Date & Time
June, 24th, 2020; 9pm PST
Agenda
- Release
- Infra
- Docs
- Misc
Minutes
Release:
- sodium SR release checklist ready and vote to approve that SR4 release is here
- planned to release 8/27
- Aluminium planned to release 9/17
- still a few problems to debug in the following CSIT jobs
- controller (seems car/people job has new string format)
- bgpcep has real issues
- 2 regressions in netconf
- OFP has a bug to report
- no idea about lisp
- genius and ovsdb might have python3 troubles
- netvirt has some real blockers
- still a few problems to debug in the following CSIT jobs
- Magnesium has nothing major going on, but will release between Sodium SR4 and Aluminium
Infra:
- info.yaml jobs are coming online now which allow us to self promotion (allowing us to add/remove committers via gerrit patch)
- Vexxhost was broken this week, but resolved rather quickly by opening a ticket and pinging mnaser on IRC
- Netvirt job count has increased from ~400 to ~500, but we are hoping the frequency will be reduced.
- Srinivas Rachakonda has further reduced the job count by only using two openstack versions as opposed to the
previous 3. (confirmed to be under 400 now for all of netvirt, 314 for CSIT)- quick cmdline to get job count: "jenkins-jobs -s odl-sandbox list -r -p jjb/ 'netvirt-*' | wc -l"
- Srinivas Rachakonda has further reduced the job count by only using two openstack versions as opposed to the
- looks like we are not getting our automated job patch from jenkins-releng user
Docs:
- no news for now
Misc:
Action Items
Jamo Luhrsen to look in to why automatic job creation patch not being generated
2 Comments
Robert Varga
re. BGP, this looks like CSIT tooling issue: https://logs.opendaylight.org/releng/vex-yul-odl-jenkins-1/bgpcep-csit-1node-userfeatures-all-aluminium/116/bgp_peer1.log.gz :
hence the CSIT driver, hence data does not get to BGP... given how these things work, shouldn't Robot be checking exit codes and scream loudly when things are aborting?!
Same thing is happening here: https://logs.opendaylight.org/releng/vex-yul-odl-jenkins-1/bgpcep-csit-1node-bgp-ingest-all-aluminium/114/manypeers_cc_play.log.gz
Robert Varga
Controller CSIT seems to be a but in RF definition:
https://logs.opendaylight.org/releng/vex-yul-odl-jenkins-1/controller-csit-3node-clustering-ask-all-aluminium/117/robot-plugin/log.html.gz#s1-s22-t7-k3-k1-k1
note how ${iter} says "21.0" not "21"
this boils down to how Builtin.evaluate works in csit/suites/controller/Clustering_Datastore/carpeople_crud.robot