#onap-int: Integration weekly meeting 1/07/2020
Meeting started by morgan_orange at 16:38:47 UTC
(full logs).
Meeting summary
- action point follow up (morgan_orange, 16:38:55)
- https://ircbot.wl.linuxfoundation.org/meetings/onap-int/2020/onap-int.2020-06-18-05.26.html
(morgan_orange,
16:39:03)
- AP1: morgan_orange krzysztof integrate
preonboarding use case in CI Master (morgan_orange,
16:39:14)
- not done yet but still in mind... (morgan_orange,
16:39:23)
- ACTION: morgan_orange
krzysztof integrate preonboarding use case in CI Master
(morgan_orange,
16:39:50)
- AP2 morgan_orange contact Marco & Brian to
get their feedback on Windriver/Intel Integration resources
(morgan_orange,
16:40:01)
- done mail exchange, green light for integration
tenant refactoring (morgan_orange,
16:40:13)
- AP3 morgan_orange create new project
Integration-Terraform (morgan_orange,
16:40:25)
- done (morgan_orange,
16:40:26)
- AP4 all complete the etherpad to list possible
scenarios (morgan_orange,
16:40:35)
- see next section (morgan_orange,
16:40:39)
- AP5 trevor provide the Bell upstream suite
(need to select the place and check legal issues first) (morgan_orange,
16:40:50)
- AP6 Pawel W confirm we can swich from http
endpoint to nonsslendpoint use case (morgan_orange,
16:40:57)
- not done (morgan_orange,
16:40:59)
- ACTION: morgan_orange
Pawel W confirm we can swich from http endpoint to nonsslendpoint
use case (morgan_orange,
16:41:04)
- AP7 thierry/Andreas discuss onapsdk scenario
structure and push first scenario in ONAP repo (morgan_orange,
16:41:11)
- done first protoype available see next
section (morgan_orange,
16:41:24)
- AP8 morgan_orange contacts OOM to precise that
it would be great to keep last 2 stable versions in CI (morgan_orange,
16:41:31)
- everybody agrees on the principle, we are
lacking resources (people and hardware) for the moment, but it would
fully make sense, maybe a question to raise to TSC/PTL (morgan_orange,
16:42:18)
- stable branches (morgan_orange, 16:43:44)
- frankfurt status stable over the last
week (morgan_orange,
16:44:00)
- https://docs.google.com/spreadsheets/d/1i3hPEqWwqggURXArrpum6I7j1Doao8HKWt-WBIG9gMg/edit?usp=sharing
(morgan_orange,
16:44:13)
- issues are known issues (due to parallelization
for vnf_) and due to hieratic behavior at AAI level for
pnf-registrate => but globally it is fine (morgan_orange,
16:45:09)
- use case: maintenance release will be needed
for some use cases.. (morgan_orange,
16:45:27)
- few patches seen on CDS (morgan_orange,
16:45:41)
- ACTION: morgan_orange
ask Yuriy to get notified when patches are all in (needed for closed
loop stuff out of the box) to self release testsuite 1.6.4
(morgan_orange,
16:46:28)
- agreed by Yuriy at the end of the
session (morgan_orange,
16:46:41)
- regarding the request for support on frankfurt
=> most of the questions dealing with frankfurt issues are
related to the installation( maria-galera resinstallation)
(morgan_orange,
16:47:03)
- OOm patch on upgrade has been reverted =>
reported to Guilin (need more testing...) (morgan_orange,
16:47:23)
- El Alto maintenance release (morgan_orange,
16:47:35)
- CI run until beginning of May (switch to daily
Frankfurt) on orange DC (morgan_orange,
16:47:44)
- Results wer OK and fixes (end of branches
mainly deal with certificate issues) => no need to launch a full
regression campaign all the more as, as all the use case owners are
no more involved in ONAP it would be impossible to deal with the
use cases. (morgan_orange,
16:47:51)
- Release manager asked if regression campaign
was needed (expected according to our status) (morgan_orange,
16:48:12)
- answer was now: CI was fine until beginning of
May, the only changes since this date deals with cert fixes
(morgan_orange,
16:48:34)
- moreover impossible to ge El Alto use case
owner (some moved to other projects). As the tests were poorly
automated, almost all should be done manually (morgan_orange,
16:49:16)
- so no resources (neither humane nor hardware)
to setup such regression campaign (morgan_orange,
16:49:54)
- admin (morgan_orange, 16:49:58)
- Windriver/Intel cleaning operation about to
start (morgan_orange,
16:50:12)
- https://wiki.onap.org/display/DW/Windriver+lab+Frankfurt+cleaning+campaign
(morgan_orange,
16:50:18)
- info resaid during PTL meeting, cleaning shall
start before the end of the week (morgan_orange,
16:50:35)
- resources for terraform based installation
created (morgan_orange,
16:50:52)
- robot-utils: repository is available => we
could set a robot linter (not done on testsuite due to lots of
errors but could be applied here) (morgan_orange,
16:51:02)
- Guilin (morgan_orange, 16:51:33)
- New JIRA board available (morgan_orange,
16:51:40)
- https://jira.onap.org/secure/RapidBoard.jspa?rapidView=221&selectedIssue=INT-1268
(morgan_orange,
16:51:46)
- cleaning also probably needed here as some JIRA
were transfered from El Alto and did not move (morgan_orange,
16:52:03)
- new robot pod archi (morgan_orange,
16:52:12)
- testsuite/oom helm chart to be moved back to
OOM (proposal to have both in // during transition phase)
(morgan_orange,
16:52:19)
- leveraging on new architecture proposed by
Daniel Rose (3 pods: 1 robot, 1 DB, 1 web server) (morgan_orange,
16:52:26)
- shall follow OOM recommendations (cloud native,
security,..) (morgan_orange,
16:52:31)
- run the tests from inside the cluster/outside
)> consequences (morgan_orange,
16:52:37)
- lucian already started working on this topic,
should take the lead as some work already done to extract robot
tests for ingress (morgan_orange,
16:53:03)
- but so far only AAI tests seem OK (morgan_orange,
16:53:09)
- other tests are calling internal components
directly preventing the tests to be run as an end user
(target) (morgan_orange,
16:53:31)
- Kzrysztof K indicates that it could be OK for
some projects but clearly not all (pnf-registrate is directly
calling the SO DB) (morgan_orange,
16:54:11)
- Kzrysztof O would like to see mark/label to
know which tests are runnable from outside (morgan_orange,
16:54:37)
- ACTION: morgan_orange
intiate a wiki page to track the status of Robot rests (runnable
from inside/outside) and get the reason cause when it is not
runnable from outside (morgan_orange,
16:55:10)
- some tests calling DCAE through HTTP Internal
VES collector pod will not work (certificates need) (morgan_orange,
16:56:16)
- some exceptions are understandable but the rule
shall be end user approach first, to be discussed case by
case (morgan_orange,
16:56:41)
- Guilin Requirements dealing with VNF automation
submitted by Lei (China Mobile) (morgan_orange,
16:57:01)
- https://jira.onap.org/browse/REQ-335
(morgan_orange,
16:57:09)
- https://jira.onap.org/browse/REQ-336
(morgan_orange,
16:57:13)
- https://jira.onap.org/browse/REQ-337
(morgan_orange,
16:57:17)
- https://jira.onap.org/browse/REQ-338
(morgan_orange,
16:57:20)
- need for automatio is shared, syncho session
with CCMC this morning (morgan_orange,
16:57:40)
- they plan to update the tools they are used too
(VTP / oclip / ) (morgan_orange,
16:57:58)
- in Integration we do not use these tools (yet)
- the use case vCPE Tosca we tried to integrated in frankfurt is
using these tools, so it could be integrated in CI (morgan_orange,
16:58:31)
- I shared the CI view including xtesting and
existing tests (morgan_orange,
16:58:50)
- I suggest also that we could maybe create an
Integration slot for Asian contributor at a more convenient timer if
needed (morgan_orange,
16:59:22)
- REQ-338 is a bit different, it deals with
certification, which is out of scope of Integration mission
(morgan_orange,
16:59:50)
- regarding REQ-335, 336 and 337, Integration
support them all the more as work has been already initiated in El
Alto in this area with the introduction of CI chains (morgan_orange,
17:00:27)
- it is also not a problem to have 2 tooling
chains (xtesting/python-sdk or robot versus VTP/oclip) (morgan_orange,
17:01:03)
- Jira had been created for the integratin of VTP
in frankfurt (morgan_orange,
17:01:17)
- next goal, integrate vCPE Tosca and see if help
is needed on their tooling chain (morgan_orange,
17:01:39)
- pythonsdk (morgan_orange,
17:01:44)
- First release available on pypi (morgan_orange,
17:01:51)
- https://pypi.org/project/onapsdk/
(morgan_orange,
17:01:59)
- thank to Michal and Sylvain for the first
version (morgan_orange,
17:02:11)
- discussion on the possible scenarios to be
created in https://git.onap.org/testsuite/pythonsdk-tests/
(morgan_orange,
17:02:28)
- First PoC proposed for scenario by Thierry
https://gitlab.com/Orange-OpenSource/lfn/onap/pythonsdk-tests
(morgan_orange,
17:02:50)
- morgan_orange created an issue to review the
scenario datamodel - avoid using the same approach than in
onap-tests (mix between sdk / scenario, with too many hardcoded
things) (morgan_orange,
17:03:42)
- pythonsdk has been developed by developers,
onap-tests by integrators => good moment to improve the sceanrio
part based on dev best practices (unit test, coverage, object
moedlization,..) (morgan_orange,
17:04:22)
- ACTION: michal sync
with thierry/Andreas/Morgan to converge to the best scenario
moedlization (morgan_orange,
17:04:43)
- Thierry implemented a scenario object as
suggested in the Issue, onboarding/instantiation OK (morgan_orange,
17:05:14)
- to be reviewed by people interested in
pythonsdk before we move the scenario to ONAP repo (morgan_orange,
17:07:54)
- new use cases (morgan_orange,
17:07:55)
- first try on check certificates: dealing with
nodport/internal/ingress => to be integrated in CI (morgan_orange,
17:08:04)
- https://gerrit.onap.org/r/c/integration/+/109207
(morgan_orange,
17:08:04)
- please committer review and merge => I will
add it it CI (morgan_orange,
17:08:47)
- preonboarding, vcpe_tosca still in the
loop (morgan_orange,
17:08:48)
- Upstream cds tests are coming (morgan_orange,
17:08:48)
- legacy tests could also be tested ..from the
cluster, use the tag.. (morgan_orange,
17:08:48)
- ACTION: morgan_orange
give a try to CDS test for next week (morgan_orange,
17:08:48)
- security test: (morgan_orange,
17:08:48)
- cleaning of the xfail lists on master?
(morgan_orange,
17:08:48)
- ACTION: morgan_orange
contact SECCOM to see if we can reset the xfail list now or if we
should wait a milestone (morgan_orange,
17:08:50)
- http enpoint => update: switch to nonssl?
(Paweł Wieczorek ) (morgan_orange,
17:08:51)
- ACTION: morgan_orange
contact Pawel W to see if we are OK to switch (morgan_orange,
17:08:51)
- which threshold fro limits? (morgan_orange,
17:08:51)
- add the question for SECCOM (morgan_orange,
17:08:51)
- check version according to SECCOM
recommentations => ask SECCOm to move their list through a file
in their repo to get a gerrit reference rather than a wiki
page? (morgan_orange,
17:08:51)
- a usecase is ready, pawel should submit it next
week (morgan_orange,
17:08:52)
- ACTION: morgan_orange
integrate the test in CI (morgan_orange,
17:08:59)
- AoB (morgan_orange, 17:09:03)
- windriver ci from gitlab.com: I was able to
install the full Openstack resources + launch the kubespray
installation but still facing some issues...to be continued
(morgan_orange,
17:09:42)
- question from Ali on the v2F2 video
(morgan_orange,
17:09:54)
- https://wiki.lfnetworking.org/display/LN/2020+June+Virtual+Recordings
(morgan_orange,
17:11:33)
- python SDK video is here (morgan_orange,
17:11:55)
- https://wiki.lfnetworking.org/download/attachments/40370243/Python%20ONAP%20SDK.mp4?api=v2
(morgan_orange,
17:11:57)
- and documentation can be found in (morgan_orange,
17:13:55)
- https://python-onapsdk.readthedocs.io/en/latest/description.html#
(morgan_orange,
17:13:55)
Meeting ended at 17:13:55 UTC
(full logs).
Action items
- morgan_orange krzysztof integrate preonboarding use case in CI Master
- morgan_orange Pawel W confirm we can swich from http endpoint to nonsslendpoint use case
- morgan_orange ask Yuriy to get notified when patches are all in (needed for closed loop stuff out of the box) to self release testsuite 1.6.4
- morgan_orange intiate a wiki page to track the status of Robot rests (runnable from inside/outside) and get the reason cause when it is not runnable from outside
- michal sync with thierry/Andreas/Morgan to converge to the best scenario moedlization
- morgan_orange give a try to CDS test for next week
- morgan_orange contact SECCOM to see if we can reset the xfail list now or if we should wait a milestone
- morgan_orange contact Pawel W to see if we are OK to switch
- morgan_orange integrate the test in CI
Action items, by person
- morgan_orange
- morgan_orange krzysztof integrate preonboarding use case in CI Master
- morgan_orange Pawel W confirm we can swich from http endpoint to nonsslendpoint use case
- morgan_orange ask Yuriy to get notified when patches are all in (needed for closed loop stuff out of the box) to self release testsuite 1.6.4
- morgan_orange intiate a wiki page to track the status of Robot rests (runnable from inside/outside) and get the reason cause when it is not runnable from outside
- morgan_orange give a try to CDS test for next week
- morgan_orange contact SECCOM to see if we can reset the xfail list now or if we should wait a milestone
- morgan_orange contact Pawel W to see if we are OK to switch
- morgan_orange integrate the test in CI
People present (lines said)
- morgan_orange (113)
- collab-meetbot (3)
Generated by MeetBot 0.1.4.