Versions Compared

Version Old Version 52 New Version Current
Changes made by

Gilles THOUENON

Gilles THOUENON

Saved on

Key

  • This line was added.
  • This line was removed.
  • Formatting was changed.

Meetings connection details

available at TransportPCE meetings
https://wikilf-opendaylight.opendaylightatlassian.orgnet/wiki/display/ODL/TransportPCE+meetings

...

  • transportpce changes under review:
    • Relation chain from Gilles related to the feature cleaning (“Fix Feature issues”, “Add artifacts to transport-pce”, “Add features -transport-pce”, “Clean for Maven files”) is comparable to what has been done in tpce/models. This Changes are running and could be merged.
    • “Increase project version” prepares S-SR4 iteration (should be released in case of future security issues)
    • “Handle missing LLDP nbr-list” could be merged
    • “Autonomous impairment aware path computation” proposed by Olivier. It complements the work initiated with catalog primitives. It allows tpce to do impairment aware path computation in accordance with the OpenROADM specifications
      • At least the PCE wiki page or/and the documentation will need to be updated with this important functional increment.
  • integration/distribution change under review:
    • “Add transportpce projects to distribution”: Gilles prepares a new change to replace the first attempt of tpce and tpce-models integration. It sticks much better to the documentation. Remains the tpce versions to adjust (with or without snapshot termination?)

...

  • The "developer guide" need to be updated because Json samples used for the service create are not integrating the evolutions of models to 10.1 (the documentation sill uses the 5.1 models). A JIRA user story will be created for next sprint.
  • One bug reported by Jonas in last August: https://jiralf-opendaylight.opendaylightatlassian.orgnet/browse/TRNSPRTPCE-681. Gilles replied today…

  • Last part of the meeting was dedicated to Sprint29 planning:

    • Bala mentioned that we should consider, not for this sprint but the sprint after, refactoring the PCE to better consider equipment capabilities and select the most appropriate operational-modes

    • Olivier will split the story proposed by Javier “T-API Renderer” because it includes also some functionalities to retrieve a topology through a SBI

    • Gilles will add a story dedicated to the move of the models to a dedicated project.

    • Sprint starts on the 19/9 and ends on the 13/10.

...

  • All changes merged during the week had already been reviewed during last meeting, except “Refactor networkmodel OpenRoadm topology part 1 to 3”, which are dedicated to solve issues highlighted by SONAR. ”Upgrade Honeynode sim version” provides latest version of Honeynode that includes latest changes associated with the handling of PMs by HoneyNode. This version is required to perform the last functional test suite proposed by Thierry.

  • Discussion about changes under review:

    • T-api Relation chain has been rebased by Javier. The 4 first changes at the bottom of the stack, up to “T-API notification functional test” could be merged if we don't solve open Gilles's comment on change 97438.

    • In ”Migrate OTN E2E functional tests to RFC 8040 step 1/2”, Guillaume initiated the work to migrate to RFC 8040 the E2E tests associated with the OTN services

    • ”Switch tests from nose to pytest” will help us handling nose deprecation with Python next versions

    • “Refactor networkmodel OpenRoadm topology part 4” is currently under analysis. Discussion on Gilles's comment, since Guillaume doesn't want we use Table structure available from odlparent dependency.

    • ”Replace tpce-topology yang by existing ordmodels” has been updated and Thierry fixed some typos

    • ”Add tests for autonomous service rerouting” is a new functional test suite which maps to the Orange demonstration under preparation. A PM going out of range in HoneyNode will generate a port state change which will trigger the corresponding notification, resulting in a service state change in the tpce service DataStore for 2 impacted services. Only one of them being re-routable. We then check that this last is rerouted correctly by transportPCE. Currently in Honeynode, when a PM is going out of range, we only manage the state change of the port, but don’t update the state of the interface. Bala mentions that interface state change is linked to the type of Alarms. In case the alarm is something like LOS then the interface shall be also out of service (och/otsi).Salem from UTD suggested to implement the change of the interface state.

  • Sprint kanban review:

    • TRNSPRTPCE-655 on Lighty.io simulator: Christophe wanted to support different model types. But as soon as some dependencies to OR models does not include the revision, lighty by default takes the latest one as the default. This being solved, when tpce is connected with lighty for interface creation, it creates loops and the Netconf connection is not released properly.  So this still needs to be solved.

...

  • changes merged durint the week:

    • All the relation chain concerning sulfur migration

    • “fix functional test issue when list in dict” allows reordering elements of a list in a dictionary

    • "pce(gnpy) for Sulfur”

  • Discussion about changes under review:
    • Thierry's relation chain:

      • “Fix bug service-create”: This is a bug dsicovered when testing the Orange demo. A service creation with an existing service name was overriding the service which has the same name, with no possibility to change anything associated with the first service. This fix introduce some check on service-name consistency

      • “Update SH test”: There were a lot of redundant variables in SH tests. This change addresses the issue and introduces a new test associated with the service name, to check its consistency.

      • “Add service-resiliency handling in service-create”: adds the handling of service-resiliency attribute in the input of the service-create rpc. Thierry addressed the comment made by Guillaume and Gilles.

      • Thierry is also working on a new functional tests suite associated with the whole relation chain corresponding to service restoration, with the goal of automating the demo we planned for the next Research Exhibition Demo in Orange. The topology will include two Xponders and 3 ROADMS. Two services will be created, only one of them being restorable. They will use the same path. A degradation of the line will be simulated on one of the node so that it goes out of service. Only the service declared as restorable will be restored by TransportPCE controller.

    • T-API relation chain:

      • Javier rebased his changes, but this needs to be done again, since the migration to Sulfur was done after.

      • Gilles tested the first 2 changes and provided some editorial comments. These changes could be merged in a short time.

      • The third change brings some issues. Gilles made some comments that will need to be addressed. Javier will try to do it beginning of next week to solve the loop issue, but he will be off part of the week.

      • Gilles could work on the change on the top of the relation chain since it is not linked to the others.

      • It should be possible top merge some of the T-API changes to be part of Sulfur.

    • Quentin proposed some changes related with the code styles. “Refactoring code issues in OpenRoadm Topology” is one of those. Some of the code style proposal might be questionable. This is the reason why it was propose to Guillaume to write the code style rules somewhere so that the contributors can check them and validate/invalidate them. Quentin made a first proposal on the wiki (https://wikilf-opendaylight.opendaylightatlassian.orgnet/wiki/pages/viewpage.action?pageId=2706650712527870)

06/16/2022

General information

...

  • 3 changes have been merged during the week, and especially a new one to update the Maven version to be used in the gate.
  • Discussion about changes under review:
    • "Adding additional kafka message": This change provides interesting elements for the purpose of Demos. Anyway it does not use "nbinotification" module. It would be preferable to use this last, but one shall note that we have planned to refactor nbiNotification (in the current sprint). After it has been refactored, the proposal of UTD which is very interesting could be integrated, so that we can rely on what was done to improve nbinotification module.
    • relation chain proposed by Thierry and related to the next Orange Exhibition Research demo:
      • "Fix issues in listeners” fix a bugs related with the handling of admin/operational state and adds some unitary tests
      • "Add service-resiliency compliance check” and “Add service-resiliency handling in service-create” allow to handle service resiliency in the service-create request, which is necessary for autonomous service rerouting.
      • "Implement an autonomous service rerouting”: implements a listener of the service data store that detects when the operation state of a service changes. That triggers a service reroute, making it into two steps: 1st, service-delete, and 2nd, service creation with a new path bypassing links and nodes affected by failures. Associated unitary tests are provided.
      • In Parallel, Thierry also developed some code in Honeynode to handle the RPC pm-interact which allows creating, reading and setting some PMs in HoneyNode. This was developed on the master branch (based on OR 7.1 device release)

    • "Updating the usage section of the developer guide”: this change was pushed by Roshan to update the documentation in which some references were obsoletes. Olivier will have a look to it before being merged.

    • relation chain related to T-API: We agreed during the sprint planning that as soon as the master branch will be migrated to Sulfur, all changes associated with T-API will be rebased on Sulfur and that the review will be launched from there

    • "Support for parameterization of SIMs attributes”: this is a new change pushed by Nathan so that the CI/CD chain could be switched from simulators to an external platform, providing some parameters for the connections to real equipments. Gilles proposed to use environment variables that can be set (see comment under Jira ticket (https://jiralf-opendaylight.opendaylightatlassian.orgnet/browse/TRNSPRTPCE-674).

06/09/2022

General information

...

  • Sulfur release: All managed project are ready to release for GA of Sulfur. On TransportPCE side, we are experiencing different issues which prevent us to migrate to Sulfur.

    • Gilles progressed but there is still a rather blocking issue with the compilation of OpenROADM models. The issue was raised to the ODL community yesterday.

    • the previous model compilation issue (https://jiralf-opendaylight.opendaylightatlassian.orgnet/browse/MDSAL-738) has been solved and integrated in MDSAL-9.0.2, released today.
    • Gilles encounters another issue with odl-parent (bad dependencies revision in maven-site-plugin) which can be solved with an overriding of pom.xml files.

  • Chlorine preparation:

    • we will have a total removal of restconf-nb-bierman. Guillaume started the migration from bierman to RFC 4080, but some work still needs to be done.

    • we will also need to migrate to java 17 (CI is currently under migration)

  • Orange had a discussion with NTT about the potential implementation of a control plane based on TransportPCE in the scope of IOWN. NTT is also in contact with UTD.

  • It is probably the good time to make a decision on the status of TransportPCE and the way it shall be handled in ODL: as a managed project or as a self-managed project. Gilles recommendation is to move to the status of managed project. Managed project are integrated in ODL as core projects. On the other hand, artifacts would probably have to be delivered earlier.

    • No objection raised, so, the decision is to move to Manage Project.

...