14:59:14 <yuval> #startmeeting Karbor
14:59:15 <openstack> Meeting started Tue Nov  1 14:59:14 2016 UTC and is due to finish in 60 minutes.  The chair is yuval. Information about MeetBot at http://wiki.debian.org/MeetBot.
14:59:16 <openstack> Useful Commands: #action #agreed #help #info #idea #link #topic #startvote.
14:59:18 <openstack> The meeting name has been set to 'karbor'
14:59:25 <yuval> hello everybody
14:59:33 <saggi> Hi
14:59:44 <edisonxy> hello
14:59:54 <yuval> saggi: sorry, thought you are away. you go :)
15:00:34 <saggi> Are we waiting on anyone?
15:01:06 <yuval> xiangxinyong, zhonghua, yinwei?
15:01:09 <yuval> chenying
15:01:29 <edisonxy> hello i am xiang
15:01:33 <zhonghua> hi
15:01:43 <saggi> #topic Checkpoint API -- Partial Delete
15:01:55 <saggi> I can't control the meeting
15:01:57 <saggi> :)
15:02:02 <saggi> yuval: You do it
15:02:13 <yuval> #topic Checkpoint API -- Partial Delete
15:02:24 <saggi> Who suggested it?
15:02:44 <yuval> chenying
15:02:44 <edisonxy> chenying
15:02:47 <zhonghua> chenying?
15:03:26 <saggi> I guess we'll postpone discussing it for next week
15:03:29 <chenying1> hi
15:03:33 <saggi> Just in time
15:04:10 <saggi> chenying1: could you explain Partial Delete?
15:04:53 <chenying1> I submit this topic, just want to discuss whether we need consider this scene。
15:05:39 <chenying1> saggi I have described the scene in the irc meeting wiki.
15:06:10 <zhonghua> The user create a plan with several volumes. These volumes don't have any relationships.
15:06:10 <zhonghua> Add these resources to one plan, the user just want to protect them at the same time point.
15:06:10 <zhonghua> So he have a request that operating the backup copies in one checkpoint separately.
15:06:31 <saggi> I don't think so. Checkpoints are read only once they are created. We might suggest creating a new checkpoint which is a clone of parts of an existing checkpoint but I don't think it's anything near the top use cases we want to support
15:06:39 <saggi> for at least the next 8 months
15:06:41 <zhonghua> https://wiki.openstack.org/wiki/Meetings/Karbor
15:07:06 <saggi> We have a lot more pending issues
15:07:10 <saggi> IMHO
15:07:17 <saggi> So unless we see real demand for it
15:07:22 <yuval> +1
15:07:25 <saggi> I think we should put it on the back burner
15:09:14 <zhonghua> saggi: need we consider it later?
15:09:23 <chenying1> This scene is from the a private Cloud product.
15:09:42 <yuval> #topic Openstack Barcelona
15:10:44 <saggi> yuval, zhonghua-lee edisonxy and I were in Openstack Summit last week
15:11:19 <saggi> We had a lot of productive talks
15:11:42 <saggi> But it was too bad that some of you couldn't make it and participate
15:12:00 <saggi> https://wiki.openstack.org/wiki/Design_Summit/Ocata/Etherpads#Karbor
15:12:07 <saggi> We summed up the meetings here
15:12:25 <saggi> But I think that for now going over: https://etherpad.openstack.org/p/karbor-ocata-roadmap
15:13:00 <saggi> We've highlighted things that we want to achieve in the Okata cycle
15:13:20 <saggi> First, I'd like to remind everyone that it's a short cycle.
15:13:29 <saggi> It ends in February
15:13:36 <saggi> So there isn't a lot of time
15:13:48 <TommyLikeHu_> yes, cinder had blocked most of the specs~
15:13:50 <Chenzeng> saggi:hope for your sharing,sending an email is a good idea
15:14:16 <saggi> Chenzeng: I will, I'm sorry I was very ill the last few days.
15:14:29 <saggi> Second, our goal for Newton was to be stable
15:14:36 <saggi> Which we failed to achieve
15:15:29 <saggi> We have a lot of problems that prevent even the most common use cases from working
15:15:54 <saggi> We need to hold back on new features and start fixing bugs.
15:16:34 <chenying1> Some patches about operationengine need be reviewed and to be merged. Now the operationengine doesn't work in karbor's master branch.
15:16:53 <saggi> chenying1 email me the links
15:17:06 <yuval> chenying1: what do you mean 'doesnt work'? did a specific patch break it?
15:17:09 <zhonghua> chenying1: yeah, there are many patches are waiting for merging.
15:17:12 <saggi> yuval, and I will make a detailed plan on what exactly it means. But first we need to make sure everyone can run and test everything, including the UI.
15:17:28 <chenying1> saggi  OK I will send a etherpad link. some patches have adressed there.
15:17:40 <saggi> chenying1, +1
15:18:36 <saggi> The things that are marked in bold are things that were accepted as part of the Okata cycle. It's still not set in stone. You can comment on the etherpad just make sure to put your name so we know who uses what color.
15:18:43 <saggi> Just don't mark things bold
15:18:58 <saggi> Just add +1\-1
15:19:06 <chenying1> 'doesnt work' means that karbor cann't run scheduled protection plan.
15:19:29 <yuval> chenying1: I understand, I was asking if a specific patch broke it
15:20:27 <chenying1> yuval: Yes not only one.  several patches.
15:20:30 <Chenzeng> saggi:I think there are some issues about taskflow used in protect service to re-consider.
15:20:50 <yuval> Chenzeng: what do you mean?
15:22:16 <saggi> Chenzeng: ?
15:22:50 <Chenzeng> yuval:for example,we don't use the parallel executor of taskflow
15:23:17 <chenying1> Yuval  What's your plan about the new design patches? IMO we need review and merge them ASAP. If there are some situation we have not think about clearly, we could submit new patch to fix it.
15:24:18 <Chenzeng> yuval,:and  some other things,I will send an email to describe details
15:24:37 <saggi> yuval: Would you like to talk about the worksession we had on the subject?
15:24:52 <yuval> chenying1: we have discussed that in the summit. The idea for ocata is to have that api for protect operation, while restore operation is unable to use such api because of heat
15:25:20 <yuval> chenying1: we were talking about removing heat, but until then, restore cannot have a pre and post for each resource
15:26:42 <chenying1> Ok I need update the message to some verdors.   They are developing the plugins based on the new interfaces.
15:27:26 <yuval> chenying1: goal for ocata is to stablize the plugin api
15:29:19 <saggi> chenying1: we already spoke about it that we don't say that the internal APIs are stable until we hit v 1.0
15:31:08 <saggi> OK
15:31:19 <chenying1> saggi  we also need think about this,  some verdors are planing to develop the integration plugins.
15:32:17 <saggi> chenying1: we already had this conversation and we agreed that we can't do that. That it makes sense to design supported interfaces properly
15:33:23 <edisonxy> yuval: does it mean your patches for protection service will not be merged in ocata?
15:33:40 <yuval> edisonxy: they will, but further changes are required to them
15:34:08 <edisonxy> understand.
15:34:10 <chenying1> So the southbound interface of karbor is high priority.
15:34:18 <saggi> It is
15:34:21 <yuval> yep
15:34:46 <saggi> We have been asking for comments for weeks
15:35:32 <edisonxy> yuval: we want to meege your patches asap
15:35:57 <saggi> zhonghua-lee, edisonxy: How was the summit for you?
15:36:12 <chenying1> about the southbound interface of karbor  I think we need gather more solutions form backup software vendors.
15:36:36 <edisonxy> yuval: and make some further modification
15:37:45 <edisonxy> saggi: very well. we learned a lot of things and people.
15:40:06 <yuval> I feel the same way. It was great you meet you edisonxy and zhonghua-lee as well!
15:40:14 <yuval> *to meet
15:42:21 <saggi> Just to sum things up we need to direct our efforts better. Getting stable and staying there is top priority for Okata. We can't afford to miss that deadline. We want to be able to say we are stable for Boston and that means being stable for Okata.
15:44:52 <yuval> #topic open discussion
15:45:48 <saggi> Any last thoughts?
15:47:19 <yuval> #endmeeting