiming | trying | 08:32 |
| Qiming | the same failure, investigating | 08:35 |
| Qiming | got it | 08:37 |
| Qiming | heat stupidity | 08:38 |
| elynn__ | Qiming, Just try 0.9.1, works for me | 08:38 |
| Qiming | yes | 08:38 |
| Qiming | sdk is trying to find out the latest version of a service api by doing a version discovery | 08:38 |
| Qiming | when it is checking heat, heat reports this: | 08:38 |
| Qiming | {"versions": [{"status": "CURRENT", "id": "v1.0", "links": [{"href": "http://192.168.42.11:8004/v1/", "rel": "self"}]}]} | 08:38 |
| Qiming | however, that is not a working endpoint url | 08:39 |
| Qiming | heat requires a 'project_id' in this href | 08:39 |
| Qiming | heat should report this: http://192.168.42.11:8004/v1/$(project_id)s | 08:40 |
| elynn__ | oh, seems a but in heat | 08:44 |
| Qiming | sdk can solve it easily | 08:46 |
| Qiming | fixing | 08:46 |
| elynn__ | That would be the better way to solve it in sdk side. | 08:50 |
| Qiming | okay fixed | 08:52 |
| elynn__ | Great, thanks! | 08:53 |
| elynn__ | I'm thinking is it better to involve senlinclient in integration tests. Now we are using tempest client. | 08:54 |
| Qiming | it was my fault | 08:55 |
| elynn__ | okay... | 09:01 |
| *** zzxwill has quit IRC | 09:01 | |
| openstackgerrit | Qiming Teng proposed openstack/senlin: Fix stack status wait https://review.openstack.org/356321 | 09:02 |
| Qiming | elynn__, ^ this is the patch at senlin side | 09:02 |
| Qiming | it was my fault when introducing the wait_for_stack call | 09:02 |
| Qiming | should have passed in stack object instead of stack id | 09:03 |
| Qiming | didn't do integration test when working on that | 09:03 |
| elynn__ | Is there any fix need to merge in sdk side? | 09:04 |
| Qiming | and the sdk side ... | 09:05 |
| Qiming | https://review.openstack.org/356323 | 09:05 |
| Qiming | this one | 09:05 |
| Qiming | gonna self approve it | 09:06 |
| elynn__ | You are always so quick! | 09:06 |
| Qiming | since the logic is simple and evident | 09:06 |
| elynn__ | Richard this didn't sleep? he is so quick to +2 | 09:09 |
| Qiming | no idea, maybe an approval during sleep | 09:31 |
| elynn__ | lol | 09:31 |
| *** elynn__ has quit IRC | 09:31 | |
| *** shu-mutou is now known as shu-mutou-AFK | 09:38 | |
| *** openstack has joined #senlin | 10:19 | |
| openstackgerrit | Qiming Teng proposed openstack/senlin: Improve server metadata format https://review.openstack.org/356302 | 10:24 |
| *** eldon has joined #senlin | 10:46 | |
| *** yanyanhu has quit IRC | 10:54 | |
| *** eldon has quit IRC | 11:05 | |
| openstackgerrit | Merged openstack/senlin: Remove garbage parameters at db layer https://review.openstack.org/356215 | 12:04 |
| *** zzxwill has joined #senlin | 12:56 | |
| *** syjulian has left #senlin | 12:58 | |
| *** zzxwill has quit IRC | 13:04 | |
| openstackgerrit | Merged openstack/senlin: Fix DB setup for unit tests https://review.openstack.org/356193 | 13:05 |
| *** zhurong has joined #senlin | 13:05 | |
| *** zhurong has quit IRC | 13:07 | |
| *** zhurong has joined #senlin | 13:09 | |
| openstackgerrit | Qiming Teng proposed openstack/senlin: Close health manager loop https://review.openstack.org/356303 | 13:13 |
| openstackgerrit | Merged openstack/senlin: Fix stack status wait https://review.openstack.org/356321 | 14:17 |
| openstackgerrit | Qiming Teng proposed openstack/senlin: Close health manager loop https://review.openstack.org/356303 | 14:33 |
| *** zhurong has quit IRC | 14:59 | |
| *** openstackgerrit has quit IRC | 15:03 | |
| *** openstackgerrit has joined #senlin | 15:05 | |
| *** ChanServ sets mode: +v openstackgerrit | 15:05 | |
| *** Drago has joined #senlin | 15:29 | |
| *** Drago has quit IRC | 15:29 | |
| *** Drago has joined #senlin | 15:29 | |
| *** Drago has quit IRC | 16:14 | |
| *** Drago has joined #senlin | 16:14 | |
| *** jdandrea has joined #senlin | 17:58 | |
| jdandrea | Anyone know if a quiescent state been proposed for Senlin? Trying to find a way to let autoscaling victims go quiescent. I imagine this is tricky since it's very app-specific, there be dragons, etc. | 17:59 |
| -openstackstatus- NOTICE: The volume for logs.openstack.org filled up rather suddenly, causing a number of jobs to fail with a POST_FAILURE result and no logs; we're manually expiring some logs now to buy breathing room, but any changes which hit that in the past few minutes will need to be rechecked and/or approved again | 19:46 | |
| *** zhurong has joined #senlin | 23:51 | |
Generated by irclog2html.py 2.14.0 by Marius Gedminas - find it at mg.pov.lt!