Backup service


(Rubén Martín) #1

Hello,

I’m curious about what Community Ops is using for backups and how can community sites learn from that or even use same infra for backups :slight_smile:

Thanks!


(Michael Kohler) #2

Not related to Community Ops, but I have two servers in different data centers. Then I do nightly rsync cronjobs for the most important directories to the other server (in both directions). This includes mysql database dumps, which get copied over as well. Of course this method only works if you have 2 servers :smiley:

Would be great indeed to hear from the Community Ops team, I think I could improve my process!


(Michael Buluma) #3

Great option @mkohler :smile: I’ve been using Rsync for months now and it serves the purpose well.

I’ve been able to get mysql dumps onto the same server. Just depends with how you set it up :smile: . But again a simpler option wouldn’t hurt.


(Yousef Alam) #4

The nice thing about AWS is that redundancy is either built-in or a button click away, so we don’t do anything too fancy. For our databases, RDS has replication nodes across availability zones with daily backups and S3 has something similar for file storage. For Discourse, the web servers are stateless so it doesn’t matter if they die, they just get swapped out for a new instance.

Ideally nobody is hosting their own databases, we put them all on RDS and take advantage of the above.


(Rubén Martín) #5

So the questions are:

  • How can communities use this backup infra?
  • What changes are needed? (if any)
  • Who should we coordinate with?