| *** zhangshuai has joined #openstack-smaug | 01:24 | |
| *** pzchen has joined #openstack-smaug | 01:31 | |
| *** pzchen has quit IRC | 01:53 | |
| *** pzchen has joined #openstack-smaug | 01:54 | |
| *** chenhuayi has joined #openstack-smaug | 02:07 | |
| *** chenhuayi has quit IRC | 02:07 | |
| *** chenhuayi has joined #openstack-smaug | 02:08 | |
| openstackgerrit | chenying proposed openstack/smaug: Fix the format of restores parameters https://review.openstack.org/320195 | 03:13 |
|---|---|---|
| yinwei_computer | ping chenying | 03:21 |
| chenying | Hi | 03:21 |
| yinwei_computer | this patch is to support parameters in protection service?\ | 03:21 |
| chenying | No this patch just fix the data stored in db. | 03:21 |
| yinwei_computer | ok | 03:22 |
| yinwei_computer | got it. | 03:22 |
| yinwei_computer | then I will have wangliuan to fix protection service | 03:22 |
| *** pzchen has quit IRC | 03:23 | |
| chenying | xiangxinyong have submitted the patch about supporting parameters in horizon. | 03:24 |
| *** zhangshuai_ has joined #openstack-smaug | 03:38 | |
| *** zhangshuai has quit IRC | 03:42 | |
| *** yuval has joined #openstack-smaug | 03:42 | |
| *** yuval has left #openstack-smaug | 03:42 | |
| *** zhangshuai_ has quit IRC | 03:47 | |
| *** zhangshuai_ has joined #openstack-smaug | 03:47 | |
| xiangxinyong | chenying is right .https://review.openstack.org/#/c/318396/ | 03:50 |
| xiangxinyong | This patch is about the parameters | 03:50 |
| *** zhangshuai_ has quit IRC | 03:58 | |
| *** zhangshuai_ has joined #openstack-smaug | 04:02 | |
| *** zhangshuai_ has quit IRC | 04:24 | |
| *** zhangshuai_ has joined #openstack-smaug | 04:24 | |
| *** zhangshuai_ has quit IRC | 04:35 | |
| *** chenying has quit IRC | 05:37 | |
| *** chenying has joined #openstack-smaug | 05:38 | |
| yinwei_computer | ping saggi | 06:15 |
| *** pzchen has joined #openstack-smaug | 06:42 | |
| chenying | ping gampel | 06:43 |
| *** yuval has joined #openstack-smaug | 06:44 | |
| yinwei_computer | hi, yuval | 06:47 |
| yuval | yinwei_computer: hey, how are you? | 06:48 |
| yinwei_computer | good | 06:48 |
| yinwei_computer | how are you doing? | 06:48 |
| yuval | well :) | 06:49 |
| yinwei_computer | hmm, will Saggi come today? | 06:49 |
| yuval | yes, I guess he will be here soon | 06:50 |
| yinwei_computer | I have some problem about this merged patch | 06:50 |
| yinwei_computer | https://review.openstack.org/#/c/319672/ | 06:50 |
| yinwei_computer | if he's not there, could you pls. check it with me? | 06:51 |
| yuval | sure | 06:51 |
| yuval | what is the issue? | 06:51 |
| yinwei_computer | I see the point this patch wants to address | 06:51 |
| yinwei_computer | but the issue it brings is, there's nothing to indicate the restore status | 06:52 |
| yinwei_computer | do you guys already get a plan to fix this issue? | 06:52 |
| yuval | there is the restore object and api | 06:54 |
| yuval | yinwei_computer: ? | 06:56 |
| yinwei_computer | ? | 06:57 |
| yinwei_computer | sorry, I don't get it | 06:57 |
| yinwei_computer | I mean the status to restore the whole checkpoint | 06:58 |
| yuval | a Checkpoint can not hold the restore status, as it can be restored more than once, in the same time | 06:59 |
| yuval | (for example) | 06:59 |
| yinwei_computer | user who is using smaug to restore, it's not reasonable to have user to check resource restore status one by one | 06:59 |
| yinwei_computer | I see the point | 06:59 |
| yinwei_computer | but the requirement is very reasonable | 06:59 |
| yinwei_computer | user needs check the whole restoration status | 06:59 |
| yuval | not sure I understand | 07:00 |
| yuval | the Restore object is not per-resource | 07:00 |
| yinwei_computer | you mean the restore object? | 07:00 |
| yuval | Protection Service have a /<project_id>/restore REST API route | 07:00 |
| yuval | yes | 07:00 |
| yuval | which starts a restore and returns a Restore object | 07:01 |
| yinwei_computer | hmm, actually I didn't see any restore object in API | 07:01 |
| yuval | would you mind taking a look again? I missed it at first too | 07:01 |
| yinwei_computer | let's confirm it with chenying | 07:02 |
| yinwei_computer | we need involve him | 07:02 |
| yinwei_computer | ping chenying | 07:02 |
| chenying | Hi | 07:02 |
| yinwei_computer | Yuval suggested we store restore status in restore object | 07:03 |
| yinwei_computer | I'm not sure if we already got this restore object model in API layer | 07:03 |
| chenying | Yes we have status field in restores vo. | 07:04 |
| yuval | I'll be right back (~10 minutes) | 07:05 |
| yinwei_computer | but restore object will get passed to protection service in restore rpc request? | 07:05 |
| chenying | But status is not updated by protect service. It only mark the rpcapi successful called. | 07:06 |
| chenying | The whole restores vo have been passed to protection service. | 07:07 |
| yinwei_computer | what does 'restores vo' mean? | 07:08 |
| chenying | restores version object. | 07:08 |
| yinwei_computer | ok, got it | 07:08 |
| yinwei_computer | so it seems that we need fix this bug | 07:08 |
| yinwei_computer | to have protection service to return status of restore object? | 07:09 |
| chenying | version object have its own curd interface. | 07:09 |
| chenying | to have protection service to return status of restore object? I think so | 07:09 |
| yinwei_computer | but there's a problem to update the restore status from restoring to restored | 07:10 |
| yinwei_computer | rpc request will return restore vo as restoring if succeeds | 07:10 |
| yinwei_computer | later when protection service finds the heat stack has finished, it needs update the status to restored. | 07:11 |
| yinwei_computer | currently we have no way to notify API from protection service. | 07:11 |
| yinwei_computer | what do you think? | 07:11 |
| chenying | I think we don't need to notify API to update. | 07:12 |
| chenying | call restore(vo).save() to updata status. | 07:13 |
| chenying | notify API to update mean we need call smaugclient in protect service. | 07:13 |
| chenying | It make this update more complex. | 07:14 |
| yinwei_computer | the question is if there's no notification, how to get restore vo status updated? | 07:14 |
| chenying | How about saving restore vo in contect? Now Does one contect mean one rpc call? | 07:17 |
| yinwei_computer | ping Saggi | 07:25 |
| yuval | Not sure what is your issue with the Restore API/RPC | 07:25 |
| yinwei_computer | we need update restore version object from restoring to restored when restore finihsed | 07:26 |
| yinwei_computer | it's not as simple as just return restore obj in rpc call | 07:26 |
| yuval | the rpc call should return the restore object when the restore has begun | 07:27 |
| yinwei_computer | this is np | 07:27 |
| yinwei_computer | but who will update the restore object status to API service? | 07:27 |
| yinwei_computer | and how? | 07:27 |
| yuval | about updating the restore vo: why not do that in create_restoration.py | 07:27 |
| yinwei_computer | the question is how | 07:28 |
| yinwei_computer | using smaug client in protection service? | 07:28 |
| yinwei_computer | or using version object API? | 07:28 |
| yuval | why not version object api? | 07:28 |
| yinwei_computer | chenying could tell you more | 07:28 |
| yinwei_computer | my point is that this patch is not an integrated one | 07:29 |
| yinwei_computer | it addresses one issue, but it takes another issue | 07:29 |
| yinwei_computer | since chenying is working on integration test, I suggest we'd better revert this patch for now and report a bug to track the multiple restoration issue. | 07:30 |
| chenying | yuval: using smaugclient or version object save API are both OK. | 07:30 |
| yuval | chenying can revert this change locally in his environment | 07:30 |
| yuval | no need to revert on master | 07:30 |
| yinwei_computer | better revert on master | 07:31 |
| chenying | my environment do not include this patch. | 07:31 |
| yinwei_computer | we need a CI to guard correctness | 07:31 |
| yinwei_computer | next time to commit this patch, we need check if it comes together with restore object status update | 07:32 |
| yinwei_computer | otherwise, we need refuse patches breaking functionality | 07:32 |
| yuval | yinwei_computer: this issue should have been dealt with in this patch: https://review.openstack.org/306311 | 07:33 |
| yinwei_computer | no | 07:33 |
| yinwei_computer | this patch updated restore status to checkpoint | 07:34 |
| yinwei_computer | saggi erased restore status from checkpoint | 07:34 |
| yinwei_computer | I understand his point. But his patch will break correctness of restoration. Or he could fix his patch with updating restore obj status. | 07:36 |
| yuval | ping saggi | 07:36 |
| yinwei_computer | 'revert' means postpone to merge it after CI is ready | 07:37 |
| yinwei_computer | let's check it when saggi comes. | 07:39 |
| *** sho__ has joined #openstack-smaug | 07:48 | |
| *** sho__ has left #openstack-smaug | 07:50 | |
| zhongjun_ | Does anyone could tell me how to install smaug? thanks | 07:57 |
| saggi | ping yinwei_computer | 08:02 |
| chenzeng | zhongjun:If you install smaug using devstack, just copy the local.cnf from smaug's repository to the devstack dir, and run stack.sh. Maybe you should update the local.cnf as you need. | 08:03 |
| zhongjun_ | saggi, chenzeng: ok, I will try it. thanks | 08:12 |
| *** pzchen has quit IRC | 08:51 | |
| *** pzchen has joined #openstack-smaug | 08:51 | |
| *** pzchen has quit IRC | 08:53 | |
| yinwei_computer | ping saggi | 09:07 |
| yuval | yinwei_computer: he will be back later | 09:08 |
| yinwei_computer | ok | 09:08 |
| yinwei_computer | I'll go for supper | 09:33 |
| *** zhongjun_ has quit IRC | 09:41 | |
| *** zhongjun_ has joined #openstack-smaug | 09:41 | |
| *** wangfeng has quit IRC | 09:47 | |
| *** zhongjun_ has quit IRC | 09:50 | |
| *** zhongjun_ has joined #openstack-smaug | 09:51 | |
| saggi | yinwei_computer: pong | 09:59 |
| *** wangfeng has joined #openstack-smaug | 10:02 | |
| *** zhongjun_ has quit IRC | 10:18 | |
| *** zhongjun_ has joined #openstack-smaug | 10:19 | |
| openstackgerrit | Merged openstack/smaug: Fix minor issues with code formatting https://review.openstack.org/319673 | 12:17 |
| openstackgerrit | Merged openstack/smaug: Add operation log endpoints to API document https://review.openstack.org/298060 | 12:18 |
| openstackgerrit | Merged openstack/smaug: [API] Add resource endpoint https://review.openstack.org/301777 | 12:18 |
| xiangxinyong | hi guys, I wrote a smaug introduce in the blog. http://www.cnblogs.com/openstackteam/p/5524661.html | 12:33 |
| xiangxinyong | welcome to visit it. | 12:33 |
| gampel | xiangxinyong:: very nice | 12:34 |
| xiangxinyong | gampel: thanks. :) | 12:35 |
| openstackgerrit | Yuval Brik proposed openstack/smaug: Reimplement the Protection workflow for new design https://review.openstack.org/310385 | 12:45 |
| openstackgerrit | Merged openstack/smaug: Fix the format of restores parameters https://review.openstack.org/320195 | 12:46 |
| openstackgerrit | Yuval Brik proposed openstack/smaug: Reimplement the Protection workflow for new design https://review.openstack.org/310385 | 13:17 |
| openstackgerrit | OpenStack Proposal Bot proposed openstack/smaug: Updated from global requirements https://review.openstack.org/319891 | 13:52 |
| *** chenhuayi has quit IRC | 13:53 | |
| *** zhongjun_ has quit IRC | 14:29 | |
| *** yuval has quit IRC | 14:42 | |
| *** yuval has joined #openstack-smaug | 15:20 | |
| *** yuval has quit IRC | 15:35 | |
| *** saggi has quit IRC | 16:02 | |
| *** xavierr has joined #openstack-smaug | 21:50 | |
| *** xavierr has left #openstack-smaug | 22:06 | |
| *** openstackgerrit has quit IRC | 23:03 | |
| *** openstackgerrit has joined #openstack-smaug | 23:03 | |
Generated by irclog2html.py 2.14.0 by Marius Gedminas - find it at mg.pov.lt!