What is the size of your Discourse instance?

Our community (Tappara.co, in Finnish) is less than one year old.

  • We have about 58k posts by 1000+ users. Images and backups are pushed to Amazon S3.
  • Our backup size today is about 217MB and growing at a rate of 0.5MB - pace will pick up as the season starts in two weeks.

According to OVH’s dashboard, our instance is already using about 46GB of data, which seems quite a lot. I perform /launcher and apt cleanup jobs regularly.

I am getting slightly worried by the storage footprint of our forum, as we are in this for the long run. What are your numbers in terms of posts, users and storage consumed?

My forum is much smaller than this, so I’m not sure my experience is relevant here, however: how many backups do you have stored? If one backup is 217 MB is it possible that 46 GB result from (too) many old backups?

I was just looking at my backups during your reply, and noticed that Discourse (1.6 stable) was not respecting my setting for the number of locally stored backups (7). There were about a dozen of unwanted backups locally stored, but that only accounts to 2GB or so.

Ping @sam

No repro – I’ve never seen that, so likely an artifact of setup or migration.

How much storage does Meta consume?

I have now re-set the value for the number of stored backus. Let’s see if it starts behaving as expected. Rather big bug, since if unnoticed this can cause massive data storage bloat (daily backups).

Try running ncdu / on your server and diving down to see which folders are taking up the disk space.

Strange. On this one instance, Discourse still does not delete the automatically created backups according to my settings (7 to keep). I have been resetting the setting with no effect.

Any ideas how to debug?

안녕하세요, 해당 유형의 문제를 확인 중인 분들을 위해 공유드립니다.
저의 경우, 오래된 discourse 도커 이미지들이 너무 많이 쌓여 있어서 발생한 문제였습니다.
아래 명령어를 실행한 결과 10GB 이상의 공간을 확보할 수 있었습니다.

docker image prune

공식 문서에 따라 다음 명령어를 사용하여 클리닝을 수행할 수 있어야 합니다.

/var/discourse/launcher cleanup