| zhenguo | morning mogan! | 00:56 |
|---|---|---|
| *** litao__ has joined #openstack-mogan | 01:09 | |
| *** openstackgerrit has joined #openstack-mogan | 01:10 | |
| openstackgerrit | Zhenguo Niu proposed openstack/mogan master: Add checks for aggregate availability_zone https://review.openstack.org/489876 | 01:10 |
| openstackgerrit | Zhenguo Niu proposed openstack/mogan master: Move baremetal dir out from engine https://review.openstack.org/492084 | 01:12 |
| *** openstackgerrit has quit IRC | 02:21 | |
| *** wanghao has quit IRC | 02:27 | |
| *** wanghao has joined #openstack-mogan | 02:28 | |
| *** openstackgerrit has joined #openstack-mogan | 02:36 | |
| openstackgerrit | liusheng proposed openstack/python-moganclient master: Add node aggregate commands support https://review.openstack.org/492002 | 02:36 |
| openstackgerrit | liusheng proposed openstack/python-moganclient master: Add commands for aggregate node actions https://review.openstack.org/492370 | 02:36 |
| openstackgerrit | liusheng proposed openstack/python-moganclient master: Add commands for aggregate node actions https://review.openstack.org/492370 | 02:37 |
| openstackgerrit | liusheng proposed openstack/mogan master: Add the missed quota object importing https://review.openstack.org/492024 | 02:40 |
| openstackgerrit | liusheng proposed openstack/python-moganclient master: Improve flavor set command to support updating flavor https://review.openstack.org/486865 | 02:55 |
| openstackgerrit | liusheng proposed openstack/python-moganclient master: Add node aggregate commands support https://review.openstack.org/492002 | 03:07 |
| openstackgerrit | liusheng proposed openstack/python-moganclient master: Add commands for aggregate node actions https://review.openstack.org/492370 | 03:07 |
| openstackgerrit | liusheng proposed openstack/python-moganclient master: Remove some useless code of flavor unset command https://review.openstack.org/492785 | 03:10 |
| openstackgerrit | liusheng proposed openstack/python-moganclient master: Add node aggregate commands support https://review.openstack.org/492002 | 03:17 |
| openstackgerrit | liusheng proposed openstack/python-moganclient master: Add commands for aggregate node actions https://review.openstack.org/492370 | 03:17 |
| openstackgerrit | Zhenguo Niu proposed openstack/mogan master: Make disabled flavors not available for common users https://review.openstack.org/492488 | 03:23 |
| zhenguo | shaohe_feng: please have a look at this https://review.openstack.org/#/c/492366/ when you got time, thanks! | 03:24 |
| *** fredli__ has joined #openstack-mogan | 03:56 | |
| openstackgerrit | Zhenguo Niu proposed openstack/mogan master: Move baremetal dir out from engine https://review.openstack.org/492084 | 04:55 |
| openstackgerrit | Zhenguo Niu proposed openstack/mogan master: Add aggregates filters when do scheduling https://review.openstack.org/491695 | 05:25 |
| *** fredli__ has quit IRC | 06:05 | |
| openstackgerrit | wanghao proposed openstack/mogan master: Manage existing BMs: Part-1 https://review.openstack.org/479660 | 06:13 |
| openstackgerrit | wanghao proposed openstack/mogan master: Introduce Cinder client into Mogan https://review.openstack.org/489455 | 06:17 |
| Xinran | zhenguo, hi, I gave the reply on your comments of rebuid patch https://review.openstack.org/#/c/490421/ Pls check it if you got time | 06:23 |
| zhenguo | Xinran: ok | 06:35 |
| zhenguo | Xinran: wrt the roll-back, you mean when setting provision state failed? | 06:41 |
| Xinran | zhenguo, yes when rebuild failed, will reset the image_uuid | 06:42 |
| zhenguo | Xinran: when rebuild failed the server should be moved to error state | 06:42 |
| zhenguo | Xinran: and you may already change the ironic node image to the new one | 06:43 |
| Xinran | zhenguo, I thind ironic has a roll-back when node redeployement failed. | 06:53 |
| zhenguo | Xinran: seems not | 06:57 |
| Xinran | zhenguo, so we should reset the ironic node image manually in mogan ? | 07:02 |
| zhenguo | Xinran: I dont' think so, if the rebuild failed, I don't think we can roll back it | 07:07 |
| Xinran | zhenguo, yes, I didn't find an ironic driver to set ironic node image ... | 07:11 |
| Xinran | zhenguo, so we just move server to error state? | 07:12 |
| zhenguo | Xinran: for now, I think it's ok | 07:12 |
| Xinran | zhenguo, ok | 07:12 |
| Xinran | zhenguo, btw I didn't pass image_uuid to ironic driver, just passed it to engine.manage, because there is a try/except :) | 07:14 |
| zhenguo | Xinran: ok, I see | 07:16 |
| zhenguo | Xinran: but do we need to set the image_uuid back if there's an exception? | 07:16 |
| openstackgerrit | Merged openstack/mogan master: Replace default_availability_zone with default_schedule_zone https://review.openstack.org/492366 | 07:19 |
| openstackgerrit | Zhenguo Niu proposed openstack/mogan master: Add aggregates filters when do scheduling https://review.openstack.org/491695 | 07:24 |
| Xinran | zhenguo, yes I think so | 07:27 |
| Xinran | zhenguo, But as you mentioned, the error is occured when rebuilding with new image, well... seems don't set back is more reasonable ... | 07:31 |
| Xinran | zhenguo, hah | 07:31 |
| zhenguo | Xinran: hah | 07:50 |
| zhenguo | liusheng: please go through the patches :D | 07:52 |
| liusheng | zhenguo: ok | 07:52 |
| zhenguo | liusheng: seems there's only one patch left for aggregates, and the server groups api has landed | 07:53 |
| *** openstackgerrit has quit IRC | 08:02 | |
| wanghao | zhenguo: ping | 08:36 |
| zhenguo | wanghao: pong | 08:36 |
| wanghao | zhenguo: I found we miss some log in some operations like delete server, do you think we should improve them now? | 08:36 |
| zhenguo | wanghao: sure :D | 08:37 |
| wanghao | zhenguo: sure, quick fix. | 08:37 |
| zhenguo | wanghao: thanks | 08:37 |
| wanghao | zhenguo: np : ) | 08:37 |
| *** wanghao has quit IRC | 08:40 | |
| *** wanghao has joined #openstack-mogan | 08:41 | |
| liusheng | zhenguo: do you think shold we set the flavor as disabled when deleting a server which still in use by some servers ? | 08:44 |
| *** wanghao_ has joined #openstack-mogan | 08:44 | |
| liusheng | zhenguo: s/deleting a server/deleting a flavor | 08:44 |
| zhenguo | liusheng: seems not | 08:45 |
| zhenguo | liusheng: only admin can do delete or set disabled | 08:45 |
| zhenguo | liusheng: but we should raise an error | 08:46 |
| liusheng | zhenguo: I mean set the flavor automatically | 08:46 |
| liusheng | zhenguo: just like a flavor in deleted state | 08:46 |
| liusheng | zhenguo: but unvisionable for common users | 08:46 |
| liusheng | zhenguo: invisible | 08:47 |
| *** wanghao has quit IRC | 08:48 | |
| zhenguo | liusheng: we can just raise an error when deleting a flavor in use | 08:49 |
| zhenguo | liusheng: not usre if admins always hope it get disabled when delete a flavor in use | 08:50 |
| liusheng | zhenguo: only the admin can delete flavor | 08:50 |
| zhenguo | liusheng: maybe they just want to try to delete it | 08:50 |
| zhenguo | liusheng: as we provide two APIs, they can choose delete or disabled, | 08:51 |
| liusheng | zhenguo: even for admin, it shouldn't delete the flavor in use I think | 08:51 |
| zhenguo | liusheng: yes, so in this case, they should disable it | 08:52 |
| zhenguo | liusheng: by themselves instead of doing automatically by us | 08:52 |
| liusheng | zhenguo: .. | 08:52 |
| liusheng | zhenguo: for common users, we can treat the disabled flavor as deleted flavor | 08:53 |
| liusheng | zhenguo: only for internal usage | 08:53 |
| liusheng | zhenguo: just personal thoughs, but maybe just raise an error to remind admins to disable is also a good choice | 08:55 |
| zhenguo | liusheng: hah | 08:59 |
| zhenguo | liusheng: we still don't raise error now when deleting flavors in use | 09:00 |
| liusheng | zhenguo: I prefer to my suggestion :(, I am a admin, so I know I am deleting a flavor, but if the flavor is inuse, I need to disable it. why don't merge these two steps to one ? | 09:01 |
| zhenguo | liusheng: but there maybe other admins like me, I dont' want to be automatically get deleted :D | 09:01 |
| liusheng | zhenguo: why ? | 09:02 |
| zhenguo | liusheng: maybe I just want to try to delete it to see if it's in use | 09:02 |
| liusheng | zhenguo: ..alright, that is a usecase | 09:03 |
| zhenguo | liusheng: if it's in use, maybe I will keep it for a while, or maybe I will disable it | 09:03 |
| zhenguo | liusheng: automatically dialbled seems a hardcoding | 09:03 |
| liusheng | zhenguo: we already support listing servers filtered by flavor uuid | 09:04 |
| zhenguo | liusheng: yes, | 09:04 |
| liusheng | zhenguo: we can treat it as a integrated deleting process | 09:04 |
| liusheng | zhenguo: so there isnt any hardcoding matter | 09:05 |
| zhenguo | liusheng: it's not a must for us | 09:05 |
| zhenguo | liusheng: I perfer to keep it as is until customers request us to change | 09:06 |
| liusheng | zhenguo: ok | 09:06 |
| *** wanghao has joined #openstack-mogan | 09:09 | |
| zhenguo | liusheng: do you think admins should be able to use disabled flavor to create a server? | 09:09 |
| liusheng | zhenguo: I don't think so | 09:10 |
| zhenguo | liusheng: so we should need to check the flavor.disabled when creating servers | 09:10 |
| zhenguo | liusheng: can't leverage the db flavor_get api | 09:10 |
| liusheng | zhenguo: you mean your new flavor patche ? | 09:11 |
| zhenguo | liusheng: yes | 09:11 |
| liusheng | zhenguo: why cannot leverage the de flavor get api ? | 09:11 |
| *** wanghao_ has quit IRC | 09:12 | |
| zhenguo | liusheng: for common users when get the flavor, it will raise not found | 09:12 |
| liusheng | zhenguo: yes | 09:12 |
| zhenguo | liusheng: as in db flavor get api it will check whether it's admin | 09:12 |
| zhenguo | liusheng: so I check the disabled there and raise disabled error to just admins | 09:12 |
| *** openstackgerrit has joined #openstack-mogan | 09:13 | |
| openstackgerrit | wanghao proposed openstack/mogan master: Add some missing logs in Mogan https://review.openstack.org/492917 | 09:13 |
| liusheng | zhenguo: but disabled error of flavor will make common users recognise wichi flavor is disabled | 09:15 |
| zhenguo | liusheng: I will make a change to move admin check to api layer | 09:16 |
| zhenguo | liusheng: then when get flavor before we check the flavor disabled it will raise not found for common useres | 09:16 |
| zhenguo | liusheng: only admin can get that disabled flavor, then raise disabled to common users | 09:17 |
| zhenguo | s/common users/admins | 09:17 |
| liusheng | zhenguo: why don' raise notfound error also ? | 09:18 |
| zhenguo | liusheng: as admins can get that flavor | 09:18 |
| zhenguo | liusheng: you are an admin, you list the flavor, then create a server with it but got not found error, | 09:19 |
| zhenguo | liusheng: then you check the flavor again, and it's there, | 09:19 |
| liusheng | zhenguo: so I am tinking another thing, should we don't display the disabled flavor defaulty for admin, but can support a --deleted option ? | 09:20 |
| liusheng | zhenguo: like Nova does ? | 09:20 |
| zhenguo | liusheng: I don't think so, don't you think it makes things more complex | 09:21 |
| zhenguo | liusheng: why do you think we can't raise the flavor is disabled to admins when creating servers | 09:21 |
| liusheng | zhenguo: if we have many flavors in an env, and some of them are disabled, I am an admin user and want to select a flavor to use, but there are many flavors cannot be sued | 09:22 |
| zhenguo | liusheng: I understand we don't want to expose that to common users, but for admins it's ok | 09:22 |
| zhenguo | liusheng: in fact, admins should not create servers | 09:22 |
| liusheng | zhenguo: why ? | 09:22 |
| zhenguo | liusheng: you even can't create servers in admin's panel on horizon | 09:23 |
| *** wanghao has quit IRC | 09:24 | |
| zhenguo | liusheng: let's go back to the question of not displaying disabled flavors default by admins, but they can also get the flavor | 09:25 |
| zhenguo | liusheng: it's not deleted, they can make it enable again | 09:25 |
| *** wanghao has joined #openstack-mogan | 09:26 | |
| zhenguo | liusheng: see my replies here https://review.openstack.org/#/c/491695/ | 09:32 |
| liusheng | zhenguo: ok | 09:33 |
| zhenguo | liusheng: we can't manage to make one request for our aggregates | 09:34 |
| zhenguo | liusheng: as every key may match with a list of aggregates which is a request with member_of the aggregates list | 09:35 |
| liusheng | zhenguo: I got it. if so, shouldn't we remove the duplicated aggregates ? | 09:36 |
| zhenguo | liusheng: how? | 09:37 |
| zhenguo | liusheng: it's just a sql filter in placement | 09:37 |
| zhenguo | liusheng: the cost there is lower then we make many loops to remove it | 09:38 |
| liusheng | zhenguo: I mean the agg filters maybe: [[agg1], [agg1, agg2], [agg2, agg3]] | 09:38 |
| zhenguo | liusheng: yes, it's ok I think | 09:39 |
| zhenguo | liusheng: but we can remove the things like [[agg1, agg2], [agg1, agg2]] | 09:39 |
| liusheng | zhenguo: got it, seems yes | 09:42 |
| zhenguo | liusheng: and if that happens, that's because the operators defined a bad aggregate | 09:43 |
| liusheng | zhenguo: looks good to me, do do you think it is valuable to add the above cleanup ? | 09:50 |
| zhenguo | liusheng: yes, | 09:51 |
| zhenguo | liusheng: or maybe later | 09:51 |
| liusheng | zhenguo: ok, wait for your new patchset | 09:51 |
| zhenguo | liusheng: I hope we can make aggregates done by this week :D | 09:52 |
| liusheng | zhenguo: ok | 09:52 |
| zhenguo | liusheng: thanks, a following up patch will up when we finished the urgent work | 09:53 |
| liusheng | zhenguo: np, ok | 09:53 |
| zhenguo | liusheng: maybe it's time for you to rebase your scheduler_hints patch :D | 09:54 |
| liusheng | zhenguo: ok, will do it later :) | 09:54 |
| zhenguo | liusheng: ok | 09:54 |
| liusheng | zhenguo: seems gate has been broken.. | 09:54 |
| liusheng | zhenguo: no unstable | 09:54 |
| liusheng | zhenguo: or unstable | 09:55 |
| zhenguo | liusheng: seems not related to us | 09:55 |
| zhenguo | liusheng: because it's release period, they all quickly +A patches :D | 09:55 |
| liusheng | zhenguo: hah, it should be | 09:55 |
| zhenguo | liusheng: lol | 09:56 |
| *** wanghao has quit IRC | 10:23 | |
| openstackgerrit | Merged openstack/mogan master: Add the missed quota object importing https://review.openstack.org/492024 | 10:30 |
| openstackgerrit | Merged openstack/mogan master: Add aggregates filters when do scheduling https://review.openstack.org/491695 | 10:40 |
| openstackgerrit | Merged openstack/mogan master: Add checks for aggregate availability_zone https://review.openstack.org/489876 | 11:32 |
| *** zhuli has quit IRC | 12:19 | |
| *** zhuli has joined #openstack-mogan | 12:19 | |
| openstackgerrit | Merged openstack/mogan master: Do not allow update flavor internal attributes https://review.openstack.org/492438 | 12:41 |
| openstackgerrit | Merged openstack/mogan master: Move baremetal dir out from engine https://review.openstack.org/492084 | 13:46 |
Generated by irclog2html.py 2.15.3 by Marius Gedminas - find it at mg.pov.lt!