#opnfv-meeting: Weekly Technical Discussion 157
Meeting started by bh526r at 14:00:12 UTC
(full logs).
Meeting summary
- Roll Call (bh526r, 14:00:38)
- Bin Hu (bh526r,
14:00:46)
- Jim Baker (bh526r,
14:01:03)
- Liang Chen (bh526r,
14:01:10)
- Jack Morgan (bh526r,
14:01:37)
- Parker Berberian (bh526r,
14:01:53)
- Al Morton (bh526r,
14:02:47)
- Daniel Balsiger (bh526r,
14:03:09)
- Mike Fix (bh526r,
14:03:24)
- Jiaqiang Zhang (bh526r,
14:03:37)
- Mark Shostak (bh526r,
14:05:00)
- Hardware Requirement from CNTT-RI (bh526r, 14:06:26)
- https://github.com/cntt-n/CNTT/blob/master/doc/ref_impl/cntt-ri/chapters/chapter04.md
(bh526r,
14:06:41)
- Mark S introduced hardware requirements listed
in above link (bh526r,
14:07:22)
- Mark S. indicated that it is still work in
progress (bh526r,
14:09:00)
- Trevor Cooper (bh526r,
14:09:22)
- Trevor Cooper (trevor_intel,
14:10:03)
- The scope is primarily "functional"
(bh526r,
14:14:24)
- "Scalability" needs scales of servers
(bh526r,
14:14:53)
- And lots of forward-looking work (bh526r,
14:17:27)
- Trevor C asked where the number 14 came
from (bh526r,
14:25:19)
- Mark S explained it was created based on
functional requirement at very high level we need to address
(bh526r,
14:25:51)
- 2 nodes per instance type, 3 controllers, quite
conservation (bh526r,
14:26:10)
- "conservative" instead of "conservation"
(bh526r,
14:26:46)
- Trevor C. said typical Pharos is 6 nodes: 3
controllers for HA, 2 compute and 1 jump server (bh526r,
14:27:33)
- Jack M asked if we need to do it in
phases (bh526r,
14:31:27)
- Mark S said that we need discussion, and it is
just an input and work in progress (bh526r,
14:32:32)
- Trevor C asked how to organize the community to
get the work done (bh526r,
14:33:26)
- Jack M. LaaS v.s. Static is just deployment
methodology. We need to understand the hardware requirement. Pharos
specs only says 6 servers (bh526r,
14:36:21)
- Cedric Ollivier (bh526r,
14:40:34)
- Sridhar Rao (bh526r,
14:40:52)
- Trevor Bramwell (bh526r,
14:41:01)
- Bin asked if "6" number in Pharos is
architecturally and technically limited, or is it just a matter of
funding (bh526r,
14:42:22)
- Jack M clarified we picked number "6" because
of OpenStack setup, and architecturally and technically it can scale
to any number (bh526r,
14:43:16)
- Ollivier also indicated that in FuncTest, it
can also divide into 3 server setup etc., so the number of servers
should be technically supported, and flexible. (bh526r,
14:44:27)
- Bin summarized that we can do it in
parallel: (bh526r,
14:51:19)
- CNTT clarify the rationale of the number of
servers, i.e. why 14 is needed and what is really the minimum
number (bh526r,
14:52:03)
- Jack specifies the scalability in Pharos next
evolution (bh526r,
14:53:13)
- Jack M suggested that CNTT specifies what is
needed for Iruya, J and K releases. (bh526r,
14:53:51)
- Pierre Lynch (bh526r,
14:54:52)
- AOB (bh526r, 14:59:42)
- None. Meeting adjourned (bh526r,
14:59:51)
Meeting ended at 14:59:55 UTC
(full logs).
Action items
- (none)
People present (lines said)
- bh526r (43)
- collabot (3)
- trevor_intel (1)
Generated by MeetBot 0.1.4.