05:03:13 #startmeeting servicevm-device-manager 05:03:14 Meeting started Tue Jul 1 05:03:13 2014 UTC and is due to finish in 60 minutes. The chair is yamahata. Information about MeetBot at http://wiki.debian.org/MeetBot. 05:03:15 Useful Commands: #action #agreed #help #info #idea #link #topic #startvote. 05:03:17 The meeting name has been set to 'servicevm_device_manager' 05:03:33 Is bob here? 05:03:53 Let's give him a minute 05:03:57 yamahata: doesn't look like he is online 05:04:37 Oops. 05:04:53 This week natarajk is also absent. 05:05:11 He'll catch up with the log 05:05:56 Okay let's start anyway. 05:06:02 #topic Announcement 05:06:21 The starckforge repo for servicevm has been created. 05:06:40 Thus gerrit review system can be used for our review. 05:07:04 #link http://git.openstack.org/cgit/stackforge/tacker/ tacker repo 05:07:17 #link http://git.openstack.org/cgit/stackforge/tacker-specs/ tacker-specs repo 05:07:33 #link http://git.openstack.org/cgit/stackforge/python-tackerclient/ python-tackerclient repo 05:08:46 So the group in gerrit system is also created. Please let me know privately your account name/email address to add tacker-core tacker-specs-core 05:09:08 yamahata: OK 05:09:16 #action candidate for core please let me know privately your account/mail address of gerrit. 05:09:28 #action yamahata add core members to gerrit system 05:10:05 Now I've created a sort of working tacker server/client in github. 05:10:17 yamahata: I think bobmel could be Bob 05:10:30 #action yamahata push the current code to review system. 05:10:43 oh. 05:10:45 Hi all. This is Bob. 05:10:53 bobmel: Hi. glad to see you. 05:11:05 bobmel: how many handles do you have :-) ? 05:11:31 s3wong: too many... :-) 05:12:20 bobmel: could you please let me know privately your account/email address of the gerrit system later. 05:12:42 yamahata: Sure 05:12:58 The a sort of working code is based on Neutron. So the it includes "uninteresting" code. 05:13:06 I think we can merge those patches at first. 05:13:17 For example deleting unnecessary Neutron specific files. 05:13:24 renaming neutron -> tacker 05:14:04 After merging tedious patches, we can start actual api discussion and patch review. 05:14:41 #link https://review.openstack.org/#/c/103724/ API review patch 05:15:01 yamahata: bobmel: have you guys thought about merging your pieces (serviceVM lifecycle mgmt / mgmt interface & device manager) together? Is that one of the goals? 05:16:52 s3wong: the current goal is to reach consensus with API, then consolidate the existing three or four implementations. 05:17:16 yamahata: OK 05:17:25 s3wong, yamahata: Yes that makes sense. 05:18:25 I think bobmem and I have close idea for API. So minor adjustment would suffice. 05:19:01 But I had a looked at DNRM document, their api is a different. So we needs discussion. 05:19:30 Unfortunately today we don't have Karthik. 05:19:39 the interceptor component probably won't have too many fans in Neutron community 05:20:13 s3wong: agree. 05:20:42 DNRM server has the feature of pooling VM. 05:21:06 I'm not sure where pooling vm live. 05:21:23 where the feature of pooling vm should live. 05:22:36 we can continue concrete discussion with gerrit. 05:23:03 about API? Or the DNRM thing? 05:23:04 #topic incubation process 05:23:09 #undo 05:23:09 Removing item from minutes: 05:23:42 s3wong: both. I think DNRM guys want to join the API discussion. 05:23:54 OK 05:24:03 Probably API needs modification based on their input 05:24:40 #topic servicevm spec/code review 05:24:51 yamahata: we do have to evaluate as a community which pieces, if any. of DNRM would make sense to integrate 05:25:06 (go on to the next topic, please. No need to undo :-) ) 05:25:15 s3wong: Agree 05:25:38 I need to have a closer look at their code. 05:26:25 yamahata: +1 05:26:28 anything else with spec/code review? 05:26:45 #action everyone have a closer look at DNRM code 05:26:50 bobmel is now bmelande :-) 05:26:58 #action yamahata have a closer look at DNRM core more. 05:27:51 Last week One of my action item is to create l3 plugin. 05:28:15 I haven't done it yet. After pushing the tacker code for review, I'll create the blueprint. 05:28:32 #action yamahata create a blueprint for reference l3-plugin with servicevm 05:29:28 #topic Open Discussion 05:29:52 Okay do we have anything to discuss? 05:30:18 all good for me 05:30:35 cool. 05:30:54 It seems one hour time slot is too long. 05:31:04 How about 30min slots from the next time? 05:31:24 yamahata: OK, either that or biweekly 05:32:03 s3wong: Surely biweek is also an option. 05:32:11 yamahata, s3wong: I think I prefer weekly but potentially make it shorter. 05:32:35 bmelande: OK. Let's go with yamahata 's original suggestion of 30 mins then 05:32:57 #agreed 30min weekly meeting 05:33:15 Regarding API: Some notifications from device manager to its "customers" could be useful 05:34:10 For example: to let a service plugin know that a VM is discovered to be "dead" 05:34:13 bmelande: interesting. Could you be more specific? Probably in gerrit review part. 05:34:34 Yes I'll comment to document. 05:34:35 bmelande: exception handling? 05:35:49 s3wong: Well possibly it could be done like that. Not sure. 05:36:43 bmelande: cool, let's comment on it on gerrit review (API) 05:37:21 In our impl. the device manager does some basic checks of VM health. Ok, I'll make a comment there. 05:37:57 any other issues? 05:38:36 no, will review API doc 05:38:47 thank you. see you next week 05:38:54 thanks 05:38:54 #endmeeting