| openstackgerrit | xu-haiwei proposed openstack/senlin: Initialize docker driver https://review.openstack.org/322692 | 00:40 |
|---|---|---|
| openstackgerrit | xu-haiwei proposed openstack/senlin: Initialize docker driver https://review.openstack.org/322692 | 01:07 |
| openstackgerrit | Qiming Teng proposed openstack/senlin: Improve DB sorting https://review.openstack.org/326194 | 01:13 |
| *** yanyanhu has joined #senlin | 01:17 | |
| *** Qiming_ has joined #senlin | 01:34 | |
| *** Qiming_ has quit IRC | 01:46 | |
| *** chenfei has joined #senlin | 01:49 | |
| openstackgerrit | Merged openstack/senlin: Initial framework for health manager listener https://review.openstack.org/325045 | 01:53 |
| openstackgerrit | Yanyan Hu proposed openstack/senlin: Add negative tempest API test for node update https://review.openstack.org/326202 | 01:54 |
| yanyanhu | hi, Qiming, around? | 01:59 |
| Qiming | yes | 01:59 |
| yanyanhu | currently, node_update API will raise 400 err when it is invoked without specifying any property to update | 02:00 |
| *** elynn has joined #senlin | 02:00 | |
| yanyanhu | but for cluster, the update request will pass silently | 02:00 |
| yanyanhu | http://git.openstack.org/cgit/openstack/senlin/tree/senlin/engine/service.py#n1525 | 02:00 |
| yanyanhu | I'm thinking maybe we should make them consistent? | 02:01 |
| openstackgerrit | xu-haiwei proposed openstack/senlin: Initialize docker driver https://review.openstack.org/322692 | 02:01 |
| yanyanhu | and profile_update now behaves like cluster_update | 02:02 |
| Qiming | ... saw "if inputs == {}" again ... | 02:02 |
| Qiming | consistency is aways good | 02:02 |
| yanyanhu | yes | 02:02 |
| elynn | Morning! | 02:02 |
| yanyanhu | morning, elynn :) | 02:03 |
| yanyanhu | so maybe we follow node_update's behavior? | 02:03 |
| yanyanhu | BTW, this kind of change will vary the API behavior, do we need to change API micro version for it? | 02:04 |
| Qiming | I'm inclined to a stricter checking | 02:05 |
| *** elynn_ has joined #senlin | 02:05 | |
| Qiming | yes, we should bump the microversion if possible | 02:06 |
| yanyanhu | I see, will fix it | 02:06 |
| Qiming | but .. | 02:06 |
| Qiming | it is not an essential api behavior change, imo | 02:07 |
| *** elynn has quit IRC | 02:08 | |
| Qiming | it is more about a change to the error response | 02:08 |
| yanyanhu | yes, that's why I'm not sure whether we need to bump api micro version | 02:08 |
| yanyanhu | yes | 02:08 |
| yanyanhu | and we don't need to change the api doc as well | 02:08 |
| Qiming | right | 02:10 |
| Qiming | if it is distinguishable from api docs | 02:10 |
| Qiming | we should definitely bump the version | 02:10 |
| yanyanhu | I see | 02:10 |
| openstackgerrit | Qiming Teng proposed openstack/senlin: Add basic schema support for profile operations https://review.openstack.org/326208 | 02:14 |
| Qiming | need some helps reviewing this patch ^^ | 02:15 |
| Qiming | it is about adding operations to profiles | 02:15 |
| openstackgerrit | Merged openstack/senlin: Add negative tempest API test for cluster update https://review.openstack.org/325777 | 02:25 |
| openstackgerrit | xu-haiwei proposed openstack/senlin: Initialize docker driver https://review.openstack.org/322692 | 02:29 |
| *** yuanying_ has joined #senlin | 02:51 | |
| *** yuanying has quit IRC | 02:51 | |
| *** Drago has quit IRC | 02:53 | |
| openstackgerrit | Qiming Teng proposed openstack/senlin: Replace 'format_time' with 'isotime' https://review.openstack.org/326213 | 02:56 |
| openstackgerrit | Yanyan Hu proposed openstack/senlin: Minor revise cluster_update and profile_update https://review.openstack.org/326216 | 03:34 |
| *** Drago has joined #senlin | 03:48 | |
| *** Drago has quit IRC | 03:48 | |
| *** Drago has joined #senlin | 03:48 | |
| openstackgerrit | xu-haiwei proposed openstack/senlin: Initialize docker driver https://review.openstack.org/322692 | 04:10 |
| *** Drago has quit IRC | 04:13 | |
| *** yuanying_ has quit IRC | 04:18 | |
| *** yuanying has joined #senlin | 04:21 | |
| *** elynn_ has quit IRC | 04:28 | |
| *** zzxwill has joined #senlin | 04:40 | |
| *** elynn_ has joined #senlin | 04:50 | |
| *** elynn__ has joined #senlin | 04:54 | |
| *** elynn_ has quit IRC | 04:56 | |
| *** zzxwill has quit IRC | 05:09 | |
| *** zzxwill has joined #senlin | 05:26 | |
| *** elynn__ has quit IRC | 05:27 | |
| Qiming | zzxwill, about Bug 1589456 | 05:27 |
| openstack | bug 1589456 in senlin "cluster creation with option "-c" won't output the execution result " [Undecided,New] https://launchpad.net/bugs/1589456 | 05:27 |
| Qiming | I cannot reproduce it | 05:27 |
| Qiming | it could be that 'openstack cluster | 05:28 |
| Qiming | create --profile server1 -c 1 mycluster_to_be_deleted | 05:28 |
| Qiming | ' command still fails | 05:28 |
| Qiming | openstackclient is using --profile for os-profiler, so there is a conflict we identified | 05:29 |
| Qiming | and that ... is the reason we didn't deprecate senlinclient command line as planned | 05:29 |
| zzxwill | Hello Qiming. I will update the code the latest and try again. The code in my environment is several-days late. | 05:30 |
| Qiming | I mean ... using openstackclient to create a cluster is problematic | 05:31 |
| Qiming | the '--profile' parameter has been used by OSC for os-profiler | 05:31 |
| Qiming | so you will have to use 'senlin cluster-create' for that purpose | 05:31 |
| Qiming | check this: http://git.openstack.org/cgit/openstack/python-openstackclient/commit/?id=530fe42589a2138278f100f791d8c6d3fbed8950 | 05:33 |
| Qiming | so --profile won't be removed until Apr 2017, so we have to use 'senlin cluster-create' for quite some time | 05:34 |
| *** elynn has joined #senlin | 05:48 | |
| *** elynn_ has joined #senlin | 05:53 | |
| *** elynn has quit IRC | 05:56 | |
| openstackgerrit | Yanyan Hu proposed openstack/senlin: Minor revise profile_update API https://review.openstack.org/326259 | 06:58 |
| openstackgerrit | Yanyan Hu proposed openstack/senlin: Minor revise cluster_update and profile_update https://review.openstack.org/326216 | 07:16 |
| openstackgerrit | Yanyan Hu proposed openstack/senlin: Add negative tempest API test for profile/policy update https://review.openstack.org/326266 | 07:16 |
| openstackgerrit | Merged openstack/senlin: Add negative tempest API test for node update https://review.openstack.org/326202 | 07:20 |
| openstackgerrit | Merged openstack/senlin: Improve DB sorting https://review.openstack.org/326194 | 07:33 |
| yanyanhu | hi, Qiming, around? | 07:33 |
| yanyanhu | I'm thinking whether we can provide operator a config option to clean all successful actions automatically of cluster after the cluster is deleted | 07:35 |
| yanyanhu | the size of action table increase very fast when large size cluster is created/deleted frequently | 07:35 |
| yanyanhu | and the action_list API fails frequently for db access timeout in that case | 07:36 |
| Qiming | yanyanhu, that is a good suggestion | 07:40 |
| Qiming | the only problem is for cluster_delete calls | 07:40 |
| yanyanhu | you mean the action for cluster_delete itself? | 07:41 |
| Qiming | callers should not check the status of the CLUSTER_DELETE action | 07:41 |
| yanyanhu | ah | 07:41 |
| yanyanhu | that's a problem | 07:41 |
| Qiming | or they should not rely solely on checking the action status | 07:41 |
| yanyanhu | yes, not found of resource become the only symbol can be used to check cluster_delete complete | 07:42 |
| yanyanhu | let me think about it | 07:43 |
| yanyanhu | or cluster_delete action is the only action that will be left in DB | 07:44 |
| elynn_ | Why not just delete all the other actions except cluster_delete action left for now? | 07:44 |
| yanyanhu | :) | 07:44 |
| yanyanhu | have the same feeling | 07:44 |
| elynn_ | So it won't break the compatible. Then we can consider how we deal with this only action. | 07:45 |
| yanyanhu | but the same problem exists for node | 07:45 |
| elynn_ | yes.. | 07:46 |
| yanyanhu | so theoretically, removing all other actions except delete actions can reduce the size of action log by nearly 50% | 07:47 |
| yanyanhu | so an option is we only leave cluster_delete action after cluster is deleted. All other cluster and node actions that have been done successfully will be removed | 07:49 |
| Qiming | more than 50% I think | 07:49 |
| yanyanhu | Qiming, yes | 07:49 |
| Qiming | we may want to delete events associated with a cluster/node as well | 07:51 |
| yanyanhu | yes | 07:51 |
| yanyanhu | there have been a even purge db api, but not very sure where we should invoke it | 07:51 |
| Qiming | there could be some reasons that we need to keep events/actions there, for auditing | 07:52 |
| yanyanhu | maybe cluster_delete is a possible position | 07:52 |
| yanyanhu | yes | 07:52 |
| Qiming | once we have a new backend for event generation, we can safely delete every action/events from db | 07:52 |
| yanyanhu | right | 07:53 |
| Qiming | auditing can be done via a different backend | 07:53 |
| yanyanhu | in that case, user has better way to track and log the workflow | 07:53 |
| Qiming | then, what is the RIGHT thing to do? | 07:54 |
| Qiming | leaving just xxx_DELETE actions there sounds a workaround | 07:54 |
| yanyanhu | maybe start from adding config option for control action/event history purging | 07:55 |
| yanyanhu | yes, that looks feasible | 07:55 |
| Qiming | an option is something we need to maintain for a long time | 07:55 |
| Qiming | just to remind | 07:55 |
| yanyanhu | understand | 07:56 |
| Qiming | just checked heat code | 07:58 |
| Qiming | senlin resources are not relying on the actions for determining whether a cluster/node is gone | 07:58 |
| elynn_ | yes | 07:59 |
| Qiming | so, maybe we should just delete all actions related to a cluster or a node | 08:00 |
| Qiming | oh, wait ... | 08:01 |
| Qiming | a NODE_DELETE action cannot be deleted if it is waited by a CLUSTER_DELETE action | 08:01 |
| Qiming | the safest bet would be delete all actions related to a cluster and all nodes that were in that cluster | 08:05 |
| openstackgerrit | Merged openstack/senlin-dashboard: Use entities.reverse() rather sorted(.., reverse=True) https://review.openstack.org/325854 | 08:10 |
| Qiming | btw, this seems interesting: http://logs.openstack.org/13/326213/1/check/gate-senlin-dsvm-tempest-api/4899d3c/logs/stackviz/#/stdin/timeline?test=senlin.tests.tempest.api.nodes.test_node_create_negative.TestNodeCreateNegativeBadRequest.test_node_create_bad_request_profile_invalid | 08:19 |
| Qiming | I'm trying to revise the event db model/api to enable event pruning | 09:04 |
| openstackgerrit | Ethan Lynn proposed openstack/senlin: Revise cluster resize test https://review.openstack.org/326308 | 09:08 |
| openstackgerrit | Ethan Lynn proposed openstack/senlin: Add API test for cluster scale_out action https://review.openstack.org/326309 | 09:08 |
| openstackgerrit | Qiming Teng proposed openstack/senlin: Remove event config option and fix prune api https://review.openstack.org/326322 | 09:22 |
| openstackgerrit | Yanyan Hu proposed openstack/senlin: Add a db_api interface for deleting actions https://review.openstack.org/326328 | 09:36 |
| openstackgerrit | Ethan Lynn proposed openstack/senlin: Add API test for cluster scale_in action https://review.openstack.org/326330 | 09:39 |
| openstackgerrit | Ethan Lynn proposed openstack/senlin: Add API test for cluster add_nodes action https://review.openstack.org/326331 | 09:39 |
| openstackgerrit | Ethan Lynn proposed openstack/senlin: Add API test for cluster del_nodes action https://review.openstack.org/326332 | 09:39 |
| openstackgerrit | Ethan Lynn proposed openstack/senlin: Add API test for cluster check action https://review.openstack.org/326333 | 09:39 |
| openstackgerrit | Ethan Lynn proposed openstack/senlin: Add API test for cluster recover action https://review.openstack.org/326334 | 09:39 |
| openstackgerrit | Ethan Lynn proposed openstack/senlin: Add API test for cluster policy_attach action https://review.openstack.org/326342 | 09:51 |
| openstackgerrit | Ethan Lynn proposed openstack/senlin: Add API test for cluster policy_detach action https://review.openstack.org/326343 | 09:51 |
| openstackgerrit | Ethan Lynn proposed openstack/senlin: Add API test for cluster policy_update action https://review.openstack.org/326344 | 09:51 |
| *** shu-mutou is now known as shu-mutou-AFK | 09:54 | |
| openstackgerrit | Ethan Lynn proposed openstack/senlin: Add API test for node recover action https://review.openstack.org/326346 | 09:59 |
| *** yanyanhu has quit IRC | 10:10 | |
| *** yanyanhu has joined #senlin | 10:10 | |
| *** elynn_ has quit IRC | 10:12 | |
| openstackgerrit | Merged openstack/senlin: Minor revise profile_update API https://review.openstack.org/326259 | 10:51 |
| openstackgerrit | Merged openstack/senlin: Minor revise cluster_update and profile_update https://review.openstack.org/326216 | 10:52 |
| openstackgerrit | Merged openstack/senlin: Add negative tempest API test for profile/policy update https://review.openstack.org/326266 | 10:53 |
| *** zzxwill has quit IRC | 12:28 | |
| *** zzxwill has joined #senlin | 12:32 | |
| *** yanyanhu has joined #senlin | 12:51 | |
| *** lixinhui_ has joined #senlin | 12:52 | |
| *** elynn has joined #senlin | 13:01 | |
| *** zzxwill has quit IRC | 13:16 | |
| *** zzxwill has joined #senlin | 13:16 | |
| *** zzxwill has quit IRC | 13:46 | |
| *** elynn has quit IRC | 13:50 | |
| openstackgerrit | Merged openstack/senlin: Replace 'format_time' with 'isotime' https://review.openstack.org/326213 | 13:54 |
| Qiming | oh, don't know if you have noticed it | 13:57 |
| Qiming | we have senlin 2.0.0.0b1 released last Friday | 13:57 |
| Qiming | senlinclient 0.5.0 released today | 13:57 |
| yanyanhu | yes, saw the msg in mail list :) | 13:58 |
| yanyanhu | nice | 13:59 |
| Qiming | oh, I was typing here, when I was supposed to type in meeting channel | 14:00 |
| yanyanhu | ... | 14:00 |
| yanyanhu | ttyl | 14:01 |
| yanyanhu | have a good night | 14:01 |
| *** yanyanhu has quit IRC | 14:01 | |
| *** lixinhui_ has quit IRC | 14:01 | |
| *** jdandrea_ has joined #senlin | 14:17 | |
| *** jdandrea has quit IRC | 14:17 | |
| *** zzxwill has joined #senlin | 14:18 | |
| *** Drago has joined #senlin | 15:10 | |
| *** Drago has quit IRC | 15:12 | |
| *** Drago has joined #senlin | 15:12 | |
| *** jdandrea_ is now known as jdandrea | 15:39 | |
| *** zzxwill has quit IRC | 16:17 | |
| *** Drago has quit IRC | 16:49 | |
| *** Drago has joined #senlin | 16:50 | |
| *** Drago has quit IRC | 16:50 | |
| *** Drago has joined #senlin | 16:51 | |
| *** Drago has quit IRC | 16:51 | |
| *** Drago has joined #senlin | 16:51 | |
| *** Drago has quit IRC | 17:08 | |
| *** Drago has joined #senlin | 17:56 | |
| *** Drago has quit IRC | 22:45 | |
Generated by irclog2html.py 2.14.0 by Marius Gedminas - find it at mg.pov.lt!