Versions Compared


  • This line was added.
  • This line was removed.
  • Formatting was changed.
Comment: top of log was truncated. Corrected.


Dublin use cases/functional requirements1h55

All Slides for VF2F - 5G U/C are here: Release 4 (Dublin) Use Cases and functional requirements

under specific use case/functional requirement

Control Loop Sub Committee Dublin Requirements45 mins

We will be discussing these 2 functional requirements:

Ease of creating analytic components and on-boarding DCAE micro services

Model driven Control Loop Design

Security Requirements for Dublin Release 20 mins

Security Requirements for Dublin Release


Zoom Chat Log Dec 12 

04:54:01 From Ranny Haiby : #info Ranny Haiby, Nokia
04:55:55 From Andreas Geissler (Deutsche Telekom) : #info Andreas Geissler, DT
04:56:29 From Chaker Al-Hakim : #info Chaker Al-Hakim, Huawei
04:56:30 From Martin Skorupski (highstreet technologies) : #info Martin Skorupski, highstreet technologies
04:59:14 From Yan Chen : #info Yan Chen, China Telecom
05:00:29 From Kenny Paul (LFN) : #info Catherine Lefevre, AT&T
05:00:32 From Jimmy Forsyth to Kenny Paul (LFN) (Privately) : hi Kenny - the usual way I find meeting info is through the Community Meetings and Calendar page - now it’s saying “Unknown macro: Calendar” on that page and not loading the calendar
05:01:14 From Bin Yang (Wind River) : #Bin Yang, Wind River
05:01:41 From Yan Yang : #info proxy Lingli, CMCC
05:01:44 From Jason Hunt : #info Jason Hunt, IBM
05:02:19 From Srinivasa Addepalli (Intel) : #info Srinivasa Addepalli, Intel
05:02:59 From Stephen Terrill : #info, Stephen Terrill, Ericsson
05:03:06 From Eric Debeau : #info Eric Debeau, Orange

05:03:34 From Avi Chapnick : # info Avi Chanick , Amdocs (Proxy for Alla Goldner)
05:07:43 From Murat Turpcu ( Turk Telekom) : #info, Murat Turpçu, Turk Telekom
05:13:12 From Kenny Paul (LFN) : #topic Architecture Requirment Summary for Dublin Release
05:17:07 From Kate iPhone : switching from my phone to my ipad
05:19:44 From DENG Hui : so after 2 months, still no conclusion from Arch about NSD
05:20:33 From Byung-Woo Jun : I thought there was a vote for NSD last Friday. What is the outcome?
05:21:34 From DENG Hui : most people prefer NSD belongs to Service,
05:21:52 From DENG Hui : does it mean we need to support NSD definitely?
05:23:04 From Keong Lim k00759777 : disaggregated
05:26:43 From Kevin Scaggs : First step - determine what domain NSD is in, then how do we address it in that domian
05:31:56 From Keong Lim k00759777 : i don't think that "service" and "resource" should be in the name of any entity.
05:32:33 From Keong Lim k00759777 : @steve in slide 4, is kafka there separate from DMaaP?
05:34:27 From Keong Lim k00759777 : 2 more kafkas inside PNDA+?
05:34:51 From Xin Miao x00396303 : No. @Keong
05:35:47 From Srinivasa Addepalli (Intel) : @Keong, there are logical blocks trying to indicate the input to analytics and output of analytics is via Kafka. But, there will be only one Kafka broker.
05:37:57 From Catherine Lefevre : is it possible to have one extra slide including owners of each topics and impacted projects or link to impacted projects? thank you
05:39:00 From Kenny Paul (LFN) : #topic Release Process Improvements for Modeling
05:43:29 From Brian Hedstrom : We may want to change the granularity of this from Component to Module
05:44:29 From Kenny Paul (LFN) : #topic M3 Modeling checklist
05:45:11 From Srinivasa Addepalli (Intel) : @Seshu and @Catherine, these pages have owners, participating companies and projects impacts. FOr K8S based Cloud region support, please see this page: and Distributed anlaytics :
05:45:40 From Kevin Scaggs : Good observation, Brian.
05:45:44 From Seshu m s71826 : Thanks Srini
05:46:06 From Seshu m s71826 : My ask was a matrix with all the functions/usecases with eh projects involved
05:46:13 From Seshu m s71826 : to make it easier to follow up
05:46:32 From Seshu m s71826 : Including the owners
05:51:04 From Kenny Paul (LFN) : #topic Jira relationships with ONAP components
05:51:22 From Srinivasa Addepalli (Intel) : @Seshu, Yes, understood. IN case you want to look at it before that summary slide is created, informaion is available in above pages. :-)
05:52:20 From Seshu m s71826 : Thanks Srini :) .. Yes, this is a good start point for sure.
05:56:02 From Keong Lim k00759777 : is there enough time for modelling to be approved, then implement in AAI, then clients to use AAI?
05:57:03 From Brian : Models are functionality freeze
05:57:05 From Brian : M2
05:57:56 From Debeau : +1 with @Brian
05:58:39 From Catherine Lefevre : +10 with Brian
05:58:50 From Brian : M3 might be consider Model "as-built" reflecting what the api's implement
05:59:13 From Catherine Lefevre : M3 artefacts code ready
06:00:29 From margaret chiosi : Should modeling feed into VNF requirements which supposedly feeds into VNF SDK/VVP?
06:01:16 From Andy Mayer (AT&T) : Yes, the M2 Model Freeze is the design model, and M3 is the Final Model "as-built"
06:01:40 From Gildas Lanilis : @Andy. Correct.
06:02:13 From Debeau : I think modeling is broader than VNF requirements
06:03:41 From Debeau : service modeling is not included in VNF requirements
06:04:12 From Kenny Paul (LFN) : #topic OOM Priorities for Dublin
06:04:49 From Srinivasa Addepalli (Intel) : Need to drop now.
06:11:06 From Pavel Paroulek : What is the relationship between the proponenti Network policies and MSB? Aren’t they solving the same thing?
06:11:34 From Pavel Paroulek : proponenti = proposed
06:12:45 From Kenny Paul (LFN) : Must drop now to prep the Board Meeting. Gildas has the Host privs, with Steve T as co-host. Recording will continue and will be posted to the wiki.
06:13:21 From Debeau : Network policies are défined at pod level while MSB acts at component level
06:14:31 From Pavel Paroulek : @Eric so access will be regulated at both levels simultaneously?
06:14:37 From Brian : is control plane separation of etcd and orch affected ?
06:19:07 From Catherine Lefevre : Team - I am going to join the LFN Board, presenting our 2018 great accomplishments
06:19:08 From Debeau : @Pavel I believe yes
06:21:34 From Brian : coffee line chit chat mentioned GlusterFS :)
06:24:27 From Debeau : I like Mike Holly Grail
06:25:08 From Keong Lim k00759777 : does it include a rollback to casablanca too?
06:25:58 From Keong Lim k00759777 : does it mean some ONAP components are running at casablanca version while others are running at dublin version concurrently?
06:27:07 From Brian : do we have backup for any persitant storage apps so we can roll back to retest the upgrade ?
06:27:25 From Brian : need a design for test discussion
06:29:14 From Keong Lim k00759777 : once rolling upgrades are available, does it mean that ONAP releases become redundant?
06:31:48 From Keong Lim k00759777 : does it mean no breaking changes can be introduced? or they need to be smeared across several releases?
06:37:10 From Michael O'Brien(LOG, Amdocs) : Logging is working directly with OOM daily on deployment, monitoring
06:38:00 From Stephen Terrill : Great. I just want to avoid uncordinated concepts that can confuse.
06:38:17 From Michael O'Brien(LOG, Amdocs) : CLMAP and LOG are looking at using the common charts from OOM for ELK, there are several other users of E and K
06:39:44 From Brian : @michael O - we really need some training on the logging dashboard - did you add a default dashboard or set of dashboards for things like SO debug, SDNC debug, SDC debug ?
06:40:32 From Brian : +1 for moving helm charts to projects
06:41:36 From Brian : timeline for the repo changes ?
06:44:23 From Michael O'Brien(LOG, Amdocs) : Adding a dashboard in definitely in my dublin queue - I parallel with getting infrastructure going for the projects - see red in the diagram and aligning the libraries with a common schema see yellow containers that should be green - still verifying each of the 200+ containers -
06:44:24 From Michael O'Brien(LOG, Amdocs) :
06:44:30 From Michael O'Brien(LOG, Amdocs) :
06:45:20 From Pamela Dragosh : +1 - Policy can participate. Thanks Mike!
06:45:21 From Michael O'Brien(LOG, Amdocs) : still a lot of work - will need the help of the projects getting reviews in to align so the dashboard is usable for requestID and invocationID transaction tracing
06:47:11 From Michael O'Brien(LOG, Amdocs) : portal, policy, so, sdnc, aai have been very helpful - we are targeting portal/sdk as a candidate when I finish prototyping with it - thanks
06:47:43 From Keong Lim k00759777 : is HEAT deploy defunct now? is it OOM deploy only from now on?
06:48:27 From Michael O'Brien(LOG, Amdocs) : TSC-25 with the LF will also helm-deploy (verifying the docker version partially) - after Mike's already running helm-verify before submission
06:48:38 From Michael O'Brien(LOG, Amdocs) :
06:48:40 From Gildas Lanilis : @Keong, Yess OOM is the way to install onap
06:52:03 From Keong Lim k00759777 : will rolling upgrade become part of s3p?
06:52:24 From Michael O'Brien(LOG, Amdocs) : Anyone with issued during deployment - please post to the group, jira, ptls, meets - as we would like to cover off anything during deployment that causes inconsistent deploys - 51/51 HC ideally should be doable 10 times in a row - towards a defined dependency/priority startup order - so we minimize contention for ram/hd/ram on startup - mike is working this via several refactors but we should all help
06:56:03 From Brian : upgrade and geo-diversity seem like the highest priority from a SP standpoint
06:58:50 From Michael O'Brien(LOG, Amdocs) : yes, backup/restore/upgrade - also Reproduction of the deployment via several yaml overrides - so we can consistently bring up the system - currently about 50% of the time I may need to undeploy/deplly a couple components that missed their config window due to a more resource heavy pod coming in earlier than the last time - the timing adjustments in the yaml are a temporary workaround for now
06:58:58 From Brian : Wednesday 10 AM EST
06:58:58 From Michael O'Brien(LOG, Amdocs) : 10am EST wed OOM
06:59:13 From Michael O'Brien(LOG, Amdocs) :
06:59:45 From Michael O'Brien(LOG, Amdocs) : we need to sync config for all prod/demo deployments - yes - currently distributed in 2 repos
07:00:07 From Michael O'Brien(LOG, Amdocs) :
07:01:00 From Michael O'Brien(LOG, Amdocs) : we are listing the completed/restarted config jobs since 1.11 k8s - we can filter these out
07:01:11 From Michael O'Brien(LOG, Amdocs) :
07:01:59 From Michael O'Brien(LOG, Amdocs) : 3 components restarted manually in 3.0.0-ONAP yesterday on 1 of my 4 envs during a demo
07:02:04 From Keong Lim k00759777 : OOM-1542 "after few days (10-12 days) ONAP pods crash automatically"
07:02:39 From Brian : good job mike
07:02:48 From Michael O'Brien(LOG, Amdocs) : that is due to 3+ issues (FS past 80%, ram, etc ) - eviction of pods starts then - causing rescheduling
07:03:09 From Michael O'Brien(LOG, Amdocs) : Mike is Excellent!!!! to work with
07:03:59 From Yang Xu : Great and important platform features, let’s work together to deliver them
07:04:01 From Alain Soleil : Glad to hear the consitent OOM deployment topic will be handled and managed as a priority, it is critical getting ready for production ONAP can be deployed and operated with success all the time
07:04:26 From Michael O'Brien(LOG, Amdocs) : looks good
07:04:34 From Brian : I can read it but that is better
07:04:44 From Stephen Terrill : #topic: SO deployment and troubleshooting for beginners, Seshu
07:05:24 From Keong Lim k00759777 : is that gitlab page also on onap wiki?
07:06:04 From Gildas Lanilis : @Keong. No. Not sure why GitLab is being used here. This is a surprised to me.
07:06:44 From Pamela Dragosh : Should be in readthedocs
07:07:05 From Gildas Lanilis : Yes. eithe rin ReadTheDocs or ONAp Wiki.
07:08:31 From Gildas Lanilis : As we were discussing weekly meeting for OOM and Integration, weekly meeting schedule is available here:
07:09:07 From Keong Lim k00759777 : seshu you have 2 microphones going now...
07:09:21 From Keong Lim k00759777 : stereo sehus
07:09:23 From Viswa KSP : @Kenny - I see 2 Seshu
07:09:39 From Stephen Terrill : ill fix
07:10:00 From Stephen Terrill : Both are unmoted
07:10:01 From Gildas Lanilis : @visma. That is bcoz Seshu is connected form 2 different devices.
07:10:15 From Michael O'Brien(LOG, Amdocs) : yes, we may want to use the official gerrit/git from ONAP Linux Foundation - not internal company infrastructure like this gitlab site - just a suggestions Seshu - "work in public..."
07:10:22 From Viswa KSP : @Gildas - Got it !!
07:10:26 From Stephen Terrill : #info. Seshu will updload to the wiki (send email) and ensure its in in readthedocs
07:10:49 From Viswa KSP : @Steve - Earlier during Orange’s session on onap
07:10:57 From Viswa KSP : Snap-tests, I did see Gitlab
07:11:08 From Michael O'Brien(LOG, Amdocs) : Any OOM documentation should be done directly in the OOM wiki - I myself do not use my corporate wiki for anything
07:11:09 From Viswa KSP : was that also contributed to Wiki / Gerrit ?
07:11:57 From Gildas Lanilis : @Visma. I do not think. Ut we have asked Seshu to publish in wiki or readthedocs.
07:12:01 From Brian : I think Dublin will be OOM only
07:12:07 From Stephen Terrill : @viswa. I think slipped by. however all material used should be uploaded to this event wiki. That is the accepted and polite practice
07:12:41 From Viswa KSP : @GIldas / @Steve - Thank you!
07:12:46 From Michael O'Brien(LOG, Amdocs) : Seshu I am having issues with SO deployment - sensitive to startup order - please post any workarounds in public only
07:13:30 From Brian : Note that integration-override.yaml overrides some of these settings for your environment like a nexus cache and pullPolicy
07:13:41 From Michael O'Brien(LOG, Amdocs) : you are missing a "make onap" in your wiki
07:14:16 From Michael O'Brien(LOG, Amdocs) : the docs are incomplete or need to be resyncd with public
07:14:55 From Brian : +1 for what Mike said
07:17:13 From Stephen Terrill : #info Some feedback in terms of make all, make ONAP; repos.
07:18:08 From Michael O'Brien(LOG, Amdocs) : ok, you had internet issues - was just thinking your team used this wiki ongoing for issues - a 1-time upload is fine
07:19:07 From Andreas Geissler (Deutsche Telekom) : What about configuration info, e.g. how to add an additional cloud instance to the OS Adapter or is this not required anymore after ESR VIM addition ?
07:19:27 From Viswa KSP : This is the most practical way for all lazy dev folks like me….
07:19:33 From Brian : that only works locally not in an OOM deployment
07:20:14 From Marcin Przybysz (NOKIA) : what with issues tracking while pod is not yest runnig, exaple in init state? how we can find out promlem?
07:20:26 From Michael O'Brien(LOG, Amdocs) : btw seshu and I are buddies and have had dinner twice - I am sorry for being hard on him - i am still in the PST time zone
07:20:48 From Viswa KSP : About cloud-config.json ??
07:21:14 From Brian : there is a REST interface to add cloud_sites as well
07:21:29 From Michael O'Brien(LOG, Amdocs) : use a -f over....yaml
07:21:49 From Michael O'Brien(LOG, Amdocs) : mikes
07:21:49 From Michael O'Brien(LOG, Amdocs) :
07:23:21 From Stephen Terrill : #info good info, good feedback to come.
07:23:51 From Michael O'Brien(LOG, Amdocs) : helm undeploy/deploy - via mikes page using a --set.enabled=false/true as well as a 2nd way
07:24:09 From Michael O'Brien(LOG, Amdocs) : --set so.enabled=true/false
07:24:10 From Brian : mike --set.enabled doesnt work well for us in integration
07:24:17 From Michael O'Brien(LOG, Amdocs) : mike is better - thanks Seshu
07:25:36 From Viswa KSP : @Seshu - Could you also put some info on live debugging into container mapping source…
07:25:42 From Viswa KSP : If it is dev oriented page…