#opnfv-doctor: doctor
Meeting started by r-mibu at 14:02:17 UTC
(full logs).
Meeting summary
- roll call (r-mibu, 14:02:27)
- Carlos Goncalves (cgoncalves,
14:02:29)
- Yujun Zhang (yujunz,
14:02:35)
- Tomi Juvonen (tojuvone,
14:02:59)
- Gerald (GeraldK,
14:03:07)
- Ryota Mibu (r-mibu,
14:03:20)
- Pedro Sousa (psous,
14:03:27)
- https://gerrit.opnfv.org/gerrit/#/c/24143/
(yujunz,
14:06:04)
- Host specific VMs list (yujunz,
14:06:42)
- Q4 hackfest (r-mibu, 14:09:31)
- https://etherpad.opnfv.org/p/doctor_meetings
(r-mibu,
14:10:05)
- yujunz and wenjuan will not attend hackfest,
but Julien might attend (yujunz,
14:11:29)
- Inspector design guideline (r-mibu, 14:16:47)
- Larry Lamers (ljlamers,
14:16:49)
- Tommy Lindgren (tommyL,
14:16:57)
- http://artifacts.opnfv.org/doctor/review/24143/design/index.html#document-inspector-design-guideline
(yujunz,
14:17:44)
- https://gerrit.opnfv.org/gerrit/#/c/24143/
(cgoncalves,
14:18:02)
- https://jira.opnfv.org/browse/DOCTOR-76
(yujunz,
14:19:15)
- https://lists.opnfv.org/pipermail/opnfv-tech-discuss/2016-October/013036.html
(yujunz,
14:19:32)
- r-mibu suspects that notifier may have taken
most time of the whole doctor test (yujunz,
14:27:12)
- r-mibu suggests detail investigation may need
to be done to see if there is alternative solution (yujunz,
14:29:58)
- r-mibu suggests that some improvement could be
done in the message queue (yujunz,
14:31:55)
- yujunz suggests to detail the specific problem
description in each bullet of guideline (yujunz,
14:35:44)
- test code refactoring (r-mibu, 14:37:01)
- https://lists.linux-foundation.org/pipermail/opnfv-tech-discuss/2016-November/013683.html
(r-mibu,
14:38:01)
- Gerald thinks the refactoring is going into the
right direction, however, due to the other high prio tasks, suggests
to delay this activity for a later release and for now focus on
smaller changes to the existing scripts (GeraldK,
14:46:17)
- Carlos will share his thoughts via email (due
to an issue his reply was not sent on the ML) (GeraldK,
14:47:12)
- Carlos: Definitely Python would be better to
profile performance rather than Bash as we have at the moment.
I propose that we work on this in parallel with the tests Bash code
refactoring activity that is ongoing. Later on, still at Danube
development time frame, we can make a decision of which one to use
by default – although for that it would be required adding more
verify jobs (2 tests branches (Bash and (GeraldK,
14:47:32)
- Yujun: Sounds like a good idea. +2 for parallel
development on both bash and Python. In fact, it helps us to create
reusable components such as sample inspector and sample
monitor. (GeraldK,
14:48:47)
- r-mibu suggests submit the code and verify it
with a non-voting job (yujunz,
14:51:36)
- performance profiler (r-mibu, 14:52:05)
- http://artifacts.opnfv.org/doctor/review/23963/scenarios/index.html#performance-profile
(yujunz,
14:52:12)
- Aaron proposes to use standard terminology such
as the ETSI FM cycle (GeraldK,
14:59:26)
Meeting ended at 15:09:03 UTC
(full logs).
Action items
- (none)
People present (lines said)
- yujunz (17)
- r-mibu (12)
- GeraldK (7)
- cgoncalves (3)
- collabot` (3)
- tojuvone (1)
- psous (1)
- ljlamers (1)
- tommyL (1)
- OPNFV-Gerrit-Bot (1)
- dwj (1)
Generated by MeetBot 0.1.4.