Back to index
4.4.0-0.okd-2020-03-28-092308
Download installer and client with:
oc adm release extract --tools registry.svc.ci.openshift.org/origin/release:4.4.0-0.okd-2020-03-28-092308
Tests:
Upgrades from:
Upgrades to:
Created: 2020-03-28 09:30:15 +0000 UTC
Image Digest: sha256:79f2415cc7b702866c0ccb974d5dabc7b85d0eacb5537680c846df7ad6e06352
Components
- Kubernetes 1.17.1
- Fedora CoreOS upgraded from 31.20200127.20 to 31.20200327.20
New images
Removed images
Rebuilt images without code change
- cvo-overrides.yaml: point OKD releases to release-controller #3367
- Update initial FCOS images to 31.20200310.3.0 #3349
- Merge fcos-devel #3338
- ignition.Config Spec v3 to runtime.RawExtension conversion #3344
- ovirt: Ask for the CA pem content ovirt API #3322
- Fix unavailable forked deps and update vendor #3238
- bootstrap: check files for duplicates before appending #3209
- start GCP services on bootstrap #3095
- pkg: always overwrite files created #3168
- Bug 1803174: OpenStack CI Install awscli from zip #3114
- don’t include duplicate files in bootstrap.ign #3100
- backport several fixes from master #3070
- pkg/asset/machines: avoid generating duplicate MC with ssh keys #3061
- pkg/asset: Force phony pull secret in place of no pull secret #3020
- GCP: add a script and a service which ensure GCP routes are set correctly #3067
- support xz-encoded images #3008
- cherrypick UPI fixes #3013
- openstack: remove “filesystem” from tf files #3012
- Full changelog
- BUG 1804826: UPSTREAM: <carry>: openshift: Add topology.kubernetes.io labels to be ignored when comparing similar node groups #127
- BUG 1805153: Ensure DeleteNodes doesn’t delete a node twice #128
- Bug 1801254: UPSTREAM: <carry>: openshift: Let Nodes() return the list of all machines #122
- Rebase 1.17.0 #121
- Cluster Autoscaler version to 1.17.0 #2603
- Bump Cluster Autoscaler version to 1.17-beta.2 #2593
- Run hack/update-kubernetes-deps-in-e2e.sh #2595
- Snapshot cluster state for scheduler every loop #2588
- Make sure bindata is generated for VPA e2e on deps update #2587
- Ignore NominatedNodeName on Pod if node is gone #2576
- Sync e2e_test file with k/k master #2580
- Remove timeout when running VPA e2e #2569
- Use git reset instead of git restore #2561
- Add nilo19 to cluster autoscaler azure owners list. #2566
- Fix a bug in initializing scheduler #2565
- Run VPA e2e tests from the e2e dir #2563
- Run with -mod vendor when building in docker #2562
- Bump Cluster Autoscaler version to 1.17-beta.1 #2558
- fix: cluster-autoscaler version in Azure autorest user agent #2555
- Validate all parameters even when using managed identity #2541
- Fix running e2e tests for VPA #2553
- Revendor CA against tip of k8s.io/kubernetes #2515
- CSR: fix a bug in GetClusterSize #2546
- Improve docs on azure autodiscover installation #2540
- support multiple groups and GPU group for baiduCloud #2539
- Add lease election object rbac rules #2534
- Remove godeps usage from Makefile #2536
- VPA: Store annotation metadata in internal structure #2535
- VPA: separate module for e2e tests #2531
- updated cluster-autoscaler readme - compatibility chart #2518
- update readme #1
- Fix metrics for list and delete instances #2533
- Rename STOCKOUT to RESOURCE_POOL_EXHAUSTED #2521
- remove temporary nodes logic #2520
- Improve performance of the Azure cloud provider #2474
- Doc: fix the Allocation Strategies link in AWS CA README #2492
- Allow “custom” DaemonSet pods #2483
- Deflake vpa ac e2e #2504
- Clarify VPA logging #2496
- CA: Make error message in scale down node draining consistent #2476
- cluster-autoscaler/gce: create a test for an instance with unknown mig #2475
- Add tests for Packet #2490
- Update VPA dependencies to k8s 1.16.2 #2486
- Add AWS ignore labels and increase max memory difference. #2462
- Fix config assignment issue #2494
- use simple string prefix placeholder match #2480
- feat: enable min/max on Azure VMSS autodiscovery #2121
- Fix ineffassign failures #2477
- Add rate limiter to the Vertical Pod autoscaler updater component #2326
- Update RBAC in Azure examples to include csinodes permission #2484
- fix golint failures #2478
- minor documentation fixes #2482
- Extract separate package for core/utils.go and split into multiple files. #2407
- doks: use providerID for node.ID #2427
- doks: fix max size check, support increase to max #2426
- Remove krzysztof-jastrzebski from OWNERS #2471
- cluster-autoscaler: fix some typos in the code #2448
- Update magnum example deployment with image built by magnum CI #2461
- Less aggresive create errors reporting for GCE #2456
- fix the webhook-address and webhook-port reference for vpa-admission-… #2455
- Generate EC2 Instance Type list in runtime #2249
- Rename GetCreatedNodesWithOutOfResourcesErrors to GetCreatedNodesWithErrors #2450
- priority expander: Documentation updates #2433
- Remove e2e tests for v1beta1 #2437
- AWS – use
session.NewSession instead of session.New to #2428
- Bump go to 1.12.10 to mitigate security issues #2420
- aws: examples: Delete trailing whitespace #2403
- Update RBAC in examples to include csinodes permission #2404
- Move default VPA deployment to 0.6.3 #2416
- Rebase VPA to distroless #2412
- Add g4dn instance type for AWS #2401
- CA: Invalidate node instances cache after deleting failed/unregistered nodes #2405
- Switch leader election mechanism to use lease objects #2012
- add prom scrap annotations #2375
- Allow Azure VMSS to scale from 0 with taints and labels #1982
- add kops instance group label to ignore list for similar node group identification. #2207
- ensure valid AWS LaunchTemplate version #2385
- Explicitly set SideEffects=None for mutating webhook #2344
- add comments and unit test #2360
- Bugfix - Allow arbitrary placeholder AWS instance names #2379
- rename clusterrole system:admission-controller to system:vpa-admission-controller #2336
- Update redis.yaml #2361
- Increase default GCE operation wait timeout to 20s #2363
- update baidu cloudprovider #2354
- update #3
- update #2
- Don’t treat lack of queried Kubernetes node in AWS cloud provider cache as error #2357
- Vertical Pod Autoscaler version 0.7.0 #2351
- Switch deployment to 0.6.1 version #2346
- Make error regexp less strict #2341
- AWS: Decrease current size only once #2296
- feat: add NV6Promo instance type to CA #2332
- Updated aws-sdk-go to 1.23.18 #2323
- cluster-autoscaler: blocked if an instance is detached from MIG #2281
- cluster-autoscaler/hack/update-vendor.sh: fix invocation of git commit #2313
- Use DO node pool autoscale API integration #2318
- Bump Cluster Autoscaler version to 1.17.0-alpha.1 #2317
- Bump Cluster Autoscaler version to 1.16 #2315
- cloudprovider/packet: clarify encoding of API token in README #2312
- Full changelog
- 4.4: Bug 1816826: Move to namespace openshift-cluster-storage-operator #35
- Bug 1808337: Update getters to match upstream #31
- Bug 1798448: Use Origin images from quay.io/openshift #28
- Bug 1797694: Add unit tests #27
- Bug 1797694: Remove unused variable and file #26
- Fix must-gather namespace names #24
- Update Deployment on image version change #23
- Remove unused deployment lister #25
- Wait for informer sync on startup #22
- Full changelog
- Bug 1813914: Disable serving for cert regeneration controller to avoid wrong dependency #798
- Bug 1812210: use localhost name instead of localhost IPs #792
- Bug 1771410: Use the new tls-server-name option in kubeconfig #786
- Bug 1797593: Revert “force cert rotation every couple days for development” #771
- Bug 1771410: Disable delegated auth for recovery #775
- Bug 1805519: skip using bootstrap etcd #774
- Bug 1805445: [release-4.4] baremetal: Explicitly advertise the host IP chosen by kubelet #773
- Bug 1797601: force cert rotation every couple days for development: aggregator-client-signer #767
- Bug 1802247: Check operator readiness before each bound token controller test #763
- Bug 1793850: Update UsingDeprecatedAPIExtensionsV1Beta1 alert #764
- Bug 1785498: regenerate-certificates command blocked by error illegal base64 data at input byte 3 #761
- Bug 1793850: Update UsingDeprecatedAPIExtensionsV1Beta1 alert #762
- Bug 1793850: Update UsingDeprecatedAPIExtensionsV1Beta1 alert #742
- Bug 1794376: Remove IPv6 dualstack feature gate #758
- Bug 1801154: manifests/: Throw away unused high cardinality apiserver duration buckets #752
- Bug 1798609: Report a metric about proxy enablement #743
- recovery: update kube-apiserver container name #755
- Bug 1798725: static pod operator cannot be composed #754
- bindata: preselect container name for oc #753
- Bug 1798214: Add a PrometheusRule to aggregate in-flight requests #737
- Bug 1693951: consume the master kubelet bootstrap credential from a separate confi… #747
- Bug 1798048: bindata: remove revision suffix from containers #750
- Bug 1797759: update library-go to use cached client #729
- Bug 1797601: force cert rotation every couple days for development #748
- Bug 1797657: increase the max-requests limits #733
- bump library-go to get controller factory #739
- manifests: set caFile for service monitor #504
- Cleanup post-bootstrap config handling #738
- Enable changing bound service account token issuer from default #727
- Add cert regeneration controller #635
- Enable bound SA tokens #718
- Full changelog
- Bug 1814729: Exclude Kubernetes control plane rules when running on IBM Cloud #712
- Bug 1772564: Bring in new alert via dependency #669
- Bug 1807140: jsonnet/rules.jsonnet: Add MultipleContainersOOMKilled alert #676
- [release 4.4] Bug 1807843: Fix more many to many errors #684
- Bug 1807843: fix many-to-many errors #682
- Bug 1807376: pkg/manifests: rename shared config’s keys to include ‘Public #680
- Bug 1805190: pkg/manifests/telemeter: bump size limit to 5mb #665
- Bug 1804012: Report system manufacturer via telemetry #661
- Bug 1804179: pkg/client: set CreationTimestamp at create time, not only at update time #657
- Bug 1803827: pkg: Generate recording rule to measure total number of series we are #653
- Bug 1794616: pkg/manifests: set modified telemeter-client command line arguments #655
- Bug 1803192: *: duplicate shared configmap in openshift-config-managed namespace #654
- Bug 1787520: fix query showing metrics collected by telemetry #652
- Bug 1801358: update prometheus-operator v0.35.1 #651
- Bug 1794616: replace up with count:up[0,1], remove node_uname_info #650
- Bug 1799539: Node exporter init collectors must be privileged #645
- Bug 1794934: Bump kubernetes-mixin to support DASD device metric reporting #644
- Bug 1799057: Send apiserver_current_inflight_requests in telemeter. #643
- BUG 1796927: Documentation: Transfer telemetry comments #640
- Bug 1796468: *: Complete new configuration scheme #633
- Bug 1796473: Revert “jsonnet: regenerate jsonnetfile.* with latest jb” #638
- Bug 1795538: Fix excessive permissions for prometheus-operator cluster-role #637
- Duplicate Grafana configmaps from monitoring to console #635
- github: add release checklist issue template #634
- go.mod,go.sum,vendor,jsonnet: Bump and lock dependencies #624
- jsonnet/kube-state-metrics.jsonnet: Fix blacklist arguments #629
- Full changelog
- Bug 1814621: storage urls observer: prevent event/log spam #339
- Bug 1811636: Fix rollout strategy to never run more than one pod per node #332
- Bug 1795163: waits for extension-apiserver-authentication before rolling out a new version #328
- Bug 1805872: 4.4 cherry-pick of switch from daemonset to deployment #325
- Bug 1806681: openshift-apiserver should never use the bootstrap etcd member #322
- Bug 1805534: openshift-apiserver should directly use the host-etcd IPs #315
- Bug 1803922: encryption: don’t degrade when daemonset is not found #311
- Bug 1801154: manifests/: Throw away unused high cardinality apiserver duration buckets #309
- Bug 1798215: Add a new label ‘apiserver’ to identify the source #307
- make static resource controller you simply pass manifests to #304
- Full changelog
- Bug 1813474: list all peers in initial-cluster #39
- Bug 1808546: If we weren’t able to get client or target member go ahead and start #37
- Bug 1808073: fix archive member name, unmask error #35
- Bug 1806752: [release 4.4] Archive data dir if unstarted 4.4 #32
- Bug 1805809: create logic for golang ETCD_INITIAL_CLUSTER #27
- Bug 1801237: *: bump etcd to v3.3.18 #21
- Automated cherry pick of #11374 on release 3.3 #11394
- Automated cherry pick of #11308 on release-3.3 #11314
- Automated cherry pick of #11247 #11253
- cherry pick “etcd_cluster_version” metric” (#10257, #11233, #11254, #11265) to release-3.3 #11261
- scripts: avoid release builds on darwin machine. #11245
- Full changelog
- Bug 1817039: [4.4] UPSTREAM: 88506: Honor status.podIP over status.podIPs, node.spec.podCIDR over node.spec.podCIDRs #24762
- Bug 1816423: UPSTREAM: 87913: Add code to fix kubelet/metrics memory issue #24753
- Lock openshift/* vendored repos to release-4.4 and update vendor to head before GA #24743
- Bug 1811916: UPSTREAM: 86886: Fix err variable shadowing issue in storage/utils #24666
- Bug 1806407: Add a delay between successive retries for service reachability check and wait for n consecutive successes. #24684
- Bug 1806697: Don’t instrument etcd operator #24669
- Bug 1813429: Restore globally-writeable /etc/passwd in tests image #24691
- Bug 1812877: fixes configmap “extension-apiserver-authentication” not found #24677
- Bug 1810722: Kubelet should not remove restart never pods until all status is reported #24649
- Bug 1811198: /readyz should start returning failure on shutdown initiation #24656
- Bug 1796660: UPSTREAM: 88146: remove duplicate pv delete calls #24625
- Bug 1811062: UPSTREAM: 79083: Add ability to reload client certificates from disk #24647
- : Bug 1811088: Fix count of failures in JUnit output #24648
- Bug 1806407: Improve upgrade availability tests and add sequential upgrade input #24589
- Bug 1809712: Don’t send no-op pod status patches #24639
- Bug 1805182: openshift-apiserver degraded due to expired certificate errors #24621
- Bug 1809222: UPSTREAM: 88663: update kube-controller-manager and kube-scheduler to match kube-apiserver defaults #24615
- Bug 1806786: UPSTREAM: 88251: Partially fix incorrect configuration of kubepods.slice unit by kubelet #24596
- Bug 1806176: kubelet dependency bumps (runc, cadvisor, and golang/sys) #24586
- Bug 1804937: builds: Only check ocm progress state #24587
- Bug 1806491: Revert: test/extended/prometheus: temporarily disable etcdInsufficientMembers” #24590
- Bug 1804514: kubelet: Make default journal format precise #24576
- Bug 1805791: picks dynamic reloading for CSR signing controllers #24583
- Bug 1801095: Fix image resolve plugin on updates and add tests #24571
- Bug 1798131: Fix CSI snapshot tests #24546
- Bug 1793340: Revert “images/tests: Globally-writeable /etc/passwd” #24556
- Bug 1792569: UPSTREAM: : admission/restrictusers ensure groups cache synced #24411
- Revert Disable flaky tests on OpenStack b390f42a42 #24490
- Bug 1803190: test/extended/prometheus: temporarily disable etcdInsufficientMembers #24547
- Bug 1803071: UPSTREAM: 87673: blank out value for unbounded client label #24541
- Bug 1768756: Run Prometheus specific tests cases at the end of the test run #24492
- Bug 1802713: disable etcd metrics check to land operator #24538
- Bug 1746149: test if oc tag can work without imagestreamtags #24505
- Bug 1722672: UPSTREAM: 82027: Enabled reading config files for vsphere e2e tests #24478
- test: Move the prometheus alerts test to the beginning of the suite #24499
- Bug 1799491: Disable Azure OLM etcd operator test due to flakiness #24512
- Bug 1767523: UPSTREAM: 84614: kubelet: add certificate rotation error metric #24513
- Bug 1798450: makes unavailableGauge metric to always reflect the current state of a service #24496
- Bug 1789266: replace deprecated commands with current secrets commands #24466
- Bug 1794885: UPSTREAM: <carry>: bootstrap user - make tokens have implicit audiences #24503
- Bug 1797634: test/extended/prometheus: test if AlertmanagerReceiversNotConfigured is firing #24277
- Bug 1796895: Don’t use annotation in openshift/csi test suite #24484
- Bug 1796715: Add a service upgrade test that verifies availability #24471
- Bug 1779303: UPSTREAM: 87242: SchedulerPredicates [Serial] validates resource limits of pods that are allowed to run - set resource requests as well #24498
- Bug 1798131: Enable CSI snapshot tests #24449
- Bug 1747479: UPSTREAM: 87813: kube-aggregator: increase log level of AggregationController API group logging #24497
- Test internal port connectivity internally for 9000-9999 #24455
- Bug 1797740: UPSTREAM: <carry>: simplify the authorizer patch to allow the flags t… #24435
- Bug 1796483: Use –kubeconfig flag for Jenkins Tests #24448
- Bug 1796234: Move the oc explain tests for networking CRDs to a new test #24477
- UPSTREAM: 85898: Fix multinode storage e2e tests for multizone clusters #24437
- test/extended/prometheus: Re-enable telemetry test #24439
- UPSTREAM: <carry>: oauth-authn: add implicit audience support #24461
- test: Set appropriate cloud defaults even when not specified on tests #24463
- image: add ImageTag to etcd test #24458
- UPSTREAM: 87432: fix flaky test “should create a ResourceQuota and ca… #24446
- test: Generate test name updates rather than calculate dynamically #24445
- Bug 1769015: Uncomment the lines to verify oc explain for networking related CRDs #24222
- Full changelog
- Bug 1800430: Sync with 1.9 release branch #26
- *: Bring in v1.9.5 #25
- Cut v1.9.5 release #1067
- internal/store/validatingwebhookconfiguration.go: Switch to v1 #1052
- *: Cut v1.9.4 #1044
- internal/store/builder.go: Fix K8s API version #1043
- fix: use the correct expectedType for networkPolicy reflector #1042
- Bug 1798938: *: Bring 1.9.4 changes #24
- Full changelog
- templates: Use new node-ip set command #1596
- Switch to back to Ignition Spec v3.0 parser #1595
- Use v3.0.0 for Ignition templates #1584
- Merge fcos-devel #1560
- Use mcd subcommand to determine node ip #1565
- non-virtual-ip: replace the script with podman wrapper #1478
- remove the etcd-member pod because we no longer need it #1500
- pkg/daemon/update.go: check for nil pointers #1498
- GCP: add a script and a service which ensure GCP routes are set correctly #1442
- pkg/server: serve config only to master in bootstrap server #1423
- Cherry-pick commits to enable cluster-etcd-operator #1415
- Full changelog
- [release 4.4] Bug 1811202: /readyz should start returning failure on shutdown initiation #81
- Bug 1804934: employ k8s label value validation when creating build pod build label… #77
- Bug 1803074: 4.4: UPSTREAM: 87673: blank out value for unbounded client label #74
- Bug 1788700: include ISI error message changes when determining new generation need #69
- Bug 1795705: return actual instantiate error for binary builds #71
- Bug 1795705: increase error logging around apiserver binary build flow #70
- Add the imagetags resource to v1 #51
- Full changelog
- Bug 1811760: Some cluster operators fail to come up because RHV CA is not trusted by a pod #42
- Bug 1803821: [ovirt] oVirt API token doesn’t get renewed if the session is expired #38
- Change image to build FROM openshift/origin-v4.0:base in Dockerfile #36
- Full changelog
- Bug 1810816: [4.4] CARRY: ovn: fix cloud load balancer rules for IPv6 #113
- Bug 1806666: merge master to release-4.4 #99
- ipv6: Make IPAddrToHWAddr() aware of IPv6 #101
- HACK: add ipv4 interfaces to pods that need them #94
- Upstream merge 2020-02-18 #93
- fix IPv6 subnet dynamic addressing #1072
- config: don’t let environment affect test results #1071
- add the ability for ovn-kubernetes to ignore a node by label #1048
- Fixes accidental override of file conf settings with defaults #1044
- Fix delete order for network policy #1068
- Remove kube-proxy in KIND deployment #1067
- remove unused methods from KClient #1053
- Ensure kind logs are always exported #1065
- ci: always export logs #1054
- Clustered database (aka RAFT) support for OVN DBs #1051
- Bug 1795454: Resync to upstream #87
- when performing static ip allocation other-config:subnet is not required #1046
- Remove the compat annotation code for OvnPodAnnotationLegacyName and OvnHostSubnetLegacy #1045
- ci: remove Travis now that GitHub actions work #1039
- J2 render all templates #1041
- ci: bump master loglevel to 5 #1040
- CNI logging and timeout fixes, plus node handler locking fix #1036
- ci: use consistent kind cluster name and fix artifact retrieval #1037
- Also run CI for pull requests #1033
- Replace Travis CI with GitHub Actions #1028
- Iptables followup #1032
- Ensure j2 does not fail to run in daemonset.sh if installed globally #1031
- Fixes specifying geneve port #1030
- fix for iptables issue in CentOS 8 #1027
- do not delete integration bridge as part of ovnkube-node cleanup #1029
- Yet more cleanups #1023
- travis: bump timeout for e2e tests #1024
- a way to specify the encap-ip for a k8s worker node #1016
- Fix kubetest to enable conformance #1018
- add support for custom MAC specified in default network-attachment #997
- Validates host subnet prefix length and default behavior #979
- ovn: read pod addresses more quickly #1012
- Enable e2e CI #1011
- Bug 1798613: OVN: Bump image to include ovn2.12 #72
- Bug 1796531: CARRY: fix race updating exclude IPs on node switch #83
- Full changelog
- Bug 1810418: Ensure service CA certs are created with unique serial numbers #111
- Bug 1803282: enable signer ca expiry timer metrics #95
- Update CA duration to 26 months to ensure upgrade after rotation #106
- Remove the option to disable automated CA rotation #107
- e2e test to check service serving cert using client pod #98
- Fix CA bundle annotation check to be case-insensitive #103
- Update new bundle injection targets to only support beta annotation #102
- Extend service CA duration to 14 months #101
- Enable injection of service CA bundle for admission webhooks #79
- Full changelog
- Bug 1813221: jsonnet/telemeter: set cpu requests, unset limits #320
- BUG 1796927: *: Move metrics whitelist to downstream usage #306
- Add annotation for new Observatorium API Route #305
- Full changelog
Source code for this page located on github