Skip to content

Latest commit

 

History

History
99 lines (66 loc) · 2.59 KB

aws_eb.md

File metadata and controls

99 lines (66 loc) · 2.59 KB
Raw

Docker Amazon Linux 2 Platform

https://docs.aws.amazon.com/elasticbeanstalk/latest/dg/using-features.migration-al.html https://docs.aws.amazon.com/elasticbeanstalk/latest/dg/docker-multicontainer-migration.html

notable changes

  • EB with Linux 2 prefers docker-compose.yml against Dockerrun.aws.json
  • Environment properties - EB will not automatically pass the enviroment variables. you need to map .env file with your docker-compose.yml source

Notes

  • elastic search
  • redis is in elastic cache
  • postgre is in RDS

debugging eb aws docker

tail /var/log/eb-activity.log -n 100
tail /var/log/docker.log -n 100

eb cli

installing

# on ubuntu 14.04
sudo apt-get install python-pip
sudo pip install awsebcli
# now you have the `eb` command, make sure to update your credentials in
# ~/.aws/config
#
#     [profile eb-cli]
#     aws_access_key_id = A******************Q
#     aws_secret_access_key = A**************************************U
#     
#     [default]
#     aws_access_key_id = A******************Q
#     aws_secret_access_key = Ar************************************XU
#     region = eu-west-1

usage

eb init --interactive
eb ssh productionMycom-env
eb deploy qaMyEnvironment #deploy currently commited code

eb logs

eb logs  # application logs (downoads them to s3 and then display)
eb logs -a  #all logs including cron 

eb status
eb open # opens website
eb local # not yet implemented but in future it vil provision the docken
         # AWS  on your localhost

eb push # alias to "git aws.push"  takes head commit of branch and
        # deploy to aws


## export / retrieve PostgreSQL dump from EBS RDS

the way how RDS backups works is that they create snapshot of the RDS
instance including data, therefore when you retrieving RDS backup you
are just rolling back state of snapshot. This is eficient enough as
backup
There is no way how to pull this snapshot from AWS. 

The only way how to acually do real `pg_dump` is to do fallowing

1. create a new micro EC2 instance and add it to *same security group as your production webservice* (...or create new security group so that this instance will have right to create connection to Postgresql 5432) # I'm not recommending SSH to one of your production EC2 instance
2. instal pg_dump on this instance 
3. `pg_dump dbname=my_db_name --username=my_db_username
   --host=somethingsomething.com --password > /tmp/dump.sql`
4. `scp` to your computer