No longer maintained: Please use `tick-charts` for up to date version.

tick-kube

A tool to deploy the InfluxData TICK stack to Google Container engine.

Included are the full set of kubernetes deployment files for the stack.

Requirements

Google Cloud SDK
- Follow the instructions in the above article.
kubectl installed
- Once you have configured gcloud, run gcloud components install kubectl

Configuration

Pull this repo:

# One extra configuration change
$ git clone https://github.com/wuchangping/tick-demo.git

Personalize the variables in /config.sh:

# Defaults to go get download. Modify to be root of repo if using git clone
BP=$(pwd)

# Name of desired cluster.
# If using an existing cluster set that cluster's name
CLUSTER=tick-demo-1

# Name of gcloud project to use
PROJECT=pravega-dev

# gcloud region
REGION=us-west1

# gcloud zone
ZONE=us-west1-b

# num GB of disk for nodes when creating cluster
DISK=100

# gcloud instance types for nodes when creating cluster
MACHINE=n1-standard-1

# Number of nodes when creating kubernetes cluster
NUM_NODES=3

# InfluxDB Disk Size
INFLUX_DISK=25GB

# Chronograf and Kapacitor Disk Sizes
OTHER_DISK=10GB

To Deploy:

This tool offers a couple of different options:

Create a new cluster and spin up the full stack
Create the full stack on an existing kubernetes cluster

Create a new cluster

To create a new cluster running the full stack and return the IP where chronograf is running:

$ ./tick-kube deploy start

This command is a shortcut that runs following commands in order:

# First set configuration for gcloud and kubectl to the settings in ./config.sh
$ ./tick-kube config set
# Next create a new kubernetes cluster with those configuration parameters
$ ./tick-kube create cluster
# Create disks for influxdb, chronograf, and kapacitor
$ ./tick-kube create disks
# Create tick runs `kubectl apply -f $filename` on all manifests and creates associated configmaps
$ ./tick-kube create tick

This command takes a few minutes. It will hang at the end waiting for a public IP for the chronograf service:

Creating tick...
tick is the full stack of InfluxData products running in production configuration
namespace "tick" created
configmap "telegraf-config" created
configmap "influxdb-config" created
deployment "influxdb" created
service "influxdb" created
deployment "kapacitor" created
service "kapacitor" created
daemonset "telegraf" created
deployment "chronograf" created
service "chronograf" created
Waiting for public IP...
NAME         CLUSTER-IP     EXTERNAL-IP   PORT(S)   AGE
chronograf   10.7.249.214   <pending>     80/TCP    1s

Once the EXTERNAL-IP is visible visit it in your web browser. The URLs for configuring Influx and Kapacitor from Chronograf are as follows:

InfluxDB: http://influxdb.tick:8086
Kapacitor: http://kapacitor.tick:8086

Create the stack on an existing cluster

Creating the stack on an existing cluster is easy as well. After setting the variables in /config.sh Run the following commands:

# First set configuration for gcloud and kubectl to the settings in ./tick-kube
$ ./tick-kube config set
# Create disks for influxdb, chronograf, and kapacitor
$ ./tick-kube create disks
# Create tick runs `kubectl apply -f $filename` on all manifests and creates associated configmaps
$ ./tick-kube create tick

This command takes a few minutes. It will hang at the end waiting for a public IP for the chronograf service:

Creating tick...
tick is the full stack of InfluxData products running in production configuration
namespace "tick" created
configmap "telegraf-config" created
configmap "influxdb-config" created
deployment "influxdb" created
service "influxdb" created
deployment "kapacitor" created
service "kapacitor" created
daemonset "telegraf" created
deployment "chronograf" created
service "chronograf" created
Waiting for public IP...
NAME         CLUSTER-IP     EXTERNAL-IP   PORT(S)   AGE
chronograf   10.7.249.214   <pending>     80/TCP    1s

Once the EXTERNAL-IP is visible visit it in your web browser. The URLs for configuring Influx and Kapacitor from Chronograf are as follows:

InfluxDB: http://influxdb.tick:8086
Kapacitor: http://kapacitor.tick:8086

Tearing Down

You can tear down either the full cluster or just the Kubernetes objects and persistent disks.

Full cluster teardown

To destroy all resources created by this demo run:

$ ./tick-kube deploy shutdown

This command takes a few minutes and will require user input at 2 different places. This command is a shortcut that runs following commands in order:

# This deletes the cluster on google compute cloud and associated LBs
$ ./tick-kube delete cluster
# This deletes the persistent disks created for InfluxDB, Kapacitor and Chronograf
$ ./tick-kube delete disk

Namespace only teardown

To destroy the namespace and the associated persistent disks run the following:

# This deletes the tick namespace and removes tick from kubernetes
$ ./tick-kube delete tick
# This deletes the persistent disks created for InfluxDB, Kapacitor and Chronograf
$ ./tick-kube delete disk

NOTE: ./tick-kube delete disk may initially fail. This is because the disks are still mounted to the hosts. Wait a couple of minutes for them to get detached or go into the Google Cloud Console and delete them manually.

Troubleshooting

If you are getting authentication errors when trying to connect to the cluster try running:

$ gcloud auth application-default login

wuchangping/tick-demo

No longer maintained: Please use tick-charts for up to date version.