14:04:11 #startmeeting Rally 14:04:12 Meeting started Mon Aug 10 14:04:11 2015 UTC and is due to finish in 60 minutes. The chair is boris-42. Information about MeetBot at http://wiki.debian.org/MeetBot. 14:04:14 Useful Commands: #action #agreed #help #info #idea #link #topic #startvote. 14:04:16 The meeting name has been set to 'rally' 14:04:37 ping andreykurilin rvasilets amaretskiy 14:04:42 hi 14:05:00 \o/ 14:05:20 o/ 14:05:30 o/ 14:05:31 ~o~ 14:05:52 sup 14:05:58 hi 14:06:22 o/ 14:06:28 sorry for late 14:06:33 remember wrong time 14:08:04 okay let's start=) 14:08:15 #topic todays agenda 14:08:51 #link https://wiki.openstack.org/wiki/Meetings/Rally 14:09:15 if somebody wants to get some meeting minutes be free to add extra points in todays meeting agenda 14:09:35 #topic [amaretskiy] New classes for HTML report charts 14:09:59 amaretskiy: I believe you don't need 3 points of meeting agenda 14:10:00 #link https://review.openstack.org/#/c/159458 14:10:05 yes 14:10:10 to discuss one piece of work 14:10:43 there are 3 patches related to reports, within today's agenda, and new classes for HTML reports is the 1st one 14:10:53 it is ready for review 14:11:22 please review it, I believe it is close to one we want.... 14:11:48 eom 14:12:05 okay 14:12:19 amaretskiy: please rewrite the agenda 14:12:40 amaretskiy: add one point and put info like close to done review required 14:13:06 okay 14:13:10 #topic [kun] vm workload 14:13:17 kun_huang: hey there 14:13:19 okay 14:13:31 actually I need vm workload recently 14:13:43 single vm workload mode is close to finish 14:13:51 only one patch is on reviewing 14:14:05 https://review.openstack.org/#/c/183228 14:14:21 but I need rally support multiple vm workload 14:14:26 kun_huang: actually pboldin is workin on that 14:14:33 kun_huang: already for a long amount of time 14:14:41 "multiple vm workload"? 14:14:44 kun_huang: yep 14:14:50 kun_huang: that is why we need all his patches 14:15:03 good 14:15:07 kun_huang: we are preparing rally step by step for multple vm workload 14:15:12 I'm using and testing his patches 14:15:35 why can't we use boot_runcommand_delete? 14:15:36 what's the next step now? 14:16:12 redixin: it works for single vm 14:16:37 some time, we need spawn 2 vm with iperf script 14:17:00 that's another kind of benchmark 14:17:14 vm a need to know how to access vm b 14:17:19 kun_huang: the next step is to create servers_ext 14:17:23 kun_huang: context 14:17:34 kun_huang: that will create various topologies of services 14:17:45 make sense 14:18:00 kun_huang: after that we will need to create small simple rally agent 14:18:11 the custom image is really like that 14:18:17 custom image context 14:18:18 kun_huang: and we will have server_ext and custome_images to create env 14:18:31 kun_huang: nope custome_image creates single image 14:18:51 kun_huang: that will be used by servers_ext to run N vms with this image with proper networking 14:18:52 custome_image without stop server 14:19:01 okay 14:19:01 kun_huang: nope 14:19:04 that's better 14:19:11 kun_huang: custom_image creates snapshot that is used later 14:19:26 servers_ext creates real servers that will run workloads 14:20:02 got it 14:20:51 kun_huang: ok any other questions? 14:20:54 iperf is kinda special case 14:21:00 we need N pairs of servers 14:21:04 no more yet 14:21:12 redixin: yes 14:22:00 redixin: it's not speical case 14:22:37 redixin: it's the thing that will be like any other multi vm workload 14:23:46 does vm need to know ip addresses of each other in any vm workload? 14:24:14 redixin: I beleive usually they should know 14:24:28 boris-42, anyways i would like to see a task configuration file first 14:24:48 Кredixin task configuration will be the same 14:24:59 redixin: just pick some contexts/sceanrio/runner 14:25:12 redixin: and they together will build proper env and trigger it 14:26:17 ok let's move on 14:26:58 yep, let's move on fist boris-42 14:27:00 redixin: #link https://docs.google.com/document/d/1FQicnj6aIdbfYts9ofYdz0dUEf9zyssWqDjQR40h268/edit 14:27:16 #topic [kun] what's the state of Solaris support? 14:27:25 oh I haven't time to do this =( 14:27:36 hah, okay 14:27:44 I found the old patch is still here 14:27:46 I will try to find time this week to create solaris cloud image 14:27:50 I've been off for the past couple of weeks so haven't looked at anything either. I'd meant to ping you Boris 14:28:07 albertw: sorry for waiting=( 14:28:13 No problem! 14:28:16 I was moving from one appramtnets to another 14:28:21 it was a bit crazy =) 14:28:52 well drop me an email or ping me if you need a hand 14:29:07 albertw: ok 14:29:28 so moving to next topic 14:29:34 #topic [ikhudoshyn] Refactor BenchmarkEngine 14:29:39 ikhudoshyn: hi there 14:29:56 just a quick note about the patch 14:30:10 it gained lots of +'s there 14:30:36 but I updated it a bit to address a good comment from Chris St.Pierre 14:30:53 so pls re-review it 14:30:55 eom 14:31:01 https://review.openstack.org/#/c/209046/ btw 14:31:29 ikhudoshyn: could you provide some info for others 14:31:35 ikhudoshyn: what does this patch 14:32:13 it extracts thread mungling from BenchmarkEngine to a separate context manager 14:32:47 does not introduces any new functionality, olny increases readability 14:33:00 that's pretty much iy 14:33:10 *it* 14:33:18 ikhudoshyn: ok great 14:33:26 thanks 14:33:31 #topic open discussion 14:33:37 do we have something to discuss? 14:34:21 Right way of submitting links #link https://review.openstack.org/#/c/209046/ 14:34:33 nothing from me 14:35:22 others?) 14:35:33 no more topics 14:35:48 no ) 14:36:04 no 14:37:44 okay 14:37:55 *all see you next week) 14:38:01 #endmeeting