Hacker Newsnew | past | comments | ask | show | jobs | submitlogin

A few possibilities:

- If you don't actually need files, design around database instead

- If you need to store files redundantly, you can use a distributed filesystem like HDFS or GridFs

- If you need to store files redundantly, you can use an IaaS like Amazon S3

- If you need to store actual files (like for hard-linking), (redundancy optional), you can use network mounted storage

- If periodic backups of the data is ok, you can run backups that ship to S3 or glacier

Just a few, offhand, I'm sure there are a number of other techniques.



> If you don't actually need files, design around database instead

That is fine if your target is eg AWS - but if you want all your infrastructure in docker, the db needs persistent storage...


That's what volumes are for -- put a volume on NFS and have your DB persist there.

Even with docker you still need roles for servers; your DB class of servers will have a cloud-config that inits the correct mount points.


Can you really run a transaction-oriented DB over NFS "in the cloud" with meaningful performance and guaranteed writes to disk in case of network/power failure?




Guidelines | FAQ | Lists | API | Security | Legal | Apply to YC | Contact

Search: