Thanks for the feedback from everyone. I’ve added a section on our docs site for setting up Litestream with a static LiteFS cluster:
Can you please provide what to do for those of us with a dynamic primary. Are people supposed to not backup their db anymore or never use promote: true?
There’s not currently a good approach for a continuous, streaming backup for dynamic primaries. The best option right now is to do periodic backups from one of your nodes in the cluster.
Does the Go API expose the export commands, so the app that is using LiteFS can call it?
Thinking maybe a cron could be setup that is managed by my app, to have it back itself up on a nightly interval or something.
The updated docs make it clear, thank you @benbjohnson
I will really miss LiteFS Cloud, outside of the one blip I had, I have used it since launch without any issues. I understand the decision around sunsetting, but I found it innovative and amazing. Totally get that the addressable market size is just too small.
I am a convert though for using SQLite as my DB. It has been a strange, and amazing experience, and I would have a hard time going back to a central DB after this.
All the litefs
CLI commands are wrappers around the litefs/http.Client
. For example, you can find the Client.Export()
here.
If you’re feeling brave you could hook into the Client.Stream()
code here and that’ll give you a continuous stream of LTX updates that you can compact together to build your own incremental backup system. Much of the LiteFS Cloud functionality relies on the compaction code in the ltx repo.
This topic was automatically closed 7 days after the last reply. New replies are no longer allowed.