About traditional database high availability, the method is like:
- master-master replication
- master-slave replication
- automatic failover
If use container to do that, we can use these ways:
busyboxto save data, but how to do when the data goes large if only use container?
- Use volume feature to sync data from container to hard disk. This way can do replication as traditional way
- Connect database from application container. Then the next thing is the same as traditional way
- Use large storage service like OpenStack Swift or AWS EBS. Maybe this way can save the data well depend on third party services.
At least, the first choice, use pure container technology, is there a way to do high availability?