Is it possible to get a "personal" backup of NeuroStars?


Before considering to place all eggs into the neurostars bucket, decided to check if it is possible somehow to regularly acquire a backup sufficient enough to later re-establish content at another discourse deployment with a compatible version etc?


@yarikoptic - we tested this when we set this up. perhaps @fitz can provide details. also this is now being backed up to dropbox as well. it would be cool if we could spin up a docker container with neurostars content, just for testing purpose.


Yes, this should be possible. My only concern would be the personal user information, but maybe we can ask people if they would have issues/concerns with distributing a backup?

I’m also thinking that we are approaching a community size would justify the cost of hosting fees. I think this would be in-line with the scaling back of the INCF infrastructure.


May be the “public” dump could be “sanitized” to not include such information.

I am asking for the use-case of e.g. a project, which initially uses neurostars, decides to go rogue and establish independent presence. Having such a backup dump would allow to quickly populate a new instance. On the other side – having such capability would help to lure other projects into using neurostars as their primary exchange/support board.


@fitz - would be happy for incf to make the transition to discourse hosting if that’s easy to do. questions:

  • would it allow us to export data?
  • would we be able to use the plugins we currently have without going to expensive plans?


Hm, yes, I think I understand. It might be a little tricky, since it looks like some of the “sensitive” things are in a couple of tables, and they’re all linked together. But let me look at it a bit…will compile a list of things to be sanitized.

But, also anyone can just grab content from the RSS feed to populate a some new forum…


Yes, the plugins is the biggest part of hosting…I have a request in to them to make sure if they can support what we currently have…

An alternative would be to switch from AWS to DigitalOcean, who work with Discourse people and have a “turn-key” VM that might work a bit better.

Also going to ask about the export. From what I read, we can still export to an S3 bucket, so the “sanitized” exports might be something that’s pulled from there…