chenzeng | gampel:if you online, please let me know. Sorry, i missed yesterday's meeting. | 01:10 |
---|---|---|
zengyingzhe | He'll online afternoon. | 01:12 |
xiangxinyong | chenying:ping | 01:26 |
yinweiishere | I missed the meeting too | 01:56 |
yinweiishere | yesterday was women's day so I went to the cinema ... | 01:56 |
*** mingyu has joined #openstack-smaug | 02:14 | |
openstackgerrit | zengchen proposed openstack/smaug: Implement rpc interfaces of OperationEngine https://review.openstack.org/276081 | 02:16 |
*** chenying has joined #openstack-smaug | 02:17 | |
openstackgerrit | WeAreFormalGroup proposed openstack/smaug: Implement cinder protection plugin https://review.openstack.org/286458 | 02:21 |
*** smile-luobin has joined #openstack-smaug | 02:31 | |
*** luobin has quit IRC | 02:34 | |
openstackgerrit | xiangxinyong proposed openstack/smaug-dashboard: Add skeleton for the smaug-dashboard https://review.openstack.org/289724 | 02:49 |
*** smile-luobin has quit IRC | 02:50 | |
*** smile-luobin has joined #openstack-smaug | 02:51 | |
openstackgerrit | zengchen proposed openstack/smaug: Implement RestAPIs of trigger. https://review.openstack.org/286406 | 02:55 |
*** mingyu has quit IRC | 03:27 | |
*** mingyu has joined #openstack-smaug | 03:38 | |
*** mingyu has quit IRC | 03:50 | |
*** mingyu has joined #openstack-smaug | 04:28 | |
*** mingyu has quit IRC | 05:28 | |
*** mingyu has joined #openstack-smaug | 05:43 | |
*** mingyu has quit IRC | 06:24 | |
*** mingyu has joined #openstack-smaug | 06:34 | |
*** mingyu has quit IRC | 06:36 | |
*** mingyu has joined #openstack-smaug | 06:52 | |
*** smile-luobin has quit IRC | 07:09 | |
*** smile-luobin has joined #openstack-smaug | 07:14 | |
*** gampel has joined #openstack-smaug | 07:21 | |
zengyingzhe | ping xiangxinyong | 07:23 |
xiangxinyong | hi | 07:24 |
yinwei_computer | yingzhe, sorry, I still have problem of your protectable layout | 07:25 |
yinwei_computer | pls. check how to get child image resources of a server through your layout | 07:25 |
yinwei_computer | ping gampel | 07:26 |
zengyingzhe | yinwei_computer, OK, we can talk about it later. I got a issure to confirm with xiangxinyong | 07:26 |
zengyingzhe | sorry | 07:26 |
yinwei_computer | np | 07:26 |
xiangxinyong | zengyingzhe: what's wrong? | 07:27 |
*** mingyu has quit IRC | 07:27 | |
zengyingzhe | xiangxinyong, the protection plugin paramters, UI is asking user to fill for each resource type or each resource instance? | 07:28 |
yinwei_computer | ping saggi | 07:28 |
*** mingyu has joined #openstack-smaug | 07:29 | |
zengyingzhe | I change an expression. UI support that user can specify parameters for each resource instance? | 07:32 |
xiangxinyong | zengyingzhe: right | 07:33 |
zengyingzhe | or the instances of one the same type have the same parameters? | 07:33 |
xiangxinyong | zengyingzhe: so far, we agree this point(this instances of one the same type have the same parameters) | 07:34 |
xiangxinyong | zengyingzhe: sorry i can not send a email of the fake ui to you | 07:35 |
xiangxinyong | but you can glance my lastest email about the smaug ui. | 07:36 |
zengyingzhe | xiangxinyong, that's OK. So the parameters are specified for resource type? not resource instance? | 07:37 |
*** smile-luobin has quit IRC | 07:44 | |
*** smile-luobin has joined #openstack-smaug | 07:45 | |
yinwei_computer | ping gampel, pls. check the following two patches: https://review.openstack.org/#/c/287572/, https://review.openstack.org/#/c/285149/ | 07:46 |
yinwei_computer | ping saggi, pls. add a pari of interfaces in checkpoint class to write and read resource definitions, then we can call them in protection plugin | 07:47 |
yinwei_computer | a pair, sorry | 07:47 |
openstackgerrit | Merged openstack/smaug: The RESTAPI of resource checkpoints https://review.openstack.org/286412 | 08:05 |
openstackgerrit | zengchen proposed openstack/smaug: RPC client recreates exception object failed https://review.openstack.org/289685 | 08:12 |
openstackgerrit | zengchen proposed openstack/smaug: Implement RestAPIs of Scheduled Operation. https://review.openstack.org/287036 | 08:12 |
gampel | hi smile-luobin | 08:13 |
gampel | hi yinwei_computer: : | 08:13 |
gampel | smile-luobin: Please rebase on [1] https://review.openstack.org/#/c/285816/ and we can merge | 08:14 |
openstackgerrit | zengchen proposed openstack/smaug: implement time trigger https://review.openstack.org/271196 | 08:16 |
xiangxinyong | zengyinzhe: yeah. we can confirm with eran. | 08:16 |
smile-luobin | OK | 08:16 |
xiangxinyong | eran: hi | 08:17 |
yinwei_computer | integrate smaug to hybrid cloud DR | 08:21 |
yinwei_computer | integrate smaug to hybrid cloud DR | 08:21 |
yinwei_computer | ping gampel, pls. check the following two patches: https://review.openstack.org/#/c/287572/, https://review.openstack.org/#/c/285149/ | 08:21 |
openstackgerrit | zengchen proposed openstack/smaug: Implement operation manager https://review.openstack.org/274339 | 08:22 |
openstackgerrit | zengchen proposed openstack/smaug: Implement executor of OperationEngine https://review.openstack.org/282263 | 08:22 |
openstackgerrit | zengchen proposed openstack/smaug: Implement RestAPIs of trigger. https://review.openstack.org/286406 | 08:22 |
openstackgerrit | zengchen proposed openstack/smaug: Implement rpc interfaces of OperationEngine https://review.openstack.org/276081 | 08:22 |
zengyingzhe | xiangxinyong, I've check the UI design you sent before. The parameters could be specified for each instance. | 08:22 |
zengyingzhe | So no problem anymore. | 08:23 |
*** smile-luobin has quit IRC | 08:32 | |
*** smile-luobin has joined #openstack-smaug | 08:32 | |
zengyingzhe | yinwei_computer, for the scenario you gave, the parent resource would be a VM, then the image plugin could construct a nova client and use this client to query the VM details, then we could get the image which this VM uses. | 08:32 |
openstackgerrit | Yingzhe Zeng proposed openstack/smaug: Implement the ProtectableRegistry https://review.openstack.org/281783 | 08:42 |
openstackgerrit | Yingzhe Zeng proposed openstack/smaug: Implement the ProtectableRegistry https://review.openstack.org/281783 | 08:45 |
openstackgerrit | zengchen proposed openstack/smaug: Implement RestAPIs of trigger. https://review.openstack.org/286406 | 08:46 |
*** openstackgerrit has quit IRC | 08:47 | |
*** openstackgerrit has joined #openstack-smaug | 08:48 | |
openstackgerrit | zengchen proposed openstack/smaug: RPC client recreates exception object failed https://review.openstack.org/289685 | 08:48 |
*** yuval has quit IRC | 08:49 | |
gampel | xiangxinyon: hi | 08:49 |
gampel | yes | 08:49 |
openstackgerrit | Yingzhe Zeng proposed openstack/smaug: Implement cinder protectable plugin https://review.openstack.org/285611 | 08:52 |
openstackgerrit | Yingzhe Zeng proposed openstack/smaug: Implement nova protectable plugin https://review.openstack.org/286542 | 08:54 |
*** smile-luobin has quit IRC | 08:58 | |
*** smile-luobin has joined #openstack-smaug | 08:58 | |
openstackgerrit | Yingzhe Zeng proposed openstack/smaug: Implement the ProtectableRegistry https://review.openstack.org/281783 | 08:58 |
openstackgerrit | Yingzhe Zeng proposed openstack/smaug: Implement cinder protectable plugin https://review.openstack.org/285611 | 09:00 |
openstackgerrit | zengchen proposed openstack/smaug: Implement RestAPIs of Scheduled Operation. https://review.openstack.org/287036 | 09:01 |
openstackgerrit | Yingzhe Zeng proposed openstack/smaug: Implement nova protectable plugin https://review.openstack.org/286542 | 09:06 |
*** mingyu has quit IRC | 09:08 | |
*** mingyu has joined #openstack-smaug | 09:12 | |
openstackgerrit | Yingzhe Zeng proposed openstack/smaug: Implement protectable RPC handlers https://review.openstack.org/285921 | 09:14 |
openstackgerrit | Yingzhe Zeng proposed openstack/smaug: Implement protectable RPC handlers https://review.openstack.org/285921 | 09:32 |
zengyingzhe | gampel, could you please check https://review.openstack.org/#/c/281783/ | 09:39 |
zengyingzhe | and https://review.openstack.org/#/c/285921/ | 09:39 |
zengyingzhe | these are two base patch, many other patches depend on them. | 09:40 |
zengyingzhe | it would be better if they could be merged sooner. | 09:40 |
zengyingzhe | saggi, please review the patches I mentioned above. Thanks. | 09:41 |
yinwei_computer | yingzhe, yes, that's the case I don't prefer | 09:48 |
yinwei_computer | it means the 'image' protectable has to visit its parent resource 'server', and should understand the result of 'get_server' | 09:49 |
yinwei_computer | it's not a good design for encapsulation | 09:49 |
yinwei_computer | the good design principle should be each layer/component should only understand its own data/logic, and there should be an isolation | 09:50 |
yinwei_computer | If you adjust the lay out and each protectable only returns its own children not its parent children, say, we have server protectable to return all of its children resources:volume, image, there's no such trouble. | 09:52 |
yinwei_computer | which means, server should only visit nova and should only understand the metadata of server; while image should only visit glance and only understand the metadata of image; | 09:54 |
yinwei_computer | gampel, what do you think? | 09:54 |
yinwei_computer | in this way, we coupled image protectable tightly with server protectable | 09:57 |
zengyingzhe | all of the plugins are isolated object, we just need to focus on the interfaces provided to outside, and the interfaces return what we want. Does it matter so much how it implements interally> | 10:02 |
zengyingzhe | And for project resource, based on your idea, the implementation of projest plugin also must be aware other services like nova and cinder, i think it's got the same problem. | 10:04 |
yinwei_computer | if the server metadata changes, say, you have to change all of its children protectables which are not server, which sounds weird | 10:05 |
yinwei_computer | and it's confusing for user to develop them | 10:05 |
yinwei_computer | protectables are plugins, which could be replaced and reimplemented by users | 10:06 |
yinwei_computer | for project, yes and no. | 10:06 |
yinwei_computer | this is because there's no project service in openstack and thus we need list all of its children resources | 10:07 |
zengyingzhe | I don't quite understand why "server metadata changes, say, you have to change all of its children protectables which are not server" | 10:09 |
yinwei_computer | If protectable is out internal logic, I don't care how to implement it because user won't change it. Since they are plugins, just have concern on the flexibility. | 10:10 |
yinwei_computer | say, server metadata changes one field | 10:10 |
yinwei_computer | it used to be server.image and now it turns to be server.image_id | 10:11 |
yinwei_computer | you need change image plugin | 10:11 |
zengyingzhe | have you seen gampel's reply after your comment. in this way there's a advantage, that if we add a new resource type dependency later, we only need to add a new plugin, no need to modify the parent plugins that already exist. | 10:17 |
zengyingzhe | this make the future maintenance much easier. | 10:17 |
zengyingzhe | but what you said also make sense. maybe we'd better pull saggi&gampel in to discuss together. | 10:18 |
zengyingzhe | i'm about go home soon, let's talk about it tomorrow? | 10:20 |
*** mingyu has quit IRC | 10:41 | |
*** smile-luobin has quit IRC | 10:52 | |
*** gampel has quit IRC | 11:03 | |
*** chenying has quit IRC | 11:11 | |
*** chenying has joined #openstack-smaug | 11:12 | |
*** yuval has joined #openstack-smaug | 11:42 | |
*** gampel has joined #openstack-smaug | 12:30 | |
xiangxinyong | gampel: sorry for late to reply. | 12:48 |
xiangxinyong | gampel:zengyingzhe and I have a question about the smaug ui. | 12:49 |
xiangxinyong | gampel:when we create a protection plan, we need to choose resources. | 12:50 |
xiangxinyong | gampel:and the parameters need be specified for each resource instance. | 12:51 |
xiangxinyong | gampel:if the resource type is the same, are the resource instance parameters the same too? | 12:52 |
xiangxinyong | gampel: e.g. Do two vm instances have the same resource parameters? | 12:53 |
*** gampel1 has joined #openstack-smaug | 13:02 | |
*** gampel has quit IRC | 13:02 | |
*** chenying_ has joined #openstack-smaug | 13:37 | |
*** chenying has quit IRC | 13:40 | |
*** gampel1 has quit IRC | 14:15 | |
*** gampel has joined #openstack-smaug | 14:19 | |
*** gampel has quit IRC | 15:43 | |
yuval | zengyingzhe: hey | 15:51 |
yuval | zengyingzhe: in nova protectable plugin, nova's configuration is in server.py. Since that there is no glance->nova mapping - in glance protectable we need to use nova's client as well, with that configuration. | 15:52 |
yuval | zengyingzhe: how about moving the nova and cinder client configuration to a central location?Maybe common.config | 15:53 |
yuval | zengyingzhe: plus the nova client version | 16:04 |
*** gampel has joined #openstack-smaug | 16:47 | |
*** gampel has quit IRC | 19:13 | |
*** gampel has joined #openstack-smaug | 19:17 | |
*** gampel has left #openstack-smaug | 19:17 | |
*** gampel has joined #openstack-smaug | 19:51 | |
*** gampel has quit IRC | 20:51 |
Generated by irclog2html.py 2.14.0 by Marius Gedminas - find it at mg.pov.lt!