| Posted | Nick | Remark | |
|---|---|---|---|
| #openstack-cyborg - 2019-05-29 | |||
| 03:58:36 | openstack | Log: http://eavesdrop.openstack.org/meetings/openstack_cyborg/2019/openstack_cyborg.2019-05-29-03.03.log.html | |
| 03:59:14 | Li_Liu | Have a good night Sundar | |
| 03:59:24 | ikuo_ | Bye, Sundar and all | |
| 04:58:59 | xinranwang | Hi guys, I have asked about the IRC tools. It's called ZNC. Here is the tutorial: https://fedoramagazine.org/never-leave-irc-znc/ | |
| 06:46:08 | Li_Liu | Wanna ask a CI/CD related question, do we have plan to have dummy device drivers for CI/CD? | |
| 12:22:17 | edleafe | I was reading the minutes from the meeting last night (well, last night for me), and xinranwang asked about a tool for continuous IRC. These are called "bouncers", and There is a free tool called 'znc' that works great. You set it up on a machine that's always on. It connects to IRC, and your IRC client connects to the znc instance. It will play back any messages you missed while disconnected. | |
| #openstack-cyborg - 2019-05-30 | |||
| 03:10:10 | Sundar__ | @Li_Liu: Re. fake drivers aka dummy devices, I have had different feedback. In the Nova spec, I put in both. However, we haven't got an owner yet. So, I am thinking of focusing on real hardware | |
| 03:11:37 | Sundar__ | #link https://etherpad.openstack.org/p/cyborg-ptg-train Liness 157-160 | |
| 03:12:43 | Sundar__ | @edleafe: Thanks. Do you know of any good free IRC bouncers for Android? | |
| 12:36:12 | edleafe | Sundar__: you don't run the bouncer on your client; you run it on a server. You would set your Android client to connect to the server running the bouncer instead of connecting to Freenode. | |
| #openstack-cyborg - 2019-05-31 | |||
| 00:44:35 | Sundar | @all, There is a discussion about whether we should use asserts in https://review.opendev.org/#/c/626057/10/cyborg/accelerator/drivers/fpga/intel/driver.py | |
| 00:45:51 | Sundar | First, raising an exception is appropriate f something is wrong with the user input or other subsystem's behavior, but asserts are good to catch internal program logic, invariants that should always ok, etc. Using asserts for these will prevent proliferation of exceptions. | |
| 00:46:08 | Sundar | Second, there are many precedents for using exceptions. E.g. | |
| 00:46:19 | Sundar | [1] https://github.com/openstack/nova/blob/master/nova/rpc.py#L97 | |
| 00:46:28 | Sundar | [2] https://github.com/openstack/nova/blob/master/nova/virt/disk/mount/api.py#L187 | |
| 00:46:37 | Sundar | [3] https://github.com/openstack/nova/blob/master/nova/objects/aggregate.py#L441 | |
| 01:09:46 | Sundar | @wangzhh I don't understand your response in https://review.opendev.org/#/c/659014/1/cyborg/tests/unit/db/utils.py Line 106 | |
| 01:10:08 | Sundar | The other comment referred to using the lspci command. | |
| 01:15:06 | wangzhh | For we use string as cpid_info, if we split it an convert to json, {"device": "2", "bus": "00", "function": "01"} is not enough. | |
| 01:15:59 | wangzhh | It should looks like {"device": "2", "bus": "00", "function": "01", "domain": "0000"} | |
| 01:17:28 | wangzhh | Am I clear? | |
| 01:19:33 | wangzhh | The cpid_info looks like 0000:00:02.0. The json_str in this patch lost domain. | |
| 01:20:14 | Coco_gao | yes, if cpid_info is a json string, what about the attach_info? | |
| 01:22:10 | wangzhh | I think it is the same. | |
| 01:24:05 | Coco_gao | And cpid_info is used to identify some of the devices, if its a json_str, then it will be hard to identify, because it' s hard to compare json_str. | |
| 01:25:26 | Coco_gao | what do you think? | |
| 01:25:39 | Coco_gao | @Sundar | |
| 01:27:23 | wangzhh | Sundar said he'll BRB in 30 min. | |
| 01:31:43 | Coco_gao | so will I | |
| 02:08:13 | Sundar | @Coco_gao @wangzhh Yes, domain should be included. attach_id also should be fixed in the same way: I'll do that. All good? | |
| 02:12:37 | Sundar | @Coco_gao: "if its a json_str, then it will be hard to identify, because it' s hard to compare json_str" Are you saying that we can store it as individual PCI BDF fields (domain, bus, etc.) in the db? That would not work for PowerVM, which represents PCI devices differently. | |
| 02:14:43 | wangzhh | Fine. Sundar. | |
| 02:16:38 | wangzhh | And do u install znc which xinran mentioned? I don't find windows version. | |
| 02:17:18 | Sundar | wangzhh: znc is linux only | |
| 02:18:07 | wangzhh | OK. Thx. | |
| 02:43:18 | Sundar | @yikun, @wangzhh, @Coco_gao and all, please respond to http://eavesdrop.openstack.org/irclogs/%23openstack-cyborg/%23openstack-cyborg.2019-05-31.log.html#t2019-05-31T00:45:51 | |
| 07:29:24 | wangzhh | @yikun Is generic driver ready? New GPU driver depends on it. Or u just responsible for spec, I'm not sure. | |
| #openstack-cyborg - 2019-06-02 | |||
| 06:46:45 | openstackgerrit | Brin Zhang proposed openstack/cyborg-specs master: Change openstack-dev to openstack-discuss https://review.opendev.org/622031 | |
| 19:53:10 | Sundar | FYI: https://review.opendev.org/#/c/662643/ "Categorize Cyborg under Accelerators and update description" | |
| #openstack-cyborg - 2019-06-04 | |||
| 05:12:52 | Sundar | @all, Shall we have a Zoom call this Thursday China time (Wed in Americas)? Agenda: Review the pending specs and patches. | |
| 05:17:21 | Sundar | Do we need the spec https://review.opendev.org/#/c/593726? We have already implemented discovery, right? | |
| 05:27:47 | ikuo_o | I can attend Thursday zoom meeting. | |
| 05:33:00 | Sundar | Thanks, ikuo_o | |
| 13:41:39 | Li_Liu | should be fine for me to attend the zoom meeting | |
| 15:13:06 | Sundar | Thanks, Li_Liu | |
| 15:30:38 | Li_Liu | Sundar, do we still have the irc meeting tonight then? | |
| 15:56:06 | Sundar | Li_Liu: Yes | |
| 18:50:27 | openstackgerrit | Donny Davis proposed openstack/cyborg master: Updating Index to reflect dependent projects and provide more information on Cyborg https://review.opendev.org/663121 | |
| #openstack-cyborg - 2019-06-05 | |||
| 03:00:48 | Sundar | Hi Biwei | |
| 03:00:55 | Biwei | Hi Sundar | |
| 03:01:09 | Sundar | Hi wangzhh | |
| 03:01:16 | wangzhh | Hi Sundar. | |
| 03:01:22 | ikuo_o | Hi Sundar | |
| 03:01:27 | Sundar | Hi ikuo_o | |
| 03:01:34 | wangzhh | Hi ikuo_o | |
| 03:01:44 | Sundar | Let's wait a min for others to join | |
| 03:01:49 | ikuo_o | Hi wangzhh | |
| 03:03:04 | Sundar | #startmeeting openstack-cyborg | |
| 03:03:05 | openstack | Meeting started Wed Jun 5 03:03:04 2019 UTC and is due to finish in 60 minutes. The chair is Sundar. Information about MeetBot at http://wiki.debian.org/MeetBot. | |
| 03:03:06 | openstack | Useful Commands: #action #agreed #help #info #idea #link #topic #startvote. | |
| 03:03:08 | openstack | The meeting name has been set to 'openstack_cyborg' | |
| 03:03:16 | Sundar | #topic Roll call | |
| 03:03:36 | Sundar | #info Sundar | |
| 03:04:24 | Sundar | wangzhh, ikuo_o, Biwei: o/ | |
| 03:04:34 | yikun | #info yikun | |
| 03:04:41 | wangzhh | #info wangzhh | |
| 03:04:45 | ikuo_o | #info ikuo_o | |
| 03:04:47 | Sundar | Hi yikun | |
| 03:04:56 | Biwei | #info Biwei | |
| 03:04:58 | yikun | hey, Sundar | |
| 03:05:07 | Sundar | Agenda: https://wiki.openstack.org/wiki/Meetings/CyborgTeamMeeting#Agenda | |
| 03:05:13 | Sundar | ANything to add to that? | |
| 03:06:05 | Sundar | #topic Python cyborg client decision | |
| 03:06:10 | wangzhh | Coco said she has other bussiness. Will miss this meeting. | |
| 03:06:24 | Sundar | Please see the thread starting from http://lists.openstack.org/pipermail/openstack-discuss/2019-May/006543.html | |
| 03:06:41 | Sundar | wangzhh, thanks for letting us know | |
| 03:07:38 | Sundar | For the client, we have 2 options: write our own plugin (as many other projects seem to have done) or integrate with openstackclient repo (less work, but need to depend on the repo owners for reviews) | |
| 03:08:56 | Sundar | For the 2nd option, there are some guidelines and rules, which seem ok to me. But we have to discuss with the repo owners about their opinions | |
| 03:08:57 | xinranwang | Hi all | |
| 03:09:04 | xinranwang | #info xinranwang | |
| 03:09:11 | Sundar | Hi xinranwang | |
| 03:09:34 | Sundar | Writing a plugin may not be that bad, given the previous examples | |
| 03:09:39 | Sundar | What do you all think? | |
| 03:09:58 | yikun | So, we have to choose one from these two option? | |
| 03:10:46 | Sundar | yikun, do you have any other option? These seem like the 2 ways that are recommended or have precedents. We could do neither and write our own client -- but that is not recommended | |
| 03:10:47 | wangzhh | yikun, I think so. Do u have other suggestions? | |
| 03:10:56 | xinranwang | I think writing our own plugin is better | |
| 03:11:16 | ikuo_o | Thanks for summarizing. | |
| 03:11:26 | xinranwang | More indenpendency | |
| 03:11:34 | wangzhh | Yes, I also vote the first one. Agree with xinran. | |
| 03:12:31 | ikuo_o | It is better to write plugin, but I heard the option is not prepared by OSC. | |
| 03:12:48 | ikuo_o | Maybe my misunderstanding... | |
| 03:12:49 | yikun | OK, just want to sure any other way to complete, and I also vote to the 1st | |
| 03:12:58 | Sundar | ikuo_o: "option is not prepared by OSC" -- what do you mean? | |
| 03:13:33 | Sundar | OSC folks are not recomending against plugin -- they gave us the options. | |
| 03:13:52 | xinranwang | please also consider the case where cyborg can run as a standalone project | |
| 03:14:24 | Sundar | xinranwang: Sure, both options should be compatible with standalone usage | |
| 03:15:17 | ikuo_o | I see. I thought the plugin option is difficult for OSC side, but maybe my misunderstanding | |
| 03:15:17 | ikuo_o | If it is realized in other works, plugin is better I think. | |
| 03:15:19 | wangzhh | Sundar, do u have any example of these two options? Maybe two different projects with two options. | |