| *** Qiming has joined #senlin | 01:00 | |
| *** Yanyanhu has joined #senlin | 01:10 | |
| openstackgerrit | Qiming Teng proposed openstack/senlin: Revise authorization doc https://review.openstack.org/280034 | 01:26 |
|---|---|---|
| openstackgerrit | Merged openstack/senlin: Revise authorization doc https://review.openstack.org/280034 | 01:44 |
| Yanyanhu | hi, Qiming, around? | 01:47 |
| Yanyanhu | about these two patches which remove unnecessary event log from cluster and node module, https://review.openstack.org/#/c/279939/ https://review.openstack.org/#/c/279938/ | 01:48 |
| Yanyanhu | by 'the information are already dumped duing action execution', you mean the log message in cluster_action and node_action module? | 01:49 |
| Qiming | Yanyanhu, yes | 01:50 |
| Yanyanhu | ok | 01:50 |
| Yanyanhu | since I didn't find event log in action modules | 01:50 |
| Qiming | when debugging things, I noticed those event calls are also creating log entries | 01:50 |
| Yanyanhu | yes | 01:51 |
| Qiming | but the data is not useful at all | 01:51 |
| Yanyanhu | understand | 01:51 |
| Qiming | this one is almost capturing all necessary info: http://git.openstack.org/cgit/openstack/senlin/tree/senlin/engine/actions/base.py#n494 | 01:51 |
| Qiming | also, we are already capturing the result of action executions here: http://git.openstack.org/cgit/openstack/senlin/tree/senlin/engine/actions/base.py#n338 | 01:53 |
| Yanyanhu | got it | 01:54 |
| openstackgerrit | Merged openstack/senlin: Remove some EVENT calls from node module https://review.openstack.org/279938 | 02:00 |
| openstackgerrit | Merged openstack/senlin: Remove some EVENT call from cluster module https://review.openstack.org/279939 | 02:01 |
| *** junxu has joined #senlin | 02:01 | |
| *** elynn has joined #senlin | 02:19 | |
| elynn | Morning! | 02:21 |
| Qiming | morning | 02:22 |
| elynn | Seems just you and me are working today? | 02:28 |
| elynn | Hi Qiming , do you have time to review this one https://review.openstack.org/#/c/273450/ ? | 02:29 |
| Qiming | elynn, I think Yanyanhu is online too | 02:31 |
| Yanyanhu | morning | 02:32 |
| elynn | Oh, just saw him :) | 02:32 |
| Qiming | elynn, it needs a rebase | 02:32 |
| xuhaiwei | I am also here, elynn | 02:33 |
| elynn | Qiming, If the codes looks good, I will rebase it soon. | 02:34 |
| Qiming | looking | 02:35 |
| Qiming | elynn, code looks fine | 02:39 |
| elynn | Thanks, I will do the rebase. | 02:40 |
| Yanyanhu | hi, elynn, I think the db issue caused by dead engine checking is still there... | 02:44 |
| elynn | Yanyanhu, yes... | 02:45 |
| Yanyanhu | about 50% chance to meet it when starting engine | 02:45 |
| elynn | Yanyanhu, you mean the warning in engine log? | 02:45 |
| Yanyanhu | yes | 02:45 |
| Yanyanhu | some complains like this: ObjectDeletedError: Instance '<Service at 0x315ca50>' has been deleted, or its row is otherwise not present. | 02:45 |
| elynn | https://bugs.launchpad.net/senlin/+bug/1538382 | 02:46 |
| openstack | Launchpad bug 1538382 in senlin "SAwarning when restart senlin-engine" [Undecided,New] | 02:46 |
| elynn | I open a bug but haven't got time to look into it. After I finished heat resources support I will investigate it... | 02:46 |
| Yanyanhu | thanks :) | 02:47 |
| openstackgerrit | Merged openstack/python-senlinclient: Apply list_formatter on two properties https://review.openstack.org/278540 | 02:47 |
| Yanyanhu | it's not that critical | 02:47 |
| openstackgerrit | Yanyan Hu proposed openstack/senlin: Remove 'force' kwarg from node_delete api/rpc https://review.openstack.org/280049 | 02:59 |
| elynn | Is there anyone working on kolla support for senlin? I notice that there is a BP for it, just curious that if this release will make it. https://blueprints.launchpad.net/kolla/+spec/senlin-container | 03:00 |
| openstackgerrit | Merged openstack/senlin-dashboard: Add receivers panel https://review.openstack.org/269528 | 03:07 |
| Qiming | woohooooo | 03:07 |
| Qiming | looks cool | 03:08 |
| Qiming | I'm looking forward to something else, the other direction, I mean | 03:14 |
| Qiming | creating kolla clusters using Senlin | 03:14 |
| Qiming | enable OpenStack HA in that way | 03:14 |
| Yanyanhu | Qiming, this is great. | 03:15 |
| xuhaiwei | kolla cluster? | 03:15 |
| Qiming | yes | 03:15 |
| xuhaiwei | go for lunch now, talk to you later, that seems interesting | 03:16 |
| Qiming | ttyl | 03:16 |
| *** Ayush has joined #senlin | 03:16 | |
| elynn | Qiming, That's cool, would like to know about it after lunch :) | 03:17 |
| openstackgerrit | Qiming Teng proposed openstack/senlin: Rework unit tests for profile service support https://review.openstack.org/280053 | 03:30 |
| openstackgerrit | Qiming Teng proposed openstack/senlin: Rework unit tests for policy service support https://review.openstack.org/280053 | 03:31 |
| openstackgerrit | Shu Muto proposed openstack/senlin-dashboard: Update URLs to Django 1.8 style https://review.openstack.org/274049 | 03:32 |
| openstackgerrit | Merged openstack/senlin: Remove 'force' kwarg from node_delete api/rpc https://review.openstack.org/280049 | 03:40 |
| *** elynn has quit IRC | 03:59 | |
| *** elynn has joined #senlin | 04:28 | |
| *** elynn has quit IRC | 04:32 | |
| *** elynn has joined #senlin | 04:33 | |
| openstackgerrit | Yanyan Hu proposed openstack/senlin: Add functional test for cluster node add/del https://review.openstack.org/280069 | 05:49 |
| elynn | Hi Yanyanhu | 05:58 |
| Yanyanhu | hi | 05:58 |
| elynn | Can you do cluster_node_add/del in latest codes? | 05:58 |
| Yanyanhu | elynn, yes, I think so | 05:59 |
| Yanyanhu | oh, I didn't test using nova server profile type | 05:59 |
| elynn | That's strange, it's always failed for me.. | 05:59 |
| Yanyanhu | just test it using fake driver | 05:59 |
| elynn | I'm using heat profile type. | 05:59 |
| Yanyanhu | ok, let me have a try | 05:59 |
| Yanyanhu | elynn, it failed as well in my env | 06:03 |
| elynn | If I create a node with a cluster id, then it works. But it won't work if I use cluster-node-add/del. | 06:03 |
| Yanyanhu | the following log msg was given: node_join_2e4bd28e [dc33c3eb] NODE_JOIN - FAILED: Node failed in joining cluster. | 06:03 |
| Yanyanhu | I guess there is some problems in heat profile or driver | 06:04 |
| elynn | Yanyanhu, yes, I got that error either, not much useful log. | 06:04 |
| Yanyanhu | let me check it | 06:04 |
| elynn | Is that because we haven't implement it yet? https://github.com/openstack/senlin/blob/master/senlin/profiles/os/heat/stack.py#L250-L256 | 06:07 |
| Qiming | elynn, those functions should return a boolean | 06:09 |
| Qiming | it was some garbage left from previous modifications | 06:09 |
| Yanyanhu | yes, that's is the reason | 06:11 |
| elynn | Oh, simply return true works. | 06:12 |
| Qiming | elynn, it was a hook left open to profile developers | 06:13 |
| Qiming | they may want to do something when joing a node into a cluster | 06:13 |
| Qiming | if there is nothing to do, these methods must return a True | 06:14 |
| Qiming | that is leading me to a different thought | 06:14 |
| Qiming | maybe we should return a True in the base class | 06:15 |
| Qiming | https://github.com/openstack/senlin/blob/master/senlin/profiles/base.py#L297-L299 | 06:15 |
| Qiming | it doesn't sound a nice design if we are forcing a developer to override the do_join, do_leave, do_get_details methods | 06:16 |
| elynn | yes, and also be careful with do_check and do_recover, maybe developers won't cover them either? | 06:17 |
| Qiming | those are some more important features I think | 06:18 |
| Qiming | if a profile doesn't support check/recover, the developer needs to explicitly say so ... | 06:19 |
| Yanyanhu | hmm... not sure which way is better | 06:20 |
| Yanyanhu | actually myself prefer the current logic :) | 06:20 |
| Yanyanhu | if a profile doesn't override do_join/leave routines, it doesn't support node join/leave | 06:21 |
| Qiming | no | 06:21 |
| Qiming | if a profile doesn't override them | 06:21 |
| Qiming | it means the profile doesn't need to take special actions during node join/leave | 06:22 |
| Qiming | I mean the base class can return a True by default | 06:22 |
| openstackgerrit | Merged openstack/senlin: Rework unit tests for policy service support https://review.openstack.org/280053 | 06:24 |
| Yanyanhu | I'm also ok with this logic. Just no sure which way is more common | 06:24 |
| Yanyanhu | for this kind of design :) | 06:25 |
| Yanyanhu | two different design logics | 06:25 |
| openstackgerrit | Qiming Teng proposed openstack/senlin: Rework unit tests for profile operations https://review.openstack.org/280080 | 06:25 |
| Qiming | em | 06:25 |
| Yanyanhu | interesting issue :) | 06:26 |
| Yanyanhu | so maybe for those cluster/node actions need to manipulate physical resource at backend (e.g. do_check, do_recover), profiles have to override related methods to support them | 06:29 |
| Qiming | yes, health checking is a very important requirement | 06:30 |
| Yanyanhu | for actions like node_join/leave which only manipulate senlin elements, we return true in routines in base class | 06:31 |
| Qiming | yep, that's my understanding | 06:31 |
| Yanyanhu | understand | 06:32 |
| openstackgerrit | Yanyan Hu proposed openstack/senlin: Check nova server status after rebuilding it https://review.openstack.org/280083 | 07:16 |
| Yanyanhu | hi, elynn, around? | 07:32 |
| Yanyanhu | are you working on the bug that happened when adding/removing heat stack node from cluster? | 07:33 |
| elynn | Yanyanhu, no | 07:33 |
| elynn | I'm working on node resource in heat. | 07:34 |
| Yanyanhu | ok, I can work on it if you don't have a plan :) | 07:34 |
| elynn | Thank you so much if you can take it :) | 07:34 |
| Yanyanhu | my pleasure | 07:34 |
| openstackgerrit | Qiming Teng proposed openstack/senlin: Refactor controllers in API layer https://review.openstack.org/280090 | 07:43 |
| openstackgerrit | Merged openstack/senlin: Check nova server status after rebuilding it https://review.openstack.org/280083 | 08:05 |
| openstackgerrit | Qiming Teng proposed openstack/senlin: Remove some dead options in engine config https://review.openstack.org/280098 | 08:13 |
| openstackgerrit | Yanyan Hu proposed openstack/senlin: Make do_join/leave in profile base return True https://review.openstack.org/280100 | 08:21 |
| openstackgerrit | Qiming Teng proposed openstack/senlin: Add node_count_by_cluster DB API https://review.openstack.org/280101 | 08:24 |
| openstackgerrit | Merged openstack/senlin: Improve param checking for event-list https://review.openstack.org/279933 | 08:43 |
| elynn | Hi Qiming , do we support 'actor' as a parameter to create receiver? | 08:46 |
| Qiming | yes | 08:46 |
| elynn | Since I see it in API doc but not support by senlinclient. | 08:47 |
| elynn | Do we need to expose it to heat resource? | 08:47 |
| Qiming | I don't see a use case there | 08:47 |
| elynn | ok, maybe don't expose it sounds good for now... | 08:48 |
| Qiming | yep | 08:48 |
| elynn | get it. | 08:49 |
| openstackgerrit | Merged openstack/senlin: Make do_join/leave in profile base return True https://review.openstack.org/280100 | 08:51 |
| openstackgerrit | Qiming Teng proposed openstack/senlin: Enforce 'max_clusters_per_project' quota https://review.openstack.org/280110 | 08:55 |
| openstackgerrit | Qiming Teng proposed openstack/senlin: Refactor controllers in API layer https://review.openstack.org/280090 | 09:09 |
| Yanyanhu | hi, Qiming, about removing unused engine config options https://review.openstack.org/#/c/280098/ | 09:19 |
| Yanyanhu | I think maybe we should make 'region_name_for_services' take effect | 09:19 |
| Yanyanhu | currently, if there are services from multiple regions registered to a single keytone, the location that physical resources created will be uncertain | 09:21 |
| Yanyanhu | unless user defines region name in context peoperty explictly in profile spec | 09:22 |
| Yanyanhu | so maybe a default region name is useful for this case | 09:22 |
| openstackgerrit | Merged openstack/senlin: Add node_count_by_cluster DB API https://review.openstack.org/280101 | 09:23 |
| Yanyanhu | oh, let me check whether sdk will fill in this value by default | 09:25 |
| Yanyanhu | seems it doesn't do this job | 09:27 |
| Qiming | we are not using it anyway | 09:28 |
| Yanyanhu | yes | 09:28 |
| Qiming | we can add such support if it is used somewhere | 09:28 |
| Yanyanhu | but I think we should enable it | 09:28 |
| Yanyanhu | since senlin should be workable in multi-region env | 09:28 |
| Qiming | yes it should | 09:29 |
| Yanyanhu | ok, will try to make it take effect | 09:30 |
| Qiming | okay | 09:30 |
| Qiming | NOTE: openstacksdk is undertaking quite some disruptive changes lately | 09:36 |
| Qiming | please don't clone the latest master for testing | 09:37 |
| Yanyanhu | ... | 09:37 |
| Qiming | it will break everything we have | 09:37 |
| Qiming | https://review.openstack.org/279618 | 09:37 |
| Yanyanhu | that patch landed? | 09:37 |
| Qiming | https://review.openstack.org/279581 | 09:37 |
| Qiming | https://review.openstack.org/279089 | 09:37 |
| Yanyanhu | I see | 09:38 |
| Qiming | https://review.openstack.org/270900 | 09:38 |
| Yanyanhu | so large amount of change is needed for senlin driver | 09:38 |
| Qiming | YES | 09:38 |
| Yanyanhu | sigh | 09:39 |
| Qiming | We will wait for the 0.8.0 release or 1.0 release of openstacksdk | 09:39 |
| Qiming | before that, please stick to 0.7.4 | 09:39 |
| Yanyanhu | will try to work on it and add the dependency | 09:39 |
| Qiming | Yanyanhu, which one? | 09:46 |
| Yanyanhu | the driver rework | 09:46 |
| Qiming | okay, let's hold on | 09:47 |
| Qiming | before 0.8.0 is released, we should not do anything | 09:47 |
| Qiming | it would be a huge change | 09:47 |
| Yanyanhu | yes | 09:47 |
| Yanyanhu | will try to figure out how many change we need | 09:47 |
| Qiming | I just sent an email to the sdk cores | 09:47 |
| Qiming | need to figure out their plan | 09:47 |
| Qiming | either they have to get this done by next week | 09:48 |
| Qiming | or they should cap the global requirements to 0.7.4 | 09:48 |
| Yanyanhu | I see | 09:48 |
| Yanyanhu | this is a disruptive change | 09:48 |
| Qiming | yep, it breaks almost all calls to sdk | 09:49 |
| Yanyanhu | yes | 09:49 |
| openstackgerrit | Qiming Teng proposed openstack/senlin: Refactor controllers in API layer https://review.openstack.org/280090 | 09:57 |
| *** elynn has quit IRC | 10:03 | |
| *** lixinhui has joined #senlin | 10:06 | |
| openstackgerrit | Yanyan Hu proposed openstack/senlin: Remove some dead options in engine config https://review.openstack.org/280098 | 10:12 |
| openstackgerrit | Yanyan Hu proposed openstack/senlin: Enable region_name_for_services option https://review.openstack.org/280132 | 10:12 |
| Yanyanhu | hi, Qiming, I download this patch and rebase mine on it https://review.openstack.org/#/c/280098/, then it will be resubmitted as well when I submit my patch? | 10:13 |
| Qiming | seems so | 10:14 |
| Yanyanhu | sigh... | 10:14 |
| Yanyanhu | any other way to rebase a patch on other guys' in-progress one? | 10:15 |
| Qiming | I don't know | 10:16 |
| Qiming | I always hate rebase ... and always get messed up | 10:16 |
| Yanyanhu | ok, seems gerrit can't handle this situation | 10:16 |
| Qiming | so I prefer merge before working on dependent patches | 10:17 |
| Yanyanhu | yes, rebase is always headache | 10:17 |
| Qiming | 'git rebase -i' sometimes doesn't work either | 10:17 |
| Yanyanhu | ok... | 10:17 |
| Qiming | you got to specify the exact change commit-id | 10:17 |
| Qiming | and do a `git rebase -i` twice ... | 10:17 |
| Yanyanhu | ... | 10:18 |
| openstackgerrit | Merged openstack/senlin: Refactor controllers in API layer https://review.openstack.org/280090 | 10:21 |
| Qiming | doing a rebase | 10:22 |
| Qiming | workflow +1 | 10:22 |
| Qiming | see if it can get merged | 10:22 |
| Yanyanhu | ok | 10:22 |
| Qiming | walking home | 10:25 |
| Qiming | bbl | 10:25 |
| Qiming | hope I can forget that I'm so hungry along the way | 10:25 |
| Yanyanhu | see U | 10:26 |
| Yanyanhu | will leave after finish those two patches | 10:26 |
| Yanyanhu | ... | 10:26 |
| openstackgerrit | Merged openstack/senlin: Remove some dead options in engine config https://review.openstack.org/280098 | 10:27 |
| *** Qiming has quit IRC | 10:30 | |
| openstackgerrit | Yanyan Hu proposed openstack/senlin: Enable default_region_name config option https://review.openstack.org/280132 | 10:33 |
| *** Yanyanhu has quit IRC | 10:44 | |
| *** Qiming has joined #senlin | 11:30 | |
| *** junxu has quit IRC | 12:10 | |
| openstackgerrit | Merged openstack/senlin: Remove useless encrypt/decrypt methods https://review.openstack.org/279986 | 12:16 |
| *** openstackgerrit_ has joined #senlin | 12:28 | |
| *** junxu has joined #senlin | 13:02 | |
| *** lixinhui has quit IRC | 13:44 | |
| openstackgerrit | Merged openstack/senlin: Enable default_region_name config option https://review.openstack.org/280132 | 13:45 |
| *** junxu has quit IRC | 14:16 | |
| openstackgerrit | zhangguoqing proposed openstack/senlin: tools/setup-service: Add email info when user-create https://review.openstack.org/279958 | 15:07 |
| *** Qiming has quit IRC | 15:11 | |
| *** openstackgerrit has quit IRC | 17:45 | |
| *** openstackgerrit has joined #senlin | 17:46 | |
| *** openstackgerrit has quit IRC | 20:17 | |
| *** openstackgerrit_ is now known as openstackgerrit | 20:17 | |
| *** openstackgerrit has quit IRC | 20:17 | |
| *** openstackgerrit_ has joined #senlin | 20:18 | |
| *** openstackgerrit_ is now known as openstackgerrit | 20:18 | |
| *** openstackgerrit_ has joined #senlin | 20:19 | |
| *** sridhar_ram has joined #senlin | 21:25 | |
| *** sridhar_ram has quit IRC | 23:21 | |
| *** yuanying_flu is now known as yuanying | 23:29 | |
| *** Qiming has joined #senlin | 23:36 | |
| openstackgerrit | Merged openstack/senlin: tools/setup-service: Add email info when user-create https://review.openstack.org/279958 | 23:46 |
Generated by irclog2html.py 2.14.0 by Marius Gedminas - find it at mg.pov.lt!