| *** wanghao has joined #openstack-mogan | 00:38 | |
| zhenguo | morning mogan! | 00:52 |
|---|---|---|
| *** litao__ has joined #openstack-mogan | 01:19 | |
| *** _liusheng has joined #openstack-mogan | 01:40 | |
| openstackgerrit | Zhenguo Niu proposed openstack/mogan master: Fix partitions error when pass down to ironic. https://review.openstack.org/504818 | 01:55 |
| zhenguo | liusheng, shaohe_feng: please have a look at this https://review.openstack.org/#/c/504818/ when you are online, thanks! | 02:00 |
| *** wanghao has quit IRC | 02:09 | |
| zhenguo | litao__: seems there's a bug for mult server creation | 02:22 |
| zhenguo | litao__: you will only check the max when scheudling | 02:22 |
| litao__ | zhenguo: Let me see | 02:26 |
| zhenguo | litao__: ok, seems min is useless | 02:26 |
| *** wanghao has joined #openstack-mogan | 02:27 | |
| litao__ | zhenguo: In scheduling ,I just check the number of instances. The max and min are checked in api layer | 02:29 |
| litao__ | zhenguo: the number of instances maybe the max or the min | 02:30 |
| openstackgerrit | wanghao proposed openstack/mogan master: Introduce Cinder client into Mogan https://review.openstack.org/489455 | 02:39 |
| zhenguo | litao__: if there's only 2 nodes available, and the min just set to 2 max to 5, should it fail all? | 02:41 |
| litao__ | zhenguo: Yes, it should be success | 02:48 |
| openstackgerrit | wanghao proposed openstack/mogan-specs master: Add Queens folder in mogan spec repo https://review.openstack.org/504826 | 02:48 |
| zhenguo | litao__; but actually, it's not | 02:48 |
| wanghao | guys, I add queens folder in our spec repo. | 02:49 |
| zhenguo | wanghao: thanks | 02:49 |
| wanghao | we can commit bp for queens now, : ) | 02:49 |
| wanghao | zhenguo: np | 02:49 |
| litao__ | zhenguo: Yes it indeed not | 02:49 |
| zhenguo | litao__: how to improve that | 02:50 |
| litao__ | zhenguo: so should It be check in scheduling? | 02:50 |
| zhenguo | litao__: you men pass both min and max there? | 02:50 |
| litao__ | maybe it is one way | 02:51 |
| zhenguo | litao__: or maybe we should move the server number check out | 02:51 |
| zhenguo | litao__: we should request the max, but if got less than max we need to check with the min | 02:52 |
| zhenguo | litao__; do you know how dose nova handle that? | 02:52 |
| litao__ | zhenguo: nova checks the number of instances in scheduling | 02:54 |
| zhenguo | litao__: it's max? | 02:54 |
| zhenguo | litao__: but seems they can successfully built the min servers | 02:54 |
| litao__ | zhenguo: you mean nova with ironic? | 02:59 |
| zhenguo | litao__: with VMs | 03:02 |
| litao__ | zhenguo: I checked the nova code, It should fail in your case above. The nova also check the pass the number of instances. | 03:02 |
| zhenguo | litao__: oh, so what's the min used for | 03:03 |
| litao__ | zhenguo: It is a little different with vm | 03:03 |
| litao__ | zhenguo: Just check the quota in api layer | 03:03 |
| zhenguo | litao__: you mean for bare metal, it will fail instead of vm? | 03:03 |
| litao__ | zhenguo: yes | 03:03 |
| zhenguo | litao__: sigh | 03:04 |
| zhenguo | litao__: but we can do some improvement | 03:04 |
| litao__ | zhenguo: In vm case , the hypervisor is always enough resources. | 03:04 |
| zhenguo | litao__: why? | 03:05 |
| litao__ | zhenguo: A host can be used by many vms. | 03:06 |
| zhenguo | litao__: but it can also be used out | 03:06 |
| litao__ | zhenguo: If it is used out , there will be also fail in scheduling | 03:08 |
| zhenguo | litao__: I mean the avaialble resources are just between min and max | 03:08 |
| zhenguo | litao__; It should be fail if less the min, but on the contrary, it should not | 03:09 |
| litao__ | zhenguo:Yes，If the min is 2， max is 5. then the quota checked is 5 in api layer, nova will pass 5 to scheduler, if the scheduler does not get enough resources, 5 instances will fail | 03:11 |
| zhenguo | litao__: you mean min max, is just for quota? | 03:11 |
| zhenguo | litao__; instead of scheduling? | 03:11 |
| litao__ | zhenguo :Yes in nova | 03:12 |
| zhenguo | litao__: sigh | 03:12 |
| litao__ | zhenguo: we can improve it in mogan | 03:12 |
| litao__ | zhenguo: Maybe in baremetal scenario, this issue have high probability | 03:14 |
| zhenguo | litao__: how do we handle the quota checking now? | 03:14 |
| litao__ | zhenguo: The instance number | 03:16 |
| litao__ | zhenguo: and the port quota | 03:16 |
| zhenguo | litao__: will check with min and max like nova, right? | 03:16 |
| litao__ | zhenguo: yes, follow the nova style | 03:17 |
| zhenguo | litao__: ok | 03:17 |
| zhenguo | litao__: we can improve it later after we tag the 0.1.0 release | 03:17 |
| litao__ | zhenguo: when do we tag the 0.1.0 release? | 03:18 |
| zhenguo | litao__: maybe today | 03:18 |
| litao__ | zhenguo: OK, I will improve it | 03:18 |
| zhenguo | litao__: I'm testing now to ensure there's no critical issues | 03:18 |
| zhenguo | litao__: ok, thanks | 03:18 |
| zhenguo | litao__: btw, can you provide bare metals now through your production? | 03:19 |
| litao__ | zhenguo: np | 03:20 |
| litao__ | zhenguo: yes, we can | 03:20 |
| zhenguo | litao__: with a separated nova or just under the same nova with VMs? | 03:20 |
| litao__ | zhenguo: Together with VMs in the same nova | 03:21 |
| zhenguo | litao__: ok, got it | 03:21 |
| litao__ | zhenguo: 😀 | 03:22 |
| zhenguo | litao__: same with FS, hah | 03:23 |
| litao__ | zhenguo: FS is a great production | 03:24 |
| zhenguo | litao__: lol | 03:24 |
| litao__ | zhenguo: hah | 03:24 |
| liusheng | zhenguo: done, I am back | 03:26 |
| zhenguo | liusheng: cool | 03:26 |
| liusheng | zhenguo: I have posted a doc patch, please help to review of you have time: https://review.openstack.org/#/c/504780/ | 03:26 |
| zhenguo | liusheng: ok | 03:27 |
| * zhenguo brb | 03:30 | |
| *** wanghao has quit IRC | 03:46 | |
| openstackgerrit | liusheng proposed openstack/mogan-specs master: Use a soft link for template.rst https://review.openstack.org/504840 | 03:48 |
| *** _liusheng has quit IRC | 03:50 | |
| zhenguo | liusheng: hi, have you built the notification docs with RTD? | 03:51 |
| liusheng | zhenguo: not sure, just test with tox -edocs | 03:51 |
| zhenguo | liusheng: you can try to build one, then we can review based on that | 03:52 |
| liusheng | zhenguo: ok, will try | 03:53 |
| zhenguo | lliusheng: thanks | 03:53 |
| * zhenguo brb | 03:53 | |
| openstackgerrit | Merged openstack/mogan master: Updated from global requirements https://review.openstack.org/504650 | 04:06 |
| *** wanghao has joined #openstack-mogan | 04:30 | |
| *** wanghao has quit IRC | 04:44 | |
| *** wanghao has joined #openstack-mogan | 05:16 | |
| *** wanghao has quit IRC | 05:31 | |
| *** wanghao has joined #openstack-mogan | 05:33 | |
| openstackgerrit | Merged openstack/mogan master: Fix partitions error when pass down to ironic. https://review.openstack.org/504818 | 05:42 |
| liusheng | zhenguo: I cannot build the RTD doc with master mogan | 06:07 |
| zhenguo | liusheng: you can build with you own repo | 06:11 |
| liusheng | zhenguo: just a private repo, without any change ? | 06:11 |
| zhenguo | liusheng: you can modify the test-requirements accroding to the building errors | 06:12 |
| liusheng | zhenguo: ok | 06:12 |
| liusheng | zhenguo: do you know is there any way to sync openstack/mogan repo to my own repo ? | 06:13 |
| zhenguo | liusheng: https://gist.github.com/niuzhenguo/f7a97b1db177c6a643fe78bb3d51fc69 | 06:13 |
| liusheng | zhenguo: I cannot open this, is this another way to github ? | 06:15 |
| zhenguo | liusheng: I can paste it via espace to you | 06:16 |
| wanghao | liusheng: ping | 06:29 |
| liusheng | wanghao: pong | 06:29 |
| wanghao | liusheng: did you have some example to add soft link for template.rst ? | 06:30 |
| wanghao | liusheng: I didn't find how to add it. | 06:30 |
| liusheng | wanghao: just use command: ln -s ../../template.rst template.rst | 06:31 |
| liusheng | wanghao: I remember many projects use like this | 06:31 |
| liusheng | wanghao, but seems for now, some project have different spec template.rst for differet version, not sure if we need that | 06:35 |
| liusheng | zhenguo: ^ | 06:35 |
| zhenguo | liusheng: we follow core projects | 06:35 |
| wanghao | liusheng: and I didn't find where to add this command, I check out zaqar and cinder spec. | 06:36 |
| liusheng | zhenguo: core projects like what I just said, hah | 06:36 |
| wanghao | liusheng: where, can you show me this? | 06:36 |
| liusheng | wanghao: I mean run this command | 06:36 |
| zhenguo | liusheng: for now, seems no need to add different tempelates | 06:37 |
| liusheng | wanghao: specs/queens/implemented/queens-template.rst | 06:37 |
| wanghao | liusheng: where to run? local computer? | 06:37 |
| liusheng | wanghao: nova-specs | 06:37 |
| liusheng | wanghao: yes | 06:37 |
| liusheng | wanghao: just a common sort link | 06:38 |
| openstackgerrit | wanghao proposed openstack/mogan-specs master: Add Queens folder in mogan spec repo https://review.openstack.org/504826 | 06:42 |
| wanghao | liusheng: that works | 06:43 |
| liusheng | wanghao: ok, thanks | 06:43 |
| wanghao | just one line: ../../templates.rst | 06:43 |
| liusheng | wanghao: yes | 06:44 |
| wanghao | liusheng: so I will update other folders too. | 06:44 |
| liusheng | wanghao: I have updated, hah, but not sure if zhenguo think we should use different templates for Ocata/Piki/... | 06:44 |
| wanghao | liusheng: okay | 06:45 |
| liusheng | zhenguo: do you know how to handle the error: AttributeError: 'Sphinx' object has no attribute 'add_html_theme' | 07:17 |
| zhenguo | liusheng: oh, I never seen that before | 07:17 |
| liusheng | zhenguo: I just added some libs in test-requirements in my onw repo according to the previous erros, after all, it raise this error | 07:18 |
| liusheng | zhenguo: what changes you made in your own repo to make it works with RTD ? | 07:19 |
| zhenguo | liusheng: just add oslo genconfig | 07:19 |
| zhenguo | liusheng: oslo policy | 07:20 |
| liusheng | zhenguo: I added oslo.policy, oslo.cache, oslo.messaging | 07:20 |
| zhenguo | liusheng: seems I don't got that errors | 07:21 |
| liusheng | zhenguo: did you enable openstackdocstheme? | 07:21 |
| zhenguo | liusheng: yes | 07:22 |
| liusheng | zhenguo: oh, werid | 07:22 |
| zhenguo | liusheng: I didn't change openstack docs | 07:22 |
| zhenguo | liusheng: I find another problem when testing mogan | 07:23 |
| liusheng | zhenguo: the error is unrelate with my patch | 07:23 |
| zhenguo | liusheng: seems console is useless if we don't support setting password | 07:23 |
| zhenguo | liusheng: you mean the repo without your patch will raise the same errors? | 07:23 |
| liusheng | zhenguo: I remember that in my previous testing | 07:24 |
| liusheng | zhenguo: but image may have default password, right ? | 07:24 |
| zhenguo | liusheng: default password is dangerous | 07:25 |
| liusheng | zhenguo: myabe for testing ? or will support password injection in the future ? | 07:26 |
| zhenguo | liusheng: we need to support password injection | 07:26 |
| zhenguo | liusheng: can you create a keypair with public key specified? | 07:32 |
| liusheng | zhenguo: you mean importing ? | 07:32 |
| liusheng | zhenguo: I remember tested ok | 07:33 |
| zhenguo | liusheng: yes | 07:33 |
| zhenguo | liusheng: I remember it's also ok, but not sure why it failed in my env | 07:33 |
| litao__ | zhenguo: Only in scheduler we known the available nodes, but the servers are allocated in db before scheduling. should we delete the severs in db if some server is not available node. | 07:34 |
| litao__ | s/is/has/ | 07:34 |
| zhenguo | litao__: you are working on that? | 07:36 |
| litao__ | zhenguo: yes | 07:36 |
| zhenguo | litao__: but users may be already aware of the db records | 07:39 |
| litao__ | zhenguo: yes, so it is not suitable | 07:40 |
| litao__ | zhenguo: Unless mogan knows the available nodes in api layer before allocated in db | 07:41 |
| zhenguo | litao__: seems not easy to fix that | 07:45 |
| zhenguo | litao__: even if you know that, it maybe occupied by others before you | 07:45 |
| litao__ | zhenguo: yes | 07:46 |
| litao__ | zhenguo: Maybe we can create servers with the same value in max and min to solve this issue. | 07:52 |
| zhenguo | liusheng: you mean just let users specify a server number instead of min and max? | 07:53 |
| zhenguo | wrong person | 07:53 |
| zhenguo | litao__ ^^ | 07:53 |
| zhenguo | liusheng: can you please confirm whether the keypair importing work? | 07:53 |
| zhenguo | liusheng: I encountered a issue which I remember already been addressed before | 07:54 |
| liusheng | zhenguo: ok, will test | 07:55 |
| litao__ | zhenguo: yes, but we can keep the current implement | 07:55 |
| zhenguo | litao__: yes | 07:56 |
| zhenguo | litao__: maybe just keep it until somebody complain it | 07:56 |
| litao__ | zhenguo: yes | 07:56 |
| zhenguo | liusheng: just found we forgot to set project_id when importing keypair | 08:02 |
| zhenguo | liusheng: which introduced with quota | 08:02 |
| liusheng | zhenguo: it also failed because the null "project_id" colomn | 08:05 |
| zhenguo | liusheng: yes | 08:05 |
| liusheng | zhenguo: not sure why, I remember I have tested ok previously | 08:05 |
| zhenguo | liusheng: yes, I also remember it | 08:06 |
| zhenguo | liusheng: when I adding partitions support | 08:06 |
| liusheng | zhenguo: ok, will fix it later | 08:06 |
| zhenguo | liusheng: oh, seems that's because I changed the db | 08:06 |
| zhenguo | liusheng: to make model and db script consistent | 08:06 |
| zhenguo | liusheng: which makes the project id a required | 08:07 |
| liusheng | zhenguo: oh, maybe should set the project_id ? | 08:07 |
| zhenguo | liusheng: you can compare create_key_pair and import_key_pair | 08:07 |
| liusheng | zhenguo: ok | 08:07 |
| zhenguo | liusheng: seems we only add that for create when intruducing quota | 08:08 |
| zhenguo | liusheng: also need to add quotas | 08:08 |
| liusheng | zhenguo: yes, in TODO list, :( | 08:08 |
| zhenguo | liusheng: also https://github.com/openstack/mogan/blob/master/mogan/common/exception.py#L457 | 08:09 |
| zhenguo | liusheng: the Keypaire | 08:09 |
| zhenguo | liusheng: I remember it's already fixed | 08:09 |
| liusheng | zhenguo: I am trying again the RTD doc with a pure Mogan with requiremnets added | 08:09 |
| zhenguo | liusheng: docs can be done later | 08:09 |
| zhenguo | liusheng: even if we cut release | 08:09 |
| liusheng | zhenguo: you mean the typo ? | 08:12 |
| zhenguo | liusheng: yes | 08:12 |
| zhenguo | liusheng: can be done with you patch | 08:12 |
| liusheng | zhenguo: https://review.openstack.org/#/c/475971/12/mogan/common/exception.py | 08:14 |
| zhenguo | liusheng: yes, that override it | 08:15 |
| liusheng | zhenguo: RTD doc still built faild | 08:15 |
| zhenguo | liusheng: we can fix that later | 08:16 |
| liusheng | zhenguo: ok | 08:16 |
| * zhenguo brb | 08:17 | |
| openstackgerrit | liusheng proposed openstack/mogan master: Fix a typo https://review.openstack.org/504879 | 08:19 |
| liusheng | zhenguo: Nova also didn't set the project_id when importing a keypair, and check the quota of a user, Nova have the concept of quota of user but we don't | 08:31 |
| liusheng | zhenguo: maybe a kaypair to be importted should belong to any project ? | 08:32 |
| zhenguo | liusheng: yes, but as we discussed before, just add a project id as a workaround | 08:32 |
| zhenguo | liusheng: we can introduce user based quota later | 08:33 |
| liusheng | zhenguo: or how about remove the constraint in db ? | 08:33 |
| zhenguo | liusheng: it's not because the db | 08:33 |
| zhenguo | liusheng: but quota | 08:33 |
| zhenguo | liusheng: we can't support user based quota now | 08:33 |
| zhenguo | liusheng: we discussed this with wanghao before | 08:34 |
| liusheng | zhenguo: no, we don't check quota for keypari importing | 08:34 |
| zhenguo | liusheng: but we check it for creating | 08:35 |
| liusheng | zhenguo: you mean quota will check all the items in db ? | 08:35 |
| zhenguo | liusheng: ? | 08:35 |
| liusheng | zhenguo: I mean we allow empty project_id | 08:36 |
| *** wanghao has quit IRC | 08:36 | |
| liusheng | zhenguo: that will broken quota check ? | 08:36 |
| zhenguo | liusheng: we don't allow empty project id | 08:36 |
| *** wanghao has joined #openstack-mogan | 08:36 | |
| zhenguo | liusheng: no | 08:36 |
| zhenguo | liusheng: you don't set project id there | 08:36 |
| liusheng | zhenguo: so why ? | 08:36 |
| zhenguo | liusheng: and the project id is introduced just for quota | 08:36 |
| liusheng | zhenguo: I mean we remove the constrant of nnullable=False | 08:36 |
| liusheng | zhenguo: for creating, it set project_id | 08:37 |
| liusheng | zhenguo: for importing, it don't set | 08:37 |
| zhenguo | liusheng: why not check quota when importing | 08:37 |
| liusheng | zhenguo: that is like what nova does | 08:37 |
| zhenguo | liusheng: it's another way of creating | 08:37 |
| liusheng | zhenguo: I juse said... | 08:37 |
| zhenguo | liusheng: nova donesn't have project id at all | 08:37 |
| liusheng | zhenguo: for importing, it use user to check quota, for creating, it use project | 08:38 |
| zhenguo | liusheng: nova? | 08:38 |
| liusheng | zhenguo: oh, sorry, both no project | 08:39 |
| zhenguo | liusheng: yes | 08:39 |
| zhenguo | liusheng: we can quick fix it by adding project id and quota reservation like create_key_pair | 08:40 |
| zhenguo | liusheng: until we improve the quota implementation | 08:40 |
| openstackgerrit | liusheng proposed openstack/mogan master: Set project_id when importing a existing keypair https://review.openstack.org/504882 | 08:42 |
| liusheng | zhenguo: ^ | 08:42 |
| zhenguo | liusheng: so quick, hah | 08:43 |
| openstackgerrit | liusheng proposed openstack/mogan master: Set project_id when importing an existing keypair https://review.openstack.org/504882 | 08:43 |
| liusheng | zhenguo: lol, a typo, ps2 comming | 08:43 |
| zhenguo | liusheng: ok | 08:43 |
| zhenguo | liusheng: it's appreciated if you can help to test all the functionalities again with the latest code, | 08:45 |
| zhenguo | liusheng: we can finish docs later | 08:45 |
| liusheng | zhenguo: back to RTD doc, still have same error even after I using a pure latest mogan with the requirements added :( | 08:45 |
| liusheng | zhenguo: ok | 08:45 |
| zhenguo | liusheng: we can talk about docs after tagging a release | 08:45 |
| liusheng | zhenguo: ok | 08:46 |
| liusheng | zhenguo: we treated reuild as a state than a action, looks a bit weird | 08:48 |
| zhenguo | liusheng: you mean provision state? | 08:49 |
| liusheng | zhenguo: yes | 08:49 |
| zhenguo | liusheng: we don't have action notion | 08:49 |
| zhenguo | liusheng: you PUT the provision state to rebuild | 08:50 |
| zhenguo | liusheng: just like you PUT the power state to on/off | 08:50 |
| liusheng | zhenguo: I know, just found in docs and API, it looks a bit weird | 08:50 |
| zhenguo | liusheng: you want to follow nova's way? | 08:50 |
| liusheng | zhenguo: not sure, just a thoughts, hah | 08:51 |
| liusheng | zhenguo: we study this from Ironic than Nova | 08:51 |
| zhenguo | liusheng: as nova's API defination looks not RESTful | 08:52 |
| liusheng | zhenguo: got it | 08:53 |
| liusheng | zhenguo: we don't support to get the access projects of flavor | 09:05 |
| zhenguo | liusheng: you mean on client side? | 09:06 |
| liusheng | zhenguo: yes | 09:06 |
| zhenguo | liusheng: oh | 09:06 |
| zhenguo | liusheng: but we already tagged the client release | 09:06 |
| liusheng | zhenguo: and do you think if we need a saperated api for the access ? | 09:06 |
| zhenguo | liusheng: seems we discussed this on Xinran's patch | 09:07 |
| liusheng | zhenguo: how about merging it with "flavor show" and | 09:07 |
| liusheng | zhenguo: oh, yes | 09:07 |
| zhenguo | liusheng: API side is ok, right? | 09:08 |
| liusheng | zhenguo: ye | 09:08 |
| liusheng | s | 09:08 |
| zhenguo | liusheng: that's ok | 09:08 |
| liusheng | zhenguo: for client, it only support set/unset the access projects, but no way to show | 09:09 |
| liusheng | zhenguo: maybe just add a field in flavor get API ? and keep current APIs | 09:10 |
| liusheng | zhenguo: and remove the FlavorAccessController.get_all api | 09:11 |
| liusheng | zhenguo: just keep FlavorAccessController.post and FlavorAccessController.delete APIs | 09:13 |
| zhenguo | liusheng: you mean no need to change client side | 09:15 |
| zhenguo | liusheng: that's a workaround | 09:16 |
| liusheng | zhenguo: yes | 09:16 |
| liusheng | zhenguo: not a workaround | 09:16 |
| zhenguo | liusheng: but in mogan, we also have other APIs like this | 09:16 |
| zhenguo | liusheng: like aggregate nodes | 09:17 |
| liusheng | zhenguo: yes | 09:17 |
| zhenguo | liusheng: we talked before, there clould be many nodes with the aggregates, so we don't return it with aggregates | 09:17 |
| liusheng | zhenguo: hmm, yes, but not sure if it is common case, Nova also show like that | 09:18 |
| zhenguo | liusheng: just keep the current way | 09:19 |
| liusheng | zhenguo: and seems there are also similar situation in openstack commands, the output of a show command including a list of stuff | 09:19 |
| zhenguo | liusheng: I don't see why that's better | 09:19 |
| liusheng | zhenguo: just thing a sapareted get_all for access projects is a bit redundant | 09:20 |
| zhenguo | liusheng: not redundant, it's a separted thing | 09:21 |
| liusheng | zhenguo: it is also looks like a property of a flavor | 09:21 |
| zhenguo | liusheng: it's hard to include all properties in one object | 09:22 |
| zhenguo | liusheng: servers which are created with the flavor, also looks like a propery of the flavor | 09:23 |
| liusheng | zhenguo: ok.. not a good example, hah | 09:23 |
| zhenguo | liusheng: lol | 09:24 |
| zhenguo | liusheng: it's ok to keep this way for now | 09:24 |
| liusheng | zhenguo: ok | 09:24 |
| zhenguo | liusheng: seems litao__ left some comments on the patch | 09:29 |
| liusheng | zhenguo: yes, I just replied | 09:29 |
| liusheng | zhenguo: may you can help to review again | 09:29 |
| zhenguo | litao__: ok | 09:30 |
| zhenguo | oh | 09:31 |
| zhenguo | liusheng: ok | 09:31 |
| zhenguo | litao__: hey, can you check the comments on liusheng's patch? we need to land that quick fix soon | 09:31 |
| litao__ | zhenguo: which patch? | 09:32 |
| zhenguo | litao__: keypairs | 09:32 |
| zhenguo | litao__: quota | 09:32 |
| openstackgerrit | Merged openstack/mogan master: Fix a typo https://review.openstack.org/504879 | 09:34 |
| litao__ | zhenguo: In the key is reserved, mogan can't reserve it again. | 09:35 |
| litao__ | zhenguo: I think this | 09:35 |
| zhenguo | litao__; you mean we should move quota commit before keypair.create? | 09:36 |
| *** wanghao has quit IRC | 09:38 | |
| *** wanghao has joined #openstack-mogan | 09:38 | |
| *** wanghao has quit IRC | 09:38 | |
| litao__ | zhenguo: You can check the quota check in server creation, when server.create() failed , use self.quota.rollback(context, reservations) to rollback | 09:39 |
| *** wanghao has joined #openstack-mogan | 09:39 | |
| zhenguo | litao__: that's because we do quota commit before object create | 09:39 |
| *** wanghao has quit IRC | 09:39 | |
| zhenguo | litao__: but for keypairs, seems we do really quota commit after object creation | 09:39 |
| *** wanghao has joined #openstack-mogan | 09:39 | |
| *** wanghao has quit IRC | 09:40 | |
| *** wanghao has joined #openstack-mogan | 09:40 | |
| litao__ | zhenguo: no, we do quota commit after object create | 09:41 |
| *** wanghao has quit IRC | 09:41 | |
| zhenguo | liusheng: for servers? | 09:41 |
| *** wanghao has joined #openstack-mogan | 09:41 | |
| *** wanghao has quit IRC | 09:42 | |
| litao__ | zhenguo: yes | 09:42 |
| *** wanghao has joined #openstack-mogan | 09:42 | |
| litao__ | zhenguo: you can check the code in _provision_servers method | 09:42 |
| *** wanghao has quit IRC | 09:42 | |
| *** wanghao has joined #openstack-mogan | 09:43 | |
| liusheng | litao__, zhenguo if we need to rollback, maybe need to fix in many places ? | 09:43 |
| zhenguo | litao__: oh, yes, understood | 09:43 |
| zhenguo | liusheng: seems yes | 09:44 |
| liusheng | litao__: thanks for clarify, got it | 09:44 |
| zhenguo | liusheng: so we need to either rollback or commit the reservations | 09:44 |
| litao__ | zhenguo,liusheng :np, Maybe update it in another patch. | 09:44 |
| zhenguo | litao__: yes | 09:45 |
| liusheng | zhenguo, litao__ yes, just want to say that | 09:45 |
| zhenguo | litao__: if you can remove the -1 there, lol | 09:45 |
| litao__ | zhenguo: OK | 09:45 |
| zhenguo | litao__: thanks | 09:45 |
| liusheng | lol | 09:45 |
| liusheng | zhenguo: another small bug, seems some API samples drafted with 2 spaces indent, and some others use 4 spaces indent | 09:46 |
| zhenguo | liusheng: we should try to clean up that | 09:46 |
| liusheng | zhenguo: yes | 09:47 |
| zhenguo | liusheng: yes, but seems docs are ok | 09:47 |
| liusheng | zhenguo: hmm, docs are consistent ? | 09:47 |
| zhenguo | liusheng: seems yes, | 09:47 |
| liusheng | zhenguo: ok | 09:47 |
| zhenguo | liusheng: but better to clean up that, maybe later | 09:48 |
| liusheng | zhenguo: ok | 09:48 |
| zhenguo | liusheng: you patch seems built failed | 09:48 |
| liusheng | zhenguo: seems devstack installed failed, not sure why | 09:49 |
| zhenguo | liusheng: just recheck for such failure | 09:49 |
| liusheng | zhenguo: ok | 09:50 |
| zhenguo | shaohe_feng: please have a look at this https://review.openstack.org/#/c/504882/ | 10:46 |
| openstackgerrit | Zhenguo Niu proposed openstack/mogan master: Avoid passing empty aggregates to Placement https://review.openstack.org/504914 | 11:21 |
| zhenguo | liusheng: please have a look at the patch above | 11:22 |
| liusheng | zhenguo: ok | 11:24 |
| zhenguo | liusheng: server group can only be supported if aggregates with affinity zone, right? | 11:25 |
| liusheng | zhenguo: seems yes | 11:26 |
| zhenguo | liusheng: but seems you got all aggregates | 11:27 |
| liusheng | zhenguo: let me check | 11:27 |
| zhenguo | liusheng: I correct that in my patch | 11:28 |
| liusheng | zhenguo: looking | 11:28 |
| zhenguo | liusheng: there are stil some problems with the scheuling | 11:35 |
| zhenguo | liusheng: for availabilty zones, aggregates, and affinity zones | 11:35 |
| liusheng | zhenguo: what's that ? | 11:37 |
| zhenguo | liusheng: needs more tests | 11:37 |
| liusheng | zhenguo: ok | 11:37 |
| zhenguo | liusheng: maybe better to cut release tomorrow | 11:39 |
| liusheng | zhenguo: I think it is ok | 11:39 |
| zhenguo | liusheng:ｈａｈ | 11:39 |
| zhenguo | liusheng: are you working on quota rollback? | 11:49 |
| liusheng | zhenguo: not yet | 11:50 |
| zhenguo | liusheng: ok | 11:50 |
| liusheng | zhenguo: need to land before release ? | 11:50 |
| zhenguo | liusheng: yes | 11:50 |
| zhenguo | liusheng: seems not many | 11:50 |
| liusheng | zhenguo: ok, let me try | 11:50 |
| zhenguo | liusheng: we only support quota for few resources | 11:51 |
| zhenguo | liusheng: ok thanks | 11:51 |
| liusheng | zhenguo: yes, np | 11:51 |
| *** litao__ has quit IRC | 12:04 | |
| -openstackstatus- NOTICE: Gerrit will be offline for the upgrade to 2.13 starting at 15:00 UTC (in roughly 3 hours) and is expected to probably be down/unusable for 8+ hours while an offline reindex is performed: http://lists.openstack.org/pipermail/openstack-dev/2017-August/120533.html | 12:06 | |
| openstackgerrit | liusheng proposed openstack/mogan master: Support quota rollback when creating/importing keypair failed https://review.openstack.org/504926 | 12:07 |
| zhenguo | liusheng: seems I got the reason why the reset nodes of our bare metal env don't work | 12:07 |
| zhenguo | liusheng: as switch config failed | 12:08 |
| liusheng | zhenguo: aha, you have fixed that ? | 12:08 |
| zhenguo | liusheng: I think so, will test | 12:08 |
| liusheng | zhenguo: ok | 12:09 |
| zhenguo | liusheng: it works, hah | 12:18 |
| liusheng | zhenguo: cool! | 12:18 |
| openstackgerrit | liusheng proposed openstack/mogan master: Normalize API samples json files https://review.openstack.org/504927 | 12:19 |
| -openstackstatus- NOTICE: Gerrit will be offline for the upgrade to 2.13 starting at 15:00 UTC (in roughly 1.5 hours) and is expected to probably be down/unusable for 8+ hours while an offline reindex is performed: http://lists.openstack.org/pipermail/openstack-dev/2017-August/120533.html | 13:37 | |
| -openstackstatus- NOTICE: Gerrit will be offline for the upgrade to 2.13 starting at 15:00 UTC (in roughly 30 minutes) and is expected to probably be down/unusable for 8+ hours while an offline reindex is performed: http://lists.openstack.org/pipermail/openstack-dev/2017-August/120533.html | 14:32 | |
| -openstackstatus- NOTICE: The Gerrit service at https://review.openstack.org/ is offline, upgrading to 2.13, for an indeterminate period of time hopefully not to exceed 23:59 UTC today: http://lists.openstack.org/pipermail/openstack-dev/2017-August/120533.html | 15:02 | |
| *** lin_yang has joined #openstack-mogan | 15:51 | |
| -openstackstatus- NOTICE: The Gerrit service at https://review.openstack.org/ is offline, upgrading to 2.13, for an indeterminate period of time hopefully not to exceed 23:59 UTC today: http://lists.openstack.org/pipermail/openstack-dev/2017-August/120533.html | 16:37 | |
| *** ChanServ changes topic to "The Gerrit service at https://review.openstack.org/ is offline, upgrading to 2.13, for an indeterminate period of time hopefully not to exceed 23:59 UTC today: http://lists.openstack.org/pipermail/openstack-dev/2017-August/120533.html" | 16:37 | |
| *** liusheng has quit IRC | 18:02 | |
| *** liusheng has joined #openstack-mogan | 18:03 | |
| *** wanghao_ has joined #openstack-mogan | 21:40 | |
| *** wanghao has quit IRC | 21:42 | |
| *** ChanServ changes topic to "Bugs: bugs.launchpad.net/mogan | Review: https://review.openstack.org/#/q/project:openstack/mogan,n,z" | 23:45 | |
| -openstackstatus- NOTICE: review.openstack.org Gerrit 2.13 upgrade is functionally complete. The Infra team will be cleaning up bookkeeping items over the next couple days. If you have any questions please let us know | 23:45 | |
Generated by irclog2html.py 2.15.3 by Marius Gedminas - find it at mg.pov.lt!