Skip to content

Latest commit

 

History

History
208 lines (134 loc) · 5.98 KB

README.MD

File metadata and controls

208 lines (134 loc) · 5.98 KB
Raw

#Helm Charts for ingest service

Can be run with datawave-helm-charts or standalone with an existing hadoop, zk, and accumulo configuration

Running with datawave-helm-charts

Deploying to a local registry

# see https://minikube.sigs.k8s.io/docs/handbook/registry/ for more details
# enable the minikube registry at localhost:5000
minikube addons enable registry

# create a pipe to the registry running inside of minikube
docker run --rm -it --network=host alpine ash -c "apk add socat && socat TCP-LISTEN:5000,reuseaddr,fork TCP:$(minikube ip):5000"

# build the images
mvn clean package -Pdocker

# create the minikube registry tags and push to the registry
mvn clean install -Pdocker -Ddocker.registry="localhost:5000/" -DpushImage -DskipDockerBuild

# use localhost:5000/ as the registry for all ingest services

##Requirements:

  • Running k8s cluster
  • helm 3+
  • configuration service address
  • zookeeper address
  • rabbitmq

##Setup: k8s cluster minikube for single instance

# install minikube
curl -LO https://storage.googleapis.com/minikube/releases/latest/minikube-linux-amd64
sudo install minikube-linux-amd64 /usr/local/bin/minikube
# start minikube
minikube delete --all --purge && \
minikube start --cpus 8 --memory 30960 --disk-size 20480 --insecure-registry="containeryard.evoforge.org" && \
minikube image load busybox:1.28
# update .bashrc for kubectl
alias kubectl="minikube kubectl --"
# test kubectl command
kubectl get pods

Minikube maintains its own internal registry. In order to push images to the internal minikube registry the minikube docker-env needs to be exposed

# source the minikube docker-env (this will only work for this terminal)
# all docker commands will then be forwarded to the minikube internals 
eval $(minikube -p minikube docker-env)

# build services as usual
mvn clean install -Pdocker

# if using the configuration service also build the datawave/config-service image

Updating values.yaml requires the following configuration. See values.example.yaml for an example values.yaml

Globals

global.registry- default registry name to append to all images

global.services.configuration- configuration service address

global.services.zookeeper - zookeeper service address

Secrets

global.secrets.accumulo - accumulo user account secret information

global.secrets.keystore - keystore name, alias, path, and password

global.secrets.truststore - truststore name, path, and password

Volumes

global.volumes.pki - pki files (secret recommended)

global.volumes.hadoop - hadoop config files

global.volumes.logs - log directory

global.volumes.datawave - datawave ingest configuration files

global.volumes.output - datawave ingest output location

global.volumes.input - bundler input location (suggested to match output location)

global.volumes.config - configuration service config volume

global.volumes.rabbit - messaging service config volume

Each volume must have the following configured

.name - the name of the resource

.destination - the internal destination for the resource

.source.type - one of secret, configmap, hostPath

.source.name or .source.path - when using secret or configmap name must be specified to reference a valid secret or configmap. When type is hostPath a source.path must be specified

###Ingest has three parts:

####Feeder feeder.feeds - An list which defines feeds to process. Each feed must have:

.name - the name of the feed

.profile - the config service profiles to apply to this feed, comma delimited

.queue - the rabbit queue to push files found to

####Ingest ingest.pools - A list of ingest processing pools. Each pool must have:

.name - a name to identify the pool

.queue - a queue to read feeder messages from (should match a queue specified in at least one feeder)

.replicas - the number of replicas to run for this pool

.profiles - the config service profiles to apply, comma delimited

Ingest must have a secret defined for its accumulo username and password. This may be controlled via

.

####Bundler bundler - defines what bundlers are in use, if any

bundler.enabled - when enabled bundlers will be deployed as a daemonset

bundler.pools - A list of bundlers to run in the daemonsets. Each bundler contains

.name - a name

.profiles - the config server profiles to apply to the bundler

See values.example.yaml for sample configuration. Any Configuration service must have the additional configs defined in example-config.

Support Services

configuration and messaging services may be deployed as part of this chart or an external configuration and messaging service may be used. .enabled - is the service enabled

.name - the service name, should be updated in global.services if being used

.replicaCount - the number of replicas to run the service

.image.repository - the image repository

.image.pullPolicy - the image pull policy

.image.tag - the image tag

Configuration config

.dir - the directory to serve config from

.port - the port to expose for the configuration service

Messaging config

Requires global.volumes.rabbitmq

.ports.amqp - the amqp port to expose

.ports.mgmt - the management port to expose

Properties on all charts containers and initContainers

.image - section for image references

.image.registry - registry override for image

.image.repository - comes after registry in image definition

.image.tag - the image tag, pulled from the Chart if not specified

.image.pullPolicy - sets the pull policy

###Suggested testing environment minikube_helm

helm install zk zookeeper/
helm install hadoop hadoop/
helm install master ingest/ &
helm install accumulo accumulo/

# update config service to pull in example-config in values.yaml, then
helm install web web/

###Install this helm chart

cd chart
helm dependency update
helm install ingest .