talos

v0.8.5 (2021-03-15)

Chore

• fix provision tests after changes to build-container

Fix

• move containerd CRI config files under /var/
• ignore 'ENOENT' (no such file directory) on mount

Images

quay.io/coreos/flannel:v0.12.0-amd64
ghcr.io/talos-systems/install-cni:v0.1.0-6-gdc32cc8
k8s.gcr.io/coredns:1.7.0
gcr.io/etcd-development/etcd:v3.4.14
k8s.gcr.io/kube-apiserver-amd64:v1.20.1
k8s.gcr.io/kube-controller-manager-amd64:v1.20.1
k8s.gcr.io/kube-scheduler-amd64:v1.20.1
k8s.gcr.io/kube-proxy-amd64:v1.20.1
ghcr.io/talos-systems/kubelet:v1.20.1
ghcr.io/talos-systems/pod-checkpointer:v0.1.0-6-gdc32cc8
ghcr.io/talos-systems/installer:v0.8.5

v0.9.0-beta.0 (2021-03-10)

New Features

• Control Plane as Staitc Pods
• ECDSA Keys for Kubernetes PKI
• Disk Encryption
• Virtual Shared IP for Control Plane Endpoint

More in the docs.

Upgrading

Please read upgrade notes before upgrading from Talos 0.8

Chore

• bump release scope to v0.9
• bump dependencies via dependabot
• update tools/pkgs/extras tags
• fix //nolint directives
• expire objects in CI S3 bucket
• fix upgrade tests by bumping 0.9 to alpha.5

Docs

• document controller runtime, resources and talosctl get

Feat

• update containerd to version 1.4.4

Fix

• do not print out help string if the parameters are correct
• mkdir source of the extra mounts for the kubelet
• properly propagate nameservers to provisioned docker clusters
• ignore connection refused errors when updating/converting cp
• align partition start to the physical sector size
• chmod etcd PKI path to fix virtual IP for upgrades with persistence

Release

• v0.9.0-beta.0: prepare release

Images

quay.io/coreos/flannel:v0.13.0
ghcr.io/talos-systems/install-cni:v0.2.0
docker.io/coredns/coredns:1.8.0
gcr.io/etcd-development/etcd:v3.4.15
k8s.gcr.io/kube-apiserver:v1.20.4
k8s.gcr.io/kube-controller-manager:v1.20.4
k8s.gcr.io/kube-scheduler:v1.20.4
k8s.gcr.io/kube-proxy:v1.20.4
ghcr.io/talos-systems/kubelet:v1.20.4
ghcr.io/talos-systems/installer:v0.9.0-beta.0
k8s.gcr.io/pause:3.2

v0.9.0-alpha.5 (2021-03-03)

Chore

• bump Go module dependencies
• properly propagate context object in the controller

Feat

• bypass lock if ACPI reboot/shutdown issued
• add --on-reboot flag to talosctl edit/patch machineConfig
• support JSON output in talosctl get, event types
• rename namespaces, resources, types etc

Release

• v0.9.0-alpha.5: prepare release

Images

quay.io/coreos/flannel:v0.13.0
ghcr.io/talos-systems/install-cni:v0.1.0-9-g302cc61
docker.io/coredns/coredns:1.8.0
gcr.io/etcd-development/etcd:v3.4.15
k8s.gcr.io/kube-apiserver:v1.20.4
k8s.gcr.io/kube-controller-manager:v1.20.4
k8s.gcr.io/kube-scheduler:v1.20.4
k8s.gcr.io/kube-proxy:v1.20.4
ghcr.io/talos-systems/kubelet:v1.20.4
ghcr.io/talos-systems/installer:v0.9.0-alpha.5
k8s.gcr.io/pause:3.2

v0.9.0-alpha.4 (2021-03-02)

Chore

• update provision/upgrade tests to 0.9.0-alpha.3

Docs

• bump v0.8 release version in the SBCs guides
• add disk encryption guide

Feat

• update linux kernel to 5.10.19

Fix

• ignore 'ENOENT' (no such file directory) on mount
• move etcd to cri containerd runner

Release

• v0.9.0-alpha.4: prepare release

Images

quay.io/coreos/flannel:v0.13.0
ghcr.io/talos-systems/install-cni:v0.1.0-9-g302cc61
docker.io/coredns/coredns:1.8.0
gcr.io/etcd-development/etcd:v3.4.15
k8s.gcr.io/kube-apiserver:v1.20.4
k8s.gcr.io/kube-controller-manager:v1.20.4
k8s.gcr.io/kube-scheduler:v1.20.4
k8s.gcr.io/kube-proxy:v1.20.4
ghcr.io/talos-systems/kubelet:v1.20.4
ghcr.io/talos-systems/installer:v0.9.0-alpha.4
k8s.gcr.io/pause:3.2

v0.9.0-alpha.3 (2021-03-01)

Chore

• bump dependencies via dependabot
• build both Darwin and Linux versions of talosctl
• bump dependencies via dependabot
• switch CI to stop embedding local registry into the builds

Docs

• update AMI images for 0.8.4

Feat

• implement etcd remove-member cli command
• update etcd to 3.4.15
• talosctl: allow v-prefixed k8s versions
• implement simple layer 2 shared IP for CP
• implement talosctl edit and patch config commands
• bump etcd client library to 3.5.0-alpha.0

Fix

• update in-cluster kubeconfig validity to match other certs
• add ApplyDynamicConfig call in the apply-config --immediate mode
• set hdmi_safe=1 on Raspberry Pi for maximum HDMI compatibility
• show stopped/exited containers via CRI inspector
• make ApplyDynamicConfig idempotent
• improve the drain function
• correctly set service state in the resource
• update the layout of the Disks API to match proxying requirements
• stop and clean up installer container correctly
• sanitize volume name better in static pod extra volumes

Refactor

• add context to the networkd
• split WithNetworkConfig into sub-options

Release

• v0.9.0-alpha.3: prepare release

Test

• add integration test with Canal CNI and reset API
• upgrade master to master tests

Images

quay.io/coreos/flannel:v0.13.0
ghcr.io/talos-systems/install-cni:v0.1.0-9-g302cc61
docker.io/coredns/coredns:1.8.0
gcr.io/etcd-development/etcd:v3.4.15
k8s.gcr.io/kube-apiserver:v1.20.4
k8s.gcr.io/kube-controller-manager:v1.20.4
k8s.gcr.io/kube-scheduler:v1.20.4
k8s.gcr.io/kube-proxy:v1.20.4
ghcr.io/talos-systems/kubelet:v1.20.4
ghcr.io/talos-systems/installer:v0.9.0-alpha.3
k8s.gcr.io/pause:3.2

v0.9.0-alpha.2 (2021-02-20)

Chore

• add default cron pipeline to the list of pipelines
• run default pipeline as part of the cron pipeline

Docs

• add link to GitHub Discussions as a support forum

Feat

• u-boot 2021.01, ca-certificates update, Linux file ACLs
• support control plane upgrades with Talos managed control plane
• add support for extra volume mounts for control plane pods
• add a warning to boot log if running self-hosted control plane
• add an option to disable kube-proxy manifest
• update Kubernetes to 1.20.4
• add state encryption support

Fix

• redirect warnings in manifest apply k8s client
• handle case when kubelet serving certificates are issued
• correctly escape extra args in kube-proxy manifest
• skip empty manifest YAML sub-documents

Refactor

• split kubernetes/etcd resource generation into subresources

Test

• enable disk encryption key rotation test
• update integration tests to use wrapped client for etcd APIs

Images

quay.io/coreos/flannel:v0.13.0
ghcr.io/talos-systems/install-cni:v0.1.0-9-g302cc61
docker.io/coredns/coredns:1.8.0
gcr.io/etcd-development/etcd:v3.4.14
k8s.gcr.io/kube-apiserver:v1.20.4
k8s.gcr.io/kube-controller-manager:v1.20.4
k8s.gcr.io/kube-scheduler:v1.20.4
k8s.gcr.io/kube-proxy:v1.20.4
ghcr.io/talos-systems/kubelet:v1.20.4
ghcr.io/talos-systems/installer:v0.9.0-alpha.2
k8s.gcr.io/pause:3.2

v0.8.4 (2021-02-19)

Fix

• drop cri dependency on etcd

Release

• v0.8.4: prepare release

Images

quay.io/coreos/flannel:v0.12.0-amd64
ghcr.io/talos-systems/install-cni:v0.1.0-6-gdc32cc8
k8s.gcr.io/coredns:1.7.0
gcr.io/etcd-development/etcd:v3.4.14
k8s.gcr.io/kube-apiserver-amd64:v1.20.1
k8s.gcr.io/kube-controller-manager-amd64:v1.20.1
k8s.gcr.io/kube-scheduler-amd64:v1.20.1
k8s.gcr.io/kube-proxy-amd64:v1.20.1
ghcr.io/talos-systems/kubelet:v1.20.1
ghcr.io/talos-systems/pod-checkpointer:v0.1.0-6-gdc32cc8
ghcr.io/talos-systems/installer:v0.8.4

v0.9.0-alpha.1 (2021-02-18)

Chore

• fix import path for fsnotify
• update artifacts bucket name in Drone
• enable virtio-balloon and monitor in QEMU provisioner
• rework Drone pipelines
• update dependencies via dependabot
• bump dependencies (via dependabot)
• add dependabot config
• update protobuf, grpc-go, prototool
• update upgrade test version used
• ci: fix schedules in Drone pipelines
• ci: update gcp templates

Docs

• update AMI list for 0.8.2
• fix typos
• update components.md
• add v0.9 docs
• add modes to validate command
• document omitting DiskPartition size
• update references to 0.8.0, add 0.8.0 AWS AMIs
• fix latest docs
• set latest docs to v0.8
• provide AMIs for 0.8.0-beta.0
• fix SBC docs to point to beta.0 instead of beta.1
• update Talos release for SBCs

Feat

• add a tool and package to convert self-hosted CP to static pods
• implement ephemeral partition encryption
• add resource watch API + CLI
• rename apply-config --no-reboot to --on-reboot
• skip filesystem for state and ephemeral partitions in the installer
• stop all pods before unmounting ephemeral partition
• bump Go to 1.15.8
• support version contract for Talos config generation
• update Linux to 5.10.14
• add an option to force upgrade without checks
• upgrade CoreDNS to 1.8.0
• implement IPv6 DHCP client in networkd
• move to ECDSA keys for all Kubernetes/etcd certs and keys
• update kernel
• mount hugetlbfs
• allow fqdn to be used when registering k8s node
• copy cryptsetup executable from pkgs
• use multi-arch images for k8s and Flannel CNI
• replace bootkube with Talos-managed control plane
• implement resource API in Talos
• update Linux to 5.10.7, musl-libc to 1.2.2
• update Kubernetes to 1.20.2
• support Wireguard networking
• bump pkgs for kernel with CONFIG_IPV6_MULTIPLE_TABLES
• support type filter in list API and CLI
• add commands to manage/query etcd cluster
• support disk image in talosctl cluster create
• update Kubernetes to 1.20.1

Fix

• correctly unwrap responses for etcd commands
• drop cri dependency on etcd
• move versions to annotations in control plane static pods
• find master node IPs correctly in health checks
• add 3 seconds grub boot timeout
• don't use filename from URL when downloading manifest
• pass attributes when adding routes
• correct response structure for GenerateConfig API
• correctly extract wrapped error messages
• prevent crash in machined on apid service stop
• wait for time sync before generating Kubernetes certificates
• set proper hostname on docker nodes
• mount kubelet secrets from system instead of ephemeral
• allow loading of empty config files
• prefer configured nameservers, fix DHCP6 in container
• refresh control plane endpoints on worker apids on schedule
• update DHCP client to use Request-Ack sequence after an Offer
• use hugetlbfs instead of none
• use grpc load-balancing when connecting to trustd
• lower memory usage a bit by disabling memory profiling
• don't probe disks in container mode
• prefix rendered Talos-owned static pod manifests
• bump timeout for worker apid waiting for kubelet client config
• kill all processes and umount all disk on reboot/shutdown
• open blockdevices with exclusive flock for partitioning
• list command unlimited recursion default behavior
• pick first interface valid hostname (vs. last one)
• allow 'console' argument in kernel args to be always overridden
• bring up bonded interfaces correctly on packet
• checkpoint controller-manager and scheduler
• correctly transport gRPC errors from apid
• use SetAll instead of AppendAll when building kernel args
• add more dependencies for bootstrap services
• pass disk image flags to e2e-qemu cluster create command
• ignore pods spun up from checkpoints in health checks
• leave etcd for staged upgrades
• ignore errors on stopping/removing pod sandboxes
• use the correct console on Banana Pi M64
• don't run LabelNodeAsMaster in two sequences

Refactor

• extract go-cmd into a separate library
• update go-blockdevice and restructure disk interaction code
• define default kernel flags in machinery instead of procfs

Release

• v0.9.0-alpha.0: prepare release
• v0.9.0-alpha.1: prepare release

Test

• trigger e2e on thrice daily
• update aws templates
• add support for IPv6 in talosctl cluster create
• clear connection refused errors after reset
• skip etcd tests on non-HA clusters

Images

quay.io/coreos/flannel:v0.13.0
ghcr.io/talos-systems/install-cni:v0.1.0-9-g302cc61
docker.io/coredns/coredns:1.8.0
gcr.io/etcd-development/etcd:v3.4.14
k8s.gcr.io/kube-apiserver:v1.20.2
k8s.gcr.io/kube-controller-manager:v1.20.2
k8s.gcr.io/kube-scheduler:v1.20.2
k8s.gcr.io/kube-proxy:v1.20.2
ghcr.io/talos-systems/kubelet:v1.20.2
ghcr.io/talos-systems/installer:v0.9.0-alpha.1
k8s.gcr.io/pause:3.2

v0.8.3 (2021-02-15)

Fix

• pass attributes when adding routes
• correctly extract wrapped error messages
• update DHCP client to use Request-Ack sequence after an Offer
• checkpoint controller-manager and scheduler

Release

• v0.8.3: prepare release

Images

quay.io/coreos/flannel:v0.12.0-amd64
ghcr.io/talos-systems/install-cni:v0.1.0-6-gdc32cc8
k8s.gcr.io/coredns:1.7.0
gcr.io/etcd-development/etcd:v3.4.14
k8s.gcr.io/kube-apiserver-amd64:v1.20.1
k8s.gcr.io/kube-controller-manager-amd64:v1.20.1
k8s.gcr.io/kube-scheduler-amd64:v1.20.1
k8s.gcr.io/kube-proxy-amd64:v1.20.1
ghcr.io/talos-systems/kubelet:v1.20.1
ghcr.io/talos-systems/pod-checkpointer:v0.1.0-6-gdc32cc8
ghcr.io/talos-systems/installer:v0.8.3

v0.8.2 (2021-02-10)

Feat

• allow fqdn to be used when registering k8s node

Fix

• kill all processes and umount all disk on reboot/shutdown
• open blockdevices with exclusive flock for partitioning

Release

• v0.8.0: prepare release
• v0.8.0-beta.0: prepare release
• v0.8.1: prepare release
• v0.8.2: prepare release

Images

quay.io/coreos/flannel:v0.12.0-amd64
ghcr.io/talos-systems/install-cni:v0.1.0-6-gdc32cc8
k8s.gcr.io/coredns:1.7.0
gcr.io/etcd-development/etcd:v3.4.14
k8s.gcr.io/kube-apiserver-amd64:v1.20.1
k8s.gcr.io/kube-controller-manager-amd64:v1.20.1
k8s.gcr.io/kube-scheduler-amd64:v1.20.1
k8s.gcr.io/kube-proxy-amd64:v1.20.1
ghcr.io/talos-systems/kubelet:v1.20.1
ghcr.io/talos-systems/pod-checkpointer:v0.1.0-6-gdc32cc8
ghcr.io/talos-systems/installer:v0.8.2

v0.9.0-alpha.0 (2021-02-02)

Chore

• bump dependencies (via dependabot)
• fix import path for fsnotify
• add dependabot config
• enable virtio-balloon and monitor in QEMU provisioner
• update protobuf, grpc-go, prototool
• update upgrade test version used

Docs

• update components.md
• add v0.9 docs
• add modes to validate command
• document omitting DiskPartition size
• update references to 0.8.0, add 0.8.0 AWS AMIs
• fix latest docs
• set latest docs to v0.8
• provide AMIs for 0.8.0-beta.0
• fix SBC docs to point to beta.0 instead of beta.1
• update Talos release for SBCs

Feat

• move to ECDSA keys for all Kubernetes/etcd certs and keys
• update kernel
• mount hugetlbfs
• allow fqdn to be used when registering k8s node
• copy cryptsetup executable from pkgs
• use multi-arch images for k8s and Flannel CNI
• replace bootkube with Talos-managed control plane
• implement resource API in Talos
• update Linux to 5.10.7, musl-libc to 1.2.2
• update Kubernetes to 1.20.2
• support Wireguard networking
• bump pkgs for kernel with CONFIG_IPV6_MULTIPLE_TABLES
• support type filter in list API and CLI
• add commands to manage/query etcd cluster
• support disk image in talosctl cluster create
• update Kubernetes to 1.20.1

Fix

• use hugetlbfs instead of none
• use grpc load-balancing when connecting to trustd
• lower memory usage a bit by disabling memory profiling
• don't probe disks in container mode
• prefix rendered Talos-owned static pod manifests
• bump timeout for worker apid waiting for kubelet client config
• kill all processes and umount all disk on reboot/shutdown
• open blockdevices with exclusive flock for partitioning
• list command unlimited recursion default behavior
• pick first interface valid hostname (vs. last one)
• allow 'console' argument in kernel args to be always overridden
• bring up bonded interfaces correctly on packet
• checkpoint controller-manager and scheduler
• correctly transport gRPC errors from apid
• use SetAll instead of AppendAll when building kernel args
• add more dependencies for bootstrap services
• pass disk image flags to e2e-qemu cluster create command
• ignore pods spun up from checkpoints in health checks
• leave etcd for staged upgrades
• ignore errors on stopping/removing pod sandboxes
• use the correct console on Banana Pi M64
• don't run LabelNodeAsMaster in two sequences

Refactor

• update go-blockdevice and restructure disk interaction code
• define default kernel flags in machinery instead of procfs

Release

• v0.9.0-alpha.0: prepare release

Test

• clear connection refused errors after reset
• skip etcd tests on non-HA clusters

Images

quay.io/coreos/flannel:v0.13.0
ghcr.io/talos-systems/install-cni:v0.1.0-6-gdc32cc8
k8s.gcr.io/coredns:1.7.0
gcr.io/etcd-development/etcd:v3.4.14
k8s.gcr.io/kube-apiserver:v1.20.2
k8s.gcr.io/kube-controller-manager:v1.20.2
k8s.gcr.io/kube-scheduler:v1.20.2
k8s.gcr.io/kube-proxy:v1.20.2
ghcr.io/talos-systems/kubelet:v1.20.2
ghcr.io/talos-systems/installer:v0.9.0-alpha.0
k8s.gcr.io/pause:3.2

v0.8.1 (2021-01-12)

Fix

• bring up bonded interfaces correctly on packet
• allow 'console' argument in kernel args to be always overridden
• pick first interface valid hostname (vs. last one)

Release

• v0.8.1: prepare release

Images

quay.io/coreos/flannel:v0.12.0-amd64
ghcr.io/talos-systems/install-cni:v0.1.0-6-gdc32cc8
k8s.gcr.io/coredns:1.7.0
gcr.io/etcd-development/etcd:v3.4.14
k8s.gcr.io/kube-apiserver-amd64:v1.20.1
k8s.gcr.io/kube-controller-manager-amd64:v1.20.1
k8s.gcr.io/kube-scheduler-amd64:v1.20.1
k8s.gcr.io/kube-proxy-amd64:v1.20.1
ghcr.io/talos-systems/kubelet:v1.20.1
ghcr.io/talos-systems/pod-checkpointer:v0.1.0-6-gdc32cc8
ghcr.io/talos-systems/installer:v0.8.1

v0.8.0 (2020-12-23)

Chore

• lower MTU to 1450 for the tests in the CI
• build ISOs earlier to launch e2e-iso as soon as possible
• add drone pipeline to upload cloud images
• bump npm ini package for security vulnerability
• update CONTRIBUTING.md
• limit unit-test run concurrency
• bump Go to 1.15.6
• bump dockerfile frontend version
• fix conform for releases
• publish Rock64 image
• enable thrice daily pipeline
• run integration test thrice daily
• output SBC images as compressed raw images
• build SBC images
• update module dependencies
• drop support for docker load
• fix metal image name
• use IMAGE_TAG instead of TAG for :latest pushes
• add cloud image uploader (AWS AMIs for now)
• bump K8s to 1.19.4 in e2e scripts with CABPT version
• build arm64 images in CI
• remove maintenance service interface and use machine service
• bump version of x/net module
• bump Go to 1.15.4
• enable gci linter
• enable nlreturn linter
• fix markdownlint command
• fix markdown-lint

Docs

• add fallback to default page description if none is set on current page
• add a note for being careful about enabling debug flag
• update Equinix Metal guide
• add architectural doc on the root file system layout
• add a note on caveats in container mode
• add storage doc
• add guide for custom CAs
• add docs for network connectivity
• improve SBC documentation
• fix typos
• add openstack docs
• ensure port for vbox and proxmox docs
• add console kernel arg to rpi_4 image generation
• add console kernel arg to libretech_all_h3_cc_h5 image generation
• provide list of AMIs on AWS documentation page
• add 0.8 docs for the upcoming release
• ensure we configure nodes in guides
• ensure gcp docs have firewall and node info
• add qemu diagram and video walkthrough
• graduate v0.7 docs
• improve configuration reference documentation
• fix small typo in talosctl processes cast
• update asciinemas with talosctl
• add proxmox doc
• add live walkthroughs where applicable
• add virtualbox deployment guide
• add high-level orientation docs for talosctl
• add a description of endpoints and nodes
• describe talos upgrade

Feat

• bump pkgs for kernel with HZ=250 on amd64
• bump Linux kernel to 5.10.1, add CONFIG_USB_ACM
• bump pkgs for kernel with CONFIG_USB_XHCI_PLATFORM
• update kernel to 5.9.13, new KSPP requirements
• reset with system disk wipe spec
• add talosctl merge config command
• add talosctl config contexts
• update Kubernetes to 1.20.0
• implement "staged" (failsafe/backup) upgrades
• allow disabling NoSchedule taint on masters using TUI installer
• add support for the Pine64 Rock64
• add TUI for configuring network interfaces settings
• make GenerateConfiguration accept current time as a parameter
• introduce configpatcher package in machinery
• suggest fixed control plane endpoints in talosctl gen config
• update kubernetes to 1.20.0-rc.0
• allow boards to set kernel args
• add support for the Banana Pi M64
• stop including K8s version by default in talosctl gen config
• add support for the Raspberry Pi 4 Model B
• implement network interfaces list API
• bump package for kernel with CIFS support
• upgrade etcd to 3.4.14
• update Containerd and Linux
• add support for installing to SBCs
• add ability to choose CNI config
• support openstack platform
• update Kubernetes to v1.20.0-beta.2
• change UI component for disks selector
• support cluster expansion in the interactive installer
• implement apply configuration without reboot
• make GenerateConfiguration API reuse current node auth
• sync time before installer runs
• set interface MTU in DHCP mode even if DHCP is not successful
• print hint about using interative installer in mainenance mode
• add TUI based talos interactive installer
• support ipv6 routes
• return client config as the second value in GenerateConfiguration
• correctly merge talosconfig (don't ever overwrite)
• drop to maintenance mode in cloud platforms if userdata is missing
• read config from extra guestinfo key (vmware)
• update Go to 1.15.5
• add generate config gRPC API
• upgrade Kubernetes default version to 1.19.4
• add example command in maintenance, enforce cert fingerprint
• add storage API

Fix

• backport fixes from 0.9 after 0.8-beta.0
• synchronize bootkube timeouts and various boot timeouts
• sync RTC in timed, sync time before fetching packet metadata
• don't overwrite PMBR
• bump blockdevice library for 2nd partitione entries copy fix
• properly define shorthand in talosctl time command
• take the first interface from the bond (packet)
• disable kmsg throttling for iso mode
• remove kmsg ratelimiting on startup
• zero out partitions without filesystems on install
• make interactive installer work without endpoints provided
• make default generate image arch dynamic based on arch
• stabilize serial console on RPi4, add video console
• make reset work again
• node taint doesn't contain value anymore
• defer resolving config context in client code
• remove value (change to empty) for NoSchedule taint
• prevent endless loop with DHCP requests in networkd
• skip board argument to the installer if it's not set
• use the dtb from kernel pkg for libretech_all_h3_cc_h5
• remove 'token creds' from maintenance service
• prevent crash in talosctl config commands
• update generated .ova manifest for raw disk size
• bump blockdevice library for mmcblk part name fix
• ignore 'not found' errors when stopping/removing CRI pods
• return hostname from packet platform
• make fingerprint clearly optional in a boot hint
• ensure packet nics get all IPs
• use ghcr.io/talos-systems/kubelet
• bump timeout for config downloading on bare metal
• update packages
• security: update Containerd to v1.4.3

Refactor

• remove setup goroutine in etcd service
• drop osd compatibility layer

Release

• v0.8.0: prepare release
• v0.8.0-alpha.0: prepare release
• v0.8.0-alpha.1: prepare release
• v0.8.0-alpha.2: prepare release
• v0.8.0-alpha.3: prepare release
• v0.8.0-beta.0: prepare release

Test

• add an extra 'node boot done' health check
• remove provision tests with Cilium CNI
• stabilize upgrade test by running health check several times
• add ISO test
• add support for mounting ISO in talosctl cluster create
• bump Talos release version for upgrade test to 0.7.1
• bump defaults for provision tests resources
• update integration test versions, clean up names

Images

quay.io/coreos/flannel:v0.12.0-amd64
ghcr.io/talos-systems/install-cni:v0.1.0-6-gdc32cc8
k8s.gcr.io/coredns:1.7.0
gcr.io/etcd-development/etcd:v3.4.14
k8s.gcr.io/kube-apiserver-amd64:v1.20.1
k8s.gcr.io/kube-controller-manager-amd64:v1.20.1
k8s.gcr.io/kube-scheduler-amd64:v1.20.1
k8s.gcr.io/kube-proxy-amd64:v1.20.1
ghcr.io/talos-systems/kubelet:v1.20.1
ghcr.io/talos-systems/pod-checkpointer:v0.1.0-6-gdc32cc8
ghcr.io/talos-systems/installer:v0.8.0

v0.8.0-beta.0 (2020-12-18)

Chore

• lower MTU to 1450 for the tests in the CI
• build ISOs earlier to launch e2e-iso as soon as possible
• add drone pipeline to upload cloud images
• bump npm ini package for security vulnerability

Docs

• add fallback to default page description if none is set on current page
• add a note for being careful about enabling debug flag

Feat

• bump pkgs for kernel with HZ=250 on amd64
• bump Linux kernel to 5.10.1, add CONFIG_USB_ACM
• bump pkgs for kernel with CONFIG_USB_XHCI_PLATFORM

Fix

• synchronize bootkube timeouts and various boot timeouts
• sync RTC in timed, sync time before fetching packet metadata
• don't overwrite PMBR
• bump blockdevice library for 2nd partitione entries copy fix
• properly define shorthand in talosctl time command
• take the first interface from the bond (packet)
• disable kmsg throttling for iso mode

Refactor

• remove setup goroutine in etcd service

Release

• v0.8.0-beta.0: prepare release

Test

• add an extra 'node boot done' health check
• remove provision tests with Cilium CNI
• stabilize upgrade test by running health check several times

Images

quay.io/coreos/flannel:v0.12.0-amd64
ghcr.io/talos-systems/install-cni:v0.1.0-6-gdc32cc8
k8s.gcr.io/coredns:1.7.0
gcr.io/etcd-development/etcd:v3.4.14
k8s.gcr.io/kube-apiserver-amd64:v1.20.0
k8s.gcr.io/kube-controller-manager-amd64:v1.20.0
k8s.gcr.io/kube-scheduler-amd64:v1.20.0
k8s.gcr.io/kube-proxy-amd64:v1.20.0
ghcr.io/talos-systems/kubelet:v1.20.0
ghcr.io/talos-systems/pod-checkpointer:v0.1.0-6-gdc32cc8
ghcr.io/talos-systems/installer:v0.8.0-beta.0

v0.8.0-alpha.3 (2020-12-11)

Chore

• update CONTRIBUTING.md
• limit unit-test run concurrency
• bump Go to 1.15.6
• bump dockerfile frontend version
• fix conform for releases

Docs

• update Equinix Metal guide
• add architectural doc on the root file system layout
• add a note on caveats in container mode
• add storage doc
• add guide for custom CAs
• add docs for network connectivity
• improve SBC documentation

Feat

• update kernel to 5.9.13, new KSPP requirements
• reset with system disk wipe spec
• add talosctl merge config command
• add talosctl config contexts
• update Kubernetes to 1.20.0
• implement "staged" (failsafe/backup) upgrades
• allow disabling NoSchedule taint on masters using TUI installer

Fix

• remove kmsg ratelimiting on startup
• zero out partitions without filesystems on install
• make interactive installer work without endpoints provided

Release

• v0.8.0-alpha.3: prepare release

Test

• add ISO test
• add support for mounting ISO in talosctl cluster create
• bump Talos release version for upgrade test to 0.7.1
• bump defaults for provision tests resources

Images

quay.io/coreos/flannel:v0.12.0-amd64
ghcr.io/talos-systems/install-cni:v0.1.0-6-gdc32cc8
k8s.gcr.io/coredns:1.7.0
gcr.io/etcd-development/etcd:v3.4.14
k8s.gcr.io/kube-apiserver-amd64:v1.20.0
k8s.gcr.io/kube-controller-manager-amd64:v1.20.0
k8s.gcr.io/kube-scheduler-amd64:v1.20.0
k8s.gcr.io/kube-proxy-amd64:v1.20.0
ghcr.io/talos-systems/kubelet:v1.20.0
ghcr.io/talos-systems/pod-checkpointer:v0.1.0-6-gdc32cc8
ghcr.io/talos-systems/installer:v0.8.0-alpha.3

v0.8.0-alpha.2 (2020-12-04)

Chore

• publish Rock64 image
• enable thrice daily pipeline
• run integration test thrice daily
• output SBC images as compressed raw images
• build SBC images
• update module dependencies
• drop support for docker load
• fix metal image name
• use IMAGE_TAG instead of TAG for :latest pushes
• add cloud image uploader (AWS AMIs for now)
• bump K8s to 1.19.4 in e2e scripts with CABPT version
• build arm64 images in CI
• remove maintenance service interface and use machine service
• bump version of x/net module
• bump Go to 1.15.4
• enable gci linter
• enable nlreturn linter
• fix markdownlint command
• fix markdown-lint

Docs

• fix typos
• add openstack docs
• ensure port for vbox and proxmox docs
• add console kernel arg to rpi_4 image generation
• add console kernel arg to libretech_all_h3_cc_h5 image generation
• provide list of AMIs on AWS documentation page
• add 0.8 docs for the upcoming release
• ensure we configure nodes in guides
• ensure gcp docs have firewall and node info
• add qemu diagram and video walkthrough
• graduate v0.7 docs
• improve configuration reference documentation
• fix small typo in talosctl processes cast
• update asciinemas with talosctl
• add proxmox doc
• add live walkthroughs where applicable
• add virtualbox deployment guide
• add high-level orientation docs for talosctl
• add a description of endpoints and nodes
• describe talos upgrade

Feat

• add support for the Pine64 Rock64
• add TUI for configuring network interfaces settings
• make GenerateConfiguration accept current time as a parameter
• introduce configpatcher package in machinery
• suggest fixed control plane endpoints in talosctl gen config
• update kubernetes to 1.20.0-rc.0
• allow boards to set kernel args
• add support for the Banana Pi M64
• stop including K8s version by default in talosctl gen config
• add support for the Raspberry Pi 4 Model B
• implement network interfaces list API
• bump package for kernel with CIFS support
• upgrade etcd to 3.4.14
• update Containerd and Linux
• add support for installing to SBCs
• add ability to choose CNI config
• support openstack platform
• update Kubernetes to v1.20.0-beta.2
• change UI component for disks selector
• support cluster expansion in the interactive installer
• implement apply configuration without reboot
• make GenerateConfiguration API reuse current node auth
• sync time before installer runs
• set interface MTU in DHCP mode even if DHCP is not successful
• print hint about using interative installer in mainenance mode
• add TUI based talos interactive installer
• support ipv6 routes
• return client config as the second value in GenerateConfiguration
• correctly merge talosconfig (don't ever overwrite)
• drop to maintenance mode in cloud platforms if userdata is missing
• read config from extra guestinfo key (vmware)
• update Go to 1.15.5
• add generate config gRPC API
• upgrade Kubernetes default version to 1.19.4
• add example command in maintenance, enforce cert fingerprint
• add storage API

Fix

• make default generate image arch dynamic based on arch
• stabilize serial console on RPi4, add video console
• make reset work again
• node taint doesn't contain value anymore
• defer resolving config context in client code
• remove value (change to empty) for NoSchedule taint
• prevent endless loop with DHCP requests in networkd
• skip board argument to the installer if it's not set
• use the dtb from kernel pkg for libretech_all_h3_cc_h5
• remove 'token creds' from maintenance service
• prevent crash in talosctl config commands
• update generated .ova manifest for raw disk size
• bump blockdevice library for mmcblk part name fix
• ignore 'not found' errors when stopping/removing CRI pods
• return hostname from packet platform
• make fingerprint clearly optional in a boot hint
• ensure packet nics get all IPs
• use ghcr.io/talos-systems/kubelet
• bump timeout for config downloading on bare metal
• update packages
• security: update Containerd to v1.4.3

Refactor

• drop osd compatibility layer

Release

• v0.8.0-alpha.0: prepare release
• v0.8.0-alpha.1: prepare release
• v0.8.0-alpha.2: prepare release

Test

• update integration test versions, clean up names

Breaking change

in pkg/provision: now NodeRequest.Type should be set
to the node type (as config can be missing now).

In talosctl cluster create add a flag to skip providing config to the
nodes so that they enter maintenance mode, while the generated configs
are written down to disk (so they can be tweaked and applied easily).

Images

quay.io/coreos/flannel:v0.12.0-amd64
ghcr.io/talos-systems/install-cni:v0.1.0-5-gcc2df81
k8s.gcr.io/coredns:1.7.0
gcr.io/etcd-development/etcd:v3.4.14
k8s.gcr.io/kube-apiserver-amd64:v1.20.0-rc.0
k8s.gcr.io/kube-controller-manager-amd64:v1.20.0-rc.0
k8s.gcr.io/kube-scheduler-amd64:v1.20.0-rc.0
k8s.gcr.io/kube-proxy-amd64:v1.20.0-rc.0
ghcr.io/talos-systems/kubelet:v1.20.0-rc.0
ghcr.io/talos-systems/pod-checkpointer:v0.1.0-5-gcc2df81
ghcr.io/talos-systems/installer:v0.8.0-alpha.2

v0.7.1 (2020-12-01)

Fix

• prevent endless loop with DHCP requests in networkd
• ignore 'not found' errors when stopping/removing CRI pods
• security: update Containerd to v1.4.3

Release

• v0.7.1: prepare release

Images

quay.io/coreos/flannel:v0.12.0-amd64
ghcr.io/talos-systems/install-cni:v0.1.0-5-gcc2df81
k8s.gcr.io/coredns:1.7.0
gcr.io/etcd-development/etcd:v3.4.12
k8s.gcr.io/kube-apiserver-amd64:v1.19.4
k8s.gcr.io/kube-controller-manager-amd64:v1.19.4
k8s.gcr.io/kube-scheduler-amd64:v1.19.4
k8s.gcr.io/kube-proxy-amd64:v1.19.4
ghcr.io/talos-systems/kubelet:v1.19.4
ghcr.io/talos-systems/pod-checkpointer:v0.1.0-5-gcc2df81
ghcr.io/talos-systems/installer:v0.7.1

v0.8.0-alpha.1 (2020-11-26)

Chore

• add cloud image uploader (AWS AMIs for now)
• bump K8s to 1.19.4 in e2e scripts with CABPT version
• build arm64 images in CI
• remove maintenance service interface and use machine service
• bump version of x/net module
• bump Go to 1.15.4
• enable gci linter
• enable nlreturn linter
• fix markdownlint command
• fix markdown-lint

Docs

• provide list of AMIs on AWS documentation page
• add 0.8 docs for the upcoming release
• ensure we configure nodes in guides
• ensure gcp docs have firewall and node info
• add qemu diagram and video walkthrough
• graduate v0.7 docs
• improve configuration reference documentation
• fix small typo in talosctl processes cast
• update asciinemas with talosctl
• add proxmox doc
• add live walkthroughs where applicable
• add virtualbox deployment guide
• add high-level orientation docs for talosctl
• add a description of endpoints and nodes
• describe talos upgrade

Feat

• support openstack platform
• update Kubernetes to v1.20.0-beta.2
• change UI component for disks selector
• support cluster expansion in the interactive installer
• implement apply configuration without reboot
• make GenerateConfiguration API reuse current node auth
• sync time before installer runs
• set interface MTU in DHCP mode even if DHCP is not successful
• print hint about using interative installer in mainenance mode
• add TUI based talos interactive installer
• support ipv6 routes
• return client config as the second value in GenerateConfiguration
• correctly merge talosconfig (don't ever overwrite)
• drop to maintenance mode in cloud platforms if userdata is missing
• read config from extra guestinfo key (vmware)
• update Go to 1.15.5
• add generate config gRPC API
• upgrade Kubernetes default version to 1.19.4
• add example command in maintenance, enforce cert fingerprint
• add storage API

Fix

• bump blockdevice library for mmcblk part name fix
• ignore 'not found' errors when stopping/removing CRI pods
• return hostname from packet platform
• make fingerprint clearly optional in a boot hint
• ensure packet nics get all IPs
• use ghcr.io/talos-systems/kubelet
• bump timeout for config downloading on bare metal
• remove 'token creds' from maintenance service
• update packages

Refactor

• drop osd compatibility layer

Release

• v0.8.0-alpha.0: prepare release
• v0.8.0-alpha.1: prepare release

Test

• update integration test versions, clean up names

Breaking change

in pkg/provision: now NodeRequest.Type should be set
to the node type (as config can be missing now).

In talosctl cluster create add a flag to skip providing config to the
nodes so that they enter maintenance mode, while the generated configs
are written down to disk (so they can be tweaked and applied easily).

Images

quay.io/coreos/flannel:v0.12.0-amd64
ghcr.io/talos-systems/install-cni:v0.1.0-5-gcc2df81
k8s.gcr.io/coredns:1.7.0
gcr.io/etcd-development/etcd:v3.4.12
k8s.gcr.io/kube-apiserver-amd64:v1.20.0-beta.2
k8s.gcr.io/kube-controller-manager-amd64:v1.20.0-beta.2
k8s.gcr.io/kube-scheduler-amd64:v1.20.0-beta.2
k8s.gcr.io/kube-proxy-amd64:v1.20.0-beta.2
ghcr.io/talos-systems/kubelet:v1.20.0-beta.2
ghcr.io/talos-systems/pod-checkpointer:v0.1.0-5-gcc2df81
ghcr.io/talos-systems/installer:v0.8.0-alpha.1

v0.7.0 (2020-11-16)

Chore

• update golangci-lint
• remove duplicate packages
• remove unused binaries
• output more logs from the installer
• update CI scripts
• move to newer release of rtnetlink with fn args
• reduce numer of steps/parallelism of Drone build
• fix the check-dirty command to abort on untracked files
• bump module dependencies in go.mod
• bump Go to 1.15.3
• add Context as param to some methods of Platform interface
• bump pkgs version
• publish list of images to release notes
• attempt to fix image pushing for GitHub
• update qemu hack script to use ISO
• fix 'push' targets
• fix edge push
• fix docker login
• fix docker login
• migrate to ghcr.io
• push edge releases on successful nightly integration
• update ntp time headers
• upgrade Go to 1.15.1
• remove extra COPY from rootfs
• update k8s modules to 1.19 final version
• upgrade Go to 1.14.8
• drop vmlinux from assets
• add a method to merge Talos client config
• bump next version to v0.6.0-beta.2
• update machinery version in go.mod
• update node.js dependencies
• re-import talos-systems/pkg/crypto/tls
• extract pkg/crypto as external module
• integrate importvet
• update capi CI manifests to use control planes
• update node dependencies
• update packages
• bump module dependencies in go.mod
• bump Go to 1.15.3
• add Context as param to some methods of Platform interface
• bump pkgs version
• publish list of images to release notes
• attempt to fix image pushing for GitHub
• update qemu hack script to use ISO
• fix 'push' targets
• fix edge push
• fix docker login
• fix docker login
• migrate to ghcr.io
• push edge releases on successful nightly integration
• update ntp time headers
• upgrade Go to 1.15.1
• remove extra COPY from rootfs
• update k8s modules to 1.19 final version
• upgrade Go to 1.14.8
• drop vmlinux from assets
• add a method to merge Talos client config
• bump next version to v0.6.0-beta.2
• update machinery version in go.mod
• update node.js dependencies
• re-import talos-systems/pkg/crypto/tls
• extract pkg/crypto as external module
• integrate importvet
• update capi CI manifests to use control planes
• update node dependencies
• update packages
• update k8s modules to 1.19 final version
• upgrade Go to 1.14.8
• drop vmlinux from assets
• add a method to merge Talos client config
• bump next version to v0.6.0-beta.2
• update machinery version in go.mod
• update node.js dependencies
• re-import talos-systems/pkg/crypto/tls
• extract pkg/crypto as external module
• integrate importvet
• update capi CI manifests to use control planes
• update node dependencies
• update packages
• re-import talos-systems/pkg/crypto/tls
• extract pkg/crypto as external module
• integrate importvet
• update capi CI manifests to use control planes
• update node dependencies
• update packages

Docs

• fix AWS guides
• address small nits
• update config reference docs
• add redirect for /docs/latest
• fix small CSS issues
• use grid instead of flexbox
• improve the config reference documentation
• improve search bar
• address small nits
• add robots.txt and fix sitemap.xml
• fix config reference types links
• move to gridsome
• link container images to our repository
• fix latest tag
• add link to latest docs
• small fixes for the config docs and air-gapped
• add guide on setting up air-gapped environment with images
• add note on settings endpoints on MacOS
• remove second meeting from README
• fix cluster name in docker docs
• add note around link-local addressing
• add ghcr.io to the registry cache docs
• add v0.7 docs
• add recommneded settings in overview
• update upgrade guide with talosctl upgrade-k8s
• update 0.6 links
• graduate v0.6 docs
• add Kubernetes upgrade guide
• add reset doc
• add QEMU provisioner documentation
• fix download link
• link container images to our repository
• fix latest tag
• add link to latest docs
• small fixes for the config docs and air-gapped
• add guide on setting up air-gapped environment with images
• add note on settings endpoints on MacOS
• remove second meeting from README
• fix cluster name in docker docs
• add note around link-local addressing
• add ghcr.io to the registry cache docs
• add v0.7 docs
• add recommneded settings in overview
• update upgrade guide with talosctl upgrade-k8s
• update 0.6 links
• graduate v0.6 docs
• add Kubernetes upgrade guide
• add reset doc
• add QEMU provisioner documentation
• fix download link
• graduate v0.6 docs
• add Kubernetes upgrade guide
• add reset doc
• add QEMU provisioner documentation
• fix download link
• fix download link

Feat

• read config from extra guestinfo key (vmware)
• bump Go 1.15.5, arm64 images in the CI
• upgrade kernel to v5.9.3
• upgrade packages
• add ISO support
• add webconfig service
• build talosctl-cni-bundle, use it in talosctl for QEMU
• skip resizing ephemeral partition if not required
• allow specifying user-disks in talosctl cluster create
• bump CoreDNS to 1.7.0
• bump Linux to 5.8.16, enable mpt3sas driver
• bump CoreDNS to 1.7.0
• encode comments as part of talosctl generated configs
• extend etcd health check on upgrade
• wipe disks faster in the installer
• upgrade Kubernetes to 1.19.3
• support MTU and route changes for DHCP
• bump packages for Linux 5.8.15 and containerd 1.4.1
• support metric values for DHCP
• bump packages version for the kernel with BBR TCP congestion algo
• handle unsupported commands being called for docker
• support disk usage command in talosctl
• bring in install-cni & pod-checkpointer from extras packages
• implement talos.shutdown=[halt|poweroff] kernel argument
• add etcd API
• allow disabling NoSchedule on master nodes
• colorize output of cluster health checks
• pull kubeconfig from the cluster on successful cluster create
• use kubeconfig merge in talosctl kubeconfig by default
• support --registry-insecure-skip-verify for cluster create
• show cluster state when talosctl cluster create finishes
• support custom filename for talosctl kubeconfig
• add support for disabling time
• add ApplyConfiguration API
• validate cluster DNS name
• build Talos images/artifacts for amd64/arm64
• bump default resource limits for talosctl cluster create
• add default install image
• add images command
• ugrade Linux kernel to 5.8.10
• allow for link local networking
• use architecture-specific image for core k8s components
• update Flannel to 0.12, support for arm64
• upgrade kubernetes to 1.19.1
• implement command talosctl upgrade-k8s
• use latest packages
• upgrade runc to v1.0.0-rc92
• upgrade containerd to v1.4.0
• remove ISO support
• add grub bootloader
• upgrade etcd to 3.4.12
• provide option to run Talos under UEFI in QEMU
• update linux to 5.8.5
• update kubernetes to v1.19.0
• make boostrap via API default choice in talosctl cluster create
• upgrade Linux to v5.7.15
• upgrade etcd to 3.4.10
• add persist flag to gen config
• upgrade Kubernetes to 1.19.3
• support MTU and route changes for DHCP
• bump packages for Linux 5.8.15 and containerd 1.4.1
• support metric values for DHCP
• bump packages version for the kernel with BBR TCP congestion algo
• handle unsupported commands being called for docker
• support disk usage command in talosctl
• bring in install-cni & pod-checkpointer from extras packages
• implement talos.shutdown=[halt|poweroff] kernel argument
• add etcd API
• allow disabling NoSchedule on master nodes
• colorize output of cluster health checks
• pull kubeconfig from the cluster on successful cluster create
• use kubeconfig merge in talosctl kubeconfig by default
• support --registry-insecure-skip-verify for cluster create
• show cluster state when talosctl cluster create finishes
• support custom filename for talosctl kubeconfig
• add support for disabling time
• add ApplyConfiguration API
• validate cluster DNS name
• build Talos images/artifacts for amd64/arm64
• bump default resource limits for talosctl cluster create
• add default install image
• add images command
• ugrade Linux kernel to 5.8.10
• allow for link local networking
• use architecture-specific image for core k8s components
• update Flannel to 0.12, support for arm64
• upgrade kubernetes to 1.19.1
• implement command talosctl upgrade-k8s
• use latest packages
• upgrade runc to v1.0.0-rc92
• upgrade containerd to v1.4.0
• remove ISO support
• add grub bootloader
• upgrade etcd to 3.4.12
• provide option to run Talos under UEFI in QEMU
• update linux to 5.8.5
• update kubernetes to v1.19.0
• make boostrap via API default choice in talosctl cluster create
• upgrade Linux to v5.7.15
• upgrade etcd to 3.4.10
• add persist flag to gen config
• add grub bootloader
• upgrade etcd to 3.4.12
• provide option to run Talos under UEFI in QEMU
• update linux to 5.8.5
• update kubernetes to v1.19.0
• make boostrap via API default choice in talosctl cluster create
• upgrade Linux to v5.7.15
• upgrade etcd to 3.4.10
• add persist flag to gen config
• upgrade etcd to 3.4.10
• add persist flag to gen config

Fix

• backport k8s 1.19.4, ghcr.io kubelet image
• r8169 driver, Go 1.15.4, maintenance service API
• remove log.Fatal from maintenance service
• address issues in webconfig
• prevent blind mode boot
• read/write human readable representations for bytes and octals
• bump type for DiskSize to be 64-bit
• properly initialize manifest in user disks creation
• remove default time server in time command
• retry connection refused errors while bootstrapping a cluster
• re-implement upgrade (install) with preserve
• revert "feat: bump CoreDNS to 1.7.0"
• stop CRI pods on upgrade with preserve
• stop etcd on any path on upgrade
• ignore transient errors in upgrade Kubernetes code
• stop ignoring EINVAL on mount
• implement preserving contents of partition on install
• correctly calculate output width in colored health reporter
• update handling of ntp disable
• address nil pointer panic
• improve logging and errors for extra manifests by URL
• random failures in cluster health checks
• apply --removable option always to get standard UEFI filename
• nil pointer panic in talosctl dashboard
• make CLI context exit immediately on second ^C
• registry auth config building
• provide unique username in generate kubeconfig
• make Flannel CNI image follow $PKGS version
• retry container image import
• update one more places which had stale reference for constants
• update the docs to fix the lint-markdown
• use images package in integration tests
• move installer image variables out of machinery
• enable --removable options for GRUB
• retry image pulling, stop on 404, no duplicate pulls
• address node package update
• validate cluster endpoint
• improve error message on empty config
• gracefully handle invalid interfaces in bond
• set environment variable for etcd on arm64
• don't enforce k8s version in talosctl cluster create by default
• tell grub to use console output
• update vmware image and platform
• don't abort reboot sequence on bootloader meta failure
• default endpoint to 127.0.0.1 for Docker/OS X
• remove udevd debug flag
• update permissions for directories and files created via extraFiles
• allow static pod files
• change apid container image name to expected value
• add syslinux to create ISO
• pass config via stdin
• handle bootkube recover correctly, support recovery from etcd
• run health check for etcd service with Get API
• ignore eth0 interface in docker provisioner
• update e2e scripts to work with python3
• retry non-HTTP errors from API server
• update qemu launcher on arm64 to boot Talos properly
• update handling of ntp disable
• address nil pointer panic
• improve logging and errors for extra manifests by URL
• random failures in cluster health checks
• apply --removable option always to get standard UEFI filename
• nil pointer panic in talosctl dashboard
• make CLI context exit immediately on second ^C
• registry auth config building
• provide unique username in generate kubeconfig
• make Flannel CNI image follow $PKGS version
• retry container image import
• update one more places which had stale reference for constants
• update the docs to fix the lint-markdown
• use images package in integration tests
• move installer image variables out of machinery
• enable --removable options for GRUB
• retry image pulling, stop on 404, no duplicate pulls
• address node package update
• validate cluster endpoint
• improve error message on empty config
• gracefully handle invalid interfaces in bond
• set environment variable for etcd on arm64
• don't enforce k8s version in talosctl cluster create by default
• tell grub to use console output
• update vmware image and platform
• don't abort reboot sequence on bootloader meta failure
• default endpoint to 127.0.0.1 for Docker/OS X
• remove udevd debug flag
• update permissions for directories and files created via extraFiles
• allow static pod files
• change apid container image name to expected value
• add syslinux to create ISO
• pass config via stdin
• handle bootkube recover correctly, support recovery from etcd
• run health check for etcd service with Get API
• ignore eth0 interface in docker provisioner
• update e2e scripts to work with python3
• retry non-HTTP errors from API server
• update qemu launcher on arm64 to boot Talos properly
• change apid container image name to expected value
• add syslinux to create ISO
• pass config via stdin
• handle bootkube recover correctly, support recovery from etcd
• run health check for etcd service with Get API
• ignore eth0 interface in docker provisioner
• update e2e scripts to work with python3
• retry non-HTTP errors from API server
• update qemu launcher on arm64 to boot Talos properly
• run health check for etcd service with Get API
• ignore eth0 interface in docker provisioner
• update e2e scripts to work with python3
• retry non-HTTP errors from API server
• update qemu launcher on arm64 to boot Talos properly

Refactor

• use gRPC for interactive installation
• bring more control to install.Manifest execution
• extract blockdevice library
• garbage collect unused constants
• deduplicate packages version in Dockerfile
• move udevadm trigger/settle to udevd healthcheck
• extract packages loadbalancer and retry
• extract cluster bootstrapper via API as common component
• move external API packages into machinery/
• rework pkg/grpc/tls to break dependency on pkg/grpc/gen
• extract pkg/net as github.com/talos-systems/net
• expose provision as public package
• remove structs from config provider
• bring more control to install.Manifest execution
• extract blockdevice library
• garbage collect unused constants
• deduplicate packages version in Dockerfile
• move udevadm trigger/settle to udevd healthcheck
• extract packages loadbalancer and retry
• extract cluster bootstrapper via API as common component
• move external API packages into machinery/
• rework pkg/grpc/tls to break dependency on pkg/grpc/gen
• extract pkg/net as github.com/talos-systems/net
• expose provision as public package
• remove structs from config provider
• move udevadm trigger/settle to udevd healthcheck
• extract packages loadbalancer and retry
• extract cluster bootstrapper via API as common component
• move external API packages into machinery/
• rework pkg/grpc/tls to break dependency on pkg/grpc/gen
• extract pkg/net as github.com/talos-systems/net
• expose provision as public package
• remove structs from config provider
• move external API packages into machinery/
• rework pkg/grpc/tls to break dependency on pkg/grpc/gen
• extract pkg/net as github.com/talos-systems/net
• expose provision as public package
• remove structs from config provider

Test

• bump Talos version for upgrade tests, bump Cilium version
• clean up integration test code, fix flakes
• add unit-test for the installer manifest
• potential fix for talosctl cluster destroy being stuck
• implement API for QEMU VM provisioner
• re-enable Cilium e2e upgrade test
• verify kubernetes control plane upgrade in provision tests
• add e2e test to the provision (upgrade) tests
• determine reboots using boot id
• add support for PXE nodes in qemu provision library
• potential fix for talosctl cluster destroy being stuck
• implement API for QEMU VM provisioner
• re-enable Cilium e2e upgrade test
• verify kubernetes control plane upgrade in provision tests
• add e2e test to the provision (upgrade) tests
• determine reboots using boot id
• add support for PXE nodes in qemu provision library
• determine reboots using boot id
• add support for PXE nodes in qemu provision library

BREAKING CHANGE

in pkg/provision: now NodeRequest.Type should be set
to the node type (as config can be missing now).

In talosctl cluster create add a flag to skip providing config to the
nodes so that they enter maintenance mode, while the generated configs
are written down to disk (so they can be tweaked and applied easily).

BREAKING CHANGE

Single node upgrades will fail in this change. This
will also break the A/B fallback setup since this version introduces
an entirely new partition scheme, that any fallback will not know about.
We plan on addressing these issues in a follow up change.

BREAKING CHANGE

Single node upgrades will fail in this change. This
will also break the A/B fallback setup since this version introduces
an entirely new partition scheme, that any fallback will not know about.
We plan on addressing these issues in a follow up change.

BREAKING CHANGE

Single node upgrades will fail in this change. This
will also break the A/B fallback setup since this version introduces
an entirely new partition scheme, that any fallback will not know about.
We plan on addressing these issues in a follow up change.

Images

quay.io/coreos/flannel:v0.12.0-amd64
ghcr.io/talos-systems/install-cni:v0.1.0-5-gcc2df81
k8s.gcr.io/coredns:1.7.0
gcr.io/etcd-development/etcd:v3.4.12
k8s.gcr.io/kube-apiserver-amd64:v1.19.4
k8s.gcr.io/kube-controller-manager-amd64:v1.19.4
k8s.gcr.io/kube-scheduler-amd64:v1.19.4
k8s.gcr.io/kube-proxy-amd64:v1.19.4
ghcr.io/talos-systems/kubelet:v1.19.4
ghcr.io/talos-systems/pod-checkpointer:v0.1.0-5-gcc2df81
ghcr.io/talos-systems/installer:v0.7.0

v0.7.0-beta.1 (2020-11-11)

Fix

• r8169 driver, Go 1.15.4, maintenance service API

Release

• v0.7.0-beta.0: prepare release
• v0.7.0-beta.1: prepare release

Breaking change

in pkg/provision: now NodeRequest.Type should be set
to the node type (as config can be missing now).

In talosctl cluster create add a flag to skip providing config to the
nodes so that they enter maintenance mode, while the generated configs
are written down to disk (so they can be tweaked and applied easily).

Images

quay.io/coreos/flannel:v0.12.0-amd64
ghcr.io/talos-systems/install-cni:v0.1.0-2-g709d580
k8s.gcr.io/coredns:1.7.0
gcr.io/etcd-development/etcd:v3.4.12
k8s.gcr.io/kube-apiserver-amd64:v1.19.3
k8s.gcr.io/kube-controller-manager-amd64:v1.19.3
k8s.gcr.io/kube-scheduler-amd64:v1.19.3
k8s.gcr.io/kube-proxy-amd64:v1.19.3
docker.io/autonomy/kubelet:v1.19.3
ghcr.io/talos-systems/pod-checkpointer:v0.1.0-2-g709d580
ghcr.io/talos-systems/installer:v0.7.0-beta.1