Release v1.13.0
2024 Apr 09
Table of contents
NSM Release v1.13.0
NSM v1.13.0 has been tested on:
Changes since last release
Please add mechanism permutations for interdomain
Status: RESOLVED.
System stability fixes and improvements
[Bug] vl3-basic example is incorrectly converted into the dashboard graphical model
Status: RESOLVED.
patchesStrategicMerge
field in kustomization.yaml files is deprecated
Status: RESOLVED.
Remove duplicated IPAM code for NSE applications
Since we’d merged https://github.com/networkservicemesh/sdk/pull/1407 we can remove copy-pasted code for NSE applications.
Update local spire binary in cmd Dockerfiles to support ARM docker testing
We are currently using spire-v1.2.2
inside our docker containers (mostly for docker tests):
https://github.com/networkservicemesh/cmd-nsmgr/blob/main/Dockerfile
Consider testing on ARM64 machines
Status: RESOLVED.
Interface is not restored after restarting the forwarder
Status: RESOLVED.
Investigate for leaks in the forwarder and nsmgr
Collect metrics from cluster and see how they change. Get pprof profiles from forwarder
Artifacts are not created with new actions on Kind
Status: RESOLVED.
Multiple services request within ext_client scenario
Status: RESOLVED.
VPP requests hang because of VPP Liveness Check
VPP API requests hang until context deadline if we get ping.WantPingFinishedEvents
after we made the request but before we get the response.
Example (Close request in cmd-nsc-vpp):
- We make [WantPingFinishedEvents](https://github.com/networkservicemesh/sdk-vpp/blob/6f452c26d4133f75226866dedabce129eb84605f/pkg/tools/heal/liveness_check.go
Question about change-nse-dinamically example
Status: RESOLVED.
Healing receives the invalid monitor event
Status: RESOLVED.
Impossible to start basic deployment on MacOS/Arm64
Status: RESOLVED.
Use the latest alpine version for the image
Status: RESOLVED.
Monitor Connection states and event types
Status: RESOLVED.
Update deps Github workflow doesn’t update dependencies of dependencies in some repos
Status: RESOLVED.
Too high probability of name collision in getNameFromConnection
Status: RESOLVED.
Logging system problems
1. Increase log file size - https://github.com/networkservicemesh/integration-tests/issues/2917
2. Tests from different suites but with the same name overwrite each other.
For example:
a. Run TestBasicSuite/TestKernel2Kernel
b. Save logs to ns-kerenel2kernel
c. Run TestMemorySuite/TestKernel2Kernel
d. ns-kerenel2kernel
logs will be overwritten
3. Logs are not complete.
For example:
a. The last bash command from one of the healing test runs at
time=2024-01-19T13:51:50Z level=info msg=1 TestRunHealSuite/TestRemote_nsmgr_remote_endpoint=exitCode
b. But the last line in cmd-nsc logs is:
Jan 19 13:50:39.017[37m [TRAC] [id:alpine-beef03a3-a58e-48a5-9100-d853c81da3de-0]
More than 1 min missed
Healing with two NSEs could keep previous IPContext values
Status: RESOLVED.
Investigate why forwarder could restart too long
Status: RESOLVED.
[R&D] Add examples with using VMWare Tanzu
Status: RESOLVED.
Documentation improving
Status: RESOLVED.
Use Github CLI to create releases
Status: RESOLVED.
TestK8sMonolithSuite/External_nsc/TestKernel2IP2Kernel is unstable
Status: RESOLVED.
Add tag when pushing to main
Currently, we only tag images: ghcr.io/networkservicemesh/govpp/vpp:v23.10-rc0-166-g4a6d39d90
It is also necessary to tag the commit from which this image was built.
Refactor “parallel” package with passing excludedTests via Options
Status: RESOLVED.
Add dual stack support
Currently vL3-endpoint only support single stack: https://github.com/networkservicemesh/cmd-nse-vl3-vpp/blob/main/main.go
CI/CD: Sometimes release specific workflows are triggering on the main branches
Have a look at release workflow runs in cmd repos, for example:
https://github.com/networkservicemesh/cmd-nsc/actions/workflows/release.yml https://github.com/networkservicemesh/cmd-nse-vl3-vpp/actions/workflows/release.yml
Release workflows are skipped for the main branches, and that’s fine, but they also can be just ignored by CI. by release prefix as it is done in the deployments.-k8s repo: https://github.com/networkservicemesh/deployments-k8s/actions/workflows/release.yaml (note: release workflow is running only on release branches)
Warnings in github actions
Status: RESOLVED.
Fix CVEs
Status: RESOLVED.
Packet cluster setup does not work
The configuration inside the servers has probably changed. Start with the latest cluster API - https://github.com/networkservicemesh/integration-k8s-packet/blob/main/scripts/setup-cluster.sh
Option to disable control characters in logs
Status: RESOLVED.
Fix heal tests on release candidate v1.13.0-rc.2
Status: RESOLVED.
Fix integrations with service meshes
Status: RESOLVED.
Add blog section
Status: RESOLVED.
Release project board
Table of contents