2022 meeting minutes
Meetings connection details
available at TransportPCE meetings
https://lf-opendaylight.atlassian.net/wiki/display/ODL/TransportPCE+meetings
12/15/2022 (last meeting for 2022)
General information
Gilles released this week Transportpce artifacts for Sulfur-SR3 and updated the S-SR3 release note
Gilles took into acount ODL feedback received on TransportPCE to ease our integration inside integreation/distribution. He added to both transportpce and transportpce-models projects a new artifacts module and features declaration module. There have been still small issues with the declaration of features in tpce, but it shall be resolved now and we shall be in time for Cl SR1. The main remaining issues could be associated with integration/distribution.
Review
Status on transportpce-models changes merged this week:
“Clean feature dependencies” (upstream dependencies), “Add artifact module”, Add features-transport-pce-models” (feature BOM) have been merged on stable/chlorine branch and backported on the master branch (not merged yet)
Status on transportpce changes merged this week:
“Handle missing spanloss in PCE without crashing” pushed by Jonas allows better handling of OMS attributes
“Fix Netconf version” solves an issue bumping to Cl-SR1
“Bump of upstream dependencies to S SR3”
“Refactor NBI notification Utils” is optimizing code using streams and allows reducing the cyclomatic complexity
“Refactor functional tests lighty support”: another way to manage the restconf url which has not exactly the same prefix with lighty.
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?)
A.O.B.
We received some feedback from Andrea Fumagalli which is in line with the feedback that provided Tianliang. Tianliang will write a JIRA ticket associated with the corrections to be applied.
12/08/2022
General information
S-SR3 has been approved by TSC and is being released. Gilles staged our stable/sulfur branch last Friday, so we are fine for this release.
Cl-SR1 should be in a short time submitted for approval by TSC. Initially, Tpce was supposed to be integrated has a managed project for Cl SR1. But some issues were identified with its integration. So its integration is being questioned. Gilles pushed some changes to solve identified issued on transportpce-models and transportpce sides. We could be ready for integration, but Gilles is not sure that it will accepted as it is. Gilles asked whether we could delay the integration of tpce as a managed project. There was no objection.
Review
Status on transportpce changes merged this week:
Relation chain about the migration to S-SR3 has been merged by Robert Varga
Relation chain concerning the migration to Cl-SR1 has also been merged.
“Bump upstream dependencies” allows aligning on the latest version of our upstream dependencies and has been pushed by Robert.
“Fix Netconf version” is also related to Cl-SR1
T-API notification Junit tests : Gilles has reworked this change pushed by Javier a long while ago to allow the change to pass the gate. Gilles mocked some classes that did not need to be instantiated.
Status on transportpce-models changes merged this week:
Different changes have been merged. It seems that Robert also did some work on this project as it was a managed project.
transportpce-models changes under review:
“Clean features dependencies” : Solves an out of memory issue rolling out the distribution. The issue was mainly due to a wrong declaration of upstream dependencies in our features. So Gilles has been cleaning the way upstream dependencies are declared
“Add artifact modules” : declaration of features and bundle eases version management use in transportPCE
transportpce changes under review:
“Fix feature issues” : does the same work for tpce integration.
“Handle missing LLDP nbr-list” has been pushed by Jonas: For Roadm-to-Roadm link discovery we could have to handle empty list which causes NPE. Gilles proposed as alternative solution to dedicate one specific method to check the different lists.
“Handle missing spanloss in PCE without crashing.” Is another change pushed by Jonas which needs to be reviewed
“Fix persistence when Karaf is started”, first change pushed last week by Jonas also needs to be reviewed.
A.O.B.
HoneyNode sends device change notifications, but the mechanism of subscription to notification is not handled. RESTCONF interface of Honeynode may be used to modify the node configuration. An example of implementation is presented in the test 13 of device_change_notifications functional test.
Tianlang submitted a paper which was accepted for OFC, describing ONE Engine and interactions with HoneyNode and TransportPCE.
Javier has also been proposing a paper “Open Disaggregated Optical Network Control with Network Management as Code” which was accepted for OFC. Nokia introduces an Automation Engine that analyzes topology changes to trigger the Continuous Delivery procedure of SDN applications on a micro-service architecture.
12/01/2022
General information
The following karaf bug https://lf-opendaylight.atlassian.net/browse/TRNSPRTPCE-701 affects S-SR3 and Cl-SR1. A workaround has been proposed migrating functional tests to lighty build.
The corresponding relation chain has been merged on master and stable/sulfur branches, so that TransportPCE is now ready to release S-SR3. Gilles will stage S-SR3.
For what concerns Cl, the dedicated relation chain to Cl-SR1 migration is also now running on the gate. Guillaume merges it on meeting, so TransportPCE is also ready to release Cl-SR1. But since we should be a managed project in Cl-SR1, remains to see how the integration may be finalized in integration/distribution and may be also autorelease.master branch is now aligned on Cl-SR1, with functional tests running wiht lighty.io runtime
Review
Status on transportpce changes merged this week:
the different relation chains related to S-SR3 and Cl-SR1 migration
a lighty refactoring of OpenRoadmInterfaces by Guillaume
transportpce changes under review:
“First persistence when Karaf is restarted” => new change pushed by Jonas to solve Bug TRNSPRTPCE-706. We need to modify returned code in associated functional test prior to merge this change. To be tested locally by everyone.
“Hack To update degree port on portMapping” has been reviewed by Guillaume who made some comments => need to be tested again on real equipment by UTD, to apply necessary modifications on it before being merged.
A.O.B.
Christophe upgraded the simulator to Cl. Apart of the notification mechanism which still does not work, which was already identified, everything is working fine.
Feedback from TianLiang on SC22 demos: Link restoration does not work if there are 2 hops on a path. The restoration is not triggered at all. This is the case not only with real equipment but also with HoneyNode. Need to be investigated...
SuperComputing : Every demos planned by UTD were successfully ran in SuperComputing. A path restoration on real devices takes about 9 minutes (tier-down, new path setting). Mainly due to the OLM timers specifed by OpenROADM.
11/24/2022
No Meeting (Thanksgiving in U.S.)
11/17/2022
General information
SC2022 has started. The restoration demo is running.
LFN Developer and Testing Forum is also currently hosted in Seatle. Robert will make some ODL presentation of interest for us during the event. But unfortunatelly, we can not attend virtually the event for free.
Stable P branch : Gilles bumped last week the branch to P-SR4 and solved the Pylint issue. (“Bump Upstream dependencies to P-SR4”)
Changes concerning S-SR3 and Cl-SR1 are not merged yet due to the karaf regression: https://lf-opendaylight.atlassian.net/browse/TRNSPRTPCE-701
Review
Status on transportpce changes merged this week:
on transportpce project, only change “Bump Upstream dependencies to P-SR4” on stable/phosphorus branch
transportpce changes under review:
"Update pre-commit and tox configuration”: Allows Commit message to be evaluated at different stages. Merged during the meeting. Could be backported now on other branches.
relation chain on Regen (“Update portMapping data for Regen xpdr-type”...): not tested with latest Honeynode version
“Hack to update degree port on portMapping” is ready to be merged.
integration/distribution:
Following an exchange with Sangwook on slack, Gilles pushed the following change “Fix transportpce.version parameter use” to fix a mistake with transportpce version. However we also need to release our Artifact for Cl, which can not be done until Jira ticket TRNSPRTPCE-701 (regarding Karaf issue) has been solved
A.O.B.
The rest of the meeting is dedicated to the Sprint-29 demo. See directly te Jira.
11/10/2022
General information
Status on the different release trains:
S-SR3 has been submitted to TSC for approval.
Cl-SR1 has also been submitted, but it will be a bit delayed because it must include TransportPCE and we are not ready. The change “Add transportpce projects in Chlorine distribution” has been submitted in the distribution/integration project so that TransportPCE can be integrated as a managed project. Gilles took into acount at least one comment. The rest of the review depends now on committers.
P-SR4 is also in preparation. Stable branch for SR4 is in preparation.
Gilles discovered an issue with the newest versions of karaf used for Cl-SR1 and S-SR3: all T-API functional tests are failing because of a regression. A Jira Ticket has been raised on transportPCE. We are not able to install the Karaf client feature. Robert Varga is working on the issue.
Review
Status on transportpce changes merged this week:
on transportpce project: “Update of user and developer guide”
on transportpce-models project: “Bump upstream dependencies to Cl SR1”, and “Upgrade project version for Cl SR1”
transportpce changes under review:
Gilles pushed a serie of changes to maintain stable/branches (P-SR4, S-SR3, Cl-SR1) which includes for each of them the upgrade of the project version and the bump of upstream dependencies to their adequate newest releases.
Relation chain of Regen feature: HoneyNode does not include the regen functionality. Christophe updated the release of HoneyNode to include regen feature so that Bala shall be able to use it for testing this chain (Using HoneyNode-pluggin-aggregator7.1.8). Bala also see that monitoring mode (on ODUCn interfaces) needs to be updated to support Regen (not-terminated, /monitoring)
“Hack to update degree port on portMapping” change has been proposed by Gilles to solve the restauration issue for the SC-2022 demo. The PortMapping now updates the port operational-state depending on both the operational-states of the port and the supported OTS interface for degrees. This looks fine to Bala, but he will review it tomorrow before we merge this change.
Relation chain associated with Kafka: we need to take time to review Nathan changes, and we understood that there is no specific urgency from UTD side to have these changes reviewed
A.O.B.
The Meeting of next week will be extended of an half to 1 hour to organize the demo of the Sprint.
11/03/2022
General information
Gilles pushed a serie of changes to update the project version to the next release development, for CL SR1, S-SR3 and P-SR4.
Review
No change merged this week.
transportpce changes under review:
Gilles proposed change 102917: Hack to update degree port on portmapping | https://git.opendaylight.org/gerrit/c/transportpce/+/102917 to solve the Bug raised by Tianliang. The first patch set still leads to some exceptions, but Gilles pushed a new patch set to remove these exceptions. TianLiang will give a try to the second patch to see if it works. UTD has to pack all equipment next Tuesday which gives limited time.
Lighty simulators : Christophe made the migration to S and his working on the migration to Cl. He experiences some issues with the PM-handling yang models when migrating to Sulfur.
10/28/2022
Dedicated meeting to TAPI v2.4 upgrade
Javier provides the recordings from the TAPI v2.4 upgrade presentation
10/27/2022
General information
Cl GA status: Cl GA distribution seems not been released yet. All upstream dependencies have been released, and it was possible for us to release both transportpce-models and transportpce projects. According to the release schedule (https://docs.opendaylight.org/en/latest/release-process/release-schedule.html), we should need to release Cl-SR1 and S-SR3, but nothing seems ready yet.
Review
Status on transportpce changes merged this week:
“Change log level in FrequenciesServiceImpl” is correcting the Level of a LOG from Error to Warn to avoid misinterpretations, especially during the Orange Research Exhibition demo
“Update Java version check in installMavenCentOS": new change from Nathan to solve a bad java revision check on the CI. Backported by Guillaume on stable/chlorine branch
transportpce changes under review:
nothing new on the 3 changes proposed by Bala on Regen functionality. Bala is currently working on a new one to add regen functional tests. All the relation chain could be merged at the same time when every changes will be ready
Guillaume and Nathan had a lot of exchanges this week on different topics : CI, Java version check, Kafka broker.
Nathan reworked his change 99746 related to send log messages outside Transportpce, in a kafka broker. Gilles raises the fact that this implementation is also outside the use of nbinotification module that had been developped for this kind of usage. Gilles would prefer we re-use as far as possible the existing base and developments provided on NBI notifications. Gilles thinks it would be preferable to use MDSAL internal notifications for communication between internal modules and towards nbiNotification module. Nathan explains that the core issue is we try to make a single Kafka producer. But there is some instability which leads to some failure of half of the tox jobs (weird failures). Nathan is about to abandon the Kafka approach. The solution he has in mind are:
Solution could be to rely on nbiNotification, but this imply to make a lot of changes in a lot of classes that use notifications.
Use the ODL karaf.Log and KarafAppender pluggin but it may imply to override ODL parent settings
Completely abandon the Kafka approach and move to an approach based on promtail...
A.O.B.
Bala experiences some issues testing equipment with last version of TransportPCE, with all what concerns capabilities announcement. The device if-supported-capabilities seems not being correctly recognized by MappingUtil class. Bala will investigage more deeply.
10/13/2022
General information
Cl GA status: Gilles attended last week TSC meeting. It was decided that we will wait for a complete integration as MSI (Managed Snapshot Integrated Project) projects (splitting transportPCE and Models) until Cl SR1. Since last Thursday, nothing happened and Cl SR0 is still not released. The release has not been approved yet.
OpenLab is officially created. This was announced today during the OpenROADM meeting. It will provide a common secure space for conducting interoperability validation of OpenROADM compliant equipment. To perform this it will work on automation of some test based on TOX and may use transportPCE as a controller being part of the test chain
Code Review
Status on transportpce changes merged this week:
“Increase project version to 7.0.0 Snapshot” will allow having the master branch now aligned with Argon.
“Update user and developer guide for Chlorine” is notably correcting the url used on REST API to be compliant with RFC 8040.
“Update link-status of service-path-list” is improving the behavior of service restoration and has been pushed for the SDLRI demo.
transportpce changes under review:
“Change log level in FrequenciesServiceImpl” is correcting the Level of a LOG from Error to Warn to avoid misinterpretations.
“Support create/delete interfaces for regen” : source and destination tp are Network tps for on Regenerators. This is used to trigger ODUCn interfaces creation on relevant ports.
A.O.B.
Lighty simulators can send notification, but Christophe still experiences some issues with the serialization. Thus, Tpce is not able to process this notification. DeviceChangeNotifications yang model includes an instance-id which is not correctly serialized.
Notifications handling for SC22 demo: we need some clarification on the streams announced in the capabilities provided by the devices. Bala mentions that the equipment may only announce the support of OpenROADM streams, even if they support Netconf notifications which is a MUST. Tpce also does not process at that time DeviceChangeNotifications on interfaces. Gilles will try to find a workaround and provide a patch that allows subscribing to the Netconf stream even if its support is not announced by the device, and that also allow handling notifications coming from state-changes on interfaces.
Next week meeting is cancelled since both Gilles and Christophe will be at the Orange SDLRI playing the demo of optical restoration.
10/06/2022
General information
Sulfur-SR2: Guillaume released tpce last week and took care of the release note this week.
Chlorine status:
yesterday the Cl release has been proposed for TSC approval. 2 members already voted no since transportpce/models and transportpce were not included in the distribution as a managed projects.
Gilles managed several tasks to prepare the integration
he pushed yesterday a change on transportpce/models to have an odl-tapi-models feature as it is already the case for odl-openroadm-models. It seems important for the ODL integration point of view because this is mainly features which need to be declared. He also pulled the stable/clorine branch and pushed a change on releng/builder project (still under review) to build models on clorine branch.
On transportpce side, Gilles merged all the relation chain on the chlorine migration step 2 (building the project using openroadm and tapi models as an upstream dependency) and pulled the stable/chlorine.
on integration/distribution project, Gilles initialized a change (and Guillaume amended it) to integrate these two projects in the distribution as managed project, but the documentation doesn't fit exactly to our projects (we don't have artifacts modules). We need support from the community on this last step. Gilles will attend the forthcoming TSC meeting to discuss this point.
Code Review
Status on transportpce changes merged this week:
Most of the merged changes of this week are associated with the second step of the migration to Cl. Let's notice two new changes: “ Reintroduce karaf restart in tapi functional tests” and “Remove openroadm and tapi models from project”. With the export of tapimodels, when we install odl-transportpce-tapi featue, restart of other transportpce bundles fails. So, currently the workaround consist in logging out and logging in again in karaf to have all bundle starting correctly. This point should be added in the clorine developer guide.
transportpce changes under review:
“PortMapping to support regen capabilities” and “Update port-mapping to support regen capabilities”: one comment from Guillaume was not addressed yet because there is still an additional change in the relation chain that addresses it but has not been pushed
“Update link status of service-path list”: it allows updating, on the service-path, the status of link resources terminated on a tp when its status changes, correcting internal notification mechanism.
09/29/2022
General information
Sulfur-SR2: Guillaume successfully released tpce in S-SR2 release. S-SR2 ODL distribution has however still not been officially released yet. The release note has not been provided yet. Guillaume will take care of it.
Chlorine: All kernel projects are more or less ready. On our side, the first step of the migration is done. The second step is well advanced (use OR and T-API models as an upstream dependency)
The bug (critical) raised last week by Bala is closed: Cl imposes to use only RFC8040 compliant urls, which implies to update Postman collections.
The DDF will be collocated with the Seattle ONE events (November).
API doc is not working on the Chlorine distribution. A Jira Ticket shall be raised first in tpce, and if we are confident that it is not on our side, then in the Netconf project. Bala will take care of it.
Code Review
Status on the new transportpce/models repo dedicated to compile openroadm and tapi models
committers rights of tpce are now also effective on this new repo. So transportpce committers are also committers on transportpce/models project.
The CI has been configured by Cedric, and merged by Guillaume on the releng/builder project
change “Add OpenROADM and TAPI models” (merged) adds models to the project.
Currently, a new relation chain to clean repo files is under review: “Add some linters via tox and pre-commit” , “Enable linting in TAPI and OR models”, “Fix Tabs and trailing blanks in OR models”, “Fix MSDOS file terminators in TAPI models”.
Guillaume proposes to rebase Gilles's change on top of this relation chain to clean every things.
Status on transportpce changes merged this week:
“Bump MRI upstreams” pushed by Robert is updating dependency for Chlorine.
“Add allure-pytest tox profile” provided by Tianlang has been merged
transportpce changes under review:
Gilles's relation chain related to the chlorine migration step 2, with compiled models imported by maven dependencies
“Remove useless yang models from-test common”, is removing an old transportpce model that is not used anymore for a while
“Update transportpce-common-types model” is adding pm-granularity to tpce models to solve some issues we have with this parameter when we use the one coming from OR models with models outside transportpce project (linkage issue)
“Fix bad imports” cleans up some imports made in tpce that were not the correct ones.
“Remove openroadm and tapi models building” removes the compilation of T-API and OR models, uploading a new bundle corresponding to the newly created transportpce/models project. There is still an issue with T-API functional tests because bundles do not restart automatically when we install odl-transportpce-tapi feature without restarting Karaf. Gilles is updating tapi functional tests with a workaround waiting to understand what happen in karaf.
Bala's releation chain to integrate the support of regen in transportpce:
“Port-Mapping to support regen capabilities” adds a regen profiles container which includes the supported regen profiles when the Xpdr-type is a regen.
“Update portMapping data for Regen xpdr-type” provides code to complement portMapping with the information required to handle Regens.
Christophe is still working on lighty simulators, providing some code that allows reaching the same functional level as what we currenlty have with Honeynode. He still has an issue to solve with notifications sent by the lighty simulator.
Olivier pushed an OpenConfig version of the SwitchPonder configuration on the LFN/HoneyNode Gitlab that will allow to test future openconfig feature developements inside transportpce
A.O.B.
A question was raised by Bala: shall we make 1.2.1 support deprecated? It seems that there are no more proposed/deployed products based on openroadm models 1.2.1. We will need to think about it, since this leads to additional complexity when handling objects in the service and network models that do have different format in R 1.2.1 and the next releases.
09/22/2022
General information
Sulfur-SR2 is underway. Guillaume staged the release.
ODL Cl release is a bit late. Transportpce master branch is already on Cl, but we still have to migrate the models to the new repo on which we don’t have any rights. Also there is no CI on this repo. The CI needs to be prepared. The export of the models may be delayed to the next distribution of ODL if we are not ready on time.
About the issue observed in the labs with code aligned on Cl (no more mounting of any equipment), Guillaume suggested that Bala raises a JIRA Ticket in TransportPCE appending all the needed traces and log (all Karaf Logs), so that investigations can be initiated relying on this ticket. App dev and Kernel project lists could be used to raise the issue based on the JIRA Ticket.
Code Review
Changes merged this week:
Re-open Drop service name from ODU Connection name
Gilles tested the relation chain provided by Thierry to improve the SDLR demonstration (path reroute), which works well, he also reviewed the code, and the changes have been merged.
Changes provided by Guillaume associated with the migration to Chlorine.
Changes under review:
“Add allure Pytest tox profiles for local use” has been reviewed and reworked by Guillaume who expects Tianlang to try the modifications on his local environment to provide feedback.
A.O.B.
Bala presented what he has done for handling regenerator in the Renderer going through a presentation.
09/15/2022
Code Review
Changes merged this week:
Except what have been merged on Sulfur branch (Rest of the relation chain concerning migration of funcional tests to RFC 8040) most of the changes are related to the migration to Chlorine on which master branch is aligned:
“Adapt transportPCE code to Chlorine” is the change with the most important code adaptations to chlorine.
Few additional small changes proposed to improve the things:
"Adapt functional test to Clorine" concerns mainly OLM functional tests (migrated before chlorine)
“Improve feature install in functional tests”: considering latest evolutions, there is no more need to restart the controller after features have been installed.
Guillaume proposed some optimization through “Refactor code after Chlorine Bump” which brings some optimization based on else if, and “Refactor Renderer OpenRoadmInterfaces step 1”. For this last change, addressing latest comments, as Ethernet CSMACD in some yang OR identities, we will need to push a correction.
Changes under review:
few changes to improve code after the migration to chlorine
“Clean poms after Chlorine bump”: provides some updates in the dependency (some RFC reference have evolved)
“Fix raw types build warning”: notification need types. If they are not specified, some warning are generated, otherwise “generic needs to be specified”
“Fix unchecked build warnings” also allows to avoid generating unnecessary warnings
==> We can now consider that step 1 of migration towards Chlorine has been accomplished
Relation Chain from Thierry associated with service restoration. Olivier reviewed them and is OK with the different changes. Gilles will test the code before merging these changes.
“Drop service name from ODU connection name” is in merge conflict. This is probably due to the migration towards Chlorine. Bala will rebase it to check.
A.O.B.
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://lf-opendaylight.atlassian.net/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.
09/08/2022
General information
Sulfur-SR2: Guillaume tryed yesterday to stage the release, but we need to check before if everything has been backported. Managed project are still not completely ready.
Code Review
Changes merged this week:
Few changes related to the gate
Migration of functional tests to RFC8040: The migration to RFC 8040 is fully accomplished with the latest “Switch functional tests to RFC8040” change that has been merged.
Alignment of master branch to Sulfur branch (S-SR2)
3 changes on the catalog (Olivier and Guillaume)
Changes under review:
Relation chain regarding the rerouting of the services still needs to be reviewed.
“Adding allure-pytest plugin”: the plugin allows to save Pytest reports, with some elements on how the tests have been performing, as well as the time needed to perform them. This can be used to improve debugging which is now very verbose. Guillaume highlights that this change will need to be reworked if we need to get the report in the CI since they cannot be accessed as they are today. Thus it brings benefit when used locally.
Chlorine migration: this is more or less finished. “Bump some dependencies to Chlorine”, “ Adapt transportPCE to Chlorine” & “Adapt functional tests to Chlorine” allow to get a +1 on Chlorine branch which demonstrates that we are in good shape for Chlorine and that everything needed was done for the migration. Guillaume added a final change “Improve feature install in functional tests” on top of those.
09/01/2022
General information
Sulfur-SR2:
According to the initial schedule plan, should have been released on the 26th of august. We are more or less ready on transportPCE side. Dependencies have been bumped. Since Pylint dependencies and runners of the gate have been updated, the code also needed to be updated (see the 2 changes “Release Python version fixed for PCE functional tests” and “ Fix new Pylint issues in functional tests”).
Chlorine should be released on the 19th of September. On our side the migration is well advanced.
Gilles splited the Video of the Avortex demo in 3 parts and pushed it to Confluence.
Code Review
Changes merged this week:
“Fix new Pylint issues in functional tests”, “Release Python version fixed for PCE functional tests” are needed for the migration to Sulfur-SR2.
Part of the new relation chain pushed by Gilles that completes and finishes the migration of functional tests to RFC8040: "Migrate end2end functional tests to RFC8040", "Refactor test_utils_rfc8040.py portmapping methods","Migrate PCE functional tests to RFC8040", "Fix few functests assert RFC8040 uncompatibilities", "Migrate hybrid functional tests to RFC8040"
Changes under review:
”Shift tox nbinotifications to with_docker profile”, “Migrate with_docker func test to RFC8040”, “Migrate TAPI functional tests to RFC 8040”: The rest of the relation chain on the migration of all remaining functional tests to RFC8040. Gilles finished a priori to fix issues on TAPI functional tests
“Adapt some dependencies to chlorine” and “Adapt transportPCE code to Chlorine”: the rest of the initial relation chain, now dedicated to Chlorine migration. First review already done by Guillaume.
"Fix checkbashisms auto install in tox for CentOS8" is proposed by Guillaume to solve issues on the gate. Still under working
The relation chain associated with the primitives to retrieve parameters from the catalog (1Initial change from Olivier + 2 refactoring changes provided by Guillaume) is ready to merge since all comments have been addressed.
”Reopen Drop service name from ODU connection name” : a patch had been provided in May to adjust (and reduce) the connection name. But the service-name was removed which sometime can leads to issues. This patch solves the issue : we don’t drop the service name from the Network and client interfaces, but only from the ODU connection name.
"Update transportPce-pce yang": part of a new relation chain from Thierry on service reroute. Thierry proposed a new RPC “path-computation-reroute-request” needed to handle intermediate phases in reroute process. ”Implement path-computation-reroute-request” provides the implementation of the RPC.
08/25/2022
General information
Sulfur-SR2: Netconf upstream dependency still missing to be able to bump to S-SR2. Otherwise, all changes already merged on master are backported on stable/sulfur
Chlorine migration:
The new dedicated transportpce/models repo is at least created
Gilles started the project migration to chlorine without exporting models first. Once it will be ok, models will be moved outside the project.
Code Review
No change merged this last week
Changes under review:
1 change from Bala: "Re-open:Drop service-name from ODU connection name" to have the service-name in the interface name, necessary when a high order termination point supports several low order interfaces
Relation chain from Thierry on service reroute: already discussed last week. Waiting feedback from Olivier.
new relation chain of 9 changes proposed by Gilles on the Chlorine migration + migration of all remaining functional tests to RFC8040
"Bump upstream dependencies to chlorine": just the bump
"Adapt some dependencies to Chlorine": to better distinguish dependency impacts on Chlorine
"Adapt TransportPCE code to Chlorine": to see code modifications brought by Chlorine
"Migrate end2end functional tests to RFC8040": pursue the RFC8040 migration started by Guillaume for all end2end tests
"Refactor functional tests": rename some functions of tests-utils to be more explicit
"Migrate pce functional tests to RFC8040": choice to only use json file as portmapping and topology input files to perform tests, to ease the migration. To that end, Gilles has converted xml input files to json input files
"Migrate hybrid functional tests to RFC8040", "Migrate TAPI functional tests to RFC8040", "Migrate nbinotifications funct tests to RFC8040": continue the rest of functional tests, but Gilles didn't have time to tests these last tests, so it remains probably some bugs...
=> so, the chlorine migratio is well advanced and at first sight, no strong issue detected...
A.O.B.
Gilles tried to load on confluence the AVORTEx demo video, but file is too big. Need to cut it...
08/18/2022
General information
Chlorine kenel projects in a good share (CSIT migration to jdk167 finished)
S-SR2 planning available at Sulfur SR2 Release Checklist - branch unlock 08/22 and deadline for SM 08/26
transportpce/models repo created https://git.opendaylight.org/gerrit/admin/repos/q/filter:transportpce https://git.opendaylight.org/gerrit/admin/repos/transportpce/models
Code Review
No change merged this last week:
Changes under review:
Waiting for feedback for the changes presented last week
WIP ODU Drop service name bug correction pushed by Bala this week
A.O.B.
question about tox to perform a few tests selectively: answer can be found in docs at https://docs.opendaylight.org/projects/transportpce/en/latest/tox-guide.html
Last meeting video cannot be seen outside Orange (teams configuration problem ?)
=> waiting for @Gilles THOUENON return : apparently he is the only one who can retrieve the original file on its personal folder and share it outside Orange.