SQLite: Backup, Restore and scale out proposal #2292
Replies: 7 comments
-
@gedw99 for your case I think mysql or postgress is more apropriate |
Beta Was this translation helpful? Give feedback.
-
@6543 did you look into litestream ? i have used stolen with postresql and it's very complicated to scale out. There is alot to getting a DB working in day 2 ops. |
Beta Was this translation helpful? Give feedback.
-
If it can be run as a external process to woodpecker, what is needed to change to woodpecker, besides adding documentation ? |
Beta Was this translation helpful? Give feedback.
-
SQLite soppprt I guess ? your not using Ent ? |
Beta Was this translation helpful? Give feedback.
-
We already do have sqlite support Whats ent? |
Beta Was this translation helpful? Give feedback.
-
For litestream i think we wont integrate it directly |
Beta Was this translation helpful? Give feedback.
-
I think we can close this? |
Beta Was this translation helpful? Give feedback.
-
Clear and concise description of the problem
The only durable state for woodpecker is the Database.
Sqlite is great for its ease of deployment but not so create if you need some HA aspects such as backup, restore and scale out to multi servers.
Suggested solution
https://github.com/benbjohnson/litestream is a golang lib that continuously backsup sqlite to s3, and can restore from it.
It can be imported into woodpecker OR it can run alongside it.
There is no changes to the way the data layer works in woodpecker.
We could add to the WoodPecker CLI to allow making calls to Litestream to do restores.
Also scale out with litestream is being worked on and also being used by some. So you could then have one READ /WRITE Woodpecker with embedded Litestream and many READONLY Woodpeckers with litestream. It would mean that multi HA would be an automatic thing for Woodpecker.
If you interested i can show examples etc.
Alternative
No response
Additional context
No response
Validations
Beta Was this translation helpful? Give feedback.
All reactions