Release v0.9.0

[marquiz]

Release Process

• [x] In the issue description, add a changelog section, describing changes since the last release.
• Local release preparations
  • [x] Perform mandatory internal release checks and preparations.
  • [x] Run make release-tests to run an extended set of tests prior to a release.
  • [x] Run git tag -a -m "CRI Resource Manager release $VERSION" $VERSION.
  • [x] Create source+dependencies tarball with make vendored-dist.
• Final verification of artefacts
  • [ ] Verify the installation of binary packages.
  • [x] Verify runnability of container images.
• Publishing
  • [x] Push the tag with git push $VERSION.
  • [x] Create a new draft release corresponding to the tag.
  • [x] Upload all artefacts to the release.
  • [x] Write the change log to the release.
  • [x] Mark the release as a non-production pre-release if necessary.
  • [x] Save as draft.
  • [x] Check that new container images are published for the tag.
  • https://hub.docker.com/r/intel/cri-resmgr-agent/tags
  • https://hub.docker.com/r/intel/cri-resmgr-webhook/tags
  • [x] Get the change log OK'd by other maintainers.
  • [x] Publish the draft as a release.
  • [x] Add a link to the tagged release in this issue.
• [x] Close this issue.

What's Changed

New dynamic-pools policy

A new policy that allows applications to be assigned to dynamically resized CPU pools. The pools are non-overlapping and are resized based on the resource requests of the containers and the actual CPU utilization of the application.

Enhancements to balloons policy

New PreferSpreadOnPhysicalCores configuration option in the balloons policy. Enabling this option for a balloon makes the CPU allocator to prefer CPUs from separate physical CPU cores.

There are also bugfixes to dynamic configuration updates.

Config status

Configuration status (was the configuration successfully applied or not) is now exposed via readiness probe of the cri-resmgr-agent. (#986)

CRI v1alpha2 dropped

Support for deprecated CRI version v1alpha2 was removed. This means that CRI-RM v0.9 requires Kubernetes v1.23 (or later) and containerd v1.6 (or later) or CRI-O v1.20 (or later).

Bugfixes

• CPU priority detection fixed (#971)
• Collecting of hugetlb statistics from cgroup v1 fixed (#968)
• Fixed sudden cri-resmgr process exit on page demotion (#1054)

List of PRs

• sysfs: add SysRoot() getter for SetSysRoot() by @askervin in https://github.com/intel/cri-resource-manager/pull/970
• cpuallocator: fix cpu priority discovery by @marquiz in https://github.com/intel/cri-resource-manager/pull/971
• Policy: create dynamic-pools policy for cri-rm by @Yugar-1 in https://github.com/intel/cri-resource-manager/pull/969
• rdt: stop trying to get container cgroup dir by @marquiz in https://github.com/intel/cri-resource-manager/pull/975
• metrics: fix collecting hugetlb statistics from cgroup v1 by @askervin in https://github.com/intel/cri-resource-manager/pull/968
• build(deps): bump golang.org/x/net from 0.2.0 to 0.7.0 by @dependabot in https://github.com/intel/cri-resource-manager/pull/976
• Bump builds to use Golang v1.19.6 by @marquiz in https://github.com/intel/cri-resource-manager/pull/981
• pkg/avx/collector: don't crash on no regexp match. by @klihub in https://github.com/intel/cri-resource-manager/pull/985
• build(deps): bump golang.org/x/sys from 0.0.0-20200202164722-d101bd2416d5 to 0.1.0 in /pkg/topology by @dependabot in https://github.com/intel/cri-resource-manager/pull/987
• agent: implement config status readiness probe. by @klihub in https://github.com/intel/cri-resource-manager/pull/986
• Migrating from CRI-RM to NRI by @luukasmakila in https://github.com/intel/cri-resource-manager/pull/988
• Add SNYK and CodeQL security scanning by @intel-k8s-bot in https://github.com/intel/cri-resource-manager/pull/991
• Bump builds to use Golang v1.20.2 by @marquiz in https://github.com/intel/cri-resource-manager/pull/982
• build(deps): bump github.com/opencontainers/runc from 1.1.4 to 1.1.5 by @dependabot in https://github.com/intel/cri-resource-manager/pull/992
• topology-aware: fix logging of cpuset changes. by @klihub in https://github.com/intel/cri-resource-manager/pull/995
• dockerfiles: (re)name suse Dockerfile as expected. by @klihub in https://github.com/intel/cri-resource-manager/pull/999
• pkg/topology: perform colonoscopy on testdata. by @klihub in https://github.com/intel/cri-resource-manager/pull/1005
• Update deps by @marquiz in https://github.com/intel/cri-resource-manager/pull/1008
• build(deps): bump pygments from 2.13.0 to 2.15.0 in /docs by @dependabot in https://github.com/intel/cri-resource-manager/pull/1013
• go.mod: update dependencies by @marquiz in https://github.com/intel/cri-resource-manager/pull/1014
• chore: remove refs to deprecated io/ioutil by @testwill in https://github.com/intel/cri-resource-manager/pull/1012
• docs/deps: update pygments to v2.15.1 by @marquiz in https://github.com/intel/cri-resource-manager/pull/1021
• Bump golang to v1.20.7 by @marquiz in https://github.com/intel/cri-resource-manager/pull/1018
• github: replace snyk with trivy by @marquiz in https://github.com/intel/cri-resource-manager/pull/1020
• github: take golang version from go.mod by @marquiz in https://github.com/intel/cri-resource-manager/pull/1022
• scripts: drop unused stuff from docker-build-image by @marquiz in https://github.com/intel/cri-resource-manager/pull/1024
• docs: use ADD in the dockerfile to fetch go tarball by @marquiz in https://github.com/intel/cri-resource-manager/pull/1025
• Makefile: prepare binary packages as release assets by @marquiz in https://github.com/intel/cri-resource-manager/pull/1026
• cross-build: get golang binaries instead of compiling from source by @marquiz in https://github.com/intel/cri-resource-manager/pull/1027
• Use golang builtin multierror by @marquiz in https://github.com/intel/cri-resource-manager/pull/1023
• Makefile: isolate image-push from image target by @marquiz in https://github.com/intel/cri-resource-manager/pull/1028
• github: drop the turnstyle plugin by @marquiz in https://github.com/intel/cri-resource-manager/pull/1030
• github: run security scanning on release branches by @marquiz in https://github.com/intel/cri-resource-manager/pull/1029
• github: drop containerized build of docs from verify workflow by @marquiz in https://github.com/intel/cri-resource-manager/pull/1031
• [backport from nri-plugins] all: stop importing kubernetes/kubelet internals. by @klihub in https://github.com/intel/cri-resource-manager/pull/1010
• github: refactor docs building by @marquiz in https://github.com/intel/cri-resource-manager/pull/1032
• fixes: fix a few innocent codeql alert. by @klihub in https://github.com/intel/cri-resource-manager/pull/1034
• Makefile: remove -it arg from docker run by @marquiz in https://github.com/intel/cri-resource-manager/pull/1035
• Makefile: fix creation of deb release assets by @marquiz in https://github.com/intel/cri-resource-manager/pull/1036
• github: split security scanning into re-usable jobs by @marquiz in https://github.com/intel/cri-resource-manager/pull/1033
• github: use pinpointed ubuntu version on the runners by @marquiz in https://github.com/intel/cri-resource-manager/pull/1040
• Makefile: drop unwanted update-workflows target by @marquiz in https://github.com/intel/cri-resource-manager/pull/1038
• scripts: remove unused batch file for docs building by @marquiz in https://github.com/intel/cri-resource-manager/pull/1041
• github: refactor verify workflow by @marquiz in https://github.com/intel/cri-resource-manager/pull/1039
• github: add workflows for image building by @marquiz in https://github.com/intel/cri-resource-manager/pull/1037
• github: build and upload release assets in release workflow by @marquiz in https://github.com/intel/cri-resource-manager/pull/1042
• github: fix the usage of github environments by @marquiz in https://github.com/intel/cri-resource-manager/pull/1044
• github: add trivy license scanning by @marquiz in https://github.com/intel/cri-resource-manager/pull/1046
• go.mod: update goresctrl to v0.5.0 by @marquiz in https://github.com/intel/cri-resource-manager/pull/1043
• github: split code and docs CI into separate workflows by @marquiz in https://github.com/intel/cri-resource-manager/pull/1045
• pkg/topology: sync go.mod with the main module by @marquiz in https://github.com/intel/cri-resource-manager/pull/1047
• e2e: clean up vm as the last step in the static-pools test suite by @askervin in https://github.com/intel/cri-resource-manager/pull/1049
• e2e: fix topology-aware/n4c16/test09-container-exit test by @askervin in https://github.com/intel/cri-resource-manager/pull/1050
• e2e: update default distro from Ubuntu 20.04 to 22.04 by @askervin in https://github.com/intel/cri-resource-manager/pull/1052
• e2e: fix distro=opensuse to support k8s 1.27+ by @askervin in https://github.com/intel/cri-resource-manager/pull/1051
• demoter: fix sudden cri-resmgr process exit on page demotion by @askervin in https://github.com/intel/cri-resource-manager/pull/1054
• e2e: relax dynamic demotion first detection round requirement by @askervin in https://github.com/intel/cri-resource-manager/pull/1053
• e2e: fix opensuse-15.4 image URL, support distro=opensuse-15.5 by @askervin in https://github.com/intel/cri-resource-manager/pull/1055
• e2e: use latest image for debian-10 tests by @marquiz in https://github.com/intel/cri-resource-manager/pull/1056
• e2e: use Fedora 38 for the fedora test target by @marquiz in https://github.com/intel/cri-resource-manager/pull/1057
• e2e: use bridge cni plugin by default by @marquiz in https://github.com/intel/cri-resource-manager/pull/1058
• e2e: ignore terminated processes on fetching allowed resources by @askervin in https://github.com/intel/cri-resource-manager/pull/1065
• e2e: fix "ambiguous allowed resources" caused by race by @askervin in https://github.com/intel/cri-resource-manager/pull/1062
• e2e: add --wait to kubectl delete of pods and namespaces by @marquiz in https://github.com/intel/cri-resource-manager/pull/1059
• e2e: check default serviceaccount in cluster readiness check by @marquiz in https://github.com/intel/cri-resource-manager/pull/1060
• Makefile: simplify packaging tests by @marquiz in https://github.com/intel/cri-resource-manager/pull/1061
• e2e: drop vm-force-restart() by @marquiz in https://github.com/intel/cri-resource-manager/pull/1064
• e2e: support for verify --retry N by @askervin in https://github.com/intel/cri-resource-manager/pull/1069
• e2e: don't use pyexec in tests directly by @marquiz in https://github.com/intel/cri-resource-manager/pull/1066
• e2e: relax topology-aware coldstart test by @marquiz in https://github.com/intel/cri-resource-manager/pull/1068
• e2e: fix starting a custom containerd version by @askervin in https://github.com/intel/cri-resource-manager/pull/1072
• build(deps): bump golang.org/x/net from 0.8.0 to 0.17.0 by @dependabot in https://github.com/intel/cri-resource-manager/pull/1073
• balloons: PreferSpreadOnPhysicalCores & balloon-specific CPU allocator by @askervin in https://github.com/intel/cri-resource-manager/pull/1071
• go.mod: update deps by @marquiz in https://github.com/intel/cri-resource-manager/pull/1074
• test/e2e: fix incorrect Allocat[ion|or]Priority config key. by @klihub in https://github.com/intel/cri-resource-manager/pull/1077
• balloons: fix applying new configuration by @askervin in https://github.com/intel/cri-resource-manager/pull/1076
• Bump golangci-lint to v1.55.2 by @marquiz in https://github.com/intel/cri-resource-manager/pull/1082
• Bump Go to v1.21.5 by @marquiz in https://github.com/intel/cri-resource-manager/pull/1080
• pkg/sysfs: drop discovery flags by @marquiz in https://github.com/intel/cri-resource-manager/pull/1083
• github: generate codeql report on release tags by @marquiz in https://github.com/intel/cri-resource-manager/pull/1084
• go.mod: update dependencies by @marquiz in https://github.com/intel/cri-resource-manager/pull/1079
• pkg/cri: remove support for the obsolete v1alpha2 CRI protocol. by @klihub in https://github.com/intel/cri-resource-manager/pull/1081
• pkg/cri: implement container event relaying. by @klihub in https://github.com/intel/cri-resource-manager/pull/1085

New Contributors

• @Yugar-1 made their first contribution in https://github.com/intel/cri-resource-manager/pull/969
• @luukasmakila made their first contribution in https://github.com/intel/cri-resource-manager/pull/988
• @testwill made their first contribution in https://github.com/intel/cri-resource-manager/pull/1012

Full Changelog: https://github.com/intel/cri-resource-manager/compare/v0.8.0...v0.9.0

[marquiz]

Drafted a changelog. PTAL @klihub @kad @askervin

[askervin]

@marquiz , looks great to me! Thank you!

[kad]

LGTM. one small addition: for CRI v1alpha2 deprecation, minimal version of cri-o also should be mentioned.

[marquiz]

LGTM. one small addition: for CRI v1alpha2 deprecation, minimal version of cri-o also should be mentioned.

Added.

[marquiz]

Container event relaying was implemented. Let's tag? 🕺

[marquiz]

GitHub release page: https://github.com/intel/cri-resource-manager/releases/tag/v0.9.0