#fdio-csit: FD.io CSIT Project Meeting

Meeting summary

1. Agenda bashing (mackonstan, 14:58:56)
   1. Tibor Frank (tifrank, 15:00:52)
   2. Tom Herbert (tfherbert, 15:01:11)
   3. Jan Gelety (jgelety, 15:02:23)
   4. Tom Herbert - slack account requested. (tfherbert, 15:02:34)
   
   
2. Inputs from LFN and FD.io projects (mackonstan, 15:03:38)
   1. VPP - no updates. (mackonstan, 15:04:02)
   2. HC - Michal - updates to ACL modules, CSIT tests fixed and now passing. (mackonstan, 15:04:34)
   3. HC - Michal - DHCP tests started to fail, t/shooting if VPP or HC issue. (mackonstan, 15:05:03)
   4. SweetComb - no updates. (mackonstan, 15:05:30)
   5. SweetComb - Michal - no initial code in the repo, at least last week that was the case. (mackonstan, 15:06:09)
   6. DMM - Sharath - patch that needs review. (mackonstan, 15:06:35)
   7. https://gerrit.fd.io/r/#/c/15547/ (tifrank, 15:06:56)
   8. TSC - no updates (mackonstan, 15:09:13)
   9. TSC - arm install progressing after escalation on TSC calls (mackonstan, 15:10:13)
   
   
3. CSIT operational status (mackonstan, 15:10:14)
   1. LF infra - systems were down last week, but then fixed. Jenkins was down earlier today, now back. (mackonstan, 15:11:08)
   2. performance trending - Tibor - jobs failing, errors generated: out of space, some other java errors, FD.io helpdesk ticket open. (mackonstan, 15:14:14)
   3. FD.io Helpdesk #64031 (tifrank, 15:14:46)
   4. performance trending - switching to Slack (mackonstan, 15:15:18)
   
   
4. VPP code performance (mackonstan, 15:15:35)
   1. trending jobs down.. (mackonstan, 15:15:52)
   
   
5. CSIT Releases: 18.10.w47 (mackonstan, 15:16:12)
   1. https://docs.fd.io/csit/rls1810/report/index.html (mackonstan, 15:16:52)
   2. quite a few changes from last week - Atom Denverton results added, more comparisons (mackonstan, 15:17:59)
   3. https://docs.fd.io/csit/rls1810/report/introduction/report_history.html (mackonstan, 15:18:44)
   4. CSIT report versioning discussion (mackonstan, 15:29:07)
   5. https://lists.fd.io/g/csit-dev/message/3236 (mackonstan, 15:29:10)
   6. Ray, Tom, Vratko having comments (mackonstan, 15:31:25)
   7. Ray - issues with docs.fd.io report performance - rendering time in browser is very long, 30..40sec (mackonstan, 15:33:04)
   8. interactive graph take time to be rendered using js (mackonstan, 15:37:17)
   9. Ray - trading flexibility with interactive graphs vs. user experience (mackonstan, 15:38:00)
   10. need a solution, few options discussed (mackonstan, 15:39:00)
   11. Ray: opt-1 graph .png, and only on click graph dynamic js graphs (mackonstan, 15:39:24)
   12. Tibor: opt-2 split js graphs into smaller groups, currently some sections have up to 72 graphs, plotted from top to bottom, usually (mackonstan, 15:40:33)
   13. Peter: opt-3 use dash to render less graphs, but allow more dynamic views to look at the data (mackonstan, 15:43:26)
   14. ACTION: Tibor Maciek Ray Peter - work on evaluating optimization options, and delivering them in 1901 (mackonstan, 15:45:25)
   
   
6. CSIT-19.01 plan (mackonstan, 15:47:55)
   1. draft plan created on wiki (mackonstan, 15:48:49)
   2. https://wiki.fd.io/view/CSIT/csit1901_plan (mackonstan, 15:48:54)
   
   
7. CSIT Developments (mackonstan, 15:52:41)
   
8. TRex (mackonstan, 15:53:00)
   
9. FD.io CSIT lab infrastructure (mackonstan, 15:53:28)
   1. Ray - TG system and interface allocation for Atom Denverton for the upcoming new setup (mackonstan, 15:54:27)
   2. https://wiki.fd.io/view/CSIT/Testbeds:_Xeon_Skx,_Arm,_Atom.#Testbeds_Type_Breakdown (mackonstan, 15:56:50)
   3. Juraj - New Cavium ThunderX machine should be coming up this week, for VPP_Device 1-Node tests (mackonstan, 15:58:42)

Action items

1. Tibor Maciek Ray Peter - work on evaluating optimization options, and delivering them in 1901

People present (lines said)

1. mackonstan (42)
2. collabot` (3)
3. tifrank (3)
4. tfherbert (2)
5. jgelety (1)