Upgrade Guide¶

Regression warning¶

Upgrading from >=1.4.0 to 1.5.y¶

1. In v1.4, the TCP conntrack table size ct-global-max-entries-tcp ConfigMap parameter was ineffective due to a bug and thus, the default value (1000000) was used instead. To prevent from breaking established TCP connections, bpf-ct-global-tcp-max must be set to 1000000 in the ConfigMap before upgrading. Refer to the section Upgrading a ConfigMap on how to upgrade the ConfigMap.

2. If you previously upgraded to v1.5, downgraded to <v1.5, and now want to upgrade to v1.5 again, then you must run the following DaemonSet before doing the upgrade:

   K8s 1.10

   K8s 1.11

   K8s 1.12

   K8s 1.13

   K8s 1.14

   $ kubectl apply -f https://raw.githubusercontent.com/cilium/cilium/1.5.3/examples/kubernetes/1.10/cilium-pre-flight-with-rm-svc-v2.yaml
   

   $ kubectl apply -f https://raw.githubusercontent.com/cilium/cilium/1.5.3/examples/kubernetes/1.11/cilium-pre-flight-with-rm-svc-v2.yaml
   

   $ kubectl apply -f https://raw.githubusercontent.com/cilium/cilium/1.5.3/examples/kubernetes/1.12/cilium-pre-flight-with-rm-svc-v2.yaml
   

   $ kubectl apply -f https://raw.githubusercontent.com/cilium/cilium/1.5.3/examples/kubernetes/1.13/cilium-pre-flight-with-rm-svc-v2.yaml
   

   $ kubectl apply -f https://raw.githubusercontent.com/cilium/cilium/1.5.3/examples/kubernetes/1.14/cilium-pre-flight-with-rm-svc-v2.yaml
   

   See Running a pre-flight DaemonSet for instructions how to run, validate and remove a pre-flight DaemonSet.

3. Follow the standard procedures to perform the upgrade as described in Upgrading Minor Versions.

New Default Values¶

• The connection-tracking garbage collector interval is now dynamic. It will automatically adjust based n on the percentage of the connection tracking table that has been cleared in the last run. The interval will vary between 10 seconds and 30 minutes or 12 hours for LRU based maps. This should automatically optimize CPU consumption as much as possible while keeping the connection tracking table utilization below 25%. If needed, the interval can be set to a static interval with the option --conntrack-gc-interval. If connectivity fails and cilium monitor --type drop shows xx drop (CT: Map insertion failed), then it is likely that the connection tracking table is filling up and the automatic adjustment of the garbage collector interval is insufficient. Set --conntrack-gc-interval to an interval lower than the default. Alternatively, the value for bpf-ct-global-any-max and bpf-ct-global-tcp-max can be increased. Setting both of these options will be a trade-off of CPU for conntrack-gc-interval, and for bpf-ct-global-any-max and bpf-ct-global-tcp-max the amount of memory consumed.

New ConfigMap Options¶

All options available in the cilium-agent can now be specified in the Cilium ConfigMap without requiring to set an environment variable in the DaemonSet.

• enable-k8s-event-handover: enables use of the kvstore to optimize Kubernetes event handling by listening for k8s events in the operator and mirroring it into the kvstore for reduced overhead in large clusters.
• enable-legacy-services: enables legacy services (prior v1.5) to prevent from terminating established connections to services when upgrading Cilium from < v1.5 to v1.5. If enabled, the option needs to stay enabled until a user is confident that they will not need to downgrade to < v1.5 anymore. Disabling and then enabling the option is not possible without breaking the established connections from Pod to Service.

Deprecated Options¶

• --conntrack-garbage-collector-interval has been deprecated. Please use the option --conntrack-gc-interval which parses a duration as string instead of a integer in seconds. Support for the deprecated option will be removed in 1.6.
• legacy-host-allows-world option is now removed as planned.
• monitor-aggregation-level: Superseded by monitor-aggregation.
• ct-global-max-entries-tcp: Superseded by bpf-ct-global-tcp-max.
• ct-global-max-entries-other: Superseded by bpf-ct-global-any-max.
• prometheus-serve-addr from the cilium-metrics-config ConfigMap is superseded by prometheus-serve-addr from the cilium-config ConfigMap.
• --auto-ipv6-node-routes was removed as planned. Use --auto-direct-node-routes instead.

Upgrading from >=1.2.5 to 1.4.y¶

1. Follow the standard procedures to perform the upgrade as described in Upgrading Minor Versions.

Changes that may require action¶

• The --serve option was removed from cilium-bugtool in favor of a much reduced binary size. If you want to continue using the option, please use an older version of the cilium-bugtool binary.
• The DNS Polling option used by toFQDNs.matchName rules is disabled by default in 1.4.x due to limitations in the implementation. It has been replaced by DNS Proxy support, which must be explicitly enabled via changes to the policy described below. To ease upgrade, users may opt to enable the DNS Polling in v1.4.x by adding the --tofqdns-enable-poller option to cilium-agent without changing policies. For instructions on how to safely upgrade see Upgrading DNS Polling deployments to DNS Proxy.
• The DaemonSet now uses dnsPolicy: ClusterFirstWithHostNet in order for Cilium to look up Kubernetes service names via DNS. This in turn requires the cluster to run a cluster DNS such as kube-dns or CoreDNS. If you are not running cluster DNS, remove the dnsPolicy field. This will mean that you cannot use the etcd-operator. More details can be found in the kube-dns section.

New ConfigMap Options¶

• enable-ipv4: If true, all endpoints are allocated an IPv4 address.
• enable-ipv6: If true, all endpoints are allocated an IPv6 address.
• preallocate-bpf-maps: If true, reduce per-packet latency at the expense of up-front memory allocation for entries in BPF maps. If this value is modified, then during the next Cilium startup the restore of existing endpoints and tracking of ongoing connections may be disrupted. This may lead to policy drops or a change in loadbalancing decisions for a connection for some time. Endpoints may need to be recreated to restore connectivity. If this option is set to false during an upgrade to 1.4.0 or later, then it may cause one-time disruptions during the upgrade.
• auto-direct-node-routes: If true, then enable automatic L2 routing between nodes. This is useful when running in direct routing mode and can be used as an alternative to running a routing daemon. Routes to other Cilium managed nodes will then be installed on automatically.
• install-iptables-rules: If set to false then Cilium will not install any iptables rules which are mainly for interaction with kube-proxy. By default it is set to true.
• masquerade: The agent can optionally be set up for masquerading all network traffic leaving the main networking device if masquerade is set to true. By default it is set to false.
• datapath-mode: Cilium can operate in two different datapath modes, that is, either based upon veth devices (default) or ipvlan devices (beta). The latter requires an additional setting to specify the ipvlan master device.
• New ipvlan-specific CNI integration mode options (beta):
  • ipvlan-master-device: When running Cilium in ipvlan datapath mode, an ipvlan master device must be specified. This is typically pointing to a networking device that is facing the external network. Be aware that this will be used by all nodes, so it is required that the device name is consistent on all nodes where this is going to be deployed.
• New flannel-specific CNI integration mode options (beta):
  • flannel-master-device: When running Cilium with policy enforcement enabled on top of Flannel, the BPF programs will be installed on the network interface specified in this option and on each network interface belonging to a pod.
  • flannel-uninstall-on-exit: If flannel-master-device is specified, this determines whether Cilium should remove BPF programs from the master device and interfaces belonging to pods when the Cilium DaemonSet is deleted. If true, Cilium will remove programs from the pods.
  • flannel-manage-existing-containers: On startup, install a BPF programs to allow for policy enforcement on pods that are currently managed by Flannel. This also requires the Cilium DaemonSet to be running with hostPID: true, which is not enabled by default.

Deprecated ConfigMap Options¶

• disable-ipv4: Superseded by enable-ipv4, with the logic reversed.
• legacy-host-allows-world: This option allowed users to specify Cilium 1.0-style policies that treated traffic that is masqueraded from the outside world as though it arrived from the local host. As of Cilium 1.4, the option is disabled by default if not specified in the ConfigMap, and the option is scheduled to be removed in Cilium 1.5 or later.

Upgrading from 1.2.x to 1.3.y¶

1. If you are running Cilium 1.0.x or 1.1.x, please upgrade to 1.2.x first. It is also possible to upgrade from 1.0 or 1.1 directly to 1.3 by combining the upgrade instructions for each minor release. See the documentation for said releases for further information.
2. Upgrade to Cilium 1.2.4 or later using the guide Upgrading Micro Versions.
3. Follow the standard procedures to perform the upgrade as described in Upgrading Minor Versions.

New ConfigMap Options¶

• ct-global-max-entries-tcp/ct-global-max-entries-other: Specifies the maximum number of connections supported across all endpoints, split by protocol: tcp or other. One pair of maps uses these values for IPv4 connections, and another pair of maps use these values for IPv6 connections. If these values are modified, then during the next Cilium startup the tracking of ongoing connections may be disrupted. This may lead to brief policy drops or a change in loadbalancing decisions for a connection.
• clean-cilium-bpf-state: Similar to clean-cilium-state but only cleans the BPF state while preserving all other state. Endpoints will still be restored and IP allocations will prevail but all datapath state is cleaned when Cilium starts up. Not required for normal operation.

Export the current ConfigMap¶

$ kubectl get configmap -n kube-system cilium-config -o yaml --export > cilium-cm-old.yaml
$ cat ./cilium-cm-old.yaml
apiVersion: v1
data:
  clean-cilium-state: "false"
  debug: "true"
  disable-ipv4: "false"
  etcd-config: |-
    ---
    endpoints:
    - https://192.168.33.11:2379
    #
    # In case you want to use TLS in etcd, uncomment the 'ca-file' line
    # and create a kubernetes secret by following the tutorial in
    # https://cilium.link/etcd-config
    ca-file: '/var/lib/etcd-secrets/etcd-client-ca.crt'
    #
    # In case you want client to server authentication, uncomment the following
    # lines and add the certificate and key in cilium-etcd-secrets below
    key-file: '/var/lib/etcd-secrets/etcd-client.key'
    cert-file: '/var/lib/etcd-secrets/etcd-client.crt'
kind: ConfigMap
metadata:
  creationTimestamp: null
  name: cilium-config
  selfLink: /api/v1/namespaces/kube-system/configmaps/cilium-config

In the ConfigMap above, we can verify that Cilium is using debug with true, it has a etcd endpoint running with TLS, and the etcd is set up to have client to server authentication.

Download the ConfigMap with the changes for 1.5.3¶

$ wget https://raw.githubusercontent.com/cilium/cilium/1.5.3/examples/kubernetes/1.10/cilium-cm.yaml

$ wget https://raw.githubusercontent.com/cilium/cilium/1.5.3/examples/kubernetes/1.11/cilium-cm.yaml

$ wget https://raw.githubusercontent.com/cilium/cilium/1.5.3/examples/kubernetes/1.12/cilium-cm.yaml

$ wget https://raw.githubusercontent.com/cilium/cilium/1.5.3/examples/kubernetes/1.13/cilium-cm.yaml

$ wget https://raw.githubusercontent.com/cilium/cilium/1.5.3/examples/kubernetes/1.14/cilium-cm.yaml

Verify its contents:

---
apiVersion: v1
kind: ConfigMap
metadata:
  name: cilium-config
  namespace: kube-system
data:
  # This etcd-config contains the etcd endpoints of your cluster. If you use
  # TLS please make sure you follow the tutorial in https://cilium.link/etcd-config
  etcd-config: |-
    ---
    endpoints:
      - https://cilium-etcd-client.kube-system.svc:2379
    #
    # In case you want to use TLS in etcd, uncomment the 'ca-file' line
    # and create a kubernetes secret by following the tutorial in
    # https://cilium.link/etcd-config
    ca-file: '/var/lib/etcd-secrets/etcd-client-ca.crt'
    #
    # In case you want client to server authentication, uncomment the following
    # lines and create a kubernetes secret by following the tutorial in
    # https://cilium.link/etcd-config
    key-file: '/var/lib/etcd-secrets/etcd-client.key'
    cert-file: '/var/lib/etcd-secrets/etcd-client.crt'

  # If you want to run cilium in debug mode change this value to true
  debug: "false"

  # If you want metrics enabled in all of your Cilium agents, set the port for
  # which the Cilium agents will have their metrics exposed.
  # This option deprecates the "prometheus-serve-addr" in the
  # "cilium-metrics-config" ConfigMap
  # NOTE that this will open the port on ALL nodes where Cilium pods are
  # scheduled.
  # prometheus-serve-addr: ":9090"

  # Enable IPv4 addressing. If enabled, all endpoints are allocated an IPv4
  # address.
  enable-ipv4: "true"

  # Enable IPv6 addressing. If enabled, all endpoints are allocated an IPv6
  # address.
  enable-ipv6: "false"

  # If a serious issue occurs during Cilium startup, this
  # invasive option may be set to true to remove all persistent
  # state. Endpoints will not be restored using knowledge from a
  # prior Cilium run, so they may receive new IP addresses upon
  # restart. This also triggers clean-cilium-bpf-state.
  clean-cilium-state: "false"
  # If you want to clean cilium BPF state, set this to true;
  # Removes all BPF maps from the filesystem. Upon restart,
  # endpoints are restored with the same IP addresses, however
  # any ongoing connections may be disrupted briefly.
  # Loadbalancing decisions will be reset, so any ongoing
  # connections via a service may be loadbalanced to a different
  # backend after restart.
  clean-cilium-bpf-state: "false"

  # Users who wish to specify their own custom CNI configuration file must set
  # custom-cni-conf to "true", otherwise Cilium may overwrite the configuration.
  custom-cni-conf: "false"

  # If you want cilium monitor to aggregate tracing for packets, set this level
  # to "low", "medium", or "maximum". The higher the level, the less packets
  # that will be seen in monitor output.
  monitor-aggregation: "none"

  # ct-global-max-entries-* specifies the maximum number of connections
  # supported across all endpoints, split by protocol: tcp or other. One pair
  # of maps uses these values for IPv4 connections, and another pair of maps
  # use these values for IPv6 connections.
  #
  # If these values are modified, then during the next Cilium startup the
  # tracking of ongoing connections may be disrupted. This may lead to brief
  # policy drops or a change in loadbalancing decisions for a connection.
  #
  # For users upgrading from Cilium 1.2 or earlier, to minimize disruption
  # during the upgrade process, comment out these options.
  bpf-ct-global-tcp-max: "524288"
  bpf-ct-global-any-max: "262144"

  # Pre-allocation of map entries allows per-packet latency to be reduced, at
  # the expense of up-front memory allocation for the entries in the maps. The
  # default value below will minimize memory usage in the default installation;
  # users who are sensitive to latency may consider setting this to "true".
  #
  # This option was introduced in Cilium 1.4. Cilium 1.3 and earlier ignore
  # this option and behave as though it is set to "true".
  #
  # If this value is modified, then during the next Cilium startup the restore
  # of existing endpoints and tracking of ongoing connections may be disrupted.
  # This may lead to policy drops or a change in loadbalancing decisions for a
  # connection for some time. Endpoints may need to be recreated to restore
  # connectivity.
  #
  # If this option is set to "false" during an upgrade from 1.3 or earlier to
  # 1.4 or later, then it may cause one-time disruptions during the upgrade.
  preallocate-bpf-maps: "false"

  # Regular expression matching compatible Istio sidecar istio-proxy
  # container image names
  sidecar-istio-proxy-image: "cilium/istio_proxy"

  # Encapsulation mode for communication between nodes
  # Possible values:
  #   - disabled
  #   - vxlan (default)
  #   - geneve
  tunnel: "vxlan"

  # Name of the cluster. Only relevant when building a mesh of clusters.
  cluster-name: default

  # Unique ID of the cluster. Must be unique across all conneted clusters and
  # in the range of 1 and 255. Only relevant when building a mesh of clusters.
  #cluster-id: 1

  # Interface to be used when running Cilium on top of a CNI plugin.
  # For flannel, use "cni0"
  flannel-master-device: ""
  # When running Cilium with policy enforcement enabled on top of a CNI plugin
  # the BPF programs will be installed on the network interface specified in
  # 'flannel-master-device' and on all network interfaces belonging to
  # a container. When the Cilium DaemonSet is removed, the BPF programs will
  # be kept in the interfaces unless this option is set to "true".
  flannel-uninstall-on-exit: "false"
  # Installs a BPF program to allow for policy enforcement in already running
  # containers managed by Flannel.
  # NOTE: This requires Cilium DaemonSet to be running in the hostPID.
  # To run in this mode in Kubernetes change the value of the hostPID from
  # false to true. Can be found under the path `spec.spec.hostPID`
  flannel-manage-existing-containers: "false"

  # DNS Polling periodically issues a DNS lookup for each `matchName` from
  # cilium-agent. The result is used to regenerate endpoint policy.
  # DNS lookups are repeated with an interval of 5 seconds, and are made for
  # A(IPv4) and AAAA(IPv6) addresses. Should a lookup fail, the most recent IP
  # data is used instead. An IP change will trigger a regeneration of the Cilium
  # policy for each endpoint and increment the per cilium-agent policy
  # repository revision.
  #
  # This option is disabled by default starting from version 1.4.x in favor
  # of a more powerful DNS proxy-based implementation, see [0] for details.
  # Enable this option if you want to use FQDN policies but do not want to use
  # the DNS proxy.
  #
  # To ease upgrade, users may opt to set this option to "true".
  # Otherwise please refer to the Upgrade Guide [1] which explains how to
  # prepare policy rules for upgrade.
  #
  # [0] http://docs.cilium.io/en/stable/policy/language/#dns-based
  # [1] http://docs.cilium.io/en/stable/install/upgrade/#changes-that-may-require-action
  tofqdns-enable-poller: "false"

  # wait-bpf-mount makes init container wait until bpf filesystem is mounted
  wait-bpf-mount: "false"

  # Enable legacy services (prior v1.5) to prevent from terminating existing
  # connections with services when upgrading Cilium from < v1.5 to v1.5.
  enable-legacy-services: "false"

Add new options¶

Add the new options manually to your old ConfigMap, and make the necessary changes.

In this example, the debug option is meant to be kept with true, the etcd-config is kept unchanged, and legacy-host-allows-world is a new option, but after reading the Version Specific Notes the value was kept unchanged from the default value.

After making the necessary changes, the old ConfigMap was migrated with the new options while keeping the configuration that we wanted:

$ cat ./cilium-cm-old.yaml
apiVersion: v1
data:
  debug: "true"
  disable-ipv4: "false"
  # If you want to clean cilium state; change this value to true
  clean-cilium-state: "false"
  legacy-host-allows-world: "false"
  etcd-config: |-
    ---
    endpoints:
    - https://192.168.33.11:2379
    #
    # In case you want to use TLS in etcd, uncomment the 'ca-file' line
    # and create a kubernetes secret by following the tutorial in
    # https://cilium.link/etcd-config
    ca-file: '/var/lib/etcd-secrets/etcd-client-ca.crt'
    #
    # In case you want client to server authentication, uncomment the following
    # lines and add the certificate and key in cilium-etcd-secrets below
    key-file: '/var/lib/etcd-secrets/etcd-client.key'
    cert-file: '/var/lib/etcd-secrets/etcd-client.crt'
kind: ConfigMap
metadata:
  creationTimestamp: null
  name: cilium-config
  selfLink: /api/v1/namespaces/kube-system/configmaps/cilium-config

Apply new ConfigMap¶

After adding the options, manually save the file with your changes and install the ConfigMap in the kube-system namespace of your cluster.

$ kubectl apply -n kube-system -f ./cilium-cm-old.yaml

As the ConfigMap is successfully upgraded we can start upgrading Cilium DaemonSet and RBAC which will pick up the latest configuration from the ConfigMap.

Affected versions¶

• Any version of Cilium running on Linux 4.10 or earlier

When a CIDR policy with too many unique prefix lengths is imported, Cilium will reject the policy with a message like the following:

$ cilium policy import too_many_cidrs.json
Error: Cannot import policy: [PUT /policy][500] putPolicyFailure  too many
egress CIDR prefix lengths (128/40)

$ cilium policy import too_many_cidrs.json
Error: Cannot import policy: [PUT /policy][500] putPolicyFailure  Adding
specified prefixes would result in too many prefix lengths (current: 3,
result: 32, max: 18)

The supported count of unique prefix lengths may differ between Cilium minor releases, for example Cilium 1.1 supports 20 unique prefix lengths on Linux 4.10 or older, while Cilium 1.2 only supports 18 (for IPv4) or 4 (for IPv6).

Mitigation¶

Users may construct CIDR policies that use fewer unique prefix lengths. This can be achieved by composing or decomposing adjacent prefixes.

Solution¶

Upgrade the host Linux version to 4.11 or later. This step is beyond the scope of the Cilium guide.

Affected versions¶

• Cilium 1.2 and 1.3 when using DNS Polling with toFQDNs.matchName policy rules and upgrading to cilium 1.4.0 or later.
• Cilium 1.4 or later that do not yet have L7 DNS Proxy (preferred) policy rules.

Mitigation¶

Deployments that require a seamless transition to DNS Proxy may use Running a pre-flight DaemonSet to create a copy of DNS information on each cilium node for use by the upgraded cilium-agent at startup. This data is used to allow L3 connections (via toFQDNs.matchName and toFQDNs.matchPattern rules) without a DNS request from pods. Running a pre-flight DaemonSet accomplishes this via the --tofqdns-pre-cache CLI option, which reads DNS cache data for use on startup.

Solution¶

DNS data obtained via polling must be recorded for use on startup and rules added to intercept DNS lookups. The steps are split into a section on seamlessly upgrading DNS Polling and then further beginning to intercept DNS data via a DNS Proxy.

Policy rules may be prepared to use the DNS Proxy before an upgrade to 1.4. The new policy rule fields toFQDNs.matchPattern and toPorts.rules.dns.matchName/matchPattern will be ignored by older cilium versions and can be safely implemented prior to an upgrade.

The following example allows DNS access to kube-dns via the DNS Proxy and allows all DNS requests to kube-dns. For completeness, toFQDNs rules are included for examples of the syntax for those L3 policies as well. Existing toFQDNs rules do not need to be modified but will now use IPs seen by DNS requests and allowed by the toFQDNs.matchPattern rule.

apiVersion: cilium.io/v2
kind: CiliumNetworkPolicy
metadata:
  name: "tofqdn-dns-visibility"
spec:
  endpointSelector:
    matchLabels:
      any:org: alliance
  egress:
  - toEndpoints:
    - matchLabels:
       "k8s:io.kubernetes.pod.namespace": kube-system
       "k8s:k8s-app": kube-dns
    toPorts:
      - ports:
         - port: "53"
           protocol: ANY
        rules:
          dns:
            - matchPattern: "*"
  - toFQDNs:
      - matchName: "cilium.io"
      - matchName: "sub.cilium.io"

[
  {
    "endpointSelector": {
      "matchLabels": {
        "app": "test-app"
      }
    },
    "egress": [
      {
        "toEndpoints": [
          {
            "matchLabels": {
              "app-type": "dns"
            }
          }
        ],
        "toPorts": [
          {
            "ports": [
              {
                "port": "53",
                "protocol": "ANY"
              }
            ],
            "rules": {
              "dns": [
                { "matchPattern": "*" }
              ]
            }
          }
        ]
      },
      {
        "toFQDNs": [
          { "matchName": "cilium.io" },
          { "matchName": "sub.cilium.io" }
        ]
      }
    ]
  }
]

Upgrade steps - DNS Polling¶

1. Set the tofqdns-enable-poller field to true in the cilium ConfigMap used in the upgrade. Alternatively, pass --tofqdns-enable-poller=true to the upgraded cilium-agent.
2. Add tofqdns-pre-cache: "/var/run/cilium/dns-precache-upgrade.json" to the ConfigMap. Alternatively, pass --tofqdns-pre-cache="/var/run/cilium/dns-precache-upgrade.json" to cilium-agent.
3. Deploy the cilium Running a pre-flight DaemonSet helper. This will download the cilium container image and also create DNS pre-cache data at /var/run/cilium/dns-precache-upgrade.json. This data will have a TTL of 1 week.
4. Deploy the new cilium DaemonSet
5. (optional) Remove tofqdns-pre-cache: "/var/run/cilium/dns-precache-upgrade.json" from the cilium ConfigMap. The data will automatically age-out after 1 week.

Conversion steps - DNS Proxy (preferred)¶

1. Update existing policies to intercept DNS requests. See DNS Policy and IP Discovery or the example above
2. Allow pods to make DNS requests to populate the cilium-agent cache. To check which exact queries are in the DNS cache and when they will expire use cilium fqdn cache list
3. Set the tofqdns-enable-poller field to false in the cilium ConfigMap
4. Restart the cilium pods with the new ConfigMap. They will restore Endpoint policy with DNS information from intercepted DNS requests stored in the cache