#opnfv-meeting: Weekly Technical Discussion
Meeting started by bh526r at 13:00:14 UTC
(full logs).
Meeting summary
- Roll Call (bh526r, 13:00:41)
- Bin Hu (bh526r,
13:00:47)
- David McBride (dmcbride_,
13:00:59)
- Fatih Degirmenci (fdegir,
13:01:33)
- Manuel Buil (mbuil,
13:01:51)
- Daniel Balsiger (bh526r,
13:02:18)
- Mark Shostak (bh526r,
13:02:58)
- Lincoln Lavoie (bh526r,
13:03:32)
- Trevor Cooper (bh526r,
13:03:45)
- Georg Kunz (bh526r,
13:03:57)
- Parker Berberian (bh526r,
13:04:16)
- Rabi A (Vodafone) (bh526r,
13:04:36)
- Lincoln Lavoie (UNH-IOL) (lylavoie,
13:04:55)
- Continue to Discuss New Release Process (bh526r, 13:05:44)
- David McBride started to introduce his
proposal (bh526r,
13:06:44)
- Pierre Lynch (bh526r,
13:15:25)
- Mark S asked if MS2 is t-2, and MS1 is t-5,
does it make sense to move MS2 a few weeks? (bh526r,
13:17:06)
- David M said, yes, this is tentative proposal,
and we can move it up a few weeks (bh526r,
13:17:38)
- Rabi indicated that specs from CNTT should be
able to indicate those details (bh526r,
13:29:20)
- Manuel asked what will be specified in common
references, i.e. version of ODL, or VM needs to start in 20
seconds? (bh526r,
13:30:24)
- Manuel asked if those performance criteria will
be defined in release criteria. (bh526r,
13:31:07)
- David agreed (bh526r,
13:31:18)
- Mark S asked detailed explanation of project
types (bh526r,
13:35:30)
- When discussing slide #11 the diagram, Bin
asked 2 questions (bh526r,
13:39:53)
- (1) Installers are scenario-based projects, how
will installers decouple from scenario-based projects? (bh526r,
13:44:18)
- (2) details of how CNTT implemented in
CI? (bh526r,
13:44:54)
- Mark commented that speaking from CNTT, if CI
is prescribed, and the concern is that if CNTT includes CI specs,
how will it satisfy different system? (bh526r,
13:50:29)
- Bin indicated 3 different logical defects of
slide #11 (bh526r,
13:51:15)
- (1) From multiple installer perspective, what
is the root cause of the issue that David has observed? PS the issue
David described (multiple CIs because of installers in upstream) is
unclear whether or not those are issues (bh526r,
13:55:02)
- (2) Without knowing the root cause, we set the
goal of moving installers from upstream to downstream, it is quite
prescriptive. (bh526r,
13:55:54)
- "diagram" (bh526r,
13:57:26)
- Bin suggested that CI part needs to be
consolidated with long-term evolution to CircleCI (bh526r,
14:06:59)
- And decouple implementation part from CI,
because of those unknown details (bh526r,
14:10:03)
- Trevor asked how those Release WG meetings were
going, and whether or not projects/PTLs were engaged. (bh526r,
14:10:42)
- The attendance of release WG meetings were
low (bh526r,
14:10:56)
- Trevor suggested to engage PTLs (bh526r,
14:11:05)
- Bin suggested that if we need to engage those
PTLs one by one, that's the work we should do. (bh526r,
14:11:33)
- add (3) Without knowing the answers from Bin's
prior questions of (1) and (2), David agreed all those details are
unknown. So without knowing the details, it is immature to propose
slide #11 diagram because prior was not info'ed (bh526r,
14:12:54)
- Bin suggested to have table to compare prior
release MSs and new MSs for easy understanding. David agreed
(bh526r,
14:14:10)
- Bin suggested that we can move forward with
Iruya release planning with those improved steps (MSs). But CI
related are premature as indicated above, and we should consolidated
with long term CI evolution to CircleCI. (bh526r,
14:15:54)
- AOB (bh526r, 14:16:12)
- Next week, by default, the same agenda as
today. Continue discussion of new release process will be 1st agenda
item, followed by new project proposals. (bh526r,
14:17:01)
Meeting ended at 14:17:05 UTC
(full logs).
Action items
- (none)
People present (lines said)
- bh526r (46)
- collabot` (4)
- fdegir (3)
- dmcbride_ (1)
- georgk (1)
- lylavoie (1)
- mbuil (1)
Generated by MeetBot 0.1.4.