09:00:45 #startmeeting large_scale_sig 09:00:45 Meeting started Wed May 15 09:00:45 2024 UTC and is due to finish in 60 minutes. The chair is ttx. Information about MeetBot at http://wiki.debian.org/MeetBot. 09:00:45 Useful Commands: #action #agreed #help #info #idea #link #topic #startvote. 09:00:45 The meeting name has been set to 'large_scale_sig' 09:00:59 ping amorin 09:01:11 I think Felix mentioned he would not be able to make it today 09:01:27 #topic Rollcall 09:01:45 songwenping_: welcome! 09:02:03 ttx: thanks 09:02:34 While we wait for others to join, could you introduce yourself? What are you working on? Why are you interested in the large scale SIG? 09:03:27 i am Cyborg PTL :) 09:04:29 You told me you were interested in helpign with the Large Scale doc, especially the performance side 09:04:33 my company Inspur recently have some optimization on large scale test. 09:05:17 and we want contribute it to the Large Scale doc. 09:05:55 Looks like amorin missed the meeting reminder and it will just be the two of us today. The others will catch up reading the meeting logs 09:06:03 #topic Large Scale doc 09:06:23 Let's discuss in more details what you want to optimize on the docs side 09:06:48 #link https://docs.openstack.org/large-scale 09:07:40 songwenping_: Do you think that would fit in the existing pages, or we would create somethign specific around performance? 09:09:07 i am not sure, first i will tell you some results. 09:09:14 yes please! 09:11:44 songwenping_: can you share the results now? 09:12:05 please wait 09:12:23 a few seconds. 09:13:42 we support 3000 compute nodes in one aggregate and Concurrent create 3000 vms and the success rate is 100%. 09:15:17 sounds pretty good to me 09:16:49 the current record is 1000 compute nodes? 09:17:58 I don;t think we really have a record, but we certainly discouraged larger aggregates as you'd get delays / failures 09:19:16 Maybe the right approach you should write a document describing all your findings, and then we can discuss how to best integrate it in the doc 09:19:59 One approach might be to integrate it into existign documents, another approach would be to create a separate "performance" specific document 09:20:08 this is our actual use case. 09:20:33 But it's hard to choose between the two approaches until we see the doc, so maybe that's the first step 09:21:28 When ready you can post it to the openstack-discuss mailing-list with the [largescale-sig] prefix so that we atre sure to see it 09:21:36 are* 09:22:01 and then we'll read it and discuss how to integrate it at the next meeting? 09:22:15 i prefer to create a seperate doc as it's easy to review. 09:23:45 OK, the format is up to you (could be etherpad, could be proposed Gerrit change against large-scale git repo, could be email...). Just let us know by email when ready :) 09:24:20 sure, i will post the the mail first. 09:24:26 Perfect 09:24:47 I want amorin and felix.huettner 's eyes on it before we get too far in discussing where to integrate it 09:25:06 They are the experts, I just play one on Mastodon 09:25:32 songwenping_: anything you'd like to add? 09:25:51 haha, nothing else. 09:26:58 Alright. I skipped the discussion about the the next OpenInfra Live episode since we are missing other members of the SIG 09:27:07 #topic Next meeting(s) 09:27:59 With our new schedule the next meeting will be on IRC at 9:00 UTC on June 19 09:28:19 songwenping_: I'm assuming you will be able to post your doc for review before then? 09:28:42 no problem 09:28:46 #action songwenping_ to post his performance findings doc for review ahead of next meeting 09:28:56 Perfect! 09:29:07 #info next meeting will be on IRC at 9:00 UTC on June 19 09:29:12 #topic Open discussion 09:29:33 Anything else to discuss? Sorry we did not get more people today, there is usually at least 3 of us :) 09:31:07 OK, if nothing else, we'll close early. Thanks songwenping_ for joining and sharing your findings! 09:31:45 #endmeeting