It looks like my community is generating data roughly at a pace of 100MB’s per month, maybe a little more. With linear projection, my backup size in December would be 1.2GB and in December 2026 around 12GB. The images seem to be ~50% of the whole package. Thats a lot of data. The previous decade (2…

Make sure you tell backup to exclude downloads, if you want smaller backups. It really depends how many uploaded images there are. Images are huge, relative to text and db data.

Yes, S3 migration & exclusion of attachments in backups is on my development roadmap. That will save 40..50%, but the backup will still be huge in a couple of years. How big is yours here at Meta, or how do you manage backups of your commercial customers with a lot of data?

Interesting comment - never thought of that and indeed the default value for email logs is 365d. The backup however does not contain any log files, only sql dump and the uploads folder. Would altering this setting reduce the size of the .sql? Any reason why would I want to log a full year of email…

I have also a large backup files (3MB per user!? already compressed) so I was wondering where this comes from. I investigated some tables I stumbled over a huge amount of entries for email_reject_auto_generated and email_reject_empty in the table. Try: psql -d discourse SELECT count(email_type) as …

Aha, is this something we need to improve @zogstrip ?

I’m not sure, I think it could be related to the fact that I had some emailing issues at the beginning where potentially several meaningless email things are stored or something. So if nothing goes wrong with emailing this 1year setting might be fine. With my rough understanding of discourse: why is…

Looking at some random examples, it seems like post_timings takes up 30-50% of the database space… Maybe it’s a good idea to introduce a ‘minimal backup’ which doesn’t include tables like post_timings, post_search_data etcetera. Even posts.cooked could be regenerated if you’d really want to keep th…

Wow, cleaning this data up reduced my backups from 600MB to under 40MB … but again: I had an email configuration problem several months ago where basically my email provider and discourse played ping pong and since that is disabled all is fine but I didn’t know that discourse stores this (that long)…

I have to repeat my question: The default setting is to keep email logs for 1 year (365d). To me that sounds an awful long time, but since Discourse generally comes with sane defaults, I have to ask why would I keep them for so long? If it consumes significant backup space, I would reduce it to 31…

How big is your Discourse backup?

Support

karussell (Peter) 2016 年4 月 22 日 19:28 4

Could this be related to some big logs? E.g. could it help to reduce delete_email_logs_after_days?

话题		回复	浏览量
Recommended backup process for very large forum? Support	6	1452	2017 年7 月 11 日
Is there any reason why daily backups would see a reduction in size? Support	25	1709	2019 年7 月 23 日
Backup Growth rate Support	8	1005	2019 年6 月 2 日
What is the size of your Discourse instance? Self-hosting hosting	8	3558	2022 年1 月 13 日
How migrate from Discourse to another platform of community? Migration	21	2192	2023 年5 月 23 日

How big is your Discourse backup?

相关话题