Lee Anjella Macabuhay present latest version of our automatic performance (daily) plots. The last issues have been fixed. The variation in the plots is mainly caused by environmental issues when running at daytime. According to Daniel Hanrahan the results form the 2am run are most stable. Lee Anjella will change the trigger to time based so the data wil only include the 2am runs. Daniel Hanrahan is fixing the limits based on teh latest results and wil also work on choosing a small set of representative test to plot and examin on a daily basis.
CPS Sonar reported 10 new issues. Most of these seem to be caused by the upgrade to Java17. Priyank Maheshwari will address those and Halil Cakal will address the issue related to his recent commit
CPS-1789
-
Getting issue details...STATUSGerard Nugent is close to delivering the Springboot 3.0 upgrade, this wil required developers to use a Java17 JDK when building locally! Newer version cause build problems (for spotBugs mvn plugin). Gerard wil notify with details when it is ready
Gerard Nugent agreed to push the SpringBoot 3 changes today for review and merge soon hopefully. The failign test can temporary be ignore and the Code coverage adjusted as needed to follow up in separate commit(s)
CPS-1784
-
Getting issue details...STATUS
- a question came up on consistency of the ordering of records returned when using the descendants option. Rudrangi Anupriya to check and confirm the consistency.
scenario 4 that a DMI reports some cm handles are 'pending' is NOT a valid scenario; a DMI can only reject or accept cm handles. Need a code update to fix this. Halil Cakal to fis this
demo still use incorrect datastore but code fix is pending in review Toine Siebelink confirms
outcome from dmi to ncmp is inefficient if many (all) cm handles are not accepted for the same reason. The reason wil be repeated in the json for each cm handle We could use a similar grouping schema as is used when forward outcome from ncmp to client. Priyank Maheshwari to decide if this will be changed now.
The demo could be approved to show that different responses from DMI are grouped per message, the current demo uses only only one error message 3 times it would be betert to use different message, one of the at last occurring twice to show the grouping when status details are forwarded to client...
scenario 3 is also based on an actual response from dmi saying all are pending this is not valid Halil Cakal will prepare a new demo where the 'pending' state wil occur and a message set by NCMP about te pending state...
Gerard Nugent DMI Being upgraded to Java 17 too now (some build issues)
CPS/NCMP Code Coverage now up over 98% however recent code is still missing coverage for error scenarios. It remains the responsibility for the whole team to double-check their own new code for coverage.
ONP Release planning is on hold as the impact of ONAP 'disaggregation' is being analyzed and the process are updated to cater for that.
Concurrent DB Layer (alternative) possibly for same data. Dataspace and/or anchor might have attribute to indicated with persistence implementation(s) is to be used
Different kind of persistency ( mongodb or elasticsearch )
mirror the config into json and come with some patterns network wide.
dataspace needs to be schema free for this kind of data.
GET and QUERY api to get the hosted data in elasticsearch (might not support all the features as PostGress impl but could have additional features. Would required documentation updates to explain.
Admin and DataPersistency service.
Lukasz Rajewski will work out a more detailed proposal for the arhictecutre based on a their PoC to discuss with CPS team when ready...
(limited) Software update
Daniel Hanrahan worked on performance bug
CPS-1716
-
Getting issue details...STATUS
. Fix almost completed which reduced memory significantly and increases performance of many query-features
a recent commit for the batch-operations user story causes issue with Kafka Messaging in general. Sourabh Sourabh and Halil Cakal will discuss with Sourabh the commit can be reverted if needed
Gerard Nugent still issues with move to Java 17, seems we still use Sqgger instead of OPenAPI to generate our interfacecode
Several user stories will be affect by the latest proposal to align with CNFC Cloud Event standards (see CPS Events Structure#CNFCCloudEventalignment) Its recommend to merge the current open patches and update as part of a new commit after header fields have been agreed meeting with kieran mccarthy on )
Gerard Nugent is working to move our projects to Java17 and SpringBoot 6. This should solve any dependency related issues that we see in the quality reports at the moment
functionality works great but output format is stil to change. Need to update documentation (london release especially) to state it is not completed yet!
works as expected, Documentation to be completed. Edge-case scenarios not need documented as limitation unit-test for those is good enough. Just 2 notes to add in documentation:
contains function is case-sensitive
leaf-names are not validated, silently ignored if not existing....
Priyank Maheshwari has some issue with deployment of DMI plugin which is block the London Release. Toine will look for additional support if needed.
New R13 Planning page: CPS R13 Release Planning Different table format, now using epics, please ensure any user story is linked to the correct epic! Page still WiP but teams should be using it from now on.
Ericsson and Fujitsu REQ-Tickets for London have been closed and remaining user stories transferred to new REQ tickets for Montreal
DT/TechMahindra are meeting tomorrow to close their London ticket and create a new one for Montreal
