| Posted | Nick | Remark | |
|---|---|---|---|
| #openstack-cyborg - 2018-02-07 | |||
| 15:37:47 | zhipeng | if one gets longer we will prolong | |
| 15:38:15 | zhipeng | since Mon have Board meeting and people still coming in | |
| 15:38:30 | zhipeng | I would suggest the Nova-Cyborg discussion happening on Tuesday afternoon | |
| 15:38:37 | zhipeng | I will double check with the nova cores | |
| 15:39:53 | SundarN | Sounds good to me | |
| 15:40:08 | zhipeng | everyone understands my proposal ? | |
| 15:40:25 | zhipeng | office hour in the morning, session discussion in the afternoon | |
| 15:40:33 | zhipeng | on Mon and Tue | |
| 15:40:49 | SundarN | Will that be enough time in the afternoon for all the topics? | |
| 15:41:06 | zhipeng | SundarN yes since some of them will be quick :) | |
| 15:41:33 | SundarN | ok | |
| 15:41:43 | Li_Liu | #info Li Liu | |
| 15:42:08 | zhipeng | if there are no questions or strong objections | |
| 15:42:15 | zhipeng | let's settle on that schedule | |
| 15:42:26 | zhipeng | I will send an email to the mainling list, just for the record | |
| 15:42:44 | zhipeng | now the second item, team dinner | |
| 15:43:05 | zhipeng | Tuesday night sounds reasonable ? | |
| 15:43:27 | Vipparthy | ok | |
| 15:43:38 | SundarN | Other teams are also planning dinner, but they haven't selected days yet :) | |
| 15:44:09 | zhipeng | yep, we are early so we select early :P | |
| 15:44:36 | zhipeng | If Tuesday is good then please clear your calendar for the night | |
| 15:44:44 | SundarN | ok, I'm fine with that | |
| 15:44:53 | zhipeng | great | |
| 15:45:47 | zhipeng | sounds like a great plan for Dublin :) | |
| 15:46:26 | SundarN | Yes, looking fwd to meeting you there :) | |
| 15:47:48 | zhipeng | same here :) | |
| 15:47:52 | zhipeng | #topic AoB | |
| 15:47:59 | zhipeng | any other business | |
| 15:50:06 | zhipeng | Dutch will be in Dublin ? | |
| 15:50:11 | zhipeng | Vipparthy ? | |
| 15:51:21 | Dutch | I think so, waiting on final confirmation. | |
| 15:52:30 | zhipeng | looks like we are gonna have a big team gathering than the previous one :P | |
| 15:53:48 | Li_Liu | I am still waiting for my VISA... no promises... | |
| 15:57:12 | Vipparthy | am not sure, but will certainly check | |
| 15:57:55 | zhipeng | hope all of you could got the VISA on time | |
| 15:59:28 | zhipeng | okey thx everyone for the meeting ! | |
| 16:00:02 | zhipeng | since Queens finished and we will have Chinese new year next week, I will cancel the meeting next week | |
| 16:00:16 | zhipeng | but if you have any questions just shout out on this channel | |
| 16:00:29 | zhipeng | #endmeeting | |
| 16:00:31 | openstack | Meeting ended Wed Feb 7 16:00:29 2018 UTC. Information about MeetBot at http://wiki.debian.org/MeetBot . (v 0.1.4) | |
| 16:00:32 | openstack | Minutes: http://eavesdrop.openstack.org/meetings/openstack_cyborg/2018/openstack_cyborg.2018-02-07-15.00.html | |
| 16:00:33 | openstack | Minutes (text): http://eavesdrop.openstack.org/meetings/openstack_cyborg/2018/openstack_cyborg.2018-02-07-15.00.txt | |
| 16:00:34 | openstack | Log: http://eavesdrop.openstack.org/meetings/openstack_cyborg/2018/openstack_cyborg.2018-02-07-15.00.log.html | |
| 19:25:32 | persia | Good day. The PTL nomination period is ending in a little over 4 hours, and there has still not been a nomination for a Cyborg PTL. I fear I arrive during the wrong timezone, but if anyone is awake and able to nominate someone as PTL, that would be appreciated. | |
| 19:26:10 | persia | My understanding is that this would be the first Cyborg PTL election, so I'd be happy to help with process if folk have questions. | |
| 20:25:13 | crushil | persia, zhipengh[m] already put in his nomination for PTL | |
| 20:25:56 | crushil | Oh | |
| 20:26:11 | crushil | You're talking about going through git | |
| 20:26:12 | persia | I don't see anything at http://git.openstack.org/cgit/openstack/election/tree/candidates/rocky/Cyborg | |
| 20:26:19 | crushil | Makes sense | |
| 20:26:23 | crushil | I can do it for him | |
| 20:26:26 | persia | Yes. That needs doing, or I need to report to the TC that there is no nomination. | |
| 20:26:32 | persia | That would be lovely, if you don't mind. Thanks a lot. | |
| 20:26:48 | crushil | Ya, np | |
| 20:26:54 | crushil | Will let you know when it's done | |
| 20:45:27 | crushil | persia, Done | |
| 20:45:28 | persia | crushil: Found it. Thanks! | |
| 20:45:31 | crushil | :) | |
| 20:51:37 | dims | crushil : thanks! | |
| 20:51:52 | dims | persia : thanks! | |
| #openstack-cyborg - 2018-02-14 | |||
| 06:53:14 | openstackgerrit | Zhipeng Huang proposed openstack/cyborg master: queens release doc fix part 2 https://review.openstack.org/544187 | |
| 07:25:14 | openstackgerrit | Zhipeng Huang proposed openstack/cyborg master: queens release doc fix part 2 https://review.openstack.org/544187 | |
| 07:59:24 | openstackgerrit | Zhipeng Huang proposed openstack/cyborg master: queens release doc fix part 2 https://review.openstack.org/544187 | |
| 08:30:42 | openstackgerrit | Andreas Jaeger proposed openstack/cyborg master: Fix doc build warnings https://review.openstack.org/544208 | |
| 08:31:28 | openstackgerrit | Zhipeng Huang proposed openstack/cyborg master: queens release doc fix part 2 https://review.openstack.org/544187 | |
| 08:32:00 | openstackgerrit | Andreas Jaeger proposed openstack/cyborg master: Remove ChangeLog https://review.openstack.org/544209 | |
| 09:00:50 | openstackgerrit | Merged openstack/cyborg master: Fix doc build warnings https://review.openstack.org/544208 | |
| 09:07:38 | openstackgerrit | Merged openstack/cyborg master: Remove ChangeLog https://review.openstack.org/544209 | |
| 09:07:39 | openstackgerrit | Merged openstack/cyborg master: queens release doc fix part 2 https://review.openstack.org/544187 | |
| 09:44:56 | openstackgerrit | Andreas Jaeger proposed openstack/cyborg master: It's releasenotes - plural https://review.openstack.org/544228 | |
| 10:50:22 | openstackgerrit | Merged openstack/cyborg master: It's releasenotes - plural https://review.openstack.org/544228 | |
| 11:44:18 | openstackgerrit | Zhipeng Huang proposed openstack/cyborg master: queens release doc fix part 3 https://review.openstack.org/544253 | |
| 14:16:46 | Guest73111 | Hi All, My Name is Rahul and i am working on Cyborg for my project. I have some questions on the latest code of cyborg. | |
| 14:17:14 | Guest73111 | I saw FPGA integration code uploaded in Cyborg github | |
| 14:17:32 | Guest73111 | along with a test framework for unit test | |
| 14:18:19 | Guest73111 | My question is, | |
| 14:22:04 | Guest73111 | 1. As i can see, Deployble and accelerator operations happening in two different ways. RPC serive is peridically checking for deployble and once a new get introduced, it is insert or update into database. On the other hand, Accelerator operations solely dependednt on curl command. How this two are related. I cant see any connection in codebase between Deployble and accelerator operations | |
| 14:24:03 | Guest73111 | 2nd question, How program function of IntelFPGADriver will get called. In test framework, IntelFPGADriver getting mocked and using that obj, program function is getting called. But in codebase, there is no call to the fpga driver's function | |
| 14:24:29 | Guest73111 | How, it will get called and from which component? is it agent? | |
| 14:32:00 | zhipengh[m] | Hi Rahul | |
| 14:33:47 | zhipengh[m] | For question 1, accelerator refers to the data model of the basic functionality of an accelerator card , for example a typical PF | |
| 14:34:04 | zhipengh[m] | Deployable refers to the VFs | |
| 14:35:55 | zhipengh[m] | That is why you see RPC calls for deployable (called by cyborg internal modules ) , and curl for accelerator (called by outside world ) | |
| 14:37:14 | zhipengh[m] | Per question 2, the program functionality is not supported in Queens :) and will be in Rocky | |
| 14:39:16 | Guest73111 | First i have a question for the last answer, As you said, queens will not support program functionality. Then, how a function ( which is written inside IntelFPGADriver), will get called? | |
| 14:39:51 | Guest73111 | i mean, how you are planning to trigger any function (written inside intelFPGADriver) | |
| 14:41:10 | Guest73111 | Let say, i want to use discover function written in intelFPGSDriver.. How it will get called? There is no call in current code, other than call from test framework | |
| 14:42:28 | zhipeng | so for our current code, we treat Intel FPGA card as a QAT card | |
| 14:42:45 | zhipeng | which is just a PCIe card with pre-programed logic | |
| 14:43:14 | zhipeng | so there will be no discovery of the functions, so to speak | |
| 14:43:33 | zhipeng | you could use the driver to discover the overall metadata of the whole card | |
| 14:47:03 | Guest73111 | Thanks Zhipeng for the prompt answer. This is so far understood. | |
| 14:47:49 | zhipeng | Guest73111 no problem :) | |
| 14:49:11 | Guest73111 | what i try to understand is in the future, if i use cyborg api(which just represent accelerators), how this will be interacting with the deployables. | |
| 14:53:57 | zhipeng | Guest73111 so we plan to support programability in Rocky, that means user should be able to call cyborg-api to initiate a new VF in a FPGA card | |
| 14:55:11 | zhipeng | so from api point of view, there should be no differency | |
| 14:55:20 | zhipeng | the user just created a new acceleration resource | |
| 14:55:56 | zhipeng | however the difference will be reflected on the data model, which means deployable tables will represent this newly created VF | |
| 14:57:57 | openstackgerrit | Zhipeng Huang proposed openstack/cyborg master: queens release doc fix part 3 https://review.openstack.org/544253 | |
| 15:12:28 | NokMikeR | Hi | |