AWS - noi-techpark/documentation GitHub Wiki
Amazon Web Services
This is just a getting-started guide to use AWS for our projects.
Types of instances
For example:
t2
with reduced CPU cycles on demandm4
one CPU full access (costs more)
Elastic Beanstalk
Simple miniservice to provide standard-images. Configuration layer over EC2 for instances, S3 for storage, etc. (uses general AWS services).
NB: Do not shut it down via EC2 instances, because it would be setup by EB instantly. Must be configured via Elastic Beanstalk inside the web-frontend of AWS.
New instances contain the default standard-image again, volumes are destroyed and reset. Other instances on AWS have a persistent volume/disk, that contains all things that you stored on it.
Down or upgrade an instance
Choose a instance, ex. Jenkins Instance state / STOP (not terminate, which means destroy) Instance settings - change instance type Instance state / START Now with same volume, because we configured it to have a persistent disk.
Use volumes/disks
Volumes/disks are fixed and can be changed from one instances to another see -> Elastic Block Store / Volumes: Detach/attach volume
Problems:
- No linux images from marketplace allowed to change volumes (except you create an AMI / Amazon Machine Image out of your marketplace image / use HAV or HVM virtualization type / rest default **)
- If mounted through UUID, the old and new volume have the same UUID (for recovery), use /dev for one and UUID for the other
- Check volumes if not attached, cost but are not used
** Now we can launch it by creating an instance out of the AMI Eventually change associate/disassociate Elastic IPs if you wanna switch from an old instance to a new one
Automated chronjobs for backups
- See Cloudwatch / Rules inside the AWS console webapp
- Usually through AWS lambda, but also predefined jobs exist.
Recovery of Virtual Machines
- Create snapshot with cloudwatch
- Create an AMI from the snapshot
- New instance out of that AMI