How to manage state with a Kubernetes Application

NOVEMBER 6, 2019
Creating a stateful application with
K8S and AWS DB Services
Bahubali Shetti
@Shetti
Director of Cloud Developer Advocacy VMware

App
Cloud
Database
Kubernetes
? Self or Fully Manage
Stateless
Application Data
(Application State)
Not easy to move
Maximizing portability with Kubernetes – BUT what about the Data?

Self Managed vs Fully Managed….
?

Find it at cloudjourney.io (look for the github site)
Typical Microservice App - AcmeShop

Parameterizing the database end points in K8S
Code Instrumentation for App-DB connectivity
...
…
spec:
volumes:
- name: acmefit-order-data
emptyDir: {}
containers:
- image: order:latest
name: order
env:
- name: ORDER_DB_HOST
value: 'order-mongo'
- name: ORDER_DB_PASSWORD
valueFrom:
secretKeyRef:
name: order-mongo-pass
key: password
- name: ORDER_DB_PORT
value: '27017’
…
…
FROM bitnami/python:3.7
MAINTAINER Bill Shetti
"billshetti@gmail.com"
ENV ORDER_DB_HOST="localhost"
ENV ORDER_DB_PORT="27017"
ENV ORDER_DB_PASSWORD=""
ENV ORDER_DB_USERNAME=""
ENV PAYMENT_HOST="localhost"
ENV PAYMENT_PORT="9000"
# needed for mongo client
RUN install_packages mongodb-
clients
COPY ./requirements.txt
/app/requirements.txt
RUN pip3 install -r requirements.txt
…
…
…
…
from os import environ
if environ.get('ORDER_DB_USERNAME') is not
None:
if os.environ['ORDER_DB_USERNAME'] !=
"":
mongouser=os.environ['ORDER_DB_USERNA
ME']
else:
mongouser=''
else:
mongouser=''
if environ.get('ORDER_DB_HOST') is not None:
if os.environ['ORDER_DB_HOST'] != "":
mongohost=os.environ['ORDER_DB_HOST']
else:
mongohost='localhost'
else:
mongohost='localhost’
…
…
K8S Yaml Dockerfile Python order.py code

Libraries and connecting to the DB
Code Instrumentation for DB
import pymongo
from pymongo import MongoClient
from pymongo import errors as mongoerrors
client=MongoClient(mongouri)
#uri=username:password@host:port
Or
client=MongoClient(host=mongohost, port=int(mongoport),
username=mongouser, password=mongopassword)
import redis
rConn=redis.StrictRedis(host=redishost, port=redisport,
password=redispassword, db=0)
Lots of standard libraries (go, python, etc) with significant support

Installation and management – Several options
Setting up your own containerized DB
Initialization
kubectl create secret generic order-mongo-pass
--from-literal=password=<value>
Secrets
Create
kubectl apply -f order-db-total.yaml
kubectl apply -f config-map.yaml
Automates deployment of single node
or replica sets for mongodb
Enables set up of alerting, monitoring
Optional persistence and storage
configuration
Easy scale
(ISH)
OperatorsSimple K8S Create
Early Days
(Beta/Alpha)
(still adding features)
etc
(ISH)

Installation and management – Several options
HA Configured by default

So what’s hard? – Keeping state persistent
apiVersion: storage.k8s.io/v1
kind: StorageClass
metadata:
name: fast
provisioner: kubernetes.io/aws-ebs
parameters:
type: io1
containers:
- name: mongo
image: mongo
ports:
- containerPort: 27017
volumeMounts:
- name: mongo-persistent-storage
mountPath: /data/db
volumeClaimTemplates:
- metadata:
name: mongo-persistent-storage
annotations:
volume.beta.kubernetes.io/storage-class: "fast"
spec:
accessModes: ["ReadWriteOnce"]
resources:
requests:
storage: 100Gi
Storage Provisioning and Management

Infra and DB expertise or App and business focus?
What’s important to you?
Infra Skills App Skills
?

AWS Options
Fully Managed Databases

Simple to implement
Creating HA arch – operational overhead
No support for sharding
No encryption
Operationally Expensive
Simple to implement
Built in HA with read replicas, multiple primaries, failovers, etc
Easly scalable
Sharding support
Encryption at rest/intransit
K8S
Operator
VS
√
Redis and AWS Elasticache
Managed Databases

Built in sharding & replica sets for easy scaling
Still have to manually add nodes
Manage backup manually or with tools (OpsManager, CloudManager etc)
Manage Upgrades etc
K8S
Operator
Easy setup (only compatible with MongoDB 3.6)
Managed sharding, replicas
Managed scale (up to 64TB)
Easy backups - AWS
Handles 100ks reads/writes/sec
Easy setup – deploys on AWS/Azure/GCP
Managed sharding, replicas
Managed scale
Easy backups – AWS/Azure/GCP
Pure Mongo experience with latest and greatest features
VS
√
√
DocumentDB and MongoDB Atlas
Managed Databases

apiVersion: storage.k8s.io/v1
kind: StorageClass
metadata:
name: fast
provisioner: kubernetes.io/aws-ebs
parameters:
type: io1
containers:
- name: mongo
image: mongo
ports:
- containerPort: 27017
volumeMounts:
- name: mongo-persistent-storage
mountPath: /data/db
volumeClaimTemplates:
- metadata:
name: mongo-persistent-storage
annotations:
volume.beta.kubernetes.io/storage-class: "fast"
spec:
accessModes: ["ReadWriteOnce"]
resources:
requests:
storage: 100Gi
Storage Provisioning and Management
Keeping state persistent?
Using a managedDB

aws --region us-east-2 elasticache create-cache-cluster --cache-cluster-id my-cluster --cache-node-type cache.r4.large --
engine redis --engine-version 3.2.4 --num-cache-nodes 1 --cache-parameter-group default.redis3.2
aws --region us-east-2 docdb create-db-cluster --db-cluster-identifier mongoeq --engine docdb --master-username bill --
master-user-password password1
OR
https://github.com/awslabs/aws-service-operator
AWS CLI with automation
AWS service operator
(ISH)
Installation and management –
Using a managed DB

Parameterizing the database end points in K8S
Code Instrumentation for DB
...
…
spec:
volumes:
- name: acmefit-order-data
emptyDir: {}
containers:
- image: order:latest
name: order
env:
- name: ORDER_DB_HOST
value: 'order-mongo'
- name: ORDER_DB_PASSWORD
valueFrom:
secretKeyRef:
name: order-mongo-pass
key: password
- name: ORDER_DB_PORT
value: '27017’
…
…
FROM bitnami/python:3.7
MAINTAINER Bill Shetti
"billshetti@gmail.com"
ENV ORDER_DB_HOST="localhost"
ENV ORDER_DB_PORT="27017"
ENV ORDER_DB_PASSWORD=""
ENV ORDER_DB_USERNAME=""
ENV PAYMENT_HOST="localhost"
ENV PAYMENT_PORT="9000"
# needed for mongo client
RUN install_packages mongodb-
clients
COPY ./requirements.txt
/app/requirements.txt
RUN pip3 install -r requirements.txt
…
…
…
…
from os import environ
if environ.get('ORDER_DB_USERNAME') is not
None:
if os.environ['ORDER_DB_USERNAME'] !=
"":
mongouser=os.environ['ORDER_DB_USERNA
ME']
else:
mongouser=''
else:
mongouser=''
if environ.get('ORDER_DB_HOST') is not None:
if os.environ['ORDER_DB_HOST'] != "":
mongohost=os.environ['ORDER_DB_HOST']
else:
mongohost='localhost'
else:
mongohost='localhost’
…
…
K8S Yaml Dockerfile Python order.py code
Insert Document DB URL INFO HERE

import pymongo
from pymongo import MongoClient
from pymongo import errors as mongoerrors
client=MongoClient(mongouri)
#uri=username:password@host:port
Or
client=MongoClient(host=mongohost, port=int(mongoport),
username=mongouser, password=mongopassword)
import redis
rConn=redis.StrictRedis(host=redishost, port=redisport,
password=redispassword, db=0)
Steps and hurdles – code instrumentization
Using a Managed DB
URL FROM AWS Services

19
www.cloudjourney.io
@cloudjourneyio
Bahubali Shetti - @Shetti

How to manage state with a Kubernetes Application

How to manage state with a Kubernetes Application

More Related Content

What's hot (20)

Similar to How to manage state with a Kubernetes Application (20)

Recently uploaded (20)

How to manage state with a Kubernetes Application