2023 meeting minutes

available at TransportPCE meetings
https://lf-opendaylight.atlassian.net/wiki/display/ODL/TransportPCE+meetings

• New ODL web site is here

• ODL K-SR1 has been been released. TransportPCE release note had not been backported on stable/potassium branch (branch not created yet when change pushed). Guillaume do it on live. Idem for the download page update.

• Guillaume want to raise the GHA issue to the TSC. He will try to connect to today's meeting to explain its position.

No change merged  these two last weeks.

transportpce changes under review:

• “Spectrum assignment skipped the last node in path”: new change proposed by Joakim to fix the PCE bug TRNSPRTPCE-772. Guillaume made a review and even a small refactoring in change PCE graph spectrum assignment slight refactoring (Ib4210d7f) · Gerrit Code Review (opendaylight.org). Gilles would like to have at least a JUnit test to check the bug fix.

• "Refactoring of T-API according to 2.4 Release": new change still in WIP to optimize the current code under development.

• New ODL web site

• TransportPCE K-SR1 had been the 17 of November. ODL distribution seems not being released yet in K-SR1…

Status on changes merged  since last meeting:

• “Bump upstream dependencies to K-SR1”

transportpce changes under review:

• “Migration to TAPI 2.4 Step 1 to 3”: Guillaume relaunched is review. Olivier will look at the comment to address them.

• “Don’t use NotificationListener (Node registration)”: All functional test using Netconf notifications are failing. Gilles checked that everything is correct on our side. The problem comes from NETCONF project: we need to wait for the implementation of all related methods in NETCONF project.

• New bug on PCE reported by Smartoptics (jira.opendaylight.org/browse/TRNSPRTPCE-772)

• Github actions: We have now a verification through Github action that has been pushed by the LFN. This will complicate a bit the process.

Status on changes merged  since last meeting:

• “Bump upstream dependencies to K-SR1” is now passing the gate since some transportPCE models have been updated.

• “Bump jgrapht dependencies to 1.5.2"

transportpce changes under review:

• “Migration to Tapi 2.4 Step 1, 2, and 3”:  Olivier has rebased his changes. Any review is the welcome.

• “Don’t use notification listener (nbiNotification) ”:  still needs to be reviewed and reworked.

• “Update transportpce release note”:  Gilles pushed a release note that needs a little correction prior to be merged

• Ar-SR3 was released last Monday

• K-SR1 should has been released but is a bit delayed.  Gilles has already backported given number of changes merged on master, but most of them are failing on the gate (big unstability of the gate these last weeks??)

• Gerrit version has changed with given number of drawbacks:

  • a new github action job that just check the yaml file of the projet must deliver a V+1 => we have lost the high and clear visbility we had before on the gerrit wall!

  • Gilles has lost its direct CR+2 shortcut button!! => it's very painful

Status on changes merged  since last meeting:

• Gilles's change on bug-768

  • “Fix Pce bug to select the correct client port”. A new bug shall be open to describe the issue of trib-port and trib-slot selection recently raised by Bala.

  • “Complement B100G”. Associated tests have been added

• “Update tox documentation link in tox-guide”

• “Fix PortMapping bug for 7.1 device-model”

transportpce changes under review:

• The 3 following changes concerning the PCE are ready to be merged.

  • “Add pce-constraint-mode enum in Pce Yang model” introduces 2 modes to handle the constraint in Path computation (Strict/loose). In strict mode, needed in a second iteration (when GNPY fails in validating the path and propose an alternative one using hard constraints of include type to express it), the constraints need to be handled following a specific order which corresponds to the exact sequence of element.

  • “Fix PCE bug to select the path proposed by GNPy” solves a recently identified bug in the PCE. It has been reviewed by Guillaume.

  • “Bump Jgrapht upstream dependency to 1.5.2” : allows relying on more complex graph handling multiple edges.

• “Migration to Tapi 2.4 Step 1, 2, and 3” :  Olivier has rebased his changes and refreshed the status of comments.

• Gilles asked why in 7.1 device configuration has several time exactly the same profile (using same trib-port and trib-slots) with different names  (ABCD-3-BY-100G-1/2/3). The explanation may be that there is an issue in oducn-rate (wrong copy paste). Bala will verify and reply to Gilles.

• Potassium (GA) was officially released last Monday.

• Ar-SR3 was in the process of being approved, but the TSC decided to restart a release process because an issue leads to a too big size of the Jar file.

Status on changes merged  since last meeting:

• the two changes form Bala, after being x times rebased: “Add json assert dependency” and “Nbi notification support for service-result-rpc"

• 2 changes from Matej on notification listener

• 2 changes related to the project maintenance: "Fix the requirements.txt file for tox" and "Use mvn quick profle in other tests Karaf build”

transportpce changes under review:

• “Fix Pce bug to select the correct client port” was pushed by Gilles who would like Bala to test it on his platform.

• “Complement B100G-end2end functional test” provides an additional test that allows verifying that the fix corresponding to previous change really solves the bug . This test applies to OTN. This change and the previous are ready to be merged.

• “Don’t use notification Listener (NodeRegistration) ”: It seems there is a real issues with it (not a false negative V-1 as often): some of the tests did not path (hybrid tests). Gilles will review it next week.

• “Bump project version to 7.3.1 SNAPSHOT” allows starting the next development iteration before releasing Ar-SR3.

• “Fix Pce bug to select the path proposed by GNPY” : TRNSPRTPCE 769 describes a new bug that observed Gilles while working on other changes. This change solves associated bug. To make this, he needed to introduce two way of handling a constraint (next change)

• “Add pce constraint mode enum in pce yang model” : Gilles introduce an enum that specifies 2 ways of handling a constraint (such as include node), strict and loose. In the strict mode, the constraint shall be expressed in a specific order. In loose mode, the constraint might be partial and not respect a specific order.

• “Bump jgrapht upstream dependency to 1.5.2” : we have been using for years 1.2.0 version of jgrapht-core which is very old. Using 1.5.2 allows to use more efficient algorithm to compute the Dijkstra on the graph (YenKShortestPath library). This is required to solve several identified bugs.

• ODL Potassium GA should have been released.

• Ar-SR3: TransportPCE artifact were released beginning of this week. Ar-SR3 ODL distribution is under approval by the TSC

• K-SR1should come up very soon now.

Status on changes merged  since last meeting:

• End of Gilles's relation chain on the migration of the RpcService (JIRA-752)

• Changes from Matej on the equivalent migration for the NotificationListener for the Service Handler module ("Don’t use Notification Listener (Service Handler) “)

• One fix on inventory module, linked to the new karaf version already used in Ar-SR3 release: “Fix inventory module configuration file”. We have a new jdbc dependency update in last Karaf version which introduced an API rupture to connect to Maria DB. This has been fixed in this change.

• Another fix for the bug-766 reported last week by NTT: “Create OC path at 200Gbps and 300 Gbps”. Intermediate rates were available for Ethernet services but not for OC service types.

• Few changes related to the project maintenance:

  • “Update aggregator POMs”: skips deploy and install maven steps for feature aggregators

  • “Delay execution of the first PCE functional test“ is an attempt to solve an issue we have with PCE tests regularly failing on the gate. Execution of this first step is 10 seconds delays. We will see from usage if it improves the gate behavior

  • “Use Maven quick profile in tox-verify job”: allows accelerating the compilation process (around 30% less time needed) activating the -Pq Profile. This profile disables tests, code analysis, Javadoc generation and so on.

  • “Fix test/requirements.txt” solves an issue with the installation of dict2xml for tox.

• On stable/argon and stable/potassium branches, main important changes from master have been backported.

transportpce changes under review:

• “Don’t use Notification Listener “: changes from Matej for TapiProvider and NbiNotification are almost ready to merge.

• “NBI notification support for service-result-rpc”:  Is almost ready to be merged.

• “Add Json assert dependency”: is ready to be merged.

• “Migration to TAPI 2.4 Step 1, 2 and 3” was rebased and is ready for review.

• “Fix PCE bug to select the correct client-port” addresses JIRA TRNSPRTPCE-768. Gilles introduced a clientZ and a ClientA to differentiate ports on both the A and Z end. This change seems to solve the bug for OTN. Gilles would like AT&T to test it on real equipment. The bug may still exist for transponder, and would imply deeper refactoring. Given the current XPDRA node configuration, the port-circuit-pack-name will need to be stated in A and Z end of the service-create body. The selection of correct ports implies that the field port-device-name, circuit-pack-name and port-name are correctly filled, otherwise there could have some confusion on the port to use depending on the naming convention used in the devices. During the testing in the labs Bala also discovered that for an OC channel or an optical-tunnel from PP to PP if a specific SRG port is stated, the PCE selects the first available port rather that the one stated in the service-create RPC input. Thus, he would like that we consider also solving this issue (OC/optical-tunnel creation) when we go deeper in the refactoring.

• Bala has an issue handling the requesting-interface-properties container of temp-service-create, since with current code we handle in the same way service-create and temp-service-create.

• Migration of lighty node : Christophe tries to reproduce the problem he experienced with when statement and Maintenance-signal in OpenROADM models. He is working on serializer/deserializer. Bala confirms that the equipment he is working on, have the maint-test-signal container present in the configuration even when enabled is set to false coming.

• ODL Potassium GA should have been released.

• Ar-SR3 shall be released at the end of the month (At the same time as K-SR1). Gilles started working on preparing Ar-SR3.

• No ODL presentation forecasted for the next DDF at that time.

Status on changes merged  since last meeting:

• “Bump upstream dependencies to Ar-SR3” on transportpce-models

• “Do not use RpcService in XXX module” end of FGilles's relation change (same modification made in the different module hosting RPC registrations). It allows changing the way to register RP. We were registering all RPC at one time. We can now register each RPC separately.

• “Update lighty.io clone url”, “Fix lighty-core build”  “Update lighty.io build” allow improving the Lighty implementation on the gate.

transportpce changes under review:

• “Bump upstream dependencies to Ar-SR3” on transportpce side. Gilles started to work on bumping dependencies to AR-SR3, but he observed a regression (https://lf-opendaylight.atlassian.net/browse/TRNSPRTPCE-765): the inventory is not working anymore (this probably comes from the newest Karaf version 4.4.4). We need DataSource that is no longer provided by Karaf. Under analysis.

• “Bump project version”: ready for a while, but could not be merged due to the branch freeze.

• “Don’t use Notification Listener (YYY module)” changes are proposed by Matej  to complement the work done by Gilles on the implementation of RPCs, and make corresponding changes on the notification registrations.

• “Nbi notification support for service-result-rpc” : Bala still needs to address the Guillaume's comments and will rebase it after.

• “Fixup checkstyle violations”: some internal classes changed to final.

• “Migrate device notifications test to lightynode” and “Use Karaf instead of lighty” can be abandoned since they have been integrated in other changes. HoneyNode is still used for device 7.1 tests. (hybrid functional test section)

• We received feedback from NTT that is currently using TransportPCE. They are struggling creating OC-service terminated on ROADMs PPs at 200Gbps. Bala says that this should be supported by TransportPCE for this rate, but we need to check whether the MC will be correctly configured (by default 75 GHz).

• Bala mentioned he is still experiencing the issue where the port selected by tpce for a service is not the one specified in service-create.

  
  
  

• ODL Potassium distribution release is about to be approved by the TSC.

  • stable/potassium branches should be unlocked soon

• Status on changes merged  since last meeting:

  • “Fix new Pylint 3.00 issues”: the lastest version of Pylint  detected an issue in our functional test library. Gilles had has to refactor one method implementation. He removed the “finally” block.

  • "Update lighty.io clone URL" was proposed by @Ivan Hraško  to update the lighty repo url. Guillaume merged it quickly, but since the repo strucutre has changed a litle bit, Guillaume pushed a new change to readapt the lighty build inside TransportPCE (https://git.opendaylight.org/gerrit/c/transportpce/+/108252)

  • Start of Gilles's new relation chain on the migration of the RpcService to move forward on https://lf-opendaylight.atlassian.net/browse/TRNSPRTPCE-752 tichet.
    
    

    • contains a refactoring in networkmodel module to convert  NetworkUtilsImpl into an OSGi Component

    • Also a complete refactoring of the OlmPowerServiceRpcImplTest class

    • migration of RpcService for networkmodel and olm modules

• transportpce changes under review:

  • rest of Gilles's relation chain on the RpcService migration

  • Matej (Pantheon Technologies) is currently doing the same kind of migration for NotificationListeners.

  • “NBI notifications support for service-result-rpc”: after the rebsase, this change has passed the gate. Since then, Guillaume added 2 new comments...

  • “Bump project version” changes are awaiting for the branch unlocking when K distribution will be released

• ODL Potassium distribution release is still under progress. TPCE is done, and the release notes are also done

• next LFN Developer & Testing Forum is Nov., 13-16, 2023, in Budapest, Hungary

  • https://wiki.lfnetworking.org/pages/viewpage.action?pageId=101351543

  • https://events.linuxfoundation.org/lfn-developer-testing-forum/

•  

  • to submit a session proposals before October 23rd!

  • It is free to attend, however registration is required

• Status on changes merged  since last meeting:

  On transportpce/models:

  • 4 changes related to the K GA release process, and the stable/potassium branch cut

  On transportpce:

  • Idem, changes related to the K GA release process, new stable/potassium branch cut…

  • "Reactivate disabled functional tests"

• transportpce changes under review:

  • 3 changes ready to be merged

    • 2 changes for the project version change on stable/potassium branches, but branches are locked…

    • "Add DS op netconf connector support to test_utils": related to the migration of functional tests to lightynode

  • Bala's change on IPoWDM use case

    • "NBI notification support for service-result-rpc": the issue compiling the odl-transportpce-tapi feature comes from the fact that the change has not been rebased since the bump to the latest K dependency

  • Gilles has just started to work on the JIRA TRNSPRTPCE-752

• Potassium release is underway. Today should have been the day for releasing K. A critical regression on Restconf was reported last week (see https://lf-opendaylight.atlassian.net/browse/AAA-265). This was solved by Robert. We should be ready by early next week for TransportPCE.

• Status on changes merged  since last meeting:

  • on transportpce/models, change on bump  to the latest K upstream dependencies

  • on transportpce side, Bala's change on "Fix bug for temp-service-delete" and few Guillaume's changes on refactoring.

• transportpce changes under review:

  • “Skip tests in failure before deps version bump” which disables some test shall not be fully reverted after the correct migration to K is confirmed.

  • “NBI notification support for service-result-rpc” has some Unit tests failing, because of the order of information in Json file. Olivier suggests to follow the last patch of Migration to Tapi2.4 Step3 which corrects this issue.

  • “Add DS op Netconf connector support to test_utils” is currently WIP but could be now reviewed by Christophe.

  • “Migrate device notifications tests to lightynode”. Sometime tests work sometime not. Gilles has investigated it and think that the modification done in the configuration DataStore induces a notification from the node which is treated to rapidly to update the PortMapping. The device does not have the time to propagate the change to the operational data store, before the controller tries to read it, which leads to a failure.

• One potential bug reported by Tianliang: when mounting a transponder, portmapping tries to get otn information on a client 100GE port. To be investigated by Bala...

• One PCE bug reported during the summer by Bala needs to be investigated...

• Venkatrangan is the new TSC chair for ODL

• Final step to release of Ar SR2 ODL distribution - has been approved by TSC

• Potassium release is due in 2 weeks. Code freeze is planned next Monday with the branch cut

• Status on changes merged  since last meeting:

  • on transportpce/models, Gilles merged last friday Robert's change (https://git.opendaylight.org/gerrit/c/transportpce/models/+/107642) bumping upstream dependencies on the latest K version. This is probably the reason why everything  has failed on the master branch for several days. The equivalent change on TransportPCE side pushed by Gilles is still failing (https://git.opendaylight.org/gerrit/c/transportpce/+/107672).

  • on transportpce side,Gilles merged the 2 changes concerning the upgrade of openroadm service and network models to release 13.1.

  • Given the fact that the gate was broken (with Robert change on transportpce/models, the single-feature-tests of odl-transportpce-tapi feature was failing, etc...), Guillaume pushed the following change “Skip Tapi features and functional test”. It was merged but should be reverted. Gilles suspects that the TAPI tests were not passing the gate because we could have a delay in getting upstream dependencies available in tpce from the snapshot repository??

  • “Refactor tox.ini” allows to have compatibility with tox 3 and tox 4. Guillaume uses a syntax in variables declaration that avoid using space or comma.

  • “Clean some functional test comments & logs” & “Refactor Networkmodel test utils” (basic refactoring removing intermediate variables, etc) have been merged.

• transportpce changes under review:

  • some new changes on code style refactoring proposed by Guillaume are ready to be merged (“Fix come compilation warning” & “Refactor SH ServiceDataStoreOperationsImpl”)

  • functional test migration towards lighty-netconf-simulators: during the summer, Christophe pushed a big change to leverage the development of new simulators based on lighty.io. Guillaume split it into the following changes:

    • "Add lightynode support to testutils.py”:  Guillaume added some variable so that using a shell we can choice the simulator to use

    • “Switch tests sims to lighty.io by default”: in this change default testing environment is set to lighty.

    • “Migrate device notifications test to lightynode” is the last change pushed by Christophe to migrate tests to lighty.io. But this change does not pass the gate.

  • K migration: Gilles pushed 2 changes that are not passing the gate yet.

    • “Bump upstream dependencies to latest version”: need to be deeply analyzed... 

    • “Declare as final private classes” allows conforming to last checkstyles version. V-1 seems being related to the gate instability.

  •  Bala reported issues with his changes related to the IPoWDM use case. They need to be rebased again.

• master branch of transportpce/models and transportpce have migrated to K

• Status on changes merged  since last meeting:

  • on transportpce/models, Gilles merged the two changes migrating openroadm service and network models to release 13.1

  • on transportpce side,

    • "Bump lighty build to 19.0.0-SNAPSHOT for K"

    • "Bump upstream dependencies to Potassium GA"

    • "Declare SuppressFBWarnings in imports"

    • "Skip the temp-service-create if already exists"

• transportpce changes under review:

  • several changes are ready to be merged:

    • "Improve Xtesting installation procedure": ansible improvment where all the necessary catalog to install are specified in the requirement file

    • " Use version 13.1.0 of openroadm-service models"

    • "Use version 13.1.0 of openroadm-network models"

  •  "Fix bug for temp-service-delete": still few comments still open

  • migration of functional test to lighty-simulator: 2 initial changes pushed by Christophe ("Use of karaf instead of Lighty" and "Use lightynode sims in tox"). Gilles reworked them and merge them in the first one.

  • change "Force some ordm models snapshot versions" => shall be abandonned

• Status on changes merged  since last meeting:

  • Gilles reverted on transportpce/models the addition of Service model 13.1 that were blocking the K migration. He has been working on consolidating it. He is now considering also upgrading the Network model, since many objects defined in the common model are used in both the service and the network models. Thus, migrating these two last to the same release will be more convenient. A solution to have R13.1 model of services will be provided soon.

  • Last week, Bala had to quickly merge an extension of the OpenROADM org-openroadm-port-types@2020-03-27.yang to include some interfaces that do not exist in R7.1. He needs this to support the device in his labs. One shall note this is also needed to have the functional test for regen pass.

• transportpce changes under review:

  • Migration towards Potassium:  Gilles reworked Robert's initial change that now passes the gate
    
    

  • “ Fix bug for temp-service-delete” : Bala is still working on it, and will push a new patchset probably by the end of today.

  • “ NBI notification support for service-result-rpc”. NbinotificationListenerImpl needed to be modified (onPublishNotificationProcessService needed to be enriched with AtoZ and ZtoA container). Anyway, if these 2 containers appear in the log, they do not appear in the notification. Javier proposed some additional help to Bala on this topic.

• work in progress:

  • Olivier presented rapidly the migration towards T-API 2.4 (Step1 and Step2). He is working on the JUnit test to provide (Step3) an adaptation so that the relation chain passes the gate.

• Gilles released Argon SR2 last weekend and pushed few changes under integration/distribution project to solve CI/CD issues. Remains release note to do.

• Status on changes merged  since last meeting:

  • Bump to the latests Potassium netconf released version on transportpce-models

  • Bump to Ar-SR2 upstream dependencies

  • "Bug in freq-map topology update"

  • "Add regenertor if-cap in 7.1 common models": Bala pushed and merged quickly this very small deviation from the official OpenROADM model to be able to mount real 7.1 device implementing regen functionality. Real device implements this small modification.

• transportpce changes under review:

  • Migration towards Potassium: netconf dependency has been updated to 6.0.1, but some functional tests are now failing (TAPI, PCE...). Under analysis by Gilles

  • "Bump lighty build to 19.0.0-SNAPSHOT for K":  allows to activate K lighty build

  • "Skip the temp-service-create if already exists": Gilles reviewed it and added a specific JUnit test to check the code modification. Also he simplified the log message to be more concise.

  • "Fix bug for temp-service-delete": Guillaume made few comments. Gilles is also reviewing it.

  • "NBI notification support for service-result-rpc": Bala modified the yang model to add a-to-z and z-to-a informations computed by the PCE. Unfortunatelly, this information does not appear in the notification received on the kafka serveur. Gilles will test it on his side...
    
    

  • 2 new changes from Christophe related to lightysims. They need to be reworked a litle bit to be gather.

• work in progress:

  • ATT still working on IPoWDM Use Case demo

  • Bala reported a PCE bug for 4x100G muxponder, especially when client ports on A and Z terminations are differents.

• TSC election: news to come

• We are close to release Ar SR2, but not sure the upstream dependencies are ready.
  
  

Merged changes (Tpce/tpceModels):

• “Bug in Freq map topology updates” and has been merged. The frequency updates were not happening when activating a service after a temp-service create was exercised in a first step. This was because the Boolean variable used to identify the fact that a service is temporary was not refreshed.

• “Update temp-service-list” : when a temporary service is migrated to a standard service, exercising a service-create, the service needs to be moved from the temp service list to the standard service list. It was not previously removed from the service list when the service creation was successful. This change replaces the old one which was abandoned since Bala found a way to handle it more efficiently than the way it was proposed in a first step.

• “Add if-cap support for if-OTU4-ODU4 ports (7.1)”, presented during last meeting was merged during the meeting.

• “Improve TPCE Xtesting tests cases”, “Add a workaround note to XTesting INSTALL.txt” and “Add a XTesting entry to combine all test cases” have been merged - they complete the (dummy) tests suite initially proposed and propose a working step-by-step installation notice to perform locally.
  
  

Changes under review:

• “Add if cap support for if-OTU4-ODU4 ports (7.1)”, presented during last meeting is ready to be merged.

Changes under progress:

• “NBI notification support for service-result-rpc” : the service-result-rpc notification is used to notify a higher layer controller and provide details about the optical tunnel activation process. It was used between transportPCE modules, but not sent externally. Bala changed the code so that this notification can be used in the standard way to provide attributes such as operational modes, spectral range and power range to be used on the pluggable to a higher layer controller so that it can relay it to the end-terminal controller.
  
  

• Christophe has pushed a new commit for Lighty including the different version of OpenROADM (1.2.1, 2.2.1, and 7.1) on GitHub. He is currently upgrading the CI/CD chain from HoneyNode to Lighty.

• Gilles has applied to be part of the next TSC. Mail have been sent for voting, but the procedure to do so is not that clear!

• Ar SR2 is planned for the 27th but waiting for the release of upstream netconf-5.0.7 version. Gilles will be on vacation during the release of Ar SR2. Guillaume proposed to handle it.

• Status on changes merged  since last meeting:

  • "Add INSTALL.txt and ansible playbook for Xtesting” and “Xtesting Dockerfile init” to automatically set up a complete testing environment. For the moment, only pre-commit is launched.

• transportpce changes under review:

  • “Add if-cap-support for if-OTU4-ODU4 ports (7.1)” has been pushed by Bala to support additional equipment. This change should be merged quickly (nothing identified that prevents it).

• changes under work in progress:

  • “Bump upstream dependencies to Potassium GA” was pushed by Robert few weeks ago to migrate TPCE to Potassium. The first unitary test was failing. Gilles has reworked this change. The build is successful, but all the functional tests are failing. Gilles noticed that there is a big change in the way the controller connects to the NETCONF device, which currently prevents tpce to connect to devices through NETCONF connector. Gilles sent a mail on the development mailing list about this. He got an answer saying that this had been identified and has already been solved in netconf project. We just need to wait for next NETCONF release (netconf-6.0.1) to benefit from the correction.

  • “NBI notification support for service-result-rpc”: nothing new on this change pushed by Bala.

• Bala is still investigating the issue associated with the update of frequency map in the topology which seems to depend on the environment.

• Gilles plans to work on the introduction of latest OR service model before he leaves for vacation.

• Tianliang has seen that some of the equipment of UTD labs are very slow when trying to retrieve the configuration from the devices (operational data-store of a ROADM using RESTCONF api), which leads to a timeout exception. He asks if increasing the timer is a good option. The same configuration can be retrieved using NETCONF. Gilles answered that the timer can be configured for R/W transactions in ODL.

• the deadline for the 2023 election for attending a seat at the OpenDaylight Technical Steering Committee (TSC) is today. Gilles has applied for attending the TSC..

• Argon SR2 is planned for July 27

• nothing new since last week

• Bala reports an issue with topology update: when we create a service, frequency map would no longer be updated, what would make a second service creation failing. Bala will confirm/deny this issue

• Gilles didn't have time to update new service models 13.1 on transportpce/models project. He will do it next week

• Tianliang raises the question to create with TPCE a service over an hybrid optical infrastructure composed of 2.2.1 and 7.1 ROADMs. Should work, but xml config files in 7.1 to load in honeynode simulators are may be not correct. It is possible that some specific 7.1 model configurations are missing; the existing configurations had been simply fast-forward migrated from the 2.2.1 configurations.

• Tianliang raises the question of the TPCE compatibility with the service creation on a 7.1 switchponder at 400Gbps. At first sight, some new developments should be required in the renderer module. Updating the OpenROADM white papers with the latest model releases would be very helpful..

• 2023 election process has begun for the ODL TSC. Self-nominations shall be submitted before the 14^th of July. Gilles will probably submit his candidature so that we can get at least one representative of TransportPCE as Guillaume is leaving the TSC.

• Status on changes merged  since last meeting:

  • 3 changes from Robert have been submitted by Robert: one concerns the migration of transportpce models towards Potassium. The second  “Migrate nullness adaptations” from Javax to Jdt, which is part of ODL parent, provides some optimizations. The third “Optimize Binding-> normalizedNode Transform” also provides optimization and avoids performing serialization twice