Networking plugin for pod networking in Kubernetes using Elastic Network Interfaces on AWS.
Download the latest version of the yaml and apply it to the cluster.
kubectl apply -f aws-k8s-cni.yaml
Launch kubelet with network plugins set to cni (--network-plugin=cni), the cni directories configured (--cni-config-dir
and --cni-bin-dir) and node ip set to the primary IPv4 address of the primary ENI for the instance
(--node-ip=$(curl http://169.254.169.254/latest/meta-data/local-ipv4)).
It is also recommended that you set --max-pods equal to (the number of ENIs for the instance type ×
(the number of IPs per ENI - 1)) + 2; for details, see vpc_ip_resource_limit.go. Setting --max-pods will prevent
scheduling that exceeds the IP address resources available to the kubelet.
The default manifest expects --cni-conf-dir=/etc/cni/net.d and --cni-bin-dir=/opt/cni/bin.
Alternatively there is also a Helm chart: eks/aws-vpc-cni
See here for required IAM policies.
makedefaults tomake build-linuxthat builds the Linux binaries.unit-test,format,lintandvetprovide ways to run the respective tests/tools and should be run before submitting a PR.make dockerwill create a docker container using the docker-build with the finished binaries, with a tag ofamazon/amazon-k8s-cni:latestmake docker-builduses a docker container (golang:1.16) to build the binaries.make docker-unit-testsuses a docker container (golang:1.16) to run all unit tests.
There are 2 components:
- CNI Plugin, which will wire up the host's and pod's network stack when called.
ipamd, a long-running node-Local IP Address Management (IPAM) daemon, is responsible for:- maintaining a warm-pool of available IP addresses, and
- assigning an IP address to a Pod.
The details can be found in Proposal: CNI plugin for Kubernetes networking over AWS VPC.
For help, please consider the following venues (in order):
- Amazon EKS Best Practices Guide for Networking
- Troubleshooting Guide provides tips on how to debug and troubleshoot this CNI.
- Search open issues
- File an issue
- Chat with us on the
#aws-vpc-cnichannel in the Kubernetes Slack community.
For all Kubernetes releases, we recommend installing the latest VPC CNI release. The following table denotes our minimum recommended VPC CNI version for each actively supported Kubernetes release.
| Kubernetes Release | 1.25 | 1.24 | 1.23 | 1.22 |
|---|---|---|---|---|
| VPC CNI Version | v1.11.4+ | v1.9.3+ | v1.8.0+ | v1.8.0+ |
Upgrading (or downgrading) the VPC CNI version should result in no downtime. Existing pods should not be affected and will not lose network connectivity.
New pods will be in pending state until the VPC CNI is fully initialized and can assign pod IP addresses. In v1.12.0+, VPC CNI state is
restored via an on-disk file: /var/run/aws-node/ipam.json. In lower versions, state is restored via calls to container runtime.
When a worker node first joins the cluster, there is only 1 ENI along with all of the addresses on the ENI. Without any configuration, ipamd always tries to keep one extra ENI.
When the number of pods running on the node exceeds the number of addresses on a single ENI, the CNI backend starts allocating a new ENI using the following allocation scheme:
- If the number of current running Pods is between 0 and 29, ipamd will allocate one more eni. And Warm-Pool size is 2 eni * (30 -1) = 58
- If the number of current running Pods is between 30 and 58, ipamd will allocate 2 more eni. And Warm-Pool size is 3 eni * (30 -1) = 87
For example, a m4.4xlarge node can have up to 8 ENIs, and each ENI can have up to 30 IP addresses. See Elastic Network Interfaces documentation for details.
For a detailed explanation, see WARM_ENI_TARGET, WARM_IP_TARGET and MINIMUM_IP_TARGET.
VPC CNI makes use of privileged mode (privileged: true) in the manifest only for its init container. This elevated privilege is required to set the networking kernel parameters.
The Amazon VPC CNI plugin for Kubernetes supports a number of configuration options, which are set through environment variables. The following environment variables are available, and all of them are optional.
Type: Boolean as a String
Default: false
Specifies whether IPAMD should allocate or deallocate ENIs on a non-schedulable node.
Type: Boolean as a String
Default: true
Specifies whether NodePort services are enabled on a worker node's primary network interface. This requires additional
iptables rules, and the kernel's reverse path filter on the primary interface is set to loose.
Type: Boolean as a String
Default: false
Specifies that your pods may use subnets and security groups that are independent of your worker node's VPC configuration.
By default, pods share the same subnet and security groups as the worker node's primary interface. Setting this variable
to true causes ipamd to use the security groups and VPC subnet in a worker node's ENIConfig for elastic network interface
allocation. You must create an ENIConfig custom resource for each subnet that your pods will reside in, and then annotate or
label each worker node to use a specific ENIConfig. Multiple worker nodes can be annotated or labelled with the same ENIConfig, but
each Worker node can be annotated with a single ENIConfig at a time. Further, the subnet in the ENIConfig must belong to the
same Availability Zone that the worker node resides in.
For more information, see CNI Custom Networking
in the Amazon EKS User Guide.
Type: String
Default: k8s.amazonaws.com/eniConfig
Specifies node annotation key name. This should be used when AWS_VPC_K8S_CNI_CUSTOM_NETWORK_CFG=true. Annotation value
will be used to set ENIConfig name. Note that annotations take precedence over labels.
Type: String
Default: k8s.amazonaws.com/eniConfig
Specifies node label key name. This should be used when AWS_VPC_K8S_CNI_CUSTOM_NETWORK_CFG=true. Label value will be used
to set ENIConfig name. Note that annotations will take precedence over labels. To use labels, ensure there is no annotation with key
k8s.amazonaws.com/eniConfig or defined key (in ENI_CONFIG_ANNOTATION_DEF) set on the node.
To select an ENIConfig based upon availability zone set this to topology.kubernetes.io/zone and create an
ENIConfig custom resource for each availability zone (e.g. us-east-1a). Note that tag failure-domain.beta.kubernetes.io/zone is deprecated and replaced with the tag topology.kubernetes.io/zone.
Type: Integer as a String
Default: 9001
Used to configure the MTU size for attached ENIs. The valid range is from 576 to 9001.
Type: Boolean as a String
Default: false
Specifies whether an external NAT gateway should be used to provide SNAT of secondary ENI IP addresses. If set to true, the
SNAT iptables rule and off-VPC IP rule are not applied, and these rules are removed if they have already been applied.
Disable SNAT if you need to allow inbound communication to your pods from external VPNs, direct connections, and external VPCs,
and your pods do not need to access the Internet directly via an Internet Gateway. However, your nodes must be running in a
private subnet and connected to the internet through an AWS NAT Gateway or another external NAT device.
Type: String
Default: prng
Valid Values: hashrandom, prng, none
Specifies whether the SNAT iptables rule should randomize the outgoing ports for connections. This setting takes effect when
AWS_VPC_K8S_CNI_EXTERNALSNAT=false, which is the default setting. The default setting for AWS_VPC_K8S_CNI_RANDOMIZESNAT is
prng, meaning that --random-fully will be added to the SNAT iptables rule. For old versions of iptables that do not
support --random-fully this option will fall back to --random. To disable random port allocation, if you for example
rely on sequential port allocation for outgoing connections set it to none.
Note: Any options other than none will cause outbound connections to be assigned a source port that is not necessarily
part of the ephemeral port range set at the OS level (/proc/sys/net/ipv4/ip_local_port_range). This is relevant for any
customers that might have NACLs restricting traffic based on the port range found in ip_local_port_range.
Type: String
Default: empty
Specify a comma-separated list of IPv4 CIDRs to exclude from SNAT. For every item in the list an iptables rule and off-VPC
IP rule will be applied. If an item is not a valid ipv4 range it will be skipped. This should be used when AWS_VPC_K8S_CNI_EXTERNALSNAT=false.
Type: Integer as a String
Default: 1
Specifies the number of free elastic network interfaces (and all of their available IP addresses) that the ipamd daemon should
attempt to keep available for pod assignment on the node. By default, ipamd attempts to keep 1 elastic network interface and all
of its IP addresses available for pod assignment. The number of IP addresses per network interface varies by instance type. For more
information, see IP Addresses Per Network Interface Per Instance Type
in the Amazon EC2 User Guide for Linux Instances.
For example, an m4.4xlarge launches with 1 network interface and 30 IP addresses. If 5 pods are placed on the node and 5 free IP
addresses are removed from the IP address warm pool, then ipamd attempts to allocate more interfaces until WARM_ENI_TARGET free
interfaces are available on the node.
If WARM_IP_TARGET is set, then this environment variable is ignored and the WARM_IP_TARGET behavior is used instead.
Type: Integer
Default: None
Specifies the number of free IP addresses that the ipamd daemon should attempt to keep available for pod assignment on the node. Setting this to a non-positive value is the same as setting this to 0 or not setting the variable.
With ENABLE_PREFIX_DELEGATION set to true then ipamd daemon will check if the existing (/28) prefixes are enough to maintain the
WARM_IP_TARGET if it is not sufficient then more prefixes will be attached.
For example,
- if
WARM_IP_TARGETis set to 5, thenipamdattempts to keep 5 free IP addresses available at all times. If the elastic network interfaces on the node are unable to provide these free addresses,ipamdattempts to allocate more interfaces untilWARM_IP_TARGETfree IP addresses are available. ENABLE_PREFIX_DELEGATIONset totrueandWARM_IP_TARGETis 16. Initially, 1 (/28) prefix is sufficient but once a single pod is assigned IP then remaining free IPs are 15 hence IPAMD will allocate 1 more prefix to achieve 16WARM_IP_TARGET
NOTE! Avoid this setting for large clusters, or if the cluster has high pod churn. Setting it will cause additional calls to the
EC2 API and that might cause throttling of the requests. It is strongly suggested to set MINIMUM_IP_TARGET when using WARM_IP_TARGET.
If both WARM_IP_TARGET and MINIMUM_IP_TARGET are set, ipamd will attempt to meet both constraints.
This environment variable overrides WARM_ENI_TARGET behavior. For a detailed explanation, see
WARM_ENI_TARGET, WARM_IP_TARGET and MINIMUM_IP_TARGET.
If ENABLE_PREFIX_DELEGATION set to true and WARM_IP_TARGET overrides WARM_PREFIX_TARGET behavior. For a detailed explanation, see
WARM_PREFIX_TARGET, WARM_IP_TARGET and MINIMUM_IP_TARGET.
Type: Integer
Default: None
Specifies the number of total IP addresses that the ipamd daemon should attempt to allocate for pod assignment on the node.
MINIMUM_IP_TARGET behaves identically to WARM_IP_TARGET except that instead of setting a target number of free IP
addresses to keep available at all times, it sets a target number for a floor on how many total IP addresses are allocated. Setting to a
non-positive value is same as setting this to 0 or not setting the variable.
MINIMUM_IP_TARGET is for pre-scaling, WARM_IP_TARGET is for dynamic scaling. For example, suppose a cluster has an
expected pod density of approximately 30 pods per node. If WARM_IP_TARGET is set to 30 to ensure there are enough IPs
allocated up front by the CNI, then 30 pods are deployed to the node, the CNI will allocate an additional 30 IPs, for
a total of 60, accelerating IP exhaustion in the relevant subnets. If instead MINIMUM_IP_TARGET is set to 30 and
WARM_IP_TARGET to 2, after the 30 pods are deployed the CNI would allocate an additional 2 IPs. This still provides
elasticity, but uses roughly half as many IPs as using WARM_IP_TARGET alone (32 IPs vs 60 IPs).
This also improves the reliability of the EKS cluster by reducing the number of calls necessary to allocate or deallocate private IPs, which may be throttled, especially at scaling-related times.
Type: Integer
Default: None
Specifies the maximum number of ENIs that will be attached to the node. When MAX_ENI is unset or 0 (or lower), the setting
is not used, and the maximum number of ENIs is always equal to the maximum number for the instance type in question. Even when
MAX_ENI is a positive number, it is limited by the maximum number for the instance type.
Type: String
Default: DEBUG
Valid Values: DEBUG, INFO, WARN, ERROR, FATAL. (Not case sensitive)
Specifies the loglevel for ipamd and cni-metric-helper.
Type: String
Default: /host/var/log/aws-routed-eni/ipamd.log
Valid Values: stdout or a file path
Specifies where to write the logging output of ipamd. Either to stdout or to override the default file (i.e., /var/log/aws-routed-eni/ipamd.log).
Type: String
Default: /var/log/aws-routed-eni/plugin.log
Valid Values: stderr or a file path
Specifies where to write the logging output for aws-cni plugin. Either to stderr or to override the default file (i.e., /var/log/aws-routed-eni/plugin.log).
Stdout cannot be supported for plugin log, please refer to #1248 for more details.
Note: If chaining an external plugin (i.e Cilium) that does not provide a pluginLogFile in its config file, the CNI plugin will by default write to os.Stderr. The output of cmdAdd are available in the Kubelet logs.
Type: String
Default: DEBUG
Valid Values: DEBUG, INFO, WARN, ERROR, FATAL. (Not case sensitive)
Specifies the loglevel for aws-cni plugin.
Type: String
Default: 127.0.0.1:61679
Specifies the bind address for the introspection endpoint.
A Unix Domain Socket can be specified with the unix: prefix before the socket path.
Type: Boolean as a String
Default: false
Specifies whether introspection endpoints are disabled on a worker node. Setting this to true will reduce the debugging
information we can get from the node when running the aws-cni-support.sh script.
Type: Boolean as a String
Default: false
Specifies whether the prometheus metrics endpoint is disabled or not for ipamd. By default metrics are published
on :61678/metrics.
Type: String
Default: eni
Specifies the veth prefix used to generate the host-side veth device name for the CNI. The prefix can be at most 4 characters long. The prefixes eth, vlan, and lo are reserved by the CNI plugin and cannot be specified. We recommend using prefix name not shared by any other network interfaces on the worker node instance.
Type: String
Default: {}
Example values: {"tag_key": "tag_val"}
Metadata applied to ENI helps you categorize and organize your resources for billing or other purposes. Each tag consists of a custom-defined key and an optional value. Tag keys can have a maximum character length of 128 characters. Tag values can have a maximum length of 256 characters. These tags will be added to all ENIs on the host.
Important: Custom tags should not contain k8s.amazonaws.com prefix as it is reserved. If the tag has k8s.amazonaws.com
string, tag addition will be ignored.
Type: Boolean as a String
Default: true
Specifies whether ipamd should configure rp filter for primary interface. Setting this to false will require rp filter to be configured through init container.
NOTE! AWS_VPC_K8S_CNI_CONFIGURE_RPFILTER has been deprecated, so setting this environment variable results in a no-op. The init container unconditionally configures the rp filter for the primary interface.
Type: String
Default: ""
Specifies the cluster name to tag allocated ENIs with. See the "Cluster Name tag" section below.
Type: String
Default: ""
Specifies the cluster endpoint to use for connecting to the api-server without relying on kube-proxy. This is an optional configuration parameter that can improve the initialization time of the AWS VPC CNI.
Type: Boolean as a String
Default: false
To enable security groups for pods you need to have at least an EKS 1.17 eks.3 cluster.
Setting ENABLE_POD_ENI to true will allow IPAMD to add the vpc.amazonaws.com/has-trunk-attached label to the node if the instance has the capacity to attach an additional ENI.
The label notifies vpc-resource-controller (https://github.com/aws/amazon-vpc-resource-controller-k8s) to attach a Trunk ENI to the instance. The label value is initially set to false and is marked to true by IPAMD when vpc-resource-controller attaches a Trunk ENI to the instance. However, there might be cases where the label value will remain false if the instance doesn't support ENI Trunking.
Once enabled the VPC resource controller will then advertise branch network interfaces as extended resources on these nodes in your cluster. Branch interface capacity is additive to existing instance type limits for secondary IP addresses and prefixes. For example, a c5.4xlarge can continue to have up to 234 secondary IP addresses or 234 /28 prefixes assigned to standard network interfaces and up to 54 branch network interfaces. Each branch network interface only receives a single primary IP address and this IP address will be allocated to pods with a security group(branch ENI pods).
Any of the WARM targets do not impact the scale of the branch ENI pods so you will have to set the WARM_{ENI/IP/PREFIX}_TARGET based on the number of non-branch ENI pods. If you are having the cluster mostly using pods with a security group consider setting WARM_IP_TARGET to a very low value instead of default WARM_ENI_TARGET or WARM_PREFIX_TARGET to reduce wastage of IPs/ENIs.
NOTE! Toggling ENABLE_POD_ENI from true to false will not detach the Trunk ENI from an instance. To delete/detach the Trunk ENI from an instance, you need to recycle the instance.
Type: String
Default: strict
Valid Values: strict, standard
Once ENABLE_POD_ENI is set to true, this value controls how the traffic of pods with the security group behaves.
-
strictmode: all inbound/outbound traffic from pod with security group will be enforced by security group rules. This is the default mode if POD_SECURITY_GROUP_ENFORCING_MODE is not set. -
standardmode: the traffic of pod with security group behaves same as pods without a security group, except that each pod occupies a dedicated branch ENI.- inbound traffic to pod with security group from another host will be enforced by security group rules.
- outbound traffic from pod with security group to another host in the same VPC will be enforced by security group rules.
- inbound/outbound traffic from another pod on the same host or another service on the same host(such as kubelet/nodeLocalDNS) won't be enforced by security group rules.
- outbound traffic from pod with security group to IP address outside VPC
- if externalSNAT enabled, traffic won't be SNATed, thus will be enforced by security group rules.
- if externalSNAT disabled, traffic will be SNATed via eth0, thus will only be enforced by the security group associated with eth0.
NOTE!: To make new behavior be in effect after switching the mode, existing pods with security group must be recycled. Alternatively, you can restart the nodes as well.
Type: Boolean as a String
Default: false
If ENABLE_POD_ENI is set to true, for the kubelet to connect via TCP (for liveness or readiness probes)
to pods that are using per pod security groups, DISABLE_TCP_EARLY_DEMUX should be set to true for amazon-k8s-cni-init
the container under initcontainers. This will increase the local TCP connection latency slightly.
Details on why this is needed can be found in this #1212 comment.
To use this setting, a Linux kernel version of at least 4.6 is needed on the worker node.
You can use the below command to enable DISABLE_TCP_EARLY_DEMUX to true -
kubectl patch daemonset aws-node -n kube-system -p '{"spec": {"template": {"spec": {"initContainers": [{"env":[{"name":"DISABLE_TCP_EARLY_DEMUX","value":"true"}],"name":"aws-vpc-cni-init"}]}}}}'
Type: Boolean as a String
Default: false
To enable prefix delegation on nitro instances. Setting ENABLE_PREFIX_DELEGATION to true will start allocating a prefix (/28 for IPv4
and /80 for IPv6) instead of a secondary IP in the ENIs subnet. The total number of prefixes and private IP addresses will be less than the
limit on private IPs allowed by your instance. Setting or resetting of ENABLE_PREFIX_DELEGATION while pods are running or if ENIs are attached is supported and the new pods allocated will get IPs based on the mode of IPAMD but the max pods of kubelet should be updated which would need either kubelet restart or node recycle.
Setting ENABLE_PREFIX_DELEGATION to true will not increase the density of branch ENI pods. The limit on the number of branch network interfaces per instance type will remain the same - https://docs.aws.amazon.com/eks/latest/userguide/security-groups-for-pods.html#supported-instance-types. Each branch network will be allocated a primary IP and this IP will be allocated for the branch ENI pods.
Please refer to VPC CNI Feature Matrix section below for additional information around using Prefix delegation with Custom Networking and Security Groups Per Pod features.
Note: ENABLE_PREFIX_DELEGATION needs to be set to true when VPC CNI is configured to operate in IPv6 mode (supported in v1.10.0+). Prefix Delegation in IPv4 and IPv6 modes is supported on Nitro based Bare Metal instances as well from v1.11+. If you're using Prefix Delegation feature on Bare Metal instances, downgrading to an earlier version of VPC CNI from v1.11+ will be disruptive and not supported.
Type: Integer
Default: None
Specifies the number of free IPv4(/28) prefixes that the ipamd daemon should attempt to keep available for pod assignment on the node. Setting to a non-positive value is same as setting this to 0 or not setting the variable.
This environment variable works when ENABLE_PREFIX_DELEGATION is set to true and is overridden when WARM_IP_TARGET and MINIMUM_IP_TARGET are configured.
Type: Boolean as a String
Default: false
Setting DISABLE_NETWORK_RESOURCE_PROVISIONING to true will make IPAMD depend only on IMDS to get attached ENIs and IPs/prefixes.
Type: Boolean as a String
Default: false
Setting ENABLE_BANDWIDTH_PLUGIN to true will update 10-aws.conflist to include upstream bandwidth plugin as a chained plugin.
Type: Boolean as a String
Default: false
Setting ANNOTATE_POD_IP to true will allow IPAMD to add an annotation vpc.amazonaws.com/pod-ips to the pod with pod IP.
There is a known issue with kubelet taking time to update Pod.Status.PodIP leading to calico being blocked on programming the policy. Setting ANNOTATE_POD_IP to true will enable AWS VPC CNI plugin to add Pod IP as an annotation to the pod spec to address this race condition.
To annotate the pod with pod IP, you will have to add "patch" permission for pods resource in aws-node clusterrole. You can use the below command -
cat << EOF > append.yaml
- apiGroups:
- ""
resources:
- pods
verbs:
- patch
EOF
kubectl apply -f <(cat <(kubectl get clusterrole aws-node -o yaml) append.yaml)
Type: Boolean as a String
Default: true
VPC CNI can operate in either IPv4 or IPv6 mode. Setting ENABLE_IPv4 to true will configure it in IPv4 mode (default mode).
Note: Dual-stack mode isn't yet supported. So, enabling both IPv4 and IPv6 will be treated as an invalid configuration.
Type: Boolean as a String
Default: false
VPC CNI can operate in either IPv4 or IPv6 mode. Setting ENABLE_IPv6 to true (both under aws-node and aws-vpc-cni-init containers in the manifest)
will configure it in IPv6 mode. IPv6 is only supported in Prefix Delegation mode, so ENABLE_PREFIX_DELEGATION needs to be set to true if VPC CNI is
configured to operate in IPv6 mode. Prefix delegation is only supported on nitro instances.
Note: Please make sure that the required IPv6 IAM policy is applied (Refer to IAM Policy section above). Dual stack mode isn't yet supported. So, enabling both IPv4 and IPv6 will be treated as invalid configuration. Please refer to the VPC CNI Feature Matrix section below for additional information.
Type: Boolean as a String
Default: false
VPC CNI uses iptables-legacy by default. Setting ENABLE_NFTABLES to true will update VPC CNI to use iptables-nft.
Note: VPC CNI image contains iptables-legacy and iptables-nft. Switching between them is done via update-alternatives. It is strongly recommended that the iptables mode matches that which is used by the base OS and kube-proxy.
Switching modes while pods are running or rules are installed will not trigger reconciliation. It is recommended that rules are manually updated or nodes are drained and cordoned before updating. If reloading node, ensure that previous rules are not set to be persisted.
Type: String
Default: empty
Specify a comma-separated list of IPv4 CIDRs that must be routed via main routing table. This is required for secondary ENIs to reach endpoints outside of VPC that are backed by a service.
For every item in the list, an ip rule will be created with a priority greater than the ip rule capturing egress traffic from the container. If an item is not a valid IPv4 CIDR, it will be skipped.
Type: String
Default: empty
Specify the EC2 endpoint to use. This is useful if you are using a custom endpoint for EC2. For example, if you are using a proxy for EC2, you can set this to the proxy endpoint. Any kind of URL or IP address is valid such as https://localhost:8080 or http://ec2.us-west-2.customaws.com. If this is not set, the default EC2 endpoint will be used.
| IP Mode | Secondary IP Mode | Prefix Delegation | Security Groups Per Pod | WARM & MIN IP/Prefix Targets | External SNAT |
|---|---|---|---|---|---|
IPv4 |
Yes | Yes | Yes | Yes | Yes |
IPv6 |
No | Yes | No | No | No |
This plugin interacts with the following tags on ENIs:
cluster.k8s.amazonaws.com/namenode.k8s.amazonaws.com/instance_idnode.k8s.amazonaws.com/no_manage
The tag cluster.k8s.amazonaws.com/name will be set to the cluster name of the
aws-node daemonset which created the ENI.
The tag node.k8s.amazonaws.com/instance_id will be set to the instance ID of
the aws-node instance that allocated this ENI.
The tag node.k8s.amazonaws.com/no_manage is read by the aws-node daemonset to
determine whether an ENI attached to the machine should not be configured or
used for private IPs.
This tag is not set by the cni plugin itself, but rather may be set by a user to indicate that an ENI is intended for host networking pods, or for some other process unrelated to Kubernetes.
Note: Attaching an ENI with the no_manage tag will result in an incorrect
value for the Kubelet's --max-pods configuration option. Consider also
updating the MAX_ENI and --max-pods configuration options on this plugin
and the kubelet respectively if you are making use of this tag.
For VPC CNI >=v1.12.0, IPAMD have switched to use an on-disk file /var/run/aws-node/ipam.json to track IP allocations, thus became container runtime agnostic and no longer requires access to Container Runtime Interface(CRI) socket.
- Note:
- Helm chart >=v1.2.0 is released with VPC CNI v1.12.0, thus no longer supports the
cri.hostPath.path. If you need to install a VPC CNI <v1.12.0 with helm chart, a Helm chart version that <v1.2.0 should be used.
- Helm chart >=v1.2.0 is released with VPC CNI v1.12.0, thus no longer supports the
For VPC CNI <v1.12.0, IPAMD still depends on CRI to track IP allocations using pod sandboxes information upon its starting.
- By default the dockershim CRI socket was mounted but can be customized to use other CRI:
- The mountPath should be changed to
/var/run/cri.sockand hostPath should be pointed to CRI used by kubelet, such as/var/run/containerd/containerd.sockfor containerd. - With Helm chart <v1.2.0, the flag
--set cri.hostPath.path=/var/run/containerd/containerd.sockcan set above for you.
- The mountPath should be changed to
- Note:
- When using a different container runtime instead of the default dockershim in VPC CNI, make sure kubelet is also configured to use the same CRI.
- If you want to enable containerd runtime with the support provided by Amazon AMI, please follow the instructions in our documentation, Enable the containerd runtime bootstrap flag
L-IPAMD(aws-node daemonSet) running on every worker node requires access to the Kubernetes API server. If it can not reach
the Kubernetes API server, ipamd will exit and CNI will not be able to get any IP address for Pods. Here is a way to confirm if
aws-node has access to the Kubernetes API server.
# find out Kubernetes service IP, e.g. 10.0.0.1
kubectl get svc Kubernetes
NAME TYPE CLUSTER-IP EXTERNAL-IP PORT(S) AGE
kubernetes ClusterIP 10.0.0.1 <none> 443/TCP 29d
# ssh into worker node, check if worker node can reach API server
telnet 10.0.0.1 443
Trying 10.0.0.1...
Connected to 10.0.0.1.
Escape character is '^]'. <-------- Kubernetes API server is reachable
If you think you’ve found a potential security issue, please do not post it in the Issues. Instead, please follow the instructions here or email AWS security directly.