| *** wanghao has joined #openstack-mogan | 00:39 | |
| wanghao | morning! | 00:53 |
|---|---|---|
| zhenguo | morning mogan and wanghao | 01:15 |
| *** wanghao_ has joined #openstack-mogan | 01:17 | |
| *** wanghao has quit IRC | 01:21 | |
| liusheng | morning | 01:21 |
| openstackgerrit | liusheng proposed openstack/python-moganclient master: Format the output image uuid to name https://review.openstack.org/498998 | 01:21 |
| zhenguo | liusheng: maybe better to format the server/networks | 01:41 |
| zhenguo | liusheng: seems currently we it's ports instead of networks | 01:41 |
| liusheng | zhenguo: you mean the network commands ? | 01:41 |
| zhenguo | liusheng: the Api | 01:42 |
| zhenguo | liusheng: like you GET servers/networks but actually get a list of ports | 01:42 |
| zhenguo | liusheng: currently it's like GET servers/networks/ports | 01:43 |
| liusheng | zhenguo: maybe the server's networks is its netowrk info which is its ports info | 01:43 |
| zhenguo | liusheng: seems a bit weird | 01:44 |
| liusheng | zhenguo: but servers/networks/ports seems a bit redundant | 01:45 |
| zhenguo | liusheng: yes | 01:46 |
| liusheng | zhenguo: maybe servers/interfaces ? | 01:46 |
| liusheng | zhenguo: in Nova's server body, it also call the interfaces "network" | 01:47 |
| zhenguo | liusheng: we need to make the notion clear about ports, interfaces, networks, and addresses in mogan | 01:48 |
| liusheng | zhenguo: maybe networks of a server won't make users confused, hah, not sure | 01:48 |
| zhenguo | liusheng: maybe yes | 01:49 |
| liusheng | zhenguo: it is easy to make users realize that is network configurations of a server, including ip, mac, net, floatingip, blabla | 01:50 |
| zhenguo | liusheng: ok , let me check the API again | 01:50 |
| openstackgerrit | Zhenguo Niu proposed openstack/mogan master: Address bad indentation in docs https://review.openstack.org/499003 | 01:54 |
| openstackgerrit | Xinran WANG proposed openstack/mogan master: support specifying port_id when attaching interface https://review.openstack.org/494121 | 02:01 |
| zhenguo | liusheng: please have a look at this https://review.openstack.org/#/c/499003/ | 02:10 |
| liusheng | zhenguo: ok | 02:11 |
| zhenguo | liusheng: the 'aggregate list node' still conflict with 'aggregate list' right? | 02:16 |
| liusheng | zhenguo: yes.. :( | 02:17 |
| zhenguo | liusheng: but seems it works | 02:18 |
| liusheng | zhenguo: really ? | 02:19 |
| liusheng | zhenguo: let me try | 02:19 |
| liusheng | zhenguo: yes, don't know how it don't conflict any more, hah | 02:20 |
| zhenguo | liusheng: so we can keep it? | 02:20 |
| liusheng | zhenguo: yes, sure | 02:22 |
| zhenguo | liusheng: ok | 02:22 |
| zhenguo | liusheng: maybe we need to add the CLI docs | 02:22 |
| liusheng | zhenguo: yes, we need to improve the client docs | 02:23 |
| *** litao__ has joined #openstack-mogan | 02:25 | |
| Xinran | zhenguo, do you know why I cant find created key in DB. It's always empty set | 02:25 |
| zhenguo | Xinran: new created? | 02:26 |
| Xinran | zhenguo, yes | 02:27 |
| zhenguo | Xinran: I just tried, it presents in the key_pairs table | 02:27 |
| Xinran | zhenguo, in my env, it's empty.. | 02:28 |
| Xinran | maybe some problem in my env | 02:28 |
| zhenguo | Xinran: seems yes, | 02:28 |
| Xinran | zhenguo, ok, let me check | 02:29 |
| zhenguo | Xinran: for the key_name task, you can save the key with server when we create the server db record, | 02:29 |
| zhenguo | /me brb | 02:37 |
| openstackgerrit | liusheng proposed openstack/python-moganclient master: Format the output image uuid to name https://review.openstack.org/498998 | 02:38 |
| *** Xinran has quit IRC | 02:40 | |
| *** Xinran has joined #openstack-mogan | 02:41 | |
| litao__ | zhenguo: which is better? cast or call for managing server , what do you think? | 03:00 |
| zhenguo | litao__: I would like call | 03:00 |
| zhenguo | litao__: and I don't like the way to create db record first | 03:00 |
| zhenguo | litao__: we should prepare/check all things before that happen instead of leaving a dead server when error | 03:01 |
| litao__ | zhenguo: Maybe call ironic API and neutron api wastes time | 03:01 |
| zhenguo | litao__: a little time | 03:01 |
| zhenguo | litao__: which is acceptable | 03:01 |
| litao__ | zhenguo: maybe OK | 03:04 |
| zhenguo | litao__: which make things simple | 03:04 |
| zhenguo | litao__: you can do some basic check on api side, then to engine which constructs network info and set ironic node instance info, then create the db record | 03:05 |
| zhenguo | litao__: seems another API call can also manage the node before you set instance_uuid there, right? | 03:06 |
| zhenguo | litao__; seems collisions may happen | 03:06 |
| litao__ | zhenguo: If the db record failed, should we rollback the ironic node instance info? | 03:07 |
| zhenguo | litao__: yes | 03:07 |
| zhenguo | litao__: you just override the instance_uuid, right? | 03:09 |
| litao__ | zhenguo: so i should lock the node when managing it | 03:09 |
| zhenguo | litao__: do you think collisions may happen | 03:09 |
| litao__ | zhenguo: maybe | 03:09 |
| zhenguo | litao__: yes, seems you should lock it, and instead of just overridding, maybe better to check the instance_uuid existence, as admins may already remember the node uuid and manage it again and again | 03:10 |
| litao__ | zhenguo: The get_manageable_servers method ensure to get the node without instance_uuid | 03:13 |
| zhenguo | litao__: seems yes :D | 03:13 |
| zhenguo | litao__: so you will still list all nodes and find the one you want? | 03:14 |
| zhenguo | litao__: can we provide a get single manageable node interface? | 03:14 |
| litao__ | zhenguo: so I should also clear the instance _uuid if manage failed? | 03:15 |
| zhenguo | litao__: also ironic node list API but specify the node_uuid, not sure if we can | 03:15 |
| zhenguo | litao__: yes, seems only when creating db failed? | 03:15 |
| litao__ | zhenguo: yes， get method is ok | 03:16 |
| zhenguo | litao__: ok | 03:16 |
| litao__ | zhenguo: Implemented in API? | 03:16 |
| litao__ | zhenguo: or just internal method | 03:16 |
| zhenguo | litao__: just driver interface | 03:17 |
| litao__ | zhenguo: OK | 03:19 |
| *** wanghao_ has quit IRC | 03:32 | |
| *** wanghao has joined #openstack-mogan | 03:33 | |
| * zhenguo brb | 03:40 | |
| zhenguo | liusheng: hi, currently active node without instance_uuid will not got to placement, right? | 03:50 |
| liusheng | zhenguo: yes | 03:52 |
| zhenguo | liusheng: so it's ok for manageable servers | 03:52 |
| liusheng | zhenguo: may need to imporove as we disscussed | 03:52 |
| zhenguo | liusheng: yes | 03:53 |
| zhenguo | liusheng: I tested the server group patch, it works well | 03:54 |
| zhenguo | liusheng: just left some nits inline | 03:54 |
| liusheng | zhenguo: ok, will update that according to your comments | 03:54 |
| zhenguo | liusheng: thanks | 03:54 |
| * zhenguo brb | 03:54 | |
| openstackgerrit | liusheng proposed openstack/mogan master: Use server group in scheduler https://review.openstack.org/496151 | 04:07 |
| Xinran | Xinran, ok, let me try thanks | 04:10 |
| *** wanghao_ has joined #openstack-mogan | 04:57 | |
| *** wanghao has quit IRC | 04:58 | |
| *** shaohe_feng has quit IRC | 05:46 | |
| *** XueFengLiu has quit IRC | 05:47 | |
| *** XueFengLiu has joined #openstack-mogan | 05:47 | |
| *** shaohe_feng has joined #openstack-mogan | 05:51 | |
| *** wanghao has joined #openstack-mogan | 05:59 | |
| *** XueFengLiu has quit IRC | 06:02 | |
| *** XueFengLiu has joined #openstack-mogan | 06:02 | |
| *** wanghao_ has quit IRC | 06:02 | |
| zhenguo | liusheng: hi | 06:19 |
| zhenguo | liusheng: about the log messages, why print the filtered nodes? | 06:20 |
| liusheng | zhenguo: hi | 06:20 |
| zhenguo | liusheng: it's just because you filter other aggregates first | 06:20 |
| liusheng | zhenguo: I think it is useful we printing more info if a InvalidNode error return | 06:20 |
| zhenguo | liusheng: it's not useful at all, and I got confuse with that info | 06:21 |
| liusheng | zhenguo: why ? | 06:21 |
| zhenguo | liusheng: It's just because we filter other aggregates first in our internal process | 06:21 |
| liusheng | zhenguo: no, it is not related with the order | 06:22 |
| zhenguo | liusheng: ? | 06:22 |
| zhenguo | liusheng: you print the filtered nodes before filter afffinity zone | 06:22 |
| liusheng | zhenguo: ... it is because you know the order of process | 06:23 |
| zhenguo | liusheng: why users need that info | 06:23 |
| zhenguo | liusheng: I know you have how many nodes left in the cloud? | 06:23 |
| liusheng | zhenguo: it is help to find the reason about why InvalidNode | 06:23 |
| liusheng | zhenguo: ok, will remove it | 06:24 |
| zhenguo | liusheng: we should not treat affinity zone special as during filtering it's not different with other aggregates | 06:24 |
| openstackgerrit | liusheng proposed openstack/mogan master: Use server group in scheduler https://review.openstack.org/496151 | 06:25 |
| openstackgerrit | Zhenguo Niu proposed openstack/mogan master: [WIP]Add root disk partitions support https://review.openstack.org/499039 | 06:27 |
| *** wanghao has quit IRC | 06:35 | |
| *** wanghao has joined #openstack-mogan | 06:36 | |
| litao__ | zhenguo: There is no state transition in managing node? | 06:45 |
| litao__ | zhenguo: right? | 06:46 |
| zhenguo | litao__: yes | 06:46 |
| litao__ | zhenguo: But If add some other process such as vif plug later, is it ok? | 06:46 |
| zhenguo | litao__: seems ok | 06:48 |
| litao__ | zhenguo: ok | 06:48 |
| zhenguo | litao__: but you should take care about the server status, it can be active or stopped according to the power state | 06:48 |
| litao__ | zhenguo: of course | 06:49 |
| zhenguo | litao__: ok | 06:49 |
| openstackgerrit | Xinran WANG proposed openstack/mogan master: support specifying port_id when attaching interface https://review.openstack.org/494121 | 07:05 |
| *** wanghao_ has joined #openstack-mogan | 08:14 | |
| *** wanghao_ has quit IRC | 08:15 | |
| *** wanghao has quit IRC | 08:15 | |
| *** wanghao has joined #openstack-mogan | 08:16 | |
| openstackgerrit | Merged openstack/mogan master: Address bad indentation in docs https://review.openstack.org/499003 | 09:21 |
| *** dims has quit IRC | 09:28 | |
| *** wanghao has quit IRC | 09:35 | |
| *** wanghao has joined #openstack-mogan | 09:35 | |
| *** wanghao has quit IRC | 09:36 | |
| *** wanghao has joined #openstack-mogan | 09:36 | |
| *** wanghao has quit IRC | 09:36 | |
| *** wanghao has joined #openstack-mogan | 09:37 | |
| *** wanghao has quit IRC | 09:37 | |
| *** wanghao has joined #openstack-mogan | 09:38 | |
| *** wanghao has quit IRC | 09:38 | |
| *** wanghao has joined #openstack-mogan | 09:38 | |
| *** wanghao has quit IRC | 09:39 | |
| *** wanghao has joined #openstack-mogan | 09:39 | |
| *** wanghao has quit IRC | 09:40 | |
| *** wanghao has joined #openstack-mogan | 09:40 | |
| *** wanghao has quit IRC | 09:40 | |
| *** wanghao has joined #openstack-mogan | 09:41 | |
| Xinran | Hi, all . Does anyone know the command line to sync db ? | 09:57 |
| openstackgerrit | Merged openstack/mogan master: Use server group in scheduler https://review.openstack.org/496151 | 09:57 |
| openstackgerrit | Merged openstack/mogan master: Add socat console support https://review.openstack.org/493836 | 09:57 |
| openstackgerrit | Merged openstack/mogan master: support specifying port_id when attaching interface https://review.openstack.org/494121 | 09:58 |
| *** dims has joined #openstack-mogan | 09:58 | |
| openstackgerrit | Merged openstack/python-moganclient master: Format the output image uuid to name https://review.openstack.org/498998 | 10:01 |
| zhenguo | Xinran: as we don't add db migration scripts, dbsync doens't work for us now :( | 10:17 |
| openstackgerrit | Zhenguo Niu proposed openstack/mogan master: [WIP]Add root disk partitions support https://review.openstack.org/499039 | 10:44 |
| *** Jeffrey4l has quit IRC | 12:07 | |
| *** liusheng has quit IRC | 12:07 | |
| *** liusheng has joined #openstack-mogan | 12:07 | |
| *** Jeffrey4l has joined #openstack-mogan | 12:07 | |
| *** dims has quit IRC | 17:14 | |
| *** litao__ has quit IRC | 17:17 | |
| *** dims has joined #openstack-mogan | 17:18 | |
| *** dims has quit IRC | 19:28 | |
| *** dims has joined #openstack-mogan | 19:31 | |
Generated by irclog2html.py 2.15.3 by Marius Gedminas - find it at mg.pov.lt!