OOM Meeting Notes - 2022-01-19
Agenda:
log4j:
only [DMAAP-MR] Remediate log4j bug | https://gerrit.onap.org/r/c/oom/+/126414 left
regression:
the version 10.6.5-debian-10-r28 solved the problem
databases:
mariadb @Mahmoud Abdelhamid :
working on CDS and policy
postgresql: @A. Seaudi
waiting for abdel corrections
service mesh:
SDC: restarted the gate on it and LGTM
CDS: should be OK to merge when SDC is merged
SO: needs a refresh but should be OK to merge
DCAE: @Andrew Lamb is working on it
Policy: @Adrian Matthews is working on it
proposal: @Sylvain Desbureaux to present a demo / slide on different authentication modes with Istio
azure kubernetes version:
azure3 was on a bad shape
tried version 1.22.4 (target of SECCOM) and it seems to work
kept the version and will see if next gates are good
if it's OK, will move azure4 to same version (1.22.4)
Helm version has not been touched for now (need to work on helm push first)
ipv4/v6 dual stack
Azure has a beta version for dual stack
if Nokia teams nows how to enable it via ansible, we can use it on the gate (@damian.nowak)
strimzi poc:
Goal : replace dmaap MR with cncf compliant kafka deployment named strimzi (https://strimzi.io/)
several issues:
some ONAP components are talking directly with kafka (and not via message router):
CPS
DCAE HV VES
and maybe more
Strimzi deployment is using CRD BUT some stuff needs to be created and in running state before other resources to be created
clusterOperator and entityOperator needs to be created and in running state
then we can create KafkaUser and Topic
if done too soon, there are not created...
maybe we could "rearrange" onap deploy script to first deploy data oriented workloads (common mariadb, common cassandra, common postgresql AND KafkaCluster, ClusterOperator and EntityOperator) and then "ONAP" (the rest)