Files
armada/charts/tiller/values.yaml
Phil Sphicas a3f11e5873 Tiller: listen on localhost by default
This change introduces a configuration option to control whether Tiller
listens on any IP addresses (the previous default), or binds only to
127.0.0.1 (the new default).

The same option is used for both the Armada and Tiller charts:
    .conf.tiller.listen_on_any (default: false)

The affected tiller command line argument is:
    -listen 127.0.0.1:port (if false)
    -listen :port (if true)

Listening on any address allows Helm client direct access to Tiller, via
'helm --host pod_ip:port'.

Listening on localhost does prevent connections directly to the pod IP,
but it does not preclude the use of 'kubectl port-forward' to establish
a connection to Tiller.

The Tiller container in the Armada pod exists only to service Armada via
127.0.0.1. The Helm client automatically sets up port forwarding (if it
has access to the Kubernetes API). As a result, this change should be
non-impacting. However, the previous behavior can be restored by setting
.conf.tiller.listen_on_any=true.

Change-Id: Id308976bac21cc521e8470516ce49ebd1942da68
2021-04-22 20:29:02 +00:00

127 lines
3.4 KiB
YAML

# Copyright 2017 AT&T Intellectual Property. All other rights reserved.
#
# Licensed under the Apache License, Version 2.0 (the "License");
# you may not use this file except in compliance with the License.
# You may obtain a copy of the License at
#
# http://www.apache.org/licenses/LICENSE-2.0
#
# Unless required by applicable law or agreed to in writing, software
# distributed under the License is distributed on an "AS IS" BASIS,
# WITHOUT WARRANTIES OR CONDITIONS OF ANY KIND, either express or implied.
# See the License for the specific language governing permissions and
# limitations under the License.
# namespace: "kube-system"
labels:
node_selector_key: ucp-control-plane
node_selector_value: enabled
dependencies:
static:
tiller_deploy:
images:
tags:
tiller: gcr.io/kubernetes-helm/tiller:v2.16.9
pull_policy: "IfNotPresent"
local_registry:
# NOTE(portdirect): this tiller chart does not support image pulling
active: false
exclude:
- tiller
deployment:
# NOTE: Current replica is hard-coded to 1. This is a placeholder variable
# for future usage. Updates will be made to the chart when we know that
# tiller is stable with multiple instances.
replicas: 1
# The amount of revision tiller is willing to support. 0 means that there is
# no limit.
tiller_history: 0
conf:
tiller:
verbosity: 5
storage: null
# Only postgres is supported so far
sql_dialect: postgres
sql_connection: null
trace: false
# Note: Defaulting to the (default) kubernetes grace period, as anything
# greater than that will have no effect.
prestop_sleep: 30
# To have Tiller bind to all interfaces, allowing direct connections from
# the Helm client to pod_ip:port, set 'listen_on_any: true'.
# The default setting 'listen_on_any: false' binds Tiller to 127.0.0.1.
# Helm clients with Kubernetes API access dynamically set up a portforward
# into the pod, which works with the default setting.
listen_on_any: false
port: 44134
probe_port: 44135
pod:
mandatory_access_control:
type: apparmor
tiller:
tiller: runtime/default
security_context:
tiller:
pod:
runAsUser: 65534
container:
tiller:
allowPrivilegeEscalation: false
readOnlyRootFilesystem: true
probes:
tiller:
tiller:
readiness:
enabled: true
params:
failureThreshold: 3
initialDelaySeconds: 1
periodSeconds: 10
successThreshold: 1
timeoutSeconds: 1
liveness:
enabled: true
params:
failureThreshold: 3
initialDelaySeconds: 1
periodSeconds: 10
successThreshold: 1
timeoutSeconds: 1
resources:
enabled: false
tiller:
limits:
memory: "128Mi"
cpu: "100m"
requests:
memory: "128Mi"
cpu: "100m"
mounts:
tiller:
tiller:
volumes:
- name: kubernetes-client-cache
emptyDir: {}
volumeMounts:
- name: kubernetes-client-cache
# Should be the `$HOME/.kube` of the `runAsUser` above
# as this is where tiller's kubernetes client roots its cache dir.
mountPath: /tmp/.kube
network_policy:
tiller:
ingress:
- {}
egress:
- {}
manifests:
deployment_tiller: true
service_tiller_deploy: true
network_policy: false