typhoon/bare-metal/container-linux/kubernetes/cl/controller.yaml

221 lines
9.7 KiB
YAML
Raw Normal View History

---
systemd:
units:
- name: etcd-member.service
enable: true
dropins:
- name: 40-etcd-cluster.conf
contents: |
[Service]
Environment="ETCD_IMAGE_TAG=v3.4.7"
Environment="ETCD_IMAGE_URL=docker://quay.io/coreos/etcd"
Environment="RKT_RUN_ARGS=--insecure-options=image"
Environment="ETCD_NAME=${etcd_name}"
Environment="ETCD_ADVERTISE_CLIENT_URLS=https://${domain_name}:2379"
Environment="ETCD_INITIAL_ADVERTISE_PEER_URLS=https://${domain_name}:2380"
Environment="ETCD_LISTEN_CLIENT_URLS=https://0.0.0.0:2379"
Environment="ETCD_LISTEN_PEER_URLS=https://0.0.0.0:2380"
Environment="ETCD_LISTEN_METRICS_URLS=http://0.0.0.0:2381"
Environment="ETCD_INITIAL_CLUSTER=${etcd_initial_cluster}"
Environment="ETCD_STRICT_RECONFIG_CHECK=true"
Environment="ETCD_SSL_DIR=/etc/ssl/etcd"
Environment="ETCD_TRUSTED_CA_FILE=/etc/ssl/certs/etcd/server-ca.crt"
Environment="ETCD_CERT_FILE=/etc/ssl/certs/etcd/server.crt"
Environment="ETCD_KEY_FILE=/etc/ssl/certs/etcd/server.key"
Environment="ETCD_CLIENT_CERT_AUTH=true"
Environment="ETCD_PEER_TRUSTED_CA_FILE=/etc/ssl/certs/etcd/peer-ca.crt"
Environment="ETCD_PEER_CERT_FILE=/etc/ssl/certs/etcd/peer.crt"
Environment="ETCD_PEER_KEY_FILE=/etc/ssl/certs/etcd/peer.key"
Environment="ETCD_PEER_CLIENT_CERT_AUTH=true"
- name: docker.service
enable: true
- name: locksmithd.service
mask: true
- name: kubelet.path
enable: true
contents: |
[Unit]
Description=Watch for kubeconfig
[Path]
PathExists=/etc/kubernetes/kubeconfig
[Install]
WantedBy=multi-user.target
- name: wait-for-dns.service
enable: true
contents: |
[Unit]
Description=Wait for DNS entries
Wants=systemd-resolved.service
Before=kubelet.service
[Service]
Type=oneshot
RemainAfterExit=true
ExecStart=/bin/sh -c 'while ! /usr/bin/grep '^[^#[:space:]]' /etc/resolv.conf > /dev/null; do sleep 1; done'
[Install]
RequiredBy=kubelet.service
RequiredBy=etcd-member.service
- name: kubelet.service
contents: |
[Unit]
Description=Kubelet via Hyperkube
Wants=rpc-statd.service
[Service]
Environment=KUBELET_CGROUP_DRIVER=${cgroup_driver}
ExecStartPre=/bin/mkdir -p /etc/kubernetes/cni/net.d
ExecStartPre=/bin/mkdir -p /etc/kubernetes/manifests
ExecStartPre=/bin/mkdir -p /opt/cni/bin
ExecStartPre=/bin/mkdir -p /var/lib/calico
ExecStartPre=/bin/mkdir -p /var/lib/kubelet/volumeplugins
ExecStartPre=/usr/bin/bash -c "grep 'certificate-authority-data' /etc/kubernetes/kubeconfig | awk '{print $2}' | base64 -d > /etc/kubernetes/ca.crt"
ExecStartPre=-/usr/bin/rkt rm --uuid-file=/var/cache/kubelet-pod.uuid
Inline Container Linux kubelet.service, deprecate kubelet-wrapper * Change kubelet.service on Container Linux nodes to ExecStart Kubelet inline to replace the use of the host OS kubelet-wrapper script * Express rkt run flags and volume mounts in a clear, uniform way to make the Kubelet service easier to audit, manage, and understand * Eliminate reliance on a Container Linux kubelet-wrapper script * Typhoon for Fedora CoreOS developed a kubelet.service that similarly uses an inline ExecStart (except with podman instead of rkt) and a more minimal set of volume mounts. Adopt the volume improvements: * Change Kubelet /etc/kubernetes volume to read-only * Change Kubelet /etc/resolv.conf volume to read-only * Remove unneeded /var/lib/cni volume mount Background: * kubelet-wrapper was added in CoreOS around the time of Kubernetes v1.0 to simplify running a CoreOS-built hyperkube ACI image via rkt-fly. The script defaults are no longer ideal (e.g. rkt's notion of trust dates back to quay.io ACI image serving and signing, which informed the OCI standard images we use today, though they still lack rkt's signing ideas). * Shipping kubelet-wrapper was regretted at CoreOS, but remains in the distro for compatibility. The script is not updated to track hyperkube changes, but it is stable and kubelet.env overrides bridge most gaps * Typhoon Container Linux nodes have used kubelet-wrapper to rkt/rkt-fly run the Kubelet via the official k8s.gcr.io hyperkube image using overrides (new image registry, new image format, restart handling, new mounts, new entrypoint in v1.17). * Observation: Most of what it takes to run a Kubelet container is defined in Typhoon, not in kubelet-wrapper. The wrapper's value is now undermined by having to workaround its dated defaults. Typhoon may be better served defining Kubelet.service explicitly * Typhoon for Fedora CoreOS developed a kubelet.service without the use of a host OS kubelet-wrapper which is both clearer and eliminated some volume mounts
2019-12-29 20:17:26 +01:00
ExecStart=/usr/bin/rkt run \
--uuid-file-save=/var/cache/kubelet-pod.uuid \
--stage1-from-dir=stage1-fly.aci \
--hosts-entry host \
--insecure-options=image \
--volume etc-kubernetes,kind=host,source=/etc/kubernetes,readOnly=true \
--mount volume=etc-kubernetes,target=/etc/kubernetes \
--volume etc-machine-id,kind=host,source=/etc/machine-id,readOnly=true \
--mount volume=etc-machine-id,target=/etc/machine-id \
--volume etc-os-release,kind=host,source=/usr/lib/os-release,readOnly=true \
--mount volume=etc-os-release,target=/etc/os-release \
--volume=etc-resolv,kind=host,source=/etc/resolv.conf,readOnly=true \
--mount volume=etc-resolv,target=/etc/resolv.conf \
--volume etc-ssl-certs,kind=host,source=/etc/ssl/certs,readOnly=true \
--mount volume=etc-ssl-certs,target=/etc/ssl/certs \
--volume lib-modules,kind=host,source=/lib/modules,readOnly=true \
--mount volume=lib-modules,target=/lib/modules \
--volume run,kind=host,source=/run \
--mount volume=run,target=/run \
--volume usr-share-certs,kind=host,source=/usr/share/ca-certificates,readOnly=true \
--mount volume=usr-share-certs,target=/usr/share/ca-certificates \
--volume var-lib-calico,kind=host,source=/var/lib/calico,readOnly=true \
Inline Container Linux kubelet.service, deprecate kubelet-wrapper * Change kubelet.service on Container Linux nodes to ExecStart Kubelet inline to replace the use of the host OS kubelet-wrapper script * Express rkt run flags and volume mounts in a clear, uniform way to make the Kubelet service easier to audit, manage, and understand * Eliminate reliance on a Container Linux kubelet-wrapper script * Typhoon for Fedora CoreOS developed a kubelet.service that similarly uses an inline ExecStart (except with podman instead of rkt) and a more minimal set of volume mounts. Adopt the volume improvements: * Change Kubelet /etc/kubernetes volume to read-only * Change Kubelet /etc/resolv.conf volume to read-only * Remove unneeded /var/lib/cni volume mount Background: * kubelet-wrapper was added in CoreOS around the time of Kubernetes v1.0 to simplify running a CoreOS-built hyperkube ACI image via rkt-fly. The script defaults are no longer ideal (e.g. rkt's notion of trust dates back to quay.io ACI image serving and signing, which informed the OCI standard images we use today, though they still lack rkt's signing ideas). * Shipping kubelet-wrapper was regretted at CoreOS, but remains in the distro for compatibility. The script is not updated to track hyperkube changes, but it is stable and kubelet.env overrides bridge most gaps * Typhoon Container Linux nodes have used kubelet-wrapper to rkt/rkt-fly run the Kubelet via the official k8s.gcr.io hyperkube image using overrides (new image registry, new image format, restart handling, new mounts, new entrypoint in v1.17). * Observation: Most of what it takes to run a Kubelet container is defined in Typhoon, not in kubelet-wrapper. The wrapper's value is now undermined by having to workaround its dated defaults. Typhoon may be better served defining Kubelet.service explicitly * Typhoon for Fedora CoreOS developed a kubelet.service without the use of a host OS kubelet-wrapper which is both clearer and eliminated some volume mounts
2019-12-29 20:17:26 +01:00
--mount volume=var-lib-calico,target=/var/lib/calico \
--volume var-lib-docker,kind=host,source=/var/lib/docker \
--mount volume=var-lib-docker,target=/var/lib/docker \
--volume var-lib-kubelet,kind=host,source=/var/lib/kubelet,recursive=true \
--mount volume=var-lib-kubelet,target=/var/lib/kubelet \
--volume var-log,kind=host,source=/var/log \
--mount volume=var-log,target=/var/log \
--volume opt-cni-bin,kind=host,source=/opt/cni/bin \
--mount volume=opt-cni-bin,target=/opt/cni/bin \
--volume etc-iscsi,kind=host,source=/etc/iscsi \
--mount volume=etc-iscsi,target=/etc/iscsi \
--volume usr-sbin-iscsiadm,kind=host,source=/usr/sbin/iscsiadm \
--mount volume=usr-sbin-iscsiadm,target=/sbin/iscsiadm \
docker://quay.io/poseidon/kubelet:v1.18.1 -- \
--anonymous-auth=false \
--authentication-token-webhook \
--authorization-mode=Webhook \
--cgroup-driver=$${KUBELET_CGROUP_DRIVER} \
--client-ca-file=/etc/kubernetes/ca.crt \
--cluster_dns=${cluster_dns_service_ip} \
--cluster_domain=${cluster_domain_suffix} \
--cni-conf-dir=/etc/kubernetes/cni/net.d \
--exit-on-lock-contention \
--healthz-port=0 \
--hostname-override=${domain_name} \
--kubeconfig=/etc/kubernetes/kubeconfig \
--lock-file=/var/run/lock/kubelet.lock \
--network-plugin=cni \
--node-labels=node.kubernetes.io/master \
--node-labels=node.kubernetes.io/controller="true" \
--pod-manifest-path=/etc/kubernetes/manifests \
--read-only-port=0 \
--register-with-taints=node-role.kubernetes.io/master=:NoSchedule \
--volume-plugin-dir=/var/lib/kubelet/volumeplugins
ExecStop=-/usr/bin/rkt stop --uuid-file=/var/cache/kubelet-pod.uuid
Restart=always
RestartSec=10
[Install]
WantedBy=multi-user.target
- name: bootstrap.service
contents: |
[Unit]
Description=Kubernetes control plane
ConditionPathExists=!/opt/bootstrap/bootstrap.done
[Service]
Type=oneshot
RemainAfterExit=true
WorkingDirectory=/opt/bootstrap
ExecStart=/usr/bin/rkt run \
--trust-keys-from-https \
Introduce cluster creation without local writes to asset_dir * Allow generated assets (TLS materials, manifests) to be securely distributed to controller node(s) via file provisioner (i.e. ssh-agent) as an assets bundle file, rather than relying on assets being locally rendered to disk in an asset_dir and then securely distributed * Change `asset_dir` from required to optional. Left unset, asset_dir defaults to "" and no assets will be written to files on the machine that runs terraform apply * Enhancement: Managed cluster assets are kept only in Terraform state, which supports different backends (GCS, S3, etcd, etc) and optional encryption. terraform apply accesses state, runs in-memory, and distributes sensitive materials to controllers without making use of local disk (simplifies use in CI systems) * Enhancement: Improve asset unpack and layout process to position etcd certificates and control plane certificates more cleanly, without unneeded secret materials Details: * Terraform file provisioner support for distributing directories of contents (with unknown structure) has been limited to reading from a local directory, meaning local writes to asset_dir were required. https://github.com/poseidon/typhoon/issues/585 discusses the problem and newer or upcoming Terraform features that might help. * Observation: Terraform provisioner support for single files works well, but iteration isn't viable. We're also constrained to Terraform language features on the apply side (no extra plugins, no shelling out) and CoreOS / Fedora tools on the receive side. * Take a map representation of the contents that would have been splayed out in asset_dir and pack/encode them into a single file format devised for easy unpacking. Use an awk one-liner on the receive side to unpack. In pratice, this has worked well and its rather nice that a single assets file is transferred by file provisioner (all or none) Rel: https://github.com/poseidon/terraform-render-bootstrap/pull/162
2019-12-05 07:10:55 +01:00
--volume config,kind=host,source=/etc/kubernetes/bootstrap-secrets \
--mount volume=config,target=/etc/kubernetes/secrets \
--volume assets,kind=host,source=/opt/bootstrap/assets \
--mount volume=assets,target=/assets \
--volume script,kind=host,source=/opt/bootstrap/apply \
--mount volume=script,target=/apply \
--insecure-options=image \
docker://quay.io/poseidon/kubelet:v1.18.1 \
--net=host \
--dns=host \
--exec=/apply
ExecStartPost=/bin/touch /opt/bootstrap/bootstrap.done
[Install]
WantedBy=multi-user.target
storage:
directories:
- path: /etc/kubernetes
filesystem: root
files:
- path: /etc/hostname
filesystem: root
mode: 0644
contents:
inline:
${domain_name}
Introduce cluster creation without local writes to asset_dir * Allow generated assets (TLS materials, manifests) to be securely distributed to controller node(s) via file provisioner (i.e. ssh-agent) as an assets bundle file, rather than relying on assets being locally rendered to disk in an asset_dir and then securely distributed * Change `asset_dir` from required to optional. Left unset, asset_dir defaults to "" and no assets will be written to files on the machine that runs terraform apply * Enhancement: Managed cluster assets are kept only in Terraform state, which supports different backends (GCS, S3, etcd, etc) and optional encryption. terraform apply accesses state, runs in-memory, and distributes sensitive materials to controllers without making use of local disk (simplifies use in CI systems) * Enhancement: Improve asset unpack and layout process to position etcd certificates and control plane certificates more cleanly, without unneeded secret materials Details: * Terraform file provisioner support for distributing directories of contents (with unknown structure) has been limited to reading from a local directory, meaning local writes to asset_dir were required. https://github.com/poseidon/typhoon/issues/585 discusses the problem and newer or upcoming Terraform features that might help. * Observation: Terraform provisioner support for single files works well, but iteration isn't viable. We're also constrained to Terraform language features on the apply side (no extra plugins, no shelling out) and CoreOS / Fedora tools on the receive side. * Take a map representation of the contents that would have been splayed out in asset_dir and pack/encode them into a single file format devised for easy unpacking. Use an awk one-liner on the receive side to unpack. In pratice, this has worked well and its rather nice that a single assets file is transferred by file provisioner (all or none) Rel: https://github.com/poseidon/terraform-render-bootstrap/pull/162
2019-12-05 07:10:55 +01:00
- path: /opt/bootstrap/layout
filesystem: root
mode: 0544
contents:
inline: |
#!/bin/bash -e
mkdir -p -- auth tls/etcd tls/k8s static-manifests manifests/coredns manifests-networking
awk '/#####/ {filename=$2; next} {print > filename}' assets
mkdir -p /etc/ssl/etcd/etcd
mkdir -p /etc/kubernetes/bootstrap-secrets
mv tls/etcd/{peer*,server*} /etc/ssl/etcd/etcd/
mv tls/etcd/etcd-client* /etc/kubernetes/bootstrap-secrets/
chown -R etcd:etcd /etc/ssl/etcd
chmod -R 500 /etc/ssl/etcd
mv auth/kubeconfig /etc/kubernetes/bootstrap-secrets/
mv tls/k8s/* /etc/kubernetes/bootstrap-secrets/
sudo mkdir -p /etc/kubernetes/manifests
sudo mv static-manifests/* /etc/kubernetes/manifests/
sudo mkdir -p /opt/bootstrap/assets
sudo mv manifests /opt/bootstrap/assets/manifests
sudo mkdir -p /opt/bootstrap/assets/manifests/crds
sudo mv manifests-networking/{crd,cluster}*.yaml /opt/bootstrap/assets/manifests/crds 2>/dev/null || true
sudo mv manifests-networking/* /opt/bootstrap/assets/manifests/
rm -rf assets auth static-manifests tls manifests-networking
- path: /opt/bootstrap/apply
filesystem: root
mode: 0544
contents:
inline: |
#!/bin/bash -e
Introduce cluster creation without local writes to asset_dir * Allow generated assets (TLS materials, manifests) to be securely distributed to controller node(s) via file provisioner (i.e. ssh-agent) as an assets bundle file, rather than relying on assets being locally rendered to disk in an asset_dir and then securely distributed * Change `asset_dir` from required to optional. Left unset, asset_dir defaults to "" and no assets will be written to files on the machine that runs terraform apply * Enhancement: Managed cluster assets are kept only in Terraform state, which supports different backends (GCS, S3, etcd, etc) and optional encryption. terraform apply accesses state, runs in-memory, and distributes sensitive materials to controllers without making use of local disk (simplifies use in CI systems) * Enhancement: Improve asset unpack and layout process to position etcd certificates and control plane certificates more cleanly, without unneeded secret materials Details: * Terraform file provisioner support for distributing directories of contents (with unknown structure) has been limited to reading from a local directory, meaning local writes to asset_dir were required. https://github.com/poseidon/typhoon/issues/585 discusses the problem and newer or upcoming Terraform features that might help. * Observation: Terraform provisioner support for single files works well, but iteration isn't viable. We're also constrained to Terraform language features on the apply side (no extra plugins, no shelling out) and CoreOS / Fedora tools on the receive side. * Take a map representation of the contents that would have been splayed out in asset_dir and pack/encode them into a single file format devised for easy unpacking. Use an awk one-liner on the receive side to unpack. In pratice, this has worked well and its rather nice that a single assets file is transferred by file provisioner (all or none) Rel: https://github.com/poseidon/terraform-render-bootstrap/pull/162
2019-12-05 07:10:55 +01:00
export KUBECONFIG=/etc/kubernetes/secrets/kubeconfig
until kubectl version; do
echo "Waiting for static pod control plane"
sleep 5
done
until kubectl apply -f /assets/manifests/crds -R; do
echo "Retry Custom Resource Definition manifests"
sleep 5
done
until kubectl apply -f /assets/manifests -R; do
echo "Retry applying manifests"
sleep 5
done
- path: /etc/sysctl.d/max-user-watches.conf
filesystem: root
contents:
inline: |
fs.inotify.max_user_watches=16184
passwd:
users:
- name: core
ssh_authorized_keys:
- ${ssh_authorized_key}