#opendaylight-meeting: tsc
Meeting started by colindixon at 17:00:05 UTC
(full logs).
Meeting summary
- roll call and agenda bashing (colindixon, 17:00:13)
- colindixon (colindixon,
17:00:16)
- https://wiki.opendaylight.org/view/TSC:Main#Agenda
Today’s agenda (tbachman,
17:00:34)
- https://meetings.opendaylight.org/opendaylight-meeting/2015/tsc/opendaylight-meeting-tsc.2015-03-19-17.00.html
Last week’s meeting minutes (tbachman,
17:00:45)
- dfarrell07 for cdub and Red Hat (dfarrell07,
17:02:11)
- Chris Price (almost on audio) (ChrisPriceAB,
17:02:21)
- LuisGomez (LuisGomez,
17:02:29)
- ACTION: colindixon to
continue to follow AD-SAL deprecation between VTN, OVSDB, and
controller (are ther others?) (colindixon,
17:03:07)
- ACTION: colindixon
keep the thread on best practices alive (colindixon,
17:03:13)
- edwarnicke (edwarnicke,
17:03:51)
- ACTION: dlenrow to
follow up on the mail about how to procede with persitence
(colindixon,
17:03:55)
- regXboi (partly here) (regXboi,
17:04:16)
- jmedved (jmedved,
17:04:18)
- mohnish (mohnish,
17:04:25)
- ACTION: LuisGomez to
start drafting a list of qualifiers for an issue for being a blocker
to an SR - edwarnicke to assist (colindixon,
17:04:39)
- ACTION: rovarga
(and/or Thanh) to add an something to a backlog somewhere about
measuring latency between merge and CSIT completely finishing
(colindixon,
17:08:08)
- LuisGomez says that this may be easier in
Beryllium assuming we move Zuul (colindixon,
17:08:43)
- updates (colindixon, 17:09:33)
- ACTION: zxiiro to
create revert patches for autorelease <relativePaths> and
abandon existing patches. Will resolve <relativePaths> inside
of autorelease scripts instead. (zxiiro,
17:09:59)
- wojciech would like to add discussing the AAA
project to the agenda (tbachman,
17:10:30)
- colindixon will cover requests to the TSC from
the board (colindixon,
17:10:57)
- edwarnicke requests to cover the weather
briefly (colindixon,
17:11:05)
- Events (colindixon, 17:11:07)
- phrobb says that the hackfest is coming up
4/15-4/16; 49 people are registered; those who want to start with
action items to work on should start to formulate what that looks
like and send an agenda (tbachman,
17:11:13)
- phrobb says for the ODL summit in July:
notifications to speakers are going out now (not all out yet — in
process today and tomorrow) for accepted presentations and
tutorials (tbachman,
17:11:56)
- http://www.opendaylight.org/news/events
ODL Events web page (tbachman,
17:12:24)
- Lithium and Stable/Helium (tbachman, 17:12:52)
- https://lists.opendaylight.org/pipermail/tsc/2015-March/002769.html
summary from phrobb (colindixon,
17:13:22)
- gzhao is still waiting for 5 offset 2 projects
to report (tbachman,
17:13:42)
- https://docs.google.com/spreadsheets/d/1KPpO9LH539Vlcoa4RvLa6PPCdLifi5JD-ihRhlybqeo/edit#gid=1196332566
the project stats spreadsheet (colindixon,
17:13:49)
- reservation says they’ll report soon; core
tutorials not as concerned about; 2 other projects still to
report (tbachman,
17:14:14)
- colindixon says the most concerning one is AAA,
which is on the agenda (tbachman,
17:15:02)
- gzhao says discovery is also in red
(tbachman,
17:15:08)
- zxiiro says there are 5 projects still to
complete the SR3 merging/tagging (tbachman,
17:15:27)
- Youcef Laribi (Youcef,
17:15:29)
- zxiiro says l2switch can merge, which is
holding up other projecgts (tbachman,
17:15:46)
- colindixon asks if there’s anything else around
projects moving to JJB, etc. (tbachman,
17:16:17)
- zxiiro says Defense4All has moved to JJB; just
plugin2oc is left (tbachman,
17:16:31)
- tykeal points out that OpFlex is also not in
JJB, but they’re working with them already (tbachman,
17:17:13)
- ACTION: LuisGomez to
start a thread about how to unblock integration testing until
patches are merged (colindixon,
17:18:14)
- System Integration and Test (tbachman, 17:18:27)
- https://wiki.opendaylight.org/view/CrossProject:Integration_Group:Performance_Test:Results
Performance Test results in lab by Integration group (tbachman,
17:19:03)
- LuisGomez says they’ll be measuring performance
from now on and will be reporting in this wiki, and will report
bugs (tbachman,
17:19:23)
- LuisGomez says he’ll keep track of the bugs and
bring them up if they’re not being filed by developers (tbachman,
17:20:12)
- mohnish says longevity is also important — asks
if it’s also being tested (tbachman,
17:20:56)
- LuisGomez says they’re working on setting up
longevity testing, but nothing currently; Spirent may also help with
this effort (tbachman,
17:21:20)
- jmedved asks if LuisGomez is using the scripts
he provided a while back (tbachman,
17:21:46)
- LuisGomez says they’re using the scripts for
some things (tbachman,
17:21:57)
- jmedved says the codecs for REST have changed
recently — should retest to see if there’s significant performance
improvements (tbachman,
17:22:47)
- Infrastructure (tbachman, 17:24:09)
- the sandbox has been having issues because of
how much stuff has been going through releng (colindixon,
17:24:35)
- tykeal says odlforge has been put on the
backburner until releng stuff gets worked out (tbachman,
17:25:18)
- AAA status (tbachman, 17:25:25)
- https://lists.opendaylight.org/pipermail/tsc/2015-March/002775.html
the status from AAA (colindixon,
17:26:07)
- Wojciech says that the token caches looks like
it will make it (colindixon,
17:27:36)
- Wojciech said they had a dependency on the
persistence project, which also has resource changes, which puts
this in jeopardy (tbachman,
17:27:37)
- Wojciech says there are only three people
currently on the project; there is a choice of an MD-SAL AAA
keystore, or completeing the keystone project (notes MD-SAL AAA
would be “experimental") (tbachman,
17:28:35)
- Mahesh (from HP) asks for using MD-SAL for AAA
— with in-memory data store, is that sufficient? (tbachman,
17:29:13)
- colindixon says we’re going to enable the
clustered data store by default, but without clustering (i.e.
single-node), so it will be persistent by default for Lithium
(tbachman,
17:29:39)
- Mahesh asks if AAA can then use the MD-SAL and
meet the Lithium goals (tbachman,
17:30:05)
- Wojciech says he doesn’t feel comfortable
committing to this deliverable by Lithium — though they can make
progress there (tbachman,
17:30:30)
- colindixon asks who is dependent on delivery of
these features (e.g. SNMP for keymanager)? (tbachman,
17:31:01)
- Wojciech says other than the general public
interested in deploying AAA in clusters, he’s not aware of
any (tbachman,
17:31:21)
- Wojciech says SNMP is no longer dependent as
they’ve gone with a very simple mechanism for storing their
credentials (tbachman,
17:31:39)
- colindixon asks why keystone federation was
critical for Lithium? (tbachman,
17:32:35)
- edwarnicke says we need to make sure we have
something that works following the standard “out of the box”
deployment model — keystone radically changes this model
(tbachman,
17:33:12)
- edwarnicke says that mestery says it’s
super-important to talk sooner rather than later to openstack
keystone folks in order to make sure there aren’t any issues
there (tbachman,
17:33:40)
- ACTION: regXboi,
Wojciech. edwarnicke, and Mahesh to organize meeting to talk about
AAA keystone federation (tbachman,
17:35:09)
- Wojciech asks for guidance from TSC on
“chicken-and-egg” problem (tbachman,
17:35:31)
- ACTION: edwarnicke,
rovarga, and ttkacik1 to look into bootstrapping AAA (tbachman,
17:35:46)
- Mahesh said for AAA, there was a dependency on
SQLite, which is not present on all the platforms; asks if we’re
removing that dependency (tbachman,
17:36:51)
- edwarnicke says that there’s a patch that
mlemay has provided that *almost* works when using H3 (tbachman,
17:37:09)
- there’s a patch that mlemay provided that will
help us to move to Karaf to 3.0.3 (colindixon,
17:37:17)
- edwarnicke says it would be huge if someone can
pick that patch up and complete it (tbachman,
17:37:33)
- Mahesh said HP will try to pick up this
item (tbachman,
17:37:43)
- Weather (by edwarnicke) (tbachman, 17:37:58)
- https://wiki.opendaylight.org/view/Weather#Clustered_Data_Store_Switch_over
Clustered Data Store Switchover on Weather page (tbachman,
17:38:26)
- edwarnicke says there’s one change that’s
positive that folks need to be aware of — controller restarts will
come up with persisted data, by default (tbachman,
17:38:52)
- rovarga says the BGP-PCEP project currently
won’t work with the clustered data store (tbachman,
17:40:01)
- ACTION: rovarga to
add the fact that BGP-PCEP currently won’t work with clustered data
store to weather page (tbachman,
17:40:33)
- Board Requests for the TSC (tbachman, 17:41:13)
- https://lists.opendaylight.org/pipermail/tsc/2015-March/002806.html
the board requests of the TSC (colindixon,
17:41:27)
- colindixon says there are 6 requests from the
TSC in that email (tbachman,
17:41:47)
- ChrisPriceAB asks who sets the requirements on
projects (tbachman,
17:45:21)
- colindixon says he hopes its a combination of
the TSC and projects (tbachman,
17:45:35)
- edwarnicke says that can get political. You
need a consistent measure of performance and stability that can be
tracked (tbachman,
17:45:59)
- edwarnicke says it’s not clear that there are
clear definitions for how you measure performance and
stability (tbachman,
17:46:16)
- ChrisPriceAB says that instead of setting a bar
at the outset as far as a use case being achieved in some way, can
we set up a set of metrics (tbachman,
17:46:49)
- ChrisPriceAB asks if we’re being to approve
requests that we don’t yet know what they are (tbachman,
17:47:47)
- colindixon says we’ve been tasked to meet these
requests, and explain how we will meet them. (tbachman,
17:48:09)
- VOTE: Voted on "Shall
the TSC have a two hour meeting next week (4/2) given the addition
discussions needed?" Results are, +1: 6, -1: 3 (phrobb1,
17:55:23)
- abhijitkumbhare asks about project maturity —
expresses concern that very few projects will be mature in the
Beryllium time frame (tbachman,
17:55:31)
- ChrisPriceAB asked about the difference between
items one and five (tbachman,
17:57:11)
- colindixon says one is about lifecycles and
five is about execution (tbachman,
17:57:24)
- regXboi_committer says the whole set of
requirements is “borderline poison” (tbachman,
17:58:42)
- regXboi says if this comes from the community,
that’s good; but coming from top-down, it’s an anathema to
open-source (tbachman,
17:59:12)
- regXboi says items #2 and #3 declare winners
and losers, and do so in a sneaky, underhanded way (tbachman,
17:59:32)
- Uri recommends consulting board members who
were involved in the discussion, as there was a long discussion
around these items (tbachman,
18:00:07)
- Uri says there was only an intent to provide
focus and direction and drive deliverables — can’t be open-ended,
b/c no investment can be completely open-ended (tbachman,
18:00:36)
- colindixon encourages folks to go to the
mailing list to discuss these items (tbachman,
18:00:52)
- phrobb1 says the next board meeting is next
Wednesday, April 1st (tbachman,
18:01:11)
- colindixon says we’re unlikely to have a
response in time for the next board meeting (tbachman,
18:01:25)
- colindixon says his ask of TSC members and
committers is to not knee-jerk react to these, but come with the
“best” interpretation that can help move the project forward
(tbachman,
18:02:00)
- cookies (tbachman, 18:02:57)
- edwarnicke points out that the extra hour for
TSC members who do development is extra painful (tbachman,
18:04:01)
Meeting ended at 18:04:06 UTC
(full logs).
Action items
- colindixon to continue to follow AD-SAL deprecation between VTN, OVSDB, and controller (are ther others?)
- colindixon keep the thread on best practices alive
- dlenrow to follow up on the mail about how to procede with persitence
- LuisGomez to start drafting a list of qualifiers for an issue for being a blocker to an SR - edwarnicke to assist
- rovarga (and/or Thanh) to add an something to a backlog somewhere about measuring latency between merge and CSIT completely finishing
- zxiiro to create revert patches for autorelease <relativePaths> and abandon existing patches. Will resolve <relativePaths> inside of autorelease scripts instead.
- LuisGomez to start a thread about how to unblock integration testing until patches are merged
- regXboi, Wojciech. edwarnicke, and Mahesh to organize meeting to talk about AAA keystone federation
- edwarnicke, rovarga, and ttkacik1 to look into bootstrapping AAA
- rovarga to add the fact that BGP-PCEP currently won’t work with clustered data store to weather page
Action items, by person
- colindixon
- colindixon to continue to follow AD-SAL deprecation between VTN, OVSDB, and controller (are ther others?)
- colindixon keep the thread on best practices alive
- edwarnicke
- LuisGomez to start drafting a list of qualifiers for an issue for being a blocker to an SR - edwarnicke to assist
- regXboi, Wojciech. edwarnicke, and Mahesh to organize meeting to talk about AAA keystone federation
- edwarnicke, rovarga, and ttkacik1 to look into bootstrapping AAA
- LuisGomez
- LuisGomez to start drafting a list of qualifiers for an issue for being a blocker to an SR - edwarnicke to assist
- LuisGomez to start a thread about how to unblock integration testing until patches are merged
- regXboi
- regXboi, Wojciech. edwarnicke, and Mahesh to organize meeting to talk about AAA keystone federation
- rovarga
- rovarga (and/or Thanh) to add an something to a backlog somewhere about measuring latency between merge and CSIT completely finishing
- edwarnicke, rovarga, and ttkacik1 to look into bootstrapping AAA
- rovarga to add the fact that BGP-PCEP currently won’t work with clustered data store to weather page
- zxiiro
- zxiiro to create revert patches for autorelease <relativePaths> and abandon existing patches. Will resolve <relativePaths> inside of autorelease scripts instead.
People present (lines said)
- tbachman (134)
- colindixon (30)
- odl_meetbot (17)
- dfarrell07 (12)
- phrobb1 (10)
- edwarnicke (7)
- ChrisPriceAB (6)
- LuisGomez (4)
- regXboi (4)
- tykeal (3)
- mohnish (3)
- Youcef (3)
- rovarga (2)
- jmedved (2)
- zxiiro (2)
- abhijitkumbhare (2)
- snoble (1)
- hideyuki (1)
- phrobb (0)
Generated by MeetBot 0.1.4.