09:00:34 <saggi> #startmeeting smaug 09:00:35 <openstack> Meeting started Tue Aug 2 09:00:34 2016 UTC and is due to finish in 60 minutes. The chair is saggi. Information about MeetBot at http://wiki.debian.org/MeetBot. 09:00:36 <openstack> Useful Commands: #action #agreed #help #info #idea #link #topic #startvote. 09:00:38 <openstack> The meeting name has been set to 'smaug' 09:00:43 <saggi> Hi everyone 09:00:59 <xiangxinyong> hello 09:00:59 <zhonghua-lee> hi 09:01:11 <yuval> hey 09:02:03 <saggi> Are we waiting on anyone? yiwei, chenzeng, etc 09:03:15 <yuval> chenzeng should be here, he wanted to discuss his suggestion 09:03:37 <chenying> yinwei is off the office today. 09:03:46 <zengchen> yuval:yes 09:03:52 <saggi> Do you guys know where everyone is? 09:04:29 <xiangxinyong> :) 09:04:33 <zhonghua-lee> I am here :) 09:04:42 <zengchen> yuval:as i descirbed on the launchpad, we may need to add a new executor 09:05:08 <saggi> OK let's start 09:05:15 <saggi> #topic mascot selection 09:05:25 <saggi> We didn't get the koala mascot 09:05:31 <saggi> Kolla won the coin toss 09:05:39 <zengchen> saggi:i am online 09:05:41 <zhonghua-lee> is there any option? 09:05:53 <saggi> Their name is also an alternative name for a Koala so they have the same claim on the animal. 09:06:40 <chenying> we use Karbor as the name or voting a new name? 09:06:41 <zhonghua-lee> so we have to pick another one? 09:07:02 <saggi> We can change the name since we did not announce the name yet 09:07:05 <saggi> We suggested a tardigrade http://apod.nasa.gov/apod/ap130306.html which is the most resilient animal on earth. 09:07:10 <saggi> Not as a name but a mascot 09:07:40 <zhonghua-lee> saggi:what's that? 09:08:18 <saggi> https://zh.wikipedia.org/wiki/%E7%BC%93%E6%AD%A5%E5%8A%A8%E7%89%A9%E9%97%A8 09:10:16 <zhonghua-lee> looks good 09:10:38 <zhonghua-lee> except looks like maggot 09:10:42 <saggi> :) 09:10:47 <zhonghua-lee> forgive me 09:10:52 <chenying> It looks fearsome. 09:10:57 <saggi> That's the problem. It fits thematically but it not as cute. 09:11:26 <yuval> imo it is quite cute 09:11:28 <chenying> Can we pick another cute animal? 09:11:31 <zhonghua-lee> :) 09:11:41 <saggi> Yes 09:11:41 <xiangxinyong> :( 09:11:57 <saggi> But it needs to be somewhat related. 09:12:13 <zhonghua-lee> can we use dragon? 09:12:25 <saggi> They don't allow mythical animals. 09:12:33 <zhonghua-lee> oj 09:12:35 <zhonghua-lee> ok 09:12:46 <xiangxinyong> I guess this animal is from the mars 09:12:49 <chenying> can we use panda? 09:12:49 <xiangxinyong> :) 09:13:12 <saggi> yuval: Where is the list of the taken mascots? 09:13:23 <yuval> https://www.openstack.org/project-mascots 09:13:36 <yuval> they just added another batch of taken mascots 09:13:57 <saggi> The problem with panda is that the name is to generic. 09:14:07 <saggi> And that the WWF already uses it as a mascot 09:14:17 <xiangxinyong> what about the glass crab? 09:14:22 <xiangxinyong> http://image.baidu.com/search/detail?ct=503316480&z=0&ipn=d&word=glass%20crab&hs=0&pn=5&spn=0&di=25493893821&pi=&rn=1&tn=baiduimagedetail&ie=utf-8&oe=utf-8&cl=2&lm=-1&cs=1366933940%2C1128410029&os=3481574606%2C599075615&simid=0%2C0&adpicid=0&ln=30&fr=ala&fm=&sme=&cg=&bdtype=15&oriquery=&objurl=http%3A%2F%2Fpre01.deviantart.net%2F736a%2Fth%2Fpre%2Ff%2F2016%2F005%2F3%2F8%2Fglass_toon_crab_by_laticis-d9mu0k1.jpg&fromurl=ippr_z2C%24qAzdH3FA 09:14:22 <xiangxinyong> zdH3Fg5pj_z%26e3Bpwwksj_z%26e3Bv54AzdH3Fr5fpAzdH3F8nC0AzdH3F4w6t3jkj6ptg2_z%26e3B1jetwgpw6p_z%26e3Bv54AzdH3Fw6pAzdH3FLwvj-Lw1y-cb8mn09an&gsm=0 09:14:36 <yuval> btw Quokka taken 09:15:30 <chenying> crab? 09:15:37 <yizhihui> camel? 09:16:12 <xiangxinyong> yizhihui: welcome to join the weekly meeting 09:16:21 <yizhihui> :) 09:16:24 <zhonghua-lee> camel, good idea 09:16:37 <yizhihui> or Black Mamba 09:17:11 <zhonghua-lee> ohh...no 09:18:15 <chenying> golden monkey 09:18:17 <zhonghua-lee> I think we can not get the conclusion this time 09:18:52 <saggi> Please send suggestions if you have any. 09:18:59 <xiangxinyong> I guess we can submit the mascot in this link. https://etherpad.openstack.org/p/smaugmascot 09:19:04 <chenying> http://s15.sinaimg.cn/mw690/65d9d9a5tx6CFUsY8z4be&690 golden monkey 09:19:14 <saggi> #topic Add a new executor for Operation Engine service 09:19:17 <zhonghua-lee> saggi: +1 09:20:07 <saggi> ping zengchen, this is your item 09:20:34 <zhonghua-lee> https://bugs.launchpad.net/smaug/+bug/1608415 09:20:34 <openstack> Launchpad bug 1608415 in Smaug "Add a new executor for Operation Engine service" [High,Triaged] - Assigned to zengchen (chenzeng2) 09:20:36 <chenying> It is better add the mascot with a picuter link. 09:21:02 <zengchen> saggi: 09:21:03 <tommylike_321> monkey +1 09:21:08 <zengchen> saggi:ok 09:21:34 <zengchen> do you have seen it. 09:21:59 <yuval> zengchen: sounds like a good idea 09:22:32 <saggi> zengchen: It's a great idea 09:22:41 <saggi> Looks simple enough to implement 09:22:53 <saggi> and have it retrofitted so we don't need the threaded one 09:22:59 <zengchen> yuval:ok, if you all agree with it, i will try to implement it. 09:23:13 <saggi> zengchen: will it come instead of the threaded one 09:23:14 <saggi> ? 09:23:50 <zengchen> saggi:we can keep the thead, because we can set the config to choose which to be used. 09:23:51 <zhonghua-lee> is it a addition or default? 09:24:22 <zengchen> zhonghua-lee:i suggest to be default. 09:24:28 <saggi> isn't the green thread one more powerful. Why would you want the threaded one? 09:24:44 <zhonghua-lee> zengchen: thanks 09:24:51 <chenying> I only care about the executor whether meet all the test scene of engine. 09:25:21 <zengchen> saggi:you mean delete the codes of thread? 09:25:30 <saggi> yes 09:26:14 <zengchen> saggi:i suggest to wait a time. when the new executor runs, we can try to delete it. 09:26:34 <zhangshuai_> +1 09:26:38 <saggi> zengchen: OK, but I'd rather not have a lot of code goes unused. 09:27:09 <zengchen> saggi:ok, i will keep the codes clean. 09:27:38 <saggi> # topic Scheduled Operation Improvements Proposal 09:27:45 <saggi> ping xiangxinyong 09:27:53 <xiangxinyong> hello saggi 09:28:01 <xiangxinyong> https://etherpad.openstack.org/p/scheduledoperation 09:28:16 <xiangxinyong> could you take a look at this link? 09:28:22 <saggi> I already did 09:28:26 <xiangxinyong> thanks 09:28:54 <gsagie> So what is Smaug mascot? 09:28:58 <gsagie> Chenying? 09:29:20 <xiangxinyong> saggi:what do you think about it? 09:29:28 <chenying> We will add the suggested mascot to the link. 09:29:29 <saggi> It's fine by me as long as you can set retention time and max checkpoint to be infinite 09:29:37 <chenying> https://etherpad.openstack.org/p/smaugmascot 09:29:47 <zhonghua-lee> I think it's reasonable to add the description 09:30:02 <saggi> gsagie: We already selected a mascot: https://media.licdn.com/media/p/2/000/18e/3dd/1d99b9c.jpg 09:30:34 <xiangxinyong> yeah. i need to add description into scheduled operation 09:30:38 <yuval> saggi: :P: 09:30:51 <zengchen> xingxinyong:+1 09:31:07 <zhangshuai_> xiangxinuong:+1 09:31:15 <saggi> xiangxinyong: +1 09:31:25 <saggi> The rest can be added if they are optional 09:31:25 <xiangxinyong> saggi: about retention_duration and max_backups 09:31:43 <chenying> add description into scheduled operation--- I also think add the description filed to other resources. like plan resotre 09:31:57 <xiangxinyong> yeah. description is optinal 09:32:02 <zhonghua-lee> chenying:+1 09:32:07 <saggi> chenying: Also retention and max_backups 09:32:15 <saggi> xiangxinyong: 09:32:20 <saggi> should be optional 09:32:27 <chenying> name and description field have different meaning. 09:33:06 <saggi> xiangxinyong: I think we are all in agreement that these are good additions as long as they are optional. 09:33:16 <chenying> yes it is optional 09:33:16 <xiangxinyong> understood. 09:33:21 <saggi> Is retention time per checkpoint? 09:33:39 <xiangxinyong> but how to implement max_backups and retention_duration for scheduled operation? 09:33:47 <xiangxinyong> do you have some suggestions? 09:34:09 <zhonghua-lee> add new Operation? 09:34:17 <saggi> You run the operation periodically looking for checkpoints that you created that are older than duration. 09:34:57 <saggi> You will need to keep record of all created checkpoints in the operation db since querying the bank is slow. 09:35:13 <yuval> I don't understand: if I manually create a checkpoint, and then later a scheduled operation is defined with max_backups, it may delete my checkpoint? 09:35:31 <saggi> If it was created by this operation 09:35:47 <xiangxinyong> I guess we need to launch a delete checkpoint request from the operation engine service 09:35:50 <zhonghua-lee> saggi: who will trigger this operation? Operation engine? 09:35:55 <saggi> yes 09:36:01 <saggi> Since it's policy 09:36:07 <xiangxinyong> I guess operation engine service will trigger 09:36:41 <chenying> all retention time and max backups are the parameters of policys. 09:37:43 <xiangxinyong> OK. I will write a spec about this features 09:38:03 <xiangxinyong> thanks guys 09:38:05 <zhonghua-lee> xiangxinyong:+1 09:38:29 <chenying> xiangxinyong:+1 need a spec about these parameters. 09:38:31 <zhonghua-lee> xiangxinyong: some guys may want to know the detaild flow 09:38:42 <yuval> How will the scheduled operation find all the checkpoints? 09:38:54 <yuval> originated from it? 09:38:57 <chenying> call the list checkpoint api 09:39:12 <xiangxinyong> chengying:+1 09:39:14 <saggi> yuval: When it creates the checkpoint it will need to keep the ID returned 09:39:26 <yuval> chenying: and list ALL checkpoints? 09:39:32 <yuval> chenying: we don't have filter on metadata 09:39:43 <yuval> saggi: keep where? 09:39:49 <saggi> operation engine db 09:39:53 <xiangxinyong> yuval: but i guess we filter it by plan 09:39:55 <saggi> Since it's local 09:40:24 <chenying> make sure we have save the checkpoint id to the operation engine db 09:40:46 <chenying> not the operation engine db some operation db table. 09:40:59 <zhonghua-lee> xiangxinyong: do you consider the use case: user can create it manually 09:41:17 <zengchen> i suggest xiangxinyong give a complete desgin. 09:41:21 <saggi> Don't design anything that filters the checkpoints. It won't scale until we decide on a querying mechanism. This will only happen for P-version, not Newton. 09:41:25 <saggi> zengchen: +1 09:41:58 <xiangxinyong> zhonghua-lee: yeah. The requirement is only about the scheduled protection 09:42:28 <xiangxinyong> OK.guys. I will specify a spec about this feature 09:42:30 <zhonghua-lee> xiangxinyong: so it is just a policy, not for the protection resource 09:42:50 <saggi> #action xiangxinyong write a BP on advance backup policy 09:42:58 <saggi> #topic Volume OPTIONS_SCHEMA Improvements Proposal 09:43:38 <saggi> Don't Huawei implement this through Cinder Backup? 09:43:52 <chenying> i note that SAVE_SCHEMA not being used when creat checkpoint data. 09:44:11 <xiangxinyong> saggi: there is some differenece 09:44:46 <xiangxinyong> we could add emc volume plugin or eisoo volume plugin. 09:45:12 <xiangxinyong> https://etherpad.openstack.org/p/volumeschema 09:45:26 <xiangxinyong> guys,please review it. thanks 09:45:30 <saggi> xiangxinyong: I would like to have everything else in the reference implementation working perfectly before adding new protection plugins. 09:45:39 <saggi> Also we need these to be tested as well 09:45:52 <saggi> I the full stack tests 09:46:46 <yuval> xiangxinyong: you want these to be added to the default cinder protection plugin? 09:47:47 <xiangxinyong> yuval: yeah. I want to add a new protection volume plugin 09:48:01 <saggi> xiangxinyong: Id 09:48:49 <xiangxinyong> saggi:? 09:48:55 <saggi> xiangxinyong: I'd rather wait with alternative plugins until we have everything we gurantee for v1 solid.. 09:49:13 <zhonghua-lee> saggi:+1 09:49:32 <zhonghua-lee> saggi: do we accept other vendors to join us? 09:49:46 <xiangxinyong> saggi: good suggestions 09:49:53 <saggi> zhonghua-lee: Sure, everyone is invited 09:50:06 <zhonghua-lee> espacially the backup vendors 09:50:19 <zhonghua-lee> saggi: :) 09:50:33 <saggi> And if they spend the time on their plugins it's OK. But we, as the core team, should concentrate on the core for now. 09:50:45 <chenying> I think verdor can add their plugins. Like backup software verdor can add ther plugins. 09:50:54 <saggi> We also need to take the Cinder route and only accept plugins with proper tests and CI 09:51:04 <zhonghua-lee> saggi: OK 09:51:25 <saggi> Otherwise we have to do all the testing ourselves, which doesn't scale. 09:51:44 <zhonghua-lee> saggi:+1 09:52:57 <zhonghua-lee> according the conclusion, I think we prefer to delay this proposal 09:53:22 <yizhihui> when will v1 solid? 09:53:46 <saggi> When N is stable hopefully 09:53:53 <saggi> That the target 09:54:26 <zhonghua-lee> yizhihui: why you ask this question? 09:54:51 <zhonghua-lee> yizhihui: do you have any plan to add your plug-in? 09:55:17 <yizhihui> yes 09:55:27 <saggi> If we get there sooner we can tag and start building v2. BTW, v2 roadmap will be discussed at Bacelona. 09:55:42 <chenying> <yizhihui is from a backup software vendors. They plan to develop a new plugin about app backup. 09:56:02 <zhonghua-lee> saggi: yizhihui: I guess you can add it as you please 09:56:14 <saggi> roadmap is here https://etherpad.openstack.org/p/smaug 09:56:15 <zhonghua-lee> saggi: what's your opinion? 09:56:21 <yizhihui> ok 09:56:27 <saggi> zhonghua-lee: about what? 09:56:57 <zhonghua-lee> saggi: yizhihui is from 3rd backup vendor, he wants to add a plug-in into Smaug 09:57:09 <zhonghua-lee> saggi: can he add it now? 09:57:43 <saggi> You can add anything to Ocata. it's just candidates 09:57:57 <saggi> We will filter the list in Barcelona 09:58:30 <zhonghua-lee> saggi:OK 09:59:01 <saggi> We will accept 3rd party plugin code even now. But we can't gurantee it will ship as supported in N. 09:59:32 <saggi> Also, it will require dedication as things are moving fast and breakage may oocure in the plugin interface before N. 09:59:36 <yizhihui> saggi: I got it 09:59:39 <zhonghua-lee> saggi: +1 10:00:03 <saggi> We are out of time 10:00:18 <saggi> #topic open discussion 10:00:20 <yuval> mascot? 10:00:21 <saggi> any last words? 10:01:08 <saggi> Thank you everybody 10:01:12 <xiangxinyong> Thanks 10:01:14 <zhonghua-lee> thanks 10:01:15 <xiangxinyong> good bye 10:01:16 <saggi> yizhihui: Hope to see you next time as well 10:01:18 <saggi> :) 10:01:23 <saggi> #endmeeting