================================ #opnfv-vswitchperf: Vswitchperef ================================ Meeting started by SRao at 16:04:44 UTC. The full logs are available at http://ircbot.wl.linuxfoundation.org/meetings/opnfv-vswitchperf/2018/opnfv-vswitchperf.2018-12-05-16.04.log.html . Meeting summary --------------- * Development Update (SRao, 16:04:55) * Added a patch to affinitize stressor-vm cpus (SRao, 16:05:16) * added a patch to fix variable name. (SRao, 16:05:30) * Al Mort9on (acm, 16:06:47) * Al Morton (acm, 16:07:04) * include 'before' and 'after' images for affinitize patch. (SRao, 16:07:12) * ACTION: all need reviews on (acm, 16:08:26) * IEEE NFVSDN (SRao, 16:08:55) * 3-hr tutorial on dataplane performance benchmarking, tried to convince the attendees to participate in these projects. (SRao, 16:09:31) * Majority of the participants from academia. VSPERF can be used as part of the research activity. (SRao, 16:10:07) * NFVBench Demo - Trex reported -ve dropped packets. (SRao, 16:10:46) * If someone can create a vagrantfile it would help users to try vsperf on laptop (SRao, 16:12:51) * I think that the method of reading TX and Rx counters separately may be part of the issue (with -Trex) (acm, 16:13:16) * - no vagrant experience for me, good question about is it still cool... (acm, 16:14:05) * Collectd reporting the l3-cache values - has issues of staledata, and timestamps. (SRao, 16:15:30) * I'd like to use collectd to read interface tx/rx packet counters periodically, can it do that? (acm, 16:15:56) * Containerizing VSPERF. (SRao, 16:16:44) * should we put the contents in existing repo or should we create a new repo - vswitchperf-containers? (SRao, 16:17:34) * I'd like to9 get some advice from Eddie Arrage or Alec H (acm, 16:17:45) * It would be a quick e-mail (acm, 16:18:13) * If existing one - create a folder containers, and within that we need 2 folders one for vsperf and results. (SRao, 16:18:50) * ACTION: acm write e-mail to Eddie Arrage or Alec H to get their suggestion on Container->VSPERF (acm, 16:19:02) * first define the messages to be exhanged (acm, 16:20:24) * we have four candidate rpcs to support - minimal - start tests and check status (acm, 16:20:56) * Install - upload a config file - start tests and check ststus (acm, 16:22:28) * I think we need results too (acm, 16:22:52) * reuest feedback on this (acm, 16:23:25) * that's good! (acm, 16:26:40) * ACTION: Sridhar to send .proto to AL. (SRao, 16:27:03) * ACTION: Sridhar to create results container. (SRao, 16:27:26) * - ok, I was imagining a high-level question first, and see where that goes (acm, 16:27:28) * ok, thanks! (acm, 16:29:02) * https://wiki.opnfv.org/display/DEV/Intern+Project%3A+VSPERF+Automated+Deployment+and+Results+Management (SRao, 16:29:17) * yes, I'll add that to the discussion. (acm, 16:29:37) * Eddie is the right guy! (acm, 16:29:51) * not yet (acm, 16:30:13) * is possible to results push to Acumos ? (acm, 16:30:45) * I would assume our VES guy QAlok Gupta knows (acm, 16:31:17) * so question: can any project push results data to Acumos today??? (acm, 16:32:18) * Yes, AT&T , but I don't have any contact with them, directly, so far (acm, 16:33:46) * AL: Started work on Long-Duration testing. (SRao, 16:36:32) * ACTION: Need better support of 'integrating' iperf (non-dpdk) tools. (SRao, 16:38:17) * LDT: Early Stopping, based on monitoring some metrics (SRao, 16:41:15) * OPNFV Strategy - VSPERF. (SRao, 16:42:32) Meeting ended at 16:59:30 UTC. Action items, by person ----------------------- * acm * acm write e-mail to Eddie Arrage or Alec H to get their suggestion on Container->VSPERF * **UNASSIGNED** * all need reviews on * Sridhar to send .proto to AL. * Sridhar to create results container. * Need better support of 'integrating' iperf (non-dpdk) tools. People present (lines said) --------------------------- * acm (31) * SRao (22) * bfubel (3) * collabot` (3) Generated by `MeetBot`_ 0.1.4