13:57:31 #startmeeting FD.io CSIT project meeting 13:57:31 Meeting started Wed May 6 13:57:31 2020 UTC and is due to finish in 60 minutes. The chair is mackonstan. Information about MeetBot at http://wiki.debian.org/MeetBot. 13:57:31 Useful Commands: #action #agreed #help #info #idea #link #topic #startvote. 13:57:31 The meeting name has been set to 'fd_io_csit_project_meeting' 13:57:35 #chair 13:57:36 Current chairs: mackonstan 13:57:41 #info Tibor Frank 14:03:46 #topic Agenda bashing 14:03:51 #info Peter Mikus 14:03:53 #info Jan Gelety 14:03:57 #info Dave Wallace 14:04:04 #topic Physical labs, testbeds 14:05:25 #info Vratko Polak. 14:05:44 #info SSDs keep dying: 14:05:44 - s34-t27-sut1 - SSD failure. Reported. 14:05:44 - s12-t31-sut2 - SSD failure. Reported. 14:05:44 - s37-t28-sut1 - SSD failure. Reported. 14:05:44 - s3-t21-sut1 - SSD failure. Reported. 14:07:03 #info Peter - going forward proposing to put a remedy in place: 14:07:03 - Implementing /tmp as ramdisk to decrease writing operations. 14:07:03 - Next step will be /usr/bin as ramdisk. 14:07:26 #info Peter - to date 18 SSD disks failed 14:09:17 #info Vexxhost tickets open, but no spares. RMA process unclear, so not shipped. 10 new spare drives still didn't reach the site. 14:12:56 #info Maciek, Peter - Review the SSD drive situation on the next TSC call, propose as a backup plan replacing SSDs with spinning HDDs next time particular SSD fails. 14:14:56 #info right now following testbed types are out of service: 1x 2n-skx, 1x 3n-skx, 2x 2n-cls 14:16:42 #info Peter - SSD RCA - will provide detailed info into existing email thread with ray.kinsella@intel 14:16:55 #topic Inputs from projects 14:17:07 #info VPP f0 milestone today 14:19:36 #info VSAP - Dave - gigantic patch (3k LoCs, many tests, overlap with infra) pushed into CSIT repo, asked to split into separate patches to make it reviewable and coordinate from there via gerrit, email etc. 14:20:13 #topic Releases - CSIT-2001 14:21:19 #info 2n-skx, 3n-skx diffs RCAs - Vratko - out of 5, 3 identified, 2 to be identified. One bisecting job is killing SSDs and testbeds. 14:25:31 #info voting for shipping skx RCA content, announce email before next VPP community call on 12-May, and presenting it on that call given allocation of 15min slot 14:25:52 #info 11 x +1 - everyone on the call here - voted for it 14:27:47 #topic Releases - CSIT-1908.2 14:30:47 #info report status - Tibor - report is prepared in the local dev sandbox, will push to csit rls1908.2 branch, missing rls notes so they need to be completed 14:31:13 #topic Release - CSIT-2005 14:32:32 #info wiki page is still to be updated - Maciek 14:32:42 #info Dave - will add directly into wiki 14:34:17 #link https://wiki.fd.io/view/CSIT/csit2005_plan 14:34:45 #topic Operations 14:41:30 #info Dave - Nomad update: Vanessa shared Nomad Jenkins Plugin, DAW to investigate converting to JJB YAML config 14:42:23 #info Nomad servers now have uniform admin accounts 14:43:18 #info Nomad Dockerfiles have been imported into ci-management repo 14:43:59 Nomad Dockerhub now has an account fdiotools which will be used to publish docker images 14:44:20 #info Nomad Dockerhub now has an account fdiotools which will be used to publish docker images 14:46:50 #info Nomad servers currently using random bare-metal OS's (ubuntu 16.04 / 18.04, etc). Plan moving forward is to use Ansible to manage uniform OS installations 14:50:35 #info Vanessa testing changes to using lf global macros on build images on Sandbox and opening JIRA tickets for issues encounter 14:53:21 #info Maciek - follow up with TWS meeting to discuss above: Dave, Peter, Maros, Maciek, Vratko. 14:53:41 #info Maciek - target Thu or Fri next week. 14:54:03 #topic VPP code performance 14:54:55 #info vpp bihash patches impact 14:55:01 #info Vratko - ip6 is back 14:56:15 #info Vratko - ipsec - fixes on csit side, failures caused by changes in vpp, but required changes on csit to do the right thing, some high scale tests are still failiing. 14:57:15 #info Vratko - vpp changed the default compiler from gcc to clang - introduced throughput regression on hsw 14:59:16 #topic Developments 14:59:45 #info Scapy licensing - Vratko - step1 merged 15:00:20 #info Vratko - step2 - contributors' statements 15:01:02 #info emails need to be sent - execution ownership to be reviewed on the next TSC call - ray.kinsella and TSC lead / Ed Warnicke are in the driving seat. 15:01:46 #endmeeting