OOM Meeting Notes - 2021-10-13

Agenda

• Policy Helm Charts: @Liam Fallon @Saul Gill 

  • basic clamp test has to be fixed by @Michał Jagiełło and / or @Morgan Richomme 

    • done in v9.0 from python-onapsdk

    • this version needs to be used in test containers

  • as soon as it's done, we'll retrigger the gate (actually already did, <=19 in gate queue)

  • once it's good, should allow to merge the 3 patches

    • Altered helm chart for policy-clamp-be | https://gerrit.onap.org/r/c/oom/+/124393

    • Add helm chart for policy participant | https://gerrit.onap.org/r/c/oom/+/124391

    • Add helm chart for http participant | https://gerrit.onap.org/r/c/oom/+/124390

  • routine patches should arrive in the near future

• Service Mesh: @Sylvain Desbureaux @Gareth Roper @Andrew Lamb @Adrian Matthews @Othman Touijer @Farida Azmy

  • SDC: @Gareth Roper
    
    

    • updated the chart

    • an issue is remaining on WFD BE ("true" instead of true)

  • SO: need to be reviewed

  • CPS: waiting for SDC for test on Service Mesh

  • CDS: need to be reviewed

  • SDNC:

    • first attempt from @Othman Touijer , some issue on the gate

    • need to change configuration java file for SDC client (https://gerrit.onap.org/r/gitweb?p=ccsdk/sli.git;a=blob;f=northbound/ueb-listener/src/main/java/org/onap/ccsdk/sli/northbound/uebclient/SdncUebConfiguration.java;h=4efc29b77ab9c5e0893a5c0e34fd3b674d1c5631;hb=refs/heads/master) in order to allow http for DMaaP MR and SDC

  • DCAE:

    • After Istanbul branching, plan is to remove Cloudify components, so no need to work a lot on DCAEGEN2 for now (DCAEMOD and DCAE-SERVICE can be reworked though)

  • service account: Farida Azmy
    
    

    • "Old" MR have been merged

    • new batch have been pushed:

      • [MSB] Update chart with service account | https://gerrit.onap.org/r/c/oom/+/124791

        • recheck because a lot of 502 errors on components that seems to use MSB

      • [AAF] Update chart with service account | https://gerrit.onap.org/r/c/oom/+/124512

        • recheck to be sure than Holmes issue is transient

      • [MULTICLOUD] Update chart with service account | https://gerrit.onap.org/r/c/oom/+/124856

        • wait for gate

      • [APPC] Update chart with service account | https://gerrit.onap.org/r/c/oom/+/124514

        • need to reenable APPC to be sure it works and now we wait for gate results

• Mariadb move to common instance: @Mahmoud Abdelhamid 

  • DMaaP and VFC have been merged

  • under wait of gating result:

    • [POLICY] Use common mariadb instance | https://gerrit.onap.org/r/c/oom/+/123718

  • Waiting for Job verify +1 and review from @Sylvain Desbureaux and @Krzysztof Opasiak :

    • [CDS] Update chart to use common mariadb-galera | https://gerrit.onap.org/r/c/oom/+/121936

  • Issue with nengdb and local instance: to be described by @Andreas Geißler 

• Postgresql move to common instance: @A. Seaudi 

  • most of them are not passing the gate yet

  • [DMAAP] Use common postgress for dmaap bc | https://gerrit.onap.org/r/c/oom/+/124773

  • [CPS] Use common postgres for CPS | https://gerrit.onap.org/r/c/oom/+/124463

  • [HOLMES] Use common postgres for holmes | https://gerrit.onap.org/r/c/oom/+/124781

  • [DCAE] Use common postgres for DCAE | https://gerrit.onap.org/r/c/oom/+/124776

• Cassandra move to common instance: @A. Seaudi 
  
  

  • [PORTAL] Migrate Portal to shared cassandra | https://gerrit.onap.org/r/c/oom/+/124345

    • waiting for @A. Seaudi to take into account remarks from @Krzysztof Opasiak 

• Mongodb: persistance is not well done : OOM-2864: Mongo data persistency not workingClosed

• AAF issue: when you have to many cert requests (~200), AAF doesn't provide new ones

  • AAF-1212: AAF CM limited number of certificate-requests cause ONAP components to hang Reopened

  • CronJob to clean Cassandra DB script is ongoing @Andreas Geißler 

• Gating and "future": @Sylvain Desbureaux 

  • today, spending on gating has reached $70k (whereas we had an approval for $50k till end of year)

    • knowing that staging azure is used and was not counted in

  • goal to shrink more gating and a limit is to have less than 64 PVCs per cluster (move from 5 * 32G nodes cluster to either 4*32G or 2*64G nodes cluster –> need to change pod limits as we have ~260 pods)

    • 1 is used for Minio

    • 1 is used for RWX support

    • 62 are left for ONAP

  • a lot of patches in queue these last days, (~20 all times)

    • maybe because now integration is getting gated

• Istanbul RC: @Sylvain Desbureaux 

  • Kubernetes version: a try on Azure with 1.20 shows issue with mariadb galera replication (replicas don't see the other ones)

    • let's stick with 1.19.11 for now

  • [COMMON] Bump ONAP version | https://gerrit.onap.org/r/c/oom/+/124843

    • waiting for @Krzysztof Opasiak review and +1

  • Post RC:

    • move to helm API v2 + version bump (@Fiachra Corcoran)

    • [GLOBAL] Migrate to helm v3 | https://gerrit.onap.org/r/c/oom/+/124842