#acumos-meeting: Acumos TSC
Meeting started by bryan_att at 13:05:12 UTC
(full logs).
Meeting summary
- Agenda (bryan_att, 13:05:31)
- Planning Release (bryan_att, 13:05:54)
- Jamil: presents release plan proposal
deck (bryan_att,
13:06:37)
- Nat: How to handle new projects, to allow them
time to design in the release. (bryan_att,
13:14:41)
- Nat: model training and license mgmt projects
need to be started or they will miss release B (bryan_att,
13:16:05)
- ... propose to identify PTLs now (bryan_att,
13:16:27)
- Jack: propose that we include design work in
release A for these new projects (bryan_att,
13:19:12)
- Bryan: we will need to address the
projects/repos structure first before PTL assignment, since the
current list would likely not fit the current proposal on the wiki
at
https://wiki.acumos.org/display/AC/Acumos+Architecture+Mapping+to+Repos+Jira+etc
(bryan_att,
13:24:18)
- Bryan: the proposal looks more compatible with
waterfall process projects; I am unsure if it will fit an agile
process which includes test/QA and requirements verification with
the stakeholders (Community/Project committee). (bryan_att,
13:25:38)
- Jack: describes current development practice
which is more agile (bryan_att,
13:27:05)
- Jack: code is developed on master and release
is cut periodically, e.g. weekly (bryan_att,
13:28:42)
- Jamil: since we have significant startup code
we can focus on the testing with a limited design time on new
projects (bryan_att,
13:30:04)
- Jack: proposal for 1st release name is
"Athena" (bryan_att,
13:30:30)
- Manoop: re code freeze, recommend this be later
at RC date to enable more development time, if the features being
dev'd are fewer; we can also use the proposed code freeze date as a
gate for high/critical defect correction only (bryan_att,
13:32:46)
- #info Jack: this points to a question about
open source in general: how much release functionality is just
controlled by contributions over time, and how much is release
functionality planned in advance (bryan_att,
13:34:21)
- Bryan: most large open source do have a
specific release plan phase in which new features are agreed
(bryan_att,
13:35:42)
- Manoop: PTLs can control which code goes in per
the relation to a feature ID (JIRA issue). Any code that does not
make the release will be deferred to the next release. (bryan_att,
13:37:34)
- Jamil: we also have to agree on the maintenance
plan for releases, how long do we support e.g. via maintenance
versions (bryan_att,
13:38:20)
- Scott: may be able to provide some input from
the LF release engineering (bryan_att,
13:40:09)
- Anwar: want to be sure there is adequate time
for integration testing (bryan_att,
13:41:25)
- Bryan: testing should be continual, with unit,
subsystem (integration), and even full system level test being done
before the code freeze (bryan_att,
13:42:18)
- Jack: the integration team will need to prepare
a plan long before integration start, so they are ready and can use
tools e.g. robot to automate testing (bryan_att,
13:43:17)
- Pantelis: should we consider cloud-native
support as a high level requirement, and how will we consider these
re the current code, e.g. will they significantly impact the current
code and thus should be deferred to release B, etc (bryan_att,
13:44:26)
- Jack that's where M1 is critical, and the
discussions between M1 and M2 would address that; in this 1st
release case M1/M2 are the same and that puts a challenge on large
feature proposals (bryan_att,
13:45:47)
- Pantelis: if we do see new requirements agreed
that have large changes on the code (e.g. architecturally), they
will in this 1st release be not possible and will be deferred to the
next (bryan_att,
13:47:05)
- Jack: we can separate M1/M2 if we need to
consider the release plan for new features, to decide if they are in
A or B (bryan_att,
13:47:46)
- Scott: Vanessa (release eng team) can provide
some input on what's typical (bryan_att,
13:50:23)
- Manoop: if M1 results in a feature proposal
that by M2 is agreed in scope terms, the team can then develop what
is possible given that scope and defer the rest (bryan_att,
13:53:02)
- Jack: 1st time thru the process we will be
figuring out the test processes e.g. re the nature of this project
vs others (bryan_att,
13:54:18)
- Jack: goal for today was to get agreement on
M0 (bryan_att,
13:54:44)
- Scott: the plan can evolve during the
release (bryan_att,
13:55:11)
- Jack: call for motions to approve (bryan_att,
13:55:23)
- Nat: recommend approval (bryan_att,
13:55:35)
- Ofer: are we separating M1/M2? (bryan_att,
13:55:51)
- Nat: we can modify to gove some time as
needed (bryan_att,
13:56:08)
- Vote resulted in no objections to the proposed
Release plan and M0 date (bryan_att,
13:57:23)
- AGREED: The release
plan proposal and M0 date are agreed (bryan_att,
13:57:47)
- Committee Lead nominations (bryan_att, 13:58:09)
- Currently Anwar has self-nominated for
Architecture; need nominations for Community and Security
(bryan_att,
13:59:27)
- for PTLs, the wiki page
https://wiki.acumos.org/display/TSC/PTL+Nominations shows the
current proposals and we need others to sign up as candidates - we
can have two or more and vote (bryan_att,
14:01:22)
- Jack: recommend that we add a Release Manager
to the TSC committee lead roles (bryan_att,
14:02:08)
- Nat: re the agenda, #B is (LF env refresh) is
planned for 5/22 (bryan_att,
14:02:47)
- Nat: re the Platform-OAM repo, we need to add
an additional committer for maintenance in advance of the project
startup (bryan_att,
14:03:42)
- AGREED: - another
committer can be added to Platform-OAM as requested by Nat
(bryan_att,
14:04:24)
Meeting ended at 14:04:31 UTC
(full logs).
Action items
- (none)
People present (lines said)
- bryan_att (46)
- collabot` (3)
Generated by MeetBot 0.1.4.