2018-12-20 AAI Developers Meeting Open Agenda

Davide Cherubini

Susana Sabater

OPEN

During the VF2F we have presented the CCVPN Use Case extension for Dublin Release.

In particular, we have defined the E-LAN Service (EP-LAN. EVP-LAN) sub-Use Case. More details can be found at this page (the presentation given to the Use Case SubCommittee and during the VF2F can also be found in that page).

We have also created 2 additional pages where we describe the:

1. Implementation Options
2. Initial Impact assessment

In the former page, we have identified Option 2 (for Service Creation) and Option A (for Service Modification) as the “ideal” implementation. The Impacts in the latter page are based on such an assumption.

We would be grateful if you could help us with the assessment for A&AI and if you could confirm the project commitment to CCVPN.

14th Dec: Regarding the Options and Impact assessment linked above:

AAI

1. Support Composite Service Instances (multi-operator)

As far as I know, the ServiceComposite pattern hasn't even exited the Modelling team's discussions for R3 or R4 workshops Hui Deng :

https://wiki.onap.org/display/DW/Composite+Pattern+UML+diagram

This would seem like a stretch goal given the impacts on all the downstream clients of AAI, if it would require reworking any existing schema elements.

However, if you are looking for a zero-impact method of implementation, the EdgeRules already contain service-instance-to-service-instance relationship with label "org.onap.relationships.inventory.ComposedOf", so AAI can do the composite data with this relationship.

Former user (Deleted)

IN PROGRESS

1st Nov 2018

Why is the pod for HAproxy not named (hard to figure out that there is a proxy), unsure how it is logging and where

James Forsyth creates JIRA tasks to 1. to have the pod named 2. add logging to the proxy

AAI-1807 - Getting issue details... STATUS AAI-1810 - Getting issue details... STATUS

Keong Lim

OPEN

Just had a look at this DMaaP bug DMAAP-896 - Getting issue details... STATUS which shows up two AAI-related exceptions when getting events:

INFO  org.onap.dmaap.dmf.mr.service.impl.EventsServiceImpl - KafkaConsumer is not safe for multi-threaded accesson AAI-EVENT datarouter ****** datarouter from Remote10.42.133.135
INFO  org.onap.dmaap.dmf.mr.service.impl.EventsServiceImpl - [AAI-EVENT/datarouter/datarouter] 

INFO  org.onap.dmaap.dmf.mr.service.impl.EventsServiceImpl - KafkaConsumer is not safe for multi-threaded accesson champRawEvents spike ****** spike from Remote10.42.121.41
INFO  org.onap.dmaap.dmf.mr.service.impl.EventsServiceImpl - [champRawEvents/spike/spike]

Anyone else seen or reported similar problems in AAI?

OPEN

Michael O'Brien has documented performance issues in aai-cassandra:

• Cloud Native Deployment (search for LOG-376 to find the specific references to AAI)
• https://jira.onap.org/browse/LOG-376?focusedCommentId=29358&page=com.atlassian.jira.plugin.system.issuetabpanels:comment-tabpanel#comment-29358

hector has discovered that the stress test jar (liveness probe?) in aai-cassandra is hammering the cpu/ram/hd on the vm that aai is on - this breaks the etcd cluster (not the latency/network issues we suspected that may cause pod rescheduling)

Is there something that should be tweaked in AAI config? Or documentation on the recommended setup to run the VM?

I'll come to the next AAI meet (conflicts with pomba meet) -

LOG-376 - Getting issue details... STATUS

Keong Lim

OPEN

Mike Elliott wrote in OOM Meeting Notes - 2018-12-5

f. AAI team wanted to get notified of AAI Casandra issues automatically
                i. Can we setup a Nagios or equivalent to monitor both rancher/k8 and the applications for rancher/k8 issues ?

Keep an eye out for new issues!

OPEN

Dénes Németh wrote in AAI-1104 - Getting issue details... STATUS :

In think it would be good to answer what is the meaning of the field (collection of PEMs of the CA xor URL)

Questions:

1. Is AAI intended to strictly prescribe how the fields are used and what contents are in the values?
2. Or does AAI simply reflect the wishes of all the client projects that use it to store and retrieve data?

Even if (1) is true, AAI is not really in any position to enforce how clients use the data, so really (2) is always true and we need to consult the original producers of the data and the ultimate consumers of the data to document their intended meanings.

How do we push to have documentation on the purpose and meaning of the fields in AAI?

Where does all this documentation go?

Should the documentation be backed up by validation code?

See also discussion about AAI in 2018-11-28 ExtAPI Meeting notes

29th Nov: Started on new wiki page AAI Schema Producer-Consumer Pairings

OPEN

While helping UUI team to debug a family of related issues such as AAI-1915 - Getting issue details... STATUS , found that their JSON request bodies had inserted extra lines that were not acceptable to AAI, e.g.

{
  ext-aai-network: {
   "aai-id": "VDF",
    "schema-version": "version-1",
    "resource-version": "1542082337501"
  }
}

vs

{
  "aai-id": "VDF",
  "schema-version": "version-1",
  "resource-version": "1542082337501"
}

Initially, I thought this could be due to XML-to-JSON translation error, but it could also have been a copy-paste from the output of a GET, e.g.

> GET /aai/v14/network/ext-aai-networks HTTP/1.1
{
  "ext-aai-network": [
    {
      "aai-id": "createAndDelete",
      "schema-version": "version-1",
      "resource-version": "1542247826990"
    },
    {
      "aai-id": "aaiId-2",
      "schema-version": "version-2",
      "resource-version": "1542029867153"
    }
  ]
}

In the spirit of Postel's Law ("be liberal in what you accept"), could/should AAI accept both variations of the input data above?

There could be additional validation that the element name inside the request body matches the element name in the URL of the API (similar validation is already performed for the key ID value).

It would also allow for simple methods of data transfer/migration, where AAI output is directly accepted as AAI input.

7th Dec: As per Jimmy's comment below, the difference between the "Not Accepted Input" and "Expected Input Format" is the behaviour controlled by "SerializationFeature.WRAP_ROOT_VALUE" as described in:

• https://stackoverflow.com/questions/37882478/serializationfeature-wrap-root-value-as-annotation-in-jackson-json
• https://fasterxml.github.io/jackson-databind/javadoc/2.6/com/fasterxml/jackson/databind/SerializationFeature.html

Former user (Deleted)

OPEN

Former user (Deleted)

OPEN

OPEN

OPEN

Looking at AAI usage in OOF - HPA guide for integration testing by Dileep Ranganathan, wondering whether there is a better way to bootstrap AAI test data?

Generating AAI data

Note: Required only if the Multicloud has no real cloud-regions and HPA discovery cannot happen.

If Multicloud team has data for creating the Cloud-region and doesn't have the HPA, then please update the existing data with the flavors with HPA.

1. Import the postman collection CASABLANCA_AAI_postman.json
2. To add/remove HPA Capabilities edit the flavors section in the body of PUT Cloud-Region{x}
3. Once all the necessary Use postman to add the complex and cloud regions in the order specified below
   (snip screenshot of specific sequence)
4. Use the GET requests to verify the data.
   (snip screenshot of specific sequence)

Similarly, Scott Seabolt and J / Joss Armstrong wrote for APPC Sample A&AI Data Setup for vLB/vDNS for APPC Consumption and Script to load vLB into AAI:

The below put_vLB.sh script can be used to submit the vLB data to A&AI in order to run ConfigScaleOut use case. This script and referenced JSON files are used on an AAI instance where the cloud-region and tenant are already defined.

Similarly:

• TEST-133 - Getting issue details... STATUS
• INT-705 - Getting issue details... STATUS
• vCPE Use Case Tutorial: Design and Deploy based on ONAP

Related AAI-1948 - Getting issue details... STATUS on the brittleness of the ReadTheDocs links to data files.

One for VIM: How-To: Register a VIM/Cloud Instance to ONAP and AAI-1928 - Getting issue details... STATUS

Potential issues:

• fragility of static import data file w.r.t. schema changes and version upgrades for each ONAP release?
• how "common" is this knowledge, i.e. what to load, where to get it, who else should be using it, etc?
• should it be automated/scripted, rather than manual steps to bootstrap?
• should it be a simulator program or test harness, rather than a static data file?
• should it reside within AAI CI/CD jobs for maintenance and upgrade of schema versions?
• who maintains the data itself? Is there a "data repository" which can be delegated to other teams, e.g. like documentation repository links in git?
• how many other teams have similar private stashes of AAI bootstrap data?
• does it need to be published at a stable URL to avoid linkrot?

OPEN

Under OOF Homing and Allocation Service (HAS) section, Dileep Ranganathan wrote about Project Specific enhancements:

Optimize - AAI cache

• Use MUSIC or any other alternative in memory caching like Redis etc?
• Optimize flavor retrieval from A&AI and Cache the information if necessary

See also OPTFRA-268 - Getting issue details... STATUS / OPTFRA-291 - Getting issue details... STATUS

Similarly to the "AAI too slow for Holmes" item below, this introduction of extra caching of AAI data is a worrisome development and sad indictment of the performance of the system architecture.

What can we do about this?

Would the AAI Cacher AAI-1337 - Getting issue details... STATUS help to improve performance?

OPEN

Bin Yang and Lianhao Lu (Deactivated) wrote in MULTICLOUD-274 - Getting issue details... STATUS :

HPA telemetry data collection and make it persistent in A&AI, from which OOF can leverage during its decision making process.

1.  Multi-cloud to collect the data from time-series data services like Prometheus (http://prometheus.io) or openstack Gnocchi, and push them to A&AI based on the data recording & aggregation rules.

and

The reason why we propose here is that VES mechanism doesn't store the telemetry data into A&AI. And OOF now can only get those kind of data from A&AI.

Some concerns:

• how much additional load will this place on AAI?
• will AAI cope with this load?
• is AAI suitable for "time-series data"?
• is "telemetry data" considered to be "active & available inventory"?
• should OOF access the telemetry/time-series data via other means (not AAI)?
• AAI API latency (4~6 second per request as benchmarked in CMCC lab) could be a problem

OPEN

Under POMBA Common Model, Geora Barsky and Sharon Chisholm discuss objects that seem to overlap the AAI schema:

• Network has direct relationships with PNFs and physical links
• PNF has direct relationship with p-interfaces and l-interfaces

Curious to know what is the relationship between the POMBA Common Model and the AAI schema? There seems to be an overlap in these object definitions and relationships.

Is POMBA a potential client for AAI schema services?

15th Nov: As per Geora's and Sharon's comments:

POMBA is a client of AAI APIs. It retrieves certain objects from AAI and transforms it into POMBA COMMON model which is aimed to represent flat structure of service instance representation.

and

The POMBA model enables us to normalize data from different data sources to facilitate auditing.  To save time and work, we often use A&AI as our starting point, but conceptually, this model can have more than is currently or makes sense to have in A&AI.

So, POMBA needs to evolve independently of AAI, even though it has common ancestry.

As we are having discussions about the AAI Schema Services with a view to future dynamic schema updates, e.g. via SDC modelling, we will need to be aware of the downstream impacts of such a change.

I think this needs to be added to the use cases for AAI Schema Services (and GraphGraph?) behaviours.

Keong Lim

OPEN

Keong Lim

OPEN

Comments on Orchestration Scenarios related to AAI:

Viswanath Kumar Skand Priya / kspviswa said:

Thank you Ranny Haiby & Fernando Oliveira . I agree partly, but I still have following queries.

• I agree & acknowledge that atleast for a foreseeable future, we would need a way to specify the VNFM / NFVO as part of "Design Decision", which I believe can be reflected as part of VNFD/NSD ( using some special attribute ) or as part of internal Model that SDC might build before distributing the same. SO can then use this hint to select relevant actors. My only question is, why this has to be maintained in AAI which is exclusively for runtime record? All AAI cares about is what is running in the network irrespective of how that got orchestrated. Isn't it ?

On a broader note, I would like to understand what's the original intent of AAI ( atleast in ECOMP world ) ? Are we simply assuming that, just because AAI has "available inventory" in its name, we are expecting it to keep track of cloud inventory realtime ? Because our entire story ( including the new G-FPS proposal ) is based on this assumption. Can anyone from AAI team or ATT clarify on this ?

Because AFAIK, AAI neither has the schema to host such available inventory, nor the MC has the pub/sub or polling mechanism ( today ) to refresh the cloud inventory inplace. Ofcourse those can be scoped for further releases, but my original question is, was that the original intent behind AAI or are we now including it in the scope?

and Fernando Oliveira replied:

For the first question: I think that A&AI needs to maintain the VNF instance ↔ VNFM instance and the NS instance ↔ NFVO instance relationship for subsequent life cycle operations, i.e. a scale or heal operation.  The path would be something like Event (VNF Instance, Busy)  → DCAE (policy for VNF instance) → Policy Evaluation (VNF instance, Scale-out)  → SO (VNF instance, Scale-out) → A&AI (find VNFM instance for the VNF instance) → SO (VNF instance, VNFM instance, Scale-out) → SOL003 Adapter (VNFM instance, VNF instance, Scale-out) → VNFM instance (VNF Instance, Scale-out).

As I understand, ESR has "esr-vnfm-list", which has an "esr-vnfm", which has "esr-system-info-list", which has "esr-system-info", which has a "relationship-list" that can contain relationships to "generic-vnf" and other AAI objects.

The "generic-vnf" object also contains "self-link", "ipv4OamAddress", "ipv4OamGatewayAddress", etc, which links the AAI object back to its "source-of-truth" external-system.

Is there some new data, new schema or new API that is required on top of this?

Fernando Oliveira; Apologies for my lack of knowledge, but a few comments:

1. For the VNF/VF ↔ VNFM case, I think that there needs to be a reference from a VNF/VF instance record to the specific instance of the VNFM that was used to deploy the VNF/VF. If there is already such a reference from the VNF/VF through the ESR to the specific item on the esr-vnfm-list, then I think that would be sufficient. If not, I think that would be a new requirement.
2. For the Service ↔ NFVO case, Is there an equivalent NFVO/Orchestrator list in the ESR? The esr-nfvo-list would need the same set of info as the VNFM case. If the esr-nfvo-list does exist, I think that there needs to be a reference from the Service Instance record to the specific NFVO instance that deployed the Service. Is there such a reference? If not, I think that would be a new requirement.

Bo Lv can comment more on the current ESR capabilities, but I believe there are only 3 kinds of systems so far: EMS, VNFM and third-party SDNC.

ESR could be extended to handle VNFO as another kind of system.

Keong Lim

OPEN

Adolfo Perez-Duran (Deactivated) / Adolfo Perez-Duran raised INT-769 - Getting issue details... STATUS and its many sub-tasks.

Most, but not all, of AAI container code is Java, so is it really feasible to be both platform-agnostic and CPU-agnostic? What is the actual range of platforms and CPUs targeted?

Keong Lim

IN PROGRESS

Modelling team having Service Instance thoughts by Chesla Wechsler, which will affect AAI schema.

Also referred from comments on ONAP R4+ Service Modeling Discussion Calls

9)“vhn-portal-url”?“Bandwidth”,"QoS","SLA",etc, attribtutes that not all the services need but still need to be stored in certain service instance: stored as a schemaless field on the service-instance vertex (Chesla will follow up) (my concerns: according to the call, is that ok if we set a "global-type of service" and a "customized-type of service", then mapped it with internal descriptor, and A&AI's model only stores global type in service instance's schema, but stores the customer-faced attributes of service in a schemaless way? Chesla Wechsler Kevin Scaggs Andy Mayer)

See also Modeling 2018-11-13

The service-instance already uses a "metadata" relationship, which can store an arbitrary list of key-value pairs, but perhaps AAI should extend the use of the "properties" element, which is also an arbitrary list of name-value pairs or the "extra-properties" element, which is also an arbitrary list of name-value pairs.

15th Nov: Having seen Chesla's presentation, it should be called "Model-driven schema" rather than "schemaless" behaviour, since the idea is that the changes are controlled by SDC modelling. Seems aligned to the eventual goal in AAI Schema Service Use Case Proposals and AAI Schema Service.

Team

Keong Lim

Tian Lee

ON HOLD

Security subcommittee has recommended teams move away from jackson, and will be presenting alternatives and asking for an assessment from each project. Our team will need to do an analysis - this would not be trivial, especially given how many of our repos are impacted. As of now, this would be a very high LOE for the team, we need to understand what the recommendation from the SECCOM is before we can provide better details on what the LOE would be.

Updated: Using Google gson vs FasterXML Jackson

10th Oct: Present to Seccom meeting

15th Oct: Present to PTL meeting

31st Oct: Debatable whether the cost of swapping Cassandra and changing code is worth the benefit of removing Jackson from the vulnerabilities list.

On-Hold until James Forsyth consults with other PTLs: PTL 2018-11-05

10th Dec: Mentioned as S3P in vF2F meeting: ONAP Project Developers Event, Dec 10 - 12, 2018, (Virtual Webinars)

Keong Lim

CT Paterson

IN PROGRESS

Guangrong Fu mentioned AAI in Baseline Measurements based on Testing Results:

1. Cache the AAI data and refresh them periodically so that Holmes won't have to make an HTTP call to AAI every time it tries to correlate one alarm to another.

The problem for caching is how to know when to update the cached data. Even though the access time may be fast for Holmes, the risk is using out-of-date data, so the correlations will be wrong anyway. Also, duplicating the AAI data outside of AAI is probably a bad architectural decision. Making AAI faster for these use cases would be better.

Has there been a performance analysis of where the time is spent? Could it help to use ElasticSearch (e.g. as in sparky)? Should Holmes have a batch interface to get more AAI data in fewer calls? Or a better correlation API that results in fewer calls?

31st Oct: https://lists.onap.org/g/onap-discuss/topic/27805753

1st Nov:

• Guangrong Fu will try custom queries for queries that took to long to return
• The hardware (mainly storage) influences the query speed - need to find out what hardware was the speed test conducted on (Guangrong Fu will provide HW specs)
• HOLMES-186 - Getting issue details... STATUS

Would the AAI Cacher AAI-1337 - Getting issue details... STATUS help to improve performance?

Former user (Deleted)

ON HOLD

There are 2 types of logging in the services

• one read from EELFManager
• the other Logger log = Logger.getLogger( ...

Is that correct? Shouldn't there be just 1 type?

1st Nov:

After Casablanca release investigate logging guidelines and figure out what library to use in order to unify logging within A&AI

26th Nov: See also ONAP Application Logging Specification - Post Dublin

29th Nov: how does this fit with LOG-877 - Getting issue details... STATUS ?

Former user (Deleted)

IN PROGRESS

Could we disable unused (i.e. not integrated) A&AI web services, so that the deployment is faster and the resource footprint is smaller? e.g. Champ (any other ws?)

Motivation: Decrease the resource footprint for A&AI (ONAP) deployments

Idea: we could support 2 different deployments 1. full (normal) deployment and 2. barebones deployment. The point of the "barebone" deployment would be to deploy only the essential services necessary for proper functioning of A&AI (leaving out services like cacher, sparky, graphadmin, having 1 cassandra node instead of 3 or 5 etc).

In order to reduce hardware/cloud costs (mainly the memory footprint) it could be beneficial to support a minimalistic A&AI deployment.

1st Nov:

Venkata Harish Kajur Former user (Deleted) - investigate how to disable/enable charts in A&AI so we can create a core group of pods which handle the use-cases and than extended group will all the services. Consider a group of unused/unintegrated services (like Champ). Consider other possible groups (like GUI?)

IN PROGRESS

IN PROGRESS

Keong Lim

IN PROGRESS

Following updates for AAI Developer Environment Setup - Casablanca need to update the link in step 14:

Tutorial: Making and Testing a Schema Change in A&AI

• Need separate pages for Beijing, Casablanca, Dublin, etc.
• AAI-1973 - Getting issue details... STATUS

 Harish will update the documentation - the section Tutorial: Making and Testing a Schema Change in A&AI

Need to distinguish on the wiki tutorials for:

• running A&AI (just installing for non-devs)
• seting up a dev environment to develop AAI ("Developer Onboarding" section on wiki)

James Forsyth will check if we still need to support HEAT deployment for Dublin

Manisha Aggarwal

IN PROGRESS

Discuss about the Schema Microservice

AAI Schema Service

11th Oct: Suggested Use Case Proposals for Dynamic AAI Schema Changes based on CCVPN usecase experience

1st Nov: William Reehil Robby Maharajh Venkata Harish Kajur will review requirement updates and research the open questions so that a final draft can be prepared for implementation

8th Nov: Added AAI Schema Service Use Case Proposals for discussion and planning

15th Nov: Reviewed the Requirements section in AAI Schema Service again.

Keong Lim

IN PROGRESS

William Reehil wrote introduction to (proposal for?) A&AI GraphGraph

What must the solution provide?

• a generic interface to interact with the schema and edge information, so it can be accessed via end users and microservices
• ability to be configured with any schema given in the set formats (JSON for edge rules, OXM for schema)
• ability to easily communicate to an end user a node type’s attributes and edge rules when provided with a node type as input

Looks like an API to reflect on the schema from an instance in the database.

Is there some overlap with the AAI Schema Services?

Is this leveraging AAI Schema Services e.g. as client? proxy? facade? implementation detail?

Is it an alternative to AAI Schema Services?

Update: Keong Lim I do see an overlap with the schema service mainly for the retrieval of the data, but GraphgGraph will offer more on top of that(UI, NLP), during the POC code for GraphGaph there was no schema service. Implementation details can be discussed on our call for how exactly to leverage the schema service, the poc code, and where the core logic resides for this functionality.

15th Nov: Reviewed A&AI GraphGraph and agreed that there is now overlap with AAI Schema Service that should be addressed in updating the POC version.

IN PROGRESS

James Forsyth

Bo Lv

ESR in Dublin will be bug fix and non-functional requirements.

Should ESR use AAF for RBAC in Dublin?

Clients of ESR are MultiVIM and VF-C

External system passwords are not encrypted in AAI

During registration of cloud-region, ESR calls multi cloud, but when user is deleting data from the port ESR does not call multi cloud but should (feature request?)

James Forsyth will establish a weekly AAI/ESR call - 9 PM EST 10 AM China time for several weeks to get the project on track for Dublin.

James Forsyth will create a session next Thursday 19.12. 10 AM China time for Bo Lv

Keong Lim

IN PROGRESS

Yang Xu wrote in AAF Component:

Issues and Workaround

If AAF stops working, e.g. AAI returns authorization error, bounce AAF aaf-locate and aaf-service containers. 

This looks like some problem for resiliency and healthchecks. Is there a JIRA case for this already? An earlier case mentioned the "fix" in passing AAI-1762 - Getting issue details... STATUS . Also in AAI-1772 - Getting issue details... STATUS

Helen Chen, is there a scenario for this in Casablanca Release Resiliency Testing Status?

Jonathan Gathman, Mike Elliott, how can this remedy be automated?

See also AAF-599 - Getting issue details... STATUS

• James Forsyth will ask Jonathan Gathman about this item.

IN PROGRESS

AAI-1989 - Getting issue details... STATUS

APPC-1245 - Getting issue details... STATUS

POLICY-1278 - Getting issue details... STATUS

VID-355 - Getting issue details... STATUS

Need to replace custom queries currently in use by these systems (and others?) in Dublin toward the retirement of the named-query API in Dublin

IN PROGRESS

AAI-1940 - Getting issue details... STATUS

• James Forsyth will apply the changes he outlined in AAI-1940

IN PROGRESS

IN PROGRESS

Invite Andrew Muller CT Paterson or other Amdocs people to talk about A&AI consistency (https://docs.janusgraph.org/latest/eventual-consistency.html) and how to resolve the issue (for example 48529533). Ideally on 20th Dec 2018 on the dev meeting.

The idea in Beijing was that since we are going to use Champ, we would get rid of inconsistency problems we have currently in Janus - clarify it this is true, if not figure out an approach how to deal with this issue.

Update 13th Dec : Agenda will probably be discussed on the AAI weekly on Wednesday

IN PROGRESS

James Forsyth

IN PROGRESS

https://spring.io/blog/2018/11/29/spring-boot-1-5-18-available-now - support for 1.5.x spring boot will end in Aug 2019 so we should consider upgrading.

We should update to spring 2.X.X.

On the meeting on the 10th Jan discuss if anyone has anything against it, if not then James Forsyth creates a JIRA item for Dublin for upgrading the spring in all A&AI repositories.